1 U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES PUBLIC HEALTH SERVICE FOOD AND DRUG ADMINISTRATION CENTER FOR BIOLOGICS EVALUATION AND RESEARCH INTERNATIONAL ASSOCIATION FOR BIOLOGICALS NATIONAL INSTITUTE OF ALLERGY AND INFECTIOUS DISEASES NATIONAL VACCINE PROGRAM OFFICE WORLD HEALTH ORGANIZATION + + + + + EVOLVING SCIENTIFIC AND REGULATORY PERSPECTIVES ON CELL SUBSTRATES FOR VACCINE DEVELOPMENT + + + + + WORKSHOP + + + + + Thursday, September 9, 1999 + + + + + The workshop took place in the Plaza Ballroom, Doubletree Hotel, 1750 Rockville Pike, Rockville, Maryland, at 8:00 a.m., Naomi Rosenberg, Ph.D., Kathryn Zoon, Ph.D., and Keith Peden, Ph.D., Session Chairs, presiding. 2 PRESENT: KATHRYN ZOON, Ph.D., Session Chair NAOMI ROSENBERG, Ph.D., Session Chair KEITH PEDEN, Ph.D., Session Chair JOHANNES LOEWER, M.D., Panel Chair STEPHEN HUGHES, M.D., Panel Chair RON DESROSIERS, Ph.D., Speaker BEN BERKHOUT, Ph.D., Speaker HSING-JIEN KUNG, Ph.D., Speaker MAXINE LINIAL, Ph.D., Speaker JOHN KAPPES, Ph.D., Speaker JOHN PETRICCIANI, M.D., Speaker DONALD BLAIR, Ph.D., Speaker RUTH RUPRECHT, M.D., Ph.D., Speaker EUGENE MAJOR, Ph.D., Speaker HAIG KAZAZIAN, M.D., Speaker BERNARD MEIGNIER, Speaker GIRISH VYAS, Ph.D., Speaker ALEX VAN DER EB, Ph.D., Speaker THOMAS BROKER, Ph.D., Speaker PAUL, SANDSTROM, Ph.D., Speaker JAMES McDOUGALL, Ph.D., Speaker BRIAN VAN TINE, M.D., Ph.D., Student Speaker SANDRA RUSCETTI, Ph.D., Panelist 3 PRESENT (Continued): CLIVE PATIENCE, Ph.D., Panelist LEONARD EVANS, Ph.D., Panelist DAMIAN PURCELL, Ph.D., Panelist ARIFA KHAN, Ph.D., Late Breaker 4 C-O-N-T-E-N-T-S Retroviral Recombination as an Adjunct to Viral Propagation and Pathogenesis (HIV Model), Ron Desrosiers, Ph.D. . . . . . . . . . . . . . . . . .6 Reactivation of Replication-Defective (HIV) Retroviruses, Ben Berkhout, Ph.D. . . . . . . . . 18 Retrovirus Insertion into Herpes Viruses, Hsing-Jien Kung, Ph.D. . . . . . . . . . . . . . 33 Foamy Virus Replication: Implications for Transfer of Cellular Sequences, Maxine Linial, Ph.D. . . . 57 Recombination in Lentivirus Vectors, John Kappes, Ph.D. . . . . . . . . . . . . . . . . . . 76 Panel Discussion of Viral-Viral and Viral- Cellular Interactions . . . . . . . . . . . . . . 91 Designer Cell Substrate, Telomerase Activity and Cell Immortalization, James McDougall, M.D. . . .137 Introduction to Cell DNA Issues, John Petricciani, M.D. . . . . . . . . . . . . . . . .151 Evaluation of the Transforming Activity and Tumor Inducing Capacity of Tumor Cell DNA, Donald Blair, Ph.D. . . . . . . . . . . . . . . . . . .166 Infectivity of Retroviral Provirus DNA, Ruth Ruprecht, M.D., Ph.D. . . . . . . . . . . .185 Cells and Their DNA Containing Integrated DNA Viral Genomes: Differentiated Phenotypes of the Human Central Nervous System, Eugene Major, Ph.D. . . .197 Mobile Elements in Mammalian Genomes and Their Implications for Cell Substrate Safety, Haig Kazazian, M.D. . . . . . . . . . . . . . . . . .217 Panel Discussion of Cellular DNA as a Potential Source of Oncogenic Activity or Infectious Agents . . . . . . . . . . . . . . . . . . . . .234 Introduction to Evening Session, Keith Peden, Ph.D. . . . . . . . . . . . . . . . . . .289 5 C-O-N-T-E-N-T-S (Continued) Industrial Experience with Life Polio Vaccine Prepared on the VERO Continuous Cell Line, Bernard Meignier . . . . . . . . . . . . . . . .294 Proteins of Replication-Incompetent Virions for HIV Vaccination, Girish Vyas, Ph.D. . . . . .313 Unusual Response to Apoptin of Diploid Fibroblasts From Cancer Prone Syndromes, Alex van der Eb, Ph.D. . . . . . . . . . . . . . . . . . . . . . .327 Human Papillomaviruses: a Window into Eucaryotic Cellular DNA Replication Mechanisms and Regulation, Thomas Broker, Ph.D. . . . . . . . . . . . . . .341 Facilitated Detection of Adventitious Agents Using Genetically Engineered Cell Lines, Paul Sandstrom, Ph.D. . . . . . . . . . . . . . . . .352 In Situ Transcriptional Analysis of Integrated Viral DNA, Brian Van Tine, M.D. Ph.D./Student . .368 Late Breakers: Arifa Khan, Ph.D. . . . . . . . . . . . . .376 Damian Purcell, Ph.D. . . . . . . . . . . .387 Johannes Loewer, J.D. . . . . . . . . . . .395 6 1 P-R-O-C-E-E-D-I-N-G-S 2 (8:03 a.m.) 3 CHAIRPERSON ROSENBERG: Good morning. 4 Please take your seats. We would like to start the 5 session so that we try to stay on time. 6 This morning's session is going to 7 continue. I think you'll hear many similar issues to 8 those that we heard addressed yesterday afternoon. So 9 we'll continue discussing retroviruses. 10 However, this morning we'll move more 11 toward the complex retroviruses with presentations 12 addressing issues relating to lentiviruses. 13 We'll start with our first speaker, Ron 14 Desrosiers. 15 DR. DESROSIERS: Thank you very much. 16 My laboratory -- excuse me -- my 17 laboratory uses simian immuno-deficiency virus, SIV, 18 infection of rhesus monkeys as an animal model for 19 AIDS, and I'm not exactly sure why I'm here, but I 20 think I'm here to -- 21 (Laughter.) 22 DR. DESROSIERS: -- I think I'm here to 23 give some examples of retroviral recombination from my 24 experiments or where there may have been opportunities 25 for retroviral recombination in our SIV system and 7 1 where we have not seen it. 2 So I only have three slides in the 3 carousel, so I'll try to give some explanation of each 4 of these three, and I should finish well ahead of 5 time. 6 One of the things we've been doing in this 7 SIV system is knocking out the so-called auxiliary 8 genes. These are genes that are not absolutely 9 required for the viruses' ability to replicate, but do 10 contribute to the virus' ability to replicate in 11 monkeys and to cause disease. 12 So we've done such experiments with the 13 nef gene, the vpr gene, the vpx gene, and the vif 14 gene. 15 When we introduced a premature stop codon 16 into the nef reading frame to eliminate the ability of 17 nef to be expressed and used that nef stop virus to 18 infect rhesus monkeys, there was very quickly -- by 19 two weeks after infection of rhesus monkeys, all we 20 could detect was revertant virus, that is the 21 premature taa stop codon had universally reverted by 22 two weeks of monkey infection to a stop codon, 23 indicating that there was strong selective pressure 24 for open functional forms of nef in vivo, and that 25 such revertant viruses, although they might be rare, 8 1 appear rarely as a rare event. The strong selective 2 advantage of such a virus can very quickly result in 3 that virus being the vastly predominant virus. 4 We've also made forms of virus with a 5 large, gaping deletion in the nef gene, and infection 6 of monkeys with such nef deletion virus shows that 7 this nef deleted virus is clearly an attenuated virus. 8 It replicates much less than the parental virus, and 9 is much less prone to inducing disease. 10 So one of the things that we've done is to 11 look very hard to see whether or in which ways a 12 virus, such nef minus virus, could regain pathogenic 13 potential. 14 We've looked both in cultured cells 15 producing this nef deletion virus and in rhesus 16 monkeys infected with nef deletion virus for possible 17 recombinant viruses. 18 In discussing use of such strains as 19 experimental vaccines, the question has been raised 20 whether such viruses could possibly acquire cellular 21 sequences in an analogous way, the way the Type C 22 retroviruses can acquire oncogenes or other such genes 23 to bring back nef function or to provide some 24 additional oomph to the virus to replace the loss of 25 nef. 9 1 And despite extensive efforts looking for 2 such recombinant viruses, where there would be strong 3 -- so in the absence of nef, there would be selective 4 pressure for capture of any gene that created a 5 selective advantage, and we should be able to see it 6 in vivo because of the strong selective advantage that 7 could impart. 8 And despite extensive efforts looking for 9 such viruses, we've never found such recombinant 10 viruses that have picked up extraneous sequences. 11 We've taken animals infected with nef 12 deletion virus, and we've serially passaged virus in 13 rhesus monkeys, and we have been able to increase the 14 pathogenic potential of such nef deleted SIV by serial 15 passage in monkeys, and the increased pathogenic 16 potential is not due, at least in the two lineages 17 we've looked at, is not due in either case to capture 18 of cellular sequences or extraneous sequences, but is 19 actually due to compensatory changes in SIV elsewhere 20 in the genome that allow the virus to make up for the 21 loss of the nef gene. 22 So similarly using nef minus virus, vpr 23 minus virus, vpx minus virus, and vif minus virus, 24 we've not seen any examples, despite extensive efforts 25 looking for it; we've not seen a single example where 10 1 the virus has captured an extraneous sequence and made 2 it increase selective advantage. 3 Now, a few years ago we published a paper 4 where we attempted to directly look for -- we designed 5 an experiment to see if we could directly demonstrate 6 retroviral recombination in an infected monkey. So 7 this is now -- we're not talking about recombination 8 with cellular sequences, but recombination between two 9 different SIV strains. 10 So what we did -- how do I go backwards 11 here? -- so we used two different SIV strains for this 12 experiment. We used one strain with the deletion in 13 nef, and the second strain was deleted in both vpx and 14 vpr. That is about four kilobase pairs away. 15 So we inoculated the nef deleted virus in 16 one leg, and we inoculated the vpr/vpx virus in the 17 other leg. Both of these are attenuated viruses. 18 By two weeks after infection, as reported 19 in that Journal of Virology paper, by two weeks after 20 infection all we could detect is wild type virus 21 resulting from a recombination between these two 22 different strains. 23 For recombination to occur, one needs 24 infection of the two different retroviral strains, 25 infection of the same cell, and one can easily achieve 11 1 that in cell culture, but there had been some question 2 how likely an event that would be in a whole infected 3 organism with the massive pool of cells, and obviously 4 at least initially a very, very low multiplicity of 5 infection. 6 For such recombination to occur, as I just 7 said, you need infection of a single cell with both 8 strains, and although at least theoretically there are 9 a number of or several different mechanisms by which 10 retroviral recombination can occur, the major way such 11 recombination occurs relates to the fact that virions 12 contain two RNA molecules, and when you have two RNA 13 molecules in a single virion of the two different 14 genotypes, when a cell is newly infected through the 15 reverse transcription process and the copy choice 16 mechanism in the reverse transcription process, one 17 can then get recombinant viruses, in this case 18 representing a wild type sequence and fully 19 pathogenic. 20 So these experiments illustrated two 21 things. They illustrated the ease with which 22 recombination can occur even in vivo in the large pool 23 of infected cells in the rhesus monkey. 24 And the second thing that they illustrated 25 was the enormous power of selective advantage and how 12 1 selective forces can very quickly selected for strains 2 of virus with a growth advantage. 3 And the final thing I want to talk about 4 is a virus deleted in the vif gene. So although vif 5 is generally considered an auxiliary, nonessential 6 gene, which it certainly is with HIV for growth in 7 some cell types, it turns out that vif in most cell 8 types for both SIV and HIV is very important for the 9 virus' ability to replicate. 10 In fact, for the strain of SIV that we 11 work with, the only way we've been able to grow a vif 12 minus virus is in a vif complementing cell line. 13 So this shows the growth curves of vif 14 deleted SIV, which has been produced in a vif 15 complementing cell line. Growth of this vif minus SIV 16 in the vif complementing cell line, viv CEMX compared 17 to growth of the same virus stock, same amount in the 18 parental line CEMX 174. 19 Now, the vif complementing cell line 20 contains the full vif open reading frame, and the vif 21 minus the deletion in the vif gene in the virus that 22 we used is within the vif reading frame. So it is at 23 least theoretically possible that there could be 24 recombination in this complementing cell line between 25 the vif minus viral sequences and the vif gene 13 1 present, integrated in the host cell genome used for 2 growing the virus. 3 And obviously if there was a recombination 4 for vif-plus virus, that virus would have an enormous 5 selective advantage both in cell culture and in 6 monkeys, and we could quickly, quickly see it. 7 So we've made a number of vif-minus virus 8 stocks and vif complementing cell lines. We've 9 infected at least six different monkeys with vif 10 deleted SIV grown in vif complementing cell lines, and 11 we've not seen a single example of virus regaining the 12 vif sequences from the cell line that it was grown in. 13 That's basically the sum total of my 14 experiences involving recombination either with viral 15 sequences or cellular sequences in the SIV system. 16 So I'll stop there and take questions. 17 (Applause.) 18 DR. KUNG: Hsing-Jien Kung, UC-Davis. 19 Hi, Ron. 20 DR. DESROSIERS: Hi, Hsing-Jien. 21 DR. KUNG: So much of the -- in terms of 22 detection of the recombinants, usually you need a 23 selection. So I was asking you in the nif experiment, 24 did you grow the virus on just wild type non- 25 complementing virus in an effort to pick up -- to 14 1 actually pinpoint the selection of that virus? 2 DR. DESROSIERS: Well, the -- 3 DR. KUNG: The vif, the last experiment 4 you were talking about. You said you didn't see the 5 pick-up of the host gene. 6 DR. DESROSIERS: That's correct. So vif 7 minus virus does not grow in CEMX 174 cells or other 8 cell lines, but the vif plus virus does. So we've 9 transfected vif minus SIV DNA into the vif 10 complementing cell line to make virus stocks. We've 11 expanded those virus stocks in the vif complementing 12 cell line. 13 If the vif minus virus had picked up a vif 14 gene, it would now grow in a non-complementing cell 15 line, and we've repeatedly tested our virus stocks, 16 undiluted, large amounts. We've taken undiluted virus 17 stocks and put it into six different monkeys, two 18 different stocks prepared independently. We've never 19 seen -- there there was clearly the enormous selective 20 advantage of minute amount of vif plus virus would 21 quickly grow out, and we've not seen it in the 22 examples we've done. 23 MR. BROKER: Tom Broker, UAB. 24 In that last experiment, how much flanking 25 homologous sequence might you have had, and can you 15 1 consider expanding that outward to ask whether you'd 2 favor recombination with enough flanking homologous 3 things allowing crossover? 4 DR. DESROSIERS: Yeah, I don't remember 5 off the top of my head exactly how much, but it was 6 not a lot. It was maybe 100 base pairs or so flanking 7 on each side, but you're right. I think if one wanted 8 to investigate it, I think that would be a pretty good 9 system to do it. 10 MR. BROKER: It would be great. 11 DR. DESROSIERS: could expand out further 12 and further and see under what conditions one might 13 see it. 14 MR. BROKER: Okay. 15 DR. DESROSIERS: I think in the experiment 16 when there was recombination between two retroviruses, 17 I've always kind of hoped that people would have used 18 that system a little bit more, too. One could 19 actually -- we never went to the trouble to try to 20 look. One could actually use that system to look at 21 where the recombination events were occurring by 22 putting in third base changes. You could actually map 23 where the crossover, quote, crossover events occurred, 24 and we never took the trouble to do that. 25 But I think anyone who's interested in 16 1 studying recombination in retroviruses, that would be 2 a pretty nice way to do it, but, no, we've not done 3 that experiment. 4 It was a short stretch of sequence. We 5 were actually -- for the experiments we wanted to do, 6 we wanted to avoid getting recombinant virus, and so 7 we kept the length of overlap short, but if one wanted 8 to study it, exactly right. You could just expand the 9 length and see if you had a much larger length of 10 flanking sequences, whether you had increased chances 11 of seeing recombination 12 PARTICIPANT: It's really an extension of 13 the same experiment. We know that retroviruses will 14 promiscuously package all sorts of cellular messages. 15 Do you know within the virion population that's coming 16 out of the complementing line whether you actually do 17 get vif message actually packaged in the virions that 18 might be a substrate eventually for a strand jump 19 recombination? 20 DR. DESROSIERS: Yeah, that's a -- we 21 haven't actually looked at it. My mind is turning in 22 theoretical terms, what that means. It's generally 23 believed that vif minus virus in a noncomplementing 24 cell line is capable of a single round of infection 25 events. 17 1 PARTICIPANT: Sure. 2 DR. DESROSIERS: It can get into one 3 round, but then not pass. I think there's really 4 only room in the virion -- some of our retroviral 5 experts here can correct me if I'm wrong -- but 6 there's really only room in the virions for two RNAs, 7 and I think if a virus did package cellular RNA and 8 that would leave room at most for a single viral RNA. 9 I think that would result in a defective particle, 10 would not be infectious, but still could be propagated 11 as a defective virus conceivably anyway in the 12 presence of a helper. 13 CHAIRPERSON ROSENBERG: I need to remind 14 the questioner to give their names. 15 DR. LINIAL: Maxine Linial, Hutchinson 16 Center. 17 About your second point, we have shown 18 with ALV that you can package a full length genome and 19 a cellular RNA and get recombination. So it can 20 occur. 21 But my main question was: how much 22 replication did you get with the delta vif virus, and 23 if you did see any, in what kind of cells was it 24 replicating in the monkeys? 25 DR. DESROSIERS: The vif minus virus? 18 1 DR. LINIAL: Un-huh. 2 DR. DESROSIERS: The only way we were able 3 to replicate the vif minus SIV is in the vif 4 complementing cell line. So the vif minus virus, 5 we've tested it. We can clearly generate high titred 6 stocks in the complementing cell line, and we've used 7 that to infect a variety of cell lines, primary rhesus 8 PBMC, primary macrophage cultures, and inoculated into 9 monkeys. 10 And in cell culture, the only way -- the 11 only system we detect any replication is in the vif 12 complementing cell line. 13 In monkeys, the vif minus virus is at the 14 very most -- the very least the most attenuated strain 15 we've seen. The only thing we've been able to measure 16 in the delta vif inoculated monkeys is very, very weak 17 antibody responses. 18 CHAIRPERSON ROSENBERG: Thank you. 19 The next speaker is Ben Berkhout. 20 DR. BERKHOUT: Okay. I would like to 21 continue on some of the themes that have been 22 introduced by the previous speaker, and I guess we can 23 quickly go over the introduction. 24 Of course, most of you should be familiar 25 with the idea of live attenuated viruses. These 19 1 entities should replicate and elicit supposed immune 2 response that confers lifelong protection against 3 challenge with a pathogenic wild type virus. 4 Of course, the concept has proven to work 5 in several systems, for instance, vaccinia, polio, and 6 measles, and for HIV-1 good results have been 7 described in SIV infection models of macaques by 8 several labs, including that of the previous speaker, 9 and it has been proposed recently by some groups that 10 one should start clinical trials on humans with the 11 HIV-1 versions of live attenuated viruses with 12 multiple gene deletions. 13 But there are still several safety 14 concerns. One of them is the induction of a fulminant 15 infection in immunocompromised hosts, and second, and 16 that's the one I would like to focus on today, is the 17 potential reversion of an attenuated vaccine strain to 18 a virus variant that can replicate fast and that 19 potentially can cause AIDS. 20 Now, all of the studies I will present are 21 done in tissue culture settings. So we will only look 22 at replication potential of viruses, and therefore, of 23 course, we don't have any measurements of 24 pathogenicity. 25 Now, let me start out with a purely 20 1 hypothetical graph. The idea here is to delete genes. 2 As you know, HIV-1 has nine genes; simple 3 retroviruses, only has three. So the idea is that we 4 can probably removed a couple of the accessory genes 5 and thereby we will credibly decrease the replication 6 potential of such a virus. 7 Probably that also has a consequence for 8 the pathogenicity of this virus, and I did draw 9 parallel curves. There's no evidence for that, but I 10 think it's reasonable for this graph at least. 11 Probably also the immunogenicity of these 12 viruses will slowly drop off, and the idea is to reach 13 a level that is below a threshold for the 14 pathogenicity such that this virus will not cause 15 AIDS. 16 However, of course, replication should 17 still suffice to induce a good immunogenicity, and of 18 course, these thresholds are, again, hypothetical, and 19 we don't know how big this window is. 20 Even if you managed to identify such a 21 virus that has these properties, one of the key 22 questions is how stable is this virus and will it 23 perhaps not be able to regain replication capacity and 24 thereby going over this threshold. 25 Now, why should we release genes? We 21 1 already heard in the previous talk that when you put 2 stop codons in, for instance, the nef gene, it's 3 repaired in two weeks. Even small deletions or 4 substitutions will be removed by this virus very 5 rapidly, and that was shown early on in a paper 6 published in '95 using the SIV system and virus 7 evolution in macaques. 8 A four amino acid deletion was introduced 9 in nef, and that's shown here. It was still present 10 two weeks after inoculation. However, after 17 weeks, 11 the virus managed to repair this sequence, at least to 12 fill in this gap, by duplicating the sequences that 13 are located over here. 14 Officially at this point in time, you do 15 not have the wild type sequence back, but credibly 16 after 25 weeks, we do see a wild type tyrosine appear 17 at this position. It's like the wild type, and after 18 45 weeks, we also have an aspartic acid here, and we 19 are almost back to the wild type sequence 20 demonstrating the enormous repair capacity of this 21 virus. 22 So small deletions will not suffice and 23 we, therefore, have to make bigger deletions. 24 Now, the system we are going to us is a 25 tissue culture evolution system, and I will use one 22 1 slide to introduce the system. The method we have 2 called forced evolution. Basically we start out with 3 a molecular clone of HIV-1 in which the deletions in 4 accessory gene products have been introduced. 5 We start with a massive transfection. Ten 6 to 40 microgram of molecular clone is electroparated 7 (phonetic) into cells, and then initially the cells 8 are cultured, but when there is a sign of virus 9 replication, it's seen in either syncytia or GAG E24 10 production. We start to passage the culture 11 supernatant onto fresh and infected cells. 12 Initially these large inocula up to 1 mL, 13 later on credibly reducing the amount of virus that is 14 transferred. 15 So in Section A we will be able to pick up 16 revertants, and I guess it's important to realize that 17 this reversion or evolution is a two-step process. 18 One needs mutations, and those will be introduced 19 primarily by the first transcriptase enzyme, which is 20 error prone. Of course, these mutations will be 21 introduced in a random manner, and then will get 22 selection of those virions that are better than the 23 input virus. 24 So its evolution, as I guess it must be 25 proposed by Darwin, although he used different terms. 23 1 It's variation and survival of the fittest. 2 Now, because I think a presentation may 3 not be complete without seeing an RNA stem loop 4 structure, I actually show you one example of a study 5 that we performed on a stem loop structure. That's 6 the so-called TAR hairpin that is found at the extreme 7 five prime and three prime end of HIV-1 genomes. This 8 is the wild type situation. 9 This is a mutant that is completely 10 replication defective. Here we have open to the lower 11 left-hand side of the stem, and after about six months 12 of culturing initially the cells, later on the virus 13 passage, we ended up with this revertant. It has one, 14 two, three, four substitutions. It doesn't go back to 15 the wild type sequence, but what this revertant is 16 telling us is that this hairpin structure is critical. 17 Sequences are less critical, but one needs 18 a base pair lower region here, and of course, we also 19 have some idea of what this hairpin is doing. It's 20 critical in transcription regulation of this virus, 21 but also because it's present in both ends, it plays 22 a role in the mechanism of reverse transcription. 23 So we obtained a rather complete set of 24 deleted HIV-1 genomes from Ron Desrosiers, and I will 25 today only focus on one of these deletion virions, and 24 1 that's the so-called Delta 3 virions with three 2 deletions. 3 One deletion is in the accessory vpr 4 protein, and this five prime half of the genome was 5 combined with this one, which has a deletion in the 6 nef gene. Then some of the sequences here at the 7 border of the three prime LTR are still present 8 because they are important for replication. 9 And then there is a second deletion in the 10 three prime LTR. 11 Now, it's important to realize that this 12 deletion in the LTR will be inherited in the progeny, 13 also in the five prime LTR, and the five prime LTR, as 14 you probably know, is the transcription promoter at an 15 untelldish (phonetic) virus. 16 Over here I've indicated some primers that 17 will be used in subsequent PCR analysis to actually 18 check whether these introduced solutions are still 19 present. 20 One more thing. We tried to do efflution 21 (phonetic) in tissue culture in primary cells. That's 22 indicated over here, but in particular, for this Delta 23 3 variant we found that the replication, while that 24 much delayed, that it probably will take years to 25 really obtain the fertent (phonetic) viruses. 25 1 So to solve that problem, we are going to 2 use a transformed T cell line, in this case Sup. T1 T 3 cell line, because there the replication defects of 4 this kind of mutants are less feared (phonetic). 5 There certainly is a replication, delay in 6 replication, but it's much less feared than in primary 7 cells. 8 So the virus with three deletion was 9 passaged for up to four month in the Sup. T1 T cell 10 line. Samples were taken at day 14, 53, 83, 125, and 11 we analyzed the samples with similar input virus 12 amounts in a parallel infection, and as you can see 13 over here, the initial virus or the virus present 14 after two weeks has a hard time replicating, and then 15 you see a gradual increase in replication capacity and 16 a big jump actually is seen already at day 53. 17 So we were interested to find out what has 18 happened with this virus. So we performed PCR essays, 19 PCR analysis. PCR analysis across vpr gene didn't 20 show any insertion or deletion of sequences, and in 21 fact, the sequenced region, nothing has happened. 22 This is a PCR across the nef and LTR 23 deletion, and as you can see over here, something 24 happens around day 55. We see the appearance of a 25 variant that is about 14 base pairs longer, and 26 1 apparently this variant is much more fit than the 2 input virus because it out competes the input virus 3 within a couple of weeks. 4 So at this point we got excited. We 5 thought, well, perhaps there is an insertion of 6 perhaps cellular sequences in these deleted virus, and 7 of course, we sequenced this region. 8 It turned out that the two introduced 9 deletions were maintained. The nef deletion was there 10 and also the deletion in the U3 region was perfectly 11 maintained. 12 There also was no insert of cellular 13 sequences, but what has happened is that in the core 14 promoter region, and as indicated here this is a 15 transcription start (phonetic) side, three binding 16 sites for the constitutive transcription factor, SP-1, 17 and a tandem repeat for nf kappa B binding. The 18 deletion that was introduced is indicated over here. 19 Nothing happened over there. 20 What we did see is a gross duplication of 21 all the three SP-1 sites. In addition, there are 22 seven nucleotides of unknown origin, and we have seen 23 an onser (phonetic) duplications. 24 Yesterday in one of the talks, and I must 25 say that this apparently is a very popular theme in 27 1 mutant retroviruses, they easily change the number of 2 repeat elements in their genome, and in fact, we 3 probably know why that is so, because this duplication 4 is probably introduced during the first inscription by 5 a mechanism called slippage realignment. 6 So RT enzyme has probably copied the 7 sequences over here. Then RT and the cDNA is 8 partially removed and it reanneals back to the 9 original sequences, and in one step one gets a 10 duplication of three SP-1 sites. 11 So we end up with this promoter. Does 12 this make a better transcriptional promoter? 13 (Unintelligible) in transient LTR reporter assays, and 14 what we compare is the wild type LTR, the one with the 15 U-3 deletion, and then the one that's the duplication 16 of SP-1 site. So as you compare the last two bars, we 17 tested that in the absence of the transcriptional 18 activate of protein TAT and also in the presence and 19 then in cells that were not more activated by PMA and 20 PHA. 21 In the absence of TAT protein, we, in 22 fact, see that the LTR promoter does not gain function 23 by duplicating of the SP-1 site. In fact, it gets a 24 little worse. 25 However, in the presence of TAT protein 28 1 and, of course, in virus replication in this system, 2 the TAT protein will be around. We do see a partial 3 restoration of the LTR function, and in this case 4 actually it's becoming a little better than the wild 5 type. 6 So the transcription promoter is, indeed, 7 improved in the presence of TAT protein. 8 Does this single mutation also improve the 9 replication of the delta 3 virus? Well, indeed, it 10 does, and that is shown here. So we reconstructed a 11 molecular clone with three deletions, and then in 12 addition the six SP-1 sites. So here is the wild type 13 virus. 14 If you only delete vpr there is a small 15 replication. So this is the input virus with three 16 deletions, and if you in this virus then introduce the 17 six SP-1 sites, there is a rather dramatic increase in 18 replication capacity. In fact, this revertant virus 19 replicates much faster than the virus with a single 20 vpr evolution. 21 Now, we thought it was also of interest to 22 test the replication of this virus not only in the 23 Sup. T1 T cell line, in which it's evolved, but also 24 in primary cells, and that's shown here, and there we 25 do see a different picture. 29 1 This is the wild type virus replication in 2 primary cells, and over here we see that there hardly 3 is a difference in the replication of the original 4 delta 3 virus and the one with six SP-1 sites. 5 So it seems that the SP-1 site duplication 6 is beneficial only in the Sup. T1 T cell line that was 7 used for the evolution experiments. That may have to 8 do with the pool of transcription factors that are 9 present in that particular T cell line. 10 And, for instance, it has been reported 11 that Sup. T1 cells have an extremely low amount of nf 12 kappa B, and one can imagine that perhaps, therefore, 13 SP-1 is more important for this virus. 14 So just to sum up, we show that this delta 15 3 strain is genetically unstable. It retains 16 replication capacity in tissue culture, and we have 17 seen duplication of SP-1 sites. 18 Now, this is all tissue culture, but I 19 think there are some additional evidence to suggest 20 that these viruses are unstable. The Ruprecht 21 laboratory has recently demonstrated that some of the 22 viruses, the delta 3 viruses, do eventually cause AIDS 23 in infected monkeys, and there's also some evidence 24 from cohort studies in humans where people infected 25 with a nef deleted and, therefore, attenuated virus 30 1 strain do over time show a decline in CD-4 T cell 2 numbers and, therefore, perhaps they are at risk for 3 going on to AIDS. 4 So the conclusion will be that these 5 deleted viruses are safe. 6 Now, in the context of this meeting, I 7 should say that the one example I showed you today we 8 did not find any incorporation of cellular sequences 9 in these -- in these deletion virions, and like the 10 previous speaker, in all of the evolution experiments 11 that we have done so far, we have not come across any 12 insertion of cellular sequences in these virus 13 strains. 14 Now, finally, I will discuss a putative 15 route to perhaps a more safe vaccine. What I showed 16 is that this deleted and attenuated virus is able to 17 regain replication capacity, and that's indicated by 18 this arrow. 19 Now, perhaps this virus may be a nice 20 starting point to remove additional functions, thereby 21 again going down on the replication letter. Perhaps 22 then again we can try to evolve a faster replicating 23 variant, and by repeated cycles of gene deletion and 24 subsequent evolution, we perhaps may end up with a 25 virus with only three to five genes that is able to 31 1 replicate efficiently, and that perhaps will be more 2 stable in genetic terms. 3 And with that I would like to stop. 4 (Applause.) 5 CHAIRPERSON ROSENBERG: Is there a 6 question? 7 MR. FALLEAUX: Hi. Frits Falleaux. 8 This may be a silly question for 9 someone -- 10 CHAIRPERSON ROSENBERG: Be sure to 11 identify yourself clearly. 12 MR. FALLEAUX: Yeah, Frits Falleaux, 13 right. 14 This may be a silly question for someone 15 from working with adenoviruses and also HIV, but did 16 someone think of trying to make an attenuated to HIV 17 which is, in part, replication efficient and in 18 combination with, for example, a useful promoter that 19 dries (phonetic) nef so that you have partial 20 replication in the presence of the inducer, which you 21 take away after the introduction of protection? 22 DR. BERKHOUT: People have worked along 23 these lines. For instance, one has introduced the 24 constitutive CMP promoter in the context of a 25 replicating HIV-1 virus. 32 1 One of the problems with introducing 2 exogenous sequences in this final genome is that in 3 most cases the virus doesn't accept them, and they are 4 over time kicked out of the genome. 5 So the only way to safely introduce 6 exogenous sequences in this final genome will be to 7 introduce elements that are absolutely required for 8 replication, and then I think, indeed, it would be of 9 interest to study better and inducible promoters, but 10 that hasn't been tested so far. 11 DR. EVANS: Leonard Evans, Rocky Mountain 12 labs. 13 Did you try to put that promoter, the 14 duplicate promoter back into the wild type? 15 DR. BERKHOUT: We did, and in 16 straightforward replication curves, there is no 17 difference in replication. If you do fairly sensitive 18 competition assays so the wild type virus, and the 19 wild type has six SP-1 sites, you do see that in the 20 wild type context, in fact, the fitness goes down a 21 little bit. So -- 22 DR. EVANS: It goes down? 23 DR. BERKHOUT: Yes, it does down a little 24 bit. 25 DR. EVANS: Okay. 33 1 DR. BERKHOUT: But that difference is 2 marginal. So you have to do very sensitive 3 competition assays with the two viruses to find out. 4 So six SP-1 sites are clearly beneficial in the 5 context of this deletion variant, but they don't make 6 wild type virus much better. 7 DR. EVANS: Okay. Thanks. 8 CHAIRPERSON ROSENBERG: Thank you. 9 The next speaker is Hsing-Jien Kung. 10 DR. KUNG: Thank you. 11 So in today's talks and yesterday's, you 12 have heard that retrovirus have significant potential 13 to recombine with another retrovirus or recombine with 14 hos genes under selective pressure. Today I'd like to 15 discuss with you about experiments that I would 16 summarize as experiments regarding recombination 17 between retroviruses and the herpes viruses. 18 If I can have the first slide. 19 Let's consider the following scenario. In 20 co-infected cells with the retrovirus and the herpes 21 virus, retrovirus copying to RNA, then into DNA, RNA 22 copying to DNA, and has a choice of going into the 23 cell or the genome or has the choice to go into the 24 herpes viral genome if in the co-infecting cells. 25 We can argue what is the probability. 34 1 What is the chance? Well, based on the most 2 simplistic view, the simple calculation, simply based 3 on the mass ratio, you would argue that if the herpes 4 virus -- and let's take the worst case scenario. It's 5 only present as one to ten copies -- one copy, let's 6 say, in the latent state. Based on mass ratio, you 7 would calculate that in every 10,000 integration into 8 the host genome, you would have one into this size of 9 herpes virus. 10 Now herpes virus sometimes will also 11 replicate. So it can reach -- so okay. Let's take 12 that calculation and say you'll have massive infected 13 cells, about a million cells, in fact, with 14 retrovirus. Then there should be 100 integrants if 15 herpes virus is in latent state. 16 Now, herpes virus, of course, also can 17 replicate, let's say, 2,000 copies of 10,000 copies. 18 If it were 10,000 copies, then every integration into 19 the host chromosome, there's a chance into the herpes 20 virus genome. 21 So if you calculate, you say this is not 22 too rare, but, on the other hand, you can make the 23 argument that herpes viral replication usually are 24 combined in certain subnuclear structure, and they may 25 not be as successful to the retrovirus integrative 35 1 complex, and they may be encapsulated very quickly, 2 and so all of these theoretical considerations may 3 also argue against genotypic mixing or recombination 4 between these two. 5 Okay. So I think only experiments will 6 tell. 7 Next slide, please. 8 So today I'm going to talk about 9 retrovirus integration into herpes virus. The system 10 under study is avian retrovirus using reticular 11 endothelial virus as a model system, REV. It's a 12 typical nonacute Type C retrovirus that cause T cell 13 lymphoma and the B lymphoma in chickens, and we show 14 before that the mechanism of oncogenesis is through 15 insertion or activation of C mixing. 16 Avian herpes virus, Marek's disease virus, 17 is very, very prevalent. Until 1970 that was the most 18 important economic loss in the poultry industry, until 19 the live attenuated vaccine was developed. 20 Now, both viruses infect T cells, infect 21 the same T cell types. The tumors are derived from 22 very similar T cell types. So co-infection did exist 23 and has been observed and reported in many papers, and 24 in fact, many of the retroviruses were isolated from 25 Marek's diseased chickens, including the chicken 36 1 syncytia virus strain of the REV and avian 2 myeloblastosis virus, some of you probably know, that 3 carries the mip gene. 4 So there is a preponderance of evidence 5 that these two viruses coexist in chickens. 6 So the implication of the study -- I will 7 show you some experimental evidence, but the 8 implication of the study is, of course, that you can 9 generate a hybrid virus with altered gene expression 10 patterns, which I will show you, and the phenotypes, 11 emergings of new pathogens, if you will, and can 12 certainly sometimes broaden the host range because 13 some of the herpes virus, indeed, can carry retroviral 14 information. 15 So with that in mind, let's first tell you 16 a little bit more about the herpes virus, this Marek's 17 disease virus. This virus, although it's 18 lymphotropic, really the structure is very much 19 similar to alpha herpes virus with the repeating 20 sequence flanking the unique, long reaching, and 21 repeating sequence flanking unique short. 22 And just in a typical alpha herpes virus, 23 you are reaching and encode mostly structural and 24 replication enzymes. In fact, this virus, now Lucy Li 25 has entire sequence and, indeed, show a strong 37 1 correlation with herpes simplex virus. 2 Unique small region also encodes some of 3 the genes, however. They are not important in the in 4 vitro replication. 5 Now, the interesting part are usually 6 confined in the more, but divergent region, in the 7 repeat region, and we now know that in the Marek's 8 disease virus that we have worked in the past few 9 years, that it encodes a protein that we can call it 10 oncogene, called the mac (phonetic), which is in the 11 Joan Foss family of loosing zipper protein, and if you 12 remove this, the virus becomes nononcogenic. 13 Okay. So in this sense the virus is 14 similar to herpes simplex virus, but it's not, due to 15 some of the coding sequences in the repeat region. 16 Okay. But for our purpose, I'd like to 17 point out this virus is extremely oncogenic, probably 18 most potent oncogenic herpes virus. It causes tumors 19 within four to six weeks in experimental animals, and 20 these tumors are of polyclonal origin. Okay. So it 21 smells like this oncogene does do something in a 22 direct way. 23 But today we'd like to talk about how this 24 virus can be used as a template for retroviral 25 insertion and what happens after that. 38 1 Just to give you a sense, this virus is 2 probably, again, the most successful live attenuated 3 virus vaccine against oncogenic virus. I'll tell you 4 a little bit about the vaccine. 5 The serotype 1 is oncogenic strand, the JM 6 MD11 and GA stand, but serotype 2 and 3 are vaccine 7 strands. They share about 70 percent homology with 8 serotype 1. However, they are not oncogenic in 9 chickens. 10 Now, there's another way of making this 11 virus attenuated virus, by taking serotype 1, okay, 12 and simply passage. It's a mysterious way, but it 13 worked every time. You passage for a long time. It 14 could be five to ten years experiments, but you would 15 get attenuated virus. 16 People still do not know exactly what 17 happened, but there is a correlation of expansion with 18 certain repeat sequence in the region. We now know 19 these attenuated viruses actually still maintain the 20 mac oncogene. So what happens most likely is that 21 these viruses do not replicate very well. Therefore, 22 they cannot induce T cell lymphoma in chickens. 23 In vitro, however, after long passage 24 viruses tend to replicate quite well. 25 Okay. So we said that moving to the real 39 1 experiment. Now, our story began with this particular 2 experiment. They were intended to study our REV -- at 3 the time it was still retrovirologist -- REV insertion 4 of the T cell lymphoma. 5 So Bob Isford took some of the T cell 6 lymphoma generated by Marek's disease virus and looked 7 for whether REV virus is present or not. 8 It turned out REV viruses are not present, 9 but there was some surprise in that in the low passage 10 of serotype 1 virus, not serotype 2 or 3, that he 11 actually could detect hybridization. This 12 hybridization was under 30 percent mismatched, but he 13 could detect hybridization against REV or TR. 14 And this suggests to us there are some 15 sequences related to REV, and we call it ALTR remnant 16 present in the present day serotype 1 viruses. So 17 this virus has never seen REV recently, but you can 18 see that they do have some sequence. 19 The fact that they show specific bands and 20 they can map in the band D and band F region, which is 21 close to the repeat sequence, convinced Bob Isford at 22 that time that these are not a fluke. These are 23 probably real homologous sequence even though it has 24 diverged significantly. 25 There are the high passage one now, 40 1 attenuated one. You can see that the sequence begin 2 to diverge because of expansion of the repeat 3 sequence. 4 However, in this, this is something else. 5 I will come back to this. This GM high virus has 6 multiple LTR related sequence. I'll come back to 7 this. 8 So then Bob Isford began to sequence of 9 them. So just to give you summary, indeed, there are 10 several patch homologies ranging from 70 to 81 11 percent, with some nucleotides ranging from 22 to 33 12 nucleotides. 13 Now, if there's only one site, you can say 14 this is very skeptical, but with all of these sites 15 clustered together in the right order and many of them 16 diverging at the junction of the retroviral genome, 17 and that suggests to us this probably is real. 18 So we took that as LTR remnants. There 19 are also some sequence related to the retrovirus, but 20 the LTR sequence was most, most prevalent. 21 And this stretch of sequence is most 22 interesting. This stretch of sequence turned out to 23 be that it's in the enhanced region of the REV LTR. 24 This sequence now is present as in the enhancer region 25 of the herpes virus alpha tif or the VP-16 equivalent 41 1 of the Marek's disease virus, and it's a T cell tropic 2 enhancer. 3 So with that we figured that in nature 4 this had happened, and let's see what happens, whether 5 in recent -- this is what we call the ancestral 6 insertion. We like to see whether in recent time 7 whether there was any evidence of recent insertion, 8 and we recall this one, the same picture. 9 This one has a multiple one, and now 10 again, this is a virus derived from low passage, 11 simply by culture them for five years. Okay? This is 12 a passage of 211, and it's completely attenuated, and 13 they can be potentially used vaccine, but this is a 14 vaccine experiment that went exactly as intended. 15 So what happens once we discover there's 16 some retroviral insertion, and these LTR hybrids under 17 stringent condition and still stick to the filter 18 paper. So we know these are more recent insertion. 19 And Dick Witter, our collaborator, 20 collected all these viruses. So he then did the 21 following smear by cloning this separately. So with 22 that he actually was able to separate these clones and 23 these LTR sequences of segregate, indicating indeed 24 they are part of the viral genome, and they are 25 genetically stable because they have been passaged for 42 1 about 200, 200 passages. 2 And so what happened is that during this 3 long term propagation you'll feed DEF, duck embryo 4 fibroblasts, the primary cells. We talked about it 5 two nights ago, that DEF, and it turns out in passage 6 about 87, that DEF they used for fuller infection by 7 Marek's disease virus turned out to contain reticular 8 endotheliosovirus. So during this evident of clone 9 mixing fraction, that retrovirus integrate that, but 10 now they become very stable. 11 Now, you would argue why do they persist 12 for so long. Well, when Dick Witter compared the 13 replication rate with the wild type, they all in vitro 14 replicate much better. We do not know whether that's 15 due to LTR insertion. However, the LTRs seem to be a 16 persister, and the LTR integrants seem to dominate the 17 culture at passage 211. 18 So this was a five year experiment ten 19 years before we began the study. We could not 20 control. We could not add any control. So Dick 21 Witter and I think that if this happened nature, on an 22 evolution or scale; if it had happened in the ten 23 years prior to our experiments, can we do a more 24 controlled experiments by doing a co-infraction 25 (phonetic), and can we make this work in five weeks or 43 1 five days? 2 So the experimental protocol turned out to 3 be very straightforward, that you mix the MDV, 4 infected DEF with REV. MDV is cell associated virus. 5 So it's actually a little bit difficult to do 6 experiment, but if you mix them together, you passage 7 them every week, and then by feeding fresh DEF, okay, 8 and then we took individual passage mass culture, 9 isolate MDV, and in fact just isolate the cell and 10 look for MDV mini chromosome imposed fiogel 11 (phonetic), and the free retroviral DNA should run out 12 of the gel, and then you probe with REV LTR to monitor 13 the kinetics of integration. 14 It turned out it's not difficult at all. 15 Okay. We have repeated this several times now. 16 Basically after, in fact, five weeks you can see a 17 little bit. This is a southern blot. So it's not a 18 most sensitive method, but you can see that retroviral 19 LTR integration increase, okay, after passage 16. 20 It's a huge amount radioactively. 21 This is just a load to show you that MDV 22 DNA were loaded at about the same amount. 23 Now, if you do TRP CPCR, you could 24 actually detect within one to two passage. Okay. So 25 the integration certainly can be very efficient, 44 1 although as I said, this is radioactive. You really 2 do not know what is the population of the integrants 3 versus no integrated one. 4 Now, this increasing intensity, of course, 5 are due to two reasons. One is that REV is still 6 present. It can infect more herpes virus, but, 7 secondly, of course, it may be the herpes virus with 8 LTR, in fact, replicated better in some fashion. 9 There is no selective pressure except in vitro 10 replication. So this may actually contain the LTR 11 integrate certain places that can enhance the 12 replication. 13 So what Dan Jones did and Rhonda Koss did 14 was -- were to actually look at the insertion sites, 15 and then something rather interesting was revealed. 16 So by looking at -- this is 17 kilobase genome, but 17 they looked at the integration sites. Integration 18 sites are tightly clustered with two insertions in the 19 GD region, but others are tightly clustered around the 20 -- close to the boundary between repeat and the unique 21 sequence. 22 Now, at Alsets (phonetic), if you look at 23 the sequence, okay, just very briefly, you found 24 actually these are the individual integration sites. 25 I do not mean to have you look at the sequence, but 45 1 look at the arrow indicating that there's no sequence 2 facility. It's a regional facility. 3 Okay. Now, at the outset we knew the 4 integration should not be totally random because we 5 were selecting viruses, replication virus. So any 6 insertion in the essential genes would disrupt its 7 ability and we may not be able to pick up. So, again, 8 this underscored the importance that if you want to 9 study some recombinant, that the selection pressure 10 turned out to be an important one. 11 So we figured that maybe these are the 12 sequence, the regions. There are no coding sequence. 13 Therefore, it integrates better, but that still cannot 14 fully account for this tight cluster because the U.S. 15 region shown by Robby Morgan and Marc Purcell can be 16 completely deleted, yet in vitro replication, and 17 still very viable. 18 So we then began to think this actually 19 also looked very much like insertion of mutagenicity 20 oncogene that we studied before, especially the RB 21 oncogene in avian erythroleukemia in terms of the 22 tight clustering. 23 So we began to entertain the hypothesis 24 maybe this can activate some of the genes near the 25 boundary. 46 1 But before I say that, before I show you 2 the data that I have, we have to isolate the virus in 3 order to do the experiment. This is a simple whole 4 cell PCR mapping insertion sites. We did not have the 5 virus yet. 6 So in the past few years Dick Witter was 7 able to isolate the virus. This was a heroic effort 8 because the retrovirus that we used had no markers, no 9 selection marker. It's a wild type retrovirus. So it 10 depends on how prevalent these integrations are in the 11 whole mass population. 12 So Dick was able to isolate several, and 13 I'm just going to talk about one, quote, RM-1. For 14 the first retrovirus, the REV MDV hybrid virus number 15 one, and this is very interesting because it contains 16 a solo LTR. No other retrovirus sequence are present, 17 and it integrates at a hot spot, the tight cluster 18 area. 19 And this shows the retrovirus duplication 20 of the MDV genome. So it's authentic integration. 21 It has a solo LTR, our first in this 22 region, and then it homogenized to the other region as 23 well. 24 Now, the most interesting is the phenotype 25 of this virus. This virus, okay, RM-1 -- these other 47 1 two are the clones -- had everything. Everything else 2 is very similar to the wild type in that it can 3 replicate very well inside bursa, inside T cells in 4 the chicken. 5 The only thing different is it has no 6 oncogenicity. Okay? So this is the viral clone that 7 did the test, in vivo test, and this is the wild type, 8 okay, the oncogenicity seven out of eight, five out of 9 eight, eight out of the eight, and this is the 10 attenuated one. Of course, after long passage it also 11 has no oncogenicity. 12 But this virus N and attenuated virus 13 differ significantly. This virus can replicate very 14 well, whereas the attenuated virus does not, do not 15 replicate very well in vivo. 16 Okay, and as a result, when Dick Witter 17 did the challenge protection experiment, so in fact, 18 it was RM-1, then challenged with the virulent, very 19 virulent MDV, it turned out this has much better 20 protectivity than the other attenuated strand. 21 And this is showing here -- maybe we could 22 just locate the protections. So this was done by 23 taking the RM clone, infect the first as a vaccine. 24 Then you challenge it with the very virulent virus, 25 and it turned out the protection is 100 percent, 48 1 whereas the current vaccine virus is 34 percent, which 2 this is the Basta (phonetic) vaccine virus, about 78 3 percent, and so on and so forth. 4 And the reason is that it replicate very 5 well inside the chicken, the only things that cannot 6 cause oncogenesis, and I'd be happy to speculate to 7 the reason why that's the case, but also because it 8 can spread very well. So it serves as very good 9 protection. 10 Okay. Now, I don't mean to say this is 11 the vaccine virus because it causes other associated 12 diseases. Other than oncogenicity, it does cause 13 thymic atrophy, like the wild type. So it cannot be 14 a vaccine yet, but this certainly gives us some clue 15 as how perhaps to manufacture better vaccine for MDV. 16 But in the context of this discussion, 17 this shows that LTR insertion can change the phenotype 18 of the virus. We look at other regions, whether 19 there's any gross change. There's no gross change, 20 but we cannot rule out point mutations. 21 So now, at the molecular level, we'd like 22 to see whether it activates anything. Indeed, this 23 RM-1 has a solo LTR integrate in there, and these are 24 the northern blot to show, indeed, there is LTR 25 insertion in here, and there is a transcript with a 49 1 link to LTR from here to here. That would transcribe 2 polysystronic message carrying soft 2 open reading for 3 US-1 and US-10. 4 Since this is the closest to the five 5 prime end, we think this product may be relevant. We 6 don't have evidence to show that this is relevant. 7 We're just beginning to do that, but I can tell you a 8 little bit about it. This is a novel sequence, unique 9 to the oncogenic herpes MDV strain virus. The 10 sequence is novel. So we do not know the function, 11 but it does share some homology with US-22 gene family 12 of this cytomegalovirus as well as HHV-6. 13 In the case of HHV-6, this particular 14 protein, it's not the same protein, but it shows some 15 similarity. It has been implicated in the 16 transactivation of HIV. 17 So we think it may be a co-factor of the 18 transcriptional factor. Indeed, recent experiments 19 show that by Hous Chan's lab that this is a 20 nucleoprotein. 21 I cannot tell you more about this simply 22 because we don't have much data about this protein 23 yet. So let me give you a conclusion. 24 So what I have shown you is that we have 25 shown retroviral insertion of herpes virus at least in 50 1 this system at several levels. We show there's 2 ancestral insertion; that avian retrovirus REV LTR 3 remnants are present in the serotype 1 MDV, indicating 4 that infection into the herpes virus in nature 5 probably happened after the divergence between the 6 vaccine virus and this virus. 7 I also show you in the acute co-infection 8 experiments REV insertion into MDV detectable as early 9 as second passage. REV insertion sites are non- 10 randomly distributed. REV LTR insertion of mutagen of 11 MDV genes demonstrated MDV with altered pathogenicity 12 in the phenotypes also generated. 13 I'd like to take this in a broader context 14 and tell you about what happened since our original 15 discovery. Now this has been repeated by several 16 laboratories, and that demonstrate that herpes 17 retrovirus insertion to herpes virus. 18 First, that our own actor (phonetic) has 19 extended REV insertion into the oncogenic strand 20 experimentally to the vaccine strand, herpes virus of 21 turkey, and we can see its insertion with no problem, 22 again within one to two passages. 23 But here there's a very interesting clone. 24 One of the clones -- I'm sorry. So this you saw 25 already, but this is REV insertion into HVT. 51 1 There's one clone here that actually 2 carries the full length of REV. In fact, we were able 3 to show this can produce virus after transfection. So 4 this is not anything specific for the oncogenic strand 5 of MDV. 6 We also showed that rav 1, rouse 7 (phonetic) associated virus, can also integrate in the 8 MDV or HVT. Okay. So, again, there is nothing 9 special about REV. 10 This is probably very significant. From 11 Japan, Hiria's group about two years ago showed that 12 this MDV virus isolate from chicken actually carry 13 endogenous REV 0. Now, this experiment did not see in 14 vitro culturing at all. So it was an in vivo isolate, 15 and REV 0 very nicely integrated MDV, again, at a hot 16 spot that we show for REV. 17 Furthermore, there are two more 18 laboratories that I actually did not update this 19 slide, show last year that the avian erythroblastosis 20 virus LTR also integrates into MDV and another REV 0- 21 like sequence integrated into MDV. So this has been 22 repeated in the four laboratories independently. 23 Recently Eric Davidson in Israel did in 24 vivo experiments to see whether in vivo recombinant 25 can be detected or not, and they used PCR knowing that 52 1 the cluster region -- so they designed PCR primary 2 that can easily detect recombinant, and they were able 3 to detect about 20 of them and were helping them 4 analyze it. 5 Now, a year ago this paper is rather 6 profound to us. It's not retrovirus insertion of 7 herpes virus, but this group found that REV can 8 integrate into fowlpox virus, and the pox virus is a 9 vaccinia virus group. 10 This is profoundly significant to us 11 because this shows retroviruses can integrate into a 12 virus that only has a cytoplasmic life cycle. They 13 may not need the DNA pk. You may not other things in 14 the nucleus to do the job, but you can. 15 And again, I emphasize this was an in vivo 16 isolate. The virus never sees in vitro culture. So 17 its recombination in vivo. 18 Then finally, George Miller's lab found in 19 EBV infected cells there is a fusion of cDNA. They 20 did an isolated virus. So we do not know the fate of 21 the virus. They do find a junction that carried both 22 retroviral LTR and EBV sequences. 23 So this may happen in other systems as 24 well, and I do not -- I have no reason to believe it's 25 a special for chicken, but I think, again, selection 53 1 is very important. If you do not have selection, this 2 kind of integration come and go. 3 I think I will stop here. Thank you very 4 much. 5 (Applause.) 6 PARTICIPANT: I have a comment and a 7 question. I think for the sake of the audience it's 8 also worth noting that there is evidence for ancestral 9 capture in a number of herpes viruses, ancestral 10 capture of cellular genes by a process involving 11 retroviruses. 12 So, for example, if you look at the 13 members of the gamma-2 herpes virus group, they -- 14 from new world primates, old world primates, and 15 humans -- those herpes viruses have genes for 16 dihydrofolate reductase, cyclin D, and a few other 17 genes, in all cases lacking introns (phonetic), the 18 lack of introns suggesting it's acquired by a process 19 involving reverse transcription, and that's likely to 20 have been acquired by a process involving co-infection 21 of cells with a retrovirus. 22 My question has to do with why do you 23 think -- it looked to me like that many of your 24 examples of capture were pieces of LTR and not whole 25 LTRs. You showed one example of a single LTR. Why do 54 1 you think there are so many -- I mean, how does that 2 happen, and why does it just have small pieces or a 3 single LTR, and is there any evidence in any cases 4 that these LTRs or LTR pieces are actually driving 5 expression of some viral gene? 6 DR. KUNG: Okay. First, I did snow in RM- 7 1. I just went through so quickly. I'm sorry. The 8 LTR is promoting insertion activate transcript. Okay? 9 Secondly, in terms of why it's LTR, I 10 think that what happens, LTI/LTR direct recombination 11 which especially in herpes virus, that you have the 12 flip-flop of the R region. So the recombination is 13 very, very acute. You have direct recombination 14 between LTI/LTR you would delete the sequence of. In 15 fact, the full length sequencing, the herpes virus 16 genome, it's not very stable. It's about ten 17 kilobase. So the virus has a tendency in my mind to 18 spit out the extra sequence, and LTR seems to be 19 harmless at least. Yeah. 20 MR. MINOR: Philip Minor from NIBSC. 21 I'm completely ignorant about 22 retroviruses. The integrations that you described 23 were all with co-infections, I guess. Is it possible 24 that you could get an integration if you were just 25 looking at an endogenous retroviral sequence that was 55 1 maybe being transcribed? Can you pick it up in that 2 way? 3 DR. KUNG: Yeah. Thank you for asking the 4 question. 5 Again, I went so quickly. There was an 6 endogenous virus REV 0 integration into the herpes 7 virus shown by the Japanese group, but also more 8 importantly is that the retrotransposon, popping -- I 9 guess later the speakers will be talking about that -- 10 it's very frequent. 11 I give you one example, not related to 12 herpes virus, but you probably know the bacula virus, 13 the autografa californa nuclear polyhedral viruses, 14 that the Freezen (phonetic) and Miller and their 15 colleagues did a beautiful study, show that the TET, 16 it's a transposal element, number four, integrating to 17 the bacula virus genome transcribed the gene and 18 contained the entire sequence. 19 So, yes, my feeling is it's -- all we're 20 talking about is selection. If you have a selection, 21 I think you will be able to detect those endogenous 22 transposable -- retrotransposable like, and I think 23 that's also related to Ron's comment. 24 Many of the herpes viruses captured, 25 especially like the KSS-3, Kaposi's sarcoma, capture 56 1 a lot of cellular homologs, and many of them are 2 internalist (phonetic) even though herpes virus are 3 known to be able to splice out intron. 4 So my feeling is that some of them may not 5 be entirely due to retrovirus, but could be due to 6 retrotransposable. 7 DR. PEDEN: Keith Peden, CBER. 8 You may have said, but does the 9 integration site have the hallmarks of genuine 10 retroviral integration, the duplications at the ends? 11 DR. KUNG: Oh, yes, yes. 12 DR. PEDEN: It does? 13 DR. KUNG: Duplication at the end, yeah. 14 MR. COFFIN: John Coffin, Tufts. 15 Just to clarify a little bit your answer 16 to the previous question, I think it was asked whether 17 the endogenous viruses could actually be moving 18 without an infection cycle, and the answer probably 19 is no, for all that we know about endogenous 20 retroviruses. You still need to get -- in order to 21 move them from one place to another, you still need to 22 -- they still need to undergo a complete replication 23 cycle. 24 DR. KUNG: Yes. 25 CHAIRPERSON ROSENBERG: I think we need to 57 1 move on because we are somewhat behind, although I 2 don't want to cut this short. 3 The next speaker is Maxine Linial. 4 DR. LINIAL: Okay. Today I'm going to 5 talk mostly about foamy viruses, but I wanted to start 6 off raising a couple of issues, some work in my lab 7 about packaging of avian retroviruses. 8 Okay. So for many years my lab has been 9 interested in defining the minimal packaging region of 10 the avian retroviruses, and this is a schematic of 11 what from the literature appears to be defined as the 12 minimal packaging regions of NLV HIV and the ALV 13 viruses, and we had previously defined a packaging 14 region of about 160 nucleotides from the five prime 15 end of the genome, which are sufficient for packaging. 16 That is, you can place this region on any heterologous 17 RNA, and that RNA will be packaged with a high 18 efficiency into retroviral particles. 19 And in fact, we find that such a small 20 region on a neo or a hygro RNA is packaged only about 21 2.7 times worse than the intact ALV genome. So we 22 believe this is a sufficient packaging region. 23 More recently we've done a series of 24 experiments on this 160 nucleotide region and have, in 25 fact, found that we can delete off the entire three 58 1 prime end, leaving an 82 nucleotide region with this 2 set computer predicted secondary structure that has 3 several stem loop regions, and by mutagenesis RNAase 4 protection and looking at a variety of viruses, we 5 know that there is one, two, three stems that are 6 important, and that possibly the region that's 7 involved in protein binding and recognition of this 8 RNA may be only a four nucleotide loop. 9 So we're getting very close to 10 understanding what the structure of the packaging 11 region looks like for this virus. 12 And one other point I want to make is that 13 retroviral packaging seems to be a hierarchy of 14 sequences. The retrovirus most avidly packages 15 sequences with its own psi regions, but then, of 16 course, we know that vector RNAs containing psi can be 17 packaged with very high efficiency, and probably after 18 that the retrovirus is packaged cellular RNAs or other 19 elemental RNAs containing psi-like sequences, as we 20 heard for MLV. VL-30 sequences have psi-like 21 sequences and are avidly packaged, but they also 22 package random cellular RNAs. 23 And work from my lab and a variety of 24 other labs has shown that such RNAs can be a player in 25 packaging cell lines, and we did a lot of work on a 59 1 quail packaging cell line which contains a single 2 integrated RSC lacking psi, and the unique thing about 3 this packaging cell line is that the provirus is 4 exceedingly active and produces about 100 times more 5 virus than any of the other packaging cell lines that 6 we looked at, and because it made so many particles, 7 we could so that what was packaged was random cellular 8 RNAs, and that we could also show that these cellular 9 RNAs could be reverse transcribed in new cells and 10 integrated into the genome with detectable frequency. 11 And we could find about 100 using neo as 12 a cellular RNA. We could find about 100 transductions 13 of the neo gene in an infection with this virus, and 14 we propose that this kind of event can occur with any 15 packaging system if it's looked for, but this kind of 16 a looking for integrations in the new cell is not 17 generally assayed in any of these packaging systems. 18 So now I'd like to turn to foamy viruses. 19 This is one of the seven genera of retroviruses, the 20 spuma retroviruses. They're a tightly knit family 21 that's fairly divergent from all the other groups of 22 retroviruses. 23 One aspect of the virus that's very 24 interesting is that unlike, for instance, the gamma 25 retroviruses which are know -- can be pathogenic in 60 1 their native host with long latency, the gamma 2 retroviruses or the alpha retroviruses are unlike the 3 lentiviruses which are known to be pathogenic and 4 accidently infected hosts but generally not in their 5 natural host. 6 The spuma viruses or foamy viruses are not 7 pathogenic in any host but spar. So that there is a 8 whole variety of these viruses, and their life long 9 infections in their host, and there's no 10 pathogenicity, and when there's an accidental 11 infection, for instance, several of these simian 12 viruses are known to infect people. 13 Again, there appears to be no 14 pathogenicity in those hosts, and this is a very 15 interesting question, is why the life style of this 16 virus is so different from those of its other 17 retroviral cousins. 18 The genome of the foamy virus is very 19 similar structurally to other retroviruses, has gag, 20 pol, and env genes. It also has several open reading 21 frames and two known products, the transactivator 22 protein called tas, which is absolutely required for 23 transcription from the LTR promoter, and uniquely to 24 this group of viruses there's a second promoter in the 25 envelope gene, the internal promoter, which seems to 61 1 be responsible for transcription of the accessory 2 genes, tas, self, and a second gene which is a spliced 3 variant from tas and bel 2 called bet (phonetic). Bet 4 is a major product of this virus. It's completely 5 dispensable in tissue culture, but is believed to play 6 some important role in vivo, although we have no idea 7 what the role of this protein is. 8 Some features of this virus that make it 9 very different than other retroviruses is, first, the 10 pol protein is not made as a gag-pol fusion. It's 11 made from its own splice pol message. This makes it 12 more similar to the hepadenoviruses that goes to the 13 retroviruses. Again, it has an internal promoter 14 unlike all the other retroviruses. 15 And a third feature is that you cannot get 16 particle egress from the cell with gag alone. You 17 must have the envelope protein. Again, we're similar 18 to the hepadenoviruses like HBV than to the 19 retroviruses. 20 Foamy viruses have been isolated from a 21 variety of species. It's extremely prevalent in cats, 22 both domestic and wild cats. Recent studies say about 23 70 percent of individuals cats are infected. It's 24 highly prevalent in many bovine flocks. 25 Recently a virus has been isolated from 62 1 horses, although the prevalence of this virus is not 2 clear at the moment, and from a variety of primate 3 species, in fact, every primate species that's been 4 studied does have a foamy virus. 5 And in some groups of primate in primate 6 centers, essentially all of the individuals infected 7 -- Jonathan Allen at Southwestern, I think, has shown 8 that all of the baboons there do have foamy virus. So 9 although there aren't so many studies in the wild, 10 these viruses have been isolated from wild animals, as 11 well. 12 There have been several isolates from 13 human, and I'd like to speak briefly about that. The 14 type species is called HFV, human foamy virus. This 15 was isolated from a nasopharyngeal carcinoma cell 16 culture from a patient from Kenya many years ago. 17 Recent studies, however, from the group at 18 CDC have clearly shown that this human virus is 19 basically a chimpanzee virus, and interestingly, it 20 clusters very closely with isolates from 21 Schweinfurthiae chimpanzees, which is the only 22 chimpanzee that's present in East Africa where the HFV 23 isolate came from, which clearly suggests that HFV was 24 in this culture because the patient might have had 25 contact with the chimp and acquired the virus through 63 1 a zoonotic infection. 2 And all the other isolates from people are 3 clearly linked to having been bitten by a monkey or a 4 chimpanzee. 5 So one interesting thing about HFV is 6 despite the lack of pathology in vivo, as far as we 7 know, in culture you get two types of infection. The 8 first and most dramatic is a cytopathic infection of 9 fibroblast, and this was what gives the foamy virus 10 their name because these cells become highly 11 multinucleic and syncytia, multi-syncytia form. 12 But there's also a second type of 13 infection, a long term, persistent infection, and this 14 is seen and our lab has found it in a variety of human 15 cell lines, T cells, erythroid cells, monocytic cells, 16 et cetera, and in these long term, persistent 17 infections you see absolutely no CPE. The cells 18 become infected. They grow perfectly normally, and 19 you would not know that they were foamy virus infected 20 without doing PCR or assaying the virus. 21 So this is clearly a problem when one 22 deals with material from primates. One needs to 23 assume that the primates are probably infected with 24 foamy virus, and although many isolations of cell 25 lines from primates lead to cytopathicity and clearly 64 1 show their foamy virus, since we know very little 2 about the cells in vivo that are actually infected, 3 it's very possible that you could have cell lines from 4 primate cultures that are infected with foamy virus 5 without CPEs. 6 This is an example. This is an indicator 7 cell line where we have LTR driving betagal in a 8 hamster fibroblast line, and when you infect these 9 cells with foamy virus, you can see that that turns on 10 the betagal expression from the tas transactivator. 11 You get highly cytopathic cultures. 12 And here's an example of a really giant 13 multi-nucleus syncytia that can occur at high 14 multiplicity. On the other hand, when we infect a 15 variety of human cell lines, these cell lines grow for 16 years. This only goes up to 40 weeks, but we've grown 17 these cells. They continually produce virus, but they 18 are never cured of the infection, and there's no CPE. 19 Here we show that in this experiment we 20 could not infect a V cell, but this is not a problem 21 of infectivity. This is a problem of viral 22 replication. In fact, now that we have vectors marked 23 with gfp and work from other laboratories as well 24 shows that there are basically no vertebrate cells 25 that are immune to foamy virus infection. Everything 65 1 from fish on upwards, and in fact, in humans no cell 2 types to be seen to be immune to foamy virus. So 3 whatever the receptor is, it's extremely widespread. 4 So the life cycle of foamy virus is very 5 similar to that of other retroviruses with a couple of 6 striking exceptions. As I mentioned or I will mention 7 again, the virus, most of the viral budding is through 8 the endoplasmic reticulum, and although some virus 9 does bud from the plasma membrane, again, in order for 10 the virus to bud, there must be glycoprotein. 11 If you have an M minus mutant, virus does 12 not bud from the cell, and this tends to be a 13 cytopathic event. If there is no envelope, the cells 14 die very rapidly. 15 The other important thing to mention is 16 that work from our lab, as well as another lab, 17 strongly suggests that reverse transcription in this 18 virus is a late step in infection, and that means that 19 the functional genome in foamy viruses is really DNA 20 rather than RNA. 21 And this is done using AZT as an inhibitor 22 and show that the stuff that is sensitive to AZT is a 23 late step rather than an early step in the life cycle 24 of other retroviruses. 25 Another point to be made is that there are 66 1 huge numbers of intracellular particles. In fact, 2 most foamy virus is intracellular. Only about one to 3 five percent of the virus buds from the cell. The 4 rest of it is cell associated, and this has suggested 5 that perhaps like hepadenoviruses, there could be some 6 type of recycling step where some of these 7 intracellular particles get back into the nucleus, and 8 I'll get to that in a moment. 9 So in terms of the HFV genome, by doing 10 very sensitive PCR and RT PCR, we found that about 25 11 percent of the particles released from cells contain 12 apparently full length, double stranded DNA. The AZT 13 experiment strongly suggests that the functional 14 genome is DNA so that even though there are a large 15 number of RNA particles, we don't believe that these 16 are infectious, and they're probably remnants of 17 abortive reverse transcription events. 18 And we've also been able to show that if 19 you extract DNA from extracellular particles, it is 20 infectious if you put it back into cells with 21 lipofectamine. 22 Despite the fact that the foamy virus 23 functional genome is DNA, what is packaged is RNA, and 24 this is RNA's protection experiment using wild type 25 human foamy virus or a deletion mutant in the pol 67 1 gene, looking at RNA's protection. 2 And I haven't shown you the controls here 3 for particle numbers, but when you look, and this is 4 RNA's protection that really only just looks at RNA in 5 the particles and not DNA. If we RNA, we don't see 6 any nucleic acid in the particle, and the same is true 7 for the pol mutant. 8 And what we found from this study when we 9 compared the amount of nucleic acid packaged in the 10 pol mutant and wild type, it's exactly the same when 11 it's normalized to the number of particles. 12 So you don't need polymerase to get genome 13 into the particle. Of course, this is completely 14 dead. So what's packaged is RNA, and DNA just occurs 15 sometime during assembly or egress from the cell. 16 We also believe there are large numbers of 17 particles in the intracellular particles that contain 18 DNA. So when you look at a copy number of foamy virus 19 persistently infected cells or acutely infected cells, 20 you can find hundreds or thousands of copies of DNA 21 per cell, and this has made it very difficult to look 22 at integrated genomes in these cells. 23 This is a schematic of what we think the 24 foamy virus looks like. Instead of having an RNA 25 genome, it has a DNA genome. 68 1 Another feature of the foamy virus which 2 I didn't mention, but which is also very striking is 3 that the gag polyprotein is not cleaved except for 4 four KD at the C terminus. So you never get 5 maturation to caps at matrix and nuclear caps as in 6 other retroviruses. Basically we believe that this 7 gag protein is multi-functional, that the carboxy end, 8 which has many basic residues, probably behaves like 9 the core protein of hepadenovirus and interacts with 10 the DNA genome in the particle, and that probably the 11 amino terminus behaves somewhat like matrix and part 12 of this protein also behaves like tapsin (phonetic). 13 So we were interested in looking at 14 integration, and to do this we used a chronically or 15 persistently infected H-92 cell line, and we cloned 16 out single cell clones of this virus, and in these 17 experiments, these southern blot experiments, we cut 18 with NAG-1, which cuts ounces in the genome, and use 19 the bet probe. 20 And if you do such an experiment, you can 21 see a large number of -- a large amount of viral DNA. 22 This is the cut DNA here. This is the small amount of 23 DNA that was not cut. 24 And in many experiments this obscures the 25 background of integration. So what we do is we treat 69 1 the cells for several weeks with AZT, and this gets 2 rids of all the intracellular DNA in the particles, 3 and then you can easily see the integrated copy number 4 in these single cell clones, and what's very striking 5 is that there is a large number of integrated DNAs. 6 And we've counted upwards of 20 bands per 7 cell of, as I said, single cell clones of foamy virus 8 infected, which were leukemia cells. We've confirmed 9 that these are single integrations by cloning out the 10 junction fragments and showing that they're not 11 duplications or repeats. 12 These are bona fide new integrations. 13 We've also done FSH analysis. This isn't a very good 14 slide, and the arrows are not pointing to all of the 15 integrated copies. You'll just have to take my word 16 for it. There's a rough correlation between the 17 integration number that we see by a southern blot and 18 by FSH analysis, and we do see up to about 20 19 integrated copies per cell genome of this virus as 20 well. 21 And we have been very interested in the 22 mechanism of these multiply integrated pro viruses. 23 I'm not showing you the data. There is, in fact, a 24 NLS sequence within the gag genome foamy virus which 25 Axel Rethwelm's lab shows behave as an NLS. So after 70 1 infection, much of the newly synthesized gag protein 2 goes back into the nucleus. 3 And by using a mutant in the NLS, so that 4 if you delete the NLS the virus can grow fine in 5 tissue culture, it does have a slightly lower titre, 6 and we've shown that by deleting the NLS we prevent 7 accumulation of all of those multiple integrated 8 copies, and we now have clones of cells that only have 9 one, two, or three copies per cell instead of ten or 10 20 copies. 11 So this implicates that there could be 12 some kind of intracellular recycling mechanism that's 13 responsible for the high copy number. However, what 14 complicates this is that we've recently found that the 15 foamy viruses also have another unexpected feature, 16 and that is we have made a vector in which we have 17 replaced the bet gene with GFP, and in this case we 18 also put in the RSV strong promoter. 19 And when we use a vector to infect either 20 BHK cells or these erythroleukemia cells, we get a 21 high infectivity. This is shown with the GFP 22 fluorescent, but surprisingly if we compare 23 infectivity of either these uninfected H-92 cells or 24 this single cell clone that has about 20 integrated 25 copies, we find there's basically no difference in the 71 1 ability to reinfect the uninfected -- the infected 2 cells versus the uninfected cells. 3 So at least in these persistently infected 4 cells that have huge copy numbers, they're not immune 5 to super infection, and this is very surprising for a 6 retrovirus. 7 So, therefore, we can't say whether we're 8 getting accumulation of all of these integrated copies 9 by an intracellular pathway or an extracellular 10 pathway, and this also suggests that even if a cell is 11 infected by foamy virus, it would not be immune to 12 superinfection by another foamy virus or more foamy 13 virus. 14 The other point about foamy -- foamy 15 viruses are not very well studied. There's very 16 little information about the packaging sequences of 17 these viruses. Several groups have tried to make 18 viral vectors, and so far this isn't an attractive 19 genome for viral vectors. 20 For one thing, the virus, as I said, is 21 probably not pathogenic in either accidental or 22 natural hosts. It has a very large genome, greater 23 than 11 KV. It has at least one gene that we believe 24 we can delete without -- at least in vitro -- without 25 any untoward effects on the virus. It has two 72 1 promoters, so it's a very flexible genome. 2 And another point is work from Germany has 3 shown that in infected monkeys you can find foamy 4 virus DNA in every organ in the body, including the 5 brain, although there's very low viral replication in 6 vivo. 7 So it might be a good gene delivery target 8 to evoke a wide variety of organs. So in vector 9 development several groups have found that, in fact, 10 there are probably at least two packaging or two sys 11 acting regions in the RNA, and one is at the five 12 prime end of the genome where you would expect a psi 13 sequence to be, but surprisingly there's also a region 14 in the pol gene that's required for transfer of vector 15 sequences. 16 Whether this is another packaging sequence 17 or has another sys acting RNA function is not known. 18 For instance, since pol is not made as a gag-pol 19 fusion protein and the method of incorporating pol 20 into the particles does seem to require a gag-pol 21 interaction, but we can't rule out that perhaps it 22 also needs to bind to the RNA, and so this region of 23 pol could be a pol binding sequence. We don't really 24 know. 25 So the packaging regions of this virus are 73 1 not at all well understood at all. 2 So in summary, all retroviruses have 3 packaging signals. Unfortunately those of the foamy 4 viruses have not yet been delineated. Foamy viruses 5 package RNA, although the functional genome appears to 6 be DNA. 7 In some cells, at least in tissue culture, 8 foamy virus infection leads to multiple integration. 9 This is an interesting point because one would think 10 that if such a thing occurred in vivo, and we have 11 absolutely no evidence for it, that foamy viruses 12 would be all set up for promoter insertions and 13 inductions of tumors in infected animals. Yet this 14 has never ever been seen. 15 So it's possible that the replication in 16 vivo is so meager that the virus never really does 17 multiply integrate, but nobody has ever been able to 18 or nobody has ever looked in vivo for foamy virus 19 integration sites. 20 Nothing is known about the recombination 21 between foamy virus and other retroviruses. Clearly, 22 experiments that are done in monkeys, infecting them 23 with other viruses such as SIV or viral vectors, there 24 is probably foamy virus in all of those animals. So 25 it would be very interesting to know something more 74 1 about how foamy virus interacts with other types of 2 retroviruses. 3 And also any packaging cells derived from 4 primates, bovine or feline species need to assess the 5 effect of foamy virus as well. 6 And I think that's all I have to say. 7 (Applause.) 8 CHAIRPERSON ROSENBERG: We need to keep 9 the questions, I'm afraid, brief because we are 10 seriously behind. So I believe we're supposed to be 11 at the break, but we still have one more speaker. 12 PARTICIPANT: Maxine, does anybody know if 13 vaccines have been checked for foamy virus 14 contamination? 15 DR. LINIAL: As far as I know, no. 16 PARTICIPANT: You mean nobody has looked 17 or as far as you know? 18 DR. LINIAL: I don't know. There are very 19 few reagents. I mean, there are reagents for the so- 20 called human or chimp foamy virus, but as far as I 21 know, there are no good antibody reagents. 22 PARTICIPANT: There are. 23 DR. LINIAL: There are? 24 PARTICIPANT: There are? They are 25 checked. Okay. 75 1 DR. LINIAL: Are they checked for all of 2 the simian foamy viruses? 3 PARTICIPANT: No. 4 PARTICIPANT: By PCR? Is that -- 5 PARTICIPANT: (Inaudible.) 6 CHAIRPERSON ROSENBERG: Could someone 7 repeat this so that -- 8 PARTICIPANT: I think it's by a 9 combination of tests, including PCR and infectivity 10 tests; is that right? Okay. 11 DR. LINIAL: One problem is that these 12 monkeys do get cross-infected with other foamy 13 viruses. So, you know, I don't know how many you're 14 looking at. 15 PARTICIPANT: Since foamy viruses seem to 16 be breaking all of the rules, I wonder if it's worth 17 asking or is it known whether integration is required 18 for replication, whether there's significant -- since 19 there are so many copies of DNA, et cetera, whether 20 there can be significant expression and replication in 21 the absence of integration. 22 DR. LINIAL: My lab, as well as a lab in 23 Germany, have made a DD35E integrate mutants, and at 24 least in tissue culture it's completely dead. So we 25 believe integration is required. 76 1 CHAIRPERSON ROSENBERG: I think we need to 2 move on. The last talk in this session is by John 3 Kappes. 4 DR. KAPPES: Just a momentary delay. 5 Technology, MacIntosh, a little slower in booting. 6 Perhaps I'll begin a short introduction. 7 The focus of my work -- there we go -- has 8 been on the possible use of lentiviral vectors for 9 gene therapy. Off to a bad start. That is the second 10 slide. 11 The principal concern for using lentiviral 12 vectors for gene therapy is that they may recombine to 13 produce replication competent retrovirus, and 14 underlying the concerns for replication competent 15 retrovirus is genetic recombination. 16 There have been a number of different 17 types of -- and I'm going to focus really just on HIV- 18 based vectors, although I'll probably refer to them 19 many times as lenti -- but there have been a number of 20 different HIV-based lentiviral vectors produced to 21 minimize the pathogenic properties of any RCR that 22 could emerge, and those include deletions of most of 23 the accessory genes, deletions of even the TAT and REV 24 regulatory genes, the lesions in U3, and while I won't 25 focus on a lot of these details, I will focus on, 77 1 again, what I think is fundamental to understanding 2 the risks associated with generating replication 3 competent retrovirus, and that is genetic 4 recombination. 5 The assays which have been used thus far 6 include, that is, to measure recombination or, if you 7 will, really more because of the way they've been 8 applied measurements of RCR, include gag transfer -- 9 oh, TAT transfer it should be -- gag transfer, and DNA 10 mobilization of marker rescue assays. 11 It's unlikely, and I'm sure they have not 12 been suggested by the authors who published on these 13 assays to be adequate indicators of the risk 14 associated with these viruses or these viral vectors 15 in vivo. Especially that's true in the long term. 16 So today I will present data from an 17 approach that I devised to understand the risk, if you 18 will, of using HIV-based vectors through an analysis 19 that focuses on genetic recombination. 20 First, I'll present one slide, the unique 21 difference that I've used to enable the detection of 22 recombinant viruses and their analysis, and then 23 several slides I'll show the detection and 24 characterization of these recombinants, both 25 biologically and genetically, and finally, in one of 78 1 two slides I'll show how I have further disarmed or 2 dismantled or split the functions of the lentiviral 3 based vector to improve safety. 4 This depicts the three component systems 5 for HIV-based vectors, not necessarily analogous to 6 what might be thought of as third generation, but the 7 important point is that there is a packaging 8 construct, a vector construct, and an envelope 9 construct, and just for this one slide, I wanted to 10 point out in particular the TAT gene because my 11 recombination assay is based on TAT. 12 That is, if recombination occurs between 13 the vector and the packaging construct, and if TAT is 14 included, I will be able to select for the recombinant 15 using this approach. 16 If lentiviral vectors are generated 17 through transfection of 2-9-3 T cells, which is what 18 I will show in every case, genetic recombination can 19 occur during reverse transcription to generate an LTR 20 TAT containing structure. It could contain gag; it 21 could contain the full packaging construction, but 22 minimally if it contains TAT and TAT is expressed, it 23 could confer resistance to puromycin in the cell line 24 that I call Hela-puro. 25 This cell line was transduced with this 79 1 construct, which confers resistance to puromycin 2 selection when the cells infected with a virus or a 3 vector containing and expressing TAT. 4 A couple of other points worth noting for 5 this slide because it's really what I will use in 6 terms of the components that generate the vector and 7 the data I'll show in every slide, except for a couple 8 at the end. 9 The packaging construct is driven by a CMB 10 promoter. It contains gag-pol-vif. Vpr, vpu, mpf 11 (phonetic) are deleted, and also importantly nef is 12 deleted, importantly because there is no overlapping 13 sequence at this end of the genome, which will become 14 important for reasons which hopefully will be obvious 15 later. 16 So to generate the HIV vectors, I 17 transfect these three constructs into 2-9-3 T cells, 18 and that's depicted here. 19 Through transfection bioparticles are 20 generated, and these bioparticles can be used to 21 infect the Hela-puro cell line. 22 If TAT is expressed, the cell should be 23 resistant, and that's exactly what is shown here. 24 These are the colonies stained with crystal violet, I 25 believe, about nine days after puromycin selection, 80 1 and what's shown is that there is approximately 1,000 2 colony forming units per ten to the seventh infectious 3 particles. 4 Also, importantly, if this infection is 5 done in the presence of a niverapine, there are no 6 resistant colonies detected, implying that it's 7 mediated through the HIV-1 reverse transcriptase, and 8 in particular, I chose niverapine because it's 9 specific for HIV-1 RT. 10 As I suggested, recombination could occur 11 through any region of the packaging construct. As 12 long as TAT was picked up, resistant colonies could be 13 produced. So to confirm TAT was actually present, we 14 designed primers to amplify the first exon, and that's 15 what's depicted here in what I call the lenti pool. 16 This is the lenti pool. After the cells 17 were expanded, the high molecular DNA was extracted, 18 and PCR shows detection of a 219 base pair fragment 19 similar to that detected in proviral DNA. 20 Because or at least in part because data 21 had been published that showed there was no TAT 22 transfer, no gag transfer, no marker rescue using 23 lentiviral vectors, I wanted to address each of those 24 points. 25 And the first series of slides I just 81 1 showed would suggest that there is TAT transfer if the 2 system is sensitive enough to detect it. 3 This slide detects a slightly different 4 approach for looking toward gag transfer. This is the 5 path I just described for TAT transfer. Because 6 recombination can occur in many ways where an 7 infectious virus would not be generated, I chose this 8 pathway for gag transfer because I was really looking 9 for an open gag reading frame. 10 So the difference is instead of going at 11 the heel of puro cells, I infected the virus particles 12 that were generated by transfection into 2-9-3 T, and 13 in 2-9-3 T, if a recombinant forms such that you had 14 an open gag-pol rating frame, you would expect that 15 particles could be generated, and if those particles 16 were pseudotyped by transfecting VSVG into the 2-9-3 17 T cells two days after infection, then they could 18 infect the Hela-puro cell line and confer resistance, 19 and that data is shown here. 20 We detected 540 resistant colonies when 21 the virus derived by 2-9-3 T cells was used to infect 22 the Hela-puro cell line, suggesting both DNA 23 mobilization and gag transfer. 24 Another way of looking at DNA -- oh, let 25 me point out that we also confirmed from the 82 1 supernatants of the puromycin resistant cells the 2 presence of gag protein, both in culture supernatants 3 and in viral pellets. 4 Importantly, the amount of gag detected 5 was increased by about tenfold if TAT was transfected 6 into those puro resistant cells, suggesting that the 7 LTR was truly linked, that is, expressed in sys, with 8 the packaging construct. 9 Another experiment, which was just a 10 further extension of what I've already shown related 11 to the gag transfer, to more specifically and more 12 clearly differentiate between artifacts and true DNA 13 mobilization is the depicted here. 14 The recombinant, which I referred to as 15 mobilizing gag or gag transfer is depicted here. We 16 don't know at this point whether this is the exact 17 structure, but it should contain an LTR gag-pol 18 reading frame, and certainly tat and rre. If this 19 structure is respent in the Hela-puro line, it's 20 possible that when it produces particles because the 21 Hela-puro cell line contains puromycin and a packaging 22 signal further puro RNA strained, that it, too, could 23 be packaged into these particles, mobiled to heal a 24 TAT line in this case, and confer resistance to 25 puromycin. And indeed, that's the result shown here. 83 1 So the data indicate by three assays 2 recombination of the lentiviral vector components, in 3 particular, recombination between the gene transfer 4 vector and the packaging construct. So all of the 5 data I've shown just far are really just biological 6 evidence, and now I would like to turn to the genetic 7 and structure data, that also supports these 8 conclusions. 9 If a combination truly occurred between 10 the packaging construct and the vector, you'd expect 11 the two to be linked, and if I used primers in the U3 12 region and at the end of gag, I would expect to 13 amplify a product from this lengthy viral pool, which 14 is what it's called. This was the product of what I 15 call gag transfer. 16 I would expect to amplify a PCR fragment 17 of about 2,000 base pairs, and indeed, that is what's 18 shown. 19 If this PCR product was cloned and 20 sequenced, and if there was recombination into the 21 packaging construct, what would be expected is this 22 sequence, that is, the initiation codon for gag and a 23 sequence depicted here, and in ten out of ten 24 recombinants, this is what we found. 25 This sequence, it's also worth pointing 84 1 out, is different than that of the original vector 2 sequence. Remember the vector includes part of gag 3 because it's important for packaging. There are clear 4 differences between these nucleotides. So this is 5 further evidence that recombination between the vector 6 and packaging construct occurred. 7 Looking instead of here on the three- 8 pronged end of the genome, we used primers that would 9 span from TAT, shown by this yellow bar, into the U3. 10 By PCR we amplified three, maybe four small bands, and 11 this one here, which corresponds in weight to what we 12 expected, was cloned and sequenced. 13 And that analysis is shown here, and I'll 14 show this in more detail on a couple of later slides, 15 but the important point is that by sequencing through 16 this entire region, what we found is that there is 17 genetic evidence for recombination similar to what we 18 found on the five prime end on the three prime end, 19 and in fact, genetic recombination or different 20 recombinants were found. 21 There were six recombinants that were of 22 this type, one of this type, one of this type, and one 23 of this type, and what are these types? This is what 24 I show on the next one of two slides. 25 Also depicted here is how I suggest this 85 1 recombination is occurring. Remember reverse 2 transcription is going to start up here somewhere, and 3 the growing DNA chain will anneal to the RNA on the 4 three-prong end, and as it's extended, in one of the 5 situations I show I would suggest recombination 6 occurring into the poly A tail of the packaging 7 construct, in another case, into the poly A tail at a 8 different position, and in another into the poly A 9 tail at yet another position. 10 What these recombinants have in common is 11 that all of them extended up into 108 base pairs past 12 the end of U3 or the beginning of U3, depending on 13 which way you're going. 14 The other type of recombinant detectable 15 is one which went in at yet a different position, in 16 position 45 or it contained 45 As downstream of the 17 poly A signal, but it only contained 32 base pairs 18 upstream of the U3 sequence. 19 This finding was unexpected, and I thought 20 for a long time about how it might be occurring, and 21 while I still can't explain it, here's a possible 22 rational model based on the limited information I 23 have. 24 If this is the RNA tail, poly tail, from 25 the packaging construct, and this is the DNA which is 86 1 growing from reverse transcription, what I noticed in 2 all cases just prior to where the recombination even 3 occurred -- and remember eight of the recombinants 4 jumped from this position into various areas of the 5 poly A, and one of them jumped from this position into 6 I think it was number 47A downstream of the poly A 7 signal. 8 What they all have in common is a sequence 9 of Ps, which was suggested could bring those 10 structures together to allow recombination. So based 11 on this data, I think there are several safety 12 concerns which I could point out. 13 One, the regeneration of envelope 14 deficient recombinant lentivirus requires 15 recombination of only two genetic elements. That 16 would be the packaging construct and the vector. 17 The LTR packages signal, gag-pol LTR 18 recombinant contains all of the necessary viral 19 replication machinery for replication minus envelope. 20 The long term risk of stably integrated 21 structures that might look something like this is 22 unknown. Lentiviral vector recombinants may infect 23 nondividing cells if envelope is provided in trains, 24 and I think this might be particularly worthwhile to 25 consider because unlike what we've experienced with 87 1 Maloney, the lentiviral vectors, if this structure 2 exists in vivo, and it can be mobilized to nondividing 3 cells. 4 I think especially this issue is important 5 in cells transduced or CD-34 positive stem cells 6 transduced with lentiviral vectors since it could 7 represent a reservoir for ongoing gene expression. 8 Okay. Quickly just a couple more slides. 9 One possible approach to minimize the risks associated 10 with genetic recombination between the vector and the 11 packaging component I would depict, I think, in just 12 two slides, and this is really based on a study that 13 Wood, et al., published in EMBO in 1997, where instead 14 of incorporating reverse transcriptase and integrate 15 as part of the gag-pol precursor protein, these genes 16 refused to vpr and expressed in trans with an RT 17 integrate minus provirus, and what was demonstrated 18 was that the provirus or that the RT integrate was 19 packaged. It was processed, including into mature 20 integrate and RT, P-51 and P-56 components, and they 21 were fully functional, so provided an opportunity to 22 further reduce the basic components of the HIV vector 23 system by deleting RT and integrate from the packaging 24 construct, and that's what's depicted here. 25 So now the packaging construct of what I 88 1 call translentiviral vector looks like this. It's 2 broken apart into RT integrate infused with vpr, 3 provided separately from gag-pro. 4 And one slide to show data, although there 5 are others that show very similar results. If the 2- 6 9-3 T cells are infected with translentiviral vector 7 derived from transfection of all the vector 8 components, this structure can produce viral 9 particles, but upon infection of the Hela-puro cell 10 line, there are -- no resistant colonies are 11 generated, and that's due to a lack of reverse 12 transcription and integration machinery because even 13 though infection might occur in the absence of reverse 14 transcription of the genome, no resistance will be 15 conferred. 16 So finally, in conclusion, recombination 17 does occur between the packaging construct and the 18 gene transfer vector. This structure is regenerated, 19 which contains all the basic replication machinery 20 minus envelope. 21 LTR containing recombinants integrate into 22 the chromosomes of the target cells. The recombinants 23 express viral proteins, including tat, gag, and the 24 entire gag-pol precursor protein. 25 The expression of the integrated gag and 89 1 pol genes produces virus particles which can be 2 mobilized if envelope is provided in trans. 3 Recombination within the RNA poly A tail 4 may represent a mechanism by which genes without 5 homologous sequence can be mobilized, including 6 oncogenes, which we discussed at some length 7 yesterday. 8 Separating RT and integrate from the 9 packaging vector controls the regeneration of a 10 functional gag-pol structure which is absolutely 11 required for RCR and DNA mobilization. 12 Thank you. 13 (Applause.) 14 CHAIRPERSON ROSENBERG: I don't think it's 15 on. Switch on the bottom maybe. 16 DR. KUNG: I just want to comment that, in 17 fact, in support of your data there are two -- 18 PARTICIPANTS: Can't hear you. 19 CHAIRPERSON ROSENBERG: Why don't you just 20 repeat his question after or you can come up here? 21 DR. KUNG: Okay. I just want to comment 22 that in two naturally occurring avian retrovirus that 23 carry oncogenes; one is avian rhesoblastosis. One of 24 the isolate recombination takes place at the poly A 25 tail. Okay? In that case it's 2-lycine. Is there a 90 1 crossover point? 2 Okay, and the other is fips (phonetic). 3 I think Mike Bishop published that. Recombination 4 also takes place at the poly A sequence. 5 So there are two naturally occurring 6 oncogene carrying. 7 DR. KAPPES: I was unaware of that. The 8 data were very surprising. 9 CHAIRPERSON ROSENBERG: I think this 10 should be the last question because we are actually 11 even beyond the break. 12 DR. LINIAL: What kind of titres are you 13 getting from the last where you break up RT into PR 14 and RT integrate? What kind of titres of vector are 15 you getting from those constructs? 16 DR. KAPPES: Let me be sure I understand 17 the question. What kind of titres are we getting when 18 we separate RT and integrate? Consistently, and maybe 19 this isn't the best answer, but it's important I 20 believe, consistently about three to fivefold less 21 than what we derive from the lenti. Under our best 22 conditions without concentration we can get near ten 23 to the seventh per mL from the lenti. So under best 24 conditions, maybe five times ten to the sixth with the 25 translenti. 91 1 CHAIRPERSON ROSENBERG: I'd like to thank 2 all of the speakers. Hopefully there's time for a 3 short break before the panel. 4 (Whereupon, the foregoing matter went off 5 the record at 10:12 a.m. and went back on 6 the record at 10:32 a.m.) 7 DR. HUGHES: I would ask everyone to 8 please come and take their seats. 9 Mindful of the fact that we are what 10 stands between you and lunch, we're going to try to do 11 this in a timely way and conclude the session as it 12 was originally scheduled. 13 I think in keeping with the spirit that 14 the conveners of the meeting suggested for us, we need 15 to consider the issues that are before us in the 16 context of the proposal to use neoplastic cells as 17 vehicles for generating vaccine strains. 18 And I think in that context, it's probably 19 important to actually give consideration to the 20 differences in the types of vaccines that people might 21 wish to generate, for example, the difference between 22 live vaccine, kill vaccine, or sub-unit vaccine. 23 And what we've heard in the last couple of 24 days are a reasonably good description of the manifold 25 interactions that viruses have with each other and 92 1 with their cellular hosts, and I think we need to give 2 consideration to that in terms of how we would think 3 about detecting interactions between viruses and other 4 viruses or viruses and their hosts so that the 5 vaccines that would be produced would be as safe as 6 possible. 7 And as Dr. Lewis suggested when the 8 meeting was opened, we need to think in terms of what 9 we mean by being safe. We need to think about the 10 problems of detecting adventitious agents, and we also 11 need to understand and think about precisely the kinds 12 of tradeoffs that exist between understanding issues 13 of safety and issues of the benefits that are provided 14 by affected vaccines. 15 And I think with those remarks, I would 16 open the discussion both to the panel and to the 17 audience. Are there comments from the panel? Does 18 anyone have any remarks they'd like to make to begin? 19 No. The audience, are there questions? 20 We may get to lunch early. 21 (Laughter.) 22 DR. HUGHES: Surely the prospects of the 23 kinds of interactions that we've seen, interactions 24 not only between endogenous retroviruses and exogenous 25 elements, but between retroviruses and herpes viruses 93 1 must have provoked some thoughts that are relevant to 2 this issue of vaccine development. 3 Please. 4 MR. KRAUSE: Just to get things rolling, 5 I'm Phil Krause from the FDA. 6 What does the panel see as the likelihood 7 that these kinds of interactions or risks or potential 8 risks might be greater with neoplastic cells than with 9 other types of cells? 10 And also, as you say, what kinds of viral 11 vaccines does one have to be more concerned about? 12 You know, I can imagine based on what we heard that we 13 might be more concerned about these kinds of things 14 happening if one is dealing with larger DNA viral 15 vaccines or with retroviral vaccines than with, you 16 know, small RNA virus vaccines, for instance, but I 17 would be interested in what people have to say on 18 that. 19 DR. HUGHES: Ladies and gentlemen. 20 DR. EVANS: Well, I'm not sure that 21 there's much difference between a transformed cell 22 line and a regular cell line, especially like a 23 primary cell line where you've got a lot of 24 differentiated cells there at least initially, but 25 like I said, I'm not sure. I don't know. I don't 94 1 know if there's elevated copies in certain cell lines 2 of endogenous sequences that might be packaged and 3 transferred, or I don't know if there might be change 4 when you infect that cell line with some other agent 5 because you could be inducing sequences and promoting 6 transfergenic information. 7 In terms of the types of viruses that -- 8 types of vaccines you might make because of the degree 9 of heterologous pseudotyping between different types 10 of viruses in terms of retrovirus transfer, I think 11 you've got to -- anything that has an envelope protein 12 that directs the tropism of that virus you're going to 13 have to be concerned about in terms of transferring 14 the vector. 15 In terms of that having a deleterious 16 effect or whether or not you could detect it, you 17 know, and eliminate it by monitoring is a different 18 question. 19 DR. LINIAL: In terms of foamy viruses, I 20 would say that the real danger there is using things 21 that aren't cell lines because presumably you can 22 easily assay cell lines, and probably most of them 23 won't make it to cell lines if they have foamy virus, 24 but you know, using primary explants from monkeys, you 25 probably run a higher risk of having foamy virus in 95 1 your material. 2 DR. PATIENCE: I think with respect to the 3 endogenous retrovirus sequences, I think basically we 4 have to consider every potential cell line on an 5 individual basis. As far as I'm aware, at least, 6 there's no correlation between a neoplastic cell type 7 and an ordinary cell type. It really has to be looked 8 at on every cell. 9 With respect to the types of gene therapy 10 packaging systems, there's never been any reports of 11 lentiviral related sequences in the human genome, and 12 with respect to the leukemia-based virus systems 13 probably the closest family is the RTV LH system, 14 which we could see no cross-packaging by MuLV gag. 15 MR. FRIED: Mike Fried, ICRF. 16 From what we talked about yesterday, 17 what's the chance of the capture of a cellular gene or 18 an oncogene? Has anybody done experiments where 19 there's an activated rash and then take it and try to 20 transfer it or an antibiotic gene? 21 I mean is it easy to pick up or not? 22 DR. HUGHES: The best data as far as I 23 know suggest that those events happen so rarely that 24 they're not seen in cultured cells, if we're talking 25 about situations in which there's no homology between 96 1 the cellular component and the viral component. 2 You're talking about the straight acquisition. 3 What can be made to happen, depending on 4 the systems, and perhaps Dr. Kung would comment for 5 us, at reasonably frequent -- reasonably frequently in 6 infected animal systems. 7 So I think there is a substantial 8 difference in terms of the history that we have for 9 cells in culture and infected animals. 10 Hsing-Jien, you might want to speak to 11 your experiences with respect to the herpes system. 12 DR. KUNG: Okay. I guess, again, all of 13 this will vary because of selection pressure. So I'm 14 talking about still in terms of tumor selection. 15 In the case of recombination generates so- 16 called acute oncogenic containing virus, the 17 efficiency vary tremendously, but one of the, I guess, 18 champion is the avian erythroblastosis virus that 19 Harriet Robinson reports about 50 percent; you can get 20 50 percent of the tumors that are released oncogenic 21 virus. 22 And there, again, I think the reason is 23 because you select for the tumor, and another reason 24 is that in the case of RB activation, the retrovirus 25 integration utilized the five prime LTR. I think 97 1 Naomi talked about it a little bit. 2 Five prime LTR integration, therefore it's 3 five prime LTR. It goes through the packaging signal, 4 then splice into the RB gene. In so doing, you do 5 not need any deletion in between, and I think John 6 Coffin earlier recalled that this reach through type 7 of transcriptional activation can be as high as 15 8 percent. 9 So the provirus position could very well 10 to pick up a packaging sequence linked oncogene, and 11 I think that is the theoretical base why the RB gene 12 captured by retrovirus had been very frequent, and I 13 think that's a pretty general phenomenon. 14 In our lab the frequency of generating 15 those recombinants are very high. At Harriet 16 Robinson's lab, she reports about 40 percent. So it 17 is very high, but again, this is selection, you know, 18 selection tumor. 19 I guess. I don't know. Maxine, you have 20 done something that's without selection, just look at 21 packaging and recombination, right? 22 DR. LINIAL: Well, we were using selection 23 for a juggling marker, and you if you select for 24 anything with retroviruses, you can basically find it 25 at some frequency. How this would occur in the 98 1 absence of selection is unknown. 2 DR. KAZAZIAN: Haig Kazazian from Penn. 3 I just wanted to make a comment with 4 regard to the last question on endogenous sequences 5 and whether there be a difference in their movement in 6 differentiated cells versus undifferentiated cells. 7 I'll be speaking this afternoon about 8 retrotransposons, and these sequences are transcribed 9 pretty much specifically in undifferentiated cells. 10 So they're transcribed in tumor cells, and they're not 11 transcribed in differentiated cells. So you can't get 12 retrotransposons to move in differentiated cells, but 13 you can very easily in undifferentiated cells. So 14 that is a difference that these endogenous sequences 15 have. 16 MS. HOROWITZ: I'm Jill Horowitz from 17 Wyeth-Lederle Vaccines. 18 And I'd like to know if anyone either on 19 the panel or in the audience is aware of studies aimed 20 at determining whether any of the millions of 21 recipients of life vaccines have currently or have in 22 the past harbored viruses that have resulted from some 23 of these live vaccine undergoing dangerous 24 alterations. 25 DR. HUGHES: Would anyone care to comment? 99 1 DR. DESROSIERS: I'm not aware of any 2 specific examples, but I'm going to take the 3 opportunity to express an opinion, and it seconds 4 something that Maxine just mentioned. 5 In terms of adventitious agents or 6 unwanted virus populations, I worry much, much more 7 about the use of primary African green monkey kidney 8 cells or any such substrate used for production of 9 vaccines than I do about a tumor cell line or an 10 established cell line that may be extensively 11 characterized in terms of potential adventitious 12 agents or unwanted agents. 13 In terms of things like what might come 14 from African green monkey kidney cells that could be 15 problematic, I don't worry so much about the agents 16 that one can test for. I worry about the agents that 17 you can't test for, that you don't know about. So 18 even if there are no errors and no problems and all 19 the agents that you know to test for it can be tested 20 for and excluded from every lot of every vaccine, I 21 worry about the agents that you don't know about. 22 And I'll be honest and say that I'm 23 surprised that primary African green monkey kidney 24 cells continue to be used, and I'm a little bit 25 disappointed that FDA and whoever is involved had not 100 1 had a more serious effort to move away from primary 2 African green monkey kidneys. 3 We all know that there are a number of 4 neurodegenerative conditions and other conditions 5 where viral etiologies have been suspected for years 6 and no viral agent identified. Maybe they're caused 7 by viruses, but maybe they're not. 8 I think despite the fact that there may 9 not be any specific examples of viruses having been 10 introduced, more and more examples are accumulating of 11 the dangers of cross-species transmission either 12 naturally or through other practices, and that's 13 something I think we need to really worry about that 14 can be minimized by going to established cell lines. 15 PARTICIPANT: Walid Heneine (phonetic), 16 CDC. 17 In response to your question, we've been 18 looking at transmission risks of the chicken 19 retroviruses that now we know now contaminate the 20 measles, mumps vaccines and the yellow fever vaccines, 21 and to the two types of viruses that were implicated, 22 there are the endogenous type avian rhekosis 23 (phonetic) viruses and the endogenous avian viruses 24 which you will hear about in later sessions. 25 So far we have not seen any evidence of 101 1 transmissions to vaccine recipients that we have 2 studied by molecular and serologic means. 3 PARTICIPANT: You know, I think you raised 4 a number of very good points, Dr. Desrosiers, and I 5 would like to explore them with you a little bit 6 further in the sense that suppose one has decided then 7 that for new vaccines one is not going to use primary 8 cells, but then that neoplastic cells represent 9 potentially a reasonable alternative. 10 If I sort of follow your comments to their 11 logical conclusion, it sounds to me like you would say 12 then that cells which come from non-human species, 13 even if they're neoplastic, might potentially be 14 slightly more dangerous than those from humans in that 15 sense, and that one might also have to worry about 16 unrecognized adventitious agents in those kinds of 17 cells depending on their history. 18 Would you comment on that? 19 DR. DESROSIERS: What's the question 20 again? 21 (Laughter.) 22 PARTICIPANT: The implication of what you 23 just said was that the species from which a cell comes 24 matters, whether it be primary or nonprimary, primary 25 or neoplastic as one might consider as an alternative, 102 1 and there's the further implication raised by what you 2 said that, in general, one tends to worry about in 3 this context more about unrecognized adventitious 4 agents than recognized one, since you can test for the 5 recognized ones. 6 I just was wondering whether that was a 7 reasonable sort of conclusion based on what you said. 8 DR. DESROSIERS: Yeah. I wasn't referring 9 so much to species of origin. I was referring to the 10 fact that I think that it's much easier to extensively 11 characterize and analyze established cell lines for 12 what's in them and what's not in them than it is a 13 fresh kidney every time you want to do something. 14 That's what I meant. 15 DR. ENGLER: Dr. Engler, Walter Reed Army 16 Medical Center. 17 I just wanted to make a comment and ask a 18 broad question to challenge the panel, speaking from 19 the perspective of a clinician and an educator, trying 20 to translate the contents of meetings like this and 21 the recent Thimerosal meeting for regular folks, 22 whether it's providers or the patients. 23 And the complexity alone, and when people 24 say, "I worry about the unknown agents," we have to 25 worry about foamy virus transmission, lentivirus 103 1 insertion into herpes, and that all translates into, 2 for the common man, woman, and child, is there a bad 3 disease with chronic consequences that can result from 4 immunization. 5 And the Hepatitis B vaccine, the original 6 vaccine, which from my understanding and experience 7 with it was one of the purest vaccines ever made, but 8 could not really be sold very effectively to lots of 9 people, including the highest risk folks that should 10 have gotten it, like surgeons and doctors, and as you 11 all are thinking about these things and trying to help 12 CBER and the FDA, the question you have to ask 13 yourself is: if I choose to use a cell line or use a 14 process to make a vaccine, how do I respond to those 15 concerns? 16 You all may think it's a minimum concern 17 compared to jumping out of an airplane, but we're 18 dealing with a public right now that is extremely 19 phobic and growingly so about vaccines, and 20 congressional representatives who have children who 21 were adversely affected by one of our old vaccines and 22 truly believe that this is the cause of a problem. 23 So you know, yesterday the comment was 24 made, "Well, why don't you do the vaccine process and 25 spike it with your worst fear and see how it comes out 104 1 at the other end?" 2 Now, that's something I can explain to a 3 patient or a nurse or a doctor. An awful lot of what 4 you all have discussed is very hard to translate, but 5 what's left is, "Gee, I'm not sure that this is safe," 6 from your mouths, and so I would just caution before 7 large investments are made in building a vaccine that 8 you consider the audience of who you have to deliver 9 it and how do you write the vaccine information sheet. 10 DR. HUGHES: I'm certainly not by training 11 or persuasion clinically oriented. I'm a basic 12 researcher, but as I understood the charge that we 13 were brought was to try and bring out the issues that 14 have to do with real safety, with the concerns that I 15 think the public should have, that the vaccines that 16 are being made are as safe as intelligent humans know 17 how to make. 18 And I would like to be able to say to you, 19 I would like to be able to say to the people in 20 Congress who in some sense pay my salary, and I'd like 21 to be able to say to the public that you have no 22 concerns; that everything that you're going to see in 23 vaccines and in your life will be completely safe. 24 However, that would be false, and I think 25 the best job we can do under the circumstances as they 105 1 exist in the real world is to try to have a frank and 2 honest discussion of what are, in fact, real concerns 3 and dangers. 4 And as Dr. Lewis pointed out, I believe, 5 in the opening, his opening remarks, we're dealing 6 with a situation in which there are costs and benefits 7 and risks and benefits, and I don't believe that's 8 avoidable. 9 The average American, as far as I know, 10 the most dangerous thing the average American does is 11 to jump into his car and drive to work every morning, 12 and by contrast, as far as I know, the administration 13 or taking of any vaccine strain that's now approved in 14 the United States has minimal risk. 15 Now, I'm certain that people don't 16 understand that, but that seems to me to be a problem 17 that is distinct from the problem that, as I 18 understood it, we were given, which is to try and 19 figure out how to make it as safe as possible and to 20 discuss the potential risks as clearly as possible. 21 And certainly if there were some way we 22 could make your job easier of explaining, I would be 23 happy to try and do that,b ut I think any discussion 24 that isn't frank and isn't complete and doesn't have 25 some clear discussion of the potential risks is doing 106 1 a disservice to exactly the people we're trying to 2 help. 3 And I'm sorry it's complicated. I wish it 4 were not, but I don't think that's an avoidable 5 problem, and I think that if there is an answer to 6 your ultimate concern, it's to try and see that there 7 is some larger and better scientific education in 8 general in this country, which I think is appallingly 9 bad, but I don't really think that's our charge here, 10 although it's a charge that I think we all must in 11 some sense want to participate in and have as part of 12 what we do every day. 13 Would anyone else care to speak to this? 14 Is this different or the same? Are we done here on 15 this one? 16 MR. ONIONS: David Onions, Glasgow. 17 I just thought really perhaps to echo the 18 last speaker and really to echo your comments that I 19 think that the FDA is to be congratulated because I 20 think all regulations should be based on good science, 21 and what we clearly heard in the last two days is some 22 excellent science. 23 If I would just summarize where I thought 24 we were, I think yesterday we put -- I hope we put -- 25 oncogenic DNA to bed. I think perhaps there are 107 1 better assay systems that could be used as I 2 articulated, but I really don't think that's a risk. 3 I think tomorrow we're going to discuss, 4 which I think is the risk, as Ron Desrosiers said, 5 adventitious agents. I think that's still our big 6 concern in any new cell substrate. 7 But I think the novel thing that we heard 8 today, and I'd really like to put this back to the 9 panel, perhaps the novel thing we heard today was is 10 there any concern for the genetic stability of our 11 vaccines given the kinds of work we heard about today, 12 for instance, retroviral insertions or retrotransposon 13 insertions that might occur into vaccines, other 14 possible recombinations between retroviral sequences 15 and other sequences. 16 So should we be looking at other 17 mechanisms in our final lots, for instance, for the 18 genetic stability of our vaccine? Is that an issue or 19 is it not an issue? 20 That's really all I had to ask. 21 DR. HUGHES: Comments from the panel? 22 DR. PETRICCIANI: Well, I guess my comment 23 would be that there is a lot of plasticity for a lot 24 of the genetic elements that we've been discussing, 25 particularly retroviral elements, but it's quite clear 108 1 that the active replication of these viruses for the, 2 you know, reverse transcription is a process that you 3 want to avoid at all cost because that is the major 4 contributor to the plasticity of the genetic element. 5 So if we can eliminate that from a 6 vaccine, that would be a major advantage in 7 stabilizing the genetic content. 8 PARTICIPANT: I think just to try to 9 follow up on what Phil was trying to say, I'm Hanna 10 Golding (phonetic) with CBER, FDA. 11 And the issues that we are really 12 concerned with is the unknown. We are dealing with 13 new cell substrate that are transformed. We don't 14 know their history. We don't know what's the 15 etiology. We don't know what's the multi-event that 16 occurred, and the question whether some of the 17 processes or mechanisms that you have described 18 related to retroviruses might have been participating 19 in the transformation event of these cells that occur 20 in vivo. 21 What we are afraid of, that we not dealing 22 with new cell substrates; that we don't know what 23 caused their transformation, whether they're viral 24 related, whether oncogenes related, whether 25 combination of both, whether some transformation of 109 1 cellular genes into a retroviral type of elements or 2 not. 3 The question is: was the knowledge that 4 you had from your labs -- do you think at least there 5 is a way to start and design certain assays that would 6 allow companies and sponsors, CBER, to try and test 7 these type of cell lines beyond the current testing 8 that is currently available? 9 DR. HUGHES: Well, one possibility is 10 rather than to try and deconvolute the rather complex 11 history, that many cell lines, many permanent cell 12 lines, transformed cell lines have is to simply create 13 anew the kind of cell substrates that Dr. Hayflick, I 14 believe, described in his talk at the beginning of the 15 meeting, in which while you can't be sure that there 16 isn't some retrotransposition of that, you certainly 17 have an understanding of the way in which the cell 18 became a more whole and if it is transformed, 19 transformed, because it is something done in the 20 laboratory, and you can reasonably exclude under those 21 circumstances that it was a viral agent if a viral 22 agent is not employed to make the cell permanent. 23 Now, if the FDA chooses that path, it does 24 exclude a number of the standard lines that people 25 routinely use. I think given the long passage history 110 1 of some of the lines, it will probably be hard to -- 2 you can certainly look at a number of genes in these 3 cells. You can look for certain known viruses, but I 4 believe as you articulated, and in a sense a cell line 5 reflection of what Dr. Desrosiers said, it may be very 6 hard to rule out the sort of unknown agent if the 7 history of the line is not well known. 8 Anyone else? 9 DR. DESROSIERS: Yeah, I'll comment. In 10 terms of any cell line that's being considered, in 11 addition to screening for all known viral agents in 12 such a cell one, one can use differential display, RDA 13 type techniques comparing the cell line to equivalent 14 normal cells to look for the presence of new sequences 15 or additional sequences not present in normal cells. 16 For example, that is how the presence of 17 Kaposi's sarcoma herpes virus DNA was detected in 18 Kaposi's sarcoma cells. It was detected as a new DNA 19 not present in normal cells of the same type. 20 So one can use those sorts of technologies 21 to look for the presence of additional unknown DNAs in 22 the cell line. 23 DR. HUGHES: Dr. Hayflick. 24 DR. HAYFLICK: Hayflick, UC-SF. 25 Several speakers have referred 111 1 continuously to neoplastic cell lines, and I have the 2 feeling that there is a belief that all cell lines are 3 neoplastic. I'm sure that most of the audience knows 4 that that's not true, and we must be very careful 5 because there are continuously propagable, immortal 6 cell populations that do not produce neoplasia in any 7 test that has been conducted. 8 It seems to me that the decision to use 9 CHO cells for the production of non-vaccine 10 biologicals, which was based on the experimental data 11 that those cells, although they might produce 12 neoplasia, do not produce malignancies, was a very 13 wise decision. 14 But I think it's very important to 15 understand that there are continuously propagable cell 16 populations out there that are not neoplastic. 17 Several, in particular, I mentioned the other night, 18 and they are cell populations that have been 19 transfected with H TERT (phonetic) and are immortal as 20 fas as we know. There are well over 400 population 21 doublings at this time, and do not show any signs of 22 neoplasia when inoculated into proper experimental 23 animals, which brings up another final point that I 24 would like to make. 25 And that is the belief that immortal cell 112 1 populations have advantages over diploid cell strains, 2 for example, that have a limited capacity to replicate 3 might not be important in the sense of the presumed 4 advantage of immortality, and the reason for that is 5 that the numbers of population doublings that you can 6 get from a diploid cell strain, if limited to 50 7 population doublings only, yields about 20 million 8 metric tons wet weight of cells, which I think is 9 ample for the production of most biologicals needed on 10 this plant for the next decade or two. 11 And the answer to that is that WI38 still 12 exists. It's been used for the production of vaccines 13 for 37 years now. There's an ample supply of these 14 cells. 15 And furthermore, one must also be aware of 16 the fact that when immortal cell populations are used 17 for the production of biologicals, there is a window 18 of population doubling during which they are permitted 19 to be used. So despite the fact that they're 20 immortal, there is a limited window of opportunity 21 from the working cell bank to the production of the 22 final product during which population doublings the 23 cells are permitted to be used, despite the fact that 24 they're immortal. 25 Thank you. 113 1 DR. HUGHES: Dr. Coffin? 2 DR. COFFIN: One point, that one of the 3 problems with the more limited cell strains is that 4 not every agent that you necessarily might want to be 5 growing in a vaccine will grow on these. If you 6 wanted to grow HIV, for example, WI38 would not be a 7 suitable choice, and you really have very little 8 choice if you want to grow that except to go to either 9 primary human cells or to some immortalized, almost 10 certainly tumorigenic cell line. 11 And so certainly in some instances it will 12 be essential -- will not be possible to use these 13 kinds of cells to do what you'd want to do for all 14 kinds of vaccines one might want to make. 15 The actual point I stood up to make was to 16 actually return to the issue that David raised a few 17 minutes ago which had to do with endogenous elements 18 and genetic instability of vaccines and the 19 possibility of endogenous viral contamination of 20 products. 21 We now know quite a lot about the 22 endogenous virus, the endogenous retrovirus 23 composition of certain species, some birds, mice and 24 humans, in particular. And there's a famous 25 phenomenon that actually hasn't, much to my surprise, 114 1 been mentioned yet, but which I should mention in 2 endogenous retrovirology, and that is that if you take 3 a human tumor line and pass it through nude mice, a 4 very high fraction of the time this tumor line will 5 come out producing very large amounts of a very good 6 looking retrovirus. 7 As chair of IBC, I have had people walk 8 into my office showing me these Grade EMs and saying, 9 you know, "Look what we found." 10 And I say, "Yes, that's a very old 11 phenomenon that has been described many times," and 12 it's because many strains of mice contain an 13 endogenous provirus. We know which one it was. It's 14 a provirus called BXB1 that produces a replication in 15 competent xenogropic virus and which can readily 16 infect human cells, particularly under the sort of co- 17 cultivation circumstances one gets in putting tumors 18 into nude mice. 19 It would seem very obvious that cells that 20 contain this provirus would be very good ones to avoid 21 for the production of biologicals. That actually has 22 not happened. Most monoclonal antibodies that are 23 produced are produced in cells that come derived in 24 some way from BALB-C mice and actually contain this 25 provirus and are not unlikely to contain noticeable 115 1 amount of xenogropic MLD contamination, which could 2 easily infect the recipient if not very carefully 3 removed, or if they were used to grow vaccines, say, 4 DNA virus vaccines, could, indeed, contribute to the 5 sorts of genetic instability that Dr. Kung mentioned 6 earlier. 7 And in fact, this is at least a 8 theoretical possibility with virtually all mouse cells 9 which contain very large numbers of proviruses, many 10 of which have not been shown to be replication 11 competent, but which could in theory at least give 12 rise to some. 13 So I think there ought to be a lot of 14 consideration as to the species of origin of the cells 15 that are used. 16 On the other hand, for example, human 17 cells have not been shown ever to give rise to 18 replication competent endogenous viruses despite 19 considerable effort to try to find them. Some cell 20 lines give rise to particles and so on, but not 21 through replication competence. 22 So I think the species and the strain of 23 origin in the case of mice ought to be given very 24 serious consideration in deciding which of these cells 25 to use to avoid some of these issues. 116 1 MR. ONIONS: I was just going to make -- 2 David Onions -- I was just going to make one point. 3 Ron mentioned using RDA. Representation difference 4 analysis is one of the ways for looking for perhaps 5 new viral sequences in cell lines, and I wanted to 6 come back to that tomorrow in the panel discussion on 7 adventitious agents. 8 But just to make a comment on it briefly 9 now, I think in many ways it's an excellent idea. I 10 think one of the problems is that finding the partner 11 to use as the sort of direct hybridization partner is 12 a real problem with these kinds of cell lines, and 13 when you do these kinds of experiments -- and for my 14 sins I've done them -- you tend to find that, of 15 course, any rearrangement that's occurred in those 16 cell lines, and of course, there are rearrangements in 17 those cell lines, will also come out in the RDA 18 analysis. 19 So it makes it a very messy and also very 20 labor intensive approach, but I think it nevertheless 21 does have merit, and I'd like to come back to that 22 tomorrow. 23 I think one question I would just like to 24 ask the panel was we touched on pseudotyping in 25 several context, and usually when a new cell substrate 117 1 is examined, one of the requirements by the FDA and 2 other regulatory authorities is we look for retroviral 3 sequences, for instance, by the kinds of analysis that 4 Joerg will talk about later, the pertile (phonetic) 5 assay for retrovirus. 6 Then if it's positive, we look and see 7 whether those viruses are infectious for human cells, 8 and that's part of the equation that's used to 9 evaluate the safety of that cell substrate. 10 But what is not usually evaluated is is 11 there any possibility of pseudotyping with the actual 12 vaccine virus that you're putting in. We now know, of 13 course, that you do get pseudotyping, not just 14 retroviruses to retroviruses, but of course with VSV, 15 but also with paramixaviruses, with flaviviruses, and 16 so on and so forth. 17 So the question I'm putting to you is: 18 should that now be part of our safety evaluation, that 19 we actually look at the potential of pseudotyping 20 between any retroviral sequence and the substrate and 21 the viral sequences that we are actually trying to 22 make a vaccine from? 23 DR. HUGHES: Comments from the panel? 24 DR. PATIENCE: I think with respect to at 25 least the gene packaging therapy cell lines which are 118 1 being used, that has already been addressed. I think 2 probably your comment was addressed more at other 3 viruses and other types of vaccine. 4 DR. HUGHES: It also brings up, I think an 5 issue with at least respect to the first part of the 6 comment that may be one of the more vexing problems, 7 and that is I think modern technology, particularly 8 nucleic acid technology is sufficiently robust that 9 it's relatively easy to determine that any agent we 10 know and under stand can be found and, if we're 11 dealing with a cell line, avoided. 12 I think the greater concern is how do we 13 find the at least potentially adventitious agents that 14 we have yet to learn to recognize, and I'd be curious 15 to hear more about that tomorrow myself. 16 Please. 17 PARTICIPANT: I just wanted to comment on 18 the experiment proposed yesterday to spike cell 19 substrates with tumor virus DNA and look for 20 clearance. This was recently just referred to as one 21 of the most understandable parts of this meeting. 22 I also feel it's one of the things that 23 brings the greatest concern. I think asking 24 manufacturers to spike their materials with a 25 dangerous tumor virus DNA poses much more risk than 119 1 any potential endogenous things you're worried about 2 in these cell substrates. 3 So it's a bad idea from a public safety 4 risk perspective. 5 This same experiment, by the way, I think 6 is in the CBER draft guidelines. So I think it's very 7 relevant for discussion. 8 Besides being a bad idea from a risk 9 perspective, it's also not necessary from a scientific 10 perspective. It's much more likely that a spike -- 11 first of all, you could use a spike for a harmless DNA 12 sequence instead of a tumor sequence, and also it's 13 much more relevant to use PCR analysis of endogenous 14 sequences versus a spike because a spike might clear 15 differently than the endogenous sequences. 16 Thank you. 17 DR. HUGHES: As a comment, when I heard 18 the statement made that you're referring to, I thought 19 the proposal was to add something that was easily 20 scorable, and I thought it was either plasmid or 21 lambda that was proposed. 22 I would also say that it was not clear to 23 me in the discussion yesterday that that was not a 24 proposal to show that the process of clearance was 25 valid, as opposed to a manufacturing process that was 120 1 proposed. 2 MR. COOK: I made the comment. Jim Cook, 3 University of Illinois. 4 I think that you've mistaken the point. 5 The point was that an experiment could be done to find 6 out how the clearance process works. This wasn't 7 proposing to put this in some kind of vaccine, run it 8 across into the population and see if polyoma DNA 9 causes tumors in man, which it doesn't. 10 (Laughter.) 11 MR. COOK: So I think doing experiments to 12 find out whether some kind of process is safe or not 13 is perfectly valid. That doesn't mean that that has 14 to be put into some kind of production schedule. 15 I think it's still a good idea, and I'd 16 like to see the experiment done. 17 MR. LEWIS: Andrew Lewis, FDA. 18 I'd like to sort of take the panel back to 19 some more generic concepts, and I just think based on 20 what we heard this morning the question I'd sort of 21 like to have a poll of the panel would be as 22 regulatory people do we really need to be worried 23 about viral-viral and viral-cellular interactions. 24 It seemed to me from the data that was 25 presented this morning that there is reason to be 121 1 concerned about these types of thing, but perhaps I'm 2 wrong. Perhaps this is something that these are 3 experimental conditions, and maybe we don't have to 4 worry about them in the world of generating vaccines 5 in neoplastic cells or in primary cells. 6 So I'd like to have a sense of how you all 7 think about that because this sort of helps us in 8 deciding how seriously we should go about taking these 9 concerns. 10 DR. HUGHES: Well, I'll begin by saying I 11 think it would certainly, given the plasticity of 12 retroviruses, be worth paying attention to. I would 13 also say that if you're asking -- if the question 14 really has to do with neoplastic versus nonneoplastic 15 cells or permanent cells versus primary cells, since 16 the cells of many species, whether they're transformed 17 or not, whether they're permanent or not, have 18 endogenous retroviruses, and certainly things from 19 primary culture can have all sorts of exotic exogenous 20 as well as endogenous viruses; I'm not convinced that 21 there is an enormous difference that is engendered by 22 making the cells either permanent or neoplastic, but 23 that doesn't take away from the larger question of 24 whether or not that's a real issue. 25 Shall we just go down the panel? 122 1 Dr. Desrosiers? 2 DR. DESROSIERS: Pass. 3 DR. HUGHES: Ben? 4 DR. BERKHOUT: Well, regarding SIV and 5 HIV, we haven't seen so far an example where a 6 defective fire resisity (phonetic) parrots by picking 7 up cellular sequences. 8 On the other hand, limited number of 9 experiments have been done so far. So I think we 10 still should do more analysis, in particular, of the 11 monkeys that were infected with live attenuated SRV 12 strains to see if this really is impossible. 13 You know, it may need some more time to 14 actually say that is impossible. 15 DR. KUNG: I guess I have to take a 16 practical point of view. I think the theoretical 17 potential is there. It's almost impossible to avoid, 18 and I'm actually thinking we have no choices in the 19 sense that I think these things will happen, to pick 20 up DNA from retrovirus, and there are a number of 21 endogenous viruses, some of the unknown ones we don't 22 even know. So it's hard to even detect them. 23 But you know, in all likelihood, without 24 any particular selection, those events will come and 25 go, and so if I take a practical point of view, my 123 1 feeling is that it's not a serious problem. That 2 would be my feeling. There's no other choice. If you 3 had better choices, that would be fine. 4 It's very difficult to even identify those 5 things, and so I don't know what the practical 6 solution to that. It does happen. I think all on 7 this panel would say it does happen. 8 DR. LINIAL: I guess there are two issues. 9 The first, I agree with Hsing-Jien and Steve Hughes, 10 and that is the problem of retroviral contamination or 11 contribution to other viral vaccines. I think clearly 12 it's a possibility, but it sounds like, you know, all 13 the common ones are being screened for, and people are 14 doing the best they can. 15 The other issue is using retroviral 16 vectors or retroviruses as vaccines, and I think 17 that's something that really needs a lot more 18 discussion. We've heard a lot of evidence here that 19 none of these things are really safe, and I have a lot 20 of concerns about using retroviruses to put into 21 people, especially possibilities of defective 22 lentiviruses, which we know can cause disease, or even 23 in the case of foamy viruses until we know a lot more 24 about what happens when they're widespread in the 25 human population, which they aren't now. 124 1 I would be very cautious about using 2 retroviruses as agents as opposed to worrying about 3 what retroviruses are doing to other agents. 4 CHAIRPERSON ROSENBERG: I guess I would 5 make two comments. I agree with what some of the 6 other panelists have said. There's nothing that 7 really convinces me that there's a difference per se 8 between using a neoplastic or an immortal-like cell 9 for a substrate. I think it more reflects the 10 properties of the particular cell, not whether it's 11 normal or neoplastic or falls into one of these 12 intermediate categories. 13 Another point I think with regard to the 14 retroviruses. I think it's very difficult to really 15 accurately assess the risk, but we've seen examples 16 presented here, and certainly there are many others 17 where situations that involve replication of the 18 viruses or have the risk of replication of the viruses 19 clearly can lead extremely rapidly -- we've seen 20 several presentations that have emphasized how rapidly 21 changes can occur in these populations. 22 So certainly circumstances that have a 23 risk of replication, I think, must be avoided because 24 the circumstances that can follow are quite 25 unpredictable. I think that's another message that 125 1 has come through. 2 DR. KAPPES: Nava, would you like to say 3 something? 4 DR. SARVER: Sorry, yes. Nava Sarver from 5 the NIH. 6 I'd like to ask the panel and perhaps some 7 people in the audience. There has been, I think, the 8 stringentest (phonetic) case of potential risk with 9 retroviral replication, has been in the case of gene 10 therapy where they've been used extensively for 11 delivering whatever therapeutic genes or replacement 12 genes, and I'd like to know. 13 I mean, I'd like to know -- the FDA and 14 the audience and the people in the panel that have 15 been involved in gene therapy -- whether there are any 16 documented cases where some of the concerns that are 17 expressed here have been materialized, and from the 18 limited knowledge that I have, I am not aware of any 19 identified risk that has occurred in humans, you know, 20 with the application of gene therapy. 21 So perhaps this could be taken as a case 22 study and really exhausted and see what are the risks 23 and are they meritorious. I mean, can we really rely 24 on the concern that is expressed here when we 25 extrapolate this to gene therapy and learn from that? 126 1 DR. KAPPES: I think we have no choice. 2 All of what we're speaking of is highly complex. Each 3 individual question is complex and needs to be 4 addressed individually. 5 I think my simple answer to whether we 6 need to worry about these interactions: yes, and I 7 think that's the obvious answer. No one would 8 probably disagree. 9 Maybe that isn't exactly how the question 10 was posed. The implications are tremendous. I don't 11 see alternatives. 12 In the case of HIV vaccines, we have to 13 continue to look at how the virus is different, for 14 example, in comparison with what we have learned, 15 Nava, maybe as you're referring to with Maloney. 16 One issue I pointed out that's clearly 17 different with lentiviral vectors compared with 18 Maloney is lentiviruses and the recombinants 19 themselves are likely to have the capacity to infect 20 nondividing cells. So if you have recombination and 21 if that recombinant can be pseudotyped by 22 superinfection or even normal cellular receptor ligand 23 interactions, it's possible to mobilize that 24 recombinant to adjacent cells, and that's very 25 different than what you might see with Maloney. 127 1 It sounds hypothetical, theoretical, and 2 certainly it is. So I don't have -- there is no -- I 3 guess my answer is there aren't easy answers, and this 4 is the platform by which I think the FDA, in 5 particular, has established to try to address these 6 important questions. 7 DR. PETRICCIANI: I guess my comment would 8 be to compare vaccine use with gene therapy, is the 9 issue with vaccines. You're involving far more people 10 probably with the potential risk, and using a greater 11 number of division of cells, and possibly the benefit 12 is less for people than gene therapy. There's less 13 personal benefit. 14 So it changes the equation in my mind. 15 My comment about the use of retroviruses 16 is that replication is something that is to be avoided 17 if possible, but you know, if that is present, then 18 you're definitely going to transmit other genetic 19 elements. That's something that will happen 20 regardless of the cell type. 21 So the question between neoplastic cells 22 or permanently dividing cells in my mind becomes less 23 of an issue rather than the types of genetic elements, 24 the endogenous elements that will be transmitted. 25 I think that John Coffin made a good point 128 1 about the species of selection. If you're going to 2 administer a retrovirus to a human, perhaps you might 3 want to think about using a human cell line where 4 there's no new genetic element coming in from a 5 different species. That might be more advantageous 6 than choosing something completely different, like a 7 mouse. 8 DR. SARVER: I'm not trying to minimize. 9 You know, I see the difference. I mean, it's 10 definitely a vast difference between gene therapy and 11 a case where you have the benefit definitely outweighs 12 the risk, and the naive individuals that are going to 13 be vaccinated. 14 What I tried to say is perhaps the gene 15 therapy field can actually give a quantifiable number 16 to the risk, to put the probability that something, an 17 adverse effect can actually take place. 18 And granted we don't have that many number 19 of patients in gene therapy, despite all of the hype 20 that we hear, but nonetheless, it's a defined 21 population of patients, and it can be used as a 22 probability study. What are the chances and can we 23 actually put a definite number on the probability that 24 something risky is -- that something bad can take 25 place? 129 1 This is what I was driving at, whether we 2 can. I mean just to work up the numbers and to say 3 are we talking with infinitesimal probability or is 4 the probability really a quantifiable number. 5 DR. EVANS: One comment about that is 6 we're going to have to wait for a long time after 7 those experiments to see if there's any problem. I 8 mean this could go on for 40, 50 years and something 9 could come up. 10 PARTICIPANT: I was just going to make a 11 comment on that comment from the last speaker. I 12 really don't think it is a very good object lesson 13 because, first of all, the numbers are small. 14 Secondly, a high proportion of the 15 therapies that are being produced are in terminal 16 patients. So there's a high correlation between 17 having had the retrovirus vector and being dead. 18 (Laughter.) 19 PARTICIPANT: And that's not meant to be 20 a flippant comment, but it's true. 21 And I think the third comment is that we 22 know that a lot of these people have had delivered 23 BLS-30 sequences. To my knowledge, there's not been 24 a particular study that's actually followed up the 25 consequences of having BLS-30 mainly because these are 130 1 terminal patients with other problems. 2 DR. PATIENCE: I've got a few points down 3 here relating to this question and earlier with 4 respect for gene therapy and the use of cell lines, 5 such as WI38, it's not practical. Basically a lot of 6 these systems need to go back to single cell clonings, 7 and by the time you got through your 50 doublings, you 8 just wouldn't have a useful packaging cell life. 9 With respect to do we have to be worried, 10 well, I think we are looking at Maloney or at least 11 murine leukemia virus, potentially lentiviruses. A 12 lot of information is known about these viruses. So 13 we can make the systems as safe as we intellectually 14 possibly can and biologically possibly can. They're 15 well understood, which may be reassuring. 16 Coming back to David Onion's point, with 17 respect to the patient either dies or receives a gene 18 therapy vector or another treatment, I think that 19 really does touch on an important point, which is that 20 we have to look at individual scenarios with respect 21 to what species barriers we may be jumping. 22 Retroviruses, for instance, their behavior when you 23 cross the species barrier may be totally different to 24 what you can see in vitro. You in very many cases 25 just cannot predict whether they have any potential 131 1 pathogenic nature or not. 2 And then another order of magnitude on 3 from there is, yes, okay, there is a definable risk 4 with respect to the recipient of a therapy, but does 5 it not -- at least personally speaking, does it not 6 become more of a public health issue if there's the 7 potential for transmission? I think maybe that takes 8 it on to a different level of assessment. 9 DR. RUSCETTI: Yeah, one slightly 10 different point I wanted to make at least in 11 relationship to using mouse packaging cell lines, you 12 may generate recombinant viruses and there's some 13 instances where that has happened, but it's not clear 14 that that actually would be a risk, even if you did 15 give it inadvertently to patients. 16 We know from studies of mouse retroviruses 17 that there are very specific changes. These viruses 18 have to have certain sequences that make them 19 pathogenic, and you may generate lots of different 20 recombinants that don't have those. Even to cause a 21 disease in mice there really is very little evidence 22 that you can generate pathogenic retroviruses in 23 vitro. 24 For these then to cause diseases in humans 25 I think are even less likely. These viruses have to 132 1 replicate at very high titres and certain cell types 2 and perhaps carry certain envoy genes that might have 3 biological effect in human cells. 4 So I just wanted to bring up the point 5 that even though you might generate recombinant 6 viruses, they may pose no risk at all to humans, but 7 granted we don't know, and I think everybody will say 8 here we're not willing to say that there's no risk at 9 all for using these sorts of lines as substrates for 10 vaccines. 11 DR. BROKER: Tom Broker, UAB. 12 I wanted to kind of extend one 13 consideration Nava brought up, and that is there 14 really is a lot of other experience that I think this 15 particular concern today can be referred to, and one 16 of them is simply the blood bank system, the blood 17 transfusions that are very commonplace, the use of 18 clotting factors which have been produced in various 19 ways, and the use of interferons, for example, and a 20 variety of other proteins that are routinely delivered 21 to patients. 22 I think there is a lot of experience here 23 that could relate to the risk of transfer of agents 24 either from contemporary blood donors or through the 25 manufacturing process. 133 1 A second thing I wanted to bring up is 2 that there is concern about, that the general public 3 has, as well as the professional communities, about 4 simply immune reactions to the very product that 5 you're trying to transfer. In the case of vaccines, 6 that's the objective, but in the case of contaminating 7 components in a vaccine, the question is how many 8 different things can you inject into a given person 9 before you start getting negative effects. 10 Something I've experienced, I helped 11 monitor an Internet site for laryngeal papillomatosis, 12 and there is an extraordinary amount of what I call 13 Internet paranoia that is not to be dismissed, patient 14 support groups that absolutely go ballistic over the 15 most elementary rumors warning people not to engage in 16 a certain therapeutic protocol, for example. It's out 17 there, and it's going to be unavoidable. 18 The fourth thing that I wanted to mention, 19 which I'll talk a little bit about tonight and 20 tomorrow, is that a number of cell lines that we have 21 felt to be relatively stable, in fact, continue to 22 evolve. There is a certain degree of genetic 23 instability in cell lines that, while you may know 24 what they are today, ten passages from now they may 25 not be what you thought they were. 134 1 And so rather continuous monitoring of 2 what you've got is going to probably be a necessity. 3 Specifically, I believe this is related to 4 the particular growth environment that the cell line 5 has experienced in its most recent history. If the 6 demands are to divide rapidly and produce large 7 amounts of material, they will have a certain 8 phenotype. If they are put into a more quiescent 9 stage, it appears that they'll be able to alter their 10 gene expression at least, if not underlying genetic 11 changes, to kind of adapt to the demands of how much 12 growth is expected out of them. 13 And the final point I wanted to make, it's 14 a variety of different ones, of course, but that is 15 the immune status of the recipient is going to affect 16 the risk of anything that's transferred into a 17 patient. For example, there's over 21,000 solid organ 18 transplants a year, 25,000 bone marrow transplants a 19 year in AIDS and a variety of other immunosuppressive 20 disorders, and what works in the general population 21 may not work for the several hundred thousand people 22 in the U.S. who have compromised immune status where 23 reactivation of a contaminating virus, for example, 24 will be very different in that recipient compared to 25 an otherwise healthy person. 135 1 So I've raised a lot of issues, but I 2 think they all are relevant to the concerns of this 3 session. 4 Thank you. 5 DR. HAYFLICK: Hayflick, UC-SF. 6 One of the panel members made mention of 7 the fact that a limitation on the use of WI38 or other 8 human diploid cell strains is the observation that 9 they can only undergo about 50 population doublings 10 which precludes their usefulness when cloning is 11 necessary in order to produce a particular product. 12 That was true. It no longer is because 13 WI38 and other normal human diploid cell strains can 14 be immortalized by transfection with H TERT and 15 undergo hundreds and hundreds of population doublings 16 with full retention of their normal properties. 17 Thank you. 18 DR. HUGHES: All right. It is now 19 slightly past 11:30, and unless someone on the panel 20 or in the audience has an additional concern that 21 can't be dealt with after lunch, we'll go to lunch. 22 Thank you very much. 23 (Whereupon, at 11:35 a.m., the meeting was 24 recessed for lunch, to reconvene at 1:00 p.m., the 25 same day.) 136 1 A-F-T-E-R-N-O-O-N S-E-S-S-I-O-N 2 (1:07 p.m.) 3 DR. ZOON: If we could get started, we 4 will now enter into our next session, which is 5 entitled "Cellular DNA as a Potential Source of 6 Oncogenic Activity or Infectious Agents." 7 I am obviously not Rob Breiman. I'm Kathy 8 Zoon, and I'm the Director of the Center for Biologics 9 and happy to have an opportunity to chair this very 10 important session. 11 I do want to announce that we are adding 12 one additional speaker this afternoon. So our break 13 will be delayed by about 15 to 20 minutes in order to 14 accommodate an additional talk that was scheduled for 15 tomorrow, which we are now having this afternoon. So 16 I would appreciate your patience. 17 And also I will ask all the speakers, 18 because we are having an extra speaker this afternoon, 19 that please keep to your time so that we give 20 everybody an opportunity to have their full time to 21 present their information. 22 With respect to this topic, this is not a 23 new topic that either the Public Health Service has 24 been working with, and particularly the FDA, over a 25 number of years, whether it's vaccine safety, blood 137 1 safety, or new therapeutic products, including gene 2 therapy, and obviously it's not without its 3 controversy over the years. Levels, acceptable levels 4 of cellular DNA has been a subject of many 5 discussions, all of which over the years have 6 developed into a position in many cases for a number 7 of the products regulated into scientifically sound, 8 I believe, estimates of risk involved with DNA, but as 9 we enter into new areas and looking at new types of 10 applications of cell substrates, we want to insure 11 that the very best science underpins our scientific 12 decision making and regulatory policy. 13 And we appreciate the opportunity of all 14 the participants in this conference in contributing to 15 those efforts. 16 So I now have the pleasure of introducing 17 Dr. James McDougall from the Fred Hutchinson Cancer 18 Research Center, and he's going to be talking to us on 19 designer cell substrate, telomerase activity and cell 20 immortalization. 21 DR. McDOUGALL: Okay. First of all, I 22 should thank the organizers for allowing me to talk 23 this afternoon. I have to get back to Seattle. The 24 sun's shining there. 25 (Laughter.) 138 1 PARTICIPANT: It won't be. 2 DR. McDOUGALL: It won't be by the time I 3 get there, no. 4 We really got into a real interest in 5 telomerase and its inductions because we found out 6 some while ago, maybe a couple of years ago, that one 7 of the oncogenes that the high risk human papilloma 8 viruses induces this enzyme in human epithelial cells, 9 and this first slide just demonstrates the fact that 10 using a trap assay to see the repeat sequences that 11 are produced by the telomerase enzyme, that the 16E6 12 oncogene induces that enzyme, but the other oncogene 13 of papilloma virus HPV16 or 18 does not do so. E7 14 does not induce it. 15 If we put the two together, we actually 16 see an enhancement of that telomerase induction, and 17 if we look at immortalized cell lines that come out of 18 this sort of experiment, and these are in human 19 mammary epithelial cells; if we look at a long term 20 one, in fact, this one is human skin keratinocytes, 21 passage for 84 times, we see that that telomerase 22 level, if anything, increases in these cell lines. 23 One other thing that we were interested in 24 in this context was does this do anything for the 25 immortalized cell, for the ability to immortalize 139 1 cells, and does it do anything for maintaining the 2 telomere length in cells, which is the key point that 3 one's interested in here. 4 And if we look at clones of keratinocytes 5 that are just expressing HPB16 E7, you see that in 6 early and late passage, there is a loss of telomere 7 length in each one of these examples. The average 8 length is dropping. So the telomeres, as they do in 9 any replicating human cell, gradually shorten. 10 However, if we add E-6 to that mix, we now 11 see that they maintain their telomeres at a standard 12 length. So clearly E-6 induction of telomerase here, 13 and these are cells pre-immortalization or pre- 14 transformation, they maintain their telomeres quite 15 efficiently. 16 And the third thing about the human 17 papilloma virus situation is that as with many, if not 18 most, if not all cancers, they are very -- this is 19 cervical cancer associated with HBV16 or 18 -- they 20 are all positive for high levels of telomerase 21 activity. So the normal cells from the same patient 22 are negative, as one would expect. The tumor is 23 always positive. 24 So this was an observation that was of 25 interest to us in terms of HPV and led us into looking 140 1 more closely at the effects of telomerase itself, and 2 this is just a cartoon of the ends of the chromosome, 3 the telomere ends and the telomerase RNA template is 4 here, and this really should now be called with the 5 acronym H TERT, H-t-e-r-t, which is the catalytic 6 subunit of reverse transcriptase that is the active 7 part of the enzyme here. 8 So we set out to look at the effect of 9 this H TERT upon primary epithelial cells, but first 10 of all repeated the Bodner experiment from Gerund and 11 their associates, in which they put H TERT into human 12 fibroblasts, and what you see clearly is that those 13 become immortalized. 14 And as Dr. Hayflick pointed out, if you 15 had 20 tons of WI38, there's no reason why you 16 shouldn't have 2,000 tons of WI38 if you add H TERT. 17 One of the characteristics we looked at 18 was the presence of P16 in these cells, P16 which 19 always increases with senescence, and we looked at the 20 profile of P16 in these cells going through extended 21 life span. 22 Now, we've taken these up to about 300 23 population doublings, and what you see is a very 24 slight increase in the P16 levels from the earlier 25 picture, but it never reaches anything like the level 141 1 of P16 that we see in human keratinocytes. 2 By the time they've gone through 16 to 20 3 passages, they have very high levels of P16, and they 4 senesce. 5 Now, the interesting thing that really 6 came out of this and that we were clued into by the 7 papilloma virus E6/E7 results was when we looked at 8 mammary epithelial cells, we found that if you looked 9 pre-M0, this first point at which mammary cells tend 10 to senesce, we found that when you added 16E6 to these 11 and immortalize them, you lose expression of P16 or it 12 was greatly reduced, which was not the case if you put 13 E7 in. 14 And the same picture occurs with the H 15 TERT. So H TERT added to HMECs in order for those 16 cells to become immortalized, you have to reduce the 17 -- you have to have a reduction in the P16 level, and 18 if we look at skin keratinocytes, the picture is 19 exactly the same. 20 Where E7 is present, together with H TERT, 21 you do not get a reduction in P16 level, but you don't 22 need to because E7 is binding the retinoblastoma 23 protein, and therefore, if you bind the retinoblastoma 24 protein, you will actually -- sorry. Here -- you're 25 actually achieving the same result as if you were 142 1 knocking out P16, but always in the case of H TERT 2 alone in order for the cells to become immortalized, 3 you have to down regulate P16, and that holds in a 4 number of different types of human epithelial cell 5 lines. 6 So the situation is that if you put H TERT 7 into fibroblasts, then clearly they carry on, and 8 there are no major phenotypic changes observed in 9 those cells, but if you put H TERT into epithelial 10 cells, human epithelial cells from any source 11 whatsoever, they have to lose expression either of P16 12 or they have to have the RB pathway knocked out in 13 order for those cells to survive and become 14 immortalized. 15 Now, loss of P16, as I'm sure most and 16 perhaps everybody here knows, is a major feature of 17 many tumors. Up to about 50 percent of human tumors 18 looked at have lost P16 expression in one way or 19 another, and that's certainly true in a number of cell 20 types, and I've just put a number of references to 21 show this effect. 22 So the P16 locus is part of the N4A locus, 23 and P16 is actually one of the two transcripts from 24 this locus. So there's P16 and P14, and we were 25 interested in the fact that when we looked at P16 in 143 1 most of the cell lines, P16 was lost, but P14 remained 2 transcribed and unaffected. 3 However, in more recent experiments that 4 we've been looking at in adenoidal epithelial cells we 5 now find that some of those have lost P14, but still 6 express P16. So it looks as though you can use one or 7 the other of these transcripts to allow this feature 8 to occur. 9 So, again, all of the H TERT immortalized 10 epithelial cells that we look at have lost expression 11 of one or other of these genes. 12 And as you know, these genes have very 13 important roles in the cell cycle pathway, P16 here 14 and P14 here, with its inhibitory effects on MVM2. So 15 these are extremely important genes. 16 Now, we were interested in how P16 was 17 being down regulated in these H TERT cells. One of 18 the ways we approached it was just to look for loss of 19 heterozygosity, and obviously we could do that in a 20 number of ways. This is just using comparative genome 21 hybridization where you compare the cell line that 22 you're interested in with a known normal cell line, a 23 competitive hybridization onto a normal karyotyp, 24 and when you do that, you should always see a line 25 that goes right down the middle here, showing that 144 1 there's no loss or gain of any sequences in either of 2 the cell lines that you're looking at. 3 But what you see on here with a little bit 4 of noise on some of the chromosomes is a clear loss of 5 the short chromosome line, and in fact, this cell line 6 when we look more closely has an isochromosome of the 7 long arms of nine, but has lost the short arms, and 8 that's where P16 locus maps. 9 So clearly loss of heterozygosity is one 10 of the mechanisms for this. 11 Now, a second mechanism that goes right 12 back to some of the talks yesterday is methylation. 13 Showing again a number of H TERT immortalized cell 14 lines from in this cases keratinocytes or some of the 15 adenoidal epithelial cells, and in every case when we 16 look at these, we see that at least one and sometimes 17 both alleles where there's not a lot of heterozygosity 18 are, in fact, the promoter is heavily methylated over 19 time. 20 And so by the time any of these cells 21 become immortalized, there's clearly methylation of 22 the promoter of P16. So clearly two mechanisms are 23 operating in these H-term immortalized human 24 epithelial cells to actually down regulate P16. 25 We haven't looked at P14, but I suspect 145 1 the same sort of results will show for that. Sorry. 2 I went the wrong way. 3 Now, one other feature that concerned us 4 is that, as I said, P16 loss is very common in many 5 tumors. What about the phenotype of H TERT 6 immortalized epithelial cells as opposed to the 7 fibroblasts where we have not seen phenotypic changes, 8 and I qualify that by saying "yet." 9 One of the ways to ask this question is to 10 look at organotypical RAFT cultures in which we put 11 the epithelial cells onto a collagen and fibroblast 12 containing substrate and allow the cells to 13 differentiate and produce what amounts to essentially 14 a normal skin. 15 Since we're putting these genes on with 16 vectors, we also look to see what the vector, which is 17 the LXSN retroviral vector that we use in these 18 systems, and you can see essentially the result is the 19 same. Even though this is a slightly smaller RAFT 20 culture, basically the differentiation pattern is 21 identical. 22 And just to show you with HPV, if we add 23 E7 to this type of RAFT, we begin to see very definite 24 changes in that there's not a clear difference between 25 basal cells and surface cells, and there are obviously 146 1 replicating cells high in the differentiating surface, 2 which is obviously abnormal and what you see in 3 dysplasia, for example, in cervical carcinoma. 4 What happens when we look at the H TERT 5 cells? Well, again, we can find divisions higher up 6 in the level, but equally importantly, if we look at 7 cytokeratin, in this case Cytokeratin 1, so here's a 8 hemotoxin and eosin stain of a RAFT, and here it is 9 with the TERT in it, showing that there clearly is a 10 change in the phenotype here. 11 And if we look at the Cytokeratin 1 12 expression, you can see that it's different in the 13 TERT RAFT. It's lost from these basal cells, and it's 14 clearly very positive in these upper layers, whereas 15 here it's much more evenly distributed. 16 And in one other example, if we look at an 17 integrin, Alpha 6, Beta 4, we see the same sort of 18 changes, where Alpha 6, Beta 4 is normally expressed 19 in normal epithelial cells just in these basal and 20 super basal cells. In the TERT immortalized 21 keratinocytes, it's expressed right through the 22 surface. So clearly a phenotypic change is going on 23 here. 24 And my last example of the fact that there 25 are clearly phenotypic changes in these cells is that 147 1 if we do a karyotypic analysis of a number of cell 2 lines, and this is just a subset of the ones we've 3 looked at, we find some that clearly are quite normal, 4 and as far as we can see, there is no karyotypic 5 change in these by any mechanism we've looked at. 6 But there are some cell lines that clearly 7 are changing. Here's the isochromosome 9-1 with the 8 P16. This nevertheless has lost P16, but has done so 9 only by methylation on both alleles, and here is 10 another H TERT cell line which has really quite 11 massive karyotypic changes in it, another one here 12 with an extra chromosome. 13 So clearly there are phenotypic changes in 14 terms of cell differentiation, and there are changes 15 in the keriotypes of these cells. So H TERT 16 immortalization is not completely benign. 17 So conclusions from this: telomerase, 18 whether by E6 or H TERT, is not sufficient to 19 immortalize human epithelial cells unless this pathway 20 is inactivated. 21 Secondly, we can get inactivation through 22 E7 binding of RB by methylation of P16 promoter or by 23 chromosomal deletion. 24 Thirdly, we can show that E6 activates H 25 TERT, but we're not clear about how that works. 148 1 Although E6 induces myc and myc activates telomerase 2 in some cells, that doesn't seem to be the pathway for 3 E6, since we know that E7 induces myc to quite high 4 levels, but does not result in the activation of 5 telomerase. So clearly there are two different 6 pathways here. 7 We've found that epithelial cells 8 immortalized by TERT can exhibit chromosomal 9 abnormalities, and lastly, that epithelial cells 10 immortalized by E6, E7 or H TERT with P16 loss have 11 similar changes from the normal differentiation 12 program seen in organotypic culture, and these are 13 remarkably similar to what we see both in experimental 14 systems and in cervical carcinoma. 15 Thanks. 16 (Applause.) 17 DR. ZOON: This presentation is now open 18 for discussion and questions, and if I could ask each 19 of the participants to identify themselves when they 20 get to the microphone. 21 Dr. Fried. 22 DR. FRIED: Mike Fried. 23 What happens in the P16 knockout mice? 24 Are the epithelial cells immortal? 25 DR. McDOUGALL: I cannot remember. I 149 1 think they're perfectly normal actually. I don't 2 really know the answer to that. So I won't speculate. 3 DR. FRIED: Okay, and what do you think is 4 the difference between the fibroblasts and the 5 epithelial cells when you add H TERT? I mean one goes 6 and one doesn't. 7 DR. McDOUGALL: I do not know. Again, 8 that's something that we're looking into and trying to 9 think up the right experiments for, but we don't know. 10 DR. HAYFLICK: Is this working? Yes, I'd 11 like to -- this is Hayflick, UC-SF. 12 I'd like to congratulate you on a series 13 of very elegant experiments. I think, however, that 14 one must emphasize that telomerase expression is 15 necessary, as far as we know, for immortalization of 16 cells, but insufficient for production of cancer in 17 those cells. 18 Telomerase is an immortalizing enzyme, and 19 as far as we know not a cancer causing enzyme, and the 20 evidence for this is as follows. 21 Virtually all transformed human cell 22 populations, certainly, have telomerase expression. 23 There are some exceptions, of course. 24 Furthermore, those cell populations have 25 been transformed, as you so well demonstrated, in many 150 1 cases by oncogenic viruses and also by radiation and 2 carcinogens, and as a result of that immortalization 3 telomerase is expressed. 4 But even more to the point is the fact 5 that post fertilization of human sperm and egg, 6 telomerase is expressed for the first three months of 7 gestation in all cells. That expression then 8 diminished and remains constant in all of the stem 9 cell populations, that is, normal stem cell 10 populations, during embryonic development and then 11 during maturation and adulthood. 12 So that the mere expression of telomerase 13 is insufficient in itself to produce cancer, and I 14 think this is very important to recognize. 15 It's also been shown, as you demonstrated 16 here, that as the cells undergo many population 17 doublings after H TERT administration, there are some 18 -- there is aneuploidy developing, and I was happy to 19 see that you have also made this same observation. 20 Thank you. 21 DR. McDOUGALL: Yes. Thank you. 22 Obviously I agree with everything you 23 said, except one thing, and that is that I draw a 24 parallel with what we see with E6, E7, and that 25 parallel is that E6, E7 also are not sufficient to 151 1 induce cancer. There have to be other genetic 2 changes, and my interpretation is that it's highly 3 likely that H TERT immortalized cells will eventually 4 produce those genetic changes, not because of H TERT, 5 but simply because that is a fact of cells as they 6 become immortalized and continue to replicate. 7 So I'm not suggesting H TERT is itself 8 sufficient, but I think the cells are certainly 9 capable of progressing. 10 DR. HAYFLICK: Yes, I agree. There are 11 downstream effects with P53, P16, P21, et cetera, that 12 are essential for the expression of malignancy or 13 mutagenicity. 14 Thanks. 15 DR. ZOON: Okay. Thank you very much. 16 Thank you. 17 Our next speaker I have a great pleasure 18 to introduce, and a long time colleague, is Dr. John 19 Petricciani. He's representing the International 20 Association for Biologicals, and he will be giving an 21 introduction to cell DNA issues, a subject which he 22 has become vastly familiar with over the past 20 23 years. 24 So, John, it's a pleasure. 25 DR. PETRICCIANI: Thank you very much, 152 1 Kathy. It's a pleasure to be here. 2 And as Kathy pointed out in her 3 introductory comments, this is not a new issue. In 4 fact, it's been around for decades, more decades than 5 some of us would like to remember, in fact. 6 Because the DNA issue is so ubiquitous, 7 it's already been discussed almost from the very first 8 lecture here. So what I'm going to do is skip through 9 a number of the introductory slides rather rapidly so 10 as not to bore you and put you to sleep after lunch, 11 and then we'll get into some of the other pieces of 12 this where I think there's room for some discussion. 13 It hardly needs mentioning again that 14 there is a rationale for being concerned about DNA, 15 and it relates simply to the issue of safety of 16 products that make contain even small amounts of DNA 17 because of the theoretical possibility that the DNA in 18 those products, once transferred to the patients 19 receiving the product, may impose an oncogenic risk or 20 some other type of pathology. 21 The risks, just to be a little bit more 22 specific, could include, for example, if there were a 23 complete viral genome in the contaminating DNA, 24 possible at the extreme, possible uptake and 25 expression of the viral genes to the extent that 153 1 intact virus could be replicated. 2 Of concern primarily in recent years to 3 the DNA vaccine and the gene therapy communities has 4 been the possibility of DNA causing mischief in terms 5 of insertional mutagenesis, activation or inactivation 6 of genes, up regulation or down regulations of genes. 7 So that's not an issue peculiar to cell 8 substrates and has been primarily an issue, as I just 9 mentioned, for those two other therapeutic areas. 10 The real area of concentration of concern 11 over the past several decades in the context of this 12 meeting has been on the possibility of tumor induction 13 with very specific emphasis on the possibility of an 14 oncogene being transferred from the cell substrate to 15 the recipients of the product. 16 In addition, there has been some 17 discussion of the possibility through residual 18 cellular DNA of activating proto-oncogenes or 19 inactivating tumor suppressor genes. Both of those 20 possibilities, again, are not unique to this, but of 21 special interest to the people involved in gene 22 therapy and DNA vaccines. 23 Well, there's a lot known about DNA and 24 the possible mischief that it could cause, and it goes 25 back many decades. We know that -- and I'm not going 154 1 to dwell on this because some of these have already 2 been discussed, but even going back to 30, 40 years 3 ago, it's been clear that nucleic acid extracted from 4 phage can be infectious for spiroplasts under the 5 appropriate conditions. 6 So we know that viral nucleic acids for 7 many years can, in fact, be infective. We also know 8 that viral DNA can transform cells, and there are two 9 examples out of many up there on the screen, and as 10 was mentioned yesterday several times, cellular DNA 11 was demonstrated to transform 3T3 cells in a classic 12 experiment in the 1970s. 13 Two out of 26 human tumors formed foci in 14 the 3T3 assays, but complicating the interpretation of 15 those data is also the fact that normal human DNA also 16 caused foci in that assay system. 17 It's also well known that viral DNA can be 18 infective for animals, and we're going to hear a 19 little bit later about the SIV story. So I'm not 20 going to dwell on that. 21 And finally, in another classic 22 experiment, it was shown that viral oncogene, a viral 23 oncogene, v-src, could, in fact, cause tumors in 24 chickens. 25 Now, on the other side of the coin to all 155 1 of those positive things associated with nucleic acids 2 and things to worry about are a bunch of negative 3 results, and going back to the 3T3 experiment that I 4 mentioned a moment ago, it's important to remember 5 that 24 out of 26 of those tumors, in fact, did not 6 cause foci in the 3T3 assay, and equally important is 7 the fact that very high molecular weight, human tumor 8 DNA, was required in order to get a positive, and the 9 molecular weight was around 30 million, and that 20 10 micrograms was about a minimum that would give a 11 positive readout in that system. 12 There's a recent report within the last 13 couple of years of up to 250 micrograms of hybridoma 14 DNA being negative in terms of any evidence of tumor 15 formation in mice over a two-year period. 16 Also 100 micrograms of Hela and ten 17 micrograms of T24 DNA have been negative in an anti- 18 thymocyte newborn rat assay, and it's here important 19 to point out that the T24 cell has an activated ras 20 oncogene associated with it. 21 And then there's a study which was 22 initiated actually during the last year that I was 23 still at CBER, and that was an attempt to do a study 24 in a non-human primate to try to assess the potential 25 tumorigenic risk associated with tumor cell DNA, and 156 1 in that study we used T24, which as I just mentioned 2 has an activated ras oncogene in immunosuppressed 3 rhesus monkeys, and there were both chronic over a 4 two-week period, and acute inoculation of the monkeys 5 intramuscularly, intravenously, and intracerebrally. 6 And those studies were last reported in 7 terms of a follow-up of the inoculated animals about 8 four years ago, and that was after eight years from 9 the initial inoculation. 10 It's now 12 years since the inoculations 11 were done, and we're in the process of trying to 12 locate the monkeys in their colony down in Louisiana 13 for another follow-up. 14 And finally, I just would like to point 15 out that we do have a normal daily burden in all of 16 this of about one nanogram of proto-oncogenes that 17 result from normal cell death in our own bodies. So 18 it's not as if there's a zero background that each of 19 us experiences. There is a significant background of 20 DNA, in particular proto-oncogene DNA. 21 Now, it's been pointed out by us and a 22 number of other people in recent years, again, with 23 special emphasis on the DNA vaccine area that foreign 24 DNA meets significant physiologic obstacles when we 25 encounter it, and I'm not going to go through these 157 1 because they're well described in more detail in the 2 literature. 3 The point here is simply that there are a 4 number of hurdles that DNA meet ounces it gets into a 5 person or an animal. 6 Now, there is a little bit actually of 7 human experience with what one would consider 8 potentially dangerous DNA from products. The first 9 one, which I'll mention, is a -- actually it was 10 mentioned during the first day of the meeting by Dr. 11 Hayflick, and that is there was an experiment done in 12 the mid-1950s in which adenovirus vaccine was produced 13 in retrospect in Hela cells, and it was a crude 14 vaccine with an extremely high probability that Hela 15 cell DNA was a contaminant of the vaccine. 16 It was administered to only six people, 17 and a number of years ago we asked the Army, U.S. 18 Army, which is where this study was done, to conduct 19 a follow-up study on those six people. They were able 20 to actually find the people, locate them, and do a 21 medical assessment, and that was 25 years after the 22 initial inoculation. 23 And the result of that follow-up was that 24 all six people were alive and well and without any 25 evidence of neoplastic disease. A very small number 158 1 of people, but it's important in talking about 2 potential risks to try to understand what data are 3 available to try to put things into perspective. 4 Another piece of information which, again, 5 is perhaps not as complete as one might want, but 6 nevertheless is important to recognize, is a study 7 that was done in the 1970s in which individuals that 8 had received blood transfusion from donors who were 9 later found to develop a neoplastic disease of the 10 lymphoid system, either lymphoma or leukemia, were 11 followed up to see whether or not there was an 12 increased risk in the recipients of blood from those 13 individuals who later developed neoplastic disease of 14 the lymphoid system. 15 And the results of that study showed that 16 there was, in fact, no increased risk for cancers of 17 the lymphoid system seven years after the blood 18 transfusions had been given. 19 Now, that's important also not just 20 because of the potential for transmitting potentially 21 oncogenic cells in the transfusion. It's important 22 because recently it's been shown that human blood for 23 transfusion contains a significant amount of human 24 DNA, which obviously derives from cell death during 25 the holding of the blood for transfusion in blood 159 1 banks. 2 That amount of DNA is substantial. It can 3 range from 75 to 450 micrograms of DNA. So the point 4 here is that not only were potentially oncogenic cells 5 transfused, but also the potentially oncogenic DNA in 6 the blood packs. 7 Now, as was stated a number of times, 8 these discussions are not new today, and there have 9 been a number of previous discussions, and I won't do 10 anything but highlight these very quickly. 11 The first significant one to tackle the 12 DNA issue was 15 years ago in 1984 at an FDA and NIAID 13 workshop, and at that point there was what's been 14 attributed to that meeting, a ten picagram per dose 15 outcome, and let me just show you the actual statement 16 and then make a comment about it. 17 The statement of the working group on DNA 18 stated that procedures for the production of 19 biologicals must demonstrate that no cellular or other 20 unwanted DNA molecules will be in the final product at 21 a level which would have a biological activity, that 22 is, activities which could induce changes of normal 23 cellular processes. 24 I think that is really the point. It's 25 not whether it's ten picagrams or 100 milligrams. The 160 1 point is that what we don't want in the product is DNA 2 or other unwanted materials that are going to have 3 biological effect, and that point is often in this 4 because of the numbers, whether it's 100 picagrams or 5 whatever it is, but this is really the issue. 6 A couple of years later, in 1986, WHO had 7 another working group and reassessing the data in a 8 more thorough manner than could have been done at the 9 workshop, and that's where the 100 picagram per dose 10 target came from. 11 More recently, Dr. Temin did some 12 calculations which were published in terms of risk of 13 DNA in products, particularly emphasizing DNA 14 vaccines, and came up with a risk of activating a 15 proto-oncogene of less than one in ten to the minus 16 16 per DNA molecule. 17 Now, I've translated that into the risk 18 associated with 100 picagrams of DNA so that the 19 numbers here are all normalized to 100 picagrams of 20 DNA. 21 A couple of years later Dr. Reinhardt 22 Kurth did similar calculations in terms of the risk 23 that might be associated with DNA in terms of the 24 probability of two independent activating events in a 25 single cell, and translated into a 100 picagram basis, 161 1 the risk would be one in ten to the 12th, which is 2 very essentially the same as Dr. Temin's calculation. 3 Another group from the Netherlands 4 estimated on a worst case basis for hybridoma DNA that 5 a risk of an adverse event would be one in five times 6 ten to the eight. 7 And then most recently, Dr. Krause and 8 Lewis estimated in a published paper that if cellular 9 DNA were to encode an entire viral genome, that there 10 is a possibility that it could become infectious 11 during the production of viral vaccines. They 12 estimated that if there were 100 genome copies in the 13 cell substrate and that if product contained 100 14 micrograms of cell DNA, then there would be a risk of 15 about one in 400,000 to the recipients of that 16 vaccine. 17 However, if you normalize that to 100 18 picagrams as I've done here, then the risk becomes 19 about one in four times ten to the ninth. 20 Now, that brings us up to the present, and 21 I think the question here is whether there are issues 22 on which there is still some disagreement and on which 23 there should be additional discussion. One of them 24 you've already heard about, and that is the question 25 of whether based on phenotypic characteristics of the 162 1 cells there is a relative risk associated with them in 2 terms of one cell being more tumorigenic than another, 3 and we've already had some discussion about that. 4 Another issue or another difference is in 5 terms of the amount of DNA that may be in a final 6 product, and what we haven't heard much about is 7 purification and downstream processing of various 8 products, including the live viral vaccines using 9 contemporary methods and how that may be applied to 10 reduce the amount of DNA to levels that are not 11 worrisome. 12 And then there is the issue of the size of 13 the DNA, and what are we really worrying about in a 14 product and is it of sufficient size to even be a 15 cause for concern? 16 In my mind the real question is in terms 17 of the cells themselves, is not how well or how badly 18 a cell is behaving in a tumorigenicity test, but 19 really whether the cellular impurities in the product 20 carry a risk. 21 For example, I could argue that a highly 22 malignant cell may have one copy of three activated 23 oncogenes while a very mildly tumorigenic cell has a 24 100 copies of a single oncogene, and I don't know 25 which one is worse in terms of using it as a product. 163 1 And I think the real point is to 2 understand what it is that's in the product and to 3 understand what the characteristics of the cells are 4 in terms of whether you have activated viral oncogenes 5 or other genetic elements that you can focus on and be 6 sure that they're not there in the final product or 7 minimized. 8 Well, let me summarize here and just again 9 reiterate. We do know that viral DNA can infect in 10 vitro and in vivo. We do know that viral DNA can 11 transform cells in vitro, and we do know that viral 12 oncogenes can cause tumors in vivo. 13 But at least to my knowledge, we also know 14 that tumor cell DNA has not caused tumors in any 15 animal system of which I'm aware, and that makes 16 people a little nervous because there's a disparity 17 between all of those positive results and the fact 18 that try as one might, there's been no evidence that 19 tumor cell DNA can cause tumors in animals. 20 Another point to try to remember is that 21 when we look at probabilities of adverse events, they 22 are a direct function of the assumptions that you make 23 that go into the calculations, and you can jiggle 24 around the assumptions and make things look better or 25 worse by orders of magnitude, and it's the assumptions 164 1 and the validity of the assumptions that are really 2 important in making a legitimate probability 3 assessment of adverse events. 4 Now, in the literature, when you try to 5 look at some of the studies that have been done in the 6 past, it's important to look at the details of those 7 studies and not just use the numbers that come out of 8 them, like 20 micrograms of tumor cell DNA causes foci 9 in the 3T3 assay. The devil is in the details, and 10 unless you look at what the things like the 30 million 11 molecular weight that's required, whether you look at 12 whether facilitators like DMSO or calcium phosphate 13 precipitation are needed to get the result, then 14 you're using things inappropriately and misleading 15 when you try to use those numbers in doing a 16 calculation. 17 Now, there are, I think everyone would 18 agree, less than all of the data that we might need to 19 make some really firm conclusions in this area. 20 Unfortunately, the lack of data has been with us for 21 15, 30 years, and I think if there were ever a 22 justification for a research based regulatory 23 organization, this is an example of where it is 24 because there are probably generic pieces of data, 25 basic ones. 165 1 For example, dose response data could be 2 very helpful in understanding whether, for example, 3 there is a threshold below which we really don't have 4 to worry about DNA, and those data don't exist. At 5 least I'm not aware of them. 6 So there are areas where focused 7 regulatory research could be very helpful so that we 8 don't have to come back here 15 years later and go 9 through the same thing again. 10 Now, one final comment, and that is I do 11 think that there is one very special case, and that's 12 been discussed particularly today, and that is live 13 viral vaccines produced in tumor cell lines. It 14 presents very special problems and needs to be 15 addressed with the greatest of care. 16 But I think that that is one element in a 17 very big array of possible products, and we shouldn't 18 confuse the very special case with all of the other 19 ones that are more easily addressed, and we shouldn't 20 make the whole area very difficult when some of them 21 can be handled relatively straightforward. 22 (Applause.) 23 DR. ZOON: Thank you for that very nice 24 summary. 25 Because it ran over, I would ask unless 166 1 there's a burning question that we'll move on to the 2 next presentation. 3 If I could then introduce our next 4 speaker, Dr. Donald Blair. Dr. Blair is from the 5 Frederick Cancer Research and Development Center at 6 NCI, and he will be discussing evaluation of the 7 transforming activity and tumor inducing capacity of 8 tumor cell DNA. 9 DR. BLAIR: Okay. Thank you very much. 10 The subject of the transforming potential 11 of DNA from transformed cells has been something that 12 has come up frequently from the start of this meeting. 13 In fact, this morning someone made the comment that 14 the question had finally been or the danger or hazard 15 of DNA had finally been put to bed. If that's true, 16 then I am going to tell what is probably part of the 17 bedtime story, which is what I was asked to do was to 18 review some of the data from a series of experiments, 19 most of which were done over ten years ago in 20 evaluating the ability of transformed trell (phonetic) 21 DNA to induce foci and tumors in systems, in tissue 22 culture, and in animals. 23 So what I will try to do is to sort of go 24 through some of the techniques that were used in some 25 of these experiments and try to summarize the results 167 1 that we had obtained in some studies that we did, plus 2 adding some from some others. 3 Okay. So what I'll do is review this data 4 and some of our results, and the whole question of 5 this is based on the need to have an effective way to 6 transfer mammalian DNA, transfer traits from DNA into 7 mammalian cells. 8 And attempts have been made ever since the 9 observations you could do this in bacteria were made 10 over the years, and in general, most of them were 11 relatively difficult to reproduce and difficult to 12 obtain. It wasn't really until the development of the 13 technique by Graham and Van der Eb of using calcium 14 phosphate to transfer DNA that there was really a 15 reliable way of doing transfer of DNA into mammalian 16 cells. 17 Following that observation and using those 18 techniques, there was a rapid development of and 19 demonstration that one could transfer viral DNA. One 20 could transfer the oncogenes of transforming viruses 21 and induce transformed cells using purified DNA, but 22 it was in 1979 that the observations out of Weinberg's 23 lab and also out of Mike Wigler's lab that it was 24 possible to transfer the transformed phenotype from 25 chemically transformed mouse cells to recipient cells 168 1 through transfection and through the transfer of DNA 2 and chromatin. 3 And following that, the demonstration, 4 again, by Weinberg, Cooper, and Wigler's lab that one 5 could do this from human tumors and that, indeed, the 6 material that one was transferring was the identified 7 oncogenes, such as ras. 8 These observations and this technique led 9 to a great outburst of attempts to use this technique 10 to try to evaluate whether all tumors or all DNAs were 11 capable of doing this and whether one could use this 12 technique to identify and isolate transforming genes, 13 and as probably everyone is aware, this went on for 14 some time with a varying degree of success. 15 The technique was based on the observation 16 that if one transferred this DNA onto NIH3T3 mouse 17 fibroblast, one could see foci of morphologically 18 transformed cells, some better, some easier to see and 19 some less easy to see. As we pointed out, 20 occasionally normal cells would also transfer this 21 activity. 22 But the crucial factor here in 23 transferring the human DNA was the fact that within 24 these foci, one could identify the presence of human 25 genetic material using repeat probes, and furthermore, 169 1 that one could take these initial foci, take the DNA, 2 transfer this again, and retain a defined set of human 3 sequences which eventually could lead to the 4 identification of specific transforming genes 5 associated with those sequences. 6 So the initial assay was fairly simple 7 using the calcium phosphate procedure, transfecting 8 that onto NIH3T3 cells with various treatments to 9 enhance that DNA uptake, incubating those cells until 10 one sees foci of morphological transformation. 11 The initial results, however, were 12 relatively -- this process was relatively inefficient, 13 and so a lot of people set out to try and make this a 14 little more efficient to try to enhance it, and we and 15 Wigler's group, as well, devised an assay which was 16 called the nude mouse tumor assay to try and increase 17 the sensitivity of this kind of assay, and this simply 18 summarizes the sort of final version that we and a lot 19 of other people over the years have used, and that is 20 to do the transfection of high molecular weight DNA 21 from tumors or tumorigenic cell lines, in the presence 22 of a selectible plasmid, DSV2 neo, and to select the 23 cells which had taken up DNA by selection for neo 24 resistance to accumulate a population of cells, all of 25 which contained at least some human DNA. 170 1 Those pools of colonies could then be 2 trypsinized, injected into mice to generate tumors, or 3 could be replated and looking for foci. The advantage 4 of the tumor assay at the time was that this was a 5 more biologically significant endpoint. Furthermore, 6 it appeared to be a bit more sensitive than looking 7 for foci. It avoided a heck of a lot of tissue 8 culture and time staring at microscopes, and it did 9 pick up some transforming genes that couldn't be 10 recognized by morphological transformation since they 11 induce relatively poor morphological transformation. 12 Again, the products of these would be 13 reisolated, checked for the presence of human DNA, 14 transferred to a second cycle in order to validate 15 that, indeed, this was a genuine transfer. 16 The critical features of this that we 17 found over the years were really, as has been alluded 18 to, the state of the DNA. The DNA appeared you have 19 to have it as relatively high molecular weight. 20 Fragmentation of any significant amount destroyed the 21 majority of the activity in most cases, although with 22 the appropriate restriction enzymes you could retain 23 activity with fairly small pieces due to the size of 24 the genes, as it turns out, you were transferring. 25 The recipient cells, however, in general 171 1 had to be 3T3s. They were ideal because they had a 2 relatively low spontaneous transformation rate. If 3 one used seed cultures and controlled the serum and a 4 variety of other variables, one could retain their 5 normal state. 6 However, they could be transfected stably 7 to high levels, and this appeared to be critical 8 because we tried a number of other cells, and although 9 we could in some cases transfer genomic DNA, the 10 efficiencies were low, and the amount of DNA the cells 11 retained was generally much lower than what we saw in 12 3T3. 13 So in general, this was the cell line of 14 choice that everybody has tended to use. 15 Now, in using the -- the problem both in 16 foci and in tumors was that you had a background, and 17 this simply sort of summarizes some data that we 18 accumulated over the years of controls. Most of these 19 were transfected either just with DSV2 neo or in many 20 cases with human placental DNA as a control, and the 21 percentage of tumors which arose within the first ten 22 weeks after injection of around three million 23 transfected cells back here turned out to be around 24 ten or 11 percent. 25 The feature of these tumors is we could 172 1 never recover human DNA from these in cases where we 2 had transfected human DNA, and we were unable to 3 retransfect this and generate a second round. So 4 these appeared to be the kind of spontaneous 5 transformance that Dr. Rubin talked about earlier in 6 this meeting. 7 So with that, however, there were a number 8 of cell lines clearly, which everyone is well aware, 9 which contained transforming DNA where one could see 10 foci and one could get tumors in a tumor assay which 11 arose relatively early, and one could get secondaries 12 by taking the primary tumor of these cell lines and 13 retransferring them. 14 It's interesting that most of these, with 15 the exception of MNG HOS are ras. All of the genes 16 recovered and the genes apparently responsible for 17 this transformation were ras. 18 This was a cell line which we used as a 19 control. It's a single copy MOS transformed cell 20 line, and it allowed us to make an estimate of really 21 how efficient this was. 22 We knew if we took 100 micrograms of 23 genomic DNA we could generate about a single focus, 24 which gave us an efficiency of around ten to the minus 25 seventh to ten to the minus eighth. 173 1 If we did the reconstitution experiment 2 with a clone plasmid DNA and at the maximum 3 efficiency, we could get the same single focus with 4 around ten to the minus fourth micrograms, again, 5 roughly within certainly orders of magnitude, the same 6 kind of efficiency. 7 But the efficiency is very low. It's a 8 lot of DNA being put on these cells to get very few 9 transformants, and given that we were pushing this 10 assay and others were pushing this assay, it became 11 clear that we could generate transforming genes out of 12 things that weren't transforming genes in our original 13 tumors. 14 In our case we took a cell line called OV 15 CAR 3, generated an active transforming agent which, 16 upon analysis and characterization, turned out to be 17 a fusion between sequences on Chromosome 8 and 18 sequences in the 9P23-24 on Chromosome 9. These 19 sequences were formed during the transfection. They 20 could be continuously transferred, but they were not, 21 indeed, transforming sequences in the original tumor. 22 And as more and more people pushed this 23 assay to many of the genes, which were recovered from 24 this kind of transfections, turned out to be fusions 25 or transfection generated fusions or activations, 174 1 deletions of control elements, other aspects. 2 It's interesting that at least in our 3 hands we only saw these kinds of things with DNA from 4 tumors, not with human placental DNA, which we ran 5 extensively as controls. Whether that has something 6 to do with the methylation state, which was referred 7 to earlier, I don't know, but the generation of sort 8 of spontaneous activation seemed to be more prevalent 9 in tumors. 10 However, in terms of what was negative, 11 which was part of why I was originally asked to do 12 this, because we had a lot of negative, a number of 13 cell lines pretty capable of forming tumors were 14 repeatedly negative in our assays, in addition to what 15 we expected the human fetal lung, human placenta, and 16 others have shown that cell lines like VERO and MRC5s 17 are also negative or have been negative when pushed 18 into these assays and assayed in this way. 19 However, if we summarize all of the 20 analysis that we did, we looked at about 27 cell lines 21 and 52 primary tumors and recovered a total of around 22 nine positive transfections when all was said and 23 done. Most of those are ras, five of the cell lines, 24 both of the tumors. One of them was nep and one ovc, 25 a gene which we called ovc. 175 1 I also added, because this recently 2 appeared in National Journal of Cancer a summary from 3 Dr. Janssen, where he had summarized a similar set of 4 experiments using a relatively similar assay protocol, 5 129 tumors and cell lines, most of which turned out to 6 be ras. A number of others were genes that he also 7 identified. A large percentage of these were 8 transfection generated recombinants and activations. 9 So the conclusion, I think, from our data, 10 from the data of others over the years was that this 11 was not an easy thing to do, to transfer 3T3 cells 12 with genomic DNA, and while it does detect oncogenes, 13 it does so at relatively low efficiency and most of 14 them are ras. 15 Now, as we look back now, from what we 16 know about the situation, it is perhaps not surprising 17 that that's the case because as we all know and as 18 we've heard, it takes a number of changes to make a 19 cell become a transformed cell, and this is simply 20 taken from the recent paper in Nature by Weinberg in 21 which he essentially took a normal human fibroblast to 22 a transformed cell by making the addition of 23 telomerase, large T antigen, and ras in order to make 24 the escape from senescence to get through crisis to an 25 activated series of tumor suppressors, and finally to 176 1 activate the mitogenic pathways. 2 The cell line that we used for most of 3 these studies, NIH3T3, appears to have a P16 4 inactivated, and certainly that might tend to bias 5 what genes we were able to detect, but in any case, 6 the detection of transforming genes and the ability of 7 DNA to transform 3T3 cells was relatively low even in 8 the situation where we were talking about optimizing 9 our conditions and optimizing the system in order to 10 detect transforming DNA. 11 So certainly a fraction of tumorigenic 12 cell lines were able to transform 3G3 cells. A large 13 number of them, however, a large percentage, probably 14 70 to 80 percent were not. 15 Rearrangements during transfections, 16 however, could also generate transforming genes, and 17 they've contributed to this, but the frequency even 18 there, again, was fairly low. 19 So even under optimal conditions for 20 transfer, for getting DNA into cells, the ability to 21 transfer this DNA into transformed cells is very low, 22 and since there's multiple genetic alterations in 23 rodents and human cells in order to generate these 24 cells, that's probably responsible in part, despite 25 the fact that we were getting as much uptake as we 177 1 could, why there's a relatively low frequency of 2 transformation. 3 And so the basic conclusion seemed to be 4 that, yes, the DNA was transforming, but, no, it's a 5 difficult thing to do even when you try and bias the 6 system in any way, and when you go to primary cells 7 and go to reduced levels of DNA, reduced levels of 8 uptake, the frequency became at least in our hands 9 and, I think, everyone else's as relatively 10 undetectable. 11 So I will stop at that point. 12 (Applause.) 13 DR. ZOON: Thank you very much for that 14 excellent overview. 15 Please identify yourself. 16 DR. FRIED: Mike Fried. 17 It's interesting. We note today that 18 NIH3T3 cells are not only missing P16, but they're 19 missing this P14 arf, which Jim McDougall referred to, 20 and arf is a very interesting protein because it's 21 upstream of P53, and it's actually only induced by 22 oncogenes, by myc, by ras, E1A, and others, it then 23 turns on P53. 24 So if the arf NIH3T3 cells are arf 25 negatives, and as most cell lines partially have a lot 178 1 of changes, but without arf, it allows oncogenes to be 2 expressed. 3 Well, in normal cells if these oncogenes 4 came in, P53 would be turned on. The cell would go 5 through apoptosis and be taken out as a defense 6 mechanism. So I think that explains why NIH3T3 were 7 so sensitive for those detecting oncogenes, and most 8 of the cell types aren't. 9 DR. BLAIR: No, I think that it's clear 10 that each cell with the pattern of inactivations and 11 3T3 having the particular set that it did probably 12 made it possible to do this, although I think the 13 point that they do incorporate a large amount of 14 incoming DNA and then take some up also contributes to 15 the fact that you were able to do these experiments. 16 DR. FRIED: Sure, but I'm just saying that 17 most cells picking up oncogenes would turn on P53 and 18 be taken out of the population. 19 DR. KRAUSE: Phil Krause, FDA. 20 Don, have you done experiments in which 21 you directly injected DNA from tumors into animals? 22 DR. BLAIR: We didn't try to directly 23 inject the DNA. We did some experiments directly 24 injecting cloned DNA, and that in our hands was 25 relatively clone V MOS DNA at the time, and those 179 1 experiments were not terribly efficient, and so we 2 never really tried with genomic using tumor DNA 3 directly into mice, no. 4 DR. KRAUSE: And I think I and a lot of 5 people are intrigued by the Weinberg experiment that 6 you summarized. Would you care to speculate? What do 7 you think would happen if you tried to inject those 8 three DNAs directly into an animal? 9 DR. BLAIR: I mean, I guess my feeling 10 would be that the probability of getting all three 11 into the same cell might be probably fairly low, and 12 whether you could see the responses that he sees in 13 cell culture, I would think it would be unlikely, but 14 you'd have to do it. 15 DR. KUNG: I'd like to make a -- oh, go 16 ahead. 17 DR. ZOON: Please identify yourself 18 though. 19 DR. KUNG: Oh, sorry. Hsing-Jien Kung 20 from UC-Davis. 21 I'd like to make a comment because from 22 Dr. Petricciani's talk, he mentioned the injection of 23 V-SARC DNA can cause tumor, and in fact, I did that 24 experiment, and had I known that I would be invited to 25 this FDA meeting 18 years later, I would have kept 180 1 that chicken longer, but I didn't. 2 (Laughter.) 3 DR. KUNG: But the one point I'd like to 4 make is that that experiment in that paper, we also 5 said showed that the tumor actually regressed. All 6 the tumors inject with V-SARC DNA, tumor regressed, 7 and we felt at that time was due to immune response 8 because we did not have nude chicken. We had the 9 immunocompetent chickens there. 10 And so it's really not a cancer in that 11 sense. All the tumors regress. Whereas if we 12 injected the entire rsv DNA, okay, doing direct 13 injection, then the tumor will, in fact, cause 14 mortality of the chickens. Okay? So there is a 15 difference. 16 And I'll finish the story. We never 17 published because it was a negative result. So we did 18 inject C-SARC, and C-SARC never caused tumor. The 19 assay was very sensitive in the sense that if you have 20 a point mutation 527, the C-SARC point mutants, that 21 can cause tumors. 22 And so the assay was very sensitive, but 23 it would pick up something that I would not call it 24 cancer. Okay? 25 DR. ZOON: Thank you for that comment. 181 1 The last comment/question. 2 DR. BERKOWER: Yes, Ira Berkower. 3 One of the problems is I see a contrast 4 between what you said and what Dr. Petricciani said, 5 and that is the relative risk between tumor DNA and 6 normal DNA. You had no transforming events with 7 normal DNA. So can you put an upper limit on the 8 relative risk due to the tumor DNA versus normal or 9 lower limit maybe? 10 DR. ZOON: Could we have the mic, please? 11 DR. BLAIR: I think it works now. 12 (Laughter.) 13 DR. BLAIR: I think, you know, what we saw 14 was that when we did normal DNA, placental DNA and at 15 least one or two human normal cell populations, we 16 didn't get any tumors or foci that contained DNA that 17 we could transfer. We were doing on the order of 300 18 micrograms of DNA at a shot. We probably did the 19 placental DNA ten or 20 -- maybe ten times. 20 So in that sense, we never saw anything at 21 that level that we could verify that there was a focus 22 or a tumor that was due to the transfer of the DNA. 23 So that's the only limit I could put on it. 24 DR. BERKOWER: But would you say tumor DNA 25 is at least one log more tumorigenic? It seems to me 182 1 that it is just from what I've seen in here. 2 DR. BLAIR: Well, if you talk about the 3 active ones, yes, but if you talk about the other ones 4 that are not active in our hands, no, they were no 5 more active than -- we didn't see anything. So -- 6 DR. BERKOWER: And, secondly, you know, 7 yesterday we were trying to get at this question 8 whether some tumors are more worrisome than others. 9 It would also seem from your data, as opposed to what 10 was being said on the podium yesterday that some 11 tumors are much more worrisome than others. Others 12 are almost not worrisome. 13 This would be towards getting a molecular 14 discriminator of the risk. 15 DR. BLAIR: Well, I think, you know, this 16 assay clearly is biased, and so one could argue that 17 a tumor that has an activated ras gene would be more 18 likely to be detected in this assay. In that sense, 19 tumors that had activated ras genes were more readily 20 detectable. 21 In some of those there were amplifications 22 that may have helped to increase the sensitivity of 23 the assay, but this is, I think -- you know, this is 24 within a very specific set of conditions designed to 25 optimize the detection of, you know, morphological 183 1 transformation of NIH3T3 cells. 2 DR. BERKOWER: And lastly, one thing I 3 used to take comfort in was this requirement for very 4 high integrity, high molecular weight DNA, but in 5 fact, that may be entirely spurious. That may be 6 almost an artifact, again, of the method because if 7 there is more than one step needed, all you're really 8 measuring there is how far those two genes are apart 9 from each other. You just want to get them into the 10 same cell. You're not actually measuring anything 11 about the DNA. You're measuring something about the 12 genes, that they're far apart. So they need to be on 13 the same Ps or something like that. 14 DR. BLAIR: Well, the 20 kilobase size, 20 15 to 30 kilobase size, is pretty small on the genomic 16 scale. So they're pretty close together. I don't 17 think the linkage is probably something that you would 18 -- that would be a problem in this particular kind of 19 experiment. 20 DR. ZOON: Okay. If we could have one 21 last comment, and then we need to move on. 22 DR. HUGHES: Yes. Let me respond by 23 saying that you need to take into account that even 24 though the DNA comes from an abnormal cell, if the 25 transforming principle you pick up is not present in 184 1 the cell DNA that you began with but was created by 2 transfection, those are events that will happen with 3 normal DNA, and the reason that those are seen with 4 DNA from tumor lines or cell lots and not with normal 5 DNA represents the very large additional number of 6 experiments that occurred. 7 But those are certainly events that will 8 take place with normal DNA if you do enough of it. 9 The other important point is in trying to 10 put some sort of form or assessment of risk, what 11 you're really measuring in these assays is the risk to 12 3T3 cells, which is not the same thing as the risk to 13 cells that are not immortalized and don't have lesions 14 in them which do not make them necessarily oncogenic, 15 but which certainly make them immortal and, as a 16 consequence, easier to get into a state in which you 17 can measure tumorigenicity. 18 So I don't think you can reasonably use 19 this as a measure of either the safety of normal DNA 20 or the danger of DNA from a transformed cell. 21 DR. ZOON: Thank you very much for that 22 comment. 23 And thank you, Don. 24 Our next speaker I'm very pleased to 25 introduce. We've introduced her many times at 185 1 workshops, is Dr. Ruth Ruprecht, who is from the Dana- 2 Farber Cancer Institute, and will speak to us this 3 afternoon on infectivity of retroviral provirus DNA. 4 Thank you for coming. 5 DR. RUPRECHT: Thank you for inviting me. 6 DR. ZOON: The lights, please. 7 DR. RUPRECHT: I'm speaking about the 8 infectivity for retroviral DNA. My laboratory came 9 upon this issue in a somewhat surreptitious way. 10 PARTICIPANTS: Microphone. 11 DR. ZOON: Switch it on. 12 DR. RUPRECHT: Is that better? 13 Our experiments started with a mutant of 14 SIV made by Ron Desrosiers' laboratory called SIV 15 Delta 3. This mutant is built on the backbone of SIV 16 MAC 239 and has three big deletion, one in the vpr 17 gene, and two in the nef gene. The third deletion 18 also overlaps with the LTR, the negative regulatory 19 element in it. 20 When Ron's group used this virus in 21 juvenile and adult rhesus monkeys, it was attenuated, 22 and it turned out to be protective in about 50 percent 23 of the recipients upon challenge with wild type nef- 24 plus virus. 25 We wanted to develop a neonatal macaque 186 1 monkey model, and we gave these neonatal animals the 2 live SIV Delta 3 orally. As you know, we have a 3 rather surprising result which we reported in 1995. 4 These animals developed AIDS that looked for all 5 intents and purposes exactly like the disease induced 6 by the wild type virus. 7 Follow-up data then showed that the SIV 8 Delta 3 was 100 percent pathogenic in the initial 9 cohort. We were obviously concerned that we may have 10 had a contamination reason, immunosuppressive virus, 11 and we ruled out any known immunosuppressive virus in 12 this initial series of animals. Therefore, we knew 13 that there was no wild type SIV, and we also knew that 14 there was no simian retrovirus Type D present in the 15 animals to explain this result. 16 But we couldn't test for an adventitious 17 agent that may have been present in the initial CMX- 18 174 cells, and we also could not test for an 19 adventitious agent that may have been present when we 20 grew up the virus in rhesus macaque PBMC. 21 Therefore, we decided to go to pure DNA, 22 and we felt that the truth would be in the DNA. 23 The initial four animals are shown here. 24 The first three got cell free virus. The fourth one 25 got infected blood. 187 1 We decided to directly clone the virus out 2 of animal 94-1 for 80 weeks post inoculation. At this 3 time this infant already had immunodeficiency. We 4 were able to clone virus out without any intervening 5 amplification in vitro. 6 So what we did was we took PBMC and cloned 7 the provirus genome in two halves, the five prime half 8 and the three prime half, and then reconstructed a 9 replication competent virus. 10 In doing so -- next slide, please -- we 11 realized that 90 percent of the proviral halves that 12 we cloned out were actually defective, and only a 13 small minority of the seemingly clones of right sides 14 were actually giving rise to replication competent 15 viruses. 16 Now, in order to continue studies with 17 proviral DNA, we needed to first conduct a positive 18 control experiment. We needed to make sure that wild 19 type SIV MAC 239 would yield similar virus replication 20 kinetics and disease in rhesus macaques. 21 Afterwards our plan was to then take 22 proviral DNA of SIV Delta 3 of the virus that we 23 cloned out of the animal and compare the 24 pathogenicity. 25 Earlier this year we published the results 188 1 that we obtained with the parental SIV MAC 239. I 2 will go through these results really quickly. 3 Intramuscular inoculation of 400 4 micrograms of DNA in plasmic form into one infant, and 5 500 micrograms of DNA into two adults resulted in 6 systemic infection, persistent viremia, and death 7 eventually in all three animals. 8 The viral loads in the infant and in the 9 adult were actually reasonably similar to what you're 10 seeing when infections are conducted with cell free 11 virus. 12 The same goes for cell associated proviral 13 DNA load. The kinetics are similar to what has been 14 published by other groups with cell free virus, and 15 the same statement also holds for infectious cells in 16 the PBMS. 17 So intramuscular inoculation of super 18 coiled plasmid DNA and coding full length SIV MAC 239 19 is, first of all, reproducible. We were able to 20 maintain the provirus stably in a single plasmid. 21 The three animals became viremic and they 22 all succumbed to AIDS after longer term follow-up, and 23 after this series of experiments, we felt that 24 intramuscular inoculation of cloned proviral plasmid 25 DNA is reproducible and would be a wonderful tool to 189 1 study viral pathogenesis because if you rely on pure 2 proviruses in plasmid form, it can easily quantify 3 your inoculum, and if you're looking at viral mutants, 4 you can eliminate any parameters that may affect the 5 infectivity of your inoculum. Standardization, in 6 other words, is very easily achieved with super coiled 7 plasmid DNA. 8 So after this series of controlled 9 experiments were finished, we then took the provirus 10 that we got out of infant 94-1. This provirus is 11 called SIV Delta 3-plus. We inoculated a series of 12 newborn rhesus macaques and adult macaques with this 13 SIV Delta 3-plus in DNA form and compared it with SIV 14 Delta 3, the parental vaccine strain. 15 First of all, infant VR-6 had high virus 16 load and died of AIDS. So the proviral clone, SIV 17 Delta 3-plus, is not only infectious and replication 18 competent in vivo, but it is also pathogenic, and it 19 shows very clearly that the progeny virus that was 20 isolated from the vaccine recipient is a pathogenic 21 virus. There was no other adventitious agent present, 22 and even with very sensitive PCR analysis, we've made 23 sure that there was no full length nef gene present. 24 Another interesting set of data when it 25 compared virus loads in adults inoculated with SIV 190 1 Delta 3-plus proviral DNA in comparison with adults 2 given the parental SIV Delta 3 DNA. It's actually 3 somewhat of a shocker. 4 The adult animals given the parental 5 vaccine virus DNA are shown in the dashed line. The 6 peak viral RNA levels were approximately 10,000 copies 7 per milliliter. The three animals then were able to 8 control the virus and were not viremic in long term 9 follow-up. 10 The picture in the adults given the 11 progeny virus that evolved in that long infant 12 (phonetic) is quite different. First of all, peak 13 YLRNA levels are two logs higher, and secondly, the 14 animals are persistently viremic a short time later. 15 This shows very clearly that the virus has 16 become much more aggressive as it adapted in the 17 infant, in the initial infant, and that what has been 18 happening in vivo was selection of the fittest. Even 19 though virus isolation in adult vaccine recipients of 20 SIV Delta 3 is negative, we have to know that the 21 vaccine strength persists in the lymphoid tissues. 22 The virus actually continues to replicate at all times 23 and uses the faulty reverse transcriptase to do so, 24 and as John Coffin has published and other groups have 25 published as well, lentiviral reverse transcriptase is 191 1 a lousy, error prone enzyme that generates a number of 2 mutants. 3 The virus is then selected for optimal 4 replicative capacity, and over the years a more 5 aggressive virus emerges, and then natural selection 6 in the host environment eventually will compensate for 7 the initial attenuation and replicative capacity that 8 was engineered into the vaccine strain. 9 I should say that the peak viremia that we 10 saw in these three adults is actually in the same 11 order of magnitude as the peak viremia seen with SIV 12 MAC 239. 13 Now, in terms of infectivity of DNA, a 14 number of questions should be addressed for which we 15 don't have experimental answers. First of all, what 16 is the minimal infectious dose for intramuscular 17 inoculation of cloned SIV MAC 239 plasmic DNA? 18 Second, which route of inoculation is the 19 most infectious? 20 And, third, what is the infectivity of 21 genomic DNA isolated from an SIV MAC 239 infected 22 animal? 23 Now, there are some indirect answers to 24 some of the questions regarding the infectivity and 25 the amount of DNA received to achieve systemic 192 1 infection. In 1991, Norman Leitwin and his colleagues 2 have published a series of an experiment that showed 3 the infectivity of a lambda clone containing the 4 provirus of SIV MAC 239. 5 Four cynomolgus macaques were inoculated 6 intramuscularly with 200 micrograms of this lambda 7 clone. Three out of the four animals were infected, 8 but actually long term follow-up data were not 9 provided, and we don't know whether these animals went 10 on to develop disease. 11 Now, 200 micrograms of the lambda clone is 12 equivalent to about 50 micrograms of our plasmid DNA. 13 So we were approximately ten times above the amount of 14 provirus that's used by Norman Leitwin in our own 15 experiments, which were conducted with 500 micrograms 16 of proviral DNA plasmid. 17 A number of publications have appeared 18 over the years showing the infectivity of DNA in 19 animals, and I've just summarized some of them for you 20 here. The references are in the Liske, et al., "AIDS 21 Research on Human Retrovirus," a paper that appeared 22 earlier this year. 23 I would like to focus on the SIV 24 experiments by Elizabeth Sparter because there is some 25 more quantitative information available in that paper. 193 1 But, first, polyoma virus. There is some 2 information about infectivity of plasmid DNA as 3 compared to cell free virus, and the infectivity of 4 the super coiled polyoma plasmid DNA is about four to 5 five logs lower than that of cell free virions for a 6 given route of inoculation. 7 For SIV Elizabeth Sparter and colleagues 8 have looked at intramuscular and intradermal 9 inoculation at various doses. The lower dose of 30 10 micrograms per animal was no longer fully infectious, 11 but we don't really have a threshold from these 12 experiments either. 13 To conclude, there is actually a positive 14 note also to the experiments I've described. If there 15 ever is a nonpathogenic, avirulent, live attenuated 16 AIDS vaccine, the way to go would be through 17 infectious proviral plasmid DNA because, first of our, 18 our experiments have shown that cell free virus can be 19 replaced by infectious plasmid DNA. This would bypass 20 the need for preparing virus stock in human neoplastic 21 T cell lines. Standardization is easy and is very 22 reproducible. 23 Systemic infection can be induced very 24 reproducibly. I can tell you that with 500 micrograms 25 in rhesus monkeys, given intramuscularly, 18 out of 18 194 1 animals became systemically infected. So this is 2 highly reproducible. 3 And lastly, infectious proviral plasmid 4 DNA is stable and can be stored very easily. 5 (Applause.) 6 DR. ZOON: Thank you, Ruth. 7 This presentation is now open for 8 questions, discussion. 9 Yes. 10 DR. MURPHY: I'm Brian Murphy, NIAID. 11 Have you actually taken SIV that's 12 integrated into genomic RNA and into genomic DNA, 13 taken the genomic DNA itself and not taken the 14 provirus out, clone it up, put it into lambda, put it 15 into provirus, and seeing if you can initiate an 16 infection with genomic DNA containing a complete copy 17 of SIV? 18 DR. RUPRECHT: No, we have not done this 19 yet, and to my knowledge, there are no reports of it. 20 DR. HUGHES: A brief comment to clarify a 21 misconception. 22 DR. ZOON: Please identify yourself. 23 DR. HUGHES: I'm sorry. I'm Steve Hughes 24 from ABL. 25 And that is the role of RT in engendering 195 1 mutations in any retrovirual life cycle is undefined. 2 It is unlikely that -- there are three enzymes that 3 are involved in the replication of the genome: RT, 4 the host DNA dependent polymerase, and the host DNA 5 dependent RNA polymerase. 6 While it's unlikely that the DNA dependent 7 DNA polymerase makes a significant contribution, the 8 relative contributions of RT and the host DNA 9 dependent RNA polymerase is completely unclear, and 10 although it doesn't really matter here, it's led to 11 some important misconceptions in terms of the drug 12 treatment for HIV. 13 DR. RUPRECHT: Well, what is known though 14 is that the mutation rate of lentivirus is orders of 15 magnitude higher than the mutation rate of DNA 16 polymerases. So that gives the virus a chance to 17 adapt, and as you have seen, it adapts rather quickly 18 over 80 weeks. 19 DR. HUGHES: The question is whether or 20 not it is RT that is responsible for the errors, and 21 it's entirely possible it's a cellular RNA polymerase 22 that's responsible for many of those errors. 23 DR. RUPRECHT: I think that's unlikely, 24 but that's my personal view. 25 DR. LINIAL: Maxine Linial. 196 1 Did you compare the sequence of the Delta 2 3-plus the whole cDNA and the Delta 3? And if so, how 3 many different changes were there in the genomes? 4 DR. RUPRECHT: We have sequenced the whole 5 Delta 3-plus, the entire genome, and we found some 6 very intriguing mutations, but right now without 7 sequencing more progeny viruses, you can't really see 8 yet which of the many mutations that we have seen are 9 the important ones, but that's what we're studying 10 right now. Stay tuned. 11 DR. PURCELL: Damian Purcell from 12 Macfarlane Burnet Center, Australia. 13 We did very similar experiments which I'm 14 going to present tonight, and I was curious with your 15 Delta 3 mutation, the U3 deletion particularly, did 16 you build that into the five prime end of the 17 construct? 18 Because when we failed to do that, we find 19 reversion of the virus very quickly, presumably a 20 contribution from the presence of high level of 21 plasmid DNA containing this element. 22 DR. RUPRECHT: Well, during the in vivo 23 replication, the deletions in the three prime end were 24 followed with PCR in all animals, and what we noticed 25 in every animal that went on to progress to disease 197 1 was an increasingly large deletion in the area of the 2 delta nef and delta nre. So, in other words, the 3 virus actually lost additional sequences. 4 Now, through the process of retroviral 5 replication and the mechanisms of the reverse 6 transcriptase, any deletions in the three prime LTR U3 7 region will automatically be copied into the upstream 8 element. So they will be present also in the five 9 prime LTR in the next round of replication. 10 So overall the Delta 3-plus virus lost 180 11 base pairs in the LTR, which then will also obviously 12 be lost in the five prime. 13 DR. ZOON: Thank you very much. We very 14 much appreciate it. 15 We'll now go on to the next paper by Dr. 16 Eugene Major of the National Institute of Neurological 17 Disorders and Stroke, and the title of his paper is 18 "Cells and Their DNA Containing Integrated DNA Viral 19 Genomes: Differentiated Phenotypes of the Human 20 Central Nervous System." 21 DR. MAJOR: Thank you. 22 I feel a little bit like Monty Python, I 23 think, standing here because I'm about to say, "And 24 now for something maybe not completely, but just a 25 little bit different." 198 1 I've been sitting here since Tuesday 2 night, and what I didn't want to do is to be too 3 redundant about talking about virus systems that I 4 think we're somewhat familiar with, but I would like 5 to talk to you a little bit about the human polyoma 6 viruses and the nature of the integrated DNA that we 7 can find in sequence in cells that they transform, and 8 so that would be a new virus we haven't talked about 9 yet. 10 And also then talk about -- and if we 11 could have the first slide, which is what it says on 12 this slide, and talk about cells we haven't talked 13 about yet, which are cells from the human central 14 nervous system, and a particularly unique cell line. 15 At least we think it's unique, that probably would fit 16 into that category of a design cell substrate, whether 17 it's useful for vaccine development or not I'm not 18 sure, but certainly it's one that has unique 19 properties, which was established by the introduction 20 of a viral immortalizing gene. 21 So we'll talk a little bit about copy 22 number of the human polyoma viruses in cells that it 23 transforms, and the ability to rescue the integrated 24 viral DNA from these cells. We won't talk about 25 integrated lentivirus in the human central nervous 199 1 system. We won't have time for that, and then we'll 2 go on to talk about a specific what we think is an 3 interesting cell line which we made a number of years 4 ago from the human fetal brain. 5 Just a few comments then. These 6 particular bullets here can describe SV40 DNA 7 integrated in human immortalized cells, and it's 8 relatively true, generally true for human cells 9 whether they are myoblasts, muscle cells, epithelial 10 cells, fibroblasts, central nervous system cells. 11 Generally speaking, and it's also true for 12 BK and for JC, the human polyoma viruses in cells that 13 it transforms, not human cells, but rodent cells. 14 And I should say just parenthetically here 15 that the human polyoma viruses are pathogenic agents 16 in the human population. DK itself is associated with 17 kidney infections. JC is associated with a 18 demyelinating disease. Both of diseases or both of 19 those viruses are associated with infections in 20 reasonably severely immune compromised individuals. 21 JC, for example, causes neurological 22 complications in approximately ten percent of all AIDS 23 patients. So it's a rather substantial infection. 24 In most of these cases, there is a low 25 copy number of the viral DNA which is integrated, and 200 1 frequently what we find in human cells, that it's 2 integrated in tandem copies, either head to head or 3 head to tail, and one to two viral copies. 4 It's usually located at a unique 5 chromosome. There aren't a lot of duplications there, 6 and it's been reported in the literature that SV40 7 DNA, for example, will integrate in human Chromosome 8 17, 12Q23. I'll show you evidence of 2Q35. Usually 9 it's a single integration site that takes place. 10 Integration -- this actually should say 11 does not select for a cell phenotype, and what I mean 12 here is that integration in a particular chromosome is 13 not pathomnemonic for a particular type of cell. In 14 other words, you won't get integration of these 15 particular DNAs. If it's a fibroblast in a particular 16 chromosome; if it's a myoblast it will be another 17 chromosome. That's not what we mean here. 18 The integration site, however, during the 19 life span of the cell line is generally stable. We 20 have not seem translocations, for example, of 21 chromosomal content from one chromosome to the other. 22 From very early passages to much later passage, the 23 site of integration of the chromosome remains the 24 same. 25 In all of these cases, if it's SV40 or if 201 1 it's JC or VK, the T protein is synthesized, and it 2 binds P53, the retinal blastoma gene products, and of 3 course, it interacts with the cell cyclins, and this 4 has been discussed quite a bit at the meeting already. 5 SV40 immortalized human cells, as Leonard 6 Hayflick has reminded us a number of times here, is 7 nontumorigenic in the conventional ways in which it 8 has been tested, in introduction into immunodeficient 9 rodent models. 10 Now, BK NJC, the human polyoma viruses, 11 will induce tumors in rodents, hamsters particularly, 12 rats and mice. JC itself, if inoculated 13 intracranially, will induce gliomas, Grade 4 14 glioblastoma multiform. In New World monkeys it's 15 perhaps the only good primate model for gliomas that 16 really exists. 17 But SV40 immortalized human cells is 18 generally nontumorigenic, and personally I like to 19 make the distinction as we discussed here in the 20 terminology between neoplastic, which is cancer 21 causing, and immortalized, which is a term that at 22 least we use to simply indicate that it has an 23 unlimited life span. 24 this is an archive slide. I had to dig 25 this out from the archives in my own collection here, 202 1 SV40 T protein and its immortalized human cells. All 2 of the cells contain T protein, and these can be 3 cloned out. As you can tell, this is a pretty old 4 slide because these are immunoprecipitation 5 experiments, simply demonstrating that the SV40 T 6 protein or JC or BK will immunoprecipitate along with 7 P53. 8 Now, let's turn our attention to, for 9 example, the way that we can look at the nature of the 10 integrated DNAs which are present in, let's say, BK 11 virus, the human polyoma virus, transformed hamster 12 tumor cells. 13 Now, before the advent of gene 14 amplification techniques and PCR, what we used to do 15 to be able to get these DNAs out of the transformed 16 cells is do cell fusions, and this is a technique that 17 we used to use a decade ago or so. 18 So if we wanted to take a look or even 19 grow BK virus, for example, what we would do is do an 20 experiment in which we would take BK virus transformed 21 hamster cells, and we'd physically fuse it in co- 22 cultures with a cell that BK is very permissive to, 23 which is human embryonic kidney. 24 And when we do that, you form 25 Heterokaryons and synkaryons, and you'd be able to 203 1 rescue the viral genome by excision, DNA replication, 2 and then the virus would grow, and in many cases a 3 long time ago, this is the way we used to propagate 4 these viruses. 5 When the fusions begin to form, for those 6 of you who may not be too familiar with that, you'll 7 get a heterokaryon that's formed, which is a single 8 cytoplasm which contains two nuclei, one of the 9 transformed cells, for example, and one of the 10 permissive cell. 11 Then as those two nuclei begin to fuse, 12 what happens is that the viral DNA which is integrated 13 in a particular single chromosomal location site is 14 excised. We don't know what the mechanism of that is, 15 but here you see actually a synkaryon, which is a 16 fused cell in which you have the nuclear content, the 17 chromaton from two of the different cells, and viral 18 DNA begins to replicate after it's been excised. 19 Now, we measure replication or we detect 20 replication here by using biotinylated DNA probes 21 that's specific to the viral DNA, and the chromogen 22 here is diaminobenzidine, and in the presence of strep 23 avid and horseradish peroxidase at areas where the DNA 24 is replicating, which is, of course, in the nucleus, 25 then you'll get a brown precipitate from the oxidation 204 1 of the chromogen, which is diaminobenzidine. 2 We've quantitated this technique so that 3 we can tell when we look at these particular kinds of 4 nuclei that we have several thousand, if not more, 5 copies of the viral DNA which is present in the 6 nucleus of these cells, and it's a technique we've 7 used quite a bit for diagnosis of viral infections in 8 the human brain. 9 Now, what happens with, for example, if we 10 do this experiment now with BK, but let's say JC, and 11 JC grows very much more slowly even in a fused 12 culture, that actually the nuclei in which the DNA has 13 now been excised will replicate and begin to divide. 14 The cell begins to divide even before virions are 15 made, and that's what you see here, a very high copy 16 number. 17 These fused cultures will undergo cell 18 lysis with time, and we can go into these cultures and 19 we can go ahead and look at the viral DNA which is 20 present there, whether it's BK, for example, or 21 whether it's JC. These are simply the genetic maps of 22 all the primate polyoma viruses that are before you 23 with JC. 24 And interestingly enough, at least for us 25 in looking for BK and JC which has now been excised 205 1 from the transformed cells, that the DNA that we 2 sequence is perfectly excised from whatever region of 3 the genome that it was integrated, and integration I 4 should say within the viral genome is random also. 5 Now, in order for transformation to exist, 6 to take place, the integration usually takes place in 7 sequences obviously that are not involved with the T 8 protein coating region, the early region of these 9 viral genomes. So they'll be in the late region, in 10 the capsid coating regions here or, for example, in 11 here for JC or BK. 12 However, and they're linear obviously if 13 they're integrated within the chromosome. However, 14 upon excision, replication and virion production with 15 these DNAs, what the DNAs that we get out in the 16 nucleotide sequences are exactly the ones that we put 17 in. We don't really find any cellular chromosomal 18 DNA. 19 Now, we usually take extracted virion 20 populations so there could be a selection for 21 encapsidation of these human viral DNAs that come out, 22 and it's possible that there could be free viral DNAs 23 present in these particular kinds of fusion cultures 24 that have chromosomal DNA, which we wouldn't see by 25 the particular kinds of assays. 206 1 If, in fact, however, we take these 2 viruses and we grow them in permissive cells in 3 culture, there are a lot of rearrangements that take 4 place in the genome, and in nature this takes place as 5 well. There's a lot of genetic changes that take 6 place in BK virus as it grows in the human kidney or 7 in JC virus as it grows in the kidney or grows in the 8 brain. 9 But if the DNA remains integrated in one 10 of its transformed cells, there doesn't appear to be 11 any rearrangements that take place. 12 And that's also true if we take plasmids, 13 conventional colicin derived E. coli plasmids and we 14 transfect these DNAs into cells that are very 15 permissive, JC, into cells from the human brain, BK, 16 and to kidney cells. 17 Without excising the DNA before we do the 18 experiment, if we simply transfect these DNAs that 19 are infectious, even if the plasmid DNA is still 20 present, we just take the whole nine KB or ten KB 21 vector that we made, retransfect it into cultures. We 22 allow lytic infection to take place, for example, BK 23 and HEK cells, and then we sequence the DNA when it 24 comes out, and whatever that mechanism is for 25 excision, there are no plasmid nucleotides left as 207 1 well. So it's a very specific kind of mechanism. 2 So we're certain, of course, that the DNA 3 that is integrated into these particular transformed 4 cells is wild type DNA, and we can determine what 5 those nucleotide sequences are. 6 Now, I'd like to turn your attention now 7 to a particular kind of cell line that we made because 8 we needed to have in the laboratory a consistent 9 source of cells that we wanted to work with that would 10 be a cell line or an immortalized cell line, but 11 wouldn't have the probability of being able to excise 12 its DNA. 13 And so we were able to get a hold of one 14 of the origin infected mutants of SV40, which Jasha 15 Gluzman had used and he had created some years ago in 16 the creation of a series of cell lines called COS 17 cells. 18 And that I'll describe to you is really 19 the human fetal brain counterpart of the COS cell, and 20 the advantage that we thought we would have at the 21 particular time is to make an immortalized line of a 22 reasonably difficult cell to work with in the 23 laboratory derived from human fetal brain, that we 24 would not have to worry about viral DNAs excising and 25 replicating. 208 1 So we used an origin infected mutant of 2 SV40. It cannot replicate. The immortalizing gene is 3 SV40 T protein. The cell lines were established from 4 the CNS, and I'll talk to you about the differentiated 5 phenotypic properties. 6 I'll spend just a couple of slides here to 7 show you the way we handle these kinds of cultures,a 8 nd we have been culturing cells derived from human 9 fetal brain for quite a long period of time now as 10 targets for viral infection for human CNS diseases. 11 What you see here is a heterogeneous 12 population of cells from an eight week gestational 13 aged human brain. There are cells here which 14 areneuronal in definition, in glial, as well as 15 progenitor cells. 16 We have ways to be able to physically 17 separate out the different population of these cells 18 so we can grow glial cells which are predominantly 19 astrocytes and progenitors for the oligodendrocyte 20 (phonetic), which is the myelinating producing cell 21 with CNS. The astrocyte performs a variety of 22 functions in the human brain. It performs part of the 23 blood-brain barrier. It biochemically modulates the 24 number of neurotransmitters. It produces chemokines, 25 cytokines. It's a real work horse. 209 1 And actually the human brain consists 2 predominantly of the astrocyte. Fifty-five percent of 3 all of the cells in the human brain is the astrocyte. 4 About 35 percent are neurons. These are a neuronal 5 population. 6 This is just an Omarski optics of a phased 7 contrast, viable cells of the human brain. Here are 8 the cells that have attached to plastic, at the glial 9 cells, the cells here that have the round body and the 10 process bearing cells are progenitor neurons. 11 And viruses, different kinds of viruses 12 will go ahead and infect different cell types, what's 13 here. This is just another view of the same kind of 14 culture, and we phenotypically characterize these 15 cells as either being fliel or neuronal by different 16 markers. 17 And what you see here is a digitized image 18 of phenotypic markers in these cells that we grow. 19 The astrocyte, for example, in glial cells will 20 express an intermediate filament called glial 21 fibrillary eosinic protein, and that's what's marked 22 in the green here. The red actually in these cells is 23 marked by an antibody which we recently made to 24 another intermediate filament called nestin, which is 25 a marker for progenitor stages of differentiation. 210 1 It's not unique to the brain, although the nestin that 2 we used as an antigen to make this antibody we closed 3 from the sequences from the human developing train. 4 We can actually identify cell types here 5 that have both markers, have both nestin, for example, 6 and FGAP, and we know that these cells are in a 7 differentiated pathway to either go to a glial cell or 8 in other cases to go to a neuron, and so we 9 characterize these cells this way. 10 So we have this availability in the 11 laboratory. We were fortunate enough to be able to 12 get one of the origin defective mutants from Jasha 13 Gluzman, and for those of you who may not remember, 14 but what was done is that there was a restriction at 15 the nucleus site in the replication origin of the SP40 16 DNA. This is just in comparison to the human polyoma 17 viruses and the deletions made in that origin. 18 The one we used to introduce into human 19 fetal brain cultures had an 11 base pair deletion in 20 the origin here. So the virus was dead. It could not 21 replicate its DNA. It would never make proteins. 22 So we took that construct. We inserted it 23 into a 12 week gestational age, mixed culture of human 24 brain, and this is what we got. This simply goes 25 ahead and gives you something of an idea of the time 211 1 frame within which we did this from the initial time 2 of the transfection that was taking place. We did not 3 clone these cells. We simply collected those cells 4 which survived as a matter of growth. 5 And after a period of about five passages 6 or so in culture, they passed through, I guess, what 7 would be called the replicated senescence period or 8 M0, perhaps M1. All of the cells became T protein 9 positive. Those were the cells that grew out. They 10 did enter a crisis phase later on in passages 31 or 11 38. This would represent perhaps 80 cell doublings in 12 tissue culture, and then from then on, developed into 13 a permanent or immortalized cell line. 14 And these are just phenotypically what the 15 cells looked like. They have some interesting 16 biological characteristics, however. 17 All of these cells, regardless of whether 18 they're at this, the early stage, or the post crisis 19 stage are anchorage dependent for growth. They're 20 absolutely non-tumorigenic in rodents or in non-human 21 primates at least under the systems in which we have 22 inoculated them. They're contact inhibited for 23 growth. They're somewhat serum dependent for growth. 24 For our particular purposes in the 25 neurobiology of some of these cells, they are quite 212 1 interesting because they produced nerve growth factor, 2 brain derived neurotropic factor. They produce a 3 great amount of the vascular endothelial cell growth 4 factor. They're Class 2 positive, Class 1 positive, 5 CE-4 negative. 6 But when you look at the susceptibility to 7 virus infections, these particular cells are 8 susceptible to JC virus and to BK. They're also 9 susceptible to HIV-1, HTLV-1 as well, to all of the 10 herpes viruses. Well, we haven't tested EBB, but 11 certainly CMV, herpes Type 1 and Type 2, varicella 12 zoster virus, which isn't on this list, as well as 13 vesicular stomatitis virus, which can be pseudotyped 14 into HIV lentiviral vectors. They're actually 15 exquisitely sensitive to those. 16 So all of these cell types or these cell 17 types are sensitive to a variety of different kinds of 18 viruses. 19 Now, we cloned out -- the fact that it was 20 non-tumorigenic, there was nothing we could do with 21 these cells to make them tumorigenic. So we thought, 22 well, there might have been mutations in the T protein 23 coating region of the DNA that was inserted into the 24 chromosome of these particular cells. So we cloned 25 out just the T protein coating sequences, and that's 213 1 what you see here in this particular slide. 2 And we found actually there were five 3 nucleotide changes in the DNA in these cells compared 4 to the DNA that we initially put into the transfection 5 experiment, and some of these nucleotide changes would 6 predict differences in amino acid calling for the 7 protein, but we haven't sequenced the protein, so 8 we're not sure. 9 But they don't really fall in locations 10 which could tell us much about any differences in the 11 biology of what T protein may be doing in these cells, 12 but at least we know that, and we do know that in FSH 13 staining, which was done by Tom Glover at Ann Arbor -- 14 you probably can't see that very clearly here -- that 15 there is a single integration site of the SV40 DNA in 16 Chromosome 2Q35 location, here, and by sequencing 17 analysis, we know the location along the SV40 DNA, 18 which that integration had taken place, and we also 19 know that it's in a direct tandem type of copy. 20 Now, we began to be interested in this 21 cell line for a whole variety of other purposes here, 22 and in collaboration with many other laboratories that 23 actually have this cell line, because we made this 24 cell line back in 1985. 25 I will go through this relatively quickly 214 1 here. 2 If we continued to grow the SV40 T protein 3 immortalized human fetal brain cell line, which we 4 call SVG, in standard laboratory medium E-MEM in fetal 5 bovine serum, it's phenotypic characteristics are that 6 of a glial cell, and these are simply phenotypic 7 markers that we use by flow cytometry or staining in 8 order to be able to define the type of cell that it 9 is. 10 They have fragments of cholera toxin, 11 cholera toxin and tetanus toxin, another surface 12 marker, which is gliactocil sulfatide called A2B5 on 13 their surface, and this really defines a type of cell 14 which is either terminally glial or it's a progenitor 15 to a glial cell. 16 However, if we treat this cell with 17 molecules that induce cyclic EMP, for example, like 18 forscalin (phonetic), and on the next slide I'll show 19 you that, or if we take these cells off of serum and 20 put them in medium which we use in the laboratory as 21 well as other laboratories use that are selected for 22 neurons, that we get a shift in phenotype. 23 And what happens is that we lose the A2B5 24 population. We increase the cholera toxin population, 25 and we also increase the number of cells which are 215 1 cholera toxin/tetanus toxin positive in the neuronal 2 phenotype of these cells. 3 Cholera toxin/tetanus toxin denotes a cell 4 which is in a neuronal lineage, and this was something 5 of a surprise to us because we had thought that after 6 so many passages in culture, that in fact this would 7 be a stable phenotype of these cells, and the presence 8 of the SV40 T protein would lock those cells into a 9 certain time in which they were immortalized. 10 DR. ZOON: Could you please summarize? 11 DR. MAJOR: Okay. I'll skip that then, if 12 I have to summarize this, and I'll say, for example, 13 these are simply cells that we can treat to force them 14 into a neuronal phenotype. This is what the cells 15 look like, for example, if we put them on neural basal 16 medium. These are the cells as they look like if 17 they're in glial medium. In the presence of serum, of 18 course, they will go ahead and divide. 19 If we infect these cultures with JC, for 20 example, you can't see that, but JC will infect these 21 cells into glial phenotype because JC is tropic for 22 glial cells, and it's at a molecular level. 23 However, if we take these cells and we 24 induce the neuronal phenotype, this is the same parent 25 cell; then infect those, then JC will not infect that, 216 1 and it's not a matter of attachment or entry because 2 we've already shown that, in fact, susceptibility to 3 JC infection of these particular cells is at the level 4 of transcriptional control, a very much different type 5 of cell. 6 So let me then summarize by saying at 7 least that here what is surprising to us is that after 8 many passages in culture, 50, 60, 70 -- some other 9 laboratories have passed this out 150 times -- what we 10 seem to have been able to do is to immortalize a 11 population of human central nervous system progenitor 12 cells. 13 That these cells still are multi-potential 14 for the different phenotypes of the cells that we can 15 grow either in a glial lineage or in a neuronal 16 lineage, and we use these now back and forth in this 17 particular lineage patterns to ask questions of 18 neurobiology, to ask questions of cell susceptibility 19 to the human viruses, and for a variety of other kinds 20 of questions that can be asked. 21 And, again, the surprising thing is that 22 even in the presence of SV40 T protein, which is 23 continually made in these cells, there's still multi- 24 potential for differentiation, and they respond to 25 signals that the human brain responds to also in 217 1 lineage pathway commitments as it develops. 2 And so this is actually from page 19 of 3 the information that Andy Lewis had sent out to all of 4 us, talking about perhaps getting to know a little bit 5 more about the progenitor nature of the cells and 6 designing specific kinds of cell lines that may or may 7 not be useful for certainly vaccine development, but 8 certainly cell lines that are extremely helpful for us 9 in understanding other kinds of questions at least in 10 neurobiology. 11 Thank you. 12 (Applause.) 13 DR. ZOON: Thank you. Thank you very 14 much. 15 If we have time for one quick question, 16 comment? No. Thank you very much. I appreciate it. 17 Our last presentation in this session is 18 by Dr. Haig Kazazian of the University of Pennsylvania 19 School of Medicine, who will speak to us on mobile 20 elements in mammalian genomes and their implications 21 for cell substrate safety. 22 Thank you for coming. 23 DR. KAZAZIAN: Thank you. 24 Well, I'm going to say some things about 25 endogenous DNA. I thought I'd start off by telling 218 1 you a little bit about retroelements and L1 biology, 2 and then I'll tell you a little bit about an assay for 3 retrotransposons that we use, and something about the 4 number of active human retrotransposons, and I'll try 5 to give you an estimate of the insertion frequency in 6 germ lines and tell you what I think we know about the 7 potential for insertion frequency in somatic cells. 8 So let me go ahead and start. 9 DR. ZOON: Do you want me to get this 10 slide? 11 DR. KAZAZIAN: There we go. Retroelements 12 are sequences in the DNA that have been put back into 13 the DNA by reverse transcription, and there are a 14 number of different types of retroelements, but 15 retroelements probably make up something on the order 16 of a third of human DNA. So a large fraction of human 17 DNA. 18 There are non-LTR, non-long terminal 19 repeat, or poly-A retrotransposons. They have no long 20 terminal repeats. They have a three prime poly-A. 21 They usually have two open reading frame, and they're 22 reverse transcribed by an unusual mechanism, target 23 prime reverse transcription or nick and prime 24 mechanism. 25 And an example of these is mammalian L1 or 219 1 line elements, and we'll be talking about those. 2 Long terminal repeat retrotransposons, we 3 don't have any of these in humans that we know of. 4 They have long term repeats. They have two open 5 reading frames. They have reverse transcription prime 6 by a TRNA, so very different from this group of 7 retrotransposons as far as their reverse transcription 8 goes. 9 An example of this group is TY1 of yeast. 10 Retroviruses, as you've been hearing a lot 11 about, have long term repeats. They usually have 12 three open reading frames. So they have an envelop 13 gene that gets them to be infectious. There's no 14 envelope gene in these retrotransposons. 15 Reverse transcription in retroviruses is, 16 again, primed by a TRNA, like that of LTR 17 retrotransposons. A good example is HIV-1 of human. 18 There are non-autonomous retroelements in 19 humans. These are transcribed, but contain no 20 functions for retrotransposition. Examples of these 21 are alu elements, small nuclear RNA pseudogenes and 22 process pseudogenes, and so they need reverse 23 transcription and help by other elements, perhaps non- 24 LTR retrotransposons. 25 So what we're talking about is 220 1 retrotransposition which is an element being 2 transcribed into an RNA, reverse transcribed, and 3 integrated at a new site in the genome. So this is a 4 copy and paste mechanism, duplicative mechanism for 5 actually expanding the genome. 6 And the major retroelement, the major 7 player in the human genome is the so-called L1 or line 8 element, and when I made up this slide, I said 9 approximately 100,000. There are new estimates that 10 say there are up to 400,000 or so of these line 11 elements in the human genome. These are database 12 estimates, which are probably more accurate now, now 13 that we have something like 12 to 13 percent of the 14 human genome in the databases. 15 The majority of L1s are truncated at their 16 five prime end. That means they're short, and they 17 have mostly three prime end sequences and not five 18 prime end sequences, but about 3,000 L1s are full 19 length, and some of these can retrotranspose. 20 So what does a full length L1 look like? 21 Here's the structure of that element. It's six 22 kilobases in size. It's got at the five prime end an 23 internal promoter sequence which is relatively poorly 24 defined at the moment, but is mostly present in the 25 first 100 base pairs. 221 1 It has a first open reading frame which 2 makes a protein which is likely to be an RNA binding 3 protein. It has a second open reading frame, which is 4 quite long, most of the element, and within this 5 second open reading frame is an endonuclease domain, 6 a sequence which encodes an endonuclease; a reverse 7 transcriptase domain which encodes a reverse 8 transcriptase, and a conserve domain at the C terminal 9 end, which may be a zinc knuckle sequence. 10 It's got a very short, three prime 11 untranslated region, and then it ends in a poly-A 12 tail, and it's usually flanked by target site 13 duplications. 14 So in the early 1990s, we were able to 15 isolate the precursors for a couple of insertions that 16 occurred in humans, and these full length precursors 17 were good candidates for active retrotransposable 18 elements. 19 So in the mid '90s, John Moran came into 20 the lab, and he devised this assay for L1 21 retrotransposition. So in this assay then a full 22 length element has in its three prime UTR inserted a 23 marker cassette or a retrotransposition cassette which 24 will tell us when a retrotransposition event has 25 occurred. 222 1 And this particular cassette has a 2 backwards neomycin gene which is split by a forward 3 intron. So here's the spliced donor site and the 4 spliced acceptor site of that intron, which is in this 5 transcriptional direction. 6 A cassette is also flanked by a promoter 7 for the neomycin gene and a poly-A signal for the 8 neomycin gene. 9 So when transcription occurs from this 10 promotor, we go through this intron in this direction. 11 We can then splice it out, and then with reverse 12 transcription, we can have this neogene now 13 transcribed and put back into the genome in the right 14 orientation now with no intron disrupting it and 15 driven by its own promoter, and having its own poly-A 16 signal, we can get G4 and 8 resistant cells. 17 So this, as it turns out, is the only way 18 you can get G4 and 8 resistant cells, is with a 19 retrotransposition event using this reverse 20 transcriptase. 21 So we put this receptor into a piece of 22 plasmid and a nice receptor zone goes into the cell 23 and into the nucleus and has about 20 or so copies, 24 and we infect Hela cells with the constructs by 25 lipofection, and we have on this P-sep (phonetic) 223 1 element a hygromycin resistant gene. So we hit the 2 cells with hygromycin. 3 And so we end up with cells that are 4 expressing the marked L1 that have been transfected, 5 and we outgrow and plate dilutions in medium 6 containing G41A, and the G41A resistant cells then 7 contain an L1 retrotransposition event. 8 And here is an example of the kind of 9 assay that we get when we put in an active element. 10 It turns out this was a precursor of an insertion into 11 a Factor A gene causing hemophilia in a child. 12 So you can see that this is with plating 13 out something like a million transfected cells and 14 getting something on the order of 1,000 positive 15 retrotransposition events. Each one of these blue 16 specks is a colony of cells which contains an 17 individual retrotransposition event, and that was 18 proven by cloning a number of these events and showing 19 that they look like the retrotransposition events that 20 are seen in an endogenous human situation. 21 Here is a construct that has a big 22 deletion in the element, and here is a construct that 23 simply has a point mutation in the reverse 24 transcriptase domain of the element. 25 I should say that we now have on the order 224 1 of nine of these active human elements in the 2 laboratory, and the extent of retrotransposition 3 varies considerably. So with some of those elements 4 we get much less retrotransposition than is seen here, 5 and with others we get much more. 6 So we have one element in the lab now that 7 will retrotranspose at a frequency of about one in 8 every ten transfected cells, a very high frequency. 9 I should also say that in other 10 experiments we've shown that the site of 11 retrotransposition is relatively unbiased. That is, 12 the insertions into introns and genes is about the 13 frequency that one would expect for the size of 14 intronic sequences or the proportion of intronic 15 sequences within the genome. 16 So it turns out that there have been on 17 the order of 13 natural L1 insertions in humans 18 causing disease, and it turns out that 12 of those 13 19 have come from a very small subset of L1 elements 20 within the genome. 21 And so we decided to see if we could make 22 an estimate of the number of active human L1 elements 23 in the genome, and we got Gary Swyrbl to do a count of 24 the number of full length elements of this subset in 25 the haploid genome. He came out with a number of 225 1 about 80. 2 We isolated a number of these subset 3 elements specifically from the genome and found out 4 that roughly a quarter of them were capable of 5 retrotransposition. 6 So from that we could make an estimate 7 that the number of active human L1 elements in the 8 genome was on the order of 40 per diploid genome, and 9 so you could make a little range on that estimate and 10 say perhaps 30 to 60 active elements in the genome, 11 and yet now we're sitting with hundreds of thousands 12 of copies that have built up over evolutionary time. 13 So I just wanted to give you an idea of 14 the number of recent retrotransposition insertions in 15 humans and mouse. I mentioned that L1 insertions, the 16 number here was 12. I know of another one now. So 17 let's say it's 13, and there have been a number of 18 non-L1 insertions, particularly alu insertions, in 19 humans, and we think that those are driven by L1s, 20 reverse transcriptase, but there's no direct evidence 21 for that yet. 22 But in any case, if we add those up, we 23 get on the order of 30 or so human insertions that 24 have been seen that cause disease. That's the only 25 way we would know of them, and as you'll see on the 226 1 next slide, that makes up something on the order of 2 one in every 600 mutations. 3 In the mouse it's very different. 4 Spontaneous diseases that have been characterized in 5 the mouse are on the order of only 200 or so, and so 6 the number of insertions or the frequency of 7 insertions causing mutations in the mouse is on the 8 order of ten percent, very much higher than it is in 9 humans. So I guess we're lucky in that regard. 10 So when does L1 retrotransposition occur? 11 Well, L1 RNA and the orf-1 protein have been found in 12 primary spermatocytes in leptotene and zygotene 13 stages. This is miotic prophase, and this is mouse 14 work. This is not human work. 15 Timing of retrotransposition events is 16 really unknown, but it's thought to be early or during 17 germ line production. 18 So we can make an estimate of the 19 frequency of retrotranspotion insertions in the human 20 genome. As I said, the number of known insertions is 21 on the order of 28. This was back in March I made 22 this estimate. It's now 31 that I know of, but the 23 number of mutations in the database is higher. 24 So the number of mutations in the database 25 when I made this calculation back in March was 16,650. 227 1 Okay? That was from the human gene mutation database. 2 So that gave us a fraction of mutations that are 3 insertions of 28 over 16,650 or about one in 600. So 4 most of the mutations in humans are point mutations, 5 but occasionally you get an insertion, one in 600 6 insertions. 7 So one can then go back and see, well, 8 what's the frequency of insertions in the average 9 genome, and one can do that by a couple of 10 calculations. 11 The estimated frequency of spontaneous 12 mutations in man is about ten to the minus nine per 13 nucleotide per year. So we can go from there. There 14 are three times ten to the ninth nucleotides per 15 genome, and I used here a generation time of 30 years 16 for another calculation. I said it was 25, but let's 17 say it's 30. It's a little long, but 30 years, and 18 times ten to the minus nine per nucleotide per year. 19 That's this number. 20 And that equals about 90 mutations per 21 haploid genome per generation, or let's go to that and 22 say it's 90 mutations per sperm producing an 23 individual upon fertilization. That's total 24 mutations, not insertions, total mutations. Okay? 25 Ninety per sperm. 228 1 So the frequency of retrotransposon 2 insertions per sperm would be one in 600. Remember I 3 said one in 600 is the frequency of insertions per all 4 of the mutations. One in 600 times that 90 or about 5 one in six. So about one insertion per six sperm. 6 If you use the 25 number for the 7 generation time, you get one in eight sperm. Okay. 8 So that's about one insertion, new insertion, in every 9 six haploid sperm. 10 Since these insertions are random, less 11 than five percent of the insertions should be 12 deleterious because exons comprise one percent of the 13 genome. Other sequences important for gene expression 14 likely account for maybe another one or two percent of 15 the genome, but it should be that most insertions 16 shouldn't cause a problem. 17 Okay. So how about somatic insertions and 18 what do we know about endogenous somatic insertions? 19 And the answer is not much. 20 Okay. So I know of one L1 insertion 21 that's likely to have been involved in the etiology of 22 cancer. There was one L1 insertion into an exon of 23 the APC gene, adenomenous polyposis coli gene reported 24 by Nakamura and his group about five years ago, and 25 that particular insertion was only seen in the tumor, 229 1 not seen in the constitutional tissue of the normal 2 colon, so the colon cancer tumor. 3 There was one alu insertion intoan MLVI-1 4 or 2 gene which was associated with leukemia. It was 5 in a cell line. As I recall, the patient was not 6 available for study. So it's not clear whether that 7 was a somatic event or a germ line event. 8 Okay. There have been a few germ line 9 events in tumors of alu insertions, but they're not 10 somatic events. 11 Okay. We do know that L1 transposes at 12 high frequency in Hela cells. I showed you that. All 13 right. So that's something. 14 And we do know from the work of Tom 15 Fanning that L1 is expressed in ovarian, testicular, 16 and breast cancers. Okay. So it's expressed in 17 cancers, and not all, but in some cancers. 18 So that about ends what I know. 19 Retrotransposon insertions occur in tumors with some 20 frequency, although their observation has been pretty 21 rare, and clearly retrotransposon insertions occur in 22 cell lines. We've seen that in the laboratory. 23 The incidence of endogenous somatic 24 retrotransposon events is unknown as far as I'm 25 concerned at this point. 230 1 That's all I have to say. Thank you. 2 (Applause.) 3 DR. ZOON: Thank you. 4 Go ahead, Phil. 5 DR. KRAUSE: Phil Krause, FDA. 6 That was a very nice talk. One mechanism 7 by which people would like to consider testing lines 8 for the presence -- cell lines for the presence of 9 adventitious agents is obviously highly sensitive 10 reverse transcriptase assays. I guess two questions 11 which are related. 12 One of them is if you were to test a 13 cancer cell line which you say might be more likely to 14 contain L1 elements, would that cause it to be 15 positive in a highly sensitive RT assay? 16 And I guess the next question is: is that 17 a good thing or a bad thing? On the one hand, does 18 that tell you that the cell line is more dangerous? 19 On the other hand, does it prevent you from detecting 20 or does it then make the assay less valuable because 21 that's a potential alternative explanation for a 22 positive RT result? 23 DR. KAZAZIAN: Okay. First of all, you 24 should do the assay in such a way that you might 25 detect an L1 reverse transcriptase so that it's going 231 1 to be different from a -- I mean you could do a 2 specific assay because the reverse transcriptase with 3 L1 is going to have different properties than a 4 retroviral reverse transcriptase. Okay? So you 5 should try to set it up as a relatively specific assay 6 for the L1 reverse transcriptase. 7 Secondly, you're going to have a hard time 8 finding it because people have had a very difficult 9 time finding endogenous L1 reverse transcriptase. 10 Okay? 11 So it's very likely that if you find it, 12 it's probably, you know, with your assay, without 13 looking for a specific L1 reverse transcriptase, that 14 it's going to be a retroviral reverse transcriptase. 15 DR. ZOON: Johannes. 16 DR. LOEWER: Johannes Loewer, Germany. 17 Perhaps I missed the point. Is the 18 transposition activity of the construct you have 19 described restricted to human cells or does it occur 20 in every cell? 21 DR. KAZAZIAN: Okay. I didn't mention 22 that, but we have shown that it is functional in mouse 23 LTK minus cells, even though it's a human element. 24 You can put the human element into mouse cells. You 25 can also put the mouse element into human cells and in 232 1 mouse cells. It's functional in LTK minus cells. 2 It's functional in hepatocarcinoma cells. 3 I heard from Maxine Linial's lab that it 4 was functional in quail cells? Okay. So, yes, it's 5 functional in quail cells. It's functional in Chinese 6 hamster ovary cells from another lab that has it. 7 However, we have not been able to get it 8 functional in differentiated fibroblast type cells. 9 Okay? That's what I mentioned this morning. 10 Yes. 11 DR. ONIONS: David Onions. 12 I really enjoyed the talk. I've got a 13 quick comment and then a question. 14 If I remember correctly, I think Don Cohen 15 showed that there were line elements in a tumor called 16 transmissible venereal tumor, which is a canine tumor, 17 and I think he's got a line item in the upstream of 18 myc, which is in the -- 19 DR. KAZAZIAN: Yes, that's correct. Okay. 20 DR. ONION: But just another example. 21 DR. KAZAZIAN: Right. 22 DR. ONIONS: But the question was that the 23 remarkable transposition frequency that you're seeing 24 in Hela cells. Did you follow the fate of any of 25 these clones that you had? And ask the question: was 233 1 there retransposition? And if so, what was -- 2 DR. KAZAZIAN: Okay. I'll tell you what 3 we have done though. We haven't done that because all 4 of the insertions that we saw were highly truncated 5 insertions. 6 DR. ONIONS: Okay. 7 DR. KAZAZIAN: So they were missing then 8 their five prime end, and so they would lack promoter 9 activity. Okay? 10 DR. ONIONS: Okay. 11 DR. KAZAZIAN: But what we have done, an 12 experiment where we've looked to see if there's, 13 quotes, heterogeneity of the Hela cells. So that we 14 have taken cells that have had a retrotransposition 15 event and compared those and looked at the frequency 16 of retrotransposition in those cells and in another 17 group of Hela cells that have not had a 18 retrotransposition event to see if the frequencies 19 were different, and we did not see the heterogeneity. 20 We did not see a different frequency. 21 So the one group of cells was not any 22 better for retrotransposition than the other. 23 DR. ZOON: Thank you, Haig. 24 We're going to take a 15 minute break. So 25 we're going to reconvene shortly after 3:45. 234 1 Thank you. 2 (Whereupon, the foregoing matter went off 3 the record at 3:34 p.m. and went back on 4 the record at 3:54 p.m.) 5 DR. LOEWER: Good afternoon. I think we 6 should continue with the program, the panel-audience 7 discussion. 8 First what I want to do is to ask the 9 speaker to come to the podium. 10 So I think we are now complete. 11 I would like to welcome you again and 12 shortly to introduce myself. My name is Johannes 13 Loewer. I'm Deputy Director of the Paul-Ehrlich 14 Institute in Frankfurt, Germany. 15 Paul-Ehrlich Institute is an institution 16 which has more or less the same duties as the Center 17 for Biologics Evaluation and Research. I'm, 18 therefore, also a regulator, and I will lead this 19 discussion from the point of view of a regulator, not 20 of a scientist, which is not much like exclusive, I 21 guess. 22 The organizers of this meeting have 23 prepared quite a number of questions which I have 24 posed to the chairs of the different panels, and 25 because the issue we have to deal with is really very 235 1 complex and all of the discussion that we have had so 2 far show that there are many different facets which 3 influence each other. 4 So to be as constructive as possible, I 5 thought it is maybe helpful to follow the questions 6 which have been posed directly to me, and copies of 7 this question have been distributed here. 8 And I would also like to try to simplify 9 the questions in order to be able maybe to come to 10 clearer answers wherever possible. 11 So I have made a number of slides from 12 these questions, and these slides should guide us 13 through the discussion. I would like to introduce the 14 question shortly, and then we could open the 15 discussion, which should not exclude if there are 16 other urgent questions and contributions. 17 So the first question posed to the panel 18 here is, again, of course, the risk of residual DNA, 19 residual DNA which comes from neoplastic cells which 20 would be used in viral vaccines, and there are a 21 number of questions. 22 Is there a special, I would say, 23 infectious risk event, a special neoplastic risk 24 event, and is there a risk for epigenetic changes? 25 So we should keep these questions apart, 236 1 I guess, and should maybe focus on the question of 2 infectious risk event. As we have discussed during 3 the last two and a half day, one and a half day or two 4 days, there are different types of infectious events 5 maybe. 6 There may be extraneous agent, and I would 7 say there is not a different risk between neoplastic 8 cells and diploid cells for extraneous agent, but we 9 should ask the question: does transformation increase 10 the risk for infectivity for maybe a broader range of 11 viruses or is this risk for extraneous agents really 12 similar for diploid cells or not, or does it really 13 depend on the virus? 14 That's one question we should discuss. 15 The second one is there may be, of course, 16 endogenous virus or endogenous element as we have 17 heard, and is there an increased risk for the activity 18 in neoplastic cells compared to diploid cells, and 19 perhaps we could discuss this question first. 20 First I would like to invite members of 21 the panel to give their comments to the questions I 22 have just posed. 23 DR. KAZAZIAN: As far as retrotransposons 24 go, there's no infectious risk. They don't go from 25 cell to cell. They're stuck in the cell in which 237 1 they're transcribed. 2 Neoplastic risk, hard to know. I would 3 guess it would be low, but I don't think we have 4 enough data on that as yet. We'll probably have more 5 data in the next few years. 6 And the third, again -- 7 DR. LOEWER: Let's come to the later 8 question. 9 DR. KAZAZIAN: All right. Go ahead. 10 DR. MAJOR: I'm sitting next. So I have 11 to answer this, too, I guess. 12 I think, you know, as I tried to point 13 out, of course, if you have an integrated viral DNA, 14 particularly in a cell line, it can be excised. You 15 need to have a functional original of DNA replication, 16 and in many cases perhaps adventitious viruses don't 17 have the complete genome present there. So whether 18 they're there or not, I think you have to ask the 19 question whether that DNA integrated is functional to 20 potentially look at an infectious risk. 21 I think also something we probably should 22 be mindful of, too, even though we're looking at these 23 kinds of risk factors here, is that, for example, in 24 the viruses I talked about, the human ones, BK and JC, 25 they're very widely prevalent in the population 238 1 worldwide. All of us in this room have come in 2 contact with these agents. We have antibodies to 3 them, and the vast majority have these type of agents 4 already latently infected and potentially, as far as 5 we know, integrated in some fashion or other in 6 certain types of organs, as well. 7 But it seems like my impression is the 8 consensus seems to be that cell lines or neoplastic 9 cells do not represent that much of a risk here. 10 DR. RUPRECHT: I would concur with that 11 last statement. The data I presented were generated 12 with exogenous primate lentiviruses for which probes 13 do exist, and neoplastic human cell lines could be 14 screened to rule out exogenous lentiviral infection. 15 One issue though that maybe should be 16 reiterated, we presently do not know what the minimal 17 infectious doses for lentiviral DNA is in the animal 18 models. So that would be something that's relatively 19 straightforward and information that maybe should be 20 obtained. 21 We also do not know whether or not 22 infections can be induced with chromosomal DNA taken 23 from infected animals. I did mention that most of the 24 profile sequences that accumulate during the course of 25 a lentiviral infection in the case of this SIV is 239 1 actually defective. So the chance of getting 2 replication competent virus out through contamination 3 seems to me relatively remote, but we should get this 4 baseline information. 5 DR. BLAIR: I think as far as the 6 infectious agents, I think I would agree that although 7 you don't know the minimal dose, it would appear that 8 the efficiency uptake is probably fairly low so that 9 if you reduce the level to the kinds of residual 10 levels that people talk about in viral vaccines and 11 things, that this would not pose a risk. 12 I think in terms of the neoplastic risk, 13 I think clearly evidence is that to induce a 14 neoplastic event in a normal cell of any lineage takes 15 multiple events, some positive, some negative, and the 16 attempts to do this even under optimal conditions 17 requires a fair amount of effort and fairly large 18 amounts of material. 19 So, again, the risk would seem to be 20 minimal with the kinds of residual DNA that you find 21 in vaccines. 22 DR. LOEWER: John. 23 DR. PETRICCIANI: Yeah. There's very 24 little that I can add to what people have already 25 said. Just let me remind you the paper that was 240 1 recently published by Krause and Lewis, taking a worst 2 case analysis for an infectious event, if you 3 normalize that down to 100 picagrams of DNA, it falls 4 right back into the same order of magnitude of risk as 5 the oncogenic events and is really no different from 6 that. 7 The other thing that Ruth mentioned a 8 moment ago and I mentioned in my talk I think really 9 is important. If as a result of this conference 10 people feel that yet more information is needed, and 11 that's debatable, but if more information is needed, 12 I agree with Ruth 100 percent that one of the most 13 important elements is some dose response data because 14 we're operating in Never-Never Land without that, and 15 whether it's in the model that she was talking about 16 or if it's in the 3T3 assay or whatever, I think 17 that's a basic, fundamental piece of data that's been 18 absent for decades. 19 And if we're going to pursue this as an 20 intellectual discussion, we really should have that 21 background data. 22 DR. LOEWER: I have a specific question to 23 Dr. Ruprecht. 24 You mentioned it's not known with 25 chromosomal DNA which have a complete retrovirus is 241 1 really infectious or not. Would you accept such a 2 cell line for production of vaccines? 3 (Laughter.) 4 DR. LOEWER: Even if the possibility is 5 rather remote? 6 DR. RUPRECHT: Well, I would probably try 7 to get the non-provirus containing equivalent of the 8 cell line. 9 DR. LOEWER: I think this should be clear. 10 It's obvious, in my opinion, at least, cell lines 11 which have a complete virus, genomes are not good 12 candidates for production of any biologicals. I think 13 I could mention a degree on this. 14 The question is whether these cell lines 15 contain part of whatever this, for example, oncogenes 16 or other parts, but this still would exclude -- sorry? 17 PARTICIPANT: (Inaudible.) 18 DR. LOEWER: Right. 19 DR. RUPRECHT: Exogenous virus. 20 DR. LOEWER: Yeah, I guess they have only 21 sometimes part of DNA viruses integrated, not these 22 retroviruses maybe. 23 DR. KRAUSE: Dr. Loewer, how would you go 24 about proving that a cell line does not contain a 25 single copy of an infectious genome? 242 1 (Laughter.) 2 DR. LOEWER: I think, of course, you can 3 test directly for certain virus you know. I think 4 every cell line should be well characterized in the 5 cell bank, and especially the use of cell lines over 6 the years, I think, proves to some extent that they 7 contain or do not contain any extraneous viruses. 8 So I think one can be rather sure about 9 well characterized cell lines, that they do not 10 contain complete viruses. 11 DR. KAZAZIAN: Phil, I gather that if you 12 PCR'ed it up and found that it was negative, you still 13 couldn't be sure. Is that what you're getting at? 14 DR. KRAUSE: Well, what primer would you 15 sue? 16 DR. KAZAZIAN: Yeah, right. 17 DR. KRAUSE: Right. So my answer to Haig 18 is what primers would you use. 19 DR. LOEWER: Okay. I think this question 20 is not really new and maybe not specific for 21 neoplastic cells and all people dealing with 22 regulation of biologicals deal 50 years with this 23 problem. 24 DR. KRAUSE: I agree with you, of course, 25 and again, for the transcriber, I'm Phil Krause of the 243 1 FDA. 2 But I guess the question comes up if 3 there's a cell line that's neoplastic or that's 4 tumorigenic and we don't know why it's neoplastic or 5 tumorigenic because of an incomplete history or maybe 6 it came from a person or something like that, the 7 question is then: does that make us think that it's 8 more likely to contain some of these latent viral 9 sequences or does it make us a little bit more 10 suspicious, that being one of the mechanisms by which 11 a cell like that might have become neoplastic? 12 DR. LOEWER: Of course, this is an always 13 open question. To test for the unknown is always 14 rather difficult. 15 DR. RUPRECHT: Maybe if I could add one 16 thing here. If we knew from a known animal experiment 17 what the minimal dose is of chromosomal DNA, say, from 18 an SIV infected animal, and if it turned out that it 19 is ten micrograms intramuscularly, you know that you 20 never have ten micrograms in a vaccine dose. 21 If it turns out that 100 picagrams of 22 chromosomal DNA will never, you know, even if you test 23 it on 200 animals, never lead to systemic infection, 24 even if the DNA were derived from a known viremic 25 animal, then at least you would have some sort of 244 1 frame of reference. 2 DR. KRAUSE: I agree with you. That makes 3 a lot of sense. One think that makes me pause a 4 little bit is the slide that you showed and that I 5 showed a couple of nights ago that appear to 6 demonstrate that polyoma virus DNA, for instance, is 7 quite a bit more infectious than retrovirus DNA. 8 And you know, the list of viruses whose 9 DNA has been tested for infectivity, and in fact, the 10 study that Andy Lewis and I published was based on 11 retrovirus infectivity, DNA infectivity, when we 12 submitted that, we actually weren't aware of the 13 polyoma virus data. 14 But so the question is: are there other 15 kinds of DNA from other families of virus which 16 haven't even been looked at which are quite a bit more 17 infectious? 18 And so how does one really determine what 19 that baseline is? 20 DR. RUPRECHT: The answer to that is yes. 21 I showed a slide that summarized a number of animal 22 experiments with other agents, including Hepatitis B 23 in chimpanzee, HTLV-1 in rabbits, FELV, murine 24 leukemia virus, just to name a couple of the others. 25 Those response curves though are not always available. 245 1 DR. BLAIR: It seemed to me though that 2 the polyoma data was in the system where the polyoma 3 replicated, and in a sense, that's biasing. I mean, 4 if the virus can replicate in the tissue, in the thing 5 you expose it to, it's going to -- you're going to 6 much more easily pick it up than if it's something 7 that has to go in, you know, and multiply infect many 8 cells just directly as DNA. 9 DR. KRAUSE: I guess, except you're not 10 putting the virus in to start with. I'm not sure how 11 that's different from the SIV mac situation where 12 you're also putting the DNA, which also then 13 presumably once you get a virus out of that also 14 replicates in that animal. 15 PARTICIPANT: The yield of the polymer is 16 about a million particles per cell. So I think it's 17 a lot more than the retrovirus. 18 DR. MAJOR: As long as we've talking about 19 polyoma, I'd like to bring it back to the human again. 20 JC DNA is not infectious in animals, in the kind of 21 situations that it causes tumors in non-human 22 primates. 23 BK, on the other, is. The BK DNA is 24 infectious, and animals will not only make virions, 25 but they'll make antibody to the BK DNA and that 246 1 antibody will also be specific to chromosomal DNA as 2 well. We don't know that in humans, but certainly in 3 animal situations it is. 4 DR. LOEWER: Yes. 5 DR. HEINEINE: Walid Heneine, CDC. 6 No one has mentioned how much DNA we now 7 have in the licensed vaccines. I mean, how much are 8 we being exposed to? Do we have any idea in the viral 9 vaccines, like yellow fever, measles, mumps, how many? 10 Do the regulators have an idea from the manufacturers, 11 how much DNA there is? 12 DR. LOEWER: I have probably no idea. 13 Nobody that I know has mentioned it. 14 Paul? 15 DR. KRAUSE: I promise to sit down for a 16 while after this. This actually came up when we were 17 licensing the chicken pox vaccine, an it's part of the 18 safety data that was submitted for that vaccine. The 19 manufacturer, Merck, did a number of studies 20 quantifying the amount of DNA in different vaccines 21 and submitted those to us. I noticed that there are 22 people from Merck here if someone from that company 23 would like to describe those results, which are in the 24 public domain already. 25 (Laughter.) 247 1 DR. KRAUSE: But the answer to the 2 question is that -- 3 DR. SHEETS: I do have an answer. This is 4 Becky Sheets from CBER. 5 I think that the vast majority of licensed 6 vaccines, U.S. licensed vaccines, have not been tested 7 for residual DNA. Most of them are manufactured in 8 primary cells and diploid cells and nobody worried 9 about it. 10 The few that have been tested are the ones 11 that have been licensed in the last few years, 12 including varicella and Hepatitis A. 13 DR. KRAUSE: The one that has the most 14 that I'm aware of is varicella, in fact, which has a 15 little over a microgram per dose of MRC5 cell DNA, and 16 in that context several other vaccines besides the 17 ones that Becky mentioned were also tested and were 18 lower than that, but that gives you a range of what 19 one potentially could be dealing with if one is 20 talking about a live attenuated viral vaccine that 21 doesn't grow particularly well in tissue culture, for 22 which you simply can't do anything to get rid of the 23 DNA without risk to the virus that you're attempting 24 to deliver. 25 DR. HENEINE: So, I mean, the whole thing 248 1 could be an issue and may not be a big issue for some 2 vaccines, and it could be an issue for some other 3 vaccines where we know we have some DNA. 4 We have some premier data from our lab. 5 We've looked at the MMR vaccines, and we amplified the 6 chicken endogenous retroviral sequences basically in 7 an attempt to type these loci in these vaccines, and 8 what we found out in the cef produced vaccines, 9 there's very little genomic gain anyway. We are able 10 to amplify some sequences from them, but there's not 11 very strong signals. 12 And other vaccines like yellow fever, for 13 example, there's we found much higher amounts of 14 genomic DNA and much higher amounts of viral proteins, 15 as well, which you can immunoprecipitate very easily. 16 So it looks like there's a wide spectrum 17 of residual DNA and proteins that may be present in 18 these vaccines. 19 DR. LOEWER: I think this is now known for 20 a couple of years that Dr. Schuepbach, also in the 21 audience and I guess he will address this question 22 tomorrow during his talk, and your findings are, I 23 guess, repeated by a number of laboratories which also 24 find AMB and EAB sequences as well as DNA in such 25 vaccines. 249 1 But to my knowledge, at least in our 2 laboratory we have not quantified the amount of cell- 3 wide DNA in these vaccines. 4 DR. HENEINE: Yes. The relevance in those 5 vaccines that have large amounts of DNA is that we 6 have full length proviruses that are intact that would 7 be present there, and that could be taken up by the 8 cells like the mechanism we've been talking about 9 these two days. The dose would be enough to initiate 10 infections. 11 DR. LOEWER: Yes, but the point mentioned 12 earlier is that these DNAs derived from normal cells, 13 therefore, didn't create any concern so far, and at 14 least to my knowledge there are no adverse reactions 15 in the millions of vaccinated people which could be 16 attributed to the presence of this DNA. 17 That's the point. There's a question 18 here, indeed. Do we expect the same with neoplastic 19 or with DNA derived from neoplastic cells or not? 20 That is the question we cannot really answer so far. 21 DR. MAJOR: Yeah, let me ask something, 22 maybe a naive question here. I'm not sure I 23 understand exactly what the nature of the state of the 24 DNA that is the contaminant in these vaccines. Is 25 this cellular DNA which is encapsulated into virion 250 1 structural particles so that there would be a 2 difference between enveloped -- 3 DR. LOEWER: No. 4 DR. MAJOR: -- versus -- or is this DNA 5 which is just carried along with the purification 6 process? Is that what we're talking about? 7 DR. HENEINE: Yes. I mean, if you take 8 the vaccine, you resuspend it, and you extract it, and 9 you can amplify endogenous retroviral sequences from 10 it very easily, and so it must have been cellular, 11 genomic DNA that got purified with these particles. 12 Now, also you can find particle associated 13 RNA. That's something else, but in addition to that 14 you also find genomic DNA. 15 DR. LOEWER: But one has to realize that 16 there is not much purification for live viruses. 17 DR. HENEINE: And that's not true, like I 18 mentioned, for all viruses. Maybe the yellow fever 19 vaccine, the way it is produced explains why you have 20 higher levels of genomic DNA in it. 21 DR. ONIONS: Johannes, I wonder if I could 22 just ask for a clarification because rather as Ruth 23 was talking about the titration of DNA proviruses to 24 get a take, to get an infection, we can push FALV down 25 to around about ten micrograms. You don't get 100 251 1 percent take at that level, but you do get takes. 2 And you just said -- and could you clarify 3 this? -- you just said that you got full length 4 proviruses; is that correct? You've actually 5 amplified the whole -- 6 DR. HENEINE: No, no, we have not tried 7 that. The experiments were originally designed to 8 type the endogenous ALV loci in the vaccine. In other 9 words, if the locus -- are we dealing with defective 10 loci or full length infectious loci? 11 So we've used the vaccine itself as a 12 source of DNA which could be a good representative of 13 the cell substrate in which it was grown to amplify 14 and type by PCR based methods the different ALV loci. 15 DR. ONIONS: Okay. I'm sorry. 16 DR. HENEINE: But while doing those 17 experiments, we found out that some vaccines have very 18 little DNA, but others have much higher levels of DNA, 19 and the concern is whether those that have high levels 20 of DNA, whether there is DNA with full length loci 21 that are infectious, that are known to be infectious 22 loci like EV2, EV18, EV19, very known, and ALV that 23 can be taken up by cells and initiate infection. 24 DR. ONIONS: I mean that seems to be a 25 central point, is the state of sharing of the DNA in 252 1 all of these examples, and again, not just making an 2 average estimate, but actually looking at the 3 distribution of size I think is something that would 4 be relevant. 5 DR. HENEINE: I guess the other experiment 6 to do is to see if we can have any evidence of full 7 length providers that are associated with the 8 vaccines. 9 DR. HUGHES: If we're talking ALV related 10 avian viruses, rav 0 and its cogenants, they are not 11 infectious for any mammalian cell on two criteria: 12 the envelope, subgroup E does not recognize an 13 appropriate target on mammalian cell, and the 14 machinery of the virus even if you transfect in the 15 intact provirus into a mammalian cell does not produce 16 an infectious virus. 17 So while I'm not recommending having those 18 kinds of full length proviruses present necessarily, 19 I don't think they prose any particular threat to 20 human. 21 DR. LOEWER: Any further question or 22 comment? 23 DR. HAYFLICK: Hayflick, UC-SF. 24 I think it's important to keep in mind 25 that it's not only the cell substrate that might 253 1 contribute DNA and RNA, but in those instances where 2 serum and trypsin are used, there is a logical 3 expectation that it could come from one or both of 4 those sources despite the fact that the final product 5 might be diluted heavily or no serum might be used in 6 the final product, but be present in the former 7 population doublings that led up to the production. 8 I don't think that this has been 9 addressed, and I'm wondering whether anyone has, in 10 fact, determined the residual DNA content or RNA 11 content in serum. 12 DR. LOEWER: Not to my knowledge at least. 13 DR. SHEETS: I think there is an answer to 14 that, but I'm not the one that knows it unfortunately. 15 This is Becky Sheets, FDA. 16 I wanted to respond to an earlier question 17 regarding how purified are live viral vaccines. Most 18 -- I'll just generalize -- many, maybe most live viral 19 vaccines are made by inoculating virus into tissue 20 culture or into the allantoic fluid of chicken eggs, 21 and once infectious virus is produced either by cell 22 lysis or budding, then the harvest is made of the 23 supernatant or the allantoic fluid, and oftentimes 24 this harvest is filtered, and that's your viral 25 vaccine. 254 1 So that we're all on the same page, that's 2 a live viral vaccine. It's not impure or unpurified, 3 as has been said, but it is minimally purified, often 4 simply by filtration to get rid of cell debris and 5 bacteria and fungi that might be present, but it is 6 not purified to the extent that a single protein in a 7 therapeutic biologic is purified. 8 DR. LOEWER: Thank you for this 9 clarification. 10 Now I would like to ask a regulatory 11 question to Dr. Kazazian. As he mentioned, the 12 activity of retrotransposons is highest in some type 13 of tumors. For example, you mentioned testicular 14 tumors and, I guess, ovary tumors. The same is true 15 for other endogenous retrovirus, and I show a Northern 16 Blot from our laboratory which shows a number of 17 different tumor cell lines and diploid cell line. 18 This is a terata carcinoma cell line, amnio chorion 19 carcinoma. I guess this is a raptomycin sarcoma cell 20 line. There's a vero cell line, and one endogenous 21 human retrovirus, HERV-K, for example. It is 22 expressed only in the testicular tumors, and we know 23 that other endogenous retrovirus, for example, HERV-H, 24 is also very well expressed in testicular tumors. 25 Would it be your opinion that this would 255 1 exclude such cell lines for the production of 2 vaccines? 3 DR. KAZAZIAN: Well, first of all, I think 4 that there are testicular, ovarian tumors of similar 5 type that don't express retroposons, at least, and 6 what I guess I'm saying is not all tumors of the same 7 type will express. I don't know whether that's the 8 case with HERV-K or not, and I think that certainly 9 there are those that don't have the expressions 10 excluded. 11 That's a tough question. That's a tough 12 question because I don't know that expression is 13 related to retrotransposition. Clearly we don't know 14 that with HERV-K. We don't know that HERV-K is 15 retrotransposable at this point, and I don't know that 16 there is a correlation between expression and 17 retroposition for sure. 18 DR. LOEWER: Now that you mention it, 19 expression is a prerequisite for transposition. 20 DR. KAZAZIAN: Yes. 21 DR. LOEWER: Yeah. John. 22 DR. KAZAZIAN: It is requisite, but I 23 don't know that it completes it. 24 PARTICIPANT: Regarding endogenous viruses 25 like this one and the ALVs, I don't think it's been 256 1 discussed much so far that these viruses as virtually 2 all of them that's found in the genome are very 3 extensively methylated and expressed at very low 4 levels. The endogenous provirus that gives rise to 5 the infectious rav 0, in the form in which it's 6 inherited seems to be expressed at a level of maybe a 7 tenth of a copy of RNA per cell; if you then allow 8 that to be replicated, it goes up by a factor of 1,000 9 or more over that. 10 So that actually provides -- if you're 11 worried about these things, that actually provides a 12 margin of safety because that's the form in which 13 these kinds of things at least will be passed in 14 recipient cells if they contaminated the vaccine 15 products. 16 DR. KAZAZIAN: And presumably that's also 17 what's holding down retrotransposition of line 18 elements of methylation. 19 PARTICIPANT: Presumably. It's been 20 argued that this may be a mechanism to protect the 21 genome against these. 22 DR. KAZAZIAN: Right. 23 PARTICIPANT: But if so, it's doing a 24 pretty lousy job. 25 (Laughter.) 257 1 PARTICIPANT: In fact, it probably allows 2 their increase because it reduces the counter- 3 selective effects. 4 PARTICIPANT: As a bit of a follow-up to 5 an issue I raised this morning about how you culture 6 the cells may affect genetic events within the cell, 7 have you ever compare the activation of lines, for 8 example, in stationary cells versus rapidly cycling 9 cells? 10 A lot of, for instance, prophage will 11 activate as E. coli approach the stationary, and there 12 may be very, very big differences between the 13 activation of events in cells depending on how they're 14 being handled at that time. 15 DR. KAZAZIAN: Yes. We have not, but Karl 16 Schmid has looked at both alus, principally looking at 17 alu transcription. This is only transcription. It's 18 not retrotransposition. So looking at alu 19 transcription, and he's looked at line transcription, 20 and it goes up, as I recall now, in serum starvation, 21 goes up in heat shock, goes up in adenovirus 22 infection. That's what I can tell you. 23 PARTICIPANT: Yeah. That's sort of what 24 I kind of expected. 25 DR. KAZAZIAN: So it's way up. 258 1 PARTICIPANT: A stress like environment 2 where somebody wants to bail out of the genome, that's 3 the perfect trigger for these things to start thinking 4 about moving. 5 DR. KAZAZIAN: Thank you. 6 DR. LOEWER: So let's go on in discussion 7 to come to the neoplastic risk event, to leave the 8 infectious risk event, and we had quite a number of 9 talks and discussions already on this issue, and I do 10 not know whether somebody on the panel would like to 11 comment again. 12 We know from the 3T3 assays that there is 13 a neoplastic risk rendered at least in tissue culture. 14 I had the impression from the talk today that NIH3T3 15 test is mainly a test for activated ras and to a 16 lesser degree for other oncogenes. 17 But I think what is intriguing is that is 18 difficult or impossible to find a neoplastic risk by 19 injecting animals with neoplastic DNA. 20 So what about epigenetic and genetic 21 changes, since this is an issue, I guess, mainly to 22 the people working on methylation? Can the 23 methylation pattern -- I think that's the background 24 of this question -- can the methylation pattern maybe 25 change by transfection or by transfection of 259 1 neoplastic DNA? 2 And is this a risk for the recipient cell 3 with respect to development of a tumor? Are there 4 data available in this respect? 5 DR. DOERFLER: Walter Doerfler from 6 Institute of Genetics in Cologne. 7 Why you probably approach this question 8 partly to me at least, it's difficult to answer. Of 9 course, we have pursued this possibility very 10 systematically in some of the transfected cell lines 11 that we have produced and see changes in methylation 12 at quite distinct places. Whether this is a general 13 phenomenon, of course, we haven't amassed enough data. 14 On the other hand, I feel one could 15 probably predict for the foreseeable future that a lot 16 of people will study this or similar issues with the 17 increasing availability of DNA chip (phonetic) 18 technology because DNA methylation is one thing, but 19 changes in expression patterns is not so difficult to 20 do. 21 Say I make a lambda transgenic cell line 22 and a non-transgenic cell line or take a lambda DNA 23 transgenic mouse and non-transgenic mouse, same organ, 24 same mouse strain, and I compare cDNA hybridizing back 25 to a huge DNA panel, say, having 20, 30, 40,000 DNA 260 1 dots on it, which technically is not impossible to do. 2 It's actually quite simple to do. It's probably 3 expensive right now. 4 So I would expect a lot of work in this 5 direction will be done within the next few years. 6 Certainly we plan to do a lot of that sort of work. 7 Whether this will answer the question is another 8 problem, of course, but we will get at least an 9 overview, and we perhaps will be able to predict in a 10 more defined and a more scientific way what's going on 11 in these cells. 12 And I suppose this will also help evaluate 13 some of the other because if you found quite distinct 14 differences, well, maybe your suspicion would be 15 stronger that also oncogenic events might ensue. 16 I understand, although I've only heard it 17 in talks and never seen the data myself, that there 18 are a number of labs now who have looked in transgenic 19 in knockout mice, yeah, I think in knockout animals, 20 and they do find changes in genes other than the one 21 knocked out. 22 I cannot swear that this is true because 23 I haven't seen the data, but I have heard from, as one 24 says, reliable sources that this is so. Perhaps other 25 people have information on that. It would be 261 1 interesting. 2 DR. KAZAZIAN: I would just say that I'm 3 sure that you all know that with expression analyses, 4 that when you do these kinds of things, I've heard -- 5 again, I haven't done them myself -- but you get 6 changes in hundreds of cDNAs both in the up direction 7 and in the down direction, and it's a very complex 8 analysis. 9 PARTICIPANT: Or a viral infection. 10 DR. KAZAZIAN: Or in viral, right. 11 PARTICIPANT: I liked your approach of 12 asking the questions about what kind of cell line 13 could one rule in or out, and I wasn't sure that there 14 was a complete answer to this question of specifically 15 if there were an intact viral genome of some nature 16 that could be defined by any of you, would that rule 17 out the use of that cell. 18 And likewise you asked Dr. Kazazian 19 whether he thought that the expression of HERV-K would 20 rule out the use of a cell, but I wasn't sure that 21 there was a clear consensus on that. 22 DR. KAZAZIAN: Well, I said I didn't think 23 so because I don't know of any HERV-Ks that are 24 active. Okay? 25 PARTICIPANT: Okay. So maybe there is a 262 1 consensus on that. 2 DR. LOEWER: But the same is true, for 3 example, for IAPs when, for instance, atag (phonetic) 4 particles and cell lines producing IAPs are accepted 5 for the production of biologicals, but not for 6 vaccines insofar as CHL cell (phonetic), for example. 7 Yeah. 8 PARTICIPANT: We've been thinking about 9 producing live attenuated virus vaccines in vero cells 10 for use by the mucosal route. Give it a couple of 11 drops in the nose, and we were wondering and I'm 12 wondering whether you feel that it is safe. 13 These live virus vaccines will have 14 cellular DNA in it, vero cell DNA. I think, Dr. 15 Blair, you had mentioned that vero cell DNA was given 16 in your assay and was not found to be tumorigenic in 17 the 3T3 cells, and I think as far as some of the 18 examples Phil has been using, if you take polyoma 19 virus DNA and give it parenterally, it is not only 20 infectious, but it is tumorigenic. That same DNA when 21 given mucosally has no biological activity, you know, 22 when you give it the maximum amount that you can grow 23 up and actually give to a mouse. 24 So would it be reasonable to conclude, 25 based on these kinds of data that do exist, that 263 1 certain continuous cell lines, such as VEROs, which 2 are poorly tumorigenic would be perfectly safe as far 3 as the DNA issue in terms of transferring infectivity 4 or neoplastic risk, would be perfectly safe to use via 5 a mucosal route? 6 I personally believe that it absolutely 7 would be, and I believe that there are other 8 neoplastic cell lines that could also be considered 9 for this route. 10 What is the feeling of the group in this 11 first cell line such as this? 12 This is a very important point because 13 there are vaccines for very important viruses can be 14 made in this cell line and almost only in this cell 15 line. 16 DR. LOEWER: So as a member of a 17 regulatory authority, it wouldn't be wise for me to 18 give you a clear-cut answer. 19 (Laughter.) 20 PARTICIPANT: So you're being asked as 21 advisors to the regulatory -- 22 DR. LOEWER: The members of the panel, of 23 course, are free to answer what they believe is 24 correct. 25 (Laughter.) 264 1 PARTICIPANT: Maybe in support of what you 2 said about it's difficult to give a clear-cut answer, 3 I think we have the practice. We make oral polio 4 vaccine, as I will tell later, on vero cells, and it's 5 purified, and the amount per dose is probably in the 6 range of one to ten picagrams. 7 Now, having said that, it doesn't say that 8 for another virus systems one could reach such figures 9 because the virus system, the processing, the 10 downstream processing will probably greatly affect, 11 and the cost that you want to pay for that downstream 12 processing, for the losses that were incurred during 13 the purification process will probably decide how much 14 residue of DNA will be there. 15 Besides, that should be put in parallel 16 place in parallel to the basic decisions of whether we 17 could or we could not be happy with, say, the presence 18 of cDNA. It's not clear-cut. It's not black and 19 white. It's really a quantitative issue. 20 PARTICIPANT: (Inaudible.) 21 PARTICIPANT: Can't hear. 22 PARTICIPANT: Thank you. 23 I want going to question the assumption 24 that live viral vaccines necessarily have to be impure 25 because modern methods of production should blend the 265 1 possibility to provide even conventional whole 2 particle or live vaccines in a purified form, and 3 we've hear the answer to that, the vero cells. 4 Now, over 20 years ago I produce 5 experimental cold adapted in flame devirus (phonetic) 6 vaccines in chick eggs, and these, in fact, were 7 purified products. It was then purified. 8 The reason we were only purifying them at 9 that time wasn't safety. It was to enable -- we had 10 more control over the product. I could adjust the 11 titre, clinical trials, but there's also a practical 12 reason. It enabled me to have product which had 13 tropic sucrose in it, which improved the infectivity, 14 but it also improve the palatability for the 15 volunteers, and I very soon found out that the 16 volunteers did not like having allantoic fluid put 17 down their noses, and a purified product was much more 18 acceptable. 19 So I just risk the question, the 20 assumption that the live virus vaccines need to be 21 pure. They couldn't be purified. 22 DR. LOEWER: So personally I am not 23 absolutely sure even with new technologies it would be 24 possible to purify live vaccines or live virus much 25 better because this may be true, for example, for 266 1 naked viruses, polio or so, but with enveloped 2 viruses, maybe herpes viruses or even rabies, it may 3 be more difficult, and they may even incorporate 4 similar DNA or nucleic acids which cannot be purified 5 away. 6 It's at least a possibility which has to 7 be checked. 8 So CBER, of course, quite a number of 9 questions this group, and this is the next question 10 they asked. Consider factors that might contribute to 11 increased safety concerns associated with residual 12 cell substrate DNA, and one of the question is -- this 13 was posed several times during this meeting and 14 wasn't, I guess, really answered -- was aggressiveness 15 or tumorigenicity of tumor cells is a factor which 16 increases safety concern. 17 That means the more aggressive, the more 18 risky or isn't it true? 19 So my personal view to say it, having 20 learned that the NIH3T3 test really recognizes only a 21 certain spectrum of oncogenes, the question would be 22 whether these tests in animals also select only for 23 certain aspects, for certain oncogenes. The question 24 is how well it models the situation in human. 25 And, therefore, personally I would think 267 1 it's difficult to extrapolate from the data in animals 2 to the possible risk in humans. So that would be my 3 argumentation, but I would like to hear other 4 opinions. 5 So if there are no opinions, I guess -- 6 (Laughter.) 7 DR. PETRICCIANI: Johannes. 8 DR. LOEWER: Yeah. 9 DR. PETRICCIANI: I think with regard to 10 the aggressiveness or the degree to which a cell is 11 tumorigenic in one or another animal system, as I 12 mentioned during my presentation, I don't see that as 13 an important criteria with regard to whether or not a 14 cell should or should be not more or less acceptable. 15 I think the real issue is as we've said 16 many times. The issue is what cell contaminants are 17 in the final product. Whether they came from a very 18 aggressive cell or a less aggressive cell probably 19 doesn't make any difference. 20 The example that I gave was you could have 21 one cell that's very aggressive with a few activated 22 oncogenes. You could have one that's much less 23 aggressive in those systems for whatever reason that 24 has 100 activated oncogenes, and I don't know that you 25 can say one is more or less better than the other as 268 1 far as the cell substrate, just as one example. 2 I think using that as a criteria is 3 fraught with problems. 4 DR. LOEWER: Dr. Hayflick. 5 DR. HAYFLICK: Hayflick, UC-SF. 6 I'd like to return to the point I made 7 earlier relevant to the use of serum and trypsin in 8 many, but certainly not all of these vaccines, and 9 that is the likelihood that there can be carried over 10 into cell culture sources of DNA and RNA, for example, 11 from live virus vaccines that are given to most cattle 12 these days. 13 There are a number of vaccines that are 14 used worldwide, certainly in this country, to prevent 15 a number of virus diseases of cattle. Those virions 16 are carried over in the serum that is then used for 17 the manufacture of vaccines. 18 In addition to that, most cattle in this 19 country and in other countries are given anti- 20 helminths and other chemicals that are known to have 21 breakdown products identical to known carcinogens that 22 are also carried over in the serum to a source of 23 product development or manufacture. 24 And I think very little attention has been 25 given to this. As long as we're talking about 269 1 aggressiveness and potential tumorigenicity, I think 2 that this matter should be seriously considered. 3 Thank you. 4 DR. LOEWER: Thank you. 5 I think we have already discussed the next 6 two questions here, the presence of proviral DNA, the 7 presence of latent/occult viral genomes, and there is 8 a question whether there are other factors which could 9 increase the concern. 10 I have here a list which was put together 11 already 15 years ago or so. One is amplification for 12 oncogenes, which John already mentioned. Then there 13 is maybe the risk for enrichment during purification, 14 especially when the sequences are incorporated into 15 the viruses themselves. 16 And there may be a certain risk from 17 substances which could enhance the uptake. For 18 example, so far as I know nobody has tested the effect 19 of adjuvants on the uptake of DNA itself, but of 20 course, one has to realize that live viral vaccines do 21 not contain adjuvants. Only inactivated viruses, and 22 this inactivation of viruses alters the DNA, may be 23 inactivated. 24 So these are other factors which have been 25 considered at the beginning of the '80s at least. 270 1 I don't know whether somebody wants to add 2 some other factor or some other possible risk. So we 3 will shortly continue. 4 This was also discussed so far that the 5 amount of residual DNA, if it it's as low as possible, 6 which may contribute to a lesser concern. 7 Then there is mentioned the possibility of 8 the physical state of residual DNA, whether 9 incorporated in viruses or not, but it's of high 10 molecular weight or it's cut in pieces, and personally 11 I'm not sure what the influence of the physical state 12 really is. 13 The larger the DNA molecule, the higher 14 the probability that a complete genus is coded for. 15 It contains a complete gene, but maybe less 16 effectiveness of incorporation, and also some less 17 possibility of integration. 18 And the shorter the pieces, and we have 19 heard yesterday that even very, very small pieces can 20 integrate and cause tumors; that means the smaller the 21 DNA, the more single DNA molecules, the higher the 22 risk for indication maybe. So there may be two sides 23 of the coin. 24 And the question of DNA clearance removal 25 was also addressed yesterday, I guess, and we have 271 1 discussed it shortly here. 2 Is there any further comment from the 3 panel or from the audience to these questions? 4 So it's not the case. So this was a 5 question, I guess, which was not addressed in our 6 session or the session as the members or the speakers 7 have presented their data, namely, the usefulness of 8 the model system, and so I will switch this, and we'll 9 come to maybe as a last point of this discussion to 10 the quantitative aspects. 11 And you may have read the paper by Krause 12 and Lewis with some considerations with respect to the 13 quantitative issues, and the first question here is: 14 would it be acceptable that in a frequency of one 15 pair, one million doses, if it is an acceptable level 16 of risk? In other words, would it be acceptable that 17 one of one million vaccines developed a tumor because 18 of the tumor DNA, which may be -- 19 DR. KAZAZIAN: By the way, how are you 20 going to prove that? How are you going to prove that, 21 you know, one in a million is due to the tumor -- I 22 mean due to the vaccine as opposed to a natural cause? 23 DR. LOEWER: You are completely right. I 24 will not defend this. I just mention what he has 25 written. 272 1 (Laughter.) 2 DR. LOEWER: And indeed -- 3 DR. KAZAZIAN: So who wrote that anyway? 4 (Laughter.) 5 DR. LOEWER: Indeed, that is the major 6 question. This would never be detected in the 7 background of tumors. That's absolutely clear, and it 8 may be only a theoretical value, which could be a goal 9 to reach. It will never be detected in -- 10 DR. KAZAZIAN: The problem is that the 11 jury might pay the suer for that one million even 12 though we don't know. It's part of the background. 13 DR. LOEWER: So, indeed, I think this 14 highlights the problematics or the problems with all 15 of these theoretical calculations. It's questionable 16 whether it's worse to do so. 17 Phil, you make a comment? 18 DR. MINOR: Well, I'm not sure you 19 wouldn't pick up one in a million per vaccine dose. 20 When you think how many doses children get and you're 21 talking about in my country it's like half a million 22 children a year get this stuff, and you can pick up 23 adverse reactions at one in 100,000, and vaccines get 24 withdrawn for that kind of reason. 25 DR. LOEWER: But it depends. 273 1 DR. MINOR: But, yeah, there are things 2 like meningitis is huge, as well, and you can pick 3 that up, and you can withdraw vaccines on the basis of 4 that. 5 DR. LOEWER: It may depend, I think, on 6 the patient period. If it takes 40 years to develop 7 the tumor you will never detect it, but if it takes 8 only two years, it may be possible. 9 DR. MINOR: Well, I think alternatively 10 you might get a very nasty surprise in 40 years. 11 DR. LOEWER: Yeah. 12 PARTICIPANT: Yes. 13 (Laughter.) 14 DR. LOEWER: Okay. 15 PARTICIPANT: I think the idea is to try 16 to think about these issues conservatively based on 17 the idea that if one is going to start manufacturing 18 vaccines in new cell substrates, you would like to be 19 able to assure the public that you've considered the 20 risks down to a very low level, and so even if it 21 occurred at this level we would not be able to detect 22 it against the background, that doesn't mean that we 23 as a regulatory agency don't have an obligation to the 24 general public to be able to tell them -- 25 PARTICIPANT: At least tell them that. 274 1 PARTICIPANT: Well, we'd like to know 2 that, too, that the risks are at negligible levels. 3 DR. LOEWER: So usually the figure one to 4 one million is more or less accepted figure in 5 biologics. For example, the risk for transmission of 6 HIV by blood donations is also in this order, and it's 7 more or less accepted or the risk of polio OPV 8 associated, poliomyelitis is also a disorder. 9 PARTICIPANT: On the other hand, just to 10 follow up, I'm not sure that right now the public is 11 willing to accept one in a million risk of vaccine 12 associated paralytic polio, and I'm certain that the 13 public is not willing to accept a one in a million 14 risk of HIV infection among healthy children who are 15 just coming in to get their routine shots for school. 16 DR. LOEWER: Okay. 17 DR. LEWIS: Yeah, I think that those 18 figures are based on what we could possibly generate 19 from an experimental or a validated test system, not 20 what we're going to expect in the population at large. 21 In other words, the issue is how to come to grips with 22 estimating the risk or the perceived risk, not 23 necessarily the risk, because at this point in time 24 I'm not sure we have any risks here, but these are 25 perceived risks, and if we're going to go out and 275 1 convince people to give these products to their 2 children and to their patients, then we have to have 3 some level of information upon which to make a 4 decision. 5 At least from my own personal perspective 6 to say "I think" or "my opinion is" in a situation 7 like this is not going to be good enough. We have to 8 have data. 9 Now, the question is: how do we generate 10 that data and what do we use that data to do? 11 And I think an estimation like this is 12 based on a simple attempt to come to grips with how to 13 estimate this level of risk so that when we try to 14 convince ourselves as a regulatory body that it is 15 safe and we want to put the stamp of FDA approval on 16 this or, to quote myself the other night, the FDA says 17 this is okay, we need to have a concrete set of 18 information upon which to make that decision. 19 Otherwise it becomes a simple exercise or article of 20 faith. 21 And at this point in time I don't think 22 I'm willing to sign my name to something that's an 23 article of faith. We need some data, and the question 24 is: how do we generate that data? 25 And what we're, you know, hoping to get 276 1 from discussions like this is how to proceed to get 2 that information. 3 DR. LOEWER: Dr. Hayflick. 4 DR. HAYFLICK: Hayflick, UC-SF. 5 We already have a baseline for the 6 exception of what I read up there to be infectious 7 risk events in respect to the fact that we are living 8 today with an attenuated polio virus vaccine that, if 9 my memory serves me well, has resulted in one 10 acceptable case of paralytic polio per three million 11 doses or something in that range. 12 And, in fact, there are more cases of 13 paralytic polio occurring in this country per year 14 from the vaccine than from the wild type. We already 15 have such a baseline. 16 DR. LOEWER: Thank you. 17 I would like to come to a final point 18 here, and this will address as the problems with all 19 of these calculations, and what I've shown here is the 20 calculation for the activation of, quote, oncogenes by 21 intracellular foreign DNA, and this calculation was 22 published by Reinhardt Kurth after some discussions in 23 our institute with Phil Minor sitting over there and 24 starting to laugh. 25 It demonstrates, I guess, when you run 277 1 into problems by all these calculations, and I would 2 like to go through very quickly to the difficult 3 aspects here. 4 The calculation starts with a quantitative 5 aspects of the confection of the zarc gene (phonetic) 6 in the wing of chickens, which are already mentioned 7 here and the probability to create a tumor with 8 respect to the per molecule of DNA or three times ten 9 to the minus 12. 10 And the efficiency of the cellular uptake 11 was taken from some studies to be ten to the minus 12 nine. That means that the probability that an 13 intracellular DNA may integrate in the chromosome at 14 three times ten to the minus three, but anybody can 15 easily recognize that there's many assumptions, and 16 I'm not sure how representative these data really are, 17 especially the uptake of injected DNA in the cell as 18 more attachment. 19 The probability that if an integration 20 occurs it's integrated into or close to a portal 21 oncogenous, easier to calculate from the size of the 22 genome and the size of the portal oncogene, and this 23 is in the range of ten to the minus five. 24 The probability that this integration 25 leads to an activation of the gene is taken here as 278 1 ten to the minus two, and this, I guess, is pure 2 judgment. It's not based really on data, but now you 3 can multiply all of these probabilities, and you will 4 end up with the probability of a single intracellular 5 DNA molecule to activate a single portal oncogene in 6 a size range of three times ten to the minus ten. 7 It's pretty low, but we all know that the 8 way -- that is a multi-step way to tumor information, 9 and this calculation it was assumed that two 10 independent activation events are required, and this 11 leads to a number of ten to minus 19. 12 And if you'll look in the recent 13 publication of Nature, we have seen the picture, their 14 claims that at least four different steps would be 15 necessary, and then we apply this figure to this 16 calculation, and we end up with ten to the minus 38 17 here in this case. 18 And then one starts to wonder why tumors 19 at all occur if this is such a rare event, and this 20 again shows that all of these calculations are very 21 difficult, and one has to take great care to take 22 these figures and to draw conclusions from this. 23 And please also, my problem for you if you 24 want to find figures to convince the public on the 25 risk, it will be always difficult to have figures and 279 1 to rely on these figures. They may lead to a false 2 sense of security or knowledge. 3 So I would always be very care of these 4 type of calculations. 5 Comments? 6 DR. BROKER: Tom Broker, UAB. 7 I want to kind of approach these kinds of 8 numbers in a very different way, and that is from the 9 perspective of human papilloma virus infections that 10 lead to cervical or penile carcinomas. 11 One of the more remarkable facts of this 12 family of diseases is that since 1980 more people have 13 died of HPV disease than have died of AIDS. I mean 14 this is a disease that takes out three-quarters of a 15 million people per year. 16 Now, the kinds of calculations, they just 17 don't seem to square with what we know is the reality 18 of that disease process. For example, vast numbers of 19 people, hundreds of millions, harbor HPVs, and the 20 process through which neoplasia occurs is, first, the 21 establishment of the viral genome in the infected 22 tissue, followed some time later in a small 23 percentage, usually one percent of the patients, where 24 an integration event takes place, and that's where I 25 think we're interfacing with this kind of table. 280 1 The integration, it turns out, is somewhat 2 selected. It's going into an actively transcribed 3 region of the genome. There's further selective 4 events for the active expression loci, which then 5 leads to an up regulation of genes that foster 6 somewhat more rapid cell cycling. 7 In the presence of mutagenic agents, and 8 over a period of time, probably ten years or more, 9 these genes get more and more activated. There's a 10 little bit of increase in P53, of E6 taking out P53, 11 and so forth, and it's a progressive neoplastic 12 process. 13 And so I think you can't have 14 instantaneous numbers, but the bottom line is three- 15 quarters of a million people die every year of this 16 family of diseases. 17 DR. LOEWER: I personally believe that 18 such type of calculations cannot be applied for virus, 19 tumor viruses. In the case of HPV, of course, virus 20 produce actively oncogenes, E6 and E7, and this is not 21 taken into account in all of these calculations. 22 These were just calculated, passive activation of an 23 oncogene. 24 PARTICIPANT: Johannes, I was just going 25 to re-echo that last point. I mean the assumption 281 1 here is that they are independent events, and I don't 2 think that is the case necessarily in transformation 3 events. 4 For instance, if you introduce nat and myc 5 (phonetic) into a cell, the probability that that cell 6 population then goes on to cumulate the other hits is 7 heightened because we're into a replicated advantage. 8 So I think that final assumption is just 9 not correct. Once you've initiated a process, you're 10 on a cascade of other events are required, but they're 11 not necessarily independent anymore. 12 DR. LOEWER: That's a chronic observation, 13 yeah. 14 PARTICIPANT: Maybe I can ask you to 15 clarify. Are you rejecting the use of a quantitative 16 approach? Because as Andy just came up here and 17 pointed out, the alternative to at least attempting 18 some kind of a quantitative approach is to simply go 19 based on your gut feeling, and our question is: is 20 that good enough? 21 So while maybe these numbers give us too 22 much reassurance, which I think is what you're 23 implying, the question is if these numbers came out 24 differently and this number were ten to the minus five 25 or something like that, wouldn't that mean we need to 282 1 go back and work a little harder? 2 DR. LOEWER: So that's really my personal 3 opinion. I would be very careful with numbers in this 4 respect because we are so many ambiguities, and I 5 personally prefer more qualitative approach to this 6 question, but this, of course, can be discussed. 7 John. 8 DR. PETRICCIANI: I think the issue of 9 trying to quantitate to get a more firm fix on a given 10 situation is attractive and it's seductive. The 11 problem in applying a quantitative approach in this 12 area to me lies in the uncertainties of all the things 13 that go into that final number so that when you come 14 out with a number, it's mushy because you have so much 15 uncertainty in virtually every variable that you put 16 in it. 17 It doesn't mean that it's of no value, but 18 I think at least in my mind what you'd come out with 19 is a qualitative impression of whether you're on one 20 end of the spectrum or in the middle or at the end. 21 But I agree with you, Johannes, that in 22 this arena I'd be pretty reluctant to use a number 23 because it doesn't have enough meaning to it. 24 PARTICIPANT: That may very well be right, 25 but then the implication of that is that if you can't 283 1 say that you think these risks are very low, less than 2 one in a million or something like that, but you all 3 say, "Well, these risks exist. You know, we see these 4 things happening in tissue culture," there's a chance 5 that this stuff can happen. What you're really saying 6 then is if we don't have a way of getting a handle on 7 what these risks are, that really if the best we can 8 do is qualitative, then can we take these risks? 9 DR. PETRICCIANI: Well, let me go back to 10 two points that I made in my talk. Number one is that 11 in this analysis and all of the other ones, they are 12 absolute worst case. They're not best case by any 13 stretch of the imagination, and if you used what I 14 think is perhaps a more relevant piece of data, and 15 that is human tumor DNA as opposed to viral oncogene 16 DNA, you'd have zero up on the top because there isn't 17 a shred of evidence that I know of that human tumor 18 DNA causes tumors in any animal system. 19 The other point is that it would be 20 helpful -- you're getting tired of me saying this -- 21 but it would be helpful to have some basic generic 22 pieces of information like DNA dose response to help 23 actually get a better handle on those quantitative 24 questions. 25 PARTICIPANT: Yeah, I was going to save 284 1 this for tomorrow, but I'm getting more and more 2 irritated in a sense over actually the last ten years 3 when I last attended a meeting to discuss these issues 4 at which this kind of calculation was done by both me 5 and Howard Clemens' calculation. The same thing was 6 cited using Hsing-Jien Kung's data, and here we are 7 ten years still using Hsing-Jien's classical data when 8 for more than ten years the basic experiments to get 9 a handle on some of these numbers have been obvious 10 and doable. 11 They're not rocket science. It is really 12 very simple, basic stuff of taking this and that DNA 13 and treating it this and that way and injecting it in 14 this and that animal. Any one of us in this room 15 could design a suitable set of experiments, but nobody 16 has done it for ten years. 17 Why not? Why not? It could probably be 18 done on less than 1,000 mice and probably for less 19 than a million dollars spread out over three or four 20 years. 21 That may be a lot of money to the FDA, but 22 to the NCI this is what you ought to be supporting, 23 this kind of experiment. This is a very small amount 24 of money. 25 (Applause.) 285 1 DR. ONIONS: So again really echoing -- 2 David Onions -- echoing what John has just said, I 3 mean I do think those systems do exist, and I made a 4 case yesterday for a variety of transgenic models, and 5 I think those do offer the prospect of giving you the 6 kind of sensitivity that you need and to give you the 7 kind of dose response curves that John has referred 8 to. I think that is important. 9 I mean, you can force a system rather like 10 the V-SARC experiments and take the worst case, but 11 then you can go and do the experiment that I think 12 John has wisely said. That is, you take tumor DNA 13 from a tumor cell line or cell lines where you 14 actually use the substrates, inject those in, and 15 you're using systems that are highly primed to develop 16 neoplasia, and because of the time scale and because 17 of the sensitivity, you don't have to go into very 18 high numbers either. 19 So I mean, I think those systems exist. 20 You might have to use multiple models, but I think 21 they exist, and I think the reason they have not been 22 done is actually because the academics like me who 23 develop these systems do it for other reasons. We're 24 looking for new oncogenes. So that kind of experiment 25 just gets in the way. 286 1 But, I mean, if somebody now said that's 2 a useful experiment to do, then we should do it. 3 PARTICIPANT: I'd just like to say one 4 more thing about the numbers and the models. I'm a 5 very poor person to be talking about mathematics 6 because I can't add two and two, but I think from a 7 simply perception point of view, as an experimentalist 8 and now as a regulator, to get at the issues that Dr. 9 Petricciani was mentioning, we have to start 10 somewhere. We have to think about how to approach 11 these things from a quantitative perspective because 12 that's where the future is. 13 If we continue to do this on an ad hoc, 14 believe type basis, we're not going to make any 15 progress, and 15 years from now we're going to be 16 right back here addressing the same questions again. 17 We have to start somewhere. 18 As crude as the numbers may be, if we 19 begin to make the attempt, we can find out what we 20 know. We can find out what we don't know. We can 21 pose the questions to ask to improve the numbers as we 22 go along. 23 If we don't make that attempt, we're not 24 really making progress, and I think that has been the 25 sort of thing that has been driving us for the past 287 1 five or six months, is we've tried to struggle with 2 how to come to grips with thinking about these issues 3 and putting together some sort of approach that made 4 sense. 5 This is simply a beginning, and we're 6 going to make mistakes, but the thing that is, we have 7 to be very careful about the kind of mistakes we make. 8 We can make mistakes in calculations, but if we make 9 mistakes in delivering products that are unsafe, the 10 consequences of that are going to be incalculable, 11 both for the regulatory process and for the public 12 health and for the companies that are unfortunate 13 enough to be associated with that. 14 So I would, again, I would make the plea 15 that our job is to try to figure out the best models 16 we can at the time, realizing that they're not 17 perfect. They probably will never be perfect, but the 18 job is to start somewhere and attempt over time to 19 improve them as information becomes available. 20 DR. LOEWER: I didn't want to say that 21 quantitative considerations wouldn't be helpful. I 22 think they are definitely helpful to find out and to 23 evaluate the different factors which may influence the 24 whole process, but I have to admit the same ideas like 25 John came into my mind. 288 1 These issues have been discussed in the 2 early '80s, and in the meantime the number of 3 experiments which help to clarify these issues is very 4 limited, and maybe one of the outcomes of the 5 quantitative approach may be to design adequate 6 experiments to perform and to find solutions to the 7 questions we have posed. 8 I think we are already half an hour late, 9 and I think I shall conclude this afternoon's panel- 10 audience discussion. 11 I would like to thank the speakers again 12 and also all the members of the audience who have 13 contributed. 14 Thank you very much. 15 (Applause.) 16 (Whereupon, at 5:10 p.m., the afternoon 17 session was concluded, to reconvene at 7:00 p.m., the 18 same day.) 289 1 EVENING SESSION 2 (7:08 p.m.) 3 DR. PEDEN: Good evening. I think we'd 4 better get started since we have to be out of here by 5 ten o'clock. 6 This is a miscellaneous session, and the 7 topics are diverse, and at the end of the formal 8 presentations of the program, there are going to be 9 some shorter presentations, but so we'll see how we 10 go, how many we get through. We have to be out of 11 here by ten. 12 And I'm going to ask the speakers this 13 time to abide by the lights that now are going to 14 appear on the podium. So I've brought it down to 15 15 minutes. So perhaps we can get everything done by 15 16 minutes. 17 So I'm going to use the Chairman's 18 prerogative, although I have one other thing to say. 19 This is a very special day today. It only occurs once 20 every decade, right, and the question, I suppose, is 21 where were you when it last occurred. So you can be 22 thinking about that. 23 What I want to do first, if I may have the 24 first slide, I'm going to take the Chairman's 25 prerogative and -- is someone over there? Could I 290 1 have the first slide? All right, and the lights down 2 a bit. 3 Because part of the thing we have to deal 4 with is quantitation of things, what I want to do 5 first is just give a couple of slides briefly on 6 methods that we've been developing at the FDA on two 7 topics. 8 The first is the sensitive reverse 9 transcriptase assay called PERT that Joerg Schuepbach 10 coined as the product enhanced reverse transcriptase 11 assay, and we've adapted that to the real time 12 quantitative PCR machine, the taq man machine, and 13 just go on and give another example of an assay we're 14 developing to look at primate polyoma viruses as an 15 example of just what sort of quantitative assays we 16 can do. 17 Since it hasn't been covered yet, this is 18 the quantitative real time PCR assay, and it's a very 19 convenient assay, and what it depends on is a reporter 20 and a quencher, and when these are on the same probe, 21 the reporter does not fluoresce. 22 But then after PCR, with the taq DNA 23 polymerase, it gets excised because of the five prime 24 and three prime nucleus activity. It releases the 25 quencher, releases the fluorochrome, and the machine 291 1 registers that. So it's a real time PCR, and the 2 assay detects the cycle at which the product is 3 detected. 4 So in the first example we use the PERT 5 assay, and since, again, no one has presented this, I 6 just run it through briefly. It's the standard RT 7 assay where the reverse transcriptase either comes 8 from a retrovirus or purified RT, and the presence of 9 a known RNA sequence, for example, MS2 RNA and a 10 primer, you get a cDNA synthesized, and then the taq 11 polymerase by PCR, you can amplify this sequence since 12 it's known, and you get double strand cDNA products, 13 and then you can assay it by various methods. 14 And what we're talking about now is the 15 taq man PERT assay, the TM PERT assay, and I just want 16 to show you what this assay can do. 17 When we look at three different purified 18 reverse transcriptases from AMV, Maloney murine 19 leukemia virus RT and HIV RT, this assay -- and make 20 dilutions -- this assay, as you can see, is linear 21 over about six orders of magnitude, and when you use 22 equal amounts of the RT according to the manufacturer, 23 these two enzymes fall on the line, and the HIV, in 24 fact, is off the line, indicating that the assay, in 25 fact, if the units are correct as determined by the 292 1 manufacturer, is probably more sensitive on the MS2 2 RNA than it is on the synthetic templates. 3 But in any case, the assay is highly 4 sensitive, and it's linear over, let's say, about six 5 or seven logs, and we can detect, going down here to 6 make the calculation, we can detect the equivalent of 7 RT of a single virion, not infectious virion, just a 8 single virion. 9 So the sort of things we want to do or use 10 this assay for us you can assess the retrovirus 11 contamination in cell substrates, biologicals, and 12 vaccines. You can use this assay, and we're testing 13 to see how efficient this is, to monitor retroviral 14 clearances from such products as monoclonal 15 antibodies, and also detect and identify unknown 16 retroviruses. 17 So it's a very useful assay. The problem 18 with this assay, as everybody who's worked with it 19 knows, is it's so sensitive and it detects DNA 20 polymerase's activity. So you're never sure that the 21 assay you're detecting is, in fact, pure retrovirus, 22 a true retrovirus or DNA polymerase. 23 And we have not been able yet using this 24 assay to discriminate between those two. We're still 25 working on that. 293 1 In the second example using the taq man, 2 we've designed primers and probes for the primate 3 polyoma viruses for 40 BK and JC-V, in this case in 4 the late region of the genome, and then when we show 5 an example of this, we're using dilutions of SV 40 DNA 6 down to about in this case you can register ten 7 molecules of SV 40 DNA. This is a cloned SV 40 DNA. 8 Viral DNA is the same. 9 And in these blue dots are dilutions of 10 cos cell DNA which carries a single copy of SV 40 11 genome, and this should be one copy, two copies, ten 12 copies, and 100 copies, and as you can see, it falls 13 on that line. So from this we can conclude that, in 14 fact, cos cells do have a single copy of SV 40 DNA. 15 So what we've done so far is test the 16 sensitivity and specificity of each of these primer 17 sets. With SV 40 we can detect about one to ten 18 molecules reproducibly. With the BK-V primers we're 19 only at the moment at about ten to 100 copies per 20 assay, and in the JC-V it's about one to ten. No the 21 JC-V primers pick up also BK-Vs. So now we have to 22 both increase the sensitivity of this assay and try to 23 improve the specificity of this assay. 24 So just in summary, we can say that the 25 developed primers for the three primate polyoma 294 1 viruses, it can detect reproducibly between ten and 2 100 copies and sometimes down to one. The reaction is 3 linear over at least seven to eight logs, and primers 4 so far specific for SV 40 and BK-V, and JC-V primers 5 cross-react with BK-V. 6 And using a series of SV 40 transformed 7 mouse and hamster lines, we've shown that, in fact, 8 the assay can detect different quantities of the SV 40 9 genome per cell. 10 So this is just an example of sort of 11 quantitative assays that are being developed and are 12 going to be very useful in determining the level of 13 quantitation -- the level of contamination of various 14 products in cell substrates and vaccines. So that's 15 what I'm going to say, and other people will cover the 16 assay a bit later on. 17 I think we should move on to the first 18 talk, and that's by Bernard Meignier from Pasteur 19 Merieux, according to the program, and he's going to 20 talk about industrial experience with live polio 21 vaccine prepared in the VERO cells. 22 And, again, I would try to ask you to keep 23 to the 15 minutes. 24 MR. MEIGNIER: Can I have the first slide, 25 please? 295 1 I would like to thank the organizers for 2 their invitation to share with you this example of the 3 use of -- no, that's not the one. I'm Bernard 4 Meignier -- so to share with you that example of the 5 use of continuous cell line applied to the 6 manufacturing of vaccines. 7 When we have the slide, I wish also to 8 acknowledge the colleagues of mine that helped me 9 assembling that presentation, and I would also like to 10 say as an introduction that I was not personally part 11 of the development of the VERO cell system, nor of the 12 manufacturing of those vaccines. 13 However, I will take the liberty of using 14 the "we," "we do this," "we don't do this," "we did 15 this," or "we did that" -- 16 (Laughter.) 17 MR. MEIGNIER: -- just to express how 18 collective an activity could be. It's just the 19 liberty of language, nothing else. 20 Forward. It looks like this time ring. 21 Good. 22 So the VERO cell story starts in 1962. 23 The two Japanese scientists there on March 27th, one 24 to establish a system, a cell system where they could 25 grow SV 40 virus in a lytic system to work more easy. 296 1 They could establish one line. From that 2 line there was a derivation of various other sub-lines 3 in Japan, and then two years later Simizu brought one 4 of those branches to NIH, and again, in the States 5 there were various lines that were spread among some 6 laboratories. 7 One of those branches, LT3 and VDOT 8 (phonetic) ATCC was deposited by Drs. Hann and Riem at 9 pass 113, as you can see here. ATCC, characterized at 10 cells scale, the top (phonetic), and two, 121, and 11 from there can supply cells -- you have the code 12 number there -- supply cells at the level of per such 13 one in 24 (phonetic), and that's where in 1979 14 Institut Merieux, which then became Pasteur Merieux 15 Connaught for various mergers, that's where we 16 started; they started at the time. 17 And from that passage 124 derived as shown 18 here, primary cell bank, and then from there working 19 cell bank, and established a scheme of production, 20 production being done at the passage one and 42. 21 In parallel to that, one bank was 22 established at the level of the working cell branch 23 that was made available to WHO. That bank is 24 deposited, in part, in ATCC, in part at ICAC, and is 25 available or could be supplied upon request through 297 1 WHO. 2 Now, these cells, just a few words on the 3 technology. These cells are grown on microcarrier 4 cells. Microcarrier is a small particle on which the 5 cells are grown. 6 There were cells encourage dependent. 7 They would not grow unless you got them -- especially 8 they could not grow in suspension. So in our case, 9 the microcarrier is made of DEAE dextron with an 10 optimized charge. The brown (phonetic) name is called 11 cytodex, and it's made by Pharmacia. 12 And what you're seeing here is one of 13 those small particle sizes, about, say, 100 microns or 14 slightly more, and actually you don't see the 15 microcarrier. What you do see are the cells. This 16 was taken from a culture, and you do see the cells 17 that are entirely covering the beads. 18 The prime scale at which we are operating 19 is 2,000 liters, and a full fermenter contains when 20 it's ready for being effective, contains only half a 21 dozen trillion cells, which is a fairly large amount. 22 Now, those cells, to maintain the safety 23 subjected to quality control and all the activity of 24 development of the banks and all of the QC that goes 25 with them and all the characterization that was done 298 1 was established in relationship -- I should say in 2 collaboration -- with authority BWHO that has the 3 asset of regulations and guidelines with national 4 authorities, including FDA in this country. 5 The sets of QC that you do see here, it's 6 a busy slide, but just to say that some of QC are just 7 done once in a while. We are still working with the 8 same primary cell bank from the beginning. We make 9 seven working cell banks, while some other controls 10 are done routinely on each and every batch of 11 production. 12 And as you can see from the list -- I 13 won't read it -- but most of those assays which are 14 done both on the cell themselves and on -- I'm sorry 15 -- on the supernatants; most of those cells really 16 have to do with looking for adventitious agents, be 17 they bacteria, be they fungi, and mostly be they 18 viruses. 19 The basis is hem (phonetic) absorption 20 inoculation to animals and inoculation to cells. ADC 21 means human diploid cells. The PMKC are primary 22 monkey cells, and NPCC is cell, a continuous line from 23 monkey kidney. 24 Attention is paid also to assays to look 25 for tumorigenicity as shown here and here, and that's 299 1 done only theoretically when we make a working cell 2 bank. The cells are assayed, of course, on the cell 3 bank itself and are also assayed on a special pool of 4 cells that is made ten passage doubling after the 5 production level. That is to optimize the chance or 6 the probability that we have to pick up 7 tumorigenicity. 8 And, by the way of tumorigenicity, out of 9 the number of assays that were done, I selected this 10 table because it fits with things that were said and 11 discussed on several occasions during this meeting. 12 The test here consists of injecting 13 newborn rats with various amounts of cells, one of ten 14 million for rats, and keeping the rats 15 immunosuppressed by injection of antithymocytic 16 antibodies. 17 And they are then monitored for the 18 development of nodules at the site of injections or 19 metastases at the draining lymph node or metastases in 20 the lung. 21 And what you can see, there are two 22 messages basically. One is that -- it's not 23 unexpected, by the way -- but one is that the amount 24 of cells that are injected do actually influence the 25 results. This is really typical here of the sort of 300 1 limit of positivity of metastases, this one here, and 2 that always causes the question for the sort of 3 arbitrary design that you can take for a test. That 4 was said before. 5 And the other message is that clearly 6 something happens. The cells change with the level of 7 passages, and something happened between that level of 8 pass -- the passage that was tested and that level of 9 passage in the sense that before the subcu. nodules 10 were -- there were no metastases and histological -- 11 it's not historical. It's also historical, but it was 12 histological. The pathologist could see here 13 regressing signs on the lesions, on the nodules wide 14 beyond that passage or from that passage on, they do 15 see progressive nodules, and you do see metastasis. 16 That's not really original. It's been 17 said before, but it's always interesting to see 18 numbers. What I showed is really the set of pre-T 19 controller assays (phonetic) that are mandated by 20 regulations or guidelines or points to consider. 21 There are specifications attached to them, and they 22 have to pass for the banks to be qualified or for the 23 production to be accepted for release. 24 Now, in addition to that, a number of 25 assays were done on the banks at various levels using 301 1 one of several of the techniques that are shown here 2 to look for viruses or for viral sequences. 3 And the next slide shows the set of 4 viruses that were specifically searched. Most of the 5 QC assays are rather broad. You go for the unknown. 6 In that case, those studies were looking specifically 7 for viruses of simian region. For obviously reasons 8 we are dealing with simian cells and for some human 9 viruses that, say, were causes of concern for us and 10 for health authorities. All of those searches were 11 negative. 12 I did not mention that there is another 13 set of analyses that goes with your in-process 14 analyses, like the monitoring, the visual inspection 15 of the cell 12 passages day after day, et cetera. 16 That also brings useful information as far as the 17 status, the cleanliness, if you wish, of the cultures. 18 A few years ago, Florian Horaud published 19 in 1992 his attempts or his -- his attempts to look 20 for the expression by Northern Blotting of some of the 21 oncogenes. He was really trying to -- he wanted or he 22 wished at the time to try and explain how come that 23 those cells have, let's say, a normal profile and, 24 however, have an indefinite life span as far as we 25 know. In his analyses, there was no signal that was 302 1 beyond the background. 2 Now, just a few words on the way we use 3 those cells for the manufacturing of OPV. We choose 4 OPV for tonight because we also make IPV, inoculated 5 polio and rabies with the VERO cell system, but those 6 have been described several times before, published. 7 So it's a bit more fun with this one. 8 In principle it's very simple as 9 manufacturing. You take the VERO cells, and as was 10 mentioned from the working cell bank, say, about 200 11 million cells through five passages. The cells are 12 just trypsinized like regular culture from the beads 13 and then plate seeded, if you wish, on the more beads. 14 You go in five passages to the level of 15 use for the production. Once the cells, the culture 16 is established, it is seeded. It is inoculated with 17 one or the other of the virus types taking from the 18 working seed that has been also controlled for being 19 clean, what it claims, et cetera, and you just keep 20 the culture for a couple of days until the cells are 21 lysed and progeny viruses are released, and that 22 constitutes the harvest. 23 And there is a set of steps to concentrate 24 and purify that virus. That's a code for filtration, 25 two steps of concentration through pettycon (phonetic) 303 1 and ion exchange per chromatography. 2 The purpose of that is to concentrate so 3 that industrially we have manageable volumes of virus 4 suspension and purify so that we have clean material. 5 This suspension is tested or titred for 6 infectious titre, and based on the result of that and 7 knowing what regulation asks to be put in the 8 monovalent vaccine or in the trivalent vaccine, then 9 there is dilution to adjust the potency. It is 10 filtered, mixed with other valences and that is the 11 vaccine. 12 Again, the viruses or virus suspension are 13 subjected to a number of assays, and here, again, it's 14 the same pattern. There is a set of assays that are 15 done during the process as in-process control that 16 contain things that you will not see here because 17 these are the ones that are asked or requested by 18 regulation. 19 For example, in-process control has -- we 20 monitor the amount of DNA. It's not requested 21 officially, but it is monitored as a way to follow the 22 efficacy of the purification. We do monitor the 23 content of protein. We do monitor the mayput 24 (phonetic) test, et cetera. 25 The set of QC assays that were done, I 304 1 will just focus on the prediction level. We are not 2 really dealing with the virus itself tonight -- has a 3 set of analyses to show that we have the right virus, 4 what we claim we have, in terms of its polio on the 5 right site. 6 It is also to determine that we have the 7 right quantity, and then there is a number of analysis 8 to show that we have no more than what we claim. 9 There is no adventitious agents, and there is the two 10 analyses, RCT4 and level virulence in monkeys. The 11 purpose of those analyses is to show or to be sure 12 that throughout the manufacturing process the polio 13 virus, which is a live attenuated virus, again, did 14 not change in the pattern of attenuation. 15 Another important feature in the 16 development was to show that the virus, the vaccine 17 that was produced on VERO cells was equivalent to the 18 vaccine produced in primary monkey kidney cells, and 19 that, again, was done through a variety of approaches, 20 the clinical monitoring antibody references and the 21 pattern of isolations. 22 Just the last or the one before last, this 23 slide is just to show the power of that technology 24 which enables us now to produce 440 -- more than 400 25 million monovalent doses per year. That's the number 305 1 for '98. The number for '99 will be in the same 2 range. 3 More important than that, at the scale 4 that we operate, 2,000 liter scale, we have a capacity 5 for about ten times more. That will be impossible to 6 run with monkey primary cells. 7 And the last thing I would like to add is 8 that beyond suppressing monkeys for the production, 9 using that scale, using the scale that permeates the 10 VERO systems reduces also the use of monkeys for QC 11 because the neurovirulence test can be applied to much 12 larger batches. We need to use less monkeys on the, 13 say, dose or median dose basis. 14 And I think I will stop here. Thank you 15 for your attention. 16 (Applause.) 17 DR. PEDEN: Thank you. 18 Does anybody have a question they wish? 19 David? 20 DR. CASHMAN: Neil Cashman from Toronto. 21 A trivial question. Is the entire batch 22 of vaccine per year -- does that come out of the 2,000 23 liter production hit? 24 MR. MEIGNIER: One fermenter makes one 25 batch. That's the practical and the regulatory rule. 306 1 DR. CASHMAN: How many fermentations do 2 you do a year? 3 MR. MEIGNIER: In principle one a week. 4 One week provided you have the right set of fermenters 5 to feed the final fermenter. 6 DR. ONIONS: David Onions. 7 Bernard, incredibly impressive performance 8 in producing those number of doses. 9 MR. MEIGNIER: I tell them. 10 (Laughter.) 11 DR. ONIONS: But as this meeting is about, 12 you know, concerns of perhaps new adventitious agents, 13 could I just ask? We now know that lots of serum 14 batches have contained fetal calf serum batches and 15 calf serum batches contain bovine polyoma virus, and 16 we know that VERO cells are permissive, and there's 17 also good serological evidence that that virus is 18 potentially zoonotic. 19 Do you do any testing for that virus now 20 or what was the position? 21 MR. MEIGNIER: Yes. (a) It's the sort of 22 multi-level process. (a) The calf serum is obtained 23 from donor calves that are monitored for their health 24 and that fermenting (phonetic), as we say, hervs or 25 frons (phonetic) with sanitary status that is 307 1 monitored. 2 These batches of serum are tested for a 3 number of agents, BVD. I'm not sure for you -- BVD 4 testing. They have a set of -- 5 DR. ONIONS: Yeah. 6 MR. MEIGNIER: -- bovine viruses that they 7 look for, and third, the calf serum is irradiated 8 before being used. 9 DR. ONIONS: Okay. Oh, sure. 10 MR. MEIGNIER: We hope that if you go down 11 that way, it turns out without problem. 12 PARTICIPANT: Could I ask very quickly 13 what your seed virus is grown in, and is that in 14 monkey kidney or is that VERO cells? 15 MR. MEIGNIER: That's a tough one. 16 PARTICIPANT: I mean the point is if 17 you're going -- if you have a VERO cell system, you're 18 introducing whatever is in your monkey kidney cells if 19 that's your seed. On the other hand, if you have a 20 VERO cell seed, you've undergone a couple of passage 21 levels from the Sabin strains. So it's of interest 22 from two points of view. 23 MR. MEIGNIER: I must say I do not 24 remember. I believe, but I'm not sure. If you want 25 the exact information, I'll be glad to -- 308 1 PARTICIPANT: Okay. 2 MR. MEIGNIER: -- convey the question. 3 I believe it's done on -- it was done -- 4 new burns (phonetic) were prepared on VERO cells, but 5 my recollection is also that they kept the same level 6 of passages of the virus because of the concerns for 7 changes in givens (phonetic). 8 PARTICIPANT: Bidion (phonetic) from 9 Provirus. 10 Have you tried looking to see how much of 11 the virus which is produced by the VEROs are active? 12 MR. MEIGNIER: Active in terms of? 13 PARTICIPANT: In terms of nucleic assay 14 and also particles. 15 MR. MEIGNIER: No, I don't think that was 16 done. 17 PARTICIPANT: It was not, and have you 18 tried an NCO-free medium for the production? 19 MR. MEIGNIER: I missed the question. 20 PARTICIPANT: NCO-free medium for the 21 production? 22 MR. MEIGNIER: No. The cells are grown in 23 serum, in a medium that contains serum, and serum is 24 removed for the virus production. 25 PARTICIPANT: And comparing the adduced 309 1 matter with like using eggs or other kind of systems 2 for growing, how would you compared those, in vivo and 3 in vitro? 4 MR. MEIGNIER: I don't think eggs were 5 ever used for the polio production. 6 PARTICIPANT: Or any kind of -- 7 MR. MEIGNIER: If you compare the sort of 8 standard that other companies or other institutes 9 still use are primary monkey kidney cells or human 10 diploid cells. Now, the comparison is clearly in the 11 scale. In the capacity -- the main feature of the 12 system is that using the VERO cells of microcarriers, 13 one can apply the ferment of the technology so that 14 you can keep the cells because those particles are 15 small enough; you can keep them stirring culture 16 medium, and you can scale up relatively easily I would 17 say, not to say that it's easy, but relatively easy, 18 while future in retrovirals, for example, is extremely 19 tedious, and when you want to scale up, you just have 20 to multiply the number of bottles, the broken 21 incubators (phonetic), the people to handle them, all 22 by robots, but still it's tedious, cumbersome. 23 PARTICIPANT: So this is showing that all 24 of the virus which is produced is active more or less? 25 MR. MEIGNIER: I really cannot -- I don't 310 1 have data to really answer that one. 2 PARTICIPANT: Okay. Thank you. 3 PARTICIPANT: Do you have any evidence 4 that you have an interseception (phonetic) in oral 5 polio vaccine? 6 MR. MEIGNIER: I missed the question. 7 PARTICIPANT: Interseception rate in 8 children that received the polio vaccine. 9 MR. MEIGNIER: The immunization rate? 10 PARTICIPANT: No, interseception is 11 adverse event in vaccination regard. 12 MR. MEIGNIER: I don't think it was looked 13 specifically at the time. The only thing I can say is 14 we have, as mandated by law, there is a special 15 department that recalls side effects that all may be 16 related to vaccines, and despite the high number of 17 doses that are sold and used -- I guess they were used 18 if they were sold -- no record was made of side 19 effects associated with those vaccines. 20 But I don't think they ever -- there was 21 a specific study to look to address that question. 22 DR. MYERS: Martin Myers, the National 23 Vaccine Program. 24 Relative to what we're discussing here, do 25 you have any long-term follow-up on a cohort of 311 1 children specifically looking at malignancy attack 2 rates? 3 MR. MEIGNIER: No, there is no follow-up. 4 DR. EGAN: Bill Egan from FDA. 5 That was also my question about what the 6 clinical safety profile was that was studied with the 7 vaccine, sort of in keeping with this current meeting. 8 MR. MEIGNIER: Well, at that time the only 9 follow-up that there is is the regular market -- post 10 market, if you wish, surveillance and the reports that 11 doctors may make that they found one or the other 12 strange association. 13 And I agree that it does not account for 14 possible putative evidence that occur not only in 15 people who received the vaccine ten, 20 -- ten years 16 ago or more. 17 DR. EGAN: Okay. Thank you. 18 DR. PEDEN: Last one. 19 DR. MERTEN: Merten from Genethon in 20 Paris. 21 When I worked in the institute in Paris, 22 I developed different media for different cell lines, 23 and I want to respond to the question concerning cell 24 free (phonetic) media. 25 The virus that grow rather easily in cell 312 1 free media and stay attached as cells, NCO-free media, 2 and so you can grow them in serum free medium and 3 they stay always attached and you can grow them in 4 microcarriers. 5 And the titres you will get out with polio 6 virus, for instance, are more or less the same as when 7 you use a class production process from -- from area 8 (phonetic) for instance. The titres are about, for 9 the Type 1 virus, about ten to the eight P few 10 (phonetic) per liters. 11 DR. RUSSO: I think we need to discuss the 12 scale of your growth because in our experience in 13 Merck -- I'm Carlo Russo from Merck -- the VERO cells 14 in the serum free medium, large scale, are now growing 15 very well. 16 PARTICIPANT: Okay. This was a laboratory 17 scale because in Pasteur Institute and in Paris we had 18 only laboratory scales, and in the scales of 1.5 19 liters, using five grams per liter of cytotex 1, 20 cytotech 1, microcarrier, they attach very well, and 21 they grew very well. We had no problems with this. 22 DR. PEDEN: All right. Thanks. 23 Do you want to respond? 24 MR. MEIGNIER: The only comment is we have 25 tried to, too, to use the medium, and the difficulty 313 1 still, as you know -- I really wasn't aware -- the 2 difficult is still the passaging of the cells. 3 It's true that there was some good growth 4 in cell free medium, but in the industrial scale, the 5 kind of handling, the passaging or associating the 6 cells and passing them still would be difficult to 7 really master, if you wish. 8 DR. PEDEN: Okay. Thank you, Dr. 9 Meignier. 10 So the next talk is by Girish Vyas on the 11 proteins of replication-incompetent virions for HIV 12 vaccination. 13 DR. VYAS: This is a cartoon of the famous 14 virus, HIV. There's a lot more known about it because 15 our government has been spending almost $1 billion on 16 everything that we can think of about biology, the 17 structure of the virus. 18 Three things have been learned in the last 19 ten years. First is that this grows in unlike 20 Hepatitis B and Hepatitis C virus -- this virus can be 21 cultured both in cell lines, as well as in normal, 22 peripheral blood mononuclear cells, or PBMCs. 23 In fact, all of the structure of proteins, 24 particularly the envelope proteins, the gp120 and 25 gp41, these two proteins are critical in design of 314 1 vaccines, and the first generation of vaccines have 2 been designed with cloned proteins which have an 3 inherent problem of antigenic radiation that results 4 from traditional mutations that occurred in gp120 or 5 all replicating RN viruses have a propensity to 6 mutate. So that's the built in handicap with some of 7 the recombinant vaccines. 8 The background is that limited progress in 9 the development of a safe and effective AIDS vaccine 10 has been accomplished. The structure of HIV-1 11 envelope and some antibody neutralizing epitopes has 12 been defined. A successful AIDS vaccine should in 13 this, both CTL, cytotoxic T responses and neutralizing 14 antibody responses against primary virus isolates and 15 not laboratory virus isolates, and finally, expanded 16 studies of the vaccine's alarming progress in non- 17 human primates and in Phase 1 and Phase 2 clinical 18 trials. These include vaccination with DNA and 19 attenuated pox viruses which induce virus specific CTL 20 in non-human primates, and in combination with the 21 booster immunization with recombinant subunit vaccines 22 which induce neutralizing antibodies against primary 23 isolates. 24 Again, this background some naive guy like 25 us from blood banking field though that perhaps it may 315 1 be a good idea to have a perfectly human blood based 2 vaccine as an alternative, and let me just 3 recapitulate what are the ideal requirements for an 4 AIDS vaccine. 5 First is the efficacy in preventing 6 transmission by mucosal and parenteral routes. 7 Secondly, safety in profile with minimum 8 risk of adverse reactions, even unscreened, real world 9 populations. 10 Third is long-lived protective effect of 11 many years after successful immunization. This is a 12 wish list. 13 Low cost, allowing widespread vaccination 14 in developing countries. 15 Stability and ease of administration, 16 facilitating mass immunization campaigns in developing 17 countries with minimal infrastructure. 18 And finally, protective immunity against 19 diverse virus isolates, preventing need for many virus 20 isolate specific vaccines. 21 Now, we thought it's rather simple to take 22 advantage of the biological property of replication in 23 competent viruses or what John Holland used to call 24 defective, interfering particles, or DI particles. 25 Any RN virus each time it replicates makes 100,000 316 1 times perhaps more noninfectious viruses than a 2 replication competent virion, and this replication 3 incompetent virion or RIV is an acronym that we have 4 given predominantly all of the replicating virions in 5 all viral isolates. 6 So prepared, whole virus vaccine is our 7 approach, and triple inactivation of this primary 8 isolates of prevalent HIV strains, pooled and expanded 9 in vitro by full culture with peripheral blood 10 mononuclear cells as the cell substrate, and I 11 emphasize the word PBMC as a cell substrate. 12 As you know, each time you get a unit of 13 transfusion, four million American get 14 million 14 units of blood and blood products, and each time you 15 get a transfusion, you get basically 400 micrograms 16 equivalent of human DNA because these cells 17 disintegrate. They are perfectly useless and, 18 therefore, some time in near future blood bankers are 19 going to be actually excluding the peripheral blood 20 mononuclear cells from blood for transfusion. 21 So these cells as a substrate becomes 22 cheap, available, widely donated, and blood bankers 23 can uniquely do this. My own background is actually 24 in blood banking. 25 Secondly, instead of fetal calf serum in 317 1 culturing the virus, we have shown that human AV serum 2 works equally well and, in fact, probably slightly 3 better than fetal calf serum in HIV cultures. So it's 4 human blood based vaccine. 5 The prestored leukodepletion (phonetic) 6 makes sense because immunological and virological 7 hazards of transfusions are minimized, and quality 8 control systems are now in place in blood centers in 9 the country. Ten percent of blood donated in this 10 country is leukodepleted. I understand that in 11 England, they are going to start routine depletion of 12 leukocytes from blood for transfusion. 13 So this is a ready resource that's 14 available for using culturing virus. 15 Let me just focus that, please. 16 Okay. Well, I want you to focus your 17 attention on the fact that any virus isolate is 18 composed of the replication incompetent virions or in 19 vitro virions and the natural virions of the 20 replication incompetent virions, and for the purpose 21 of the vaccine, the most important things are gp120 on 22 the surface and gp41, the trans-brain antigen. 23 Both antigens are present in the 24 replication incompetent virions as well as in 25 replication competent virions. The built in advantage 318 1 of making a blood based vaccine is that in a given 2 unit of plasma, for every virus that is replication 3 competent, there are 100,000 virions of this type that 4 are replication incompetent, and by in vitro expansion 5 of this virus each time that we do the expansion, we 6 are creating more of this replication incompetent 7 virions in vitro, and that's really the principle that 8 we thought would be an advantage in making a blood 9 based human vaccine or a blood based HIV vaccine. 10 And the genetic analysis can be done with 11 prototype relevant strains by interlooplets (phonetic) 12 analysis, and this would be done at Uttering 13 (phonetic) Memorial Blood Bank in San Francisco. 14 The principles are that plasmas is 15 procured as the primary source of HIV, negative for 16 HCV and HBV, and the PBMCs -- I'm sorry. I'll come to 17 that -- plasma pooled and concentrated for viral 18 expansion in peripheral blood mononuclear cells 19 prepared from CMV negative blood donors and AB serum 20 is used as a supplemental medium. 21 So this is basically the principal 22 components of the culture system. The HIV is 23 cultured, is concentrated from primary virus 24 inactivation using a compound called Sino-wiring N 25 (phonetic). 319 1 Now, blue-green algae is a bacterium which 2 has an 11,000 Dalton protein that it makes, and this 3 has been cloned and sequenced and large quantities are 4 being produced by National Cancer Institute. Its 5 unique property is that it interacts with gp120 almost 6 irreversibly, but it only selectively inactivates the 7 replication incompetent -- I mean, sorry -- 8 replication competent virions. 9 So after the expansion is achieved, the 10 Sino virion is bound to magnetic particles used 11 Steptoff (phonetic) in biotin. That is, this is 12 biotylated. It's bound to Steptoff and encoded, and 13 it removes the infectious virions so that the leftover 14 virions, which are lot of viral protein, is actually 15 culture negative in vitro. 16 This work has been presented. Some of you 17 who may have been at the meeting that we organized in 18 San Francisco in March on blood safety know this data 19 that were presented there. 20 The secondary inactivation is specific for 21 inactivating residual RNA that is still present in 22 some of the replication incompetent virions, and this 23 can be achieved by nucleophilic means or with dyes, 24 such as dimethyl methylene blue or psorlens. Psorlens 25 is now in advanced area clinical files, and perhaps a 320 1 preferred compound that does not bind to any proteins, 2 but specifically binds to the RNA. 3 And then finally, the purification of the 4 inactivated virus from all the components in the 5 culture system, as well as the inactivating chemicals 6 is achieved by molecular C ring (phonetic). 7 After this is put into final containers 8 for terminal viral inactivation, the pressure cycling 9 technology, this is basically a hydrostatic pressure 10 that is repeatedly applied to the virions in the final 11 container that disrupts the virions and disintegrates 12 into nonprevalent bond proteins. 13 So basically that's where the word or the 14 title comes, that HIV proteins are produced through 15 this pressure cycling technology. 16 After that we put the adjuvant and the 17 immune response tested in mice shows that such 18 antigens are quite immunogenic. 19 Safety and efficacy testing is going to be 20 done in chimpanzees, and finally this is what I say we 21 are going to be doing. CGMP facilities for 22 manufacture of the RIV vaccine, as we call it, in an 23 FDA approved facility, and finally Phase 1 and Phase 24 2 clinical trials in humans are planned. 25 And finally, I want to show the 321 1 investigators at UC-SF are myself; Cliff Lawless, 2 Chief of Immunology; Manish Gandhi is a pathologist 3 and transfusion medicine specialist at our place. 4 Mike Boyd at NCI is the one who cloned and 5 provides the Sino virion N. 6 Mark Manak at the BBA Biotech is the one 7 who has the pressure cycling technology for 8 inactivation of the final product. 9 Mike Bush and Eric Diewart were both 10 responsible for purifying for genetic analysis of the 11 pool against standard strains of HIV. 12 And finally, the safety testing in 13 chimpanzees will be done in Southwest Foundation by 14 Chris Mouffy (phonetic). 15 And finally, I want you to make any 16 thoughtful suggestions you can make so that this 17 project has the maximum benefit of peer input. 18 Thank you. 19 (Applause.) 20 DR. PEDEN: Thank you for keeping on time. 21 Are there any questions? 22 PARTICIPANT: Can you say something about 23 where you HIV is actually going to come from Girish? 24 I mean, it seems to me that the blood banking people 25 have spent the last 15 to 20 years getting rid of 322 1 intravenous drug abusers and HIV infected gays, and 2 you're going to be out there hunting for them again; 3 is that right? 4 (Laughter.) 5 PARTICIPANT: I mean, are they saying 6 thank you to you for this? 7 DR. VYAS: Well, actually we have more 8 than adequate resource of plasma that is HIV positive. 9 Yes? 10 DR. BROKER: In the protocol, you had 11 mentioned possibly putting psorlens into the 12 preparation near the end; is that correct? 13 DR. VYAS: That's correct. 14 DR. BROKER: That's a potent mutagen. It 15 seems like you're kind of going backwards on this in 16 terms of safety. 17 DR. VYAS: Well, there are three 18 alternatives, and any one of the three, which is RNS 19 specific inactivation otherwise. That is, there is 20 RNA there. Nobody would simply accept final virion as 21 a primary inactivating agent as adequate. Certainly 22 my colleagues at FDA would not approve of it. 23 So we wanted to be sure that we include 24 one step which is specific for nucleic acid and does 25 not in any chemical way alter the proteins, and either 323 1 it can be a means that has been used for foot-mouth 2 disease, virus, for many, many years or it could be 3 psorlens. 4 Now, we have not completely made a 5 decision as to which of the three alternatives we are 6 going to go, but putting inactivation of HIV has been 7 successfully done with psorlens, and then in the final 8 stage, they are able to remove the free chemicals to 9 molecular series. 10 DR. BROKER: Yeah, I just would point out 11 that there have been some adverse effects in using 12 psorlens in dermatology for psoriasis. It tends to be 13 used only in the extreme. It's an interpolating dye, 14 and if it for any reason passes through to the 15 patient, I'd be quite concerned. 16 DR. VYAS: Your point is well taken. In 17 fact, the amount of psorlens that's found to the 18 residual nucleic acid is so exquisitely small that it 19 has no adverse effect so far, in fact, much less than 20 a unit of blood. 21 The clinical trials for virally 22 inactivated blood is in Phase 2 and Phase 3 clinical 23 trials, and those trials have not had any adverse 24 reactions as far as I know. So I'm not discouraged by 25 all the points that you made. We are aware of it, and 324 1 I think that since we are removing the chemical 2 inactivating agents, we probably are safe. 3 And in the amount that is there in the 4 residual inactivation, it's probably trivial. 5 DR. BROKER: Okay. Thank you. 6 DR. VYAS: Sure. 7 DR. PEDEN: That question was by Tom 8 Broker. 9 Please give your name when you ask the 10 question. In the back. 11 DR. RUSSO: Carlo Russo from Merck. 12 How stable is your inactivation? How 13 stable do you think it is going to be? 14 And you know, in the context of this 15 meeting where there is lots of concern in the way we 16 validate the product that we're going to use for a 17 vaccine, how are you going to validate your release 18 assay for your preparation? 19 DR. VYAS: I'll come to Merck for doing 20 that work. 21 (Laughter.) 22 DR. RUSSO: That's not a bad idea. We 23 have a terrific validation procedures, but I would 24 like to have an announcement from you. 25 DR. VYAS: I just want to mention that the 325 1 primary work of this nature has been done in my lab. 2 The major NIH support for this study is impending, and 3 if that gets approved, then the R&D work will be done 4 at the University of California and BBA Biotech. 5 The validation issues and the manufacture 6 issues are tertiary, and we have not yet come to grips 7 with those issues yet. 8 DR. LEWIS: Yeah, Lewis, FDA. 9 Have you given any possible thought to the 10 use of human stem cells as a source of substrate? It 11 might be less of a problem. You might be able to 12 generate them in large volumes and from a very limited 13 source that would be a more amenable testing than 14 would be pooled human peripheral blood monocells -- 15 mononuclear cells. 16 DR. VYAS: Our major consideration was 17 actually having the Third World countries' ability to 18 make this vaccine. I think the stem cell biology and 19 propagation in vitro is still not available in the 20 majority of Third World countries. 21 So leukofiltration is one way of getting 22 rid of leukocytes, and we have actually been able to 23 get leukocytes out of the filters and been able to 24 culture them and grow HIV into that. So that's a 25 first step. 326 1 The point out stem cells is a very good 2 point, and I think ultimately that would be one of the 3 ways that we want to go about it. 4 DR. JOHN LEWIS: Lewis from Merck. 5 I'm curious to know if you know what the 6 physical product of your high pressure dissociation is 7 and whether you have quality control assays that would 8 assure the consistent manufacture of that. 9 DR. VYAS: Unfortunately you were not in 10 San Francisco in March when this work was presented by 11 the BBI people, and they have the protect technology 12 and the quality control in place for pressure cycling 13 technology and validation that introduces six logs of 14 HIV infectivity. 15 As such, after the primary inactivation, 16 in vitro we cannot demonstrate any infectivity in the 17 vaccine. The second inactivation with psorlens or MEs 18 is an added, and I call it tripolene (phonetic) 19 activated with special cycling technology. 20 But the immunogenicity is the main 21 criterion we are using in the animal model, and as 22 long as it is immunogenic and it produces neutralizing 23 antibodies, those are our standard criteria for useful 24 product. 25 DR. PEDEN: Okay. Thanks. We have to 327 1 move on. 2 The next talk is by Alex van der Eb, 3 unusual response to apoptin of diploid fibroblasts 4 from cancer prone syndromes. 5 DR. VAN DER EB: I would like to talk 6 about apoptosis producing protein that we call 7 apoptin, and which some of you may have heard, and the 8 unusual effects of apoptin on diploid cells of certain 9 cancer prone individuals. 10 However, before I do that, I have to 11 summarize some of the properties of apoptin, and if I 12 can have the first slide, please, apoptin is a 13 protein of the chicken virus, chicken anemia virus. 14 The chicken anemia virus is a small avian pathogen 15 that causes a lot of economic damage in the poultry 16 industry. It causes clinical signs in young chickens, 17 although older chickens are immune for the virus. 18 Anemia, it causes anemia, and by 19 destruction of the erythroblastoid cells, 20 immunodeficiency by depletion of the thymocytes. 21 And in the early stages of the work, which 22 was actually intended to try and make a vaccine 23 against the virus, we found that CAV causes these 24 disease symptoms by inducing apoptosis. 25 In order to clone the virus in the final 328 1 DNA was a circular, single stranded genome which 2 encodes a single MRNA, and on this MRNA there are 3 three open reading frames, coding for what we call 4 VP1, VP2, and VP3. 5 So we asked the question: which of these 6 three open reading frames causes the apoptosis? And 7 we transfected cells with each of the open reading 8 frames and found that it was VP3. So the smallest of 9 the three proteins causes apoptosis. 10 So how does VP3 causae apoptosis? There 11 are many pathways of apoptosis, and the next slide 12 shows you the major pathway that is mediated through 13 P53, when P53 is properly activated will cause a 14 signal that causes activation of the CAV spaces. IS 15 is one of the CAV spaces, and that will then 16 eventually lead to apoptosis. 17 There are several proteins that regulate 18 this process, like VCL2 and BAG1, which are negative 19 regulators. 20 So we asked the question which does 21 apoptin, as we call this protein, that is encoded by 22 VP3 -- does that need VP3? Is it inhibited by VCL2? 23 Does it need half spaces for apoptosis induction? 24 And the answer was initially rather 25 surprising. P53 is not needed. Also cells without 329 1 P53 go in apoptosis, VCL2 and BAG 1, which are strong 2 anti-apoptotic proteins, do not inhibit apoptin 3 induced apoptosis, and a number of CAV space 4 inhibitors, including CrmA, did not prevent apoptosis 5 induction by apoptin. 6 So that was surprising and indicated that 7 apoptin had an apoptosis pathway of its own. However, 8 recently we found that the P35 gene, an anti-apoptotic 9 gene or maculovirus and which is an inhibitor of CAV 10 spaces, of the downstream CAV space also inhibit 11 apoptin, indicating that apoptin also activates the 12 CAV spaces. 13 When we did these studies, we realize that 14 what we ha tested so far, what we had tested in 15 initial studies was a number of cell lines that all 16 went in apoptosis, and all of these cell lines were 17 actually neoplastic cells. 18 And so what about the non-neoplastic, 19 normal, diploid cells? And then came a surprise. It 20 turned out that the normal cells, diploid, fibroblast 21 endothelial cells, T cells, smooth muscle cells, et 22 cetera do not go in apoptosis. They are completely 23 resistant. 24 That was a surprise, and why would apoptin 25 not cause apoptosis? Well, it turns out to be the 330 1 localization probably of the protein. 2 The next slide shows you a tumor cell, an 3 osteosarcoma cell which is transfected two days ago 4 with apoptin, and you can see here the protein, the 5 fluorescence of the protein in the nucleus, and it is 6 just present in the nucleus throughout the nucleus, 7 and the DNA staining of the same cell shows that 8 everything is still normal. 9 However, the days after this second day, 10 an increasing number of cells appeared that have a 11 condensed and fragmented fluorescence as you can see 12 here, and if you look at the DNA staining, you can 13 also see that it is abnormal, and this cell is an 14 apoptotic cell. 15 So this happens in cancer cells. So what 16 about the normal diploid cells? Well, it turned out 17 that the apoptin is not present in the nucleus, but 18 stays in the cytoplasm, and even if you looked many 19 days after transfection, it still stays in the 20 cytoplasm. 21 They even have transgenic mice now that 22 express apoptin in a number of tissues and everything 23 seems normal. So it is not toxic for normal cells. 24 Now, what would happen if you transformed 25 these cells with SV 40, for example? We have the same 331 1 cells formed by SV 40, and then we are completely 2 sensitive to apoptin. So what would happen if you co- 3 transfect normal diploid cells now with SV 40, the 4 antigen and transforming gene that completely 5 transforms the cell? 6 And then it turns out that these cells 7 become sensitive. So here is the diploid VH10 8 fibroblasts that are co-transfected with apoptin and 9 the SV 40 P antigen, and here you see that the cells 10 that were transfected with apoptin 3T3 alone and 11 neofactor did not go in apoptosis, whereas cells that 12 were transfected with VP3 and actually 4T, large T, 13 here went in apoptosis and will eventually reach the 14 100 percent if you wait long enough. 15 So this means that also a transient 16 expression of an oncogene can cause sensitization of 17 normal diploid cells to apoptin. 18 The next question that we asked was: what 19 would happen if you treat normal diploid cells with a 20 carcinogenic agent? And since we have an irradiation 21 program in the lab, we irradiated with ultraviolet 22 light or with actually ionizing radiation, and we 23 found that if you do that, the diploid cells still 24 remain completely insensitive to apoptin, and this is 25 actually what was expected because a single radiation 332 1 exposure will not transform the cells, of course, in 2 a single step, and maybe you induce a few mutations 3 and no more than that. 4 So up until recently we have found a 5 category of diploid cells from cancer prone patients 6 that do react with apoptin, and before I go into that 7 part, I have to briefly introduce something about 8 irradiation or radiation program or radiation 9 research. 10 Oh, incidentally, this is what happens if 11 you transfect with T antigen a normal cell, T antigen, 12 and apoptin, and this is two days after transfection. 13 In the normal cells after one day the apoptin is still 14 present in the cytoplasm, but after two days -- and 15 this is after two days -- there is a kind of a 16 transfection. You see that it moves into the nucleus. 17 So it is actually the movement of the protein, the 18 transport of protein to the nucleus that may be 19 essential, and one day later, everything is in the 20 nucleus. 21 And now radiation. It is, as you probably 22 know, that radiation of cells, particularly UV 23 irradiation, causes a large number of responses, most 24 of which are transient, such as activation of plasma 25 membrane, activation of genes that are normally 333 1 induced by growth factors, secretion of growth factor, 2 stabilization of P53, and so on and so on. 3 Also, ultraviolet light induces so-called 4 SOS-like phenomena, such as a reactivation, an aus 5 (phonetic) mutagenesis, ER and EM. They are called 6 like that because these phenomena they resemble, the 7 SOS responses that were first described in E. coli in 8 response to irradiation. 9 Now, we asked the question, and this is 10 several years ago: do these, in particular, EM 11 somehow contribute to carcinogenesis? And we decided 12 to try and study that. 13 First I would like to show you what ER and 14 EM are. ER, an ounce (phonetic) reactivation, and EM 15 are both measured with viruses. They are either 16 measured with DNA viruses, like HSV, or actually for 17 virus and ER, an ounce (phonetic) of reactivation, is 18 the phenomenon that your re-irradiated virus survives 19 better in your re-irradiated cells than in 20 unirradiated cells. In fact, it survives about two 21 times better. And that means that ER by definition is 22 two. 23 EM also makes use of the virus, but in 24 this case the virus is not irradiated, and if you 25 allow it to replicate this unirradiated virus and you 334 1 re-irradiate it or unirradiated cells, it turns out 2 that the mutations that are accumulated in the virus 3 are about two to three times more higher in your re- 4 irradiated cells, indicating that your re-irradiated 5 cells have a kind of mutator activity, and that is 6 about two or a little more than two. 7 The effect is transient, and here you see 8 ER, an ounce reactivation, in a cell. It returns 9 after a few days going back to one, and EM also 10 returns to one, and they have a maximum expression of 11 one day after irradiation. 12 So do any of these phenomena play a role 13 in carcinogenesis? So we decided to measure ER and EM 14 in cells and diploid cells from cancer prone 15 individuals, and we have measured two different cancer 16 prone individuals, and in one of which we saw 17 abnormalities, and not with EM, as we expected, but it 18 was with ER. 19 And ER turned out to be extremely high in 20 diploid fibroblasts from individuals who are cancer 21 prone as a result of a germ line mutation in a tumor 22 suppressor gene, and here you see, for example, 23 Aniridia cells from an Aniridia patient, just normal 24 diploid cells who have a mutation in one of the 25 Willems tumor genes, and as you can see here, the ER 335 1 is extremely high. 2 Now, this is an only single point, but you 3 can believe me that this is reproducible, and the 4 correct slide with more points in the curve I couldn't 5 find, and it's always disappeared in the lab. 6 So surprisingly this was ER and not EM 7 that was correlating with cancer prone. So we then -- 8 this is a list of the syndromes from which the diploid 9 fibroblast show the extremely high ER phenomenon, ER 10 super plus, as we call it. Retinal blastoma, Li 11 Fraumeni, Euro Frondromitosis (phonetic), Willems 12 tumor, et cetera, et cetera, et cetera, and we have 13 measured now more, and all of them show this 14 abnormally high ER phenomenon. 15 And then we decided to test these cells 16 also for their sensitivity to apoptin without and with 17 radiation, and we have -- and this is quite recent -- 18 we have three of these cell strains in culture: Li 19 Fraumeni cells that have one inactive LEU (phonetic) 20 and one wild type LEU, and they are ER super plus. 21 A Lynch Type 2 syndrome familiar in breast 22 and ovarian cancer, ER super plus. The gene is in 23 this case, as far as I know, still unknown. 24 And there was a dysplastic Nevus syndrome, 25 a familiar melanoma, which is also ER super plus and 336 1 which has mutated P16. 2 Each of these cells, these two cell lines, 3 we have two different individuals from the same 4 family, and for the Li Fraumeni syndrome, we had one 5 patient who was affected and one individual from the 6 family who had two wild type alleles. 7 The first experiment already showed that 8 the normal irradiated fibroblasts are completely 9 resistant to apoptin. This was actually not 10 unexpected, of course, but after irradiation with 11 ultraviolet light, we found quite a surprise, and that 12 was that these cells become completely sensitized to 13 apoptin just as tumor cells. 14 Here you see the reaction of the PH10 15 normal diploid cells, which remained resistant to 16 apoptin, and they remain so much longer, as long as 17 you can keep the cells alive. 18 However, this is a Li Fraumeni cell, and 19 that is the Lynch Type 2 syndrome, ovarian/breast 20 cancer syndrome. They become completely sensitive 21 within four to five days. 22 And the next slide shows the more complete 23 study of all the cells together, and in the left panel 24 you see the cancer prone syndromes here, fibroblasts 25 from the cancer prone syndromes, and here first the 337 1 cells, the normal cells, which have nothing special or 2 the Li Fraumeni patient, the Li Fraumeni individual 3 who had two wild type, family member who had two wild 4 type P53 molecules, genes. 5 You can see that these remain completely 6 resistant to apoptin after re-irradiation, whereas the 7 other ones go completely in apoptosis as if they are 8 cancer cells. 9 In contrast, if you transfect and you re- 10 irradiate the cells with a desmid control gene instead 11 of apoptin, then all of the cells remain completely 12 resistant and do not go into apoptosis, indicating 13 that the effect is specific for the 3T3 or apoptin 14 gene. 15 The next slide shows just a dose response 16 and where you can see that at ten Joules you already 17 get the maximum reaction of the Li Fraumeni cells in 18 this case, whereas at five Joules there is still 19 nothing. 20 If you transfect with desmid, again, 21 nothing happens, and if PH10 normal cells, even at 25 22 Joules, show no reaction whatsoever compared to 23 transfection with desmid. 24 And then the last slide shows you here 25 that the effect on these cells from cancer prone 338 1 individuals is transient, like many of the other 2 injury induced phenomena. It completely disappears 3 again after four to five days and returns to zero, to 4 the same level as the PH10 control cells. 5 So in conclusion, the diploid cells of 6 these cancer prone syndromes that have one allele of 7 a tumor suppressor gene have two abnormal properties 8 and have lost one allele tumor suppressor gene, have 9 two abnormal properties. 10 First, they show very high ER response, 11 and, second, they express what are also found in 12 cancer cells and which render them susceptible to 13 apoptin, which is a cancer specific protein. 14 And essentially it means that apoptin is 15 allowed to be transported to the nucleus, and both 16 properties, irrespective of the tumor suppressor gene 17 because it's found in many different tumor suppressor 18 genes now also irrespective of the type of 19 irradiation. 20 We are trying now to identify what is 21 induced by radiation in these cells and that is 22 different from what is caused by irradiation in normal 23 cells. 24 (Applause.) 25 DR. PEDEN: Thank you. 339 1 And we have time for a couple of 2 questions. 3 DR. FRIED: Did you say you had a 4 transgenic with -- 5 DR. VAN DER EB: Yes. 6 DR. FRIED: Does it get cancer? 7 DR. VAN DER EB: This is something that we 8 still have to study. The transgenic mice exist, and 9 it is expressed particularly in the lymph nodes and 10 other lymph organs, and what we have to do is see what 11 happens if you, for example, try to induce lymphoma by 12 leukemia viruses. They should be resistant. 13 PARTICIPANT: That was Mike Fried. 14 Could you give your name, please? 15 DR. CASHMAN: Neil Cashman. 16 You've adroitly avoided any mention of the 17 sequence of this gene. 18 DR. VAN DER EB: Right. 19 DR. CASHMAN: Can you comment at all? 20 DR. VAN DER EB: Yes. 21 DR. CASHMAN: Are you free to comment? 22 DR. VAN DER EB: I have the slide with me, 23 but of the protein sequence of the gene. It's the 24 small protein, 120 amino acids. It does not resemble 25 any other known protein in the cell bank. It has two 340 1 nuclear localization signals, in the right-hand C 2 terminal part, and it has a nuclear export signal in 3 the internal part, and that's about it. There's 4 nothing else, no. 5 I have it. It's published also. 6 DR. COFFIN: John Coffin. 7 Ordinarily one thinks of induction of 8 apoptosis as an antiviral mechanism. 9 DR. VAN DER EB: Yeah. 10 DR. COFFIN: Can you speculate on why it 11 might be of value to the virus to do such a thing? 12 DR. VAN DER EB: I have no idea. This is 13 a good question. 14 One possibility is that this virus is a 15 very simple virus. It has only three genes, and it 16 has no other means to get out of the cell, and then 17 the best way to get out of the cell or to infect other 18 cells is to induce apoptosis. Then the cell fragments 19 and is taken up by other cells, and maybe that's the 20 way it's expressed. I don't know. No, I don't know. 21 DR. PEDEN: Thank you, Alex. 22 The next speaker is Tom Broker from UAB, 23 and his title is "Human Papillomaviruses: a Window 24 into Eucaryotic Cellular DNA Replication Mechanisms 25 and Regulation." 341 1 DR. BROKER: Okay. Thank you very much. 2 I'd like to share with you several things 3 regarding papilloma virus DNA replication, and what 4 I'd like to do is start by very briefly reviewing two 5 aspects of HPV biology. 6 The first is that the virus infects a 7 squamous epithelium through a wound allowing direct 8 access of mature virus particles from the environment 9 to the germinal layers of the skin, typically the 10 basal cells or the parabasal cells. 11 The virus would be harbored in these cells 12 and maintained more or less indefinitely by 13 replicating in synchrony with the normal cell cycle, 14 and it would be latent in that capacity in most cases. 15 Occasionally induction of viral early, as 16 well as late gene expression can commence at about the 17 third layer up where it becomes so the so-called 18 spinus epithelium with the delayed early, late 19 transcription, viral DNA replication occurring here, 20 and assembly into mature virus particles, and then 21 naturally a release through desquamation of the 22 cornified envelopes. 23 Now, periodically the genes that normally 24 affect the onset of reentry into S phase can 25 inadvertently become expressed in cells that have not 342 1 yet withdrawn from the cell cycle, and that can lead 2 to dysplasias, and through an accumulation of cellular 3 mutations and probably some viral mutations, you can 4 progress over a period of years toward higher grade 5 dysplasias and carcinoma in situ. 6 I want to emphasize that there are several 7 modes of replication of HPVs. They're in the wound 8 healing phase when there is literally a stripped off 9 epithelium. There is a re-epithelialization across 10 the wound bed that originally allowed access in the 11 virus to these layers, and at that time there's a 12 transient up regulation of viral oncogenes E6 and E7. 13 However, once the full thickness skin has 14 been reestablished, E6 and E7 transcription stops at 15 this level, and I'll show you that, and the E6 and E7 16 oncogenes only are transcribed from this layer up in 17 a full thickness epithelium. 18 So we have establishment replication 19 during the first phases of wound entry; secondly, 20 maintenance replication here; and third, vegetative 21 replication there higher up. 22 This is the HPV18 genome. It's pretty 23 typical of most viruses. The proteins that I'll be 24 discussing today are the origin binding protein E1 and 25 E2. What you'll also see is E2 plays a role in 343 1 transcriptional regulation at the promoter right 2 there, and that the E1 is the only enzyme encoded by 3 this virus and is the preferred target, therefore, for 4 drug discovery. It is a helicase with some remarkable 5 properties. 6 It together as a multimer binds to the E6 7 promoter region, which is synonymous with the 8 replication origin right here. E7 protein, which I'll 9 talk about a bit today and a bit tomorrow, interacts 10 with the retinoblastoma protein and triggers S phase 11 entry, and as many of you know, the E6 protein 12 interacts with P53 and blunts some of the repair 13 responses and also can blunt movement toward 14 apoptosis, allowing the virus time to replicate. 15 Now, in a normal, full thickness, squamous 16 epithelium, in this case cervix, without any viral 17 infection, I want to point out that the only DNA 18 replication that the normal cells are engaged in are 19 the parabasal or transid amplifying cells, as marked 20 by PCNA stain. 21 All cells above the parabasal layer do not 22 have the oxyribonucleocide triphosphates, DNA 23 polymerase, topoisomerases, PCNA, or any of the other 24 replicative enzymes. Yet the virus replicates up 25 here, and that is the dilemma the viruses have to 344 1 face. How do they reactivate S phase entry and the 2 reinduction of all the replication machinery to 3 support their own replication? 4 This is one piece of evidence from a 5 natural laryngeal papilloma virus from a child in dark 6 field illumination, and the probe we used was 40-67 7 transcripts, and as you can see, they're basically 8 excluded from the basal cells. This is a dermal read, 9 and the signals are in the spinus layer of the 10 epithelium for E6 and E7 once the wound healing has 11 been complete. 12 Okay. Now, I very briefly want to remind 13 you of the normal cell cycle and that it is controlled 14 by the retinoblastoma protein being bound to the E2F 15 enhancer protein, which itself is in complex with a 16 cofactor called the DP1 or DP family of E2F cofactors. 17 That would be the normal state in the G1 18 phase or G0. However, upon mitogenic signals, cyclin 19 dependent kinases, primarily cyclin D and DCK4, can 20 phosphorylate RB, causing it to be released from E2F, 21 and then E2F DP1 complexes can up regulate the 22 transcription of the replication enzymes and certain 23 cell cycle progression proteins like cyclin E. 24 Conversely, the papilloma viral E7 protein 25 can disrupt this association in the absence of the 345 1 kinase activity, and therefore, the virus can trigger 2 the up regulation of the necessary enzymes for its own 3 synthesis. 4 And this is shown here in a couple of 5 examples. This is a genital condyloma, and I want you 6 to see in this cross-section of the papilloma that, 7 again, the basal cells do not have PCNA. The 8 parabasal cells do. Then the lower spinus cells do 9 not have PCNA, and the upper spinus cells, once again, 10 do have PCNA. 11 So you have this natural cell cyclin here, 12 and the viral induced S phase reentry toward the upper 13 layers of the epithelium. 14 This phenomenon is also seen with other 15 important enzymes, in this case DNA polymerase alpha 16 in the natural laryngeal papilloma; again the tritium 17 or S35 silver grains are predominant in the upper 18 layers of the skin, and I want to particularly 19 emphasize that there's a massive over expression of 20 the polymerase, as there is with PCNA, in the upper 21 layers. That will become particularly important. 22 It's something I'll have to say probably in tomorrow 23 morning's talk. 24 Now, the second thing Andy asked me to 25 comment on is some of the model systems that we have 346 1 for investigating in an experimental way. This is the 2 epithelial raft culture, and very briefly, one makes 3 a collagen matrix, imbeds dermal fibroblast in this, 4 and then you can take either dispersed keratinocytes, 5 primary human keratinocytes or established 6 immortalized keratinocytes, and spread them on the 7 surface, and at that point you can raise the entire 8 thing up into the air. 9 Think about your own skin. You have an 10 air surface, and you have a blood surface, and when 11 you do this, the exposure to air sets up a capillary 12 action, calcium gradients and other gradients of 13 probably mytogens, and so forth, and you quickly 14 develop a stratified and differentiated squamous 15 epithelium. 16 To show that this actually can work not 17 only with disbursed cells, but with punch biopsies, 18 we've been able to take skin from any anatomical site 19 in the body, place it on the raft, which is right 20 there. This is a punch biopsy from a excess cervical 21 epithelium, and these cells will grow out across the 22 collagen surface and very rapidly stratify and 23 differentiate appropriate to the tissue of origin. 24 You can do the same thing with papilloma 25 itself, and you always recapitulate the type of skin 347 1 you started with. In this case, the natural laryngeal 2 papilloma will grown out and become an in vitro 3 papilloma. 4 The third thing we've been able to do is 5 introduce papilloma virus genes or reporter genes into 6 the keratinocytes through retrovirus mediated gene 7 transfer process. Denise Galloway and Jim McDougall 8 did this in which they placed the papilloma virus 9 E6/E7 oncogenes directly under the control of the 10 upstream LTR. 11 With Louise Chan in my laboratory, we 12 chose to place the reporter gene or the E6 or E7 genes 13 under the control of their own promoter at a rather 14 downstream location, in which case we were able to 15 very exactly recapitulate the differentiation 16 dependent express of the E6 promoter. 17 Parenthetically I would note because it 18 reflects a change in thinking, people had always felt 19 that the loss of an E2 gene upon HPV integration would 20 invariably lead to the up regulation of the E6 21 promoter, which it can repress in basal cells. This 22 construct has no E2 gene in it, and indeed, the E6 23 promoter driving LAX-E (phonetic) in this case is only 24 in the upper layers where we expected it to be. 25 So E2 has absolutely nothing to do with 348 1 the repression of the E6 promoter in these types of 2 cells, unless other cellular changes have occurred, 3 which in this case they're PHKs, and they haven't 4 engaged in those mutations yet. 5 To show that this system works and that we 6 can do genetic dissection of HPV, we have an empty 7 vector here. We have an E7 expressing vector here, 8 and as you saw in the natural papillomas, polymerase 9 alpha is up regulated; PCNA is up regulated; and 10 bromodeoxyurasyl incorporation, all occur in the upper 11 compartment of the full thickness skin, which 12 developed in just one week after the viral infection 13 and lifting of the rafted air surface. 14 And notice again the basal cells tend to 15 have low grade expression of these induced genes. 16 They're highly up regulated in the differentiated 17 compartment. 18 What we noticed upon doing a variety of 19 analyses is that while many of the cells do have up 20 regulated PCNA, only a subset of them are capable of 21 incorporating bromodeoxyurasyl in response to these up 22 regulated genes. This will become important for 23 tomorrow's talk, but notice that we always compare a 24 natural papilloma with an E7 raft culture, and only a 25 subset of the superficial cells become competent for 349 1 DNA replication. In other words, all warts, as well 2 as rafts, are a mosaic of cells that are capable of 3 engaging in DNA synthesis and S phase entry, and 4 others which appear to be inhibited for reasons I'll 5 discuss tomorrow. 6 Now, I want to give the bottom line of 7 what we now know about the assembly of the replication 8 complex so you know where I'm going with several of 9 the data slides. 10 Very briefly, the viral E2 protein is the 11 main ORI binding protein, and as I'll show you, it 12 associates with the nuclear matrix. It then recruits 13 the viral E1 protein, the helicase, and one of the 14 most remarkable observations that we've made recently 15 in collaboration with Doug Seer and Jack Griffith is 16 that the assembly of this complex depends on heat 17 shock protein 70 and heat shock protein 40, and in 18 fact, heat shock protein 40, the co-chaperon for 19 HSP70, in fact, remains a permanent component of the 20 active replication complex. We presume it's true also 21 in general eucaryotic replication, but it certainly is 22 with the virus. 23 The E1 protein, in turn, can recruit DNA 24 polymerase alpha, and when that assembly has occurred, 25 cyclin E, CDK2 moves in, and as you'll see 350 1 phosphorylates this complex. 2 I would like to show you some of the data. 3 Very briefly, the E1 protein has a series of domains. 4 At this end is an ATPase, which is essential to the 5 helicase action. There's a DNA binding motif, and as 6 you'll see, very, very critical, a cyclin E binding 7 motif right here which has great bearing on our 8 understanding of Hela cells. 9 The E2 protein has three basic domains, 10 the DNA biding domain, a hinge, and a transacting 11 domain. As I'll show you in a moment, the hinge 12 anchors this entire complex to the nuclear matrix. 13 What we did is put green fluorescent 14 protein tags on each of the domains of the E2 protein, 15 full length protein and the various combinations of 16 one or two portions of the E2 protein. We were able 17 to show function was retained despite the GFP tag, and 18 bottom line is the hinge unexpectedly was the thing 19 that targeted the E2 protein to the nucleus. 20 If you didn't have the hinge domain, as in 21 this column, the proteins remain cytoplasmic. If we 22 did have the hinge, they got nuclear. If we had the 23 N terminal domain and the hinge together, we began to 24 see nuclear foci. If we had the whole protein, the E2 25 protein went to the pods or nuclear domain 10s of this 351 1 protein. 2 And we were able to identify the anchor to 3 the nuclear matrix as a basic motif in the center of 4 the hinge that's highly conserved among all papilloma 5 viruses. If we mutated that domain, we got a 6 cytoplasmic distribution. If it was wild type, it 7 went to the nuclear foci. 8 And this is the model then. The hinge 9 anchors the E2 protein to the matrix. The C terminal 10 domain binds to the viral DNA. The N terminal domain 11 associates with the matrix and also helps recruit the 12 E1 protein into the complex, as you can see here. 13 E1 protein and the bromodyoxyurasyl 14 replication foci are coincident. 15 Now, I want to show you one really 16 remarkable thing about the nuclear domain, about the 17 assembly, yeah, yeah, and that is that the DNA J 18 protein of E. coli and K protein are the things that 19 help the lambda ONP proteins assemble at the lambda 20 ORI. DNA J and DNA K also help the assembly of the E. 21 coli replication components, and they're highly 22 conserved all the way up to eucaryotic and human 23 cells. 24 This domain right there helps the helicase 25 assembly. This was done with Jack Griffith's lab at 352 1 UNC. In the absence of the J protein, the E1 protein 2 by itself can form a hexamer ring around the DNA, 3 either a super coil or an open circle. You'd throw 4 the HSP40 in, and you get a dihexamer, which we 5 believe is a bidirectional helicase. 6 Secondly, the E2 protein helps recruit the 7 RPA, the single stranded binding protein, and we feel 8 this assembly of E2, E1, the J domain right here 9 creates the preinitiation bubble at which you end up 10 with four replication strands, two leading strands, 11 two lagging strands, and the complex bound with RPA. 12 I think I'll stop at this point, and I'll 13 be able to show you the role of cyclin E tomorrow. 14 Thank you. 15 (Applause.) 16 DR. PEDEN: Thank you, Tom. 17 We'll have maybe one question. 18 Well, I guess we have to move on. So the 19 next talk is by Paul Sandstrom, "Facilitated Detection 20 of Adventitious Agents Using Genetically Engineered 21 Cell Lines." 22 DR. SANDSTROM: What I'm going to present 23 tonight is somewhat of a revisiting of some previously 24 published data in an attempt to try to develop it in 25 the direction of the theme of this meeting, and that's 353 1 the detection of adventitious agents which may be 2 present in cell substrates that's used for vaccine 3 production. 4 Largely speaking, it's a proof of concept 5 model which demonstrates approximately a practical 6 application of a cellular phenomenon that's received 7 considerable attention over the last ten years, that 8 being program cell death. 9 I'm not sure what I'm pointing at here. 10 Okay. This project was done in the 11 laboratory of and under the direction of Dr. Tom Folks 12 from the HIV and AIDS and retrovirology branch of CDC, 13 and I myself have the, I guess, notoriety of being, I 14 think, probably the only Canadian citizen who's 15 employed by the American Public Health Service and who 16 is working for the Canadian equivalent of the Public 17 Health Service and stationed in Canada with all of the 18 rights and privileges that come with working for the 19 two largest bureaucracies in the Western Hemisphere. 20 (Laughter.) 21 DR. SANDSTROM: Okay. It's readily 22 obvious, I think, to probably everybody here that when 23 you go about trying to isolate a potential 24 adventitious agent, that the failure to isolate 25 doesn't necessarily mean the agent isn't there. It 354 1 just means that possibly the culture system wasn't 2 sufficient adapted to allow for the agent to be 3 isolated. 4 So what's needed for isolation is, number 5 one, it would be good if you had an adventitious 6 agent, or at least good depending on which side of the 7 fence you're on, that the virus has the ability to 8 bind, fuse, or otherwise gain entry into the cell; 9 that the wiring and forming of the cell are such that 10 it can support a complete cycle of virus replication; 11 and of particular importance or I should say of equal 12 importance, I guess, is that the cell is a good host, 13 and it stays alive long enough for the virus to 14 complete its replication cycle. 15 Apoptosis or program cell death is 16 involved, in part, not exclusively, but in part as a 17 first line host defense against pathogenic viruses, 18 and efficient and rapid apoptosis followed by viral -- 19 following viral infection possibly acts to restrict 20 virus replication and, if unchecked, will ultimately 21 result in an aborted viral infection, and this is sort 22 of the ultimate goal that the cell had in mind when 23 developing the programs. 24 So as a result of this intense restrictive 25 pressure which apoptosis places on virus replication, 355 1 many viruses have evolved to develop their own 2 repertoire of borrowed proteins, which act to block 3 the cellular programs. 4 It's this that I think we're interested in 5 here, that when trying to isolate an adventitious 6 agent, along with all the other elements that I 7 identified earlier, you're always going to be running 8 up against the possibility that the cells that you're 9 trying to isolate it in are attempting to undergo 10 apoptosis and, therefore, limiting your ability to 11 isolate the virus. 12 So to get around and to take a look at 13 this, what we did is to use the, I guess, historical 14 adventitious agent of HIV, coming from the HIV and 15 Retrovirology Branch, and we took a cell line which 16 was exquisitely sensitive to support HIV replication. 17 This is a sub T1 cell line. 18 One of the reasons we chose sub T1 is that 19 it is, or at least the clones that we were using 20 lacked expression of the very strong anti-apoptosis 21 gene, VCL2, and so what we did, quite simply, was to 22 take VCL2, place behind a constitutive promoter, and 23 to transfect it into the cells, and then go on to 24 select clones. 25 So what we have here is our positive 356 1 control. This is a Western Blot, positive control, 2 and two control clones which lack VCL2 expression, 3 which are paired with two clones which express 4 significant amounts of VCL2. 5 And we went on to analyze these and 6 characterize these in terms of the ability of VCL2 to 7 block apoptosis in sort of the nonviral, but 8 traditional, apoptosis models, and we were able to 9 identify that the expression of VCL2 in this case was 10 associated with enhanced resistance to program cell 11 death. 12 Now, intuitively it would appear as if you 13 over express a cell, a protein which would block 14 apoptosis, that you would, in essence, increase the 15 viability of those cells. If anything you would 16 increase the viability of those cells upon infection. 17 What we found, in fact, was the opposite, 18 and this graph illustrates viability which was 19 characterized by using a probe for mitochondrial 20 function. So it's not actually measuring cell lysis 21 in this case. 22 And what we saw is that if you gave these 23 -- infected these cultures at the low MOI of 24 infection, and this is about one in every 10,000 25 cells, although we were able to go significantly lower 357 1 than this, what we saw is that for about the first 2 four or five days everything appeared pretty much the 3 same in the control and the VCL2 cultures. 4 However, at about five days, there is this 5 precipitous drop in viability that was observed in 6 both of the VCL2 forms, whereas the control forms kept 7 on going out past ten days before they, too, started 8 to die due to viral mediated cytopathic effects. 9 And this effect on cell viability was 10 recapitulated on what we saw with regard to the effect 11 on viral replication, where both of the VCL2 clones 12 that we were looking at seemed to support a much more 13 robust viral kinetics, peaking significantly in 14 advance of what we saw in the case of the control 15 clones. 16 This larger spike in virus here is 17 probably due to just more cells being present in the 18 culture at that point. 19 So it's somewhat counter intuitive until 20 you sort of sit back and look at what was going on in 21 the culture, and what we observed was that -- this 22 doesn't show up all that well -- but these two clones 23 here represent -- all four of these cultures are 24 infected with the same low MOI of HIV, and what we saw 25 was in the two control clones you had a very nice 358 1 confluent lot of single cells, no evidence of viral 2 mediated cytopathic effects taking place, whereas in 3 the case of the cells which were over expressing VCL2, 4 we had these very large syncytia appearing. 5 Just to show that it wasn't due to a cell 6 specific effect, we carried out similar experiments on 7 another cell line, this being the A301 T cell line, 8 exact paralleling the experiments I just outlines, and 9 once again saw the exact same effect, that the VCL2 10 expressing clones showed a significant enhancement and 11 cytopathic effects going on in the culture. 12 What was significant though was the nature 13 of the typopathic effects. Using a slightly different 14 model, and this one here was an attempt to get away 15 from the issue that comes up with viral kinetics that 16 would happen during a spreading infection, what we did 17 was we took a cell line which constitutively expressed 18 HIV. This is the A301 cell line, which is quite 19 commonly used in HIV research, and this cell lines 20 throws off just gobs and gobs of virus and expresses 21 on its surface viral proteins, and we cultured that 22 with our sub T clones, and what we found is that all 23 the clones would roughly make about the same amount -- 24 that in all of these cultures we could identify about 25 the same amount of syncytia. 359 1 However, in the case of the control cells, 2 and the black represents viability here; in the case 3 of the control sub T cells, that most of these 4 syncytia were dead, and this was done initially by 5 just trypan blue, whereas in the case of the VCL2 6 clones, syncytia were maintaining their integrity. 7 Their membranes weren't being disrupted. 8 We also ran a routine for -- the cell line 9 itself is RT negative. So what we did is use other 10 measures for virus presence in the supernatant, and 11 found that in both of these cultures. We've seen 12 roughly equivalent amounts of virus. 13 Now, if we examined -- this is the tunnel 14 method for measuring apoptosis, which basically is an 15 in situ measurement of DNA fragmentation. What we 16 found was that red cells here represent cells in which 17 the DNA hasn't defragmented, whereas this green 18 appearance is through a number of steps is 19 representative of cells where the DNA has been 20 fragmented through apoptosis. 21 What we found was -- and these are 22 syncytia here -- in the case of the VCL2 cultures we 23 have these nice, health looking syncytial masses, 24 whereas in the case of the sub T clones we've seen 25 that these cells are undergoing apoptosis. So this is 360 1 probably more graphically represented. 2 In the next slide, if we look at actual 3 EMs, and once again this isn't really showing up all 4 that well, but this is a syncytia, and if the 5 reproduction was a bit better, what you'd see is that 6 the cellular nuclei are complete intact. They're sort 7 of wedged together, crushed together inside of the 8 cell, but each nuclei generally represents what you'd 9 see in a single cell, whereas in the case of the 10 control cell, we see this very characteristic 11 condensation and marginalization of the chromatin, 12 which is diagnostic of apoptosis, as well as 13 vacuolization of the cytoplasm and membrane webbing 14 going on. 15 And this below just represents sort of a 16 more general view of the field, showing that we're not 17 just looking at one cell that in the case of VCL T 18 have multiple, health looking syncytia with no 19 apoptosis, whereas in the case of the control 20 cultures, most of the syncytia that are present are 21 undergoing apoptosis. 22 And this is just to add in that no matter 23 whether you have VCL2 expression or you don't, if the 24 syncytia reaches a significantly large size, what 25 happens is you start to witness necrosis taking place 361 1 with rupturing of the cellular membranes, release of 2 the cytoplasmic content, swelling of the nuclei. 3 So are the effects that we're looking at 4 here the enhancement of HIV replication and 5 cytoplasmic effects strictly due to syncytia 6 formation? Apparently not. 7 If we take a look at -- in this case here 8 what we did is we went back to our original sub T 9 clones, and these are our two controls, which lack 10 VCL2 expression, and these are our two clones that 11 have enhanced VCL2 expression and effective MOI of 12 one, and this essentially means that every cell in the 13 culture receives an infectious hit, and 48 hours later 14 took a look at the level of viral replication, and 15 what we see is in the case of the VCL2 cultures we 16 have significant -- and this is a Western Blot looking 17 for the HIV protein and the cell associated HIV 18 proteins -- that we have significantly higher levels 19 of HIV protein present in the clones, in the VCL2 over 20 expressing clones. 21 So overall the point that I'm trying to 22 make here is that in the case at least for HIV and 23 possibly other lentiviruses, we have looked at other 24 retroviruses and see similar effects; that the 25 engineering of cell lines which over express in this 362 1 case VCL2, although there is a range of anti-apoptosis 2 change, both cellular and viral, which probably gives 3 similar effect and, in fact, we have seen this exact 4 same phenomena with other cellular genes which are 5 VCL2 mimetics, such as glutathione peroxidase, as well 6 as just throwing in chemical mimetics of VCL2, but we 7 see a similar phenomenon taking place. 8 So the engineering of cell lines, the 9 changing of cell culture conditions in a way to block 10 apoptosis may allow facilitated isolation of 11 adventitious agents' presence in cellular substrates. 12 The other point that sort of came out of 13 this, and I had forgotten about it until I was 14 actually putting the slides together this afternoon, 15 was that when it comes to designing a cell substrate 16 for the purpose of vaccine production and deciding 17 what components need to go into the recipe, it may be 18 wise to look at apoptosis and apoptosis inhibitors. 19 Just looking at the data set which I presented, the 20 over expression of VCL2, we see remarkably higher 21 levels of viral replication, and in addition, we see 22 significant inhibition of DNA fragmentation, which are 23 two issues which I think are of interest at least from 24 this afternoon's talk. 25 So that's it. Thanks. 363 1 (Applause.) 2 DR. PEDEN: Thank you, Paul. 3 Any questions? 4 PARTICIPANT: Very interesting data. 5 Indeed, it's fascinating. I have a couple of 6 questions. 7 One is when you say that the over 8 expression of VCL2 render the cells resistant to 9 apoptosis, what stimulus do you use for apoptosis 10 besides the virus? 11 DR. SANDSTROM: I have to think about 12 this. I know that we use serum starvation, and we 13 used a flavonoid dose called Genestine, which is -- 14 I'm trying to think of the mechanism by which it 15 produces apoptosis. I think it has something to do 16 with mitochondria. 17 But with these VCL2 clones, we have looked 18 at a number of other cell lines that people had 19 engineered, but the effect is just remarkable, these 20 VCL2 clones, and I think what was most remarkable is 21 that Genestine would inhibit cellular proliferation in 22 the control cells. You'd see the same pattern. 23 Cellular proliferation would cease. Control cells 24 would die. 25 The VCL2 expressing cells would just 364 1 continue to live or at least stay intact in culture, 2 and this would go on for days or longer, weeks with 3 cells lysing and releasing their components. 4 PARTICIPANT: Yeah. My question was how 5 generalized is this observation. I assume those are 6 T cells, and perhaps using mechanism of apoptosis that 7 are specific to T cells, such as engagement of CD3 of 8 the T cell receptor may allow you to dissect whatever. 9 This is a generalized phenomenon that will apply at 10 different cell lineages or it's specific to T cells 11 and perhaps to HIV virus. Do you -- 12 DR. SANDSTROM: Yes, BCL2 isn't specific 13 for T cells. 14 PARTICIPANT: No, no. I know the BCL2, 15 but the T cells that you use in HIV are T cells and 16 the virus that is tropic for T cells. So I wonder if 17 you have other information that will allow to 18 generalize this information to other cell lineage. 19 DR. SANDSTROM: No. The only cell line 20 that we've personally looked at was T cells. 21 DR. LEWIS: Lewis, FDA. Am I on? 22 The development of syncytia in cultures 23 infected with retroviruses, lentiviruses, whatever you 24 want to call them, is very much reminiscent of the XE 25 phenomenon that was discovered by Rowe in Wally's 365 1 laboratory back in the late 60s. Do you have any 2 evidence that there's any proportion dilution with 3 these syncytia? 4 DR. SANDSTROM: I'm sorry? 5 DR. LEWIS: Do the syncytia develop in 6 response to the dilution of the multiplicity of the 7 HIV that you put on there? Have you looked at that? 8 DR. SANDSTROM: No, we didn't. I'm trying 9 to think if maybe I serendipitously looked at it, but, 10 no, not specifically. 11 DR. LEWIS: Gosh, I mean, the question 12 goes through my mind of sort of being of the classical 13 oralgia (phonetic) modes, whether you finally have a 14 CPE assay for HIV infection. 15 DR. SANDSTROM: Oh, I see. Yeah. No, I 16 think the only thing I could say to that is that in 17 our own experience we would see better syncytia at the 18 lower multiplicity of infection, but I always think of 19 it just as a more controlled infection, a more 20 controlled spreading infection rather than at the high 21 ones that just went too fast. 22 Does that answer your question? 23 DR. LEWIS: Well, no. I guess I would 24 poke at you a little bit and say suppose you went 25 below the lower multiplicity that you're using because 366 1 if the titre is ten to the seven or ten to the eight 2 in this assay and you're looking at ten to the four, 3 then you know, you're up here. 4 DR. SANDSTROM: Right. 5 DR. LEWIS: You haven't looked at the 6 range. 7 DR. SANDSTROM: Yeah. One of the problems 8 is that just because of the culture when you get down 9 too low, you reach the mathematical endpoint. There's 10 no virus for the number of cells that you have there. 11 So I think we were able to bring it down to somewhere 12 in the range of about, so to speak, five viruses in a 13 culture class, and we would see the effect. 14 The effect actually is sort of -- the 15 differences between those clones was enhanced at the 16 lower multiplicity of infection. 17 DR. LEWIS: Well, I guess if it's not, if 18 the virions are not doing it and it's not proportioned 19 dilution, then what is it? 20 DR. SANDSTROM: Yeah. 21 DR. PEDEN: It's certainly proportioned to 22 the dilution, Andrew. It's certainly proportional. 23 Is it pressing? 24 DR. COOK: Jim Cook. 25 So you're putting BCL2 into the cells. 367 1 The nuclei remain intact. The cells die early, and 2 you said your indicator of cell death was a 3 mitochondrial marker of some sort. It sounds a lot 4 like cell necrosis. 5 Did you look at anything in terms of 6 mitochondrial function, mitochondrial respiration, and 7 BCL2, as you know, is a mitochondrial membrane active 8 agent. 9 DR. SANDSTROM: Right. 10 DR. COOK: And the question is whether 11 somehow or other this is sensitizing the cells to 12 injury induced necrosis rather than apoptosis, and 13 that's why you get early cell death. 14 DR. SANDSTROM: I don't think that's 15 what's going on. I mean, we were using MTS, which is 16 largely -- I mean, basically that measures 17 mitochondrial function, and -- 18 DR. COOK: So if the mitochondria are not 19 functioning, the nuclei remain intact. That's almost 20 the definition of necrotic cell death, right? 21 DR. SANDSTROM: Well, I think it's a lot 22 of things. I mean, number one, necrosis and apoptosis 23 are endpoints on a Gray scale. So there's a certain 24 amounts of overlap, certain features that they have in 25 common. 368 1 But largely speaking, when we looked at it 2 in terms of the lack of chromatin, fragmentation, 3 marginalization, and the fact that the cellular 4 membranes stayed intact would to us say that it was 5 apoptosis that was going on or the blocking of 6 apoptosis that was taking place. 7 DR. PEDEN: All right. Thank you. 8 The next and last of the formal program is 9 by Brian Van Tine. I hope that's the correct 10 pronunciation. "In Situ Transcriptional Analysis of 11 Integrated Viral DNA." 12 DR. VAN TINE: I'd like to introduce 13 myself. My name is Brian Van Tine, and I work in the 14 laboratory of George Shaw in the Center for AIDS 15 Research at the University of Alabama. 16 I work on a technique called FICTION by 17 tyramides, and FICTION stands for fluorescence 18 immunohistic chemistry and in situ for the study of 19 neoplasms, or in our case for the study of RNA DNA in 20 proteins in a single cell at the same time. 21 Briefly, what I'd like to go over is just 22 a method which is I realize basic for most people in 23 here, but is highly informative. You pick a target, 24 whether it's a protein, an RNA or a DNA. We design 25 either an antibody or an RNA or DNA probe to it; come 369 1 in with a detection scheme, which leads to an HRP; 2 come in with a fluoro-4-tyramide and cause a covalent 3 deposition to the slides that we're interested in, and 4 why this is important is because you can kill that HRP 5 and start over. 6 So if you're interested in studying, say, 7 your HIV-1 DNA, the RNA it produces, and some of the 8 proteins it's associated with, you can now do this in 9 a single cell. 10 This work was done in collaboration with 11 NEN because the experts that went to microdetections 12 were there, and as a result a collaboration was 13 involved with Phillip Moen and a papilloma project was 14 developed. 15 And so in general, this research that I'm 16 presenting today is on papilloma HV16, which 17 predominates in cervical cancers. It's, in general, 18 widespread and prevalent, and it's a causative agent 19 in both benign papillomas and condylomas. 20 It's multiply spliced, and the early 21 proteins that we'll be concerned with today are E6, 22 E7, E1, and E2. 23 What this shows is a general map of HPB16, 24 but what's important to note is it's only 8 KB, and if 25 this virus is to integrate not in tandem, but to be 370 1 just a single copy, it is very, very difficult with 2 any sort of sensitivity to detect a single copy by 3 normal methods, and that's why this method becomes 4 important. 5 I'd like to introduce our cell line 6 players for today. The CaSki cell line has 7 approximately 600 copies of HPV integrated into 8 various places in a very complex genome. 9 The SiHa cell line has one to five, 10 depending on the literature, copies of HPV, both on a 11 Chromosome 13, and we have designed probes either to 12 E6, E7, E1, all the way through E1, and we have a 13 probe just to E2. We also have full length probe 14 specific for each cell line. 15 Can we make it any darker in here? 16 What we are -- what I'm demonstrating in 17 the top figure here is the CaSki cell line, and this 18 is a hybridization for DNA and in tat cells, and as 19 you can see, there's many, many sites in each and 20 every cell, but down here in the SiHa cell there are 21 two cites that grow up on both 13 chromosomes in this 22 genome. 23 What we did was to ask the question, you 24 know: are all the copies in the SiHa gene on? And 25 what you do is you begin with a hybridization for RNA. 371 1 What you can see, we have two RNA centers. This is a 2 G2 doublet indicative of cells in G2. This is a bit 3 brighter in intensity and is probably an overlap G2 4 signal. 5 We then RNAase the slides, denatured them, 6 and detected the DNA, and what you end up with is you 7 can see that both copies of the HPV associated with 8 SiHa cells are both on. 9 This is a metaphase analysis of the CaSki 10 cell line. It has approximately 78 chromosomes and 11 approximately 16 integration sites ranging from what 12 is probably 100 down to, based on HIV size 13 comparisons, one copy of HPV. 14 What became interesting is we then 15 proceeded in the CaSki cell line to ask which of the 16 DNA integrants were active, and surprisingly it was 17 one. Here is the DNA in situ. This is a single cell. 18 It was pre-extracted, fixed, and hybridized for RNA, 19 then denatured, and the DNA was detected, and we find 20 that this minor site right here, an emerged picture 21 down here, is the only site that is transcribed in 22 this cell line. 23 We then went on an exploration to see if 24 we could discover exactly which of the sites was 25 active, and if it was the same. So we did this by a 372 1 variety of paints where in a yellowish color on this 2 slide we have a Si-3-tyramide to HIV and we have a 3 chromosome paint in this case to nine. 4 We at the end of the day came down with 5 three minor sites, one on Chromosome 3, one on 6 Chromosome 14, and one on a Chromosome 21 and 22. We 7 haven't, because of the paints, decided which one it 8 is. 9 We then went in, back to an interphased 10 intact cell, and just to show you that there are many 11 colors of tyramides. If you work with NEM, there's 12 now about eight. 13 Here we have in blue an integrated DNA. 14 Red, we have the RNA, and in green we have Chromosome 15 13 territories, and we were looking for association, 16 and we found that there's an association between 17 Chromosome 14 and the active integrant (phonetic). We 18 have run controls on chromosomes known to contain 19 larger tandem repeats and not see that association, 20 and on some of the few chromosomes that didn't have an 21 HPV integration and found that uniformly we find that 22 only the integrant associated with Chromosome 14 was 23 active. 24 So that led us to our next question, which 25 was: is there something special about 14 and its 373 1 position in nucleus that would suggest that it be 2 specifically transcribed? 3 And so we began by looking at an SC35 4 splicing complex oligo to RNA, and we didn't find an 5 association there. 6 We looked at a U2SN RNA and did not find 7 an association, but what you see here in these darker 8 regions are nucleoli, and consistently we have an 9 association between this being turned on and the 10 nucleoli. 11 We also looked at coiled bodies and found 12 no association, but we're not quite sure to this point 13 yet why there's only one copy turned on in the cell 14 line. 15 We also found in our process of running 16 controls that not only could we detect the sense 17 transcript in CaSki cells, but there's also an anti- 18 transcript which has been previously reported in the 19 literature. Just to show you the sensitivity and the 20 specificity of this technique, this is the E6/71 21 riboprobe in the SiHa cells, and we did a riboprobe 22 here to the E2 protein which is missing in the SiHa 23 integration to show you just how clear and specific 24 this technique is. 25 What becomes important is we have now gone 374 1 to a cervical carcinoma, and this is actually a 2 formalin fixed tissue section where in red we detected 3 DNA and in green we detect RNA, and we found that 4 we're seeing the same effect in vivo where we have 5 nuclear -- we have DNAs that are both on and off, and 6 we're starting to dissect exactly if there's an 7 association between position and being on or off. 8 So, in general, in SiHa cells, HPV16, DNA 9 and RNA, they're both on, and they colocalize. 10 In the CaSki cells, the RNA colocalizes 11 only with a single integration site, and at the time 12 the slide was made, we weren't sure exactly what it 13 was, but we now know it to be 14. 14 The RNA and the DNA does not associate 15 with splicing factor domains. The RNA and the DNA 16 does not associate with coiled bodies, but they do 17 tend to localize near nucleoli. 18 It was initially assumed that all 19 integrants were transcribed and that all transcription 20 occurred in the same direction, and the CaSki cell 21 line shows evidence for transcription from one minor 22 locus, and we detected the E2, E6, and E7 transcripts, 23 were detected from both strands -- I believe that 24 should be E1 -- and we also have anti-transcriptions 25 seen in vivo. 375 1 And going with the themes that we've 2 picked up in the meeting, we have suggested some 3 mechanisms for the HPV silencing, which may either 4 deal with methylation, the chromosome positioning 5 effects, or maybe just histos. 6 And so with that I'll conclude. 7 (Applause.) 8 PARTICIPANT: Can you -- what is the 9 sensitivity? The viral promoter may be very strong. 10 Can you use this to detect the cellular RNA? 11 DR. VAN TINE: Cellular RNA, yeah. If 12 it's colocalized in one spot, you can probably, if you 13 talk to Robert Singer, who does really fancy 14 deconvolution, you can see individual copies of RNA 15 and map the direction. 16 PARTICIPANT: Is that right? 17 DR. VAN TINE: That was reported in 18 Science about six months ago. 19 PARTICIPANT: I was not too clear about 20 the fixation. Why do you not see RNA in the 21 cytoplasm? 22 DR. VAN TINE: They're pre-extracted 23 before they're fixed. So the only RNA that remains in 24 the cell is that associated with the DNA as it's being 25 transcribed. 376 1 DR. FRIED: Mike Fried. 2 If you put in five As of cytidine, do the 3 other sites come on? 4 DR. VAN TINE: As I recall they don't. 5 DR. FRIED: So you don't see 6 inspotsylation (phonetic)? 7 DR. VAN TINE: Not on the first 8 experiments that we've done. They're still in 9 process. 10 DR. PEDEN: Okay. Thanks very much. 11 That ends the formal presentation. Now, 12 as far as I know, there are a couple of people who 13 want to present -- what's this? Just one? Yes. 14 So Dr. Arifa Khan will present first, and 15 I think Johannes Loewer will present after. Is that 16 right, Johannes? Do you still want to present? It 17 depends on the stamina of the audience. 18 DR. KHAN: Can we have the first slide, 19 please? 20 DR. PEDEN: Press it. 21 DR. KHAN: Oh, Dr. Khan. 22 DR. PEDEN: It's K-h-a-n. 23 DR. KHAN: Although my talk was to make it 24 very short, but I think it's going to get longer than 25 I -- 377 1 DR. PEDEN: It's the one with two people 2 in it, Peden and Khan. 3 DR. KHAN: I just wanted to give a brief 4 presentation of our recent work on protein analysis of 5 chemically induced k BALB 3T3 cells and to present 6 this as a potential strategy for the protection of 7 infectious retroviruses in vaccine cell substrates. 8 Okay. The k BALB 3T3 cells are derive 9 from BALB 3T3 cells. These are Christian Maloney 10 sarcoma virus transformed. BALB 3T3 cells are non- 11 producers, and the reason that I selected this 12 particular cell line is that extensive work has been 13 done on this cell line actually by Stu Aaronson's 14 group on the NIH campus in the '70s, and the 15 endogenous retroviruses that can be induced from this 16 cell line have been well characterized biologically. 17 I should also mention that in terms of 18 retrovirus induction, extensive studies have been done 19 on mouse cells, and most of the studies have been done 20 by labs on the NIH campus in the early '70s, and in 21 terms of optimizing conditions for other cells, very 22 little, as far as I know, has been done, and basically 23 the conditions that have been optimized and determined 24 for the mouse cells in most cases are just directly 25 applied to other cell lines from other species. 378 1 The purpose of initiating this work was to 2 optimize -- to understand the kinetics of retrovirus 3 induction in this well characterized cell line, and 4 then to extrapolate to see whether the results from 5 this study can be used to investigate other vaccines, 6 potential vaccine cell substrates from other species. 7 In terms of the amount of the BALB 3T3 8 cells, it's known that these cells contain or con 9 produce retroviral particles of three types. One is 10 the IAP, which are the defective Type A particles 11 which are present in numerous copies in the mouse 12 genome, and these are defective sequences that are 13 associated with internal Type A particles. 14 In addition, there are two proviruses that 15 can be induced as infectious virus from these cells. 16 They have been designated as BALB virus Type 1 -- BALB 17 Virus 1, which is an entropic, ecotropic Type C virus, 18 and BALB Virus 2, which is a xenotropic virus based 19 upon infection on rat kidney cells, which is also a 20 Type C virus. 21 So these two viruses, Virus 1 and Virus 2 22 have previously been shown to be induced from these 23 cells. 24 So what we wanted to do was to revisit 25 these induction studies, applying the highly sensitive 379 1 PERT assays to study the kinetics of retrovirus 2 induction from the k BALB cells. 3 There has been extensive work to 4 investigate a variety of agents in terms of their 5 ineducability in mouse cells and in some cases even in 6 avian cells. 7 Yeah, it's from your book, John. It's 8 your table, and you're right here. Good. Any 9 comments, direct it to Dr. Coffin, please. 10 (Laughter.) 11 DR. KHAN: As you can see, there's a 12 variety of agents that can induce retroviruses, 13 chemical agents, especially halogenated pyrimidines, 14 IBU and BRDU, also other mutagens, and a bit of the 15 protein synthesis, neurological agents, and even 16 viruses can induce retroviruses. 17 So the reason I've listed this is just so 18 that one should be aware that during the production of 19 a vaccine, you know, there may be a concern that under 20 certain circumstances the use of certain agents that 21 may be useful for producing a high vaccine virus titre 22 may inadvertently also be activating a retrovirus. 23 And, therefore, it is very important to be 24 able to detect very early on if the cell substrate has 25 a potential infectious retrovirus using sensitive 380 1 techniques. 2 We investigated using IDU since it is a 3 potent retrovirus inducer, and extensive studies have 4 been done with this, some of them. 5 In this study, the k BALB 3T3 cells were 6 exposed to various concentrate -- to 20 microgram or 7 30 micrograms of IDU at 24 hours or at 48 hours, and 8 the control cells -- this is using the taq man PERT 9 assay that Dr. Peden has described earlier. 10 The control cells were negative in the 11 assay. 12 As you can see, the 30 micrograms produced 13 high RT activity early on. This is day two time 14 point. This is day four, and this is day six here. 15 So early on the high dose, even the 24 hour post 16 induction resulted in high RT production, whereas the 17 lower dose at either 48 hours or even 24 hours, did 18 produce low levels, but they were not detectable at 19 the later time points. 20 On further culturing, we saw another P 21 come up later on on day 17 with the higher dose and 22 was not detectable by the PERT assay with the lower 23 dose. 24 I should mention that these two peaks, an 25 early peak and a later peak, has previously been 381 1 identified, induced from these cells using co- 2 cultivation and marker rescue experiments. So we 3 confirmed that we could, you know, repeat these 4 studies, and using the PERT assay. 5 However, the additional data is that the 6 higher dose was necessary to be able to pick up the 7 virus early on. 8 I should also mention that in the earlier 9 studies done by Joe Anderson's group they were able to 10 also detect RT using the traditional assay. However, 11 they had to use samples that were 100-fold 12 concentrated, whereas in our case this represents one 13 microliter from a T-75 flask. So this demonstrates 14 the sensitivity of the assay. 15 And in this case, the supernatant was 16 harvested daily up to seven days, and then after every 17 three days. 18 To further investigate the particles that 19 were induced, we also analyzed some of the time points 20 by EM. This is the untreated sample. At day two 21 there is your typical Type A particle, which is at low 22 levels in our induced cells. 23 At 20 micrograms, even at 48 hours at day 24 five the only particles that were detected by EM were 25 the A particles. We did not detect any C particles, 382 1 although we know that C particles are present at this 2 concentration because we did detect it by the PERT 3 assay. The As would not be released into the medium. 4 So we would only be detecting exogenous C. So this is 5 just the sensitivity of the assay. 6 With the 30 micrograms at 24 hours at day 7 two, we saw the As. In addition, we could see a few 8 Cs being produced, as you see this one budding out, 9 but there were very few Cs at 24 hours, whereas it was 10 increased at 48 hours. 11 However, I want to mention that the higher 12 concentration and the long time of treatment, as you 13 can see, there was cell toxicity. So, however, to 14 assess that the RT that we were detecting was due to 15 retroviral particles and not due to cellular lysis, we 16 also have tested lipid in the PERT assay, and the RT 17 levels that were detected at this time point were 18 threefold above that of cellular background, 19 concentrated cell lysis. So this RT -- this doesn't 20 go back? -- anyway, so the RT that we were picking up 21 were associated with the Type C particles and not cell 22 lysis. 23 Okay. So in summary, spontaneous release 24 of virus was not detected in the untreated k BALB 3T3 25 cells. However, it has been shown that under, I 383 1 guess, certain culture conditions this can occur. 2 However, we were careful not to overgrow the cells or 3 stress the cells to avoid that. 4 Induction of endogenous Type C retrovirus 5 was dose dependent, and we found that evaluation of 6 several early and later time points was necessary to 7 completely analyze all of the inducible retroviruses. 8 And in conclusion, I would just like to 9 propose this strategy for consideration for detection 10 of endogenous infectious retroviruses in cell 11 substrates used for production of biologics. 12 The first step would be to optimize the 13 induction conditions, which would be to identify and 14 use a potent general retrovirus inducer or it may be 15 more than one, I think, depending on your cell 16 substrate and what you might expect from it. 17 To determine the optimum dose and exposure 18 time, you want to achieve high viral induction with 19 low cell toxicity. 20 Then you want to combine this with 21 sensitive detection of retroviruses from induced 22 cells. The PERT assay is very useful in this aspects, 23 and you would want to test several early and late time 24 points. 25 And of course, if you have a positive PERT 384 1 result, it then would be necessary to further identify 2 the source of the activity by EM or by co-cultivation 3 to see whether you have an infectious retrovirus. 4 I would just end with that. 5 (Applause.) 6 DR. PEDEN: Thank you, Arifa. 7 Any short questions? 8 Dr. Coffin. 9 DR. COFFIN: I would suggest adding to 10 this some attempt to characterize the genome of the 11 virus that's induced. Do you know what you're 12 actually seeing from the BALB C cells? I mean, there 13 are 60 or so proviruses that might give rise to this 14 activity. 15 DR. KHAN: Well, the BALB C cells, I 16 think, is very well characterized because, you know, 17 the two proviral genomes that give rise to these 18 viruses, I think, are known in terms of genetic -- 19 DR. COFFIN: Those are the infectious 20 ones. 21 DR. KHAN: Yes. 22 DR. COFFIN: But the particles that you're 23 seeing may actually be coming from any of a number of 24 other ones as well. It would be of some interest 25 actually to look at the envelope genes and see what's 385 1 going on. 2 DR. KHAN: As you know, we've been both 3 down this path in terms of, you know -- I think for a 4 long time -- in terms of trying to characterize all of 5 the endogenous retroviral sequences, and I think the 6 majority of them are defective, and the question 7 actually, a broad question I get with your -- which 8 you could present is that, you know, in terms of RT 9 activity or in terms of particles that are produced, 10 what is the real public health concern, and I think, 11 you know, would it be a defective particle? Would 12 it -- 13 DR. COFFIN: We don't know that. I mean 14 to be fair, we don't know that for most of the 15 endogenous C type viruses that you can see in mouse. 16 We know that you don't see whole virus that 17 corresponds to most of those, but we don't know that 18 they're defective -- 19 DR. KHAN: No, I agree. 20 DR. COFFIN: -- in any real way. 21 DR. KHAN: So in this case, the next step 22 that we need to do was to do infectivity studies to 23 see whether those peaks correspond to any infectious 24 virus, and the reason I did not -- you know, I went 25 ahead and presented this data is that for this 386 1 particular cell line, these two time points, day three 2 and day 20, are associated with infectious virus. The 3 early time point is with the phenome, BALB Virus 2, 4 and then they have infectious BALB Virus 1, endotropic 5 virus. 6 But in general, for a novel cell 7 substrate, if you had RT activity, the next would be 8 to see whether there's infectious virus. 9 DR. COFFIN: But that echovirus has to be 10 the product of recombination or mutation because the 11 proviruses themselves are defective. 12 DR. KHAN: Right. That's why you see it 13 over a long period of time. So I think -- 14 DR. COFFIN: There must be other -- there 15 may well be other defective things inducing that might 16 contribute biologically to the final product. 17 DR. KHAN: Right. You would propose that 18 once you have RT activity, then you need to follow it 19 up with infectivity experiments? Yes? 20 DR. COFFIN: Well, I would do some PCR and 21 see what sort of sequences you get in there. It's 22 fairly simple with oligo probes and so on to just the 23 PCR amplify envelope genes and see what's being made. 24 DR. KHAN: Right. You can do it, and 25 then, you know, wraps it and stuff, yeah. 387 1 DR. PEDEN: Okay. The next speaker is 2 Damian Purcell. 3 Is Purcell's carousel in the -- it is. 4 Thank you. 5 DR. PURCELL: Okay. This is going to be 6 fairly fast since half the slides have been presented 7 and duplicates in some ways a story we're already 8 heard from Ruth Ruprecht with slide differencing in 9 some of the outcomes which I'll show you. 10 So what we've done here, as Ruth did, we 11 evaluated the logistical problems of delivering a live 12 attenuated HIV vaccine strain, and we saw that this 13 would be difficult with the virus strain in our minds 14 being at the time a strain of HIV that emerged in 15 Sidney, Australia, know as the Sidney Blood Bank 16 Cohort, which is essentially a nef deleted virus, 17 which for a long time appeared to be attenuated and 18 not causing disease in patients, but just recently 19 some of the cohort of that group of patient have begun 20 to lose immune function. So it's probably following 21 the same story as the simian cases that progressed to 22 full pathology. 23 Anyway, we saw that the possibility of 24 having to culture the isolate of this virus strain 25 that we had in the lab by necessity in PBMC, we 388 1 thought that this would be a problem rather than an 2 opportunity as other people have found, and the 3 necessity to produce large lots even if we could find 4 a human cell line would cause problems in terms of the 5 approval of the vaccine, a live attenuated vaccine. 6 So we figured that the best way of dealing 7 with these issues would be to deliver a truly 8 attenuated virus, if this was one, as provirus DNA. 9 So we thought to model this in the SIV 10 system, and we took the sequence of the Sidney Blood 11 Bank Cohort patients and identified the minimal 12 deletion, which is a region in U3, and the patients 13 had other deletions in and around nef and NU3, but we 14 chose to take the minimum region of the deletion. 15 We made an alignment against the SIV MAC 16 239, which through this region is not a very strong 17 alignment, and we made a deletion that sort of 18 represented roughly the maximum amount of sequence 19 deleted out of the Sidney Blood Bank viruses, and just 20 as a comparison against the equivalent deletion, which 21 I guess we're more familiar with, the second of the 22 two deletions in SIV made by Ron Desrosiers. It sort 23 of accommodates half of that second deletion. 24 Now, we made two forms of this proviral 25 strain, one bearing the deletion in just the three 389 1 prime end, which we call SIV SBBC Delta 3. So the 2 deletion is 105 base pairs, and after some coercing 3 with the provirus, we managed to make the double 4 deleted form with the deletion also in the five prime 5 LTR with the provirus. 6 We're unsure when administering these 7 proviral constructs into the animals whether it would 8 be necessary to retain the five prime U3 promoter 9 function, but even with the deletion in the three 10 prime end, as we've heard in the meeting earlier, RNA 11 transcription commences at the R region, so effective 12 excising the first of the U3 region in this singly 13 deleted form, so that after one round of replication 14 of the virus, there's a duplication of the deletion, 15 and all progeny virus then will have the double 16 deletion after one round of replication. 17 So our design was to administer plasmid 18 DNA into eight juvenile macaque Nemestrinas. We've 19 already done one titration point. Compared to Ruth 20 Ruprecht here we're down. She used 500 micrograms. 21 We used 300 micrograms IM. 22 We've also taken the gene gun and 23 administering 15 micrograms with wild type SIV and 24 with the two different attenuated strains, the Delta 25 3; so that's with the just three prime deletion in U3. 390 1 We did an IM 300 microgram injection of this one, and 2 with the double deleted version, we loaded this up on 3 the gold beads and administered 15 micrograms of DNA 4 by gene gun. 5 So the wild type delivery in each case was 6 successful. What we're looking at is plasma RNA. 7 Three of the animals -- well, this one actually died 8 at ten weeks. This one was euthanized here with very 9 low CD-4 counts, as you'll see, and the other one also 10 was euthanized with very low CD-4s. 11 But what I want to draw your attention to 12 is with the singly deleted form of the construct that 13 we administered, one of these two animals or both of 14 them scored high virus loads, but then by between five 15 and ten weeks one of these animals, Monkey 16, had a 16 high level of virus coming back in the plasma. 17 The time points we have on the double 18 deleted virus were short. They haven't had this virus 19 in the monkeys as long, but it dwindles down very 20 rapidly. It's certainly detectable, to begin with. 21 Just, again, looking at the CD-4 counts in 22 the wild type monkeys, they lose their CD-4 very 23 rapidly. So it's a very efficient way of 24 administering virus, just as Ruth Ruprecht showed. 25 And this Monkey 16, as well as retaining 391 1 high viral load, has begun to lose CD-4 cells. 2 So just following up, the changes to the 3 sequence in the two monkeys, this is the one that does 4 not regain high viral load nor lose CD-4 counts. This 5 is the kind of changes that we commonly see when you 6 administer genuine virus with deletion. You see extra 7 deletions within the nef to increase fitness, little 8 duplications of sequence. 9 But at week five we did find a wholesale 10 duplication of 63 base pairs from the wild type part 11 of U3 into this, but it was lost through time, but the 12 other monkey, Monkey 16, by the third week we found a 13 complete replacement of the U3 here in a small 14 fraction of the clones, and as time went by, all of 15 the clones had reverted to wild type. 16 So the question is: how did this event 17 occur? It's a different form of reverse to what we 18 typically see, and we strongly suspect the presence of 19 the complete U3 within the DNA, the plasmid that was 20 administered to the animals as a bolus of 300 21 micrograms. 22 So just in conclusion, wild type SIV DNA 23 is certainly pathogenic. As low as 15 micrograms of 24 DNA administered by gene gun is also very efficient 25 even with the attenuated double deleted form of virus. 392 1 Probably the most interesting thing is 2 that the evidence of the reversion in one of the 3 animals implicates the DNA as a potential source of 4 recombination with the virus. 5 And I'll leave it at that. 6 (Applause.) 7 DR. PEDEN: Thank you, Damian. 8 Any brief questions, comments? 9 DR. PATIENCE: Damian, thanks a lot. 10 DR. PEDEN: Clive Patience. 11 DR. PATIENCE: Yes. Sorry. Clive 12 Patience, BioTransplant. 13 I don't know if I saw it through a haze. 14 It's getting late for me, but did you see quite a 15 severe dropoff in the intradermal administration of 16 you variants after one week in the CD-4 count? 17 I think it was just in the intradermal if 18 I -- 19 DR. PURCELL: Yes. 20 DR. PATIENCE: And if so, if I did read it 21 right, do you have any explanation why the CD-4 count 22 went down about fourfold, I think it was? 23 DR. PURCELL: It was probably the way -- 24 this was not numbered, the CD-4s. We were measuring 25 the percent of CD-4s. So it's a percent of the pre- 393 1 inoculation percentage, if you like. So rather than 2 counting total CD-4 numbers, we're looking -- we run 3 the samples through the fax (phonetic), and we look at 4 the ratio of CD-4s at the time point that we examine 5 and compare that back to the pre-inoculation time 6 point. 7 So it's really looking at the ratio more 8 than the actual numbers. 9 DR. PATIENCE: Okay. Do the real numbers 10 follow the same sort of pattern as well or is it 11 disturbance of the ratio of your CD-4 -- 12 DR. PURCELL: At the time we started this, 13 we didn't have the technology to count the numbers in 14 the primates, but we'll have to look at that in the 15 next set of animals, yeah. 16 DR. PATIENCE: Thanks. 17 DR. BERKHOUT: I guess the -- 18 DR. PEDEN: Berkhout. 19 DR. BERKHOUT: Ben Berkhout, yeah. 20 The unusual recombination event that you 21 see may still have occurred through RNA, although, you 22 know, I guess there are two ways. One is that perhaps 23 the transcription is fired from your plasmids starting 24 at the three prime LTR, and of course, you can also 25 have transcripts that are started at the five prime 394 1 LTR, but you get read through transcription. 2 DR. PURCELL: The possibility that we're 3 initiating transcription from here and reading through 4 an RNA transcript to here, which would then be 5 terminated, polyadenylated without requiring an 6 action. 7 DR. BERKHOUT: But if polyadenylation is 8 not efficient then -- 9 DR. PURCELL: True, and the other probably 10 more likely possibility is a read through the 11 polyadenylation site here down to this, the second one 12 that appears, making it, you know, quite a long RNA, 13 which might be code packaged. 14 So we haven't completely excluded it. We 15 need a second experiment, I think, but certainly the 16 possibility is there, and I think that the selective 17 pressure we're picking up, that piece of DNA would be 18 greater than any other cellular DNA. 19 DR. PEDEN: All right. Thank you, David. 20 I think Johannes wanted -- do you still 21 want to speak? 22 DR. LOEWER: I would, but I guess -- 23 DR. PEDEN: Did you turn off the machine? 24 So if the audience is able to take one 25 more talk. 395 1 (Pause in proceedings.) 2 DR. LOEWER: What I wanted to do is to 3 present very quickly the most recent experiments we 4 have performed to understand and moditate (phonetic) 5 molecular biology of the human endogenous retrovirus 6 families, K family. At least to my knowledge, it's 7 the only retrovirus, human endogenous retrovirus 8 family with codes for all structural viral proteins. 9 These proteins are expressed with the zone 10 (phonetic) virus family, which is able to form wider 11 particles which you can see here on the micrograph. 12 This virus is expressed mainly in terata 13 carcinoma cell lines, as I've shown this afternoon. 14 This is a Northern Blot describing as full length and 15 singly spliced N message in addition to these four 16 messages. We see also two. There is more messages, 17 which is very unusual for endogenous retrovirus. This 18 resembles the more complex retroviruses. 19 Analysis shows that one of the smaller 20 length RNAs is doubly spliced, encodes for a small 21 protein, which is located in this area which was 22 called because of this localization central open 23 reading frame. 24 And this protein resembles to some extent 25 the left protein from HIV or the next protein from 396 1 HIV-1 in the sense that it has nuclear localization 2 signal, as well as a domain which resembles the 3 (unintelligible). 4 And the question we ask, of course, is 5 that if this is a functional protein which enhances or 6 allows the export of full length simply spliced RNAs, 7 to study this we used a system developed by Barbara 8 Felmer and Charles Papalakus. It's a construct which 9 expresses HIV only if there is a responsive element 10 like the RRE and if ref or a corresponding protein is 11 provided in (unintelligible), we have replaced RRE by 12 different parts of the three prime part of the K 13 genome and have also provided the C orf protein in 14 trans. 15 The HIV expression was first -- or what I 16 have to say, we have checked here several pieces, and 17 what turned out is that there is a responsive element. 18 It's a three prime LTR which we called RcRE, analogy 19 to our XIA from HIV-1 because it has very similar 20 localization. 21 And this is shown by more fluorescence 22 data here, for example, only in the presence of what 23 we call RcRE HIV that is produced with other parts of 24 the nonresponding fragments. You can only see as a C 25 orf protein localized in the nuclei. 397 1 This can be quantified by P24 antigen 2 measurement. Here is RRE, the HIV system, and if raf 3 is present at RRE, you get a high expression of P24, 4 and the same is true if you have the responsive 5 element for C orf and C orf is present, but raf can 6 also go up to some extent on RcRE, but C orf cannot 7 work on RRE and other parts of the genome are not 8 active. 9 Next, we looked at the different domains. 10 We mutate nuclear localization signal and also mutated 11 the positive nuclear export signal, and we check the 12 shortened -- if the C terminal shortened, C orf would 13 also be active. In this case, all this construct, 14 they were used to keep lost protein (phonetic). 15 And these are the effects of the different 16 mutations. This is wild type. This is C orf 17 localized in the nuclei, and P24 in the cytoplasm with 18 mutation of the nuclear localization signal. The C 19 orf cannot enter the nucleus. It's found only in 20 cytoplasm, and there's a mutation of the nuclear 21 export signal as you can expect. The mutant is 22 localized in the nuclei as well as the glial line 23 (phonetic), but no P24 is formed. The same is true 24 with the one which is a protein which is truncated. 25 Shortly, the next question of course is 398 1 whether the same export pathway is used for HIV. HIV 2 uses the CREM pathway, and this can be blocked by a 3 type called leptomycin B, and this is also active in 4 the RcRE orf system, for example, in the presence of 5 six nanomolar leptomycin, no P24. In this case it's 6 the major core protein of rev K itself, is formed, no 7 N protein in this construct, but C orf is still 8 produced and present in the nuclei and in the 9 cytoplasm. 10 So in summary the endogenous retrovirus 11 family, we have K codes for protein which is 12 equivalent to ref and rex. That means that the 13 nuclear export pathway was not only detected by 14 reasoned lentiviruses or reasoned HIV-1 BLV viruses, 15 but already 30 million years ago when this family most 16 probably was active. 17 What also makes this family interesting, 18 observations by Nicholas over lunch and his group in 19 Germany, he has indications that this virus may be 20 orf'ed in oncogenesis and only in situation where C 21 orf is also present. 22 So the next study in the end, to find out 23 whether this look like nuclear export protein may play 24 a role without the conditions in carcinogenesis. 25 Thank you. 399 1 (Applause.) 2 DR. PEDEN: Thank you, Johannes. 3 I think because time is very short and we 4 have to get out of the room I'd like to express my 5 gratitude to the audience for staying so late, and 6 especially to the speakers. 7 So I think we should terminate this 8 evening and be back tomorrow. 9 Thank you. 10 (Whereupon, at 9:55 p.m., the meeting was 11 adjourned, to reconvene on Friday, September 10, 12 1999.) 13 14 15 16 17 18 19 20 21 22 23 24 25