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 + + + + Friday, 10 September 1999 The workshop took place in the Plaza Ballroom, DoubleTree Hotel, 1750 Rockville Pike, Rockville, MD, 20852, at 8:00 a.m., Regina Rabinovich, M.D. and Martin Myers, M.D., Session Chairs, presiding. 2 PRESENT: Regina Rabinovich, M.D. Session Chair Martin Myers, M.D. Session Chair David Onions, Ph.D. Panel Chair John Coffin, Ph.D. Panel Chair Philip Minor, Ph.D. Speaker James Robertson, Ph.D. Speaker Joerg Schuepbach, M.D. Speaker Jens Mayer, Ph.D. Speaker Thomas Broker, Ph.D. Speaker Neil Cashman, M.D. Speaker John Sedivy, Ph.D. Speaker Frits Fallaux, Ph.D. Speaker Michael Fried, Ph.D. Panelist Stephen Hughes, Ph.D. Panelist Johannes Loewer, M.D. Panelist Also Present: Gary Nabel 3 I N D E X SPEAKER PAGE Philip Minor, Ph.D. 4 James Robertson, Ph.D. 32 Jorg Schuepbach, M.D. 51 Jens Mayer, Ph.D. 66 Thomas Broker, Ph.D. 78 Neil Cashman, M.D. 102 Panel Discussion 129 Dr. Onions, Chair John Sedivy, Ph.D. 176 Frits Fallaux, Ph.D. 199 Panel Discussion 212 Dr. Coffin, Chair 4 1 P-R-O-C-E-E-D-I-N-G-S 2 8:04 A.M. 3 CHAIRPERSON RABINOVICH: Good morning. I 4 would like to welcome you back to Session 6, 5 Adventitious Viral Agents in Cell Substrates, and 6 congratulate all those that were here until 10:00 last 7 night for the latest part of the show, including Mr. 8 Harris, who put in a grueling 18-hour day in 9 yesterday, our visual aides person. 10 I would like to introduce the first 11 speaker, Dr. Phil Minor, from the National Institute 12 of Biological Standards and Control, who will be 13 giving us an introduction to adventitious agent 14 issues, both reviewing the past and current experience 15 with adventitious agent contamination of biologicals 16 in vaccines. 17 DR. MINOR: Thanks. Thanks very much. 18 Can I have the slide on, please, or do I just press it 19 here? 20 What I am going to do is to review firstly 21 all of biologicals, if you like, from an adventitious 22 agent point of view. So it won't just be vaccines. 23 In particular, I will be talking about the range of 24 source materials that people have used in preparing 25 biologicals. There will be a clear message that comes 5 1 out of that, which is that the more you use well- 2 characterized cells, the better. 3 I will also be talking about the SV40 4 story in some detail, which has been gone through a 5 number of times, but I will be going through it from 6 a particularly regulatory point of view because again, 7 there is a message there which says that if you get it 8 wrong, you will still be working on it 40 years later. 9 Finally, I will get onto the continuous 10 cell line bit right to the very end. 11 So there are a variety of source materials 12 that you can use if you are preparing biologicals. 13 They are sort of listed here, if you like. There are 14 biological materials which are made from whole 15 animals. That would include things like blood and 16 blood products. I will describe that in a moment. So 17 you can just go to a whole animal and take something 18 out and make your biological from that. 19 You can use your whole animal as a 20 substrate for growth. I will discuss that in the 21 context of things like influenza vaccines and the 22 like. 23 You can grow material on primary cells. 24 This was the main starting point for things like polio 25 vaccines in the early days, where the SV40 issue 6 1 arose. Finally, you can grow materials on well- 2 characterized cell preparation. 3 The further down the list you go on this 4 thing, probably the happier you are from the 5 adventitious agent point of view. 6 This shows some examples of contaminants 7 which have arisen when whole animals have been used as 8 source materials or the origin of the source material. 9 Most of these will be human rather than anything else, 10 but really an awful lot of the serious adventitious 11 agent problems that have arisen have arisen because of 12 materials sourced from whole animals or using pooled 13 preparation. 14 The first one on this list here is CJD, 15 Creutzfeldt Jakob Disease, which was transmitted by 16 growth hormone. The growth hormone was produced from 17 human cadaveric material. A very unpleasant disease. 18 It's almost impossible to detect the agent other than 19 by standing back and waiting for the incubation period 20 to go. 21 In France, there are still a large number 22 of cases coming through as a result of this. It may 23 well be that around 10 percent are recipients of 24 human-derived growth hormone, will actually wind up 25 going down with CJD in France. 7 1 Dura mater is another one. That should be 2 one T, not two Ts. Again, this has been shown to 3 transmit really quite readily when pooled materials 4 are actually used. Almost impossible to detect. 5 Clear, very, very serious kind of consequence of it. 6 Scrapie was first shown to be a 7 transmissible agent by the use of a TBE vaccine, which 8 was grown in the brains of sheep. TBE being tick- 9 borne encephalitis, which then transmitted scrapie to 10 a large number of the sheep that were actually 11 inoculated with it. So again, this is a whole animal 12 source material, if you like, that had quite serious 13 consequences, especially if you were a sheep. 14 Over the last 15 or 20 years or so, one of 15 the best examples of serious or disease-causing 16 transmissions of infectious agents has been through 17 human blood and blood-derived materials, clotting 18 factors in particular. In all of these things, the 19 entire alphabet soup of hepatitis viruses has been 20 transmitted by blood product. 21 In the early days at least, some of these 22 were really regarded as really a hazard, if you like, 23 of being a hemophiliac. So, for example, hepatitis C, 24 in the days when there was a non-A, non-B hepatitis, 25 it was really regarded as an inevitable consequence of 8 1 using factor 8 to treat hemophilia. I am not sure 2 that that is an acceptable way of actually doing 3 things any more. I am sure that hemophiliacs would 4 agree with that. B-19, paravirus B-19 is still 5 transmitted by clotting factors. 6 Finally, this one down at the bottom here 7 is a classic example of a transmission by a vaccine, 8 if you like, where hepatitis B was transmitted by 9 yellow fever vaccine back in the 1940s. The hepatitis 10 B actually came from the stabilizers of the albumin 11 that was actually put in there to keep it stable. 12 There is a story that Fred McCallum, who 13 is head of the Public Health Service in the United 14 Kingdom tells to the effect that he basically won the 15 war because he prevented Winston Churchill having a 16 yellow fever vaccine when he was going off to talk 17 with Stalin around 1944. 18 So most of the serious consequences really 19 come from whole animal source materials, if you like. 20 You can use whole animals as substrates. I'm using 21 the term "whole animals" in a fairly broad sense. 22 Eggs in the definition of the Animal Regulated Use Act 23 in the United Kingdom count as an animal because they 24 are embryonated. 25 For many years, rabies vaccines were 9 1 produced in mouse brain or sheep brain. They have 2 quite serious consequences, but not necessarily 3 associated with adventitial agents. You can get 4 encephalitis as a result of immune responses to the 5 non-invasic protein. 6 The Japanese encephalitis vaccine, which 7 is used for travelers in the United Kingdom, is still 8 made in mouse brain. So it's not an unusual source of 9 material, if you like. Smallpox for a long time was 10 made on the scarified flanks of calves. Like I said, 11 isn't any more. However, while these things seem 12 really quite primitive, in terms of how you make 13 vaccines nowadays, you still have a number of vaccines 14 that are made in eggs. Yellow fever is the classic 15 example, and influenza. 16 Yellow fever is not required to be grown 17 in avian leukosis-free eggs. The reason for that is 18 that there are a number of sites at which it was 19 manufactured throughout the world, where yellow fever 20 is a really very serious problem, such as Nigeria, for 21 example, South America, whatever, where SPF eggs, 22 avian leukosis-free eggs even, were really not freely 23 available. So yellow fever can in principle at least 24 be made in avian leukosis containing eggs, and in fact 25 is. I think there's no evidence that this has an 10 1 adverse consequence. But on the other hand, you 2 wouldn't necessarily want to have a virus in there 3 that you didn't know about. 4 Influenza is an actuated vaccine. Again, 5 it's not made on SPF eggs, that is, specified 6 pathogen-free eggs. They are avian leukosis free, but 7 they are not free of all the other variety of 8 pathogens that you would choose to screen for measles 9 vaccine production system, for example. 10 So even today then you have to bear in 11 mind that a large amount of vaccine that's made is 12 made on really quite crude materials, from an 13 adventitious agent point of view. It's not a trivial 14 usage. In fact, when you go through and consider what 15 vaccines are actually made on these days, they are 16 quite primitive, if you like, in some respects. 17 Primary cultures as been described 18 previously around here, are really cultures that are 19 made directly from the animal. So they are not one 20 pass. They are directly from the animal, if you like. 21 Here are a number of examples where agents are 22 actually being found or at least located in these kind 23 of cultures. 24 SV40 is one that I'm going to talk about 25 in some detail in a minute. This was in polio 11 1 vaccines in the 1950s and very early 1960s, probably, 2 a source from rhesus monkey kidney. Polio vaccines 3 are still made on monkey kidney, though they are not 4 usually on rhesus monkey kidney. It would be 5 cynomologous or something like that, for reasons which 6 I'll describe in a moment. 7 Nonetheless, a great deal of vaccine is 8 still made in primary monkey kidney cells. There are 9 reasons for that. There's a deep conservatism I think 10 about changing the vaccine production process if you 11 have a vaccine that works, largely because you are 12 dealing with a prophylactic material rather than a 13 therapeutic material. So you don't want to mess about 14 with anything if it's reasonably safe and effective. 15 I'll mention very briefly the defective 16 retrovirus story in chick embryos. I think Jim 17 Robertson will probably mention this in more detail, 18 but I will mention that just as I go by. Finally 19 recently, the FDA released a talk paper on a 20 preparation of urokinase, which is used in treating 21 the heart. This material was grown from primary 22 cultures made from aborted fetuses. I think it was 23 aborted fetuses or miscarriages, or whatever. There 24 were quite a variety of infectious agents were 25 actually found in this. I believe this one has now 12 1 been suspended. 2 The point is that there are still a large 3 number of materials which are made on really quite 4 basic culture systems, if you like, where adventitious 5 agents are a serious consideration, if you like. So 6 it's not all continuous cell lines versus the rest. 7 I mean there are -- most of the vaccines that are made 8 in the world probably come from other primary cultures 9 or eggs or things of that nature. 10 I will now talk about SV40. I'm sure in 11 this audience there are people who know far more about 12 SV40 than I do. But nonetheless, I'll talk about this 13 from what you might call the regulatory adventitious 14 agent point of view, if you like. 15 So it's a very common polyoma virus of old 16 world monkeys, and particularly rhesus macaques. The 17 difficulty with this was that when the rhesus macaque 18 monkeys are sacrificed and a primary monkey kidney 19 culture is made from him or her, as the case may be, 20 a silent infection is set up. So there is on evidence 21 of infection just by looking at the cultures. In 22 fact, these cultures can throw out as much SV40 as 23 they do polio, when you start infecting it with polio. 24 So you wind up with a culture that's just stiff with 25 adventitious agent which you really don't want. 13 1 It's able to transform non simian cells in 2 vitro, and it can be tumorigenic if you have the right 3 kind of animal that you put it into. Between 55 and 4 62, probably at least a third of all the vaccines that 5 were made on these kind of cultures, because they were 6 pooled and the like, were almost certainly 7 contaminated with SV40. It wasn't a trivial 8 contamination. It was really quite a serious 9 contamination. 10 Because it was mainly an activated polio 11 vaccine, there wouldn't have been that much live SV40 12 in it perhaps. But SV40 is more resistant to formalin 13 than polio is. So almost everybody who received the 14 shot of inactivated polio in the 1950s, which would 15 include me, would have received live SV40 in some form 16 or another. 17 So the concern is really summarized here, 18 which is basically that everybody, I mean this is my 19 own take on it, that everybody -- I mean you can argue 20 that it might not have been sort of everybody, but I 21 think it probably was. But almost everybody who 22 received the full course of polio vaccine between 1955 23 and 1965, also got live SV40 stuck into them. That's 24 millions of people basically. 25 There were epidemiological studies that 14 1 were done at the time which really didn't cause much 2 concern, but they can all be criticized. Some of the 3 studies were really quite short-term, about two or 4 three years or so, looking to see if there were cancer 5 effects basically, as a result of SV40. It may be 6 that two or three years is not enough to actually find 7 such an effect, if it actually exists. 8 The longest which was assumed was over a 9 period of about 19 years. Most of the individuals 10 involved in that study would have been oral polio 11 vaccine recipients rather than inactivated polio 12 vaccine recipients. So they have had it by mouth 13 rather than by injection. Again, you could argue that 14 that might not be the right cohort to actually be 15 looking at. 16 So while the studies were reassuring, the 17 most reassuring thing was that there was no sudden 18 surge of cancers that you can actually trace back to 19 polio vaccine usage in the United States or in Europe 20 where these things were used in a big way. So it 21 really did seem that in the long term, over about 19 22 to 20 years, there was no real cause for alarm. 23 However, in 1992, Michaili Carboni and 24 colleagues and others, a number of others, including 25 Janet Butelle down in Texas and the like, identified 15 1 SV40 sequences which were present in a variety of 2 relatively rare tumors. So mysathelia, which is the 3 asbestos tumor, osteosarcomas, pendymonas, actually 4 the young chorioid plexus tumors of children, these 5 sequences do appear to be genuine SV40 sequences. 6 Where they come from is really not quite 7 clear. Part of the argument was that you could get 8 similar types of tumors in experimental animals, like 9 hamsters. I think that is probably the only example 10 where a hamster is cited as a good model for a human 11 being perhaps. But who knows? In fact, this might 12 actually be an argument that this has got nothing to 13 do with it. 14 So the question then arises as to where 15 did the SV40 sequences come from. Of course the 16 classic response really would have been it must have 17 come from the polio vaccine because why not? 18 Now SV40 was discovered around 1961 or 19 1962 or thereabouts, 1960 perhaps. Directly it was 20 discouraged. There were precautions put in place to 21 exclude it from polio vaccines, because it was known 22 to be a tumor kind of virus, if you like. These were 23 the kind of things that were put in place. They are 24 listed in WHO requirements from about 1962 onwards. 25 They reached their final fully flowered form, if you 16 1 like, by about 1965. A number of countries certainly 2 had put this in place before that. 3 The first thing you can do is to use 4 seronegative animals as the source of cells. So you 5 can use animals that have no evidence of SV40 6 infection as your source. That really is something 7 which is now very firmly in place, which manufacturers 8 now do. 9 The second thing is, you remember I said 10 that it was the rhesus macaques with the problems. 11 The problem was that the cell cultures didn't show any 12 sign of having defect, when they were actually 13 infected with SV40. What you can do then is you can 14 use species, such as cynomolgus or pattus monkeys, 15 where the primary monkey kidney culture cell, when 16 infected with SV40, will actually wrinkle up and die 17 on you. So at least you know you have got something 18 nasty and you can throw it away. 19 Finally, you can test your control of 20 production cultures for SV40 by the same kind of 21 procedure. That is why using sesetral cells to see if 22 anything comes through. 23 Around the period that this was taking 24 place, wild caught monkeys were being used extensively 25 in vaccine production. Up to a half of the cultures 17 1 would have been thrown away because of adventitious 2 agent contamination, mainly foamy virus, but certainly 3 other things as well. I think that just illustrates 4 the kind of lack of control, if you like, over the 5 source materials that was going on, and the extent to 6 which adventitious agents are really a serious problem 7 in finding monkey kidney cultures or primary cultures 8 in general. 9 An alternative way of doing this is to 10 actually use a validated cell bank. Certainly many 11 manufacturers use MRC5, and Mary of course used vira 12 cells, as we heard last night. 13 Nonetheless, a significant, if indeed not 14 a large proportion of the world's supply of polio 15 vaccine is still made on primary monkey kidney cells, 16 which should really fit this kind of criteria for 17 excluding SV40. 18 One of the questions that then arises is 19 were these precautions good enough? What we did at 20 NIBSC, because we happened to have about 150-odd 21 batches of vaccine archived from the years, was to go 22 back and look at them by PCR. PCR of course is the 23 cat's pajamas. It's really the best technique that 24 anybody ever invented in terms of sensitivity. It's 25 probably about as good as infectivity, at least in our 18 1 hands anyway. 2 But nonetheless, we went back and we 3 looked by PCR at 133 preparations of polio vaccine 4 which had been used in the United Kingdom between 1966 5 and about 1997. What we had done was looked at all 6 batches of vaccine which had been used since 1980, and 7 all of those were free of SV40 sequences. So that 8 gives you some reassurance that these precautions were 9 actually appropriate. 10 In fact, the only preparation which had 11 any SV40 sequence in it at all was a seed virus which 12 was used by a manufacturer for making vaccine from. 13 The amount that was in there was around two logs worth 14 of genome as opposed to seven logs of genomes in a 15 really full-fledged infected preparation. So there 16 wasn't that much in there. The manufacturer had also 17 treated this stuff with toluidine blue, which is 18 supposed to kill of SV40. This was done on the advice 19 of Albert Sabin back in 1960-something or other. 20 But nonetheless, it does seem to me that 21 it's rather a foolish thing to have a seed that's got 22 SV40 sequences in it at all. I think the WHO 23 requirements have now been changed so the seed has to 24 be checked to see if it does have SV40 sequences in 25 it or not. 19 1 This particular seed was not infectious 2 SV40. We did some quite serious studies on it, like 3 transvecting the DNA into cells to see if it would 4 work, infecting monkeys with it to see if we could 5 actually get seroconversion. There was no 6 seroconversion. So there was no infectious virus 7 there that we could actually detect. But nonetheless, 8 the seed did have material in it. 9 If on the other hand you look at materials 10 from around the 1960s or from other parts of the world 11 a little bit later than that, you can pick up SV40 12 sequences quite easily. So the method would have 13 picked it up had it been there. 14 So our conclusion from that was then that 15 really as soon as these kind of precautions were put 16 in place, no SV40 would have been present in all polio 17 vaccines used, at least in the United Kingdom and I 18 would guess in the United States as well, because it's 19 after the same kind of precautions were put in place. 20 So the precautions were adequate. Which means that 21 SV40 exposure of the population through polio vaccines 22 would have stopped around 1962. 23 So what you then have is the problem of 24 the chorioid plexus and appendinoma tumors, which 25 occur in children who are around two years of age or 20 1 maybe less. You have to say well how did they get a 2 hold of the SV40 sequences? One possibility, which is 3 mooted with some enthusiasm is that maybe you are 4 getting passage of SV40 from parents who did receive 5 the SV40 contaminated polio vaccine to their children. 6 So how this stuff gets around is quite important. 7 One of the things that we have been 8 involved in is doing serological surveys of 9 populations to see who has got SV40 antibodies and who 10 hasn't. It is about a five percent seropositivity by 11 the assay that we're using at least. It seems to peak 12 at around age 10 or thereabouts, and doesn't arise 13 after that. 14 So what you could argue then is that you 15 are seeing vertical transmission from parents down to 16 their children. What you could also argue is that you 17 are not picking up SV40 specific antibodies at all, 18 and they could be other human polyomas like the BK or 19 the JC, and it's cross-reacting antibodies that we're 20 picking up. I think that is still a thing that needs 21 to be resolved. This is how we were trying to resolve 22 it. 23 We have access to a number of sinomorgous 24 breeding colonies. One of them at least is absolutely 25 riddled with SV40. It's chronically infected. They 21 1 are all infected basically. 2 So this is just four examples of this 3 particular colony. There's about another 50 or so. 4 This happens all the time. The mothers here are 5 highly sera positive to SV40, all of them. What 6 happens is that the mother and the baby stay together 7 for about six months until the baby is weaned. Then 8 the babies are taken off, no longer being babies of 9 course. They are all banged up together in one 10 gigantic sort of teenage squabbling colony. 11 At the time of weaning, the babies are 12 uniformly negative. So despite the fact they have 13 been on the mother for six months, they have not sera 14 converted to SV40. Almost immediately you bang them 15 up together, or at least within about a month or so, 16 they sera convert. So we actually have a sera 17 conversion panel here, if you like, with about 50 or 18 100 or so sera, where the babies actually were 19 seronegative and then become seropositive. 20 My view on this is probably that the 21 babies don't get infected until you bang them up 22 together. But it may be that they are infected, but 23 they are just not seropositive. So what we have to do 24 here is to fish out the virus from these animals here, 25 and see if it looks like the mother's virus or if it 22 1 looks like the other babies' virus. 2 The point about this long story which I 3 have just been telling you about SV40 is that SV40 was 4 a problem between 1955 and 1962, and it's now 1999, 5 and we still don't really know what was going on. So 6 if you actually make a mistake, it's really quite 7 serious. It may keep you occupied for the rest of 8 your working life. 9 One last quick thing or two last slides 10 here. One is about reverse transcriptase of vaccines. 11 Dr. Schuepbach will be talking later and Jim Robertson 12 will be talking in a moment about detection of reverse 13 transcriptase in chicken cell grown vaccines, such as 14 flu or yellow fever or measles, mumps, rubella. 15 This appears to be due to the presence of 16 defective non-infectious particles. There are 17 sequences from EAV and ALV both in these things, a 18 ratio of about nine to one as I understand it. It 19 does seem to me that you are not really quite sure 20 what the AV sequence is in there and what ALV sequence 21 is in there. It's probably going to vary from chicken 22 to chicken in so far as these chickens have not been 23 bred. In other words, every egg is a new experiment. 24 You are really not quite sure what you are dealing 25 with in that. I think that is quite an unfortunate 23 1 position to be in. I'm not sure how you control it. 2 Finally, this is my last slide, and this 3 has to do with characterized cells. The issues that 4 I have been dealing with really have been to do with 5 primary cells and primary cell problems where the 6 virus comes in direct from the animal origin. I think 7 there is no doubt in my mind that that's the main 8 source of concern in terms of human health. 9 Nonetheless, there are clearly problems 10 which also arise with characterized cells and the 11 continuous cell lines, in particular. We have some 12 down here. 13 Now the regulatory authorities in the room 14 will be well aware of a large number of other examples 15 of this type which don't actually get published. I 16 think that's not so good. I think this stuff really 17 should be out there in the public literature. But 18 nonetheless, these are the ones which are well known, 19 I think. 20 CHO viruses, CHO cells have defective 21 retroviruses. Manufacturers take a great deal of care 22 to actually get rid of them in the final product. So 23 they are endogenous. 24 There are examples of things like BVDV 25 contaminating cells which are growing in culture, and 24 1 also other bovine viruses contaminating cells in 2 culture, particularly when they are grown on a very 3 large scale. Whether or not that poses a hazard is 4 another matter, but clearly there must be methods in 5 place to actually detect them. 6 The classic example here was the minute 7 virus in mice, where the tpa had been grown in CHO 8 cells on a 10,000 litre stove essentially, and then 9 tiled up for an effect with minute virus of mice. Now 10 this was on the order of eight logs, as I understand 11 it, of virus per mil, and yet a 10,000 litre fermenter 12 culture. This is probably more minute virus of mice 13 in one place on the planet than has ever been the case 14 before. You might want to think how you actually get 15 rid of it actually. 16 This is a question of actually getting the 17 cells infected while they are actually burning in 18 culture. So while family cells are clearly a major 19 problem, and while whole animal sources, if you like, 20 are probably the biggest hazard which is likely to be 21 raised in terms of human health, biological aspects to 22 do with well characterized cell banks, where viruses 23 may be introduced from biological materials or they 24 may be introduced by mice walking across the top of 25 the fermenter or whatever, are nonetheless a 25 1 significant matter. It really is not totally clear 2 whether these things have an implication for human 3 health. But I think you would be wise to make sure 4 that they are not actually present. 5 That's where I stop. Thank you. 6 (Applause.) 7 CHAIRPERSON RABINOVICH: Please identify 8 yourself. 9 DR. COFFIN: John Coffin of Tufts. That 10 was a really nice summary actually, Phil. But some 11 caution might be called for in translating the results 12 of vertical transmission experiments from monkeys to 13 humans. 14 As far as we know, simian immuno 15 deficiency virus in monkey populations are not 16 transmitted vertically. Yet HIV-1 is transmitted with 17 reasonable efficiency vertically in human populations. 18 So there may be some underlying biological difference 19 that perhaps a very subtle one, that promotes this 20 kind of transmission in people, where you wouldn't see 21 it necessarily in monkey models. 22 DR. MINOR: Yes. I take you point. We 23 are doing the studies for two reasons. Firstly, to 24 look at natural transmission to monkey on the grounds 25 that it might be a model, although I take your point 26 1 entirely. But also to supply serum conversion panels 2 so that we can try and sort out specificity of 3 immunological reactions as well. I take your point 4 entirely. 5 DR. ONIONS: David Onions, Glasgow. 6 Phil, when people switched to cynomologous 7 monkeys, and I can see the reason because you can pick 8 up SV40. That's very clear. But how do you know that 9 at the same time, you have not invented a new problem, 10 that you have got another polyoma virus in that 11 species that you are not detecting. I mean has anyone 12 done redundant PCR to look? 13 DR. MINOR: I think I would choose to look 14 at a polyoma man to answer that question. Anybody? 15 DR. MAJOR: Gene Major, Bethesda. In the 16 monkeys that we have used for the human polyoma virus 17 studies, we have screened all those animals for the 18 conventional SV40, but not have addressed the question 19 of whether or not there's other additional polyoma 20 type viruses that are present there. 21 Clearly by this time, PCR technology has 22 advanced to the point that perhaps if something were 23 there whose sequences were somewhat similar to the 24 ones that are currently expressed in these animals, we 25 may have picked it up, but we certainly haven't found 27 1 anything yet. 2 DR. LEWIS: Phil, I take it that 3 seroconversion is by neutralizing antibody? 4 DR. MINOR: Sorry? Say that again. 5 DR. LEWIS: I take that seroconversion is 6 by neutralization? 7 DR. MINOR: That's right. Yes. 8 DR. LEWIS: Have you had any chance to 9 look at monkey breast milk to see how long they may be 10 treating antibodian, so that the newborns can be 11 passive immune? 12 DR. MINOR: Right. The answer is no. We 13 have discussed, for example, getting the urine out of 14 these monkeys after they are banged up together. I am 15 told that actually chasing them around the floor is 16 insuperable. I'm not sure you can actually pry a baby 17 monkey from her mother long enough to actually milk 18 her. It's a worthwhile question. I think we'll have 19 another go and see if we can do something about it. 20 There may be some resistance, however. 21 DR. BROKER: Tom Broker, UAB. I wanted to 22 follow up on that exact question. We are facing the 23 same problem with potential vertical transmission of 24 human papilloma viruses. I'll mention it later in my 25 own presentation, but briefly, it does appear there is 28 1 some protective immunity during nursing. On the other 2 hand, removal of an infant, say through adoption to 3 another family, is the highest risk factor for a child 4 acquiring laryngeal papillomatosis later in life. So 5 a two to three year delay. 6 DR. MINOR: So is the assumption then that 7 the infant is infected, but it's not infected properly 8 then? 9 DR. BROKER: It's infected vertically, 10 presumably perinatally, perhaps just before or during 11 delivery, but receive sufficient passive immunity by 12 nursing that gives it life long protection. 13 So what I would propose as a potential 14 experiment is to literally take the, if possible, take 15 the baby monkey immediately away from the mother, and 16 don't allow it to nurse, and then just have different 17 lengths of time of nursing to see if this onset of 18 seroconversion is affected by a timing mechanism. 19 Alternatively, don't ever let that baby monkey be 20 housed with other baby monkeys in the daycare center, 21 and keep it with the mother even if it's not nursing, 22 and see if it fails to seroconvert. 23 So the question is, is it getting infected 24 from its playmates. 25 DR. MINOR: That's right. 29 1 DR. BROKER: Or is it receiving a period 2 of important passive protection from the mom. 3 DR. MINOR: I figured we could do that by 4 looking at the actual strain of viruses the monkeys 5 get infected with. We have a number of different gang 6 rooms, if you like. If you get a different strain in 7 each gang room, but it's the same strain within a gang 8 room, then I think that will answer the same question. 9 You can also go back to the mother and see what kind 10 of strain she's got too. But it's a valid point. 11 AUDIENCE MEMBER: I would like to 12 reemphasize one of the important points that you made. 13 I know you didn't have time to expand on it, but I 14 think it is extremely important. That is the need for 15 those organizations who discover a new virus or some 16 contaminant, cell population used for vaccine 17 production or in a production run, to make that 18 publicly known. 19 I think that the declaration by Genentech, 20 who has published this information under their name, 21 that an NBM contamination occurred in a 10,000 22 fermenter is an act of great courage. I think that 23 that kind of courage, this declaration by other 24 companies in this field, is very necessary for the 25 health of this industry. 30 1 I understand from some of the remarks that 2 have been made that there are others that are known to 3 a small coterie of people here that have not been 4 publicly declared. I urge all of you to think about 5 this seriously because it can and will have a great 6 impact on this industry. Thank you. 7 DR. MINOR: I agree totally with that. It 8 does seem to me that sooner or later the information 9 will leak out. I think the industry looks very bad. 10 DR. VAN DER EB: Van der Eb, Leiden. Did 11 I understand it correctly that ferrisfaruses were 12 found in human embryo material that was used for 13 urokinase production? 14 DR. MINOR: I think the FDA can answer 15 this one better than me, yes. But I mean that was my 16 understanding of it. It's out on the net. 17 DR. VAN DER EB: But where does it come 18 from? 19 AUDIENCE MEMBER: I think it's a rea 20 virus. 21 DR. MINOR: It's various rea viruses, plus 22 others. 23 DR. VAN DER EB: I see. Okay. 24 DR. FRIED: Mike Fried. Was any of the 25 old vaccines from the 1960s that were contaminated, 31 1 were they PCR'd up to show that the virus was the same 2 as being found today? Because it's also possible that 3 we all have a latent SV40 type virus which likes to 4 grow in tumor cells, and that's why you find it. It's 5 a passenger. But I mean since there's polymorphisms 6 in the sequence, if you can go back to the 1960s and 7 then find out if it's the same thing that we find 8 today, it would be helpful. 9 DR. MINOR: We looked at, when you say the 10 1960s, I have to emphasize this is very early 60s. 11 Certainly the things that we have got which came out 12 positive weren't used in the UK, or they might have 13 been used somewhere else. 14 We had a Russian SV40 and we had an 15 American SV40, and we had an SV40 of unknown origin 16 all from the 1960s, and they were all different 17 basically. They were different from the 7-7-7, you 18 know, the cos kind of sequence as well. So they were 19 all unique basically, in terms of the region we were 20 looking at, which was C terminus of t antigen. 21 CHAIRPERSON RABINOVICH: Last question 22 please. 23 MS. MARCUS: Carole Marcus Sequora from 24 Bassey Consulting. 25 I just wanted to clarify that urokinase is 32 1 produced from cells. It's not aborted fetuses. It's 2 newborns who did not survive. Just for the record. 3 DR. MINOR: Thank you. 4 MS. MARCUS: It was rea virus. 5 DR. MINOR: I'm sorry about that. 6 CHAIRPERSON RABINOVICH: Thank you. Our 7 next speaker is Dr. Jim Robertson, speaking on 8 experiences with retroviruses in avian and mammalian 9 cell substrates. 10 DR. ROBERTSON: Good morning. For those 11 of you who don't know, NIBSC is CBER's cousin from 12 across the pond in the U.K. What I am going to do is 13 pick up where Phil left off and concentrate on the 14 retrovirus aspect of viral contamination. Initially 15 I will look at say biologicals in general, but 16 ultimately focusing down on the vaccine issues. 17 So I will begin with some direct 18 information regarding retrovirus situations with 19 biologicals. I will go onto look at how some of the 20 regulatory guidelines deal with the issue of 21 retroviruses. I will go into look at RTase testing, 22 which is a reasonably current them just now, and 23 finish up looking at the recent situation of the 24 finding of retroviral-like particles in avian cells. 25 So to begin with, here is a short list of 33 1 the incidence of retrovirus contamination found in 2 biologicals in general, not just vaccines. I have 3 sub-divided these into two groups here. You see this 4 upper half here, this is where we have in the past had 5 overt adventitious contamination by a retrovirus of a 6 biological. For instance, being mentioned earlier, 7 ALV, that causes virus in yellow fever vaccine by 8 virtue of producing the vaccine in embryonated eggs 9 infected with the virus. The other one that was 10 mentioned earlier by Phil, HIV and blood products. 11 The bottom half here is a quite, somewhat 12 separate type of contamination. In fact, you might 13 find it equaler to call it contamination or not. 14 Certainly these are not adventitious situations. 15 These are situations in which an endogenous 16 retroviral-like particle is present in the 17 manufacturing process. 18 In the first instance here, it's 19 established that murine hybridomas used in the 20 manufacture of monoclonal antibodies produced, 21 secretes C type particles. These have been tested in 22 a variety of other mammalian, including human cell 23 lines, and generally are not infectious. 24 The titre can be very high for these types 25 of particles. You can get 10 to the sixth particles 34 1 per mil. I've even seen 11 particles per mil in one 2 instance. So you can have a very high burden of 3 direct viral particles. 4 It is also well established that CHO 5 cells, which are used for producing 6 biopharmaceuticals, secretes C type particles. You 7 also get intertestinal type particles from these 8 cells. These are probably much more characterized, a 9 bit more work has gone into describing the particles 10 from CHO cells, sequence information from the 11 endogenous elements within the CHO genome, which is 12 producing these particles, give some ideas as to why 13 they are defective. The reading frames are 14 incomplete. There are stop signals. So you don't get 15 a proper infectious virus from these endogenous 16 elements. 17 The latter type is the only type here 18 that's dealing with vaccines, even dry vaccines 19 produce either an ovo or cef cells. I'll come back to 20 that in a few minutes. 21 From a regulation point of view, how do we 22 deal with virus contamination and retrovirus 23 contamination? There are a couple of guidelines I 24 would like to bring to your attention. The first one 25 here is an ICH guideline, which looks at viral safety 35 1 evaluation. Admittedly it is only for biotech 2 products. The scope of the guideline does comment 3 that this is not, this guideline is not applicable to 4 vaccines. But I think it is worth looking at what it 5 says about virus contamination. 6 Within the document, it describes five 7 different cases of potential contamination, starting 8 from the most desirable case, where you don't have a 9 virus present in the process in any way, down to the 10 worst scenario where you know you've got a virus, but 11 you haven't a clue what it is. 12 The guideline goes on to state what is 13 acceptable and what is not acceptable in the 14 manufacturing process. The only two cases which are 15 generally acceptable of the first two cases, Case A, 16 where you have got no virus, and Case B, where you 17 have got a non-pathogenic retrovirus. The other cases 18 are only exceptional. Generally you don't want one at 19 all. The manufacturing is not allowed when you have 20 got a virus contamination. 21 So for Case B, really what you have here 22 is a murine retrovirus is probably the only 23 contaminant acceptable in the bulk harvest. If you 24 remember this guideline is applicable only to 25 recombinant products and not to vaccines in general, 36 1 and these recombinant products are highly purified. 2 The other guideline is the WHO 3 requirements which came out recently for use of animal 4 cells as in vitro substrates. That does include 5 vaccine production. 6 When it comes to testing for retroviruses, 7 this guideline has several -- many other guidelines in 8 the past have indicated, the types of assays being 9 used for retroviruses, specific infectivity assays, 10 electron microscopy and transcriptase assays, are the 11 three general approaches for checking for retrovirus 12 contamination. 13 There may be the use of specific antigen 14 detection as is in some particular cases, but these 15 are the generally recognized methods of going about 16 picking up retros. 17 If I can concentrate now on the RTase 18 assays. The traditional type of assay involves 19 incorporating a nucleotide precursor, a labeled 20 precursor of some kind into an assay using a rather 21 synthetic type of template. Then more recently of 22 course we have the PERTs, PB RT, AMP RT type of 23 assays, which includes a PCR amplification step, with 24 the result that these type of assays are incredibly 25 more sensitive than the more, as I can call it, 37 1 traditional type of assay, and what is often quoted as 2 up to a million fold times more sensitive by virtue of 3 incorporating a polymerase chain reaction. 4 Now using this type of assay, the cat was 5 set among the pigeons. When this paper came out, I 6 might even say that the fox was set amongst the 7 chicken coop. Detection of reverse transcripted 8 activity in live attenuated virus vaccines. This 9 quite naturally caused a bit of concern as to what was 10 going on here. The vaccines indicated, the one common 11 feature was that ovine produced in eggs of some kind, 12 measles vaccine out of CEF primed cultures, similarly 13 mumps. Yellow fever and influenza in ovo. But not 14 measles vaccine out of human diploid cells or rubella 15 vaccine out of human diploid cells. So the common 16 link here seemed to be the CEF, the chicken source 17 used in the production of the vaccine. 18 We joined in the boat here and started 19 looking at this issue. Every type of hen fluid that 20 we have looked at, CF fluids or an type fluid from a 21 variety of different strains of hen have all been 22 positive in the assay for reverse transcriptase. 23 Summarily, quail, jungle fowl, and pheasants are 24 positive. 25 The types of sources of fluids which have 38 1 been negative for reverse transcriptase are listed 2 here. Some species are not positive, turkey and duck 3 cultures, quite a range of human cell lines. Simian 4 rabbits have been tested and found to be negative. So 5 the clear source of this RTase that was being picked 6 up in the vaccines is quite clearly of chicken avian 7 origin. 8 We would want to look at -- I should add 9 that this RTase was known at the time to be particles 10 associated and appears in the supernate of the cells. 11 We are going to look at this particle to see if it 12 would pick up any infectivity. In all, we looked at 13 10 different cell lines, mainly human, but including 14 rabbits and turkey. Over 21 tests and 116 passages. 15 In each case, in every test and at every 16 passage level, the cultures were negative for reverse 17 transcriptase activity. There's absolutely no 18 indication that this particle is infectious. Since 19 then, CBER and CDC have also come up with similar 20 data, including use of PBMCs. No infectivity 21 associated with these RTase containing particles. 22 Where might these be coming from? 23 Presumably they are derived from endogenous 24 retroviral-like elements in the chicken. The 25 information to date regarding such elements in the 39 1 chicken genome are quite well characterized EV loci, 2 which are related to the avian chosis virus family, 3 and more recently discovered about 10 years ago, EAV-0 4 family, which is an older element than EV, and then 5 older still, ART-CH and CH-1 elements. 6 The information at the time and pretty 7 much where it still exists is that we knew that 8 there's a line of chickens which was negative for EV. 9 It had been eliminated from the genome. This line of 10 chickens, the culture fluids were positive for RTase. 11 So we knew that it had to be at least one of these 12 elements producing RT activity. At the same time, you 13 couldn't eliminate the fact that EV might also be 14 producing RT activity. The best bet was EAV-0, given 15 the sequence information that was present at the time. 16 More recently, in the last year or two, 17 Joerg Schuepbach's laboratory has produced a good 18 evidence for the presence of EAV-0 derived RNA 19 associated with the RT particles secreted from CEF 20 cells, and then this year, Walid Heneine, CDC, also 21 produced the presence of EAV and ALV RNA. When I say 22 ALV, I mean derived from the EV loci and not exogenous 23 ALV contaminating RNA. 24 So what can happen here retrovirus-like 25 particles, defective particles being produced from 40 1 endogenous elements both from EV and EAV-0 family of 2 endogenous elements. The presence of the RNA and 3 relsconstrictase in a particulate fraction leads one 4 to come to the conclusion that we have retroviral like 5 particles in the CF fluids of the chick cells, which 6 is present in the vaccines measles and mumps. 7 The absence of infectivity in the current 8 genetic information, sequence information that we have 9 on EV loci and the EAV-0 family of endogenous elements 10 strongly indicates that these particles are defective 11 viral particles. The only question mark that remains 12 from the regulation point of view, but also scientific 13 point of view, the possibility of pseudotype formation 14 during vaccine manufacture. The current evidence 15 suggests the particles that are defective in the 16 envelope-like protein and so there's a particulate of 17 pseudotype formation with the glycoprotein of vaccine 18 viruses being grown in the CF cells. 19 So to summarize a couple of these issues 20 then, from the practical point of view, testing for 21 reverse transcriptase as an indicator of retroviral 22 contamination, these assays are evolving, changing all 23 the time. One has to take into consideration the 24 strength of the assay and the validity of the assay. 25 There may be different requirements within an assay 41 1 for different sources of RT. It may be necessary to 2 use some other sort of method to asses the 3 significance of any RT detected because we know that 4 RT activity can derive from other enzymes. Telomerase 5 is or DNA polymerase, cellular DNA polymerase is. 6 These features are not specific to the more recent 7 sensitive type of assays involving PCR, the parents, 8 and the PBRT. These features were also factors that 9 had to be considered in the more traditional types of 10 assays. 11 It is often quoted that the RT levels in 12 chick cells is very low, given that it was detected by 13 a very, very sensitive assay, and has not been 14 detected by the more traditional type of assays. 15 Certainly some preliminary data that I have got 16 suggests that it is not quite as low as we first 17 thought. Really this RT activity in chick cells and 18 ultimately in vaccines is only just below the level of 19 detection of the more traditional type of assay. In 20 fact, this was a relatively novel phenomenon 21 discovered just a few years ago. It was in fact first 22 reported 20 years ago in the late 1970s by Berne and 23 Hofschneider at the Max Planck Institute in Munich. 24 They reported the presence of a novel type of RT 25 enzyme in chick embryos and in chick cells. That was 42 1 in the days before PCR. 2 So the level -- I certainly believe the 3 level of RT and the level of these particles is 4 actually quite high in chick cell fluid. Ultimately 5 I think what we have to do is look into the need for 6 standards, standard materials in some way to assess on 7 a quantitative basis the level of RT activity in chick 8 cells, in measles vaccine, in mumps vaccine, in order 9 to come up with some meaningful conclusions regarding 10 it. 11 So to look at the RT issues from a 12 regulatory point of view, a couple of comments I would 13 make. First, that these are state of the art 14 technologies. When these highly sensitive assays 15 first came about, it posed very useful from a research 16 point of view to what use are they in a routine 17 manufacturing validated type of assay. 18 I think the time has come where yes, you 19 would say that these are state-of-the-art techniques 20 and can be and should be used for detecting the 21 presence of RT in your manufacturing process. However, 22 when it comes to, for example, chick cells, and until 23 we have a greater understanding of what the levels 24 might mean, and until standards are available, there 25 is really -- it is difficult to justify any 43 1 requirements to perform RT or PBRT assays on systems, 2 and basically here I am talking about chicks, which 3 inevitably will be positive. We know they are going 4 to be positive, that there's no great need to actually 5 require any manufacturer to do these assays. But 6 certainly there is a still a strong requirement to 7 provide evidence for freedom from retrovirus 8 contamination. This will have to derive from other 9 data. Thank you. 10 (Applause.) 11 CHAIRPERSON RABINOVICH: Thank you, Dr. 12 Robertson. 13 Any questions? 14 AUDIENCE MEMBER: Just a comment. For 15 known endogenous avian retroviruses or exogenous avian 16 retroviruses, of the cell lines, of the cells that you 17 tested for infectivity, only the turkey cells would 18 have given a positive result. I would urge for avian- 19 derived -- urge the use of those cells, and a PERT 20 assay is a sensitive readout, for detection of perhaps 21 unknown agents in these vaccines, end products, as 22 being the most sensitive, at least for avian-derived. 23 DR. ROBERTSON: Yes, yes. The turkey 24 cells are sensitive for rav, because it's virus, but 25 apparently not for the RTase. I have no idea, duck 44 1 cells are also negative, but I have no idea if duck 2 cells are susceptible to the -- 3 AUDIENCE MEMBER: They are not as good as 4 turkeys. Turkeys themselves actually are not 5 sensitive to all exogenous ALVs, but to most they are, 6 and all endogenous ones. 7 DR. ONIONS: David Onions. I really 8 enjoyed that, Jim. I just want to make a comment on 9 your comment about standards. I think as we heard 10 from Keith last night and what we're doing, and I'm 11 sure George is doing too, using the tac man technology 12 where you can actually quantitate the PCR product. 13 Then if you actually do EN counts of virus particles, 14 dilute these out, you can actually quantitate your 15 assay system and actually determine the number of 16 particles you can detect. 17 Now it seems to me that that is a useful 18 kind of standardization, and that you can then relate 19 that to if you like, a consistency of your starting 20 material, in this case the egg. 21 So I think in that case, applying those 22 techniques does have value, because it gives you a 23 kind of lock-to-lock consistency of your materials. 24 So that if something goes out of spec, then perhaps 25 something odd is going on in those materials. 45 1 DR. ROBERTSON: Whatever we want to 2 approach this standardization, one would have to do it 3 on a quantitative basis. I was sensitive earlier to 4 quote any actual figures, but to quote you some 5 figures that I have got so far, in one chick cell 6 preparation, the culture fluid, there was the 7 equivalent of the order, and this the first of 8 investigation, 10 to the 4 focus forming units of rav 9 in uninfected chick cell fluid. 10 A large current high level is when your 11 typical infection goes up to 10 to the 6th, I believe, 12 focus forming units. You are only talking about 104 13 drop, lower value. So you are not far away. It is 14 going to be difficult. If you have got an overt 15 contamination going up to 10 to the 6th, I think that 16 will be quite clear on a quantitative basis. But once 17 you drop down a bit, it is difficult to say whether 18 you have got an infection or whether it's just 19 background level of endogenous RT-derived activity 20 that you are picking up. 21 DR. MYERS: Martin Myers from National 22 Vaccine Program. As I sit and count the number of 23 immunizations that various populations receive with 24 these particles in it, repeated immunizations with it, 25 I wonder if there is any data on sero responsiveness 46 1 in longitudinal -- 2 DR. ROBERTSON: Well the reason I am 3 looking over your shoulder, we have Walid Heneine from 4 CDC. I'm not sure if you are going to say something 5 along those lines, but serologically, there is no real 6 evidence for reaction to ALVs. Epidemiologically, 7 when it first came out, we also, not ourselves but 8 epidemiological colleagues, to provide information. 9 There is no evidence again, for any increase in the 10 incidence of childhood cancers since the onset of 11 measles, mumps vaccination. 12 Walid, were you going to say something? 13 DR. HENEINE: Yes, I have. Regarding 14 transmission risks, so far the data we have where we 15 have looked at the presence of antibodies to avian 16 leukosis virus by western lot, I'd say we developed as 17 well as presence of ALV sequences and EAV sequences in 18 the peripheral blood lymphocyte from vaccinated kids 19 as well as in plasma. So far, the results have been 20 all negative for both viruses. 21 I have just one comment regarding 22 referring to these viruses in general as defective. 23 My comment is that the evidence we have so far on 24 those we have studied in a couple of vaccines between 25 in Europe and in the U.S. suggests that those, for 47 1 example, ALVs we're dealing with could be defective 2 because they come from loci that have deletions. 3 However, this may not be true for all the contaminants 4 from it we might find in other vaccines, because these 5 contaminants reflect the particles expressed from 6 these loci in the different cell substrates that Phil 7 mentioned in his talk, that the nature of these 8 particles and their phenotypes would vary depending on 9 the presence of the particular loci in these 10 substrates. 11 So just a comment, not to generalize that 12 we always should expect to have defective particles. 13 We might or we might not in certain cases. 14 DR. ROBERTSON: Yes. 15 DR. SCHUEPBACH: Joerg Schuepbach from 16 Zurich. I would also like to make a comment regarding 17 seroconversion. We have done two vaccination studies, 18 one with yellow fever vaccines, where about 120 19 individuals were tested actually for reactive 20 antibodies against HIV. The reason for this was that 21 in 1991, there have been reports that false positive 22 HIV reactions were found in people that have been 23 vaccinated against influenza. Influenza vaccine is 24 one of the vaccines which contains the EAV. 25 So we found that statistically highly 48 1 significant the vaccinees which have received the 2 vaccine which have the higher content, about 80 times 3 higher than the other vaccine which we used to use 4 too. That these patients actually had highly 5 significantly elevated antibodies to HIV 1 and 2, 6 although none of these actually became sero positive. 7 The serious bonds was highest in those individuals 8 that also had a history of measles vaccine. 9 In the second study, this was an influenza 10 vaccine where we tested two different brands, a split 11 vaccine and a rather crude vaccine. We also had a 12 response to HIV bond two in the third generation tests 13 in the vaccine which contained more in the crude 14 vaccine, which contained more of the EAV protein, and 15 again those individuals who had a history of yellow 16 fever vaccination had the highest type, the highest 17 increases. 18 We also tested yellow fever vaccinees by 19 PCR, RNA PCR and DNA PCR for EAV-0 sequences. We 20 found one out of 180 individuals in which both these 21 tests were positive as plasma on PBNC. At the moment, 22 we cannot exclude that this was the result of a 23 contamination, but we are working on that. So I think 24 that the matter is actually not as clear as has been 25 presented by others. 49 1 MS. SHEETS: Hello. I'm Becky Sheets from 2 FDA. How would you recommend that avian-derived 3 products be tested for retroviruses? EM is not very 4 sensitive. The conventional test is often inhibited 5 by the allantoic fluid, and therefore, is not 6 necessarily a valid test. How would you recommend, if 7 you don't use a PCR-based RT des? 8 DR. ROBERTSON: At the end of the day, it 9 would have to be an infectivity assay. There are also 10 some antigen, ELISAs for the viral antigen. I don't 11 have experience with those. I don't know the 12 sensitivity of them. Ultimately you are looking at an 13 infectivity assay which can be performed even on chick 14 cells, which are positive for RTase. One could assess 15 for after several passages on chick cells, looking at 16 an increase in RT activity or increase in antigen. 17 MS. SHEETS: When you said infectivity 18 test, were you talking about those specific for ALV or 19 were you talking about general tests to detect any 20 kind of retrovirus? 21 DR. ROBERTSON: Well, it would have to be 22 an avian retrovirus if one is performing the assay on 23 chick cells. Propagating the material, the test 24 material in chick cells but using either RTase in 25 general or an ELISA specific for ALV test for 50 1 increased presence of either RT or for the presence of 2 ALV antigen. 3 CHAIRPERSON RABINOVICH: Dr. Coffin, the 4 last question? 5 DR. COFFIN: Yes. I'd like to actually 6 address Dr. Schuepbach's comment. Did I understand 7 that you were basing your sero assays on the rationale 8 that there might be cross reactivity between ALV and 9 retro viruses in HIV? There is no rational basis for 10 that. There's virtually no amino acid the same 11 between those two viruses, except for some extremely 12 highly -- you know, three or four in pol and some 13 other places. Did you actually assay directly for 14 seroreactivity against ALV? It would have been a much 15 more straight-forward experiment. 16 DR. SCHUEPBACH: We agree that there is no 17 sequence homology on the nucleic acid and on the 18 protein label, but these are the results which we 19 found. We have to find an explanation for them. We 20 don't have at the moment. 21 CHAIRPERSON RABINOVICH: Thank you, Dr. 22 Robertson. 23 Our next speaker, Dr. Jorg Schuepbach, 24 from the Swiss National Center for Retrovirology. 25 Induction/activation and detection of occult viral 51 1 agents that are present in mammalian tissues. 2 DR. SCHUEPBACH: May I have the first 3 slide, please? Okay, from previous remarks, I heard 4 that I was expected to talk about these avian 5 retroviruses as well, but actually I was asked to talk 6 about the induction activation of occult viral agents. 7 So I will just have a few remarks on this other stuff. 8 So occult viral agents are agents you 9 don't detect or at least do not easily detect. They 10 may include two groups: a group that includes known 11 agents which are present at two low concentrations for 12 easy detection. The reason for these may be latency; 13 The other group consists of unknown agents. Since we 14 do not have good detection methods for these, they may 15 be present at low or also at higher concentration. 16 Viruses known for their latency or various 17 types of the herpes virus, true, they are latent in 18 various types of non-permissive cells such as neurons, 19 B cells, monocytes, PBLs, and others. They are 20 activated from these latent stages by various kinds of 21 stimulation of their host cells by differentiation, 22 agents by the differentiation of precursor cells, to 23 more mature cells. Again, by other activating agents. 24 Other viruses could be considered in 25 addition to the herpes viruses include the adeno 52 1 viruses, the adeno-associated virus and the pathyloma 2 and polyoma viruses of which we heard yesterday, and 3 will hear more in a subsequent talk. 4 Regarding the RNA viruses, I might discuss 5 the measle viruses and of course the retroviruses. 6 When we look at the mechanisms by which we 7 can activate these various viruses, it is mostly by 8 activation of their host cells, by cell stimulation, 9 by induction of cell differentiation of these cells, 10 and then by co-cultivation with cells which are 11 permissive for replication. 12 Now since we have different viruses and 13 host cells systems, these methods vary greatly among 14 the different viruses. If you have unknown viruses, 15 you really don't know what to do. 16 So the effect of such activation would be 17 that from a lonely latently infected cell, by 18 inducement of replication, a virus would spread 19 throughout the culture, resulting in virus gene 20 amplification in production of viral proteins. So 21 this would make of course the detection easy. You 22 might also have some pathogenicity which is easy to 23 detect. 24 However, our goal is actually not 25 detecting any possible virus that might be present. 53 1 The principal goal is to provide a virus production 2 system which is free of such agents. It is suggested 3 here the easiest way of achieving this is actually 4 cellular cloning. Because if you have an agent that 5 is present in only a minority of the cells, the 6 chances that you derive a clone that is free of these 7 agents is very high. 8 If by chance you hit an infected cell, the 9 descendants of that cell will all carry along the 10 virus and of course then we come into a situation 11 which makes detection of unknown viruses and also 12 known viruses much easier because either all of the 13 cells will be infected or none at all. 14 So cellular cloning, if we hit an infected 15 cell, has actually a viral gene amplification effect 16 which is comparable to virus induction activation if 17 it's successful. Most importantly, it is a procedure 18 that works for all the latent viruses except 19 endogenous retroviruses, but these are present anyway 20 in all of the cells. 21 So going on to the detection methods for 22 these agents, let's first talk about known viruses. 23 Since all the cells will be infected, we actually do 24 not need the most sensitive procedures. We do not 25 need procedures that detect the single viral copy. 54 1 What we need is broadly reactive methods which go 2 detect all the different members of a certain virus 3 group. 4 So I think techniques, old-fashioned 5 techniques like hybridization techniques on the low 6 stringency or if we want to use PCR or nucleic acid 7 based methods, we should take care that we take a lot 8 of different probes, use data generated primus, 9 multiplex PCR and so on. 10 Of course in addition, we should also do 11 the classical methods, doing cell activation and co- 12 cultivation as permissive cells, the routine detection 13 methods of broadly reactive antibodies which detect 14 all the different members. 15 For those who think that what I have told 16 so far is rubbish, and that we actually do need very 17 sensitive methods, I offer the mega PCR, which has 18 also been named catcher PCR by others. The purpose of 19 this method is to take very rare sequences among a 20 very high background of DNA or RNA. So here we 21 convert the samples of up to 500, maybe even 1 22 milligram of DNA or respectively RNA. 23 The principle is very simple. We use 24 biotinylated capture probes which bind to these 25 sequences inquest. We isolate these complexes on 55 1 coded beads, wash the rest of the DNA away, and them 2 amplify these by PCR with primus which are located 3 outside of this capture probe. 4 The advantage of this is that we 5 absolutely do have no carry over because the amplicons 6 are selected against when we do the capturing. It is 7 this type of test which I would actually like to have 8 been seen when testing in the question of 9 xenotransplantation where the PERV sequences can be 10 found in humans which have received pork material. I 11 think this will be the test, to test these questions. 12 Now using this method, it's actually very 13 sensitive. You can detect a single copy here of HIV 14 DNA. We still have double positive signal, is about 15 one copy. This serial dilution was done in the proper 16 range here. The fact that in these two, three last 17 dilutions only one of the two duplicates was positive 18 clearly demonstrates that we are in a Poisson 19 distribution. So we can detect the single copy with 20 this method in 100 microgram. 21 DNA, we have actually demonstrated that 22 there's 95 percent probability we can detect three 23 double standard HIV copies in 100 micrograms of DNA. 24 So now going on to the exclusion of 25 unknown viruses, and I will talk about retroviruses 56 1 later, we can actually use the same procedures as I 2 have already described previously. We just have to 3 take care that we really have broadly reactive 4 methods. This is true for molecular based tests as 5 well as for the more classical procedures. 6 Now coming to retrovirus detection, of 7 course also of cell cloning, here we have two 8 situations, the exogenous retrovirus may not be 9 present in none of the cells or in all of the cells. 10 The endogenous retroviruses were always present in all 11 of the cells. The known exogenous retroviruses are 12 detectible by tests for conserved sequences. Of 13 course you might also use universal pool primers for 14 unknown retroviruses -- because of the endogenous 15 retroviruses. Not all of which, or very few of which 16 are actually harmful. 17 So I think it is better at this time to 18 switch from the analysis of cells to the analysis of 19 particles. This is best done by the PERT assay which 20 has been mentioned before by several speakers. 21 Now when we devised this test in 1992, we 22 devised it as an anti-family of related tests which 23 would have in common that reversed inscriptase present 24 in a sample would be used to create from a template 25 primer combination and nucleic acid that is to be 57 1 unamplified. 2 Now in most instances, this will simply be 3 the cDNA. There are other possibilities as well. You 4 can take any nucleic amplification procedure, not just 5 PCR. You may also use ligase chain reaction or NASPA 6 or you can make use of auto replicated DNAs or RNAs in 7 order to generate amplification product, which can 8 then be assayed by different methods. 9 So since we have provided for all these 10 different methods already in 1992, we do not think 11 that it is necessary to invent new names for these 12 current assays. 13 Now this test is actually very sensitive. 14 This experiment in comparison to classical RT assay. 15 It occurred as six to seven orders of magnitude more 16 sensitive, and in a direct comparison with -- in the 17 case of HIV, where we compared the method with RT PCR, 18 detecting one copy of cDNA, we had the same dilution 19 endpoints for two different samples. 20 Actually as others, we can detect only a 21 few particles in the case of HIV. We believe that in 22 some cases we can detect even less than one particle. 23 Now this is one of the theories taken from 24 the Joerg Koenig paper in 1996, where we demonstrated 25 that the measles vaccines, the mumps vaccines, the 58 1 yellow fever vaccines, and the MMR vaccines all 2 contain activity which is about three orders of 3 magnitude higher than the background here on other 4 vaccines, and were negative. 5 Now in order to identify the viruses 6 behind these activities, we along with the PERT assay, 7 developed the method for the identification of unknown 8 retroviruses. It is based on three properties of old 9 retroviruses, namely, that they all are 10 polyadenylated, that R sequences are repeated at both 11 ends, and that cDNA synthesis has started here at the 12 primer, binding site, and that for primers, tRNAase 13 are used and the use of such tRNAase is actually very 14 much restricted among the various retroviruses. For 15 example, is just four PRNA primer equivalence. You 16 can start cDNA synthesis for all exogenous 17 retroviruses known today. 18 So what we do is that we bind the 19 retroviral RNA to poly t coated beads. Then we start 20 here, the synthesis of the cDNA with one of the 21 various t RNA primers, synthesizing the strongest of 22 DNA. Then adding a tail here, and then with anchored 23 TCR, we can amplify this sequence and submit the 24 sequencing directly. 25 Actually this method has also been used by 59 1 the group of Dr. Loewer at the Paul Ehrlich Institute, 2 and even published before us. But we have somehow 3 optimized this procedure, so in general we need less 4 than one-thousand RNA sequences, sometimes as few as 5 20 or 40, 50, in order to generate this sequence here. 6 As soon as you have it, you actually know whether you 7 are dealing with a retrovirus or not. 8 When you deal with a retrovirus, you have 9 to R sequence and then you can check with the other 10 anchored PCR. Where there is R here, it's repeated at 11 the three prime end. If it is, you can then amplify 12 the entire genome with a little bit of luck by long 13 PCR. 14 So this is what we use to identify this 15 EIV-O sequence. We have also done some other work. 16 For example, we investigated the NIH 323 cell line. 17 This was negative by convention RT tests, but positive 18 by PERT assay. We had a nice band in sucrose, and 19 then radiant. Using this procedure which we call 20 parar, we identified 23 different products, 15 of 21 these were actually retroviral sequences from four 22 different groups. Three of them were unknown 23 sequences, at least at that time. So far we have not 24 further characterized these sequences, but this is 25 still awaiting. 60 1 Now staying with retroviruses, as Dr. 2 Coffin pointed out yesterday, sometimes if you have a 3 cell line here, you are dealing with melanoma cell 4 lines which were found to be highly percipated by PERT 5 assay. We analyzed what was in there. It turns out 6 to be endogenous murine leukemia virus, and later we 7 were told that these cell lines have actually been 8 passaged in mice. 9 If you have low titres of activity, then 10 that becomes a little bit more complicated. This is 11 the analysis of primary samples from a patient with MC 12 cor cultures. No actually not cor cultures, just cor 13 cultures which were found lowly positive in the PERT 14 assay with activity in the order of two, maybe three 15 times above background. 16 Here the patterns is a little bit more 17 complicated. You have here a small peak that might 18 correspond in density to ritualized particles. This 19 one might correspond to cor particles. You have 20 another identified -- unidentified peak here. It will 21 certainly be a challenge to find out what this stuff 22 is. 23 Next, please. This is another example of 24 a primary culture where we have a very short peak at 25 the higher density. This might be for particles, 61 1 could be a different retrovirus, a different virus, or 2 just a subcellular particles containing some cell or 3 enzymes. 4 Now you will say that this test of course 5 detects only retroviruses that are released. We are 6 also worried about retroviruses that are inside the 7 cells, so stimulation may be necessary. Actually I 8 think one important question is or one possibility is 9 that actually the vaccine virus we would like to 10 produce in such a cell might activate latent 11 proviruses. So I think it is important that we 12 actually do not just test the virus production systems 13 while uninfected, but also when this seed virus has 14 been added, and then we harvest the virus. 15 Now in some cases, as in the measles virus 16 or so, this has proven very easy. We had quite a good 17 specificity. But in other cases, it might be more 18 difficult as indicated in this example, where we 19 tested a vaccine, experimental vaccinia, recombinant 20 vaccinia virus vaccine against melanoma. This was 21 found highly positive by PERT. It had actually been 22 produced by just the lysing, the infected cells by 23 ultrasonication. 24 What we now find is here in black, is the 25 vaccinia virus DNA two peaks. We have here a major 62 1 peak of RT activity which does not coincide with the 2 vaccinia virus peaks, and also is not characteristic 3 of retroviruses. So I think in this case, we can rule 4 out the presence of a retrovirus. 5 Now it may also be interesting to find out 6 whether upon induction, viruses might come out. So 7 this would add an increase of safety to the vaccine. 8 As retroviruses are regulated, you have the promoter 9 in anti sequences in the upstream LTR in the U3 10 region. Depending on the cell type, activation state 11 of the cell and the differentiation, you have various 12 sets of transcription favors interacting with this 13 enhancer regions. 14 In addition to this balance of positive 15 and negative transcription factors, you may have 16 positional effects as the chromatin structure or the 17 DNA methylation. You may now try to influence this 18 balance by tipping it by either inducing mitosis cell 19 differentiation by substances that lock inhibitors or 20 by alleviating the negative positional effects, again 21 by inducing mitosis or by inducing DNA de methylation. 22 The number of induces have been described 23 in the past. The most important ones are listed here 24 at the top, allogenated pyrimidins, the azacytidine, 25 which only both of them working only in infected 63 1 cells. I will not mention the others because of the 2 lack of time. 3 Now it depends a lot on the virus whether 4 azacytidine or the deoxy pyrimidine is preferable. 5 For example, in experiment in cell line where two 6 types of different retroviruses are produced, several 7 type A particles here. The azacytidine is certainly 8 better. But in C-type particles, these cells produced 9 IdUdr. Yes, the IdUdr is better. So you might have 10 to use a combination of these two drugs. 11 So in conclusion, I think induction 12 activation certainly serves to amplify latent viruses 13 for which improved detection. I think it is more 14 important that we early in the process of selecting 15 virus production systems be cloned B cells, and sub- 16 cloned, because this will amplify, because this really 17 facilitates detection very much. 18 In consequence of this, we do not -- I 19 think this is very important. We do not need the most 20 sensitive procedures. What we need is broadly 21 reactive procedures which will detect all the 22 different agents. 23 I also think that at the end, the only 24 important thing actually when dealing with 25 adventitious agents, not just with DNA, which might be 64 1 infectious, is that the vaccine is free of these 2 contaminant viruses and for retroviruses I believe 3 that this can be verified by the PERV assay. Thank 4 you. 5 (Applause.) 6 CHAIRPERSON RABINOVICH: We'll take just 7 a couple of questions because I would like to leave 8 the rest for the panel discussion. 9 DR. COFFIN: John Coffin. I would agree 10 that if you get preparations of vaccines that are 11 negative by all these assays, you can have a pretty 12 good level of confidence that they are not 13 contaminated with retro viruses. The problem is, if 14 you do these enough, it may well be that no vaccine 15 will pass these tests. 16 What I think is very important to add to 17 this would be one more level to your last slide. That 18 is an infectivity step. As in the example we saw 19 before when one perhaps collects a panel of cells or 20 cell lines which are pert negative, and there seems to 21 be reasonable numbers of those, and then test the 22 vaccine, the induced stuff and everything else by 23 infectivity and induction of pert activity on those 24 cells. 25 I think that would be a much more useful 65 1 and reliable test for the presence of viruses that 2 might be problematic than simply looking at the pert 3 activity in preparations with cell soups. 4 DR. SCHUEPBACH: Yes, I agree with you. 5 I actually thought that was included in those 6 conventional methods which I have listed for the known 7 viruses. Of course you should also do some studies 8 for retro viruses. 9 AUDIENCE MEMBER: You mentioned results 10 associated with particles from supernatants of primary 11 human materials. Did you try to find retro virus-like 12 sequences in these particles for para assay? 13 DR. SCHUEPBACH: Yes. These are very 14 recent results. We are in the process of doing that. 15 DR. KRAUSE: Phil Krause, FDA. One of the 16 issues in testing vaccine products is obviously what 17 tests are available and have been validated and that 18 we understand the sensitivity of. So I guess in the 19 context of thinking of highly conserved sequences to 20 which we might develop primers that could detect a 21 broad array of viruses, including some unknown related 22 viruses, what can you say about the current state of 23 the art? How good is that? How well has that been 24 validated? Is that something which if we decided 25 tomorrow we wanted to apply that to new vaccines 66 1 producing neoplastic cells, we could simply say "let's 2 do it" or is more work required? 3 DR. SCHUEPBACH: I'm actually not very 4 familiar with other viruses than retro viruses. But 5 I think these things, however they exist, should 6 clearly be developed. 7 CHAIRPERSON RABINOVICH: Thank you. We 8 will go onto our next speaker, Dr. Jens Mayer, from 9 the University of Pennsylvania. The status of HERV in 10 human cells. 11 DR. MAYER: Okay. My talk will deal with 12 -- can I have the first slide, please? Okay. My talk 13 will tell you something, I hope, about the status of 14 these human endogenous retrovirus regarding the coding 15 capacity and the expressions. Just again, it was 16 mentioned before already what is actually an 17 endogenous retro virus. HERV is created by the germs 18 of infection of an exogenous retro virus. This leads 19 to radical inheritance of this newly created virus 20 following generations. In the course of the 21 evolution, it will be also inherited to newly arising 22 species. 23 The human genome, like all mammal genomes, 24 and also some invertebrates, invertebrate genomes 25 where it has been shown, contains several families of 67 1 elements and so on. It has been estimated that about 2 one percent of the human genome of such retro origin. 3 These elements antiquated already several million 4 years ago through the genomes of human predecessor 5 species. Some present for at least 30 million years. 6 Some have been shown to be present for at least 40 7 million years. We have several indications of 8 different various families. So they were independent 9 of several exogenous retro viruses. Some of these 10 elements that are now present in the human genome 11 existed. Single copy, and some have copies, copy 12 numbers up to 1,000, per haploid genome. 13 But as I said, most of these sequences 14 were already present for a long time. Therefore were 15 targets for mutations. Most of these families then 16 became coding deficient or they do no longer encode 17 for retro R proteins. However, even if they are 18 coding deficient, many of these families are still 19 transcribed in several human tissues. Some have been 20 discolored just by virtue of their expression. 21 It also seems that the expression of these 22 sequences is regulated in certain tissues and tumors, 23 so we heard that there might be an deregulation of 24 families. It seems possible that that deregulation 25 mechanism is not present in certain tumor tissues. 68 1 Just a word regarding the nomenclature of 2 these sequences. The tRNA that was originally used in 3 the priming of the transcription process, the life 4 cycle of the exogenous vaporized, and according to the 5 amino acid and tRNA codes for, and this single code 6 for the amino acid stands dependent. This is just one 7 possible nomenclature of perts. It's still very 8 confusing. 9 I said that most retro viruses are coding, 10 HERVs are coding deficient. However, there are some 11 good described examples, especially some new examples 12 of coding in tact HERV sequences. At least there are 13 some in tact genes. We have already known for a long 14 time the so-called ERV-3 sequence that belongs to the 15 R-family. This, we agreed, pro-virus, or pro-virus 16 sequence encodes, and 1.9 KBN open reading frame. 17 That open reading frame is highly regulated to the 18 transformation of trophoplasts into sensitio- 19 trophoplasts in the placenta. So we have here clearly 20 an up-regulation during a developmental stage. 21 We have for instance, you have H-family 22 and we have about 1,000 copies of that H-family. 23 Among them are 100 copies that are still in tact 24 regarding the pro-virus structures. They have an LTR 25 gag pol env, LTR structure remaining 900 lack N gene. 69 1 There has also been reported that this HERV-H families 2 are expressed in various cell lines. We see the 3 highest expression for these elements has been 4 reported in cell lines that are derived from germ cell 5 tumors, and germ cell tumors I guess you will hear 6 some more about germ cell tumors later on. 7 Just this year, Lindeskog, Mark Lindeskog 8 reported the isolation of an intact HERV-H env gene. 9 So it is now clear that there is within the human 10 genome one intact HERV-H evn gene. It's not know so 11 far whether there are any among these many sequences, 12 whether there are any intact gag of pol sequences. 13 I would like to mention the new discovered 14 HERV-W family that has originally been reported, has 15 been isolated from retro virus by particles from 16 multiple sclerosis patients. It has also been 17 reported that these HERV-W sequences are up-regulated 18 in the placenta. Joni Blanc also reported this year 19 the isolation of an intact HERV-W in the genes. It is 20 also not known whether there are intact gag pol genes. 21 I would like to in the second part of my 22 talk, report about results for our family of clearly 23 outlines from our other HERV families in the coding 24 capacity. This is the so-called HERV-K HML-2 family. 25 This is quite complicated. 70 1 The human genome contains several families 2 that use lycine primer binding site or TRNA for primer 3 binding. They were named human MMTV-like sequences, 4 one through six. The family that we are talking about 5 is reported in more detail by Ono and co-workers and 6 the original sequence was the so-called HERV-K 10 7 sequence, which is by the new nomenclature is the 8 HERV-K HML-2 sequence. 9 We have reached about 25 to 50 copies of 10 that HERV family is present in old world monkeys, but 11 not in new world monkeys. One concludes that family 12 is present for at least 30 million years in the 13 genomes. 14 In the past, there have been reports of 15 isolation of the isolations of intact HERV HML-2 16 sequences. So there were reports about intact gag 17 sequence and intact protease sequence has been 18 reported, that is able to process that HERV-K gag 19 protein, intact pol sequences with RT activity, with 20 endonuclease activity, and have been reported and also 21 intact MRNA has been reported from the group from 22 Johannes Loewer. And also what we heard yesterday 23 evening, there is also an additional splicing product 24 from the N gene, the so-called C-ORF that still has a 25 rav-like function. 71 1 What is known already for a longer time is 2 that these particles or the cell lines are derived 3 from germs of tumors or typically testicular tumors of 4 the young man. These cell lines do produce with rav 5 particles. Boller and coworkers could show that these 6 particles are encoded by the HML-2 gag protein, 7 labeled antibodies, and recognized that gag protein. 8 If we look at patients suffering from germ 9 cell tumors, we also have some surprising results 10 regarding that HML-2 sequences. Namely, if we look at 11 the antibody status of these patients compared to 12 controls or other non-germ cell tumor types, we see 13 that mixed germ cell tumors and here especially, 14 seminomas, these patients have very high antibodies 15 directed against HERV-K gag and HERV-K N proteins. 16 These tumors or these antibody titres are already very 17 high if the tumor is clinically detected. From other 18 results, we also know that the precursors of these 19 tumors, the so-called carcinoma in situ, also 20 expresses already on the RNA level these HERV-K HML-2 21 sequences. 22 We were interested to see or to find out 23 where in the genome are these impact genes located 24 that are responsible or that cause finally the 25 production of these gag and env antibodies. As I 72 1 said, it has previously already been reported that 2 they are intact genes, but it was not possible because 3 of the high copy number of these sequences to isolate 4 or to at least chromosomally assign these intact 5 genes. We, therefore, tried to chromosomally assign 6 these intact sequences using a combination of the so- 7 called protein truncation test and using a 8 monochromosomal hybrid panels, or panel of human 9 rodent fusions, fusion cells. 10 We were able to show that there are at 11 least, still at least eight intact gag genes within 12 the human genome, and at least three intact env genes. 13 We did not publish that. There are also several 14 intact pol genes within the human genome. 15 I just want to show you how we got these 16 numbers. This is the protein truncation test that has 17 been described by Roest and coworkers in 1993. So it 18 was originally developed for the detection of APC gene 19 carriers that carry it, the APC gene. So the APC 20 lesion is characterized by trends or not completely 21 translated APC proteins. It is almost like the 3 22 prime terminus. 23 So we have three possibilities. One is 24 that both are intact, both genes are intact. The 25 carrier will carry one defective APC gene. The 73 1 defective person would carry the two defective genes. 2 The principle of the test is that the coding sequence 3 is PCI amplified, where the protochomo contains the T- 4 7 promoter and the translation initiation sequence. 5 So if this PCI product is then in vitro transcribed 6 and translated and impressed and radiolabeled amino 7 acid, electrophoresed, and then auto radiographed, you 8 will see according to the status of these donors that 9 you will have only full-length proteins, the carryover 10 also show an additional shortened protein and 11 defective people will only produce defective proteins. 12 We in principle used the same test because 13 we in principle have the same situation. We have some 14 defective gag genes within the genome. There must be 15 at least one gag or env gene because we have the 16 antibodies. So we put -- in principle used the same 17 test. 18 What we did was we are looking for the 19 presence of full-length gag genes or env genes on the 20 human chromosomes and then tested the PCI product we 21 got from the chromosomes for their coding capacity. 22 This is the result for the gag coding capacity. So 23 gag protein would result in a protein of about 73 24 kilodaltons. So these are controls that give the 25 respective proteins. 74 1 You see that there are several human 2 chromosomes that contain or produce a full-length 3 protein. There are eight human chromosomes that 4 contain at least one gag gene that contains four full- 5 length proteins. 6 I also would like to mention that we are 7 also able to demonstrate the defective gag genes if we 8 see here, these proteins that are just smaller than 9 expected. These are very likely the gag genes that 10 are defective. Stop codons within the coding 11 sequence. 12 We did the same for the HERV-K env genes. 13 We see here that three chromosomes produce a protein 14 of about 76 kilodaltons. These are the chromosomes 7, 15 19 in here on the chromosome. 16 What we also see in the gag experiments is 17 that there are additional env genes that are only on 18 the almost intact. We have here a protein that is 19 about four kilodaltons more. So this actually could 20 also be considered as an intact reading frame. 21 So we have several human chromosomes that 22 still contain gag and env genes. We have three 23 chromosomes that contain both intact gag and env 24 genes, the chromosome 7, 19, and the Y chromosome. We 25 were interested whether these chromosomes or the 75 1 intact genes on these chromosomes are derived or 2 located within one provirus or within several or 3 different positions within the particular chromosome. 4 I would like to report or tell you 5 something about what we found out for the chromosome 6 7. We were using for addressing that question, we 7 were using a chromosome-specific, chromosome-7 8 specific cosmid library. We were screening for clones 9 that contained both gag and env sequences. 10 What we finally found out, that we 11 isolated the so far least defective human endonuclease 12 on chromosome 7. We were able to characterize the 13 proviral sequence within one cosmid clone that still 14 has intact LTRs. So they regulate to the elements. 15 They are able to transcribe, as you will see. We have 16 an intact gag gene. We have an intact protease gene 17 that protease is able to cut itself from a gag 18 protease, polymer precursor protein, and is 19 furthermore able to process encoded gag proteins. So 20 it's typical retro-ized protease. 21 We know just from sequence comparison, one 22 can deduce that the endonuclease within the polymer 23 genes also acted just by sequence comparison, no 24 significant changes compared to recently described 25 active K in the nuclease. We have an intact env gene. 76 1 This intact env gene sequence has already been 2 described by Johannes Loewer's group as an MRA, which 3 also shows that this sequence is actively transcribed. 4 So this is actually an expressed provirus. 5 We have spliced on the inceptors sides the 6 corresponding position that would allow to splice an 7 M on A, and what we heard yesterday also, to splice an 8 additional soft M RNA. 9 What we see is that this proviral sequence 10 is only defective in the RT domain. It has a single- 11 based permutation within the YXDT motif. So very 12 likely, this highly important catalytic motif is -- so 13 only in reverse transcription function this probably 14 missing from that proviral. 15 Okay. We have here almost intact proviral 16 sequence. But now regarding infectivity, we had that 17 already several times I guess before. We have HERV- 18 encoded retro of particles, several cell lines, even 19 in tissues, the placenta tissue for instance. We find 20 HERV-RNA in these particles. 21 We have no infectivity so far shown for 22 any of these HERV sequences. We do not really know 23 why. There are several reasons that can be mentioned 24 for the HML-2 family. It has to be reported that the 25 env protein cannot be cleaved into the auto membrane 77 1 transmembrane domains. It is conceivable that they 2 are defective genomes that are packaged into these 3 particles, so only if they would be able to get a new 4 cell, they would only deliver defective genomes. 5 It is also not clear whether the 6 receptors, that they were once used by that, retro 7 families are still present and would still be used. 8 So what you should take home I guess is 9 that human endogenous are expressed in several tissues 10 tumor types that are highly up-regulated in certain 11 tumor types. Several HERV families are still able to 12 encode proteins, and among them, the HML-2 family that 13 still encodes all essential proteins. We have almost 14 intact HML-2 provirus within the human genome. Thank 15 you. 16 (Applause.) 17 CHAIRPERSON RABINOVICH: I think we will 18 hold questions at this point. We are going to take a 19 10-minute break now. We are going to come back and 20 finish with the last two speakers. 21 I need to figure out how to catch up time, 22 and yet leave the time for the panel discussions. I 23 ask you to do two things. Check-out time from the 24 hotel is 12:00. You should know that. They have 25 already called in a bunch of the taxis so that if you 78 1 need taxi arrangement, please let them know so they 2 can do that for you. Ten minutes we will start again. 3 (Whereupon, the foregoing matter went off 4 the record at 9:55 a.m. and went back on 5 the record at 10:10 a.m.) 6 CHAIRPERSON RABINOVICH: Is Dr. Broker 7 here? Great. 8 If you could take a seat please. The next 9 speaker is Dr. Thomas Broker from the University of 10 Alabama at Birmingham speaking on viral latency- 11 papilloma virus model. 12 DR. BROKER: Thank you very much. I would 13 like to deal with two subjects under this topic. The 14 first is a study of the prevalence of HPV in the 15 general population, and then following on Dr. Mayer's 16 pattern that you just heard, a study of some 17 endogenous sequences in papilloma virus transformed 18 cell lines with some surprising results. 19 We have done some inside 2 hybridization 20 studies of the expression of human papilloma viruses 21 in biopsies from women with HIV/AIDS who were 22 moderately immuno deficient. This is one example, but 23 fairly typical. 24 What you are seeing is a full thickness of 25 across the cervix. The various probes that we used 79 1 reveal the expression of one of the major early 2 transcripts of papilloma virus, the E4, E5. You are 3 seeing it here in bright field illumination and dark 4 field, matched pairs, basal layers right there. 5 As I indicated yesterday evening, 6 papilloma transcription is differentiation dependent 7 and occurs typically in the upper half of the skin. 8 E6, E7 messages, the delayed early oncogenes are hard 9 to see in bright field, but fairly easy to see in dark 10 field. You can see they follow a comparable 11 distribution. 12 The capsid component, L1 or L2, again, is 13 right at the very top of the last live layers of the 14 epithelium. Also to the point, the vegetative 15 amplification of viral DNA is in the upper half of the 16 epithelium. 17 Papilloma infections of the genital tract 18 in fact have been designated an official AIDS-defining 19 illness in the syndrome because of the significant 20 upregulation of HPV gene expression in women who have 21 AIDS or other immuno deficiencies. With that 22 knowledge in hand, and pictures like this, we 23 undertook the following study. 24 We decided to investigate the prevalence 25 of HPB in the population by focusing on immuno- 80 1 deficient groups. The three that we have chosen so 2 far are: women who are in enstay renal failure and in 3 need of a kidney, and most clearly ill; those then who 4 get a kidney and are pharmacologically immuno- 5 suppressed beyond their underlying illness; and those 6 with AIDS. 7 The strategy that we're using is an 8 extension of the techniques Steve Wolinsky and I 9 developed really 12 or 13 years ago when we first 10 proposed the use of degenerate primers for looking at 11 related genomes. The pair that's most commonly used 12 in the papilloma field is our original design called 13 MY911, but Louise Chao and I moved right next door. 14 We found that this region is a little too long to use 15 in form one fixed tissues, and this particular pair 16 has some wonderful restriction fragment polymorphisms 17 available that will allow us to do genotyping after 18 amplification. 19 So basically we start with the nested PCR 20 approach, outer primers and inner primers. The 21 starting material is cervico vaginal lavage, which 22 harvests cells from throughout the lower genital tract 23 of the women. We amplify and then we put it through 24 several different assays. Initially, agarose gel 25 electrophoresis to look for a 278-base amplimer. 81 1 Secondly, restriction fragment 2 polymorphisms which usually can tell us which genotype 3 is present. But if it's a pattern that we cannot 4 recognize, we will put it through sequencing. As you 5 are going to see, about half of the fragments that we 6 amplify we need to sequence. 7 The results of this study, I am going to 8 summarize. It's absolutely mind-boggling. Seventy 9 four percent of all women in the AIDS cohort have 10 clearly identifiable HPVs. We have managed to type 11 over 85 percent of these so far. Fifteen percent are 12 still under investigation. In more than half the 13 cases, the individuals yield multiple HPV types. 14 The study is longitudinal, and has been 15 going on for three-and-a-half years now. Many of the 16 members of the cohort have been sampled two up until 17 seven different times at six to 12 month intervals. 18 So that's our biggest cohort. 19 These are the renal transplant cohort. We 20 have statistically significant numbers. I would like 21 to point out that in instage renal failure, but no 22 pharmacologic suppression, about 59 percent of those 23 women have detectible HPV. Again, quite a few, a high 24 percentage have multiple infections. This carries 25 over to that portion of this group who go onto actual 82 1 transplantation. 2 Some of the remarkable outcomes of the 3 care with which we undertook the genotyping is the 4 following. In the pre-transplant population, the 5 prevalent types are those that are commonly seen in 6 the general population as causing disease, namely HPV- 7 6, 11, and 16. Those types persist in those women who 8 were pharmacologically immuno-suppressed. 9 We see a scattering of other types, but 10 the common types from prior studies are those that 11 predominate in the renal transplant cohorts. 12 In contrast, those women who are in 13 various stages of immuno-deficiency as a result of 14 AIDS, do not show the same genotype profiles. The 15 only member in common is in fact most common of all 16 genital HPVs, HPV-6. What we see instead are niche 17 homologs of the common types. For example, HPV-45, as 18 you are going to see, if a close relative of HPV-18, 19 which is often cited as a common virus. But we don't 20 see that in the AIDS cohort. 21 HPV-52 is our most common virus. It is a 22 close homolog of HPV-16, which we don't see amplifying 23 in this cohort. Most notably are the ones that I 24 indicated by stars, which are a very rare detection 25 within the general population, but in fact are most 83 1 common viruses in the AIDS cohorts. 2 In particular, we have identified 13 new 3 HPV types based on less than 75 percent sequence 4 homology to each other or to any other known papilloma 5 virus. They are all members of what has been 6 designated group A-3, which appear to be an AIDS- 7 defining subset of HPVs. 8 These can be at least considered in the 9 context of phylogenetic trees based on sequence 10 alignments in the L1 region. So, for example, HPV-16, 11 the main cause of cervical cancer in the world, is 12 seen in the renal cohort, but a very close relative, 13 52, is seen in AIDS. 14 Six and 11, that cause benign genital 15 warts and laryngeal papillomas are here. One of the 16 main groups coming up in AIDS is this group of cousins 17 of these guys. 18 HPV-70 is one of our most common types, as 19 well as 45. They are in the HPV-18 family, but 20 represent new members of this niche. The group I just 21 mentioned, A-3, that is so commonly seen in AIDS, 22 include our members jyn 2, 3, 4, all the way up to 13, 23 MM8 and 61, 72, and 83. That cluster seems to be an 24 AIDS defining group. The other ones that we have seen 25 abundantly are 51 and 53 in this arm. 84 1 Overall, in the Birmingham and generally 2 Alabama population, every virus types seen with the 3 star we have found one up to 23 times, indicating that 4 we have universal presence and also detectibility of 5 all of the known viruses within our immediate 6 population. 7 While this is up here, I also wanted to 8 point out the very large huge group of epidermal 9 dysplasia formus viruses that other labs have studied. 10 Again, it is a very rare group of illnesses, in fact, 11 only defined a few hundred times in all of medical 12 history in terms of individual patients. However, 13 there is this huge ramification of somewhat related, 14 but clearly distinct genotypes that comprise the 15 family or subgroup of viruses responsible for EV. 16 It is known that these patients all have 17 particular cell-mediated immune deficiencies. Again, 18 suggesting that particular arms of the immune system 19 are responsible for either containing or failing to 20 contain different subgroups of the papilloma viruses. 21 As we look at these women over a period of 22 time through these six month or so samples, what we 23 also find, and other labs have exactly the same 24 results, is every time we sample, you may or may not 25 see the type you saw before. It may switch. For 85 1 instance, we have this patient who had 6 plus 16, and 2 then 11 plus one that was minor and we couldn't tell, 3 then jyn 2, and then type 40, and then we had a type 4 53, but the others disappeared. 5 Everyone's experience in the field has 6 been that the viruses rise above a detectibility 7 threshold, stay there for a while, days or weeks or 8 months, and then fall below detectibility, only to be 9 replaced by a different HPV type. These are not new 10 infections. They are basically cryptic or latent 11 persistent infections that fluctuate in their levels 12 of replication and detectibility. Pretty much anybody 13 is showing that flexibility. 14 What I want to state at this moment before 15 showing the correlation with disease may sound 16 controversial, but I will stick by it. We have found 17 a brand new HPV type for every 10 people that we have 18 looked at. Philodelius and Ethel Michelle Diveres and 19 zur Hausen and Shamen in European study of tutanius 20 papilloma viruses have found a new papilloma virus for 21 just about every other person they have looked at when 22 they use the combination of nested PCR and DNA 23 sequencing. 24 Robbie Burke's group, Jill Polefski's 25 group, have very comparable experiences looking at 86 1 anal papillomas or female genital tract. 2 It is my contention right now that instead 3 of 80 HPV genotypes or 150 that have been officially 4 named, that there probably are millions of variants, 5 virtually a continuum. We feel that basically 6 everybody has their own personal micro flora, that 7 these are passively acquired or vertically acquired, 8 not necessarily sexually, but certainly possibly 9 sexually, and that they simply are part of the human 10 condition as are microflora, just as we have 11 microflora composed of bacteria and many other 12 viruses, and that they basically are utterly 13 ubiquitous. I will come back to that point in a 14 moment. 15 We did try to correlate the various other 16 medical parameters in these cohorts, especially the 17 AIDS cohort, with CD4 count, HIV virus load, other 18 infecting known STDs like herpes, chlamydia, 19 trichomonas, so forth. The one correlate that held up 20 and not surprisingly at all, was that the degree of 21 pap smear abnormality from normal, abnormal cells of 22 unknown significance, low grade dysplagias or high 23 grade dysplagias, is with CD4 count. 24 The medians, these are all the people who 25 had multiple infections, a high risk virus type, a low 87 1 risk, no virus at all, and had either normal or these 2 various abnormal pap smears. These bars here are the 3 median CD4 count in each of these groups. 4 The one place where we saw active disease, 5 low and high grade dysplagia, these by median, is when 6 people fell below the CD4 count of 200 cells per cubic 7 millimeter. 8 In summary of that data, we found that 9 it's very very possible to have negative pap smears, 10 but definitely have HPV infections. We feel these are 11 people who have not yet reactivated long enough to 12 have resulted in cytologic change as a result of 13 infection. We have on the other hand, the people with 14 overt disease by biopsy or by cytology, and the higher 15 the grade lesion, the more likely it is to see either 16 single infection or especially multiple virus types 17 present within that patient at that time. 18 So the more that we can detect the virus, 19 that is, the more it has replicated or amplified 20 throughout the population, the more cells that are 21 shedding the virus in effect, the more likely we see 22 disease. 23 So to summarize this part of the talk, I 24 feel that they are virtually ubiquitous. they are 25 typically sub-clinical, persist in or latent 88 1 infections. There are staggeringly large number of 2 genotypes if we take the care to look. I might say 3 that the reason these are typically not found is that 4 people use generic cross-hybridizing probes or have 5 cut off their probe sets. If you're not probing for 6 something, you are not going to see it. 7 Most of the viruses in this number 60, 70, 8 80 and above, are not even present within the 9 commercial probe sets. So if you aren't probing, you 10 are not going to see them, and you are going to get 11 lower numbers. 12 They can be found throughout the genital 13 tract in 60 to 75 percent of the people that we have 14 looked at who are admittedly good yielders, because 15 they are immuno-compromised, but I think this simply 16 represents the general infection in the population. 17 They can be found in oral and esophageal 18 mucosa. Utaneious types persist in hair follicles. 19 There's a wonderful study from Amsterdam by Tershaget 20 and Ingebor Boxman. She plucked hair follicles, both 21 eyebrow hairs and pubic hair, and 60 to 70 percent of 22 all people harbored EV viruses or other rare virus 23 types in their hair follicles. No disease, it's just 24 part of the human condition. 25 I believe they are vertically transmitted 89 1 perinatally, mother to baby. Some of them are clearly 2 pre-natal infections. As we know, there's long-term 3 maintenance that requires viral replication in concert 4 with host replication in the cell cycle. 5 So what I would like to do now is tell you 6 a little bit about a very unexpected observation we 7 made in Hela cells. This goes back to last night's 8 talk regarding the structure of the replication 9 complex of HPVs. 10 As you know, cyclin E is one of the key 11 checkpoints or entries into S phase. Ectopic 12 expression of cyclin E can speed up entry into S 13 phase, and it can even bypass the need for some of the 14 RB phosphorylation by cyclin D. It's simply one of 15 the key steps that needs to follow the induction of 16 the DNA replication enzymes. 17 HPV E7, the viral oncogene that in fact 18 binds RB and can help bypass that step, among the E2F 19 enhenca protein regulated genes is cyclin E itself. 20 In other words, HPV infection upregulates cyclin E. 21 So we asked whether the induction of 22 cyclin E is essential for the reactivation of 23 unscheduled cellular DNA synthesis in the upper 24 stratum of squamous cepathelium that differentiated 25 caratinocyde. I'll just summarize that data. 90 1 I got you to the point last night where 2 the E1 diheximer, the double helicase held together by 3 the HSP-70 cochaperone protein, is there. 4 The next thing that loads in the study we 5 did with Theresa Wong at Stanford, is the recruitment 6 of the cellular DNA polymerase, and showed direct 7 interactions between the helicase and the catalytic 8 sub-unit of pol alpha, P-180, as well as its P-70 sub- 9 unit. This was the first indication of what P-70 does 10 in the four sub-unit complex of pol alpha, which 11 includes two primary sub-units. The answer is, it 12 brings the polymerase to the ora itself. 13 The next thing that comes in is cyclin E, 14 CDK-2 complex, that critical S phase entry point. AS 15 a result of that, what happens is upon cyclin E 16 finding an appropriately assembled pre-initiation 17 complex, five target proteins are phosphorylated. 18 They include: the E2 protein, which appears to be 19 displaced by that event; in addition, P-70 helps 20 displace E2. So the loading of this and the 21 phosphorylation kicks this guy out. 22 Secondly, E1 is phosphorylated. These two 23 subunits of preliminary salpha that bind directly to 24 E1 are phosphorylated. When all four of those have 25 been successfully modified, the kinase phosphorylates 91 1 cyclin E itself, which is displaced and degraded by 2 ubiquitination. That enables the pre-initiation 3 complex to convert to the elongation complex. 4 In studies with Wade Harper and Jien-Ling 5 Ma at Baylor, two things were done. The first is 6 together we found that there's a cyclin binding motif 7 that the amino terminal have at the E1 protein, which 8 in fact is shared with a number of other things that 9 bind the cyclin E. That motif involves an RXL. That 10 is, an arginine something leucine motif right there. 11 In addition, their candidate 12 phosphorylation cites, the series of serine, serine, 13 serine, and threonine, mutation of any of these, the 14 motif or any of the target phosphorylation cites, 15 diminishes the capacity of cyclin E to convert the 16 pre-initiation complex to an initiation complex. So 17 the functional requirement for phosphorylation has 18 been verified. But keep in mind this location. We'll 19 come back to it in a second. 20 So we assumed that the consequence of 21 upregulation of cyclin E by E7 gene expression would 22 identify those cells that are capable of supporting 23 papilloma replication. To our amazement, we found the 24 opposition. This is our epithelial raft model. We 25 have done the same in natural papilloma lesions. Here 92 1 we monitored cyclin E expression, over expression in 2 the tissue. Here's bromo deoxy uridine incorporation 3 or PCNA upregulation. These are the match. 4 What we found is the cells that had high 5 cyclin E could not replicate. In fact, they are 6 mutually exclusive with those capable of supporting 7 DNA synthesis. Conversely, PCNA, which is upregulated 8 by papilloma 7 and cyclin E do co-localize. But we 9 see a number of cells where PCNA is present and there 10 is no cyclin E. So we have a reciprocal pattern to 11 what we expected. 12 I am just going to very briefly tell you 13 that P-21 cip, one of the inhibitors of cyclin D and 14 cyclin E, is also upregulated by E-7 expression in 15 natural condylomas or in our E-7 expression raft 16 cultures. You can see those signals in the upper 17 strata again. So we have P-21 upregulation, again, in 18 a subset of cells. 19 When we look in rafts or in natural 20 papillomas, we see that those cells that have high P- 21 21 are mutually exclusive from those capable of 22 supporting either viral or cellular DNA synthesis. 23 When we did the third pairwise combination and looked 24 at cyclin E and P-21, we found perfect colocalization 25 of those two. 93 1 So ironically, the cells that have high 2 cyclin E also have high P-21 and do not support 3 replication. This was really perplexing, except we 4 did know this inhibited that. But we assumed cyclin 5 E was in the licensing factor for engaging in 6 replication. 7 So what we came to feel is the following 8 model: that in the course of unscheduled DNA synthesis 9 reactivation, if cyclin E appeared in the appropriate 10 timing or sequence or amount, once a pre-initiation 11 complex formed, you would successfully phosphorylate 12 the target proteins, polymorases and E-1 and E-2 13 proteins, and successfully engage in elongation 14 replication. 15 Conversely, if too much cyclin E appeared 16 and it appeared in an untimely fashion, its inhibitor, 17 P-21, would recognize misassembled complex. They 18 would cross stabilize. They would both pile up to 19 high levels, and those would be defective in engaging 20 in elongation. 21 Now we put this all together by asking how 22 does this play into the establishment of immortalized 23 and transformed cells and cancers. What Wade Harper 24 had found is that when he did pull-down assays with 25 cyclin E to ask in hela cells what binds to cyclin E, 94 1 almost all the things that came down in the assay was 2 that E-1 protein from the resident HPB-18 in the cell 3 lines. Up until that point, people had thought the E- 4 1 gene was deleted from hela. In fact, it's present. 5 The entire length of the E-1 gene is still present in 6 hela. In fact, is expressed. 7 Now the functional assay that our lab did 8 was that we found that hela cell extracts could not 9 support HPB replication in our cell-free system, that 10 there was a missing factor in hela that the extract 11 needed. We could put 293 cell extracts or any other 12 cell line that we could find, they would easily 13 complement papilloma replication in vitro. But 14 anything from hela cHa caski or any other HPB 15 transformed cell line could not support it. 16 The upshot of the whole thing is that 17 every papilloma transformed cell that we studied 18 expressed a full length E-1 transcript, but in all 19 cases, the transcript had an either frame shift or a 20 stop code on partway through the gene or miss sensed 21 mutations in this vicinity, so that in tact E-1 could 22 not be made. But in all cases, it made the RXL 23 portion that interacts with cyclin E. 24 So we added a little cyclin E back to hela 25 cell extracts and immediately restored full 95 1 complementability to those extracts, establishing that 2 the missing thing in hela cells was cyclin E. 3 So our conclusion is the following. I 4 think we feel very confident about it. In the process 5 of immortalizing cells, either natural cancers or 6 attempts at making cell lines, substrates in effect, 7 it's good to have all the upregulation of DNA 8 polymerase, topasomerases, PCNA and so forth, that 9 help rapid cell cyclin. But the one thing that E-7 10 upregulates you don't want, is cyclin E. 11 So what these natural experiments did, is 12 figured out a way to sequester part, but not all, of 13 the cyclin E by putting in retaining fragments of E-1, 14 capable of mopping up that one product that's 15 upregulated that you don't want to have. That is, 16 cyclin E. 17 Brian Van Tine, last night, also indicated 18 there's evidence of some antisense in other papilloma 19 lines like cHa caski, which would again, modulate the 20 amount of E-1 that you could translate from the 21 messages that are clearly there. 22 Together we believe that to establish 23 these cell lines, whether you make them in the lab or 24 whether nature has made them for you through cancer, 25 you need to diminish the amount of cyclin E to achieve 96 1 an equilibrium where you have enough cyclin E to 2 support cellular replication, but not too much so that 3 it's an unsuccessful high level. Thank you very much. 4 (Applause.) 5 Oh, I did want to acknowledge -- could I 6 have the slides for one second -- a very, very large 7 number of collaborators. I'll leave them up here. 8 But we're very grateful for our own students and 9 collaborators at UAB, Jeff Engler, Doug Seer, Sean Van 10 Tine, Kim Towns. At Baylor, Wade Harper's lab, UNC, 11 Jack Griffith's lab, who did the things with the HSP- 12 40, and at Stanford, Theresa Wong, and our 13 collaborators at the Free University, who did a lot of 14 work on some transcriptional control that we 15 collaborated on for several years. Thanks. 16 CHAIRPERSON RABINOVICH: Thank you. Are 17 there any questions? 18 DR. FRIED: How could you be so sure that 19 with the evolution of the HPVs, are not due to new 20 infections, but to pre-existing sequences to cover? 21 DR. BROKER: I think one is the repetity - 22 - 23 DR. FRIED: Mike Fried. 24 DR. BROKER: With which these viruses are 25 appearing, especially in women who at least declare 97 1 they are not engaged in much, if any, sex. A number 2 of the epidemiologic studies have traced either the 3 frequency of recent sexual activity. Everyone who has 4 looked feels its an emergence of pre-existing 5 subclinical infections. There is some evidence in new 6 infection, but an awful lot of it simply appears. 7 Women in that stage renal failure, for 8 example, who are 38 years old and married basically, 9 are not suddenly acquiring new infections. 10 CHAIRPERSON RABINOVICH: Back microphone, 11 please. Why don't you come up to the front. 12 DR. RUSSO: Hi. Carlo Russo, from Merck. 13 Very interesting talk. 14 I have a couple of questions. One is, how 15 can you be sure that what you are sampling by PCR is 16 really an infective virus, it's not just a transient 17 presence with a virus due to the fact that you can't 18 control sexual behavior, and perhaps the woman has 19 just been exposed to a virus? 20 DR. BROKER: Well, I think one way is that 21 in a very high percentage of the people, there are 22 clear dysplasia, low and high grade. All the women 23 who have any degree of dysplasia are also biopsied, 24 and the inside 2 hybridization, as you can see, is 25 showing clear effects in the tissue. 98 1 So I don't think there's surface 2 adventitious contaminants. These are within, at least 3 a fair number, if not all, are within the cells, and 4 causing various degrees of actual overt illness. 5 DR. RUSSO: I may have missed the data. 6 Did you show the types that are associated with high 7 grade lesions, the HPV types? I didn't see on the 8 table. 9 DR. BROKER: Yes. Well basically, in this 10 immense spectrum of what's now 37 different viruses 11 that we found, those that are most typically 12 associated with low and high grade dysplasia, the 13 actual diseases, are the higher risk types. 14 DR. RUSSO: So you are not suggesting that 15 if you want to prevent cervical cancer, we should 16 focus on different types of the one already 17 identified? 18 DR. BROKER: Well the real problematic 19 thing for any clinical management, either vaccination 20 programs or small molecule drugs, is this absolutely 21 exploding number of virus types. 22 The one thing that I think is going to -- 23 and I commented a day or two ago that in the U.S. 24 alone today, there are over 250 to 300,000 people 25 immuno-suppressed just due to organ transplants, 99 1 steroid use, or bone marrow transplants or AIDS. So 2 there is an immense reservoir of particularly high 3 risk patients. 4 Nonetheless, most of the diseases are 5 still being caused by a handful of viruses like 16, 6 18, 52. So I think, at least the ones we have to 7 worry about today, are still manageable in number. 8 AUDIENCE MEMBER: I would like to ask a 9 little bit about cell substrates. Considering that 10 hela has multiple HPV integrants, I guess, are any of 11 those infectious? 12 DR. BROKER: No. 13 AUDIENCE MEMBER: Can you get them back 14 and make them infectious? 15 DR. BROKER: No. 16 AUDIENCE MEMBER: What are they missing? 17 DR. BROKER: They are all truncated within 18 E-2, at best. Although there are 30 to 50 copies, 19 depending on the hela sub-1. All the integrated 20 copies are truncated. 21 AUDIENCE MEMBER: Would that DNA be 22 transforming, even not infectious? 23 DR. BROKER: They do contain the E-6 and 24 E-7 genes. Expression of those genes, as shown by 25 studies primarily in zur Hausen's lab, must be 100 1 maintained or you no longer can cycle hela cells. 2 That is antisense to E-6 and E-7 in hela makes them 3 not cycle any more. So the driving force of hela is 4 the overt expression of E-7. 5 AUDIENCE MEMBER: Would hela cells qualify 6 as an example of a cell that should never be used to 7 make a vaccine? 8 DR. BROKER: I don't know that I would go 9 that far. It's obviously wonderful as a producer of 10 all sorts of biomedical products. 11 DR. MURPHY: I actually wanted to ask you 12 the same question, but I would phrase it in a 13 different way. 14 (Laughter.) 15 DR. MURPHY: This is Murphy from NIAID. 16 That is, do you see any reason why, you know, having 17 an intimate knowledge of hela cells and human 18 papilloma viruses, that hela cells should not be used 19 as a substrate for making live attenuated virus 20 vaccines? 21 DR. BROKER: I don't know of any evidence 22 of these genes being transduced out or in any way 23 posing a risk. I was going to save it for the panel, 24 but it occurred to me last night, I had challenged one 25 of last night's speakers about the use of psorilins as 101 1 a cross-linking agent. It gives me great concern that 2 it is a known carcinogen. However, Brian and I were 3 talking. Brian van Tine and I were talking last 4 night, and he reminded me that there are biotinylated 5 psorilins. 6 So for all the debates regarding how to 7 remove contaminating DNA, one strategy in principle is 8 throw in a biotinylated psorilin, cross link it, and 9 pass the whole thing over avidin magnetic beads or 10 batch subtraction of the DNA. 11 So in fact, that strategy may actually 12 help you deplete adventitious contaminants very, very 13 readily. So it is an alternative at least. 14 AUDIENCE MEMBER: Can I just ask you about 15 hela again? We learned last night that not every HPV, 16 if there's 30 to 50 copies, are not all active, I mean 17 in caski only one was active. 18 DR. BROKER: Yes. 19 AUDIENCE MEMBER: What is the state of 20 hela? Are they -- 21 DR. BROKER: Very, very few are active. 22 We and Wade Harper's lab are both sequencing all the 23 transcripts. This actually was done by Elizabeth 24 Schwartz and others in zur Hausen's lab in 1985, and 25 a variety of groups since then in Japan and elsewhere 102 1 have looked at the expression loci in copies in hela. 2 There appear to be three or four different 3 transcripts made from different positional integrants, 4 but the majority are silent. A few of them are 5 active. But so far, all the ones that are active have 6 truncated E-1s. They have the cyclin E binding motif, 7 but they don't have their normal carboxy terminus. 8 CHAIRPERSON RABINOVICH: Thank you very 9 much. 10 Let's go on to Dr. Cashman. Thank you for 11 being so patient. Transmissible spongiform 12 encephalopathies: vaccine issues. 13 DR. CASHMAN: It worked. My friendly A/V 14 guy explained how to do this. 15 I am Neil Cashman. I am predominantly at 16 the Center for Research and Neuro Degenerative 17 Diseases at the University of Toronto. I have a 18 special interest and a long-term research effort in 19 the expression and function of the normal cellular 20 isoform of the prion protein. 21 I am also obliged to mention that I am the 22 chief scientific officer one day a week of a little 23 biotechnology company in Montreal called Caprion. 24 I want to spend a few minutes talking 25 about prions and prion disease. We had a speaker 103 1 yesterday who said "and now for something completely 2 different." Well, how does a genomeless infectious 3 agent grab you? 4 Creuzfeldt-Jakob disease is the most 5 common human prion disease that we run into. I do 6 want to spend a few minutes talking about this so that 7 we are all on the same page with regard to public 8 health risks. Creuzfeldt-Jakob disease or CJD is 9 basically a disease you wouldn't want to wish on your 10 worst enemy. It is a completely untreatable uniformly 11 fatal disease resulting in death within six to nine 12 months of presentation. Survival over a year is 13 recorded, but it is not very frequent. 14 The presentation is usually that of a kind 15 of Alzheimers-like syndrome, with problems in memory 16 and intellectual function, but it can also present as 17 a disorder of gait and balance as well. Most people 18 have mild clonus, which is twitching of the muscles, 19 sufficiently forceful to move a joint. Other features 20 of the neuro-degenerative syndrome are reminiscent of 21 other neuro-degenerative diseases like Lou Gehrig's 22 disease and Parkinson's disease. Basically it's like 23 having every neuro-degenerative disease at once, 24 telescoped into an unmercifully short period of 25 decline. 104 1 Fortunately, it is rare. Sporadic 2 Creuzfeldt-Jakob disease occurs at about one per 3 million population per year. Also, somewhat 4 fortunately, it's not a disease of children. The 5 average incidence of CJD is in the 60s. 6 There are three recognized forms of CJD. 7 The most common being sporadic. This is a spontaneous 8 onset of CJD in an individual for which we have no 9 clue why they have developed it. There are familial 10 variants, which seems to be passed as an autosomal 11 dominant in families. That constitutes about 15 12 percent of the cases of human prion disease that 13 occur. There are iatrogenic prion diseases, which are 14 caused essentially by treatments and surgeries, well- 15 meaning, but nonetheless transmitting the disease. 16 Of course the transmissible spongiform 17 encephalopathy that even my kids know is bovine 18 spongiform encephalopathy, or so-called mad cow 19 disease. Since the early 1980s, this disease has 20 affected about 200,000 cattle in the U.K. and Republic 21 of Ireland, and a few hundred across continental 22 Europe. About 2 million cattle have been killed in an 23 attempt to stem the epidemic. This culling, as well 24 as change in policies, such as feeding ruminant to 25 ruminant -- we turned cattle into neo-cannibals -- is 105 1 resulting in a rapid decline of new cases, predictions 2 being that the epidemic in cattle may be essentially 3 stamped out in the early part of the new millennium. 4 I won't even say the new century. I'll say the new 5 millennium. 6 Unfortunately, this disease is unique, 7 unlike every other known naturally occurring prion 8 disease. It doesn't seem to obey species barriers, or 9 at least obeys them to a much lesser degree. There is 10 an outbreak of feline spongiform encephalopathy in 11 house cats. There is spongiform encephalopathy in zoo 12 animals, including primates. The primates that we are 13 most concerned about are also vulnerable to this 14 disease. 15 To date, 44 people have developed a new 16 variant of Creuzfeldt-Jakob disease, which is 17 clinically and pathologically distinct from classical 18 CJD. The statisticians predict there will be 19 somewhere between a few hundred and maybe 80,000 20 cases. This does not include the chicken little 21 predictions of the extent of the epidemic. 22 The disease unfortunately seems to strike 23 the young. There have been teenagers involved. It is 24 a relatively slower progression than classical CJD. 25 There are clinical features that are distinctive, but 106 1 I won't bore you with them this talk. The pathology 2 is also absolutely distinct, including a preter 3 natural accumulation of PrP Sc, which is this abnormal 4 amyloid protein that's been linked to infectivity. 5 This occurs both in the brain and in peripheral 6 lymphoid tissues. 7 Well, before we leave the clinical stuff 8 about CJD and prion disease, I want to kind of set the 9 stage with a sobering statistic, which is there is 10 iatrogenic transmission of this group of diseases. 11 Considering the penetrants and the young age of 12 vaccinees, this is a scary possibility. This would 13 dwarf every other iatrogenic transmission known to 14 date. 15 In humans, basically a few hundred cases 16 have been attributed to iatrogenic transmission, from 17 hormones extracted from cadaver pituitaries, from dura 18 mater transplantation, which is the tough lining of 19 the brain. But incredibly, the largest iatrogenic 20 transmission known to date, also the first documented, 21 was that in passage with a vaccine, which was a 22 vaccine for looping ill of sheep. Formal and 23 inactivated brain preparations passed sheep scrapie to 24 about 1,000 sheep. So hopefully this will not be a 25 pattern with human vaccines. 107 1 Well here is the prion hypothesis. This 2 has gone from being an object of ridicule to the 3 middle of the road interpretation of prion 4 infectivity. It has been sanctioned by the Nobel 5 Prize committee, garnering the prize for Stanley 6 Prussiner, the investigator whose ferocious work with 7 this group of disorders and with this agent has given 8 him, in my opinion, a well-deserved Nobel Prize. 9 The basic tenants of the prion hypothesis 10 are that there's a normal cellular protein, which is 11 called PrPC, which has been cloned and recognized. 12 It's expressed by just about every organism down to 13 drosophila. It is a very old gene. It's incredibly 14 well conserved in evolution. It is predominantly 15 alpha helical in secondary structure. 16 Now this normal cellular protein can adopt 17 an alternate confirmation, which is rich in beta sheet 18 structure. When this protein is in this alternate 19 confirmation, it acquires many unique physical 20 chemical properties. It becomes partially protease 21 resistant. It tends to aggregate. It's very poorly 22 soluble. Plus, it then seems to act as a catalyst for 23 recruiting more confirmational copies of itself. 24 Now whether this occurs by a kind of 25 enzymatic confirmase activity or whether this is kind 108 1 of a biological crystallization phenomenon is being 2 actively investigated. But it is clear that this 3 abnormal confirmation isoform of the protein, called 4 PrP Sc for scrapie, is capable of recruiting more 5 confirmational copies of itself from the normal 6 cellular isoform. 7 So onto vaccines. There are some concerns 8 about vaccines. I will mention three areas that need 9 to be considered. I will dwell most of the time on 10 cell substrates, which is nice of me considering this 11 is a cell substrate meeting. I will also talk briefly 12 about potential prion infectivity coming over in media 13 supplements for those cells, and in excipients, which 14 are compounds used to stabilize vaccines in their 15 final formulation. 16 In this case, luck seems to be at least 17 partially on our side, because it's not easy to infect 18 cells in vitro. It is possible to infect primary 19 neuro cultures. It is even possible to infect 20 neuronal cell lines neural blastoma. But there is not 21 much infectivity, and basically each successful 22 infection of a cell line is worth a publication or 10. 23 This may be due to the fact that cell 24 lines have very little PrPC, which is the precursor 25 for PrP Sc. The conversion of the protein from PrPC 109 1 to PrP Sc forms occurs at the surface or a post- 2 surface compartment. So in general, cell surface 3 abundance of the protein correlates with infectivity. 4 Most cell lines in my own laboratory, including hela, 5 express no more than one-tenth of the amount of cell 6 surface PrPC that a primary neuron does. 7 It has also been thought that the very act 8 of cell division itself can kind of sterilize a 9 culture because cell division can out pace the 10 relatively slow conversion and processing of PrP Sc. 11 So if you have a couple units of infectivity, they get 12 progressively diluted by having huge numbers of cells 13 that bear no infectivity. Finally, those cells may 14 die, the infected cells. 15 There is also a poorly quanitifiable role 16 for cell biology, which I put in quotation marks here. 17 Things that we really cannot quantify at this point, 18 like proper trafficking, post-translational 19 modifications of PrPC that are important in 20 conversion, and even sub-cellular distribution. The 21 protein seems to accumulate in this glycosal 22 phosphatindinol rafs at the cell surface. Some cell 23 lines don't seem to support these sort of rafs. 24 Now to make some kind of estimate about 25 the spontaneous development of prion infectivity in a 110 1 cell culture, especially a vaccine cell culture that 2 may have hundreds of trillions of cells, I am going to 3 back up and try to explore some assumptions about the 4 spontaneous development of prion disease in humans, 5 which is the species for which we have the best 6 numbers. 7 According to the prion hypothesis, an 8 occasional accidental mis-folding of PrPC to PrP Sc is 9 what triggers the recruitment process which proceeds 10 on an exponential basis. Each molecule that's 11 converted converts to more, da, da, da, da, da, da, 12 which happens on a post-translational level. No 13 genome involved. "Look, ma, no genome." 14 But sporadic disease in humans is 15 incredibly rare, one per million people per year. 16 Humans have something on the order of 100 billion 17 neurons. So one can make the kind of interesting 18 calculation that a productive infection arising from 19 a single neuron, you need about 10 to the 17th 20 neurons, 100 million neurons across a million people 21 in order to develop spontaneous scrapie. 22 But of course that's not the only way one 23 could potentially develop spontaneous CJD. Discovered 24 by familial CJD and familial prion diseases, there are 25 mutations within the open reading frame of the prion 111 1 protein that apparently predispose to this accidental 2 misfolding, such that somebody with a prion protein 3 mutation that actually results in an amino acid 4 substitution is basically guaranteed of developing the 5 disease over the course of a lifetime if he or she 6 lives long enough. 7 So could this occur in vitro? Could 8 certain cells in vitro acquire a somatic mutation 9 which is then propagated to infect an entire culture, 10 again, on a post-translational level? Well, let's run 11 some numbers on this one. 12 The mutation rate in man is about one per 13 billion basepairs per year. I thank Dr. Kazazian for 14 yesterday for pointing me to this reference. Thank 15 you very much. 16 The prion protein open reading frame is 17 really less than 1,000 basepairs. It's a relatively 18 small protein that's all contained in one exon. This 19 gives rise to a kind of pseudo calculation that a cell 20 can develop a mutant prion protein gene, a cell in 21 vivo, and a human can develop a mutant prion protein 22 gene in about one out of a million cells per year, if 23 you take one out of a billion and multiply it by 24 1,000, that's one out of a million. 25 This, just as an aside, this gives a rise 112 1 to a kind of startling calculation that all of us in 2 the audience are generating about 100,000 prion 3 mutants in our brain per year. 4 However, there must be a safety factor 5 here because the rate of prion disease arising from 6 somatic mutation cannot exceed the calculation of 7 prion disease arising from individual neurons that we 8 just went through in the last slide, which is 10 to 9 the 17th neurons per year. 10 So incredibly, somatic mutation is a 11 seriously flawed non-efficient process for producing 12 prion disease. Something on the order of one out of 13 100 billion mutations are productive of infection. 14 This may give us some comfort when we turn to the in 15 vitro scenario. 16 So let's talk about cell substrates. Is 17 it possible that spontaneous prion infectivity could 18 arise de novo in a culture? I told you that substrate 19 cells have usually less than one-tenth PrPC than 20 neurons. So if we run these calculations, by 21 misfolding one substrate cell per 10 to the 18th 22 years, and I think that is a quadrillion or something 23 like that, it's certainly comfortably larger than the 24 projected age of the universe. 25 Substrate cells, however, are less 113 1 genomically stable than primary neurons. If one says 2 that there's say 1,000-fold greater rate instead of 3 one out of a billion basepairs, one out of a million 4 basepairs can be mutated per year in a substrate cell, 5 this gives rise to a calculation suggesting that you 6 need 10 to the 14th substrate cells per year in order 7 to have one productive, i.e. spontaneous infectivity 8 arising in a culture. 9 Now this number looks incredibly large, 10 but when we think about the numbers we heard last 11 night about the production of polio virus from vero 12 cells, according to my calculations, 300 million cells 13 are used per year to generate all those vaccine lots. 14 Ten to the 14th is only 100 -- sorry, 300 trillion, 15 and 10 to the 14th is only 100 trillion. Did I get 16 that right? Please forgive me and correct me if I 17 didn't get it right. 18 So perhaps it is possible, considering the 19 enormous scale of substrate cell culture, that prion 20 infectivity could arise through somatic mutation in a 21 substrate cell, and could contaminate a vaccine 22 destined for human beings. 23 Well, there are some things to talk about 24 with this model. If this is so, how come we haven't 25 seen any vaccine transmissions yet? One of the 114 1 factors is that very few human cell line vaccines have 2 gone into humans, certainly not on a scale of vero 3 cells or primary cells that are used for culture. 4 Another unsettling thing is that if indeed 5 there is somatic mutation in a culture of human 6 substrate cells, would we ever detect it? We are 7 talking about something that would occur in one out of 8 a million cells or even one out of a thousand cells, 9 would ever be able to find by PCR or SSEP or anything 10 you could think of, a mutation at this level. 11 So aside from substrate cells, I did want 12 to touch upon a few potential sources of infectivity. 13 The media coming in contact with substrate cells are 14 potentially carrying prion infectivity. Bovine serum, 15 fetal calf serum, and newborn calf serum is used as a 16 supplement for proteins and growth factors and 17 hormones for most cell lines. Some cell lines are 18 also supported by human serum albumin. I am not aware 19 of a vaccine cell line that's supported in this 20 manner, but many recombinant proteins are supported 21 with human serum albumin. 22 There's also potential prion infectivity 23 in excipients, this last compound that's added to the 24 preparation to keep it stable before use. Many 25 childhood vaccines are stabilized with pig skin 115 1 gelatin. Pigs don't seem to be a species which 2 spontaneously develop prion disease or a species that 3 can catch prion disease via the oral route, although 4 deliberate intracranial injection of BSE infectivity 5 can produce a prion disorder. 6 Human serum albumin is also an excipient 7 in measles, mumps, rubella, and rabies vaccines. I 8 would like to spend just a few seconds talking about 9 human plasma proteins before I close, and give my 10 final advice, such as it is. 11 Human serum albumin of course comes from 12 humans. Of course it's a plasma fractionation 13 product. There has been a great deal of work trying 14 to identify potential risks of transmission of CJD 15 from human to human through blood or blood products. 16 Suffice it to say that population studies, case 17 control studies, and cohort studies have proven 18 universally negative. There is no documented 19 incidence of human CJD, classical CJD being passed 20 through blood or blood products. 21 There are of course case reports of people 22 getting a transfusion and developing CJD, but one 23 should not expect that transfusion or administration 24 of a blood product is protective against CJD. The 25 incidence of CJD in the transfused or treated 116 1 population is the same as in the non-transfused or 2 treated population. 3 However, we're in more difficult territory 4 with variant CJD. This is, I told you, an unusual 5 agent. It seems to cross species lines with impunity. 6 There are other features that are quite scary with 7 regard to human blood and human plasma products, 8 including albumin that might be used as an excipient. 9 The agent starts of course in the 10 periphery through oral exposure, suggesting a 11 prionemia. There is a huge accumulation of PrP Sc, 12 our only surrogate for infectivity, our only 13 biochemical surrogate for infectivity, in not only 14 brain, but in lymphoid tissue. 15 The agent itself has odd properties. It 16 is stable across species. It may in fact be 17 specialized or selected. I realize that these terms 18 are not often applied to a protein. I'm borrowing 19 terms from agents that contain a genome. It suggests 20 that this agent may be more virulent, especially with 21 regard to peripheral exposure. In other words, one 22 unit of classical CJD infectivity will not cause 23 disease when injected intramuscularly. One unit of 24 variant CJD infectivity may very well. There is no 25 data. 117 1 Another thing that has been noted is the 2 young age of onset of variant CJD. This has been 3 attributed to kids eating hamburgers and all kinds of 4 weird meat products. But it could also be attributed 5 to host factors which would promote infectivity in the 6 young. Since vaccinees are usually young, we have to 7 take this in mind. 8 But our greatest risk factor here is that 9 this is a new disease and we have no data. So how do 10 we minimize the risks from vaccine transmission? 11 Basically there's three ways that I can think of. I'm 12 sure that other people can think of more. It is 13 important to add prion validation to the list of 14 agents and microbes which are tested for in vaccine 15 lots. 16 This could be done two ways. The 17 biochemical marker of infectivity is PrP Sc. This 18 technology is in evolution. It appears that capillary 19 electrophoresis, some types of optimized immuno 20 blotting, and even ELISAs are reaching the point where 21 one unit of infectivity will be detectible. 22 Another important way of assay for 23 infectivity is called the bio assay in the field, in 24 which selected samples are injected into a species 25 which is capable of supporting that infectivity. That 118 1 would be non-human primates, and again, a technology 2 in evolution, transgenics engineered to express human 3 and perhaps bovine PrP. 4 There is also the possibility of trying to 5 prophylax cultures, substrate cultures with chemical 6 agents. This is also in evolution, but the classic 7 molecule in this regard is congo red, which not only 8 seems to bind to PrP Sc, but seems to dissolve 9 infectivity in vitro. New discoveries out of Byron 10 Kohe's lab that tetrapyrrole, including porphyrins and 11 phthalocyanines, can block infectivity. Perhaps some 12 of these compounds can be utilized at appropriate 13 concentrations to use as a kind of antibiotic for 14 substrate cultures. 15 Finally, the slam dunk in this area would 16 be to develop a cell line that lacks a prion protein 17 gene. The prion infectivity, whatever the hell it is, 18 seems to be absolutely dependent on the presence and 19 expression of PrPC. So if one were able to ablate the 20 prion gene out of a substrate cell, that didn't come 21 with 300 other bad pathogens, this may be a strategy 22 of obviating any prion infectivity in vaccines. 23 So I would like to summarize by saying 24 that it is possible, although not favorable, for 25 substrate cells to be infected with prions. It is 119 1 possible, considering the huge bulk of cells that are 2 cultured, 300 million a year for one vaccine, that 3 prion infectivity could potentially emerge by 4 misfolding and/or somatic mutation in vitro. 5 I will note that prion components and 6 excipients may contain prion infectivity. Although 7 this is an old story with regard to classical CJD, we 8 don't have the information for the BSC variant CJD 9 agent. We should worry, at least for the time being. 10 The remedies for this are selective 11 sourcing, avoid animals and people that could 12 potentially be brewing prion infection, biological 13 manipulation in vitro, including anti-prion agents, 14 and maybe ablating the prion gene, and then validate, 15 validate, validate. Prion infectivity should be added 16 to the list of infectivities that are excluded in 17 vaccine lots. 18 I thank you for your attention, and I 19 would be glad to answer any questions. 20 (Applause.) 21 DR. KRAUSE: Phil Krause, FDA. In keeping 22 with the idea that one presumably wants to find cell 23 substrates which carry the least risk, I guess if one 24 presumes that tumor cells have a greater risk of 25 genomic instability than non-tumor cells, are you then 120 1 implying that there's a greater sort of spontaneous 2 mutation than prion risk from tumor cells than for 3 instance primary or diploid cells? 4 DR. CASHMAN: That's a good question. I 5 guess this is basically not quantifiable. If one 6 takes a rock solid cell that enjoys all kinds of DNA 7 repair mechanisms then yes, that is less likely to 8 give rise to the mutation in the prion protein gene. 9 One area which should be investigated, I'm 10 realizing from this meeting, is to take some cell 11 lines and look at 1,000 clones a piece and see if any 12 of the prion copies have acquired mutations. So this 13 would be a piece of data that we could use to actually 14 discuss this issue. Right now, I don't have any. 15 AUDIENCE MEMBER: Along the same lines, I 16 guess the implication would be that if somebody wanted 17 to produce a vaccine in cells that are derived from 18 neurons, given the fact that they are making a lot 19 more of this, you have the potential for greater RNA 20 polymerase mutation rates, and perhaps also greater 21 risks? 22 DR. CASHMAN: Well, neurons are the best 23 factoring for making infectivity, both in vivo and in 24 vitro. Part of that is due to high levels prion 25 protein. Part of it is due to factors that haven't 121 1 been identified as of yet, prion receptor, trafficking 2 of prion protein. So I guess one would want to stay 3 away from neurons unless they came from an animal or 4 human that had the prion gene locked out or not 5 transcribed or translated. 6 CHAIRPERSON RABINOVICH: Dr. Egan? 7 DR. EGAN: As Dr. Minor mentioned before, 8 Japanese encephalitis virus is made in mouse brains. 9 Can the PrP SC of a mouse catalyze the conversion of 10 human PrPC to PrP Sc? 11 DR. CASHMAN: That is an extremely good 12 question. I neglected to mention in my talk, my 13 oversight, that there is a very prominent species 14 barrier between most prion agents, something like one 15 to a thousand, one to a million, even higher for 16 generating infectivity, especially between widely 17 differing species. 18 Now as we move to human cell substrates, 19 we will no longer be protected by the species barrier. 20 Even non-human primates have a sequence similarity to 21 convert human PrP and vice versa. 22 So yes, I think that answers it. The 23 exception to that is of course this new variant agent 24 which we're all frightened about because it doesn't 25 seem to obey species barrier. 122 1 DR. ONIONS: David Onions. Could I just 2 ask the converse of the question that you posed for 3 vaccine substrates, which is I think one that has been 4 discussed. That is, the idea of knocking out the PrP 5 gene. We know that PrP mice are viable, so it looks 6 like perhaps an interesting way to go. 7 But can you also engineer cells over 8 expressing normal PrP and use those as substrates for 9 infectivity? You mentioned that one of the problems 10 was the low level of PrPC in most of the cell lines 11 you have used. Can you not over-engineer cells so 12 that they become susceptible? 13 DR. CASHMAN: The only -- that's never 14 been done. The only data that really pertains is 15 transgenic mice. In fact, if you have knocked out the 16 prion gene, you are absolutely resistant. If you have 17 one normal copy, you have disease with a longer 18 incubation and a shorter rate of -- or a longer rate 19 of progression. If you have two normal copies, you 20 have disease at the normal time. If you have 21 transgenic 10, 20, 100 copies, then the disease 22 presents at an earlier age and is more rapidly 23 progressive. 24 As far as I know, nobody has moved that 25 observation to an in vitro paradigm. 123 1 DR. LOEWER: Johannes Loewer, Frankfurt. 2 I would like to challenge, to some extent, your 3 calculation on the risk of mutations, because they do 4 not take into account the similar biologies of prion 5 protein that are secreted from the cell, as they can 6 spread to other cells. That there's really a short 7 time induction of new PrP Sc. 8 For example, at least to my knowledge, if 9 you infect so to say cell cultures who use prion via 10 PrP, you get another multiplication of infectivity. 11 It stays more or less. If it disappears, it may be 12 more or less there's no real spread. 13 The question I have is, you are asking for 14 validation studies. What would be the material you 15 would recommend to spike? Should we spike with 16 scrapie brain material? Is this relevant for 17 purification in these cases? What would be the ideal 18 spike? 19 DR. CASHMAN: I would very much appreciate 20 some challenge from my numbers. They are new numbers, 21 so don't hold back. 22 The question of spiking, there have been 23 experiments performed at least with a purification of 24 albumin and other plasma proteins. Paul Brown and 25 Robert Roewer published an article last fall in 124 1 Transfusion detailing the amount of infectivity if you 2 start from a high amount in whole blood, what you end 3 up with in albumin. In fact, there's at least a 10 to 4 the 6th loss of prion infectivity if one follows the 5 normal protocols for purification of albumin. 6 The question is whether a single 7 infectious particle is there. The other question is 8 what happens if this single infectious particle is 9 variant CJD. If the single particle is classical CJD, 10 nothing would happen. The disease would not take 11 hold. We don't have enough information about the 12 virulence of variant CJD to be able to answer that 13 question with any confidence. 14 Did that answer your question? I have a 15 feeling it didn't. 16 DR. FRIED: Mike Fried. I think your 17 numbers also don't take into account modifier genes, 18 which could be just random mutation. 19 For instance, I understand that all the 20 people with the new variant CJD have a polymorphism of 21 one type in the PrP, the protein. Is that not true? 22 DR. CASHMAN: In fact, that is partially 23 right. There is a normal polymorphism in the prion 24 protein open reading frame at codon 117. One can 25 either have a valine or a methionine. To date, 125 1 everyone -- well, there's a nice distribution in the 2 normal population, about 50 percent are heterozygotes 3 and about 25 percent are met met, and 25 percent are 4 val val. 5 In a new variant CJD, everyone who has 6 contracted the disease to date has been met met. But 7 that in fact may be due to the fact that the met mets 8 are more susceptible to the agent. It may not be due 9 to the fact that val vals or heterozygotes are 10 resistant. 11 DR. FRIED: Sure. I am just saying that 12 that goes into your numbers, whether there's 13 modifications. 14 DR. CASHMAN: The point of modifier genes 15 is a very important point. I tried to -- 129, I'm 16 sorry. That's David Asher who was keeping me honest 17 in more ways than one. It's the polymorphism is at 18 codon 129. 19 Now what was I going to say? I was going 20 to say something. 21 CHAIRPERSON RABINOVICH: Modifier genes. 22 DR. CASHMAN: Modifier genes, yes. 23 Modifier genes have of course been proposed from 24 animal studies of infectibility and experimental 25 scrapie. Dr. Prussiner has hypothesized a protein X, 126 1 which may be a receptor or may be a chaperon that 2 somehow modifies susceptibility of an animal to prion 3 diseases. There is also a protein Y that Dr. 4 Prussiner has hypothesized. 5 I agree with you from the bottom of my 6 heart that there will be modifier genes affecting 7 susceptibility to prion diseases and the propagation 8 of prion infectivity in vitro. But we don't know what 9 they are yet. At the crude operational level of being 10 able to infect cells, yes. We can infect cells in 11 vitro. So at least some of those modifiers have to be 12 there. Did that make sense? 13 CHAIRPERSON RABINOVICH: Dr. Hayflick, 14 final question. 15 DR. HAYFLICK: Hayflick, UCSF. I was 16 intrigued by your observation that the species barrier 17 for prion transmission is less, is reduced between 18 non-human primates and humans, which would raise some 19 additional concern about the use of primary tissue, 20 and particular, primary monkey kidney tissue for the 21 production of human virus vaccines, because contrary 22 to popular belief, that tissue and any primary tissue 23 does not contain -- and I'll use primary monkey kidney 24 as an example, only cells that are derived from a 25 particular part of the kidney. 127 1 A primary monkey kidney culture consists 2 of an enormous variety of differentiated cell types 3 that compose the vascular system and neurons. So that 4 monkey neurons do play a part in the production of 5 polio virus, for example, derived from monkey kidney. 6 So that I think that it's important to mention that 7 neurons are not only a part of brain tissue in 8 considering various cell substrates. 9 Also I was wondering whether there's some 10 reason why you omitted the mention of trypsin as 11 potential source in your discussion of substrates or 12 media supplements for prion transmission. I didn't 13 see trypsin as a component. Was there some reason for 14 that omission? 15 DR. CASHMAN: Just blanking out. So thank 16 you for adding that to the list. 17 DR. HAYFLICK: The final question I have 18 is I may have misunderstood your slide in which you 19 were addressing the question of utilization of human 20 cell line, that a human cell line had not been used 21 for the production of enormous numbers of doses of 22 vaccine, for example, in order to support one of your 23 contentions. 24 Did you mean it in that respect, a human 25 cell line that is defined to be immortal and 128 1 transformed? Or did you mean any human cell 2 population? 3 DR. CASHMAN: Perhaps you could educate 4 me. Has any human immortal or neoplastic cell line 5 been used to develop large penetrants, large 6 population vaccines? 7 DR. HAYFLICK: No. Not as you have 8 defined that cell population. I was interested 9 particularly in normal human cell populations, which 10 have been used for the production of up to a billion, 11 with a B, doses of virus vaccine. But you are not 12 including in that? 13 DR. CASHMAN: Say it again. I'm sorry. 14 DR. HAYFLICK: Normal human cell 15 populations have been used for the production of about 16 three-quarters of a billion doses of human virus 17 vaccine world wide, but these are not continuously 18 propagable abnormal heteroploid cell populations. 19 These are normal finite lifetime cell populations. 20 DR. CASHMAN: So these are cell strains? 21 DR. HAYFLICK: Yes, as I defined them. I 22 realize there's a problem in understanding these 23 terms. 24 DR. CASHMAN: Which vaccines? 25 DR. HAYFLICK: Virtually all pediatric 129 1 vaccines, polio, rubella, mumps, measles, rabies, 2 adenovirus, some rhinovirus vaccines, are all produced 3 on a semi-continuous human diploid cell strain like 4 WI38 or MRC5. 5 DR. CASHMAN: Thank you. 6 CHAIRPERSON RABINOVICH: I think you 7 better clarify. 8 DR. LEWIS: Yes. To my knowledge, there 9 are really no nerve cells in the kidney. The nerve 10 cells lie on the spinal cord and porosises go down 11 there. I don't believe there are any nerve cells in 12 the kidney. Even if they are, once the nerve cell is 13 differentiated, they basically do not grow on tissue 14 culture. I think that needs to be corrected. 15 CHAIRPERSON RABINOVICH: Okay. I would 16 like to move rapidly to bring the panel members up to 17 the podium, and to invite Dr. Onions to come over and 18 run things. 19 DR. ONIONS: I notice that we now have 20 actually 45 seconds for discussion according to the 21 program. 22 (Laughter.) 23 Brilliant as this panel are, I don't think 24 they could do that. So could I have some guidance on 25 when you would like to close this panel session? 130 1 Perhaps somebody could give me some guidance. 2 CHAIRPERSON RABINOVICH: Forty five 3 minutes. 4 DR. ONIONS: Forty five minutes. Thank 5 you very much. 6 Okay. We were charged in this panel to 7 answer a number of questions. I'll come to those and 8 try and go through and cover the areas that the panel 9 will discuss. I would be very grateful for as much 10 participation from the audience as possible. 11 I thought it was just useful to pick up 12 two strands that I think came out of some of the 13 comments yesterday. One I think that's important to 14 make, and that is that vaccine production is a very 15 pragmatic process, and that once there have been lots 16 of theoretical objections to particular cell 17 substrates, particularly autogenic cell substrates, 18 there are very practical reasons for the use of cell 19 substrates that might be immortalized or neoplastic 20 from a new generation of vaccines. I don't think we 21 should lose sight of that. There are very practical 22 reasons in scale up and use that I think we should 23 bear in mind to produce therapeutic vaccines. 24 The second concern was that came out, and 25 perhaps didn't get enough airing as it should have 131 1 done, and that is there clearly is a trend in society 2 at the moment about concern in safety of vaccines. 3 That perhaps therefore focuses particularly on the 4 item we are going to discuss today, which is the 5 possibility of adventitious agents. 6 The question that the FDA asked us to 7 evaluate, or at least one of the first questions, and 8 they would like the panel to take a look at is do 9 neoplastic cells represent the greater equivalent or 10 lesser risk for the presence of adventitious agents 11 than primary cells, diploid cells, or non chunogenic 12 continuous cell lines. 13 I am not sure, given my previous comments, 14 whether I think necessarily that you can answer that 15 in a simple sense, but does anyone in the panel want 16 to sort of pick that one up first of all? 17 What I could perhaps do is to prompt 18 people, is perhaps to put up those which is just my 19 suggestion, of some of the factors that might 20 influence the risk of adventitious agent testing in a 21 variety of cell substrates. 22 DR. MINOR: I am personally convinced that 23 they are going to be better than primary cultures. I 24 am not sure whether they will be necessarily better 25 than human diploid cells or anything else if we even 132 1 get a decent banking system going. It seems to me 2 that when you get to that kind of stage, it's the 3 concern about how you find what's in there rather than 4 anything else. I think the actual extent to which you 5 can characterize them is clearly to me, it's very 6 similar. 7 DR. ONIONS: That would be a generally 8 universal statement that primary cells are likely to 9 be more difficult to characterize and therefore, if 10 you can use a cell line, that is probably the way to 11 go. 12 But I think it's also worth making a 13 countervailing point that there are still vaccine 14 strains that are very successfully produced in primary 15 cells. There are others that have been passed in 16 primary cells and therefore change them into a cell 17 substrate, the genetic stability of them. So there 18 are nevertheless countervailing arguments. 19 I think the statement is correct. That 20 is, adventitious agent testing is clearly going to be 21 more in the region -- 22 DR. ROBERTSON: Another point which one 23 could consider. Where neoplastic cells might be 24 considered more susceptible than primary cell 25 cultures, in that the primary cells are derived 133 1 specifically for vaccine production. Whereas the 2 neoplastic cells have probably been kicking around at 3 least one, if not several laboratories, before being 4 put into use as manufacturing of vaccines. 5 Because of that, they may well have picked 6 up something that you don't want to be there. Nothing 7 to do with the cell type or the origin of the cell. 8 So a virus of some different species all together 9 which you really have got to check for. 10 So if you are actually setting up a cell 11 bank of a neoplastic cell, you shouldn't just be 12 considering species of origin of that particular cell, 13 whether it's porcine, human, murine whatever. You 14 have really got to consider any virus under the sun. 15 We know there have been instances of this happening. 16 This morning there was the comment about 17 I think it was a human endogenous retrovirus which was 18 actually of murine origin, been picked up from passage 19 somewhere. 20 DR. ONIONS: There is a good example in a 21 commercial product. That of course was the Glaxo 22 Wellcome's novalma cell line which was used to produce 23 interferon, which in fact contained SNRV, and probably 24 picked up SNRV in George Kahn's lab at some point 25 during its history, I think was the general consensus. 134 1 But clearly that was unknown and the whole system was 2 used in the production for several years before it was 3 realized that perhaps this was contaminated by an 4 adventitious agent that you just would not expect in 5 this cell line. So I think that is a very good point. 6 There are issues about tumor cells. One 7 of the things that occurs to me is that -- actually it 8 does concern me quite a lot about adventitious agent 9 testing. That is that it is rather traditional in 10 character still. It is not very directed in terms of 11 its specificity in looking for certain viruses. That 12 is changing I think, but I think until recently that 13 has been the case. 14 For instance, we have known that in 15 certain tumor cell lines, that viruses that we have 16 only recognized in the last decade are certainly 17 found. For instance, well HHV-6 isn't a transforming 18 virus, but there are cell lines that carry HHV-6 that 19 have been used in the lab for many years. The same is 20 true for HHV-8, which is a transforming virus. 21 So there is a concern that we may have 22 cell substrates that are contaminated by other tumor 23 viruses. 24 Tom, would you like to pick up? 25 DR. BROKER: I think we have actually a 135 1 wonderful opportunity for a so-called natural 2 experiment. That is solid organ transplantation. It 3 turns out, as we all know now, virtually all kinds of 4 organs, not only the corneal we have just heard about, 5 but kidney, liver, pancreas, part of the intestine, 6 heart, lung, so forth, have all been transplanted. I 7 think the opportunity is that the recipient is 8 invariably immuno-suppressed until the transplant 9 takes, and then they are slowly weaned off the drug 10 like cyclosporin. 11 Yet on other occasions, the transplant 12 fails for one or another reason. One could go back 13 into failed transplants to look for the reactivation 14 of agents that came from all these different tumor or 15 tissue types I mean. 16 One example I could cite that we recently 17 encountered in the course of our kidney transplant 18 study is a pair of kidneys that went in from a five- 19 year-old boy to a 19-year old female. Within a few 20 days, the kidneys had completely become destroyed, 21 necrotic. 22 It turned out -- they suspected CMV 23 infection, but it turned out to be adenovirus. The 24 presumption, and I'm being completely hypothetical, is 25 the five-year-old boy who had died in a bicycle 136 1 accident, the donor, probably was in the age bracket 2 where adeno was just a natural infection in his 3 airway, and that these cells say from his tonsils or 4 adenoids, which were in the midst of processing the 5 adeno, became circulating, were in the kidneys, and 6 the recipient female then acquired adeno-infected 7 kidneys, and upon transplant to her, the virus 8 reactivated and just wiped out the tissues. I might 9 also say the different individual who received the 10 boy's liver also lost the liver. 11 So presumably these were entering through 12 B cells that were in any of these remote organs. 13 Nonetheless, the basic opportunity to look at organ 14 recipients I think is the experiment to ask how much 15 infectious agent is being transferred. 16 DR. RUSSO: Carlo Russo from Merck. I 17 think as you indicated, these patients are profoundly 18 immune-suppressed. Therefore, is going to be very 19 difficult to assess where this agent came from. In 20 your case, it's very well possible that the woman was 21 exposed to adeno virus. Since she was immuno- 22 suppressed, that's the reason why she got the 23 infection. 24 DR. BROKER: Well, in that case, it's why 25 I did also point out that the liver who went into a 137 1 completely different boy also wiped out. But I agree 2 with you. One doesn't know whether it's endogenous, 3 but it does give some indication of infectious agents 4 in these organs. 5 DR. ONIONS: Could I just actually take a 6 backup actually? I was about to go back to the 7 primary cell issue again. Phil gave I thought a 8 wonderful presentation. I hadn't heard parts of this 9 before. It actually started to worry me a little bit 10 actually. 11 To what extent do you need now to control 12 the kinds of colonies of these particular primary 13 colonies? I'm not sure, I mean I don't know what kind 14 of testing goes on in these colonies for a range of 15 adventitious agents. 16 Can you maybe just comment on that? I 17 mean are we dealing with inverted SPF animals? 18 DR. MINOR: Well with respect to the 19 primate, you will certainly not. But they are 20 increasingly heavily monitored. It depends very much 21 on the manufacturer and how much monitoring they do. 22 One manufacturer, for example, has only 23 recently, well in the last four or five years I guess, 24 started using colonies of monkeys that were monitored 25 for foamy virus. The result of that is being 138 1 revolutionary in terms of the number of cultures that 2 you get surviving to production. 3 You would have thought you might have 4 started this a bit earlier perhaps. But you couldn't 5 call them SPF, but they are increasingly closely 6 monitored I think. Certainly some manufacturers have 7 them more closely monitored than others. 8 But one of the difficulties with the whole 9 of adventitious agent business of course is you only 10 really find what you are looking for. That's an 11 ongoing problem. 12 Things like chickens are a different 13 matter. I mean I think this would establish what you 14 need to do to make an SPF chicken colony. But 15 primates are much more tricky. 16 DR. ONIONS: There are other cell 17 substrates out there that are used, cells like primary 18 hamster kidney cells in JV vaccines and various other 19 things. So there are I think other vaccines out there 20 that are going to come to attention because they use 21 primary cells. I think we perhaps ought to start 22 thinking of the kinds of procedures that are needed, 23 like closed colonies and embryo derivation of these 24 animals in some cases. 25 Could I move to perhaps the third element. 139 1 That is, we have heard a lot about retrovirus. 2 Retroviruses always come back to focus when we deal 3 with cell substrates. What is the panel and the 4 group's feeling here in general about the concerns of 5 using either immortalized cells or transformed cells, 6 because frequently those -- well, that's not an 7 accurate statement. Activation of transcription of 8 endogenous genes is more frequent in such cells. Is 9 that of concern or not a concern? Or do we have to go 10 cell by cell, species by species, to answer that 11 question? 12 John, would you like to make a comment? 13 AUDIENCE MEMBER: Well, I think clearly 14 there are some famous cases of activation of 15 transcription of endogenous retroviral genes and 16 genomes and tumor cells. We heard about germal 17 tumors. We have heard about recently in the news 18 about mammary tumors and probably a variety of others. 19 It is not entirely clear to me whether 20 this actually represents activation of transcription 21 of these cell lines or a fixation of a differentiated 22 state which itself is what's activating the 23 transcription. I am inclined to think probably both 24 are true in different cases. 25 One of the things that we have learned, at 140 1 least from human tumor cell lines, is that none of 2 these things that are activated have ever been shown 3 to be infectious, despite the fact that as came out in 4 the earlier talk, the probability for recombination 5 between a large variety, for example of the HERV-K 6 sequences, would seem to be rather high. If these 7 were other types of retroviruses, such as MLVs where 8 you do get that endogenous, you do get that kind of 9 thing. 10 So there's probably something else that's 11 protecting the people in the cell against actual 12 infectivity in this particular case. We don't yet 13 know what it is. 14 DR. ONIONS: I was just going to bring in, 15 I was very struck by the HERV-K story. That is to a 16 very pragmatic level. I take the point that it does 17 seem to me that it doesn't much recombination or much 18 adjustment to me, at least into a more functional 19 virus. I mean should we simply be screening any human 20 cell substrate for the expression HERV-K? Would that 21 be something that would be useful to do? 22 DR. PALLEY: That's a problem, in that you 23 will I guess you will have a hard time finding a human 24 cell line that would not express any human retrovirus, 25 so I report that it's the case for HERV-K, the special 141 1 HERV-K family. 2 DR. ONIONS: But that particular locus, I 3 mean that is not expressed, as I understood you, in 4 all? 5 DR. PALLEY: We do not know whether that 6 particular locus is expressed in some cell lines. We 7 know that if this, at least this HERV-K family, is 8 activated for some reason, that there is very likely 9 not only one locus activated within the genome, but 10 that there are several loci that will be activated. 11 So it's possible that this locus is also activated, 12 but we also then have the problem that if that locus 13 would not be activated, that there's some 14 transcomplementation. So one gag, intact gag gene 15 would complement another intact pol gene and so on. 16 So that could be a problem, but we see it then in 17 several, even in normal peripheral blood lymphocytes, 18 we have HERV-K expression. It obviously is not a 19 problem. 20 So I wouldn't see so much a problem in 21 that we have HERV expression if defective sequences 22 are expressed, and that are not coding the intact. 23 AUDIENCE MEMBER: I think we have to focus 24 on infectivity here, because if we just go looking for 25 expression of defective stuff, you always find it, and 142 1 we all go home and won't be able to do anything. It 2 might not be a bad idea perhaps, but -- 3 (Laughter.) 4 AUDIENCE MEMBER: Sounds good to me. But 5 just to push this point a little bit further on the 6 HERV-K, you know, about two-and-a-half years ago, 7 there was a report in Cell, which is a better journal 8 than I usually publish in, which claimed that HERV-K 9 env can act as a super antigen that then stimulates 10 diabetes mellitus in some people. 11 DR. ONIONS: I think that story -- just 12 before you go on, I could be wrong, but Johannes might 13 know. Is that story being modified? I'm not quite 14 sure that's -- yes. It's no longer supported. 15 AUDIENCE MEMBER: Okay. I think it would 16 be worth explaining how it's been modified. 17 DR. ONIONS: I think the retraction is in 18 fourth hit. So before you sort of put that out as a 19 paradigm. 20 DR. LOEWER: I think the main point is 21 that a couple of groups tried to repeat this data, and 22 they were not able to repeat. So it seems not to be 23 specific and even not effects on the T cell lines 24 could be repeated. So let us depart from this idea. 25 DR. ONIONS: Can I maybe get your opinion? 143 1 I really would like to get some feeling because we 2 have heard a lot, and it's scientifically really 3 interesting by the expression of these human 4 endogenous retroviruses. I think John has probably 5 just summarized it. It looks like we're saying the 6 list, this collected group here of retro, are saying 7 that as far as we are aware, at the moment these are 8 not of concern and uninfectious, and probably 9 therefore there is not a great deal of point in 10 looking for expression of these in cell substrates, 11 the very pragmatic practical point. 12 Would that be your opinion too, Johannes? 13 Johannes is nodding. That's a "yes," I 14 assume. 15 AUDIENCE MEMBER: From a research 16 standpoint, it's absolutely worth pursuing to see if 17 one can find these things eventually. But in terms of 18 vaccine issues, I don't see how we could possibly deal 19 with it now. 20 DR. SCHUEPBACH: Yes. I also would like 21 to make a comment regarding that super antigen 22 activity because we are coauthors in that paper. It 23 is true that the presence of these sequences, RNA 24 sequences in the serum we can not repeat, so it's not 25 specific for IDDM patients. But to my knowledge, our 144 1 data regarding the super antigen activity and the 2 stimulation of VP cells has not been disputed by any 3 other group. So that is still around. 4 I think that the real important topic here 5 is whether these endogenous viruses actually give rise 6 to infectious particles. I believe that with the PERT 7 assay, we actually can contribute very much to this 8 question. I think, as I pointed out, of course the 9 easiest thing is to test the super natants for 10 reversed inscriptase activity. But I think with a 11 little bit of additional work, it should also be 12 possible to define that profile of RT activity and 13 cellular, DNA polymerase activities along the 14 different fractions of the sucrose gradient and tend 15 to recognize any abnormal pattern that might be 16 associated with infectious rate of viruses. 17 AUDIENCE MEMBER: Let me just go back then 18 because obviously the question comes up if the public 19 is asked to accept a vaccine that's made in cells that 20 express HERV-K. Even if one part of the story has 21 been refuted, the question comes up. You know, are 22 there potential immunological consequences of the 23 expression of antigens from these kinds of cells which 24 are not expressed in human diploid cells? Even if 25 this story is wrong, I guess, and it sounds like it 145 1 hasn't all been refuted, the question then comes up of 2 what level of scientific data in the literature is 3 necessary to completely refute it? 4 I can imagine the outcry that could occur 5 if people believe the story. If there appears to be 6 some controversy about what part of this story has 7 been refuted, then I think one might have a public 8 confidence problem as well. 9 DR. HENEINE: David, I have a comment 10 which could be redundant, but go ahead and say it. 11 While thinking about all these questions from my 12 simple mind, it looks like if you want to compare cell 13 lines versus primary or diploid cells, the two 14 questions that were raised is which ones transmit less 15 adventitious agents or transmit less neoplasms to 16 vaccine recipients. 17 What we have heard so far about the 18 mechanisms of the neoplasms, many of those are 19 mediated by viruses or viral-like elements. So it 20 looks to me that the majority of the concerns are 21 rising from the adventitious agent group rather than 22 from other elements. 23 So therefore, in trying to make up our 24 mind, based on the available data, which one is the 25 more suitable substrate, maybe we can go very simply 146 1 with a checkpoint list on these different cells, 2 targets, which one we can test for the presence of 3 these adventitious agents known, unknown, and which 4 can be better monitored, which can be for practical 5 reasons of culture as well, and make up our mind, 6 rather than jumping right and left with different 7 issues. 8 If you can say cell lines, primary and 9 diploid, and then go one by one, all these concerns 10 that we have been talking about, and say which one is 11 more suitable for each of these points so that we can 12 conclude at the end. I mean it's one suggestion. 13 DR. SCHUEPBACH: I would like to come back 14 to the human endogenous viruses. Since we all carry 15 them and have them expressed in one or the other part 16 of our body, I do not think that this presents a 17 particular risk. Independent of whether you have a 18 super antigen activity or not in some of them, I think 19 is just the same as receiving blood from any person, 20 because they also would have these endogenous 21 retroviruses. 22 So I think the only thing that matters is 23 endogenous retroviruses from other species, not from 24 humans. 25 DR. LEWIS: This is a question. I think 147 1 we are sort of being faced with something of a dilemma 2 here, because as you very correctly pointed out, the 3 one way to look for a retro virus, adventitious retro 4 viruses or endogenous retro virus, whatever, is by the 5 PERT assay. Now if we take a situation in which our 6 HERV-K is expressed and there's RT activity in there 7 with the PERT assay, from a regulatory perspective, 8 what do you do about that? 9 I don't think you can dismiss it. I don't 10 think at this point in time that you could dismiss the 11 use of the PERT assay for looking for adventitious 12 retroviruses or any type of retrovirus activity in 13 there. 14 So the question is going to be, when you 15 find something, what do you do about it if the assay 16 is positive? 17 DR. SCHUEPBACH: I think once you have 18 activity, then you have to characterize what it is. 19 Depending on whether this is exogenous or endogenous 20 virus, steps will be taken. I think identification is 21 important. 22 DR. LOEWER: I think we shouldn't continue 23 to discuss use of this type of cell lines in absolute 24 terms, because the other side of the coin of course is 25 a product which could be made from it. This is always 148 1 to be judged in conjunction, in my opinion. For 2 example, there's the question of endogenous sequences 3 which may be active, just retro virus or line 4 elements. So far as I know, this is not the same in 5 all cell lines. It's mainly expressed in tumor cell 6 lines, and so far as I know, there is so far no need 7 to use these cell lines for vaccine production. 8 If somebody believes it's necessary, must 9 be a very special virus which can only replicate in 10 these cells, then perhaps we do not have another 11 choice to use them. 12 DR. ONIONS: Could I just go to Arifa? 13 She has been waiting very patiently. Then Steve. 14 DR. KHAN: Yes, thank you. I think it is 15 important to clarify the word "expression" in terms of 16 human cell lines and human cells. I think we all 17 expect that there will be some RNA expression in the 18 human cells from endogenous retroviral sequences. 19 However, I don't believe that you are going to get 20 particle production in the majority of the cells under 21 normal conditions. So therefore, I think the use of 22 the PERT assay would be very helpful to evaluate 23 particle production from human cells, which can then 24 further be investigated for infectivity, as opposed to 25 looking at RNA expression, which I think you would 149 1 find at some level in all human cells. 2 DR. ONIONS: I would like to endorse that. 3 I think really we're making it too complicated. I 4 think it's very simple. You use that kind of assay 5 system which has a very high sensitivity developed by 6 Yumascript here in the audience. Then if you get a 7 positive, then you go and look and see if there's 8 something there that's infectious. It is a hierarchy 9 of testing strategies, it seems to me. 10 DR. HUGHES: I am willing to take an even 11 stronger line here, which is I think akin to the line 12 that was espoused by John Coffin. 13 If your technology is sensitive enough, 14 all cells from all vertebrates are going to have 15 endogenous viruses in them, either intact or 16 defective. The question really devolves down to not 17 whether they are there, because they are. If we use 18 the most stringent criterion, are they there, you are 19 not going to be able to do anything. We'll be 20 paralyzed in terms of making vaccines. 21 So the obvious criteria is not whether or 22 not these agents are there, and in particular, I think 23 not whether they would make physical particles, 24 because my strong prejudice is if you look hard 25 enough, you are almost certain to be able to find it. 150 1 There are almost certainly ways of looking even more 2 stringently than we look now. In very large batches 3 of material prepared for vaccines, if you look hard 4 enough, you are going to find something. 5 The question is whether there is an 6 infectious agent which represents any kind of 7 pathologic threat when its present in a vaccine. We 8 have, as has been pointed out earlier, given from 9 chicken cells, some of which were clearly contaminated 10 with agents which are infectious for the chickens, if 11 not for humans, that hundreds of millions of doses 12 have been given, with as far as anyone can tell, no 13 untoward effects. 14 DR. ONIONS: Okay. I would like to bring 15 Jim in, because I am not sure that statement is -- 16 well, it could theoretically be true. Maybe if you 17 get a more sensitive technique, maybe you are correct. 18 But there is a very strong distinction if 19 you look at cell lines like MRC-5 and compare them to 20 what you see in say AB, you get signals, if you're 21 looking at MRC-5 you don't. If you look in tetra 22 anacells, you get a signal. 23 DR. HUGHES: I would argue that is with a 24 particular assay that has been tuned up to detect RT 25 in a particular way. I would be willing to wager that 151 1 if we look hard enough, we could certainly find 2 evidence of particle production in any of these cells, 3 simply because they are full of endogenous viruses 4 that -- I mean the very fact that there's obviously 5 expression is RNA present. 6 DR. SCHUEPBACH: But the important thing 7 is I think the number of particles. You have very, 8 very low number of particles, and you do a test, a 9 PERT assay from a concentrated material where you 10 pellet virus, let's say, from one liter, and assay 11 that and find a very low activity. You know for 12 production of the viral vaccine, the cell harvest, the 13 vaccine harvest will be diluted 50 to 100. You also 14 realize that most of the particles even of infectious 15 retroviruses are non-infectious. Then you certainly 16 can come to some calculation which permits you to 17 establish a level of safety where you have a very high 18 probability that this vaccine is safe. 19 DR. HUGHES: But the safety is not 20 predicated on whether or not there are physical 21 particles. The safety is predicated on whether or not 22 the particles are infectious. 23 If the particles are infectious, a very 24 small number is very important. If the particles are 25 not infectious, in particular, if the particles are 152 1 not infectious for humans, the presence of a 2 relatively large number of particles is probably also 3 irrelevant. 4 DR. ONIONS: My only caution about that is 5 that you can make assumptions about affectivity that 6 also are not true. Since Clive Patience is here -- 7 DR. HUGHES: I think you have to do the 8 test. I don't think you can make assumptions about 9 it. 10 DR. ONIONS: Well, the point I was going 11 to make actually was that Clive's group and our group 12 showed that you could actually infect cells with PERV, 13 yet those experiments have been done 20 years ago, and 14 been done by very good people, including George Tadai, 15 and were unable to show infectivity. It's just the 16 techniques have changed slightly and we could get 17 infectivity. 18 So I think that you are making -- there is 19 a straight yes or no about infectivity. That is not 20 always the case with these retroviruses. 21 DR. HUGHES: I am not trying to suggest 22 that the assays that we have for infectivity are 23 necessarily always 100 percent accurate. But what I 24 want to get away from is the idea that the presence of 25 a physical particle is somehow a measure of safety or 153 1 lack of safety. 2 I certainly agree with you that there are 3 numerous technical problems in determining 4 infectivity. But what I believe we should focus on is 5 better ways of doing infectivity assays rather than 6 better ways of doing physical assays. 7 The physical assays can be very useful if 8 they are coupled to infectivity. In fact, I believe 9 that was a statement that John Coffin made. What I 10 would propose is what I think John Coffin proposed, 11 that we actually use these really wonderful sensitive 12 techniques, but in the context of measuring whether or 13 not the viruses are infectious, not whether the 14 particles are physically present. 15 DR. SHEETS: Can I ask a very pragmatic 16 question of Dr. Hughes? 17 DR. HUGHES: Sure. 18 DR. SHEETS: I'm Becky Sheets, FDA. What 19 I hear you suggesting is that rather than testing for 20 RT activity by a physical assay, as you called it, a 21 PERT assay or a conventional RT assay, you think that 22 it would be preferable to test vaccines for 23 infectivity assays? 24 DR. HUGHES: The ability to transmit that 25 RT assay to a reasonable recipient cell. I believe 154 1 this is exactly what John -- 2 DR. SHEETS: The pragmatic question is 3 that would you do this testing lot by lot on vaccines? 4 For instance, if you were making a vaccine in a 5 primary cell substrate, for instance an egg, would you 6 test each lot of vaccine or each batch of vaccine for 7 infectivity assays? Then the really pragmatic part of 8 it is, if you are making a flu vaccine, where the 9 timing of production, the timing of testing, and the 10 timing of lot release is very tight, would you 11 recommend these infectivity assays on lot by lot for 12 primary source? 13 DR. ONIONS: We're running out of time. 14 Do you want to answer that? You have been asked a 15 question, do you test lot by lot? 16 DR. HUGHES: Very simply, if we're talking 17 avian systems, I think there are reasonable ways of 18 determining that endogenous avian viruses are not 19 infectious. My personal bias, and I mean it no more 20 broadly than that, is in the case of avian viruses, as 21 long as you carefully establish that the avian viruses 22 that are present are not infectious for human, that 23 that's not necessary. But that is my prejudice. 24 DR. SHEETS: That's fine for SPF 25 situations. I am asking this question because this is 155 1 what sponsors ask FDA. So they want to know do we 2 need to do this lot by lot, or if a cell bank, you can 3 do a one-time characterization. Of a primary system, 4 you can't do it that way. 5 DR. HUGHES: Use SPF chickens and don't 6 ask. 7 (Laughter.) 8 DR. LOEWER: I would like to make a 9 comment to Dr. Hughes' comments. They are very sound 10 in a scientific meaning, but they face regulatory 11 problems, the main problem indeed. 12 Regulatory authorities have to show that 13 there is no infectivity and the proof of non-activity 14 is always nearly impossible in a scientific sense. 15 You will always find reasons to say you were not able 16 to find infectivity. Look at HIV. If you would use 17 MRC-5, for example, or a lot of other animal cells, 18 you would never find infectivity of HIV. The same is 19 true for many situations. 20 So there a fundamental problem is to test 21 for non-activity or noninfectivity. 22 DR. ONIONS: I would like to stop this 23 because we are running out of time and there are other 24 issues. 25 I am going to take the Chairman's 156 1 privilege and just say that I actually think you need 2 multiple techniques, because I think as Johannes has 3 just said, if you have complete infectivity, you will 4 miss things as we would have missed cell lines 5 producing ver. I think you really need to have a 6 combination technique. So I think it's a belt and 7 braces situation. That's a personal view. 8 What I would like to move onto, is were 9 asked by the FDA also to consider species of origin. 10 I think really you end up in very general statements 11 here. You can argue that if you are worried about 12 adventitious agents, then clearly there are species 13 barriers to the transmission of some agents. On the 14 other hand, other agents do go across species 15 barriers, sometimes in abortive replication. They can 16 be very nasty. We of course know that herpes B going 17 across species barriers is actually lethal. Ad 12 in 18 hamsters is oncogenic. Equine herpes veras, which is 19 an alpha herpes veras is oncogenic in hamsters. There 20 are natural examples of cross-species transmission, 21 the ovine herpes veras 2 is innocuous in sheep, but it 22 kills cattle. So there are examples of these 23 heterolic transmissions being worse than natural 24 infections. 25 Is there anything that we can say, the FDA 157 1 have asked us, in a general statement about species of 2 origin? My own view is I don't think you can, but 3 does anyone want to make a statement? 4 DR. MINOR: I think sometimes it is better 5 and sometimes it's worse. 6 (Laughter.) 7 DR. ONIONS: Yes. That's exactly what I 8 think. Thank you, Phil. 9 I would like to drop the discussion now, 10 because I think that sums it up. I'll turn the phrase 11 back on the edge and say it's a case-by-case, it seems 12 to me. 13 I don't want to trespass really on 14 yesterday's, but I think maybe just to come back to -- 15 we're going to move onto assay systems in a second, 16 but I think one of the issues we're coming around to 17 in a second is latent viruses, because those seem to 18 be the real concern. It may be worth just remembering 19 some of the things that were partly discussed over the 20 last two days. That is, that the complementation of 21 defective viruses can occur. For instance, 22 adenoarectus can be complemented by HPV in hela cells. 23 We talked about psuedotype formation both 24 today and yesterday. I think just I would like to 25 make the point about pseudo formation. We talked a 158 1 lot about retrovirus retrovirus pseudotype formation, 2 but this can occur across viral species. For 3 instance, paramyxovirus is rather badly, but they can, 4 pseudo type retro viruses. So you could alter the 5 host range when endogenous agent is expressed in your 6 cells. 7 Of course there are recombinants. Some of 8 these recombinants, and we have got representatives 9 who did the work here, interesting recombinants like 10 SV-40 adnivos recombinants. 11 One of my concerns, we'll come on in a 12 second, I think is like polyoma viruses in cell 13 substrates and that potential for interaction with 14 other cells. 15 So if we can take that as a kind of 16 background, can we turn to a question of -- this had 17 come up and was discussed by several candidates, 18 unknown viruses. What are the potential candidates 19 and what kind of systems do we use to try and go 20 looking for those unknown viruses. 21 Anyone want to comment on what we should 22 be doing about novel cell substrates and you have got 23 a virus there that you don't know anything about. 24 What sort of technique should we be applying? 25 DR. PALLEY: I would like to make a 159 1 comment from that HERV field. I think if we talk 2 about possible recombinations of HERVs with some other 3 viruses that might be a little difficult, I think it 4 could be conceivable. 5 We do not know, however, it's very hard to 6 predict the outcome of whether it is possible at all. 7 We basically have or we might have situations where a 8 HERV is expressed where another virus, retrovirus is 9 also present in the same cell. I think that has to be 10 discussed or taken into consideration, that there 11 might be possible recombinations between HERVs and 12 some retrovirus status put into a particular cell. 13 However, I think it's very hard to predict whether 14 there's any possible recombination and what the 15 outcome of that recombination might be. 16 But I think there are examples where retro 17 viruses indeed recombine with each other and produce 18 some productive outcome. But it's very hard to 19 predict what, in which ways HERV case for instance, or 20 whatever HERV sequence within human genome could 21 recombine with something else. 22 DR. ONIONS: I was just going to point out 23 I think one of the things I have been very struck by 24 by some of the talks here, particularly the 25 polymerized talks, it does strike me, the comment I 160 1 made to Phil earlier about the use of cynomologous 2 monkeys and perhaps have these been screened for other 3 polyoma virus. That seems to be of concern. We know 4 that polyoma viruses are coming in as contaminants in 5 bovine serum into the primate cells, the bovine 6 polyoma virus. They worry me as potential 7 adventitious agents. I am just wondering again 8 whether we should be doing more in terms of redundant 9 PCR approaches to look for these agents in both 10 primate and non-primate cell lines. 11 Any views on that? 12 DR. ROBERTSON: This is potentially the 13 most important door, also the most difficult to deal 14 with. If you think back to what Phil was saying in 15 the first talk this morning, all these instances of 16 vital contamination, generally they occur with viruses 17 unknown at the time, viruses expected, the presence of 18 viruses in vaccines or biological preparations. 19 Potentially is not something we're talking 20 about today, it's not an endogenous virus or 21 recombination between an endogenous virus, but 22 something unknown that's going to leap up at us out of 23 the dark. Of course almost impossible to deal with. 24 But Joerg was saying this morning about 25 this is what we should be looking out for, the 161 1 unknown, and if possible using a more broadly reactive 2 type of assay rather than highly specific type of 3 assay to look for something. 4 If we knew to look for something, that's 5 fine, we can deal with it. It's what's not there 6 which causes the problem. 7 DR. SCHUEPBACH: May I add something to 8 this? I think our chances to detect such unknown 9 agents are really much better if they are present at 10 high concentration. So I don't know whether you 11 accept that concept of cellular cloning in order to 12 either get rid of these agents or to have them at the 13 very high concentrations so that their detection is 14 actually much easier. In the meantime, you can try to 15 activate the host cells by all kinds of different 16 agents. You do EM studies, you do serological 17 studies, use broadly cross-reactive antibodies. I 18 mean this is a wide field actually of methods you can 19 employ. 20 So I think using such an approach, we 21 should actually be quite capable of detecting such 22 agents. 23 DR. MINOR: Can I ask for a definition 24 here? What do you actually mean by a virus? I mean 25 what I would understand by a virus is something that 162 1 actually grows in the cell and increases in numbers. 2 If you think back to the very, very early days of 3 polio vaccines, amongst others, there was an awful lot 4 of effort put into trying to make sure that you put 5 your supernatant or whatever, into all sorts of 6 potential different systems where a virus might grow. 7 So you put it into different cell 8 substrates and you look for cytopathic effect, where 9 you can say maybe you have got a virus that doesn't 10 cause a cytopathic effect. You put it into mice, you 11 put it into eggs, you put it into everything. So it's 12 like evidence of actual growth. You see? 13 I think there's actually quite a lot of 14 effort that goes into trying to detect viruses that 15 you don't know are there, but you suspect might 16 actually grow in some system that you are going to 17 check it on. If they don't grow, I'm not sure certain 18 that you are worried about them, or even if they are 19 viruses. So what do you actually mean by an unknown 20 virus? 21 DR. ONIONS: Well, okay. Let me give you 22 an example. I think one of the areas that is going to 23 emerge as a concern are going to be circaviruses. The 24 reason I say that is that TGV may be a circavirus. We 25 now know of others. Thomas was talking about the kind 163 1 of normal flora of HPVs. It looks like we all have a 2 normal flora of these circaviruses. Certainly when we 3 started now looking in animals, we find these all over 4 the place. 5 So I think they may come up as potential 6 cell substrate contaminants, but also in our serum, in 7 our trypsin, and so on. Of course trypsin is full of 8 of course circoviruses. 9 What we know is if you look in a cell 10 substrate like PK-15, which has used in the pig 11 vaccine industry, that carries a circovirus genome. 12 It looks, it's possible semi-defective, but you can 13 introduce it as an infectious agent, but most of the 14 time it isn't inducible as an infectious agent. 15 So I think I'm not quite sure I take your 16 distinction. I think there are latent viruses there 17 that are reactivatable under certain circumstances and 18 are a concern. In fact, in the vetmurines vaccine, 19 you have to get rid of that virus. 20 DR. PALLEY: Yes. I just had sort of an 21 alternative. Virtually all the conversation has 22 revolved around using higher eukaryotic type cells as 23 the producer cell source. There is an alternative 24 that's being actively pursued in the papilloma vaccine 25 business, which is to make virus-like particles either 164 1 with a baculo virus expression system in insect cells 2 or even in E. coli. 3 Very briefly, what's done is the L-1 and 4 L-2 capsid proteins of papilloma virus self assemble 5 no matter where they are over produced. Four 6 different groups are pursuing these now. I am sure 7 they are engaged in some corporate relations. 8 But one of the neat strategies that John 9 Shiller here at NIH and others have done, is to fuse 10 peptide epitopes or even intact other genes to the L-2 11 protein at their n-terminus. It turns out that when 12 L-2 assembles into the virus-like particle, but brings 13 this n-terminal protein in with it, that other protein 14 is on the inside of the virus-like particle. So it 15 will ultimately be presented to the recipient of this 16 virus infection. It's non-genetic, but it will be 17 taken up by cells. So there is in effect no risk of 18 some bacteria phage infection running rampant in our 19 bodies. So it's a complete alternative. 20 DR. ONIONS: I think obviously there is a 21 move to sub-unit vaccines in this form of vaccines. 22 But I think it's still going to be a long time before 23 our traditional eukaryotic cell production is going to 24 be lost. But I take your point. That is a valid 25 point. 165 1 Could I just ask perhaps questions about 2 we had examples indeed from Thomas toady about using 3 redundant PCR techniques. It would be feasible to 4 screen for a number of viruses that are of concern, 5 like herpes viruses, circoviruses, polyoma viruses, 6 herpes viruses retro viruses by redundant PCR 7 techniques. These are not in FDA terms, validated 8 techniques. On the other hand, they are very powerful 9 techniques. For instance, one of my colleagues just 10 developed a herpes virus redundant PCR technique with 11 112 primer combinations that to date has picked up all 12 of the herpes viruses that's been challenged with both 13 human and animal origin. So you could go looking for 14 herpes viruses by that kind of technique. 15 Robin Weiss' group in London picked up a 16 new human herpes virus, HRV-5 by a redundant PCR 17 technique. So do any of you think that these kinds of 18 technologies should be implied to cell substrates? If 19 so, which viruses should we use? Because it is a lot 20 of work to do this kind of thing. So should we be 21 doing this or is it not necessary? Can I have 22 comments? 23 Jim, you look like you are about to say 24 something, but not quite sure. 25 DR. ROBERTSON: One would not expect to 166 1 see this type of assay risen up in a pharmacopeial 2 recommendation of any kind. But certainly they do 3 have uses at the investigational level, especially 4 with some incredibly novel cell types. 5 Everything is kind of going molecular 6 these days. All these assays are looking at things 7 from the very molecular point of view, and picking out 8 a signal say with your herpes primers, need not 9 necessarily say that you have got an infectious herpes 10 virus. 11 So go back to this argument we have been 12 having this morning about it's infectivity which 13 potentially is what we're concerned about, and not 14 picking up a fragment of a genome. 15 These are assays, again I mentioned that 16 Joerg mentioned them this morning, broadly reactive 17 molecular assays. Phil rightly pointed out that we 18 have had in place for eons broadly reactive 19 infectivity assays using tissue cultures, suckling 20 mice eggs, looking for signs of infectivity of any 21 kind. 22 DR. ONIONS: But don't you think -- my 23 feeling about that is, I tend to share, I don't think 24 one would ever use these kind of assays on a routine 25 basis. 167 1 DR. ROBERTSON: No. 2 DR. ONIONS: They might be useful in 3 establishment of a master cell bank or something, you 4 know, the first one. But I mean I would criticize, I 5 don't think that current infective assays do pick up 6 everything. I think that's the whole problem. I 7 think, for instance, that it would miss -- well, 8 polyem virus has perhaps been used. 9 DR. ROBERTSON: At the end of the day, 10 that's the weakness of anything, that you will not 11 pick up something that's not designed to pick up. You 12 will potentially miss viruses in an infectivity assay. 13 But you also potentially miss viruses with a redundant 14 PCR. If you don't pick something up, you can't say 15 it's there or not there. It's a bit of a 16 philosophical argument. 17 DR. ONIONS: That's why I'm just saying 18 shouldn't you have an adjunct to these? I mean I 19 think should you not have at least an adjunct in terms 20 of broadening the kinds of assay systems that you are 21 using? 22 DR. ROBERTSON: Sure. Oh yes. 23 DR. ONIONS: Does it have anything to do 24 with the -- I mean there are people out there who have 25 to do this for a living, rather than us who can just 168 1 sit here and pontificate about it. How does this go 2 down with the industry? What does the industry feel? 3 DR. PALLEY: Just one point regarding 4 again HERVs. 5 DR. ONIONS: Sorry? 6 DR. PALLEY: I mentioned in my talk that 7 HERV-W family that has been reported for the first 8 time this year and has been isolated from vitro by 9 particles from multiple sclerosis patients. It turned 10 out that it at least codes for an env gene. So it's 11 certainly worth -- the human genome I guess is among 12 the genomes that regarding endogenous retroviruses is 13 among the best characterized genomes besides mouse, 14 for instance. 15 I think it is certainly worth to continue 16 and even by such redundant PCR approaches and so on to 17 further characterize HERV sequences on endogenous 18 retrovirus sequences, and to see, to give then an 19 estimate whether there are any additional sequences 20 that could be harmful. So far, we did not find any 21 sequences, but it's certainly worth doing that. 22 AUDIENCE MEMBER: I think if you just kind 23 of look back at the history of biological products, 24 and maybe even going way back to when hepatitis B was 25 discovered, and then there became the ability to look 169 1 for hepatitis B in blood and blood products. There is 2 always concern about what are you going to find, and 3 what are the implications, and what are the costs, and 4 all of those things. 5 But the bottom line is as technology 6 evolved, then the discovery of reverse transcriptase, 7 and when I was still here at the agency, we applied in 8 a research setting to vaccines and first demonstrated 9 RT in yellow fever. You get concerned about again, 10 what are you going to find when you look at all these 11 cell substrates, and then with the more enhanced 12 sensitivities of these systems. You are always in the 13 same muddle. That is, is this appropriate to apply 14 across the board? Where should it be properly 15 applied? Should it be done more in the initial stages 16 of characterizing something versus a routine quality 17 control test? 18 Those things, you can't sit here in a 19 meeting or on a panel and give specific answers to 20 those questions. I think the bottom line is as 21 technology evolves, and it's going to continue to 22 evolve with more sensitivity and specificity 23 hopefully, it needs to be explored and it will find 24 its appropriate place in the overall testing, whether 25 it's in characterization or perhaps in some cases, if 170 1 it's appropriate, on a routine basis. 2 But I think the general principle of 3 applying new technology as it begins to be available 4 to look at these issues, particularly as they relate 5 to safety and the presence of adventitious agents, is 6 unassailable. 7 DR. ONIONS: I agree. Can we move on, 8 because I would like to just cover TSEs just before we 9 have lunch time. I'm desperate for some lunch. 10 If you have got something, sorry to 11 inhibit you. If you can be brief. 12 AUDIENCE MEMBER: I agree completely with 13 what John says, but would add that if we are faced 14 with a decision of whether to approve the use of 15 different types of cell substrates that are 16 tumorigenic or derived from tumors from which we don't 17 know the mechanism of transformation, we are faced 18 with not only the question of should technology be 19 applied, but is the technology as it exists today and 20 can be applied today, good enough to permit us to say 21 that it's okay to use these cells. So that is a very 22 practical question which perhaps could be answered. 23 DR. ONIONS: Can we have a view on that? 24 That is, given the technology we have today, is it 25 acceptable to use the kinds of substrates we have been 171 1 talking about, that is tumorigenic or immortalized 2 cell substrates? Are we confident that with the 3 technology we have, that we can use these cell 4 substrates safely? 5 DR. BROKER: I would just basically say I 6 think so. I think if we combine PCR with these 7 microchip or microprobe arrays, DNA chip technologies, 8 and we have the growing human genome base and the 9 analogs and a number of other species, I think we have 10 got the tools at a level of sensitivity far beyond 11 what would probably would be more than adequate. 12 DR. ONIONS: Okay. I would just like to 13 finish up, because we heard a really I think important 14 interesting talk from Neil. While perhaps the risk of 15 spongiform encephalopathies in the kind of cell 16 substrates we are concerned with is probably extremely 17 remote, the consequences of being wrong about this 18 issue are potentially devastating. So it is certainly 19 worth cautious consideration. 20 Really I think Neil in his talk, already 21 summed up these key issues about the possible origins, 22 are the mutations spontaneous or infection, and the 23 kinds of cell substrates of concern might be, it seems 24 to me, are the neuronal cells. Since it's recently 25 shown that in the peripheral introduction of TSEs, the 172 1 B-lymphocyte might be important to carriage, then 2 perhaps lymphoid cells, particularly B cells that are 3 invariably used, and since we're looking at the 4 possibility of using HIV and T cells, maybe that 5 suddenly becomes an issue. Maybe we should be looking 6 at lymphoid cells for the potential of there being 7 spongiform encephalopathies. 8 Which brings you back to the question that 9 Neil finished with. That is, what should we do? It 10 did strike me that one of the possibilities was that 11 there are now very good, very interesting new cell 12 lines being based on retinal cells, which we heard 13 from Dr. van der Eb and others, which look very, very 14 promising for the generation of anti-viral vectors. 15 But as they have, as I understand it in 16 theory, I mean just at the simplistic level, should 17 not one thing be done and just sequence the PrP gene 18 in that? The probability of having a key mutation 19 seems to me extraordinarily remote. But then it's a 20 relatively simple thing, cheap thing to do is to go 21 and sequence the PrP gene. Is that something that we 22 should do in that kind of a situation? Should we also 23 do that in T lymphocytes? It's a trivial thing to do? 24 DR. CASHMAN: I would say in this case, 25 you get what you pay for. It is a trivial experiment, 173 1 but if one out of a million or one out of a thousand 2 cells could be harboring a mutant prion protein gene, 3 the technology is a little more dicey. So yes. 4 DR. ONIONS: I was thinking of excluding 5 the origin of the familial form, which clearly occurs 6 to some inherited disorders. I mean I think you are 7 right. You can't cope with a somatic mutation. 8 DR. CASHMAN: Okay. You can't cope with 9 somatic mutation, I agree. But certainly one can cope 10 with a mutation that's in every cell, yes. 11 DR. ONIONS: Neil touched on validation 12 technology. I think that's important. I think there 13 are new techniques for doing validation of TSE 14 removal, but using a disrupted PrP protein. But I 15 don't think that's going to be applicable to quite a 16 lot of the processes that are used to produce vaccines 17 at the moment. It is by technology products, but not 18 to vaccines. 19 I just wanted to touch on testing because 20 I know that Neil has an interest in that area. It 21 seems to me that we were sort of rather optimistic a 22 couple of years ago, and indeed, there have been 23 publications by Bruno Esch and others on specific 24 antisera of PrP Sc. But those haven't held up. They 25 actually pick up aggregated protein and not, as I 174 1 understand it, strictly PrP Sc. 2 We can use, and we have been using, 3 treatment of protease followed by immuno blotting. It 4 certainly works, but it isn't that sensitive. The 5 problem, it seems to me, is that we really don't have 6 a specific test that's an in vitro test. The only 7 thing that you are left with, at the moment, it may 8 change, but at the moment is animal inoculation. 9 Would you like to comment? 10 DR. CASHMAN: I think that the bio assay, 11 with no species barrier, can detect one unit of 12 infectivity. It's the most sensitive thing we have to 13 date. But in fact, there is new technology. Mary Jo 14 Schmirr and her colleagues have developed a very 15 sensitive capillary electrophoresis technique which 16 is, if one can believe the papers, including one in 17 press, is as sensitive as bio assay. So technology is 18 evolving, and there may in fact be a specific and 19 sensitive test for PrP Sc right around the corner. 20 DR. ONIONS: Okay. On that optimistic 21 note, I am going to wrap the session up. I think 22 we're all getting edgy for lunch. Unless anyone has 23 got some burning issue that they wish to go into 24 before we go. 25 If not, I think the panel members, and 175 1 thank you all for participating. 2 CHAIRPERSON RABINOVICH: Thank you, Dr. 3 Onions. There is a light repast outside for those 4 that have been so patient. I would like to get 5 everybody back in here in 15 minutes. 6 For those of you who would like to avoid 7 the wholesale garage sale that's going to go with your 8 luggage, for those of you that haven't checked out of 9 your room, I encourage you to complete that now. 10 Thank you. 11 (Whereupon, the foregoing matter went off 12 the record at 12:33 p.m. and went back on 13 the record at 12:53 p.m.) 14 CHAIRPERSON MYERS: Back to order. I 15 would like to introduce a co-chair and a new person we 16 are very pleased to have attending the meeting, Dr. 17 Gary Nabel, who is the new Director of the Vaccine 18 Research Center at NIH. So he is going to join me in 19 chairing this session. I am probably going to 20 disappear before the end of the session to make a 21 plane. 22 The first night I got here, as some will 23 recall, I came in a little late. I ran into a couple 24 of you in getting a beer because it was after the time 25 of the close of the meeting. The discussion ensued as 176 1 to what is a designer cell substrate. What do we mean 2 by that? My first reaction to that of course was it's 3 anything that I happened to have made. Clearly at 4 this point in the meeting, it's not primary cells. 5 I suppose from a strictly semantic 6 perspective, it would be a cell substrate created with 7 specific characteristics. It could be immortal or 8 not. But I think over the last couple of days, at 9 least my thinking on this and for the purposes of this 10 discussion, by a designer cell substrate, we mean a 11 cell substrate of defined origin and with a defined 12 pedigree. 13 It is probably immortalized because it is 14 likely to have been cloned. It will be validated as 15 specific pathogen-free and at least specific pathogen 16 sought and perhaps in certain circumstances, defined 17 as non-infectious. 18 For the purposes of the next discussion, 19 we are really talking about immortalization. Jim 20 McDougall, as you know, presented his paper yesterday. 21 So we'll start this session with the first 22 paper by Dr. John Sedivy from Brown University, who 23 will talk about differences in the capacity to 24 immortalize rodent, primate, and human cells by tissue 25 culture passage or viral transformation. 177 1 DR. SEDIVY: Thanks very much for the 2 invitation. I am sorry that Jim gave his talk 3 yesterday because -- well, maybe it will jive all 4 together. 5 I was asked to give somewhat of a 6 historical overview on the issues of replicated 7 cellular senescence, and obviously the topic of 8 cellular immortalization. 9 So from a historical point of view then, 10 this is the Hayflick phenomenon. This experiment has 11 been performed in numerous labs and always with the 12 same result. This happens to be an experiment in my 13 lab. You will see a number of slides like this from 14 me today. What we're plotting here is replicated 15 lifespan, the doublings of the culture versus days. 16 We see a culture growing and then reaching a non- 17 proliferative plateau. This is what we define as 18 senescence. 19 Really the interesting point here is that 20 the correlation here of this plateau is with the 21 number of cell divisions as opposed to chronological 22 time. The question that has been plaguing this field 23 ever since its inception is well, is this really some 24 type of a terrible artifact. I don't really want to 25 get into this discussion. It really revolves around 178 1 the issue of media and media artifacts, and have these 2 really been adequately resolved today. I don't think 3 they have, especially for some more specialized cell 4 types. I think they have been pretty well resolved 5 for keratinocytes, maybe breast epofelial cells, 6 fibroblasts, et cetera. 7 One really has to keep in mind that if one 8 sees a culture that is slowly declining in its 9 proliferation, this could simply mean that 10 increasingly a larger and larger fraction of those 11 cells are withdrawing from the cell cycle. This could 12 be perfectly explained by inadequate culture 13 conditions, such that eventually on the macroscopic 14 scale, the culture has ceased proliferating. 15 There are really three arguments that have 16 been used historically to justify the claim that 17 replicated senescence is a biologically interesting 18 phenomena. Here we are plotting, again very simply, 19 the mean-like span of a species versus fibroblasts 20 replicated life span in tissue culture. As you can 21 see, there is a rather striking correlation, such that 22 animals that don't live for very long don't have cells 23 that live for very long in tissue culture. 24 The next phenomenon that one often sees 25 cited is the age of the donor plotted against -- here 179 1 is the age of the donor, and the remaining life span 2 of the cells, in this case fibroblasts taken from that 3 donor. As you can see, the points are all over the 4 place. In fact, more recently, this view has been 5 challenged by a recent paper in PNAS from Vince 6 Cristofalos, who actually claims that this correlation 7 doesn't exist. But if you read the literature, you 8 will see this coming up over and over again. 9 The one fact that seems to remain, at 10 least to my knowledge, and that is if you look at 11 these points down here, these are fibroblasts taken 12 from individuals that suffer from premature aging 13 syndromes. These are called progerias. Typically, 14 these cells have a very short life span. 15 So this really is the issue here. How do 16 we differentiate between senescence, quiescence, and 17 differentiation. I think that for the purpose of 18 discussion today, this is really not a point of major 19 interest, but for historical reasons I'll go through 20 it rather quickly. 21 Quiescence is defined as a reversible 22 process. So what we are talking about here is 23 essentially a cell cycle phenomenon. That is, we can 24 have a culture that is cycling or contains a large 25 fraction of cycling cells. Then these cells can 180 1 withdraw into the quiescence state. Then when they 2 are induced with the proper growth factors, and here 3 of course the key phrase is what are the proper growth 4 factors to elicit this phenomenon. At any rate, we 5 are talking about a reversible process. Whereas 6 senescence by definition is irreversible. 7 So then of course the very interesting 8 next question is how do we differentiate senescence 9 from terminal differentiation. I don't really have 10 answers here because in many cases, this is very 11 difficult to do in many specialized cell types. What 12 one would like to see in general is the absence of 13 features that are characteristic for terminally 14 differentiated cells. But this is not possible in 15 many cases. 16 So really this has given the impetus to a 17 search for molecular events. So then if we pose the 18 question are there molecular events that are unique to 19 senescence versus quiescence versus differentiation, 20 again, the picture is not very clear cut. I don't 21 want you to absorb this whole slide. Suffice it to 22 say that this is well, not all, but the major part of 23 the regulatory circuitry in G-1. Here you see the D- 24 type cyclance. CDK-4 and CDK-6 driving RB 25 phosphorylation, which in turn drives the second 181 1 phase, which is cyclin E production, activation of 2 CDK-2. Of course there are a lot of modifying 3 proteins here, CDK inhibitors, kineses that activate 4 the basal CDK kinase, et cetera. 5 Now this is an area that is receiving a 6 lot of attention. The general theme, at least to me, 7 it seems that there's a high degree of overlap between 8 mechanisms that regulate quiescence, senescence, and 9 differentiation. I don't think this is really 10 surprising because all these three states are 11 characterized by the absence of cell cycle 12 progression. In most cases, by an arrest in the G-1 13 or a G-0 state. 14 The one central theme is that the 15 regulation of cyclin dependent kinase activity is 16 necessary to achieve a physiological cell cycle 17 arrest. In addition to the cyclins, which are the 18 positive affecters, there is a number of CDK kinase 19 inhibitors that have been shown to play a key role. 20 The two major inhibitory pathways that act 21 on this basal cell cycle machinery are the RB pathway, 22 shown here, and also the pathway regulated by the 23 tumor-suppressor protein P53. In both of these 24 pathways, CDK kinase inhibitors have been shown to 25 play key roles. 182 1 So let me turn to the issue of 2 immortalization. We all know that senescence can be 3 overcome because quite obviously, there are many cell 4 lines out there that are very immortal. So in a very 5 simplistic and general sense, we can think of cell 6 culture in three broad categories. We can have 7 primary cells or cell strains that have a limited life 8 span and senescence after several passages. We have 9 a category of cell lines that are immortal, not 10 necessarily by the 3T3 protocol, but in general, they 11 display the characteristics of unlimited lifespan, 12 non-malignant phenotype, and in most cases by the 13 ability to become quiescent. 14 Finally, we have the large group of cell 15 lines that are derived from either tumors or have been 16 transformed by one process or another. These of 17 course also have an unlimited lifespan, but they have 18 a malignant phenotype as defined by one or more 19 criteria. They also usually cannot become quiescent. 20 This again is the Hayflick plot. What I 21 am showing here is a rodent culture, mouse in this 22 case, and human. This little bump on the curve in 23 fact is senescence for a mouse fibroblast culture. So 24 it's been known for a very long time that rodent cells 25 can overcome senescence spontaneously. 183 1 You can also see the great difference 2 between the replicated lifespan in vitro of human 3 cells that go on for a very long time. If this 4 experiment here was continued, it would level off and 5 you would see the typical Hayflick phenomenon. So the 6 human plateau up here in fact is corresponding to this 7 rather short plateau senescence in rodent cells. 8 So the relatively low frequency of 9 immortalization -- I should point out that this 10 doesn't really seem like a low frequency, but on a per 11 cell basis, it actually is an event that has a 12 frequency of 10 to the minus 5, to 10 to the minus 6. 13 It's just that the X axis is plotted in days here. 14 The fact that this immortalization can be 15 stimulated by mutagens has led to the hypothesis that 16 this in fact is a mutational event in nature. This is 17 supported by the existence of several viral genes, 18 such as, and we have heard about them here, SV40 large 19 T antigen, polyoma large T antigen, animal virus E1A, 20 HPV E6 and E7, that can cause immortalization. In 21 fact, when these genes are introduced into rodent 22 cells, they are sufficient to cause immortalization in 23 a single step. In other words, if you take a rodent 24 culture and you put SV40 large T into those cells at 25 this point, the curve would look like this. No 184 1 apparent senescence under the right culture 2 conditions. 3 So what are these viral oncogenes doing to 4 promote immortalization? Without going into a lot of 5 detail, there is a large body of evidence that now 6 indicates that these proteins interfere with the 7 function of the P53 and/or RB growth inhibitory 8 pathways. In agreement, there's a lot of data from 9 knock-out mice now recently that has shown that the 10 elimination by gene knock-out of a variety of 11 negatively acting affecters can result in apparent 12 one-step immortalization, as shown here for example. 13 To date, embryo fibroblasts from strains 14 deleted for P53, P16 inc 4A, P19 arf 1 in P21 cip 1 15 have displayed this apparent immortalization 16 phenotype. 17 So what happens in human cells? Normal 18 human cells have never been observed to spontaneously 19 immortalize. Senescent cultures do not give rise to 20 sub-populations that resume proliferation as shown 21 here. Treatment with mutagens has been shown to 22 sporadically give rise to immortalized derivatives, 23 but the frequency of these events is significantly 24 lower than that in rodent cells. 25 Let me now talk a little bit about the 185 1 phenomenon of crisis. So what happens when we put, 2 for example, SV40 large T or E1A into a human 3 fibroblast? What we get instead of immortalization, 4 is a phase of so-called extended lifespan. So here we 5 see a primary cell, the initial proliferative phase. 6 This is senescence or the Hayflick limit. The 7 introduction of a viral oncogene is going to cause an 8 extended lifespan for variable duration, typically in 9 human fibroblasts of 20 to 30 divisions. Then one 10 sees a second proliferative decline. This has been 11 designated as crisis. 12 Now this decline at the end of this 13 extended lifespan which we call crisis, this word is 14 somewhat ambiguous, because it has also been applied 15 to rodent cells. These cells do not display a two- 16 stage mortality process. So to distinguish more 17 clearly between senescence and crisis, some groups 18 have started to use the word "M1" for mortality stage 19 one, and "M2" for mortality stage two. 20 Senescence is different from crisis. 21 These are not just the same proliferative decline. 22 The main distinction is that cells in senescence or M1 23 are truly non-dividing. Whereas in crisis cultures, 24 the apparent absence of proliferation on the 25 macroscopic scale is actually the result of ongoing 186 1 cell division combined with ongoing cell death. 2 This is an experiment that was performed 3 in my lab. What we show here is that elimination of 4 the CDK inhibitor P21 in a pre-senescent normal human 5 fibroblast causes an apparent extension of lifespan 6 that is equivalent in magnitude to that elicited by 7 SV40 large T antigen. So also in human cells now we 8 have been able to do ablative intervention. That is 9 eliminate the activity of certain negatively acting 10 affecters and cause an apparent extension of lifespan. 11 In terms of cell substrate design or the 12 technology that would go into doing this, this was 13 really strictly an aside, we have now developed 14 methods -- these are really based on gene knockouts, 15 homologies reculmination gene targeting, that can be 16 used to delete entire genes, multiple genes in human 17 cells, including normal human cells. 18 So let me now turn to my last topic, which 19 is the molecular clock of aging. I think probably 20 this is where I am going to overlap with what Jim has 21 already said. As I told you, there are some older 22 observations that correlated entry into senescence 23 with the lap cell division as opposed to chronological 24 time. Quite a few years ago, this has led to the 25 proposal for the existence of some sort of a molecular 187 1 clock. Then one envisioned that the running down of 2 this clock would generate a signal that triggered the 3 senescence program. 4 Then the expression, for example, SV40 5 large T could either prevent senescence by overriding 6 a signal from this clock or by what I think is more 7 likely now in light of new evidence, actually 8 interfering with the senescence machinery itself. 9 So as you know, the currently prevailing 10 hypothesis is that the nature of the molecular clock 11 is the attrition of telomeres. This is a slide by one 12 of my dear friends, Chris Counter, who has fancifully 13 imagined H-TERP, which is the catalytic sub-unit of 14 human telomerase sitting here at the end of a 15 chromosome end. So this is a telomere here. Then 16 catalyzing the addition of the telomere heximer. You 17 can see the telomerase RNA that acts as a template for 18 that process right there. 19 Germ cells and some key stem cells are 20 known to express telomerase catalytic activity while 21 the majority of somatic cells lack this activity. The 22 estimation of telomere shortening for one generation 23 is in human cells between 50 to 100 bay spares. So 24 that's 50 to 100 bay spares per S phase. This 25 correlates reasonably well with the average telomere 188 1 length in a young human fibroblast of 18 to 20 2 kilobases and the length of 8 to 10 kilobases in the 3 senescent fibroblast. 4 I think it's an important observation that 5 senescent cells in fact contain appreciable 6 telomerase. So here we have a normal cell or a young 7 cell. We get attribution of telomerase. At this 8 point, the telomerase are maybe 8 to 10 kilobases in 9 length. This generates a signal. If the cell is now 10 driven into the extended lifespan phase, these 11 telomeres will continue to erode because telomerase is 12 not expressed in that state. Eventually one enters 13 into a crisis which is caused by erosion at the end, 14 genetic instability, et cetera, et cetera. 15 It is really the nature of this signal 16 that I think is one of the enduring mysteries of the 17 field. One can really now beginning -- we can start 18 to see the process as being composed of a clock, a 19 signal, and then the senescence machinery itself which 20 is most likely composed of the same players, CDK 21 cyclin inhibitors, et cetera, et cetera, that are used 22 in other types of responses such as differentiation 23 and quiescence. 24 The linguistic definition of senescence is 25 the state of being or the process of becoming old. 189 1 This term has therefore been used to describe 2 essentially any sort of age-related irreversible 3 proliferative decline. In light of these new 4 molecular insights, I prefer to use senescence in the 5 more restrictive mechanistic sense to designate the 6 response triggered in normal cells. I really believe 7 that senescence is an active genetically programmed 8 process that responds to an inductive signal. Perhaps 9 telomere shortening, but that is not 100 percent 10 clear. 11 How the signal is generated is not really 12 well understood. One can argue that the ensuing 13 growth arrest has the obvious advantage of preventing 14 the cell from becoming grossly genetically unstable. 15 In contrast then, I think of crisis as an 16 unphysiological state. You have to do something to 17 the cell to drive it to this point, and that it leads 18 eventually to the catastrophic breakdown of chromosome 19 stability, which is caused by critical telomere 20 shortening on many chromosome ends. 21 So now this is really just a restatement 22 of the two-stage mortality process. What I have added 23 here now is telomere length in kilobases on the Y 24 axis, the replicative age on the X axis. So here we 25 have a cell in the beginning. If this happens to be 190 1 a germ cell or a stem cell, it will maintain telomeres 2 because it will express telomerase activity. 3 Most somatic cells will start down the 4 slippery slope of telomere attrition, eventually 5 entering into a physiological state of growth arrest, 6 through which they can be driven by either the 7 expression of certain viral oncogene or the eblation 8 of certain inhibitory pathways that are intrinsic to 9 those cells. 10 The cells then enter into extended 11 lifespan. They continue to erode telomeres. They 12 enter into a state of crisis, which is characterized 13 by genomic instability. Finally, at this point, one 14 can attain a truly immortalized derivative in the key 15 step here, is the expression of telomerase catalytic 16 activity. 17 I should also point out that telomerase 18 need not be expressed at the final step. It has been 19 shown experimentally that telomerase can be 20 artificially or experimentally activated anywhere 21 along this line, and that that will lead in some cell 22 types, not necessarily all cell types, to 23 immortalization. However, I think the large body of 24 evidence suggests that at least in vivo, and by this 25 I mean during the natural development of malignancy, 191 1 the activation of telomerase activity is a relatively 2 late step. 3 So if crisis doesn't exist in rodent 4 cells, and bypass of senescence is sufficient for 5 immortalization, how does telomerase become expressed 6 in somatic cells, rodent somatic cells? The bottom 7 line here seems to be that telomerase is not very 8 strictly regulated in rodent cells and tissue. A 9 variety of rodent tissues have been shown to express 10 telomerase activity. Telomerase negative primary 11 cultures often become telomerase positive over time 12 even prior to reaching senescence. 13 In contrast, telomerase appears to be 14 regulated very stringently in human cells. Therefore, 15 telomerase activation could occur in rodent cells that 16 are undergoing immortalization either prior to or 17 after the senescence bypassing event, and could easily 18 occur in the subtle and gradual fashion so that no 19 clearly apparent downturn in proliferative capacity of 20 the ball culture would be observed. 21 In other words, one step immortalization 22 that one sees so often in rodent cells may in fact 23 require two steps, the obvious step of senescence 24 bypass and very likely a second step that may be very 25 subtle, at least in rodent cultures. That is, of 192 1 activating telomerase catalytic activity. 2 So I think that is about as good a summary 3 as I can think of in 20 minutes. I will be glad to 4 entertain questions. 5 (Applause.) 6 AUDIENCE MEMBER: Bill Egan, from the FDA. 7 When you immortalize cells, you know, after they go 8 into crisis or whatever, what becomes the length of 9 the telomere? Does it go back up to 20 kilobases? 10 What maintains the length of that telomere at a fixed 11 -- 12 DR. SEDIVY: That's a very good question. 13 AUDIENCE MEMBER: Why doesn't it become 30 14 or 40 kilobases. 15 DR. SEDIVY: In fact, it seems that 16 excessive telomere length is not good, at least in 17 human cells. It's been known for a long time that 18 many spontaneously immortalized human cell lines which 19 we love and honor like 293 and Hela, et cetera, et 20 cetera, have very short telomeres. These telomeres 21 can be maintained at a length of one to two KB. These 22 cells seems to be perfectly happy with that. 23 So I think it's more the maintenance of 24 the telomere length rather than the absolute length of 25 the telomere. 193 1 If you artificially introduce telomerase 2 catalytic subunits into fibroblasts, what one 3 typically sees is that the best clones are ones that 4 build up telomere length to about 8 to 10, 12 KB and 5 then maintain it at that level. It seems to be a 6 function of the expression level of the H-TERP gene, 7 because if one does this experiment, you see cultures 8 that very slowly erode their telomeres. They will 9 eventually senesce. 10 You see cultures that build up telomeres 11 to maybe 20, 30 kilobase in length. That doesn't seem 12 to be good for them because the rate of growth goes 13 down. So really the best cultures are the ones that 14 maintain at least in fibroblasts. So I think it's the 15 maintenance rather than the absolute length. 16 DR. HUGHES: Would you please comment on 17 Carol Greider's knockout mice? 18 DR. SEDIVY: Well, yes. I didn't get into 19 that at all because that's at least for the time being 20 -- there are some paradoxes here. Okay? The obvious 21 paradox is that mice have extremely long telomeres, 60 22 KB on average. This is the laboratory mouse. 23 Muskulorattus has perfectly normal telomeres, for 24 example. 25 In fact, this is really the other way 194 1 around because these cells live for a very short time 2 in tissue culture. When they undergo senescence, they 3 undergo senescence with telomeres that are 50 KB in 4 size as opposed to 60 KB in size. 5 So there are really two answers to that. 6 One answer is that if you look at individual telomere 7 ends, you will see that there's a certain degree of 8 heterogeneity in that in fact in all cells, including 9 mouse cells, you always see a minority fraction of 10 chromosomes that have very short telomeres. If the 11 signal is caused by a perfectly short telomere that 12 gives, for example, DNA damage-like signal, then you 13 only need one per cell to give the senescence signal. 14 So that's one explanation. I don't really know 15 whether it's correct. 16 The other explanation is that mouse cells 17 don't senesce. They neither undergo crisis or they 18 undergo senescence. In fact, that plateau that we are 19 seeing during the immortalization is not senescence. 20 It's a differentiation-like process. 21 There are people, including myself, that 22 given this kind of loose distinction between 23 quiescence, differentiation, and senescence, would 24 prefer to define senescence now as a process that is 25 triggered by telomere erosion. Obviously when a mouse 195 1 primary fibroblast culture undergoes senescence, it's 2 not doing it, probably not doing it because it's 3 receiving a telomere signal. 4 I don't know if that is -- does that make 5 sense? 6 DR. HUGHES: I had hoped you would comment 7 on the mice themselves. 8 DR. SEDIVY: The mice themselves? What do 9 you want to know about the mice themselves? They are 10 alive. 11 DR. HUGHES: Yes, I know. But in the 12 Greider experiment with telomerase knockout. 13 DR. SEDIVY: If you knock out telomerase 14 in mice, it takes six organismal generations to 15 observe lethality. Okay? What you see at each 16 generation is that the average telomere length. So 17 generation one, it's 50. Generation two, it's 40. 18 Generation three, it's 30. 19 If you take mouse and real fibroblasts at 20 any one of these generations, they senesce in vitro on 21 schedule. Is that what you wanted? 22 DR. HUGHES: (Inaudible.) 23 DR. SEDIVY: Well, I think I just offered 24 you one explanation for that. That is that what we 25 are calling senescence is not senescence. It's 196 1 something caused by some insufficiency in the median 2 that is in fact triggering a differentiation event. 3 Actually, Jim McDougall and I also don't quite agree 4 about what's happening in his keratinocyte cultures 5 because what he is calling senescence, some 6 keratinocyte biologists would prefer to call 7 differentiation. 8 DR. NABEL: Okay. If we could move the 9 questions along, maybe brief answers. Then we'll move 10 onto the next speaker. 11 AUDIENCE MEMBER: Alex van der Eb, Leiden. 12 You just already answered, I think, my question, which 13 was why do mice cells, mouse cells enter senescence 14 while they have such long telomeres? In fact, you 15 answered already part of that question. 16 Do these cells that enter a so-called 17 senescence have high levels of P21 or P16 or something 18 like that? 19 DR. SEDIVY: Yes, they do. Yes, they do. 20 AUDIENCE MEMBER: So there is a signal 21 then. 22 DR. SEDIVY: If you take a knockout mouse 23 for P21 that doesn't undergo senescence. It just 24 keeps going. But you know, that's what I was trying 25 to say. That is that op regulation of P21 is not a 197 1 molecular market for senescence. P21 is op regular 2 because of oxidated stress, osmotic stress, 3 differentiation signals. This is a very general 4 machinery that is used to establish cell cycleresce. 5 I, in fact, don't know of any molecular 6 marker that is specific for senescence. This includes 7 the famous senescence-specific betagalactocytis 8 activity. You know, you see a lot of people staining 9 cells, and they turn blue and they say it's 10 senescence. Everybody knows if you put hydrogen 11 peroxide on your cells, they turn blue as well. 12 AUDIENCE MEMBER: Just a brief comment for 13 those people who might be setting up assays that would 14 be monitoring P21 sip. We, as I showed, found that 15 P21 is elevated in a subset of the HPB infected cells. 16 We did three other related assays. One was to look 17 for P21 MRNA. It turns out it's abundant in all 18 differentiated cells. But there is a post- 19 translational control on the accumulation of P21. 20 It turns out what happens is that if there 21 is not a signal that unscheduled in a synthesis is 22 underway, namely, abundant cyclin E, then proteosomes 23 rapidly degrade the P21 that's translated. When we 24 put in proteosome inhibitors, P21 piled up in all 25 cells and all replication was blocked. 198 1 We went on to ask one additional question. 2 That is, how does P21 actually block S phase or DNA 3 synthesis. Unexpectantly, it had nothing to do with 4 blocking cyclin E activity. It turns out cyclin A, 5 CDK-2 or cyclin A CDC-2, can phosphorylate DNA 6 preliminary cell and all these other subunits I 7 showed. 8 The one thing cyclin A can't do is bind to 9 PCNA. But when the P21 sip piles up in these cells 10 that have excessive cyclin E, the way the P21 is 11 actually functioning is by binding to the PCNA and 12 blocking elongation, not initiation. 13 DR. LEWIS: This may be a naive question. 14 Is there any change in the activity of endogenous 15 oncogenes in cells at about the time they are entering 16 into senescence, especially rodent cells? 17 DR. SEDIVY: By activity, you mean level 18 of expression? I am not aware of that. 19 AUDIENCE MEMBER: What happens in 20 spragues? I mean are they different than 21 muskulorattus? Do they have shorter telomeres that 22 they go through? 23 DR. SEDIVY: What happens in spragues is 24 exactly the same that happens in muskulor. 25 AUDIENCE MEMBER: So even though one has 199 1 60 KB and one has 2 KB? 2 DR. SEDIVY: I mean, you know, this kind 3 of all argues that the length of telomeres has nothing 4 to do with this plateau in mouse cells that we define 5 as senescence. Okay? I think there's a result that's 6 kind of floating around, which is also consistent with 7 that. That is, we all know the wonderful experiment 8 of expressing telomerase in human fibroblasts which 9 causes immortalization. It doesn't do that in mouse 10 cells, which also argues that the length of telomeres 11 in mouse cells is not what is triggering this growth 12 DR. NABEL: John, I am going to just end 13 with one last question. You referred to the notion of 14 program of senescence and it being dominant. I am 15 just wondering, has anyone ever done a cell fusion 16 experiment where you have taken cells approaching 17 their limit and then fused to neo-natal cells. Is it 18 in fact dominant? 19 DR. SEDIVY: Yes. Those are very old and 20 classical experiments. In general of course, 21 senescence is a dominant state. 22 DR. NABEL: Thanks. The last talk in this 23 session is from Frits Fallaux. The title of his talk 24 is using defined adenoviral genes and primary human 25 cells for the generation of immortalized cell 200 1 substrates. 2 DR. FALLAUX: First of all, I would like 3 to thank your organization for inviting me here. The 4 subject of my talk will be on the generation and 5 characterization of new helper cell lines for the 6 construction, provocation, and protection of 7 recominance replication effective adenoviral vectors. 8 In the past few years, the interest in 9 vectors derived from human viruses. This is caused by 10 the fact that from the many years of intensive 11 fundamental research on human adenoviruses, it has 12 been found that adenoviruses have several favorable 13 characteristics, including high stability of variance. 14 The variance is very easy to grow into pure with very 15 high fibers. It has a very broad host range. 16 Importantly, it has the capacity to transduce non 17 milotic cells. This makes adenovirus a very potent 18 gene therapy. 19 It is known that it has very low kinisity, 20 and there is there ample experience with adenoviruses 21 as vaccines. 22 This slide shows a schematic 23 representation of the adenoviral genome. It is a 24 double stranded linear DNA molecular of approximately 25 36 KD, carrying several genes, flanked by inverted 201 1 herminal repeats. The genes are sub-divided in so- 2 called early genes and in late genes, depending 3 whether they are stressed early or late during the 4 lytic infection. 5 This slide shows you a scheme of the 6 classical methods to construct the common 7 adenoviruses. All currently used adenoviruses carry 8 a deletion in E-1. This renders the virus replication 9 effective, and it also provides space to insert 10 therapeutic genes. 11 Now in the old days, we used to isolate 12 the DNA from wild type adenoviruses at 5 or S-2. In 13 purified DNA, and I just -- the restriction enzyme 14 cla-1, which puts ones in area region one. We then 15 purify the large fragment. 16 In addition, it also needs the 17 construction of an adaptor plasmid which carries the 18 transcriptase unit, including geno-fenchfras, but also 19 the left inverted herminal repeat and a part of the 20 adenovirus sequence which is also present in the large 21 fragment. Pro-construction of these two moleculars in 22 so-called helper cells, and the helper cell is the 293 23 cell made by Frank Reim. Upon close inspection, 24 another mination occurs, creating now the recombinant 25 adenovirus, carrying the gene of interest at the 202 1 position that we want. 2 You can proficate these elongated viruses 3 due to the fact that the helper cells complement the 4 missing elong function. 5 Now despite the encouraging results of 6 things so far with the use of recombinant 7 adenoviruses, there are also several problems 8 associated with the use of such vectors. These 9 problems include the growth infectivity range. That 10 is, you do not only infect the target cells, but also 11 non-target cells. This may cause pathogenicity. 12 Also, the viruses are rather immunogenic. 13 We only leave off E-1, and all the other viral genes 14 are still present and can be expressed to low levels, 15 resulting in numerous responses by the host, both 16 humoral responses, antibodies, and cellular responses 17 against new cells. The cells are killed and the 18 therapeutic effect is lost within several weeks. 19 Another issue is the occurrence of 20 replication competent adenovirus, abbreviated RCA. I 21 will focus on this topic. There are various sources 22 of replication competent adenoviruses. In a sort of 23 infection, during the production of viral, or as an 24 earlier stage, or during the construction of the 25 recombinant factor, especially when you use classical 203 1 methods, if you use the large clavon fragments. If 2 the digestion if not complete, you have RCA, namely 3 the wild virus. 4 It has also been shown recently that you 5 can generate RCA by homologous recrimination because 6 the factor and the helper cell carry adenovirus 7 sequences that overlap. As a result, by homologous 8 recrimination, you can get RCA. I will focus on this 9 source of RCA. 10 Well, how does it work, homologous 11 recrimination resulting in RCA? This is a scheme of 12 a typical elongated factor. This is a scheme of the 13 integrated adenovirus sequences in the helper cells. 14 The helper cell line is 293, and more 15 recently, we made alternative cell line 911. Both 16 helper cells carry the evon A, evon B in chorion 17 regions. But in addition, they also carry sequences 18 that enclose structural protein lines, downstream of 19 evon B, and upstream of evon A, be left for determinal 20 repeat. 21 Those sequences are also present in the 22 factors. So that a sequence overlap 5 prime and 3 23 prime of the therapeutic gene. As a result, you can 24 get homologous recrimination by which the recominance 25 virus now trades its therapeutic gene for region E-1, 204 1 and becomes replication competent. Now what you can 2 do about this is to avoid the sequence overlap. We'll 3 come back to that later. 4 Well there are only a few helper cell 5 lines available when you work with recombinant 6 adenoviruses. We are fortunately in our lab to have 7 three of them, including the two in our free cell 8 line, recently an iomosa line, and even more recently, 9 the PER cell line. All three cell lines are obviously 10 of human origin. They are all derived from primary 11 diploid embryonic cells. 293 is derived from kidney 12 cells, 911 from retinoblasts, and PER C-6 as well. 13 Now when I started to work with 14 recombinant adenoviruses in our lab, which is the lab 15 of Professor van der Eb at the Leiden University, I 16 used obviously 293 cells, and I met with some 17 technical difficulties. Since we had a panel of adeno 18 virus transformed human cells, including cells of 19 kidney, lung, and retinoblasts, I decided to screen a 20 panel of cells in order to find an alternative for 21 293. 22 From this panel of cells, I selected one 23 particular retinal cell line. We named it cell line 24 911. The reason for this name was to get the 25 attention of our colleagues in the U.S. 205 1 (Laughter.) 2 Well, the construct we used to make the 3 911 cell line is shown over here. It carries the 4 adenovirus sera type 5 nucleotides 87 to 5,788, 5 including evon A and evon B. 6 Now I want to be short on the 911 cell 7 line. The most important findings were that they 8 performed very good in virus titrations. We also 9 found that the virus use of 911 are up to three times 10 higher with various viruses, also recombinant viruses, 11 three times higher than obtained from 293. 12 Some other characteristics of this cell 13 line are that they express very high levels of evon A 14 and evon B, are highly transfectable, which is 15 important when you want to construct recombinant 16 viruses at the classical method. The use of the 17 viruses are very high, as told, and they perform very 18 well in titration assays. So we concluded that 911 is 19 a good alterative for 293. 20 However, I have shown you the construct we 21 used to generate the 911 cell line. We now have a 22 situation which is similar to 293. Namely, and also 23 in 911 cells, besides evon A and evon B encoding 24 sequences, also sequences of the left inverted termin 25 are repeat, and sequences in part encoding protein 9 206 1 are present. So again, there is overlap and you can 2 create RCA. 3 So what we decided to do is start all over 4 again and make now the cell line in combination with 5 a so-called matched vector, now sequence overlap. 6 What we did was to make a so-called packaging 7 construct carrying only the evon A and evon B encoding 8 sequences in which evon A is driven by PGK, a 9 heterologous promoter, and a heterologous poly and 10 signal, and lay matched vectors that are deleted of 11 exactly that elong region which is present in the 12 packaging construct. Thus affording sequence overlap 13 and thus eliminating homologous recrimination as a 14 source of RCA. 15 This shows you one of the packaging 16 constructs we constructed. Present are adenovirus 17 sera 5, sequences four, five line to 3,511. Those are 18 only the evon A and evon B encoding sequences. Evon 19 A is driven by the human PGK promoter. Evon B is 20 under its natural promoter, and directly flanking the 21 evon B stop codon as the polyadenylation signal 22 derived from hepatitis B virus. 23 Now before we decided to transfect in this 24 vection construct into our retinoblasts, we decided to 25 do some functional assays with this construct first. 207 1 We did this because first we only had a few frozen 2 ampoules of the retinoblasts. Second, the packaging 3 construct contains several PCR fragments. 4 This slide shows you some of the functions 5 or features of adenovirus evon A. The features are 6 that domains 1 and 2 are involved in the regulation of 7 expression of genes. Evon A is known to associate at 8 the protein level with cellular proteins, P-105 RB, 9 cyclin A, P-300. I think the list is growing. All 10 these different features result in the transformation 11 and immortalization. 12 Not shown is the feature or function of 13 evon B. Evon B prevents the cells from growing into 14 apoptosis as a result of the activities of evon A. 15 Now this is the actual functional assay we 16 performed. At the left is shown the constructs we 17 tested. At the right is shown the number of colonies 18 we obtained upon transfection of one or five microprin 19 of these plasmids into baby rat kidney cells. When we 20 transfected in a construct expressing only evon A, we 21 only found on average one focus or colony upon 22 transfection of one microgram. This is very low. 23 This is caused by the fact that expression of evon A 24 in primary cells, in the absence of evon B, is toxic, 25 causing apyltosis. 208 1 You can see this in the second line. 2 Transfected this plasmid to come in plasmid in evon B, 3 and we did obtain reasonable amounts of foci. 4 Obviously also the construct we used to 5 generate 911 resulted in focus formation. 6 Unfortunately, also when you are packaging construct, 7 which we used to make the PER cells gave foci, thus 8 indicating that the packaging construct allowed the 9 functional expression of early region non-probenes. 10 So then we went to the actual experiment, 11 transfected the packaging construct into primary human 12 diploid retinalblasts. We could establish seven 13 clonal cell lines. We tested these clones for first, 14 expression of evon A and evon B proteins. 15 Now we found that all clones expressed 16 very high levels of evon A in both 55 and 21 K evon B, 17 when compared to 293 and 911. We also looked at 18 vector use. We looked at three clones, clone 3, 5, 19 and 6. As you can see, we found that the three PER 20 clones tested exhibited similar use of recombinant 21 viruses, compared to 293 and 911. Since PER C6 played 22 the highest use, we decided to analyze this clone in 23 further detail. 24 Now of course the major issue for us was 25 to test whether or not our approach to use the PER 209 1 cells in combination with matched adenovirus vectors 2 would reduce or even eliminate the generation of RCA. 3 This testing has been performed at the enzyme. 4 What I did was to amplify an RCA free 5 master stock of a typical adenovirus vector and 6 amplify it to 293 or PER C6. What I found, I can 7 summarize it for you, is that amplification of 293 8 resulted in RCA positive vector batches in 9 approximately 50 percent of the places. 10 Now for a clinical setting, that means 11 that you might consider to throw away half of what you 12 had made. In the case of PER C6, fortunately in none 13 of the batches amplified on PER C6, we were able to 14 detect RCA, not even in a large scale production 15 setting. 16 So we concluded that our strategy to make 17 a PER cell in combination with new matched vectors 18 severely reduced, maybe even eliminated the RCA 19 problem, at least by homologous recrimination. 20 The next two slides summarize some of the 21 other features of PER C6 cells that contain three to 22 five copies of the packaging construct, very high 23 levels of evon A and evon B, comparable to 293 and 24 911. Good use of the different vectors, also similar 25 to the other two producer cells. 210 1 The cell line was very stable. We have 2 now come over passage 250 actually. So far we have 3 not detected RCA, and the list of productions with 4 different vectors is still increasing. 5 We have a master cell bank available for 6 PER C6, also a working cell bank. Importantly, the 7 PER cells were made on a GLP conditions, using 8 certified U.S. bovine serum and trypsin. Currently, 9 InterGene is doing all kinds of tests which were 10 necessary for the use in the chemical setting, 11 including mycoplasma and sterility testing. In the 12 academic lab, you can simply draw the cells to 13 standard medium. 14 Finally, I would like to thank all the 15 people that are involved in this project. Number one, 16 the PER cells were generated in the lab of Professor 17 van der Eb at Leiden University, in the Applied 18 Virology Group, supervised by Dr. Gugen. The 19 packaging constructs were made by Edie von Frel of 20 InterGene, and all the downstream processing, a lot of 21 work is currently being performed by InterGene, 22 supervised by Valeria in the Adenovirus group, 23 supervised by Dr. von Laud. As I told, all the RCA 24 testing was performed at Genzyme by Kathy Hay here. 25 Thank you. 211 1 (Applause.) 2 DR. NABEL: Thank you. We'll take 3 questions. 4 DR. FRIED: Mike Fried. Do these cells 5 form tumor in nude mice? 6 DR. FALLAUX: Actually we did not test 7 that yet. However, we did test this for the 911 cell 8 line, the weakly tumorogen in nude mice. So you might 9 expect the PER cells would exhibit the same feature 10 with respect to that. 11 DR. FRIED: What is weakly tumorogen? 12 DR. FALLAUX: Compared to some of the 13 let's say adenovirus sera type 12 transformed rodent 14 cells, they are much less tumorigenic. 15 DR. SHEETS: Becky Sheets, FDA. I had a 16 similar question, but I have a couple of other. Does 17 the PER cell stay diploid or is it aneuploid? Also, 18 the individual from whom you obtained the retina, did 19 they have wild type RB genes or were they -- you know, 20 was this someone that died of retinal blastoma? 21 DR. FALLAUX: For your first question, 22 they are no longer diploid. I don't know all the 23 details, but we do find chromosome duplications I 24 guess in PER C6 because we have tested recently. 25 Your second question. This may sound 212 1 silly, but we are currently trying to find out the 2 information of the donor. The cells were isolated in 3 the early 1980s, and we're now working backwards to 4 find out those details. 5 DR. NABEL: John? 6 AUDIENCE MEMBER: 293 cells have become 7 quite popular in the laboratory for reasons that have 8 little to do with their ability to support adenovirus 9 vector replication. Have you checked these other 10 cells, for example, high levels of transvectibility in 11 being a good host for other kinds of viruses and that 12 kind of thing that make 293 cells so beloved by many 13 virologists? 14 DR. FALLAUX: In fact, all adenovirus 15 transformed cells are as highly transvectible as 293 16 cells. 17 DR. NABEL: Okay. If there are no further 18 questions, then I think we can just proceed onto the 19 panel discussion. John Coffin will chair that. If 20 the panel members want to come forward and get 21 started. 22 DR. COFFIN: By my calculation, we are 23 running almost exactly an hour late. A check with the 24 board outside reveals that we really probably can not 25 go much past 3:00 before we start to lose people quite 213 1 seriously. So we probably should shoot for an hour in 2 which we either have a lot or a little to do, 3 depending on the will of the crowd and our host. 4 This panel discussion actually as two 5 functions. One is a discussion of the last topic 6 covered. That is the designer cell substrates, two 7 talks we heard today and the one yesterday. Then 8 secondly, where we really earn the generous honoraria 9 that FDA is paying us, where we try to summarize and 10 hopefully answer some questions that might be useful 11 to our host in terms of development of policy, ideas 12 for further meetings, experimentation and so on. 13 Tentatively we'll plan to sort of split 14 the discussion half and half between these issues, but 15 I think we can play that by ear as we go along. 16 Again, I expect widespread audience participation, 17 particularly since these are topics that I myself am 18 not really actively working in and am familiar with. 19 The questions that on the first part, on 20 the designer cell substrates, that we were charged to 21 address are summarized on this overhead. Before I 22 turn it on, I want to apologize in advance for two 23 things. One is my handwriting is very bad, so you are 24 going to be subjected to that for a while. Secondly, 25 I was given a rather blunt instrument to write with. 214 1 Thirdly, of course I'm not well enough organized like 2 some of the previous chairs who have prepared these 3 ahead of time. 4 So this is a paraphrase, I hope an 5 accurate one, of the issues that were raised, that are 6 raised in the points to consider. The first regarding 7 designer cell substrates is the issue of whether cells 8 that are derived by the kinds of defined means that we 9 have seen, and we have seen the example of 10 introduction of telomerase plus or minus oncogenes, 11 viral oncogenes, or viral oncogenes alone whose 12 function among other things was to stimulate or 13 inactivate genes that are involved in senescence. 14 Whether cells that are created in this way 15 in fact offer significant safety, create safety issues 16 relative to other cell lines, whether they offer 17 advantages or disadvantages, whether we can go through 18 the sort of defined risk algorithm that was given to 19 us at the beginning of the meeting, to address these 20 and anything else. 21 So if we could get onto the first point 22 here. Are there significant safety issues relative to 23 tumor or neoplastic cell lines? In other words, 24 uncharacterized, what we should call it perhaps, 25 uncharacterized cell lines. Cell lines that have just 215 1 been handed to us either in tumors or that have arisen 2 by means we don't really understand very much about in 3 culture. 4 Would anybody on the panel care to -- 5 DR. SEDIVY: Yes. I would just like to 6 make a very brief comment about designer cells. I was 7 asked to talk about the history of immortalization. 8 In fact, what my lab works on is more related to 9 interventions, genetic interventions in human cells. 10 So you know, obviously we have talked about putting an 11 H tert and putting in various viral oncogenes. So 12 really here the issue is can we make a cell line that 13 is immortal and it has a particular spectrum of 14 phenotypes that we want by absolutely defined genetic 15 interventions. I think the answer now is yes. We can 16 do a lot better than putting in SV-40 large T or E-1A 17 or E-7, because as have heard even today, we don't 18 really know exactly down to the last T what these 19 viral proteins are doing to the cell. 20 So in fact, what we can do is we can 21 delete cellular genes using gene knockouts, and 22 produce very much the same effect. That is, we can 23 really now contemplate really designing cells without 24 the use of viral oncogenes. I think that putting an 25 H tert is obviously a necessary step, but this is a 216 1 cellular gene. So I think that's probably okay. 2 DR. ONIONS: It's really a question of 3 clarification from my point of view. But it seems to 4 me that one of the advantages of cell lines that come 5 from potential oncogenic background is they have some 6 of the features from mass culture that industrity. 7 That is, they can be grown in an anchorage-independent 8 way. They can be grown in high density in fermenters. 9 What's the position with telomerase immortalized? 10 I assume actually they are mimicking much 11 more the kinds of cell type that Dr. Hayflick would 12 define as -- I forgot what he used, apologies -- the 13 first stage of cell strains. That probably do not 14 have those particular phenotypic properties. Do you 15 know what the stage of those cells are? 16 DR. SEDIVY: Well, you know, we have a 17 really limited experience. This game has only been 18 played for a few months, maybe a year in some 19 privileged labs. I think what we're really talking 20 about here is proof of principle. In my lab, we're 21 not interested in growing cells in fermenters. We are 22 interested in cell cycle progression. But I think if 23 somebody wanted to make a cell line that grows well in 24 fermenters, I think it may be a good idea to 25 contemplate some of these new approaches. 217 1 DR. ONIONS: Yes. The point I was trying 2 to make was that if you start down the road of trying 3 to produce new cells and immortalize them, and you try 4 and use procedures that you think are of in a sense, 5 the safest, whatever that means, those might not be 6 the steps that you actually need to actually produce 7 an industrially useable cell line regardless of the 8 importance of the science. The practical end may not 9 be what you want. 10 DR. FRIED: What else besides, you said 11 you would knock out genes? I mean would this be in 12 human cell lines or instead of using the viral 13 proteins? 14 DR. SEDIVY: I think in general, the first 15 thing you have to do is you have to immortalize the 16 cell. For that reason, we put an HH tert. That is a 17 technical feature that we need because to do the 18 knockouts, we have -- you know, there's limited time 19 to do a knockout. We can do two knockouts before a 20 human fibroblast undergoes senescence. But if we want 21 to do more, and we're obviously interested in doing a 22 lot more, it is very convenient to put tert in there 23 at that point. 24 If you don't want tert, you know, tert now 25 comes flanked with lock sites, so that you can take it 218 1 out later on if you are interested in that. 2 DR. COFFIN: Can you be more specific 3 about what you would knock out? 4 DR. SEDIVY: Pardon me? 5 DR. COFFIN: Want you to be more specific 6 about what you are knocking out. Which genes have you 7 knocked out? 8 DR. SEDIVY: Well what we are interested 9 in doing is we are interested in dissecting the 10 machinery that establishes senescence. So not in a 11 single cell line, but at this point, we have P-21, cip 12 1, RB, B-53, and P-16 ink 4A knockouts in various 13 combinations. So you know, the vectors are available. 14 One interesting point about human cells 15 that is interesting to the technocrats I think is that 16 until this time, we do not see the requirement for 17 isogeneic DNA. So in fact, these vectors that we have 18 made, and some other labs have contributed to this, 19 can be used and essentially -- and have been used in 20 any human cell with equivalent frequency of 21 recombination. 22 DR. HAYFLICK: Yes. We have a lot of 23 information about the biological properties of H tert 24 transformed normal human cells now. There are the 25 ones with which Choma is familiar, are now as I 219 1 indicated several days ago, approaching, and in one or 2 two cases beyond 400 PDLs. Many of them are beginning 3 to show some signs of aneuploidy. They are still 4 anchorage dependent. Their virus spectrum seems to be 5 identical to that with which we are familiar prior to 6 H tert transformation. Finally, the studies that have 7 been done on animal inoculation of these cells 8 indicates that they are not neoplastic. So that we do 9 know that much. 10 However, we are not talking about two 11 other classes of cell populations that are important 12 candidates to consider. I mentioned these in my 13 opening remarks on Tuesday night. Namely, normal 14 human diploid cells that can be immortalized using 15 repeated treatments of exposure to chemical 16 carcinogens. We published on this on a population 17 that that explanation defines about 25 years ago. 18 It's called SUSM-1. It's freely available. It is 19 from a normal human diploid fibroblast transformed by 20 multiple exposures to MMNG. 21 A second class of cells that also falls 22 into this category of normal human cell populations 23 transformed with something other than a virus, by way 24 of example, is KMSD-6 transformed by a former student 25 of mine, using multiple exposures to cobalt 60 220 1 radiation. That cell population was also described 2 and in fact is used commercially today. It was 3 described about 25 years ago. 4 I think it is very important to realize 5 that it's possible to immortalize well-characterized 6 normal human diploid cells like WI38 and MRC5 and 7 perhaps others, utilizing non-viral means. 8 I would also like to repeat what I 9 mentioned the other day because it's been restated 10 again, and it's not accurate. That is, that there are 11 spontaneously transformed normal human cell 12 populations. I will be happy to provide the 13 references to whoever would like to have them. 14 DR. COFFIN: Any response to that? Do we 15 feel in terms of this first issue that there are 16 significant safety, differential safety issues of 17 these kind of cells relative to the relatively 18 uncharacterized lines? 19 DR. MINOR: From the point of view of the 20 infectious agent side of things, I mean I don't see 21 much difference between, you know, a brand new 22 tumorigenic cell line that appears in your hand and 23 one that's actually being designed to actually appear 24 like that. It seems to me they are both 25 uncharacterized and you would have to look at both of 221 1 them very carefully. I'm not sure that you have new 2 infectious issues simply because you designed it to be 3 transformed. 4 DR. FRIED: Something came up the first 5 day. Would it be worth getting new cells that really 6 grow well with defined media so we can avoid serum and 7 any problems that come with that as something one 8 might think about? 9 DR. COFFIN: If you avoid serum, I think 10 would be a highly desirable trait in vaccine 11 production if one could engineer that. Obviously 12 there are BSE issues, and that sort of raises what 13 might be the tip of an iceberg. Is that a practical 14 goal for production? 15 DR. FALLAUX: Actually, for the 16 production, we now can grow PER C6 in serum-free media 17 in suspension. 18 MR. LEWIS: Lewis, FDA. Are there any 19 restrictions to the telomerase immortalization? In 20 other words, there are some populations that you can 21 immortalize and others that you can't, or is this a 22 universal way of immortalizing all types of human 23 cells? 24 DR. SEDIVY: You know, I am by no means 25 expert on this, so I can, you know, basically restate 222 1 what I think is already out in the literature. That 2 is that you can definitely immortalize human 3 fibroblasts, pre-senescence fibroblasts by putting H 4 tert in. That's been shown now in a number of 5 laboratories. 6 You can also immortalize retinal pigmented 7 epithelial cells as reported by the Texas group. I am 8 not sure whether it's been published yet, but I have 9 heard that you can immortalize T cells, CD8 positive 10 peripheral lymphocytes. 11 Jim McDougall says that you can not 12 immortalize keratinocytes unless he said something 13 different yesterday, in that you need to interfere 14 with the RB pathway in addition to putting H tert. 15 There's also some indication that breast epithelial 16 cells may need an additional step to become 17 immortalized. That's all I know at this time. 18 DR. LEWIS: Since you can immortalize T 19 cells and you suspect that those would grow in 20 suspension like normal T cells, and you could grow 21 them in any large suspension culture, that you would 22 need to. 23 DR. SEDIVY: I don't think this work has 24 been published, so I think I'm just telling you 25 something that I heard at another meeting. So maybe 223 1 we should just cool it. But I would presume that if 2 they grew in suspension before tert, they would grow 3 afterwards as well. 4 DR. COFFIN: Obviously these immortalized 5 T cells would be of great interest to people who are 6 interested in growing attenuated HIV vaccine. 7 DR. FRIED: But I mean once you have 8 something that's immortalized, you could always select 9 for something that grows in suspension, or you could 10 try anyway. 11 DR. COFFIN: But when doing that, of 12 course as soon as you start selecting for these 13 additional characteristics, when you are introducing 14 new and uncharacterized genetic changes. 15 DR. FRIED: Right. 16 DR. COFFIN: That these are less important 17 than safety issues, than changes that might have led 18 to immortalization in the first place. 19 DR. SEDIVY: You know, I think I would 20 also like to second the point that was brought up 21 earlier. That is that sure, we can always make the 22 claim that we know exactly what we did to the cells 23 ourselves, but we don't know what the cells have done 24 on their own during those zillions of passages that 25 they are growing in my lab. It's definitely being 224 1 documented that H tert immortalized fibroblasts are 2 karyotypically very stable. But if you passage them 3 for long periods of time, you will find anemploy 4 these. So I don't think that that's any different 5 from any other established cell lines. 6 So it's really, you know what you did, but 7 -- 8 DR. COFFIN: But you don't know what 9 happened. To bring a point to this, if one is 10 concerned about issues of what might happen with DNA 11 from the cells that was carried along, then it sounds 12 like, it sounds from what I'm hearing like there may 13 not be a great deal of difference between using these 14 cells and using these kinds of cells as compared to 15 using relatively uncharacterized cells. 16 Although there are very good reasons for 17 making such cell lines, that this particular issue may 18 not be the most important one. 19 DR. SEDIVY: You know, you could do some 20 pretty neat things that under certain circumstances 21 may be very advantageous. For example, if somebody 22 wanted to knock out the endogenous PR gene, that could 23 be done. You could make a cell line like that. 24 DR. COFFIN: You can also use factors that 25 make these cell lines highly susceptible to viruses 225 1 you might want to be growing on them, and things like 2 that. 3 AUDIENCE MEMBER: I'll just make a comment 4 on one of the papilloma transform cell lines that 5 Margaret Stanley initially isolated. It harbors -- it 6 was from a cervical dysplasia and harbors episomal 7 HPV-16 from which E-6 and E-7 are expressed. 8 Paul Lambert sub-cloned an isolate that 9 carried generous, approximately a thousand copies of 10 this episome in a fairly homogenous state. The 11 problem with it, and it sounds very appealing to have 12 episomal maintenance of your E-7 gene. The problem is 13 it's dreadfully unstable, and nobody has been able to 14 keep it with the episome. It tends to integrate and 15 completely rearrange, and has gone aneuploid. 16 So any efforts toward trying to utilize an 17 episome are probably doomed to a degree of failure 18 because of the risk of -- well, you have to maintain 19 episomal replication in addition to your chromosomal 20 replication. It just doesn't seem to work. 21 DR. COFFIN: And this of course, this kind 22 of issue gets far more amplified when we're talking 23 about growing up 10 to the 13th. 24 AUDIENCE MEMBER: Yes, absolutely. We 25 can't keep this thing going for three months. 226 1 DR. FRIED: Also, every time you knock out 2 an allele, you have to lock out the other alleles to 3 get them both. So that means more passages, and they 4 get away. I don't know whether that's good or bad. 5 I mean if you finally end up with the cell type that 6 has a lot of positive features, it may not really 7 matter whether they are instability of chromosomal, 8 and stable or rearranged. 9 DR. ONIONS: Just as a general principle 10 about whether it's useful or not to engineer cells 11 rather than go out and select a transformed cell, a 12 pre-existing transformed cell from a tumor, it does 13 strike me that again, that it's under control and that 14 you have a number of choices. 15 The kinds of studies that PER C-6 has been 16 involved in give you a very precise engineered system 17 that's absolutely ideal for vector production. But it 18 perhaps also highlights with respect to the mistake 19 that was made. That is, that you had another 20 possibility here where you could actually choose the 21 cell. You could validate its origin. You could check 22 the person. That is the other advantage of you being 23 able to engineer materials, that you can actually pre- 24 select the actual source of the material that you 25 start with. That would have been an advantage that 227 1 was unfortunately missed in this particular case. 2 That's not to undervalue the value of these cells, but 3 it does seem to me that that's what engineering cells 4 can give you. It gives you control at each stage of 5 the process. 6 DR. FALLAUX: Can I mention that it's 7 nowadays rather difficult to take primary human, 8 especially immuno material. 9 DR. ONIONS: It was -- I understand only 10 too well. It's not at the end of the day a criticism 11 of PER C6 success, which I think are excellent. But 12 really just that where possible, that that should be 13 done. 14 DR. COFFIN: Okay. So I think we have a 15 consensus here that there's lots of useful things 16 about such cells, but that we really don't' know 17 whether they enhance any particular safety issue or 18 not. I think that's a sort of at least some sort of 19 closure on that particular point. 20 The points we were asked to address also 21 included the use of the defined risk algorithm that 22 was mentioned at the beginning of the meeting, to 23 evaluate these kinds of things. Andy Lewis had that 24 on his slide, which I have asked him to put back on. 25 This will also, I mean with this, we will 228 1 sort of segway into the general discussion as well I 2 think, because these are the issues. 3 So the question is, can we go through and 4 do this, and is it possible in this particular case, 5 just using this as an example of this kind of 6 procedure, to assess the level of risk posed by these 7 issues, infectivity, infectious and so on, 8 quantitatively. My own feeling right now is that 9 we're no where near a position to do this, certainly 10 for DNA issues. We might be able to put some sort of 11 numbers on viral issues. It's a little hard to see 12 exactly how because there's so many different ones, 13 which could have a different contribution. But maybe 14 we can get some further comments on these sort of 15 issues from the panel. 16 DR. ONIONS: My only comment, and I 17 understand why a defined risk approach was used, and 18 it certainly makes you think. I mean that's one of 19 its great virtues. I think one of the real intrinsic 20 problems, that if you applied, you can give yourself 21 a false sense of security. It would strike me that 22 Phil's story about SV40, if it turns out that is the 23 origin, SV40 in people, would have given you such a 24 false sense of security, I think, because you might 25 have come through that exercise in the 1960s. I'm not 229 1 sure how you would have predicted that that agent was 2 there a priori. 3 So I'm not sure that you can give 4 guarantees that 1 and 10 to the 6th dose is one-half 5 X, if you don't know what X you are looking for. 6 DR. SEDIVY: Yes. I mean I broadly agree 7 with that. I think it's worth trying to do some sort 8 of numerical calculations, so long as you don't 9 believe the numbers that you get out at the end of it. 10 (Laughter.) 11 Because I think one thing it will do is it 12 will tell you where you think you are confident, on 13 what stage of the process you are actually confident. 14 Then you can actually question whether your confidence 15 is misplaced or not. But I think if you come out with 16 a number, I think you are asking for trouble. 17 DR. LEWIS: Yes. I don't think that we 18 discounted that. I think in sort of going through 19 this thing, what we were trying to do was to figure 20 out exactly where we can be reasonably confident of 21 what we're doing and where we can not be confident of 22 what we're doing. But we always recognize that when 23 push comes down to shove, the bottom line is that 24 we'll always -- we can never be sure. 25 So at some point in time, it requires a 230 1 leap of faith to say this product or this cell line, 2 or whatever, can in fact be used. I think what we're 3 trying to figure out a way to do is to make sure that 4 when we get to the point or we have to make that leap 5 of faith, that it's better to find than it would be if 6 we did nothing at all. 7 So the attempt here is to develop sort of 8 a way of thinking about narrowing that margin of 9 error, or at least to develop a margin of error that 10 is better than it would be if we're just doing it on 11 intuition. 12 DR. ONIONS: I think that's absolutely 13 right. I was taken by Neil Cashman's risk assessment, 14 quantitative risk assessment today. I think again, 15 what it did, although I think he himself didn't 16 believe certain the numbers at the end of the day, it 17 makes you think about the process. I think that's 18 fine and I think I would agree with that. 19 DR. HUGHES: One of the things that's true 20 about the numbers that I've seen is no one has 21 attempted to put what I guess I would call a 22 confidence interval on the numbers. One of the things 23 that makes me uncomfortable is that I think in some 24 cases the uncertainty is as large as the number. I 25 would feel a bit more comfortable with a calculation 231 1 with which I'm fundamentally uncomfortable, if I had 2 a better notion of how uncertain people were about the 3 assumptions they were making in generating the numbers 4 in the first place. 5 DR. ONIONS: I think that's what we're all 6 saying, is I think I started it off by criticizing the 7 whole approach. I think what Phil said is what I 8 think. I think Dr. Lewis said the same thing. That 9 is, don't believe the numbers. All it is is gives you 10 a manner of approaching what are the issues, really. 11 I think that's the way it should be treated. 12 I agree. I don't think anyone should 13 believe the numbers at the end of the day. 14 DR. HUGHES: I think it might help if when 15 someone put down a number, they at least put down a 16 range of numbers, and that would generate a range of 17 confidence at the end. 18 I think what people will see when they do 19 that, is that the ability to define the confidence 20 interval is going to expand when you multiply the 21 numbers together. I think that act may in a sense 22 help define how uncertain the number actually is. 23 DR. COFFIN: I think from a regulatory 24 standpoint, what often happens is that the far end of 25 the confidence interval on the worst possible side is 232 1 taken, and then that's propagated through. You never 2 see the other side. 3 DR. ONIONS: I don't want to just go into 4 an academic discussion about risk assessment because 5 I'm not really interested in it, in that formal sense, 6 but there are two other approaches that are used. The 7 engineering industry uses a form of analysis that 8 doesn't do risk assessment like that. It actually 9 looks for holes. It looks for what could go wrong. 10 In a sense, that's really probably what we ought to be 11 doing. Then there are four mechanisms of that kind of 12 analysis. 13 The other form of analysis is the one that 14 has become fashionable in the U.K., which is this 15 concept of the precautionary principle, which 16 ultimately, it seems to me, means you don't ever do 17 anything because you never know what might happen, 18 which seems to me completely dumb. 19 DR. COFFIN: So we're voting against the 20 precautionary principle. You can't be sure of 21 anything, but you can be sure of that. 22 Are there any other points anybody would 23 like to make about this? One could say that this is 24 a useful way to organize your thoughts on this 25 subject, but shouldn't be taken as giving you either 233 1 additional grounds or comfort or discouragement, 2 unless you actually had a situation where you had 3 measurable quantities. 4 Are there any other issues or questions 5 regarding the designer cell substrate issue that 6 anybody wishes to raise? 7 DR. SHEETS: Hi. Becky Sheets, FDA. I am 8 going to ask the whole panel what I tried to ask Dr. 9 Hughes earlier. That is, the kinds of questions that 10 sponsors ask us. One question I would ask is we've 11 heard a lot of people in this meeting say that the 12 oncogenic DNA issue has been put to bed. Has it? 13 DR. COFFIN: That segways us into the 14 next, into the general discussion, which is fine. 15 Before we go into that, can we see if 16 there are any other specific issues regarding cell 17 substrates? 18 DR. SHEETS: Any questions about the 19 quantitative? 20 DR. COFFIN: If we do that, we can turn 21 off that slide, put up my next one, and then you can 22 ask the question. 23 DR. SHEETS: There's one question. 24 AUDIENCE MEMBER: Jerry Sato from Merck. 25 I understand the reluctance to put a number on 234 1 something when you have such degree of lack of 2 confidence and the assumptions that are going into it. 3 But I do think that getting an order of magnitude of 4 where we are is actually helpful in our thinking about 5 what we feel comfortable going forward with or not. 6 When you have a lack of confidence in each 7 of those areas, you also have to ask the question, 8 what are the chances that all of your assumptions are 9 wrong? In other words, are two of them wrong, three 10 of them wrong, five of them wrong, seven of them 11 wrong? Because you have to put a degree of that's not 12 likely to happen. So if you multiply the lower end of 13 the confidence interval for all those things, then you 14 will never do anything. But that's not the way it 15 works in reality. 16 So I think it would be useful for somebody 17 from the engineering community, where they design 18 bridges that aren't supposed to fall down and other 19 things, to try and put a bit little more 20 sophistication into this analysis, because in the end, 21 somebody is going to ask our community, which is the 22 regulatory community, the academic community, and the 23 industrial community, for the number or at least what 24 they thought the number was when they went forward 25 with their act of faith. Because there is a certain 235 1 amount of common sense that goes into it, which is the 2 basis of the act of faith. Then there is whatever 3 kind of quantitation we can put into it. It's the 4 combination of those two things that I think we are 5 going to have to reassure the general public about. 6 DR. HUGHES: I would recommend to you the 7 book Strategy and Conscience by Anatol Rappaport, 8 which attempts to deal with the issues having to do 9 with what was called strategic thinking, in which you 10 calculate, for example, the probability of some 11 unlikely event, such as thermonuclear war. Mr. 12 Rappaport does a very good job of making clear why 13 doing the calculations when you don't have the proper 14 data is in fact a very risky and misleading 15 proposition. 16 AUDIENCE MEMBER: I guess it might be 17 worth pointing out that there are, however, some cases 18 where you clearly do have the proper data. Right? 19 You know the sensitivity of a specific assay for a 20 specific adventitious agent, if in fact you choose to 21 figure out what it is. 22 So you can actually answer, based on those 23 kinds of questions, and based on that kind of data, 24 the specific question of how sure are you that 25 something isn't there. If a better assay comes along, 236 1 and it's still negative, then you can say by how much 2 more certain you are. 3 So just because you can not come up with 4 good estimates for some of the numbers, seems to me it 5 would be crazy to throw the baby out with the bath 6 water and claim then that you shouldn't attempt to 7 come up with good numbers for those things that you 8 can. 9 DR. RABINOVICH: Gina Rabinovich, NIH. A 10 non-regulator asking a question from experience 11 learned this summer, in which we have been dealing 12 with using a quantitative number, i.e. the numbers 13 that the Federal agencies and the global agencies has 14 set for acceptable limits of mercury, i.e. methyl 15 mercury, and then trying to attempt to understand what 16 those uncertainty factors mean for thimerosal and 17 vaccine exposure im. 18 The concern I have, and I think it has 19 been heard, is that these numbers take on a life of 20 their own. They become the standard against which 21 things are measured. So that that kind of concern 22 needs to be entered into, attempting to use the data 23 when those data do exist, but understanding the limits 24 to it. 25 DR. COFFIN: That's inherent in 237 1 regulation, that things become standard. 2 DR. LEWIS: Lewis, FDA. To follow up a 3 little bit on what Phil was saying. I think one of 4 the areas that we could approach with some confidence 5 is the ones who saw Keith Peden's data last night 6 using the tac man assay to detect JC, BK, and SV50 in 7 human tissues. 8 Now if someone comes in with substrates 9 they derive from a neuroblastoma, which we learned at 10 the DNA tumor virus meeting a year or so ago, is it's 11 usually contaminated with BK virus. We wanted to be 12 sure that there was no BK virus in that substrate. 13 Then we could apply this assay with a fairly 14 sufficient level of sophistication and say with some 15 certainty, based on the volumes and things that were 16 tested, the level at which that particular genome or 17 that particular virus is absent. So I think that we 18 sort of view that as a possible starting point for a 19 quantitative approach. 20 Now obviously you can't do that unless you 21 know exactly the probes and things that you are 22 working with, and you define the limits of their 23 ability to detect things. But I think this is one of 24 the sort of examples that was going through our mind 25 when we were thinking about how to do this. 238 1 So you start at the place where you might 2 be able to generate some relevant data that's 3 meaningful. The other stuff will fall into place as 4 we get better. 5 DR. ONIONS: I mean I think that's exactly 6 right. I think more and more that we move to assays, 7 that we get good quantification on, we can validate 8 them, and it's the sensitivity and limits of 9 detection. I think all of that is absolutely 10 essential. I mean I absolutely 100 percent concur 11 with that. I think it does add confidence to those 12 specific questions. 13 I think when you are asking specific 14 questions, then I 100 percent agree. I think my 15 concern is perhaps that where you try and make 16 assumptions, for instance, of residual DNA. I mean I 17 was the one who said I thought it should be put to 18 bed. That is because all the evidence that I had 19 heard didn't convince me that there was a risk. 20 On the other hand, I actually believe that 21 no one has done the right experiment to actually 22 convince me of that, in a formal scientific sense. So 23 we're then dealing with the area of conjecture. That 24 conjecture is based on non-quantitative data. 25 DR. COFFIN: We're leading into you, 239 1 Becky. 2 DR. SHEETS: I'm patient. 3 DR. COOK: I'm sorry. Jim Cook. As I was 4 sitting there thinking about how you would describe 5 these issues to a patient or to a group who is asking 6 you about the wisdom of using a vaccine, it seems like 7 in addition to trying to generate some logic about 8 calculations of numbers and risks, that every 9 opportunity that you have, it would be worthwhile 10 going back to history and saying well, we have done 11 virtually something like this along the way ever since 12 vaccines have been developed, and the experience with 13 this approach has been the following. 14 So maybe there could be some real numbers 15 in a historical sense, used to color or give some more 16 real meaning to these, what are otherwise theoretical 17 things, to help communicate this to the public, as 18 well as to provide some, a little bit more logic than 19 just phenomenology to the calculations that are being 20 made now. So if history is used to color the 21 estimations, that might be of some use. 22 DR. COFFIN: How comforting is it to tell 23 -- 24 DR. COOK: Say it again? 25 DR. COFFIN: How comforting is it to tell 240 1 patients that there are three or four cases of 2 paralytic polio? 3 DR. COOK: Well, I think you have to be 4 honest with them. You say look, you are one of 250 5 million people. The odds for your child is the 6 following, and I think it's a very good idea to use 7 this vaccine. 8 It is those kinds of conversations that 9 eventually lead adolescents into getting hepatitis B 10 vaccine. We're doing a miserable job of this, by the 11 way. I think as a culture, you know, if we bat 50 12 percent, we're lucky. 13 DR. COFFIN: That's clear in light of 14 things in the movies, because lately we're doing a 15 terrible job where these people show up with these 16 anecdotal cases of somebody's child gets vaccinated, 17 and then two months later is diagnosed with autism. 18 It's automatically due to the vaccine. 19 DR. ONIONS: But I think history can also 20 be a dangerous thing. I mean the British government 21 has been criticized, partly justifiably, but I think 22 partly unfairly for the problems of the way BSE 23 problem goes up. 24 But the issues concerning public health 25 were based on people were asked what is the risk, what 241 1 is the risk of the human population of the BSE 2 outbreak, when we had a few hundred cases of BSE in 3 cattle. Well, there were only a few hundred cases. 4 The general assumption was, and it was a widely shared 5 general assumption by those who were informing the 6 area, the people who worked on scrapie and so on, 7 well, scrapie has no evidence whatsoever of scrapie 8 transmission to man. We have been eating scrapie- 9 infected sheep for generations and it doesn't seem to 10 have been transmitted to people. There is no evidence 11 of that whatsoever. 12 The probable likelihood that BSE is of 13 scrapie origin that's gone through the rendering 14 process, because we have never seen a spontaneous 15 spongiform encephalopathy in cattle, so that's 16 probable origin ipso facto, you know, there's no 17 problem. 18 It wasn't quite as glib as that because 19 actually within a year, all the controls on human food 20 were in place and so on and so forth. They were badly 21 conducted, but they were theoretically in place. 22 So I think history can also be dangerous. 23 I'm not sure that we can always learn the right 24 lesson. 25 AUDIENCE MEMBER: I get your point, but I 242 1 think there is a difference between an ongoing 2 epidemic that's yet resolved, and the experience of X 3 number of hundreds of millions of doses of polio 4 vaccine. 5 So what I am saying is, if you are going 6 to make a calculation about the likelihood of one in 7 a million or one in 10 or 100 million happening with 8 a vero cell or with a primary cell, you can say well, 9 we've got experience making polio vaccine in vero 10 cells, and there have been the following hundreds of 11 million doses given, and the likelihood is that when 12 the incidence of the real disease gets so low that you 13 are finally going to see some background. That's the 14 real -- 15 DR. ONIONS: I'm sorry. You 16 misunderstand. What you are saying, I absolutely 17 agree. 18 DR. COFFIN: Do you want to continue this 19 discussion or do you want to break into -- 20 AUDIENCE MEMBER: Very similar aspect, 21 although a little bit less scientific. As I perceive 22 the discussion, of course there's this highly 23 sophisticated, highly conscientious scientific 24 community, and there's the general public on the other 25 side. 243 1 The general public, to my perception, 2 consists of at least two sub-groups. One group that 3 is sort of generally benevolent and would believe 4 scientists. But there is a very strong group that is 5 not believing scientists. They are sort of using that 6 as a political tool to attract attention. In our 7 country, we have had this experience with the Green 8 Party, that has become very influential in the 9 European Union and maybe in other countries as well. 10 Now one thing I also, since we're ending 11 and coming to the end of this meeting, would like to 12 raise, isn't it the responsibility of scientists also 13 to do something to better educate the public? I know 14 this is a utopian goal, but at least if we could 15 increase enough of people in the general public who 16 are educated or better educated in science and biology 17 and biomedic issues, we would at least have a 18 political community that might support scientific 19 issues more valuable than we have had it so far. 20 You know, in our country at least, every 21 time when the Greens demanded to stop all biomedical, 22 all gene technology research, and it course never 23 happened. After they come to power, maybe they have 24 different outlooks on life. But I think as a 25 scientific community, unless we do something at least 244 1 for the future, we might be in a very difficult 2 situation to defend certain issues. 3 If I confronted some of the violent 4 ideologically pure Greens in our country, because the 5 trick is, we have been discussing here, I'm sure they 6 would say "Shut it down because this is unsafe, 7 totally unsafe." 8 So what I am trying to recommend is we 9 have to do something to have more people in the 10 general public who can appraise and can assess the 11 difficulty and the uncertainty in any biological 12 research. We can never get down this figure to 10 to 13 the minus 80. So we have to raise understanding on 14 the other side. 15 DR. SHEETS: So have we put oncogenic DNA 16 to bed? 17 DR. COFFIN: No. I would like to use the 18 sort of summary -- I think that's in a sense, an 19 overhanging issue. We have talked about infectious 20 risks and measurements and so on considerably today 21 and in the past few days. I think an overhanging 22 issue is this oncogenic DNA issue regarding the 23 specific charge of the meeting, which is the use of 24 tumor versus other kinds of cells, tumor and 25 neoplastic whatever, transformed cells versus other 245 1 kinds of cells as substrates for vaccine production. 2 Although I think many of us here, perhaps 3 all that's here, feel this is not a risk to be really 4 concerned about in a scientific sense, I think many of 5 us here might agree that the issue is not completely 6 put to bed in the sense that we can't put any real 7 good numbers on it. 8 So now if you ask your question. 9 DR. SHEETS: Okay. Has oncogenic DNA been 10 put to bed? 11 DR. HUGHES: I will answer in two 12 different ways. I will give you my opinion, personal 13 opinion, and then I will tell you what I think should 14 be done, which is slightly different. 15 Personally, and this would apply if you 16 approached me to do something to myself, I am not 17 concerned. However, it is my view that the data that 18 we have, particularly for the consequences of putting 19 DNA into animals, is not sufficient to satisfy me as 20 a scientist. I am going to try to help my colleague, 21 Dr. Coffin, and some of my colleagues at the FDA to 22 try and organize a simple study that would be more 23 satisfying to me. 24 I think I would feel more comfortable if 25 we had more data that was of the experiments done on 246 1 a larger scale under more controlled conditions. I 2 think that would give me a greater degree of comfort. 3 DR. ONIONS: I think I almost entirely 4 concur with that comment actually. I did say a 5 comment, and said it partly to be provocative, but on 6 the other hand, I think I share that opinion. 7 I have seen nothing that would convince me 8 at the moment on the data, which there's a singular 9 lack of, or just from I suppose theoretical reasoning, 10 to suggest that this would really be a significant 11 danger. But on the other hand, there are the tools 12 now, and I say yet again, but I think some of the 13 transgenic models offer that possibility for testing, 14 whether or not DNA is a risk. It can be done in a 15 series of graded experiments from taking the worst 16 case examples of actually simply just repeating -- of 17 injecting oncogenes at various titrations into animals 18 that are already primed with oncogenes as a 19 transgenic, down to taking tumor DNA down to taking 20 normal DNA. 21 I mean those experiments, they are pretty 22 straight forward to do. I mean the interpretation 23 might be a bit more complex, but they can be done. I 24 think they are worth doing. It might help you to put 25 some limits, broad limits on the thing we're all just 247 1 making conjectures about. 2 DR. MINOR: I don't think it's been put to 3 bed either. 4 DR. COFFIN: For the same general reason? 5 DR. MINOR: Broadly speaking. I mean I 6 think it's clearly a very, very complex issue about 7 how you actually induce a tumor. So you do your 3T3 8 assays and you pick up H ras. Okay? I mean it's an 9 artifact of 3T3s or was that just a question of how 10 common H ras is. 11 If you go and put your DNA in 12 intravenously, is that the same as putting it in 13 subcutaneously, for example? I mean if you put it in 14 because it's been picked up by an envelope virus, is 15 that going to make any -- will it be picked up by an 16 envelope virus? Will that make any difference? 17 I mean it seems to me that there are so 18 many sort of loose ends to it that I don't think while 19 there is no evidence that DNA is tumorigenic, and I 20 buy that 100 percent, it doesn't seem to me that it's 21 necessarily been dealt with properly. That is why I 22 guess I am agreeing with what the previous two 23 speakers said. 24 DR. LOEWER: So as I already have said, I 25 personally believe that there's not a real big risk 248 1 with purely oncogenated DNA. Purely oncogenated means 2 three or five or six. But I realized that there is 3 still, I believe, lack of experimental data. This was 4 already mentioned by John Coffin, in saying that since 5 18 years, this question is on the table. Since 18 6 years, no additional experiments have been performed. 7 I would like to propose that regulatory 8 authorities, which are involved in regulation of these 9 biologicals, the major ones, that we should sit 10 together and to join the efforts and maybe decide on 11 experiments which can be done in the foreseeable 12 timeframe. 13 But I look forward to see what types of 14 experiments John may recommend. 15 DR. FRIED: I think most of the evidence 16 we have so far, which is limited, says at least 17 putting DNA into animals, we haven't seen anything 18 happen. 19 We have only seen in NIH3T3 cells, and we 20 now know that there is a defense mechanism in the cell 21 when it sees an oncogene. That is this P, this arf, 22 which is the alternative reading frame of P-16. So 23 it's like an immune system of the cell, specifically 24 for oncogenes. The radiation activation of P-53 is a 25 different pathway. This turns on, and arf activates 249 1 P-53, and P-53 then closes itself to go through 2 apoptosis. 3 The only positive things are in NIH3T3 4 cells. They are the classic cell where the arf gene 5 is inactivated. Probably that's why people have been 6 using that for years. They are very easy to transfect 7 because maybe even transfection kills the cells in 8 terms of P-53. 9 But that said, I would like to see a lot 10 more injection of DNA from different tumor lines into 11 animals, and to really put it to bed. 12 DR. COFFIN: It's striking to me that one 13 of the very few, if only successful experiments with 14 injected activated oncogene DNA in chickens is the one 15 that Hsing-Jien Kung did that was off-site, or in any 16 animal, that was Hsing-Jien Kung. It was actually 17 fairly efficient sort of transient transformation of 18 cells, but transformation of cells is always 19 transient. It's virtually impossible to immortalize 20 them. They are much harder then human cells to 21 immortalize. It's been done once, to my knowledge. 22 So there may be something very 23 fundamentally different going on in that model, 24 because there may be some fundamental difference in 25 chickens as compared to mice, as compared to humans, 250 1 which already have important -- 2 DR. EGAN: Bill Egan from FDA. I would 3 certainly like to work with you and other regulatory 4 authorities to try and design and our colleagues in 5 PHS, to try and design these experiments and do these 6 experiments, and get the data, and get away from the 7 remark I quoted from Maurice Hilleman from 30 some-odd 8 years ago about this debate being a philosophical or 9 ecclesiastic debate because we simply don't have the 10 data. Here we are 30 some-odd years later, you know, 11 with the same question, with the same debate. It's 12 still opinion. 13 I must say I also feel that personally, 14 myself, I don't think there is a large risk from the 15 DNA. But then again, the kinds of risks that we're 16 talking about are very small risks, very, very rare 17 events. Things like one in a million are not 18 acceptable or in many cases are not acceptable. Those 19 are hard data to get. 20 While I may not feel there is a risk to 21 me, the bottom line basically for the approval of 22 almost any of these vaccines, is would I put this into 23 my children. There it becomes a much more 24 conservative process. If putting it into my children 25 is putting it into other people's children, it's the 251 1 same thing. 2 DR. HUGHES: If you want one more piece of 3 data that should give you some comfort, it is the 4 experience attempting to make antibody to the oncogene 5 sarc, which involved putting an avian virus into a 6 number of mammalian species under circumstances in 7 which the virus absolutely does not replicate. 8 In adult animals, to my knowledge, no cell 9 growth was ever seen. The only experiments that 10 succeeded, to my knowledge, were those initiated by 11 John Burge, in which he put enormous amounts of ras 12 sarcoma virus of a subgroup that would infect 13 mammalian cells. We are talking sort of 10 to the 14 10th infectious units. Into immunologically naive 15 baby bunnies. In those animals very transiently, 16 there were small nodules which regressed. 17 In those animals, you did see antibody to 18 the oncogene sarc, implying that there was transient 19 uptake of the DNA, at least probably permanent uptake 20 of the DNA. But even under those circumstances in 21 which the delivery of the DNA is extremely efficient 22 and every copy of the DNA carries a known potent viral 23 oncogene under circumstances where the DNA will not 24 replicate but will insert, do you see any permanent 25 transformation? Again, I'm not suggesting that this 252 1 is sufficient. I am one of the advocates of more 2 experimentation. But the data that we have suggests 3 that this is not a simple process. 4 DR. EGAN: No, but I mean these are the 5 kind of data that start to put brackets around the 6 numbers for the levels of risk. 7 DR. COFFIN: Of course you have enormous 8 problems, including the fact that sarc is never seen. 9 It's a human oncogene. For many years, the most 10 popular viral model. 11 AUDIENCE MEMBER: I had a question. If 12 given the unknowns, and given the data that was 13 presented about hit and run DNA modification 14 potentials, would the panel in the context of this 15 type of vaccine development, and given the unknowns, 16 give the vaccine to someone with a strong family 17 history of malignancy or who was a cancer survivor who 18 we know is at increased risk for a second cancer, if 19 that was you or your family member? 20 DR. COFFIN: The question, to sort of 21 focus that a little bit, the question is whether we 22 would consider there to be a greater risk in certain 23 sub-populations who might have sort of pre-activated 24 oncogenes or some other fact of predisposing. 25 DR. ONIONS: I understand the question, 253 1 but it sort of arose -- I will go backwards, because 2 one of the things that used to concern me was when 3 people were doing clinical trials with rusvel vectors, 4 and they were using marker studies. That sort of 5 study did worry me, because actually what you were 6 there doing was putting something, inserting something 7 into somebody who probably already had a preexisting 8 oncogenic hit. That struck me as being dubious, 9 mildly. 10 In this situation, I would have thought 11 that unless you have got somebody with a Li-Fraumeni 12 syndrome or something, that you are dealing with 13 changes that are somatic changes in a few cells, even 14 if they are going to risk of a second cancer. 15 So the likelihood that you hit the right 16 cell is pretty low. So I wouldn't have thought it up 17 the risk -- I'm sure the risk has increased, but I 18 wouldn't have thought that risk has increased 19 significantly, unless of that sort where you've got 20 mutation. 21 DR. MINOR: It would also depend on what 22 you are trying to protect them against too, wouldn't 23 it? 24 DR. SHEETS: Before we lose our entire 25 audience, I wanted to -- I think we have gotten a 254 1 pretty clear answer on the oncogenic DNA issue. I 2 wanted to ask another kind of question that FDA has 3 asked, before we lose our entire panel. That is, are 4 there risks, additional risks that one perceives in 5 using a continuous cell line such as vero cells, 6 particularly vero cells, which is immortal but not 7 tumorigenic at the levels that vaccines are made, past 8 the level that vaccines are made. Is that worse in 9 any way than diploid cells, for a live viral vaccine? 10 DR. MINOR: This is apart from the DNA 11 issue? 12 DR. SHEETS: Well, in a continuous cell 13 line, certainly there are -- it may be aneuploid, but 14 it's not tumorigenic in animals. So you can comment 15 if you'd like about whether you think the DNA is 16 oncogenic. 17 DR. MINOR: I would say that the DNA from 18 vero is as questionable as the DNA from anything else. 19 I mean John, with whom we discussed these matters, 20 more or less said the same thing. I think that it 21 doesn't matter how malignant it is. Maybe it depends 22 on how many oncogenes you put in there. So I would 23 have thought that a vero is as questionable as 24 anything else, or is not as questionable as anything 25 else. 255 1 DR. SHEETS: So you wouldn't suggest to 2 make a live viral vaccine in vero cells? 3 DR. MINOR: I think it would depend on the 4 live viral vaccine. I mean I think OPV clearly has 5 been made in vero cells. You can scrub it clean. I 6 think you can more or less destroy anything that's 7 actually hanging on the end. 8 I have more serious thoughts perhaps about 9 things like a paramixo virus vaccine, because you 10 couldn't clean it up so much perhaps. 11 DR. SHEETS: What about the sort of crude, 12 less purified live viral vaccines, not the purified 13 vaccines like we heard about last night with OPV, but 14 the things that are just filtered cell culture 15 supernatant? 16 DR. MINOR: Right. They might be figit. 17 I'll tell you that. Which is not to say there's any 18 good reason for me to be uncomfortable with them. 19 It's just that they make me feel uncomfortable. 20 DR. COFFIN: We are very fast losing our 21 audience, so I think -- 22 DR. HUGHES: Isn't it partly the question 23 do you know the life history of your cells as opposed 24 to the state of the cells at the end? 25 DR. SHEETS: Vero cells are a bank that is 256 1 well characterized. The reason for the question is 2 that we have numerous live viral vaccines of the sort 3 I described that are being proposed to be made in vero 4 cells. Manufacturers prefer vero cells because one, 5 they can be characterized. Two, you get a high yield. 6 Three, they can be grown in the sorts of fermenter 7 culture that you heard about. 8 DR. HUGHES: I'm not particularly bothered 9 as long as I know that the sort of life history of the 10 cell. But I think the question is, if you have a cell 11 that's been in culture for a long time and has had a 12 complicated culture history, do you know that history? 13 AUDIENCE MEMBER: I asked the last panel 14 the same question. It comes down to the question 15 really is the adventitious agent issue put to bed as 16 well. Do we now have the assays in place that can 17 easily be applied to validate the freedom from 18 adventitious agents of these kinds of new cell lines? 19 The answer that Dr. Broker gave in the 20 last panel suggested that one could attempt to use DNA 21 chips and things like that, which to my knowledge 22 aren't assays that at least tomorrow I could go out 23 and do on a cell line and give me some confidence. 24 So my question to you is, sort of using 25 the standard assays that you are all aware of, without 257 1 developing further assays for this specific purpose, 2 do we have enough information to be sure that these 3 kinds of new cell lines are safe from the adventitious 4 agent perspective? 5 DR. ONIONS: That's another unanswerable 6 question, isn't it. I would just make the point that 7 I think you have to adopt somewhere between good 8 science and pragmatism. I mean you could 9 theoretically go and do representation difference 10 analysis on all these cell lines. Actually, I don't 11 think it's possible because you don't usually have the 12 partner. But theoretically you could do that. That's 13 not really a practical solution. 14 It does seem to me that we do know virus 15 types that tend to be latent in cells, and that it's 16 sensible to perhaps think of strategies of widening 17 the brief of detecting those agents, because I'm not 18 convinced that the kind of routine types of infecting 19 -- infectability assays when they work are as 20 sensitive as PCR, as just Phil pointed out. But I am 21 not convinced that always the right infectability 22 assay is present to actually detect certain agents. 23 So that you are probably relying on a combination of 24 things. Perhaps we do need to look at redundant PCR 25 for certain agents. 258 1 DR. COFFIN: I would think the producers 2 would have a big attraction, is set up the same assay 3 and use it for everything. 4 DR. MINOR: I mean I think you could also 5 argue that you have used these assays for looking at 6 human diploid cells and primary cultures, and all that 7 sort of stuff. Right? What's the difference in 8 principle in terms of adventitious agent contamination 9 between those and the cells you are looking at here? 10 I am not sure there's much difference. But are the 11 concerns as big or as little. 12 AUDIENCE MEMBER: For one, I would like to 13 thank Dr. Andy Lewis for bringing us together. This 14 has been a very stimulating week. Last night Dr. Vyas 15 showed us a picture of the thinker. But that actually 16 prompted me to remember that Rodin placed that 17 gentleman directly above the gates of hell. 18 (Laughter.) 19 The question I would like to pose right 20 now is whether we're walking in through the gates or 21 out. 22 DR. COFFIN: I would like to also second 23 the thanks to the organizers for setting this up and 24 bringing us here. 25 DR. LEWIS: Yes. On behalf of the 259 1 sponsors and those of us at CBER who worked on this, 2 we really appreciate the effort that the session 3 chairs, the panel chairs, and the speakers have put 4 into this meeting. When you attempt to put something 5 like this together, there's always a question of how 6 it's going to turn out. I think the success that we 7 have enjoyed here the past three days is a tribute to 8 the work, an incredible amount of work, that has gone 9 on on a very short period of time. 10 I think that I was very concerned when we 11 were trying to contact folks in May to do this by 12 September. For those of you who rose to the 13 challenge, I can't thank you enough on behalf of the 14 sponsors. 15 With that in mind, I hope everybody has a 16 great trip home. Get your papers in whenever you can. 17 Thank you. 18 (Whereupon, at 2:57 p.m., the proceedings 19 were concluded.)