GRADUATE SCHOOL

U.S. DEPARTMENT OF AGRICULTURE

* * *

FDA FOOD ADVISORY

COMMITTEE MEETING

* * *

FOOD BIOTECHNOLOGY SUBCOMMITTEE

OF THE FOOD ADVISORY COMMITTEE

MOLECULAR CHARACTERIZATION

CENTER FOR FOOD AND APPLIED NUTRITION

Wednesday, September 24, 2003

J.W. Marriott Hotel

1331 Pennsylvania Avenue, N.W.

Washington, D.C.

The meeting convened, pursuant to notice

at 8:36 a.m., before Acting Chairman Francis F.

Busta.

MEMBERS PRESENT:

JONATHAN ARIAS, Ph.D.

Associate Research Scientist

Department of Biological Sciences

University of Maryland, Baltimore County

1000 Hilltop Circle

Baltimore, Maryland 21250

Tel: 410-455-3470

Fax: 410-455-3875

Arias@umbi.umd.edu

BOB B. BUCHANAN, Ph.D.

Professor

University of California

Department of Plant and Microbial Biology

111 Koshland Hall

Berkeley, California 94720

Tel: 510-642-3590

Fax: 510-642-7356

view@nature.berkeley.edu

FRANCIS FREDERICK BUSTA, Ph.D.

Emeritus Professor

University of Minnesota

Department of Food Science and Nutrition

1334 Eckles Avenue

St. Paul, Minnesota 55108-6099

Tel: 612-624-3086

Fax: 612-625-5272

DOUGLAS GURIAN-SHERMAN, Ph.D.

Science Director, Biotechnology Project

Center for Science in the Public Interest

1875 Connecticut Avenue, N.W.

Suite 300

Washington, D.C. 20009

Tel: 202-332-9110 Ext. 377

Fax: 202-265-4954

dgurian-sherman@espinet.org

MEMBERS PRESENT (Continued):

ANNE R. KAPUSCINSKI, Ph.D.

Professor, University of Minnesota

Department of Fisheries & Wildlife

180 McNeal Hall

1985 Buford Avenue

St. Paul, Minnesota 55108

Tel: 612-624-3019

Fax: 612-625-8153

ark@fw.umn.edu

ABIGAIL A. SALYERS, Ph.D.

Professor of Microbiology

Department of Microbiology

University of Illinois at Urban-Champaign

B103 CLSL

601 S. Goodwin Avenue

Urbana, Illinois 61801

Tel: 217-333-7378

Fax: 217-244-8485

Temporary Voting Members:

DENNIS GONSALVES, Ph.D.

Center Director

Pacific Basin Agricultural Research Center

99 Apuni Street, Suite 204

Hilo, Hawaii 96720

Tel: 808-932-2100

Fax: 808-969-6967

Dgonsalkves@pbarc.ars.usda.gov

STEPHEN BENEDICT, Ph.D.

Department of Molecular Biosciences

University of Kansas

1200 Sunnyside Avenue

Lawrence, Kansas 66045-7534

Tel: 785-864-4007

Fax: 785-864-5294

sbene@ku.edu

MEMBERS PRESENT (Continued):

CALVIN QUALSET, Ph.D.

Genetic Resources Conservation Program

DANR Building, Hopkins Road

University of California

Davis, California 95616

Tel: 530-754-8502

Fax: 530-754-8503

coqualset@ucdavis.edu

NINA FEDOROFF, Ph.D.

519 Wartik Laboratory

Huck Institute for Life Sciences

Penn State University

University Park, Pennsylvania 16802

Tel: 814-863-5717

Fax: 814-863-1357

Industry Representative:

JAMES ASTWOOD, Ph.D.

Director, Food and Feed Safety Policy

Monsanto

800 North Lindbergh Boulevard

St. Louis, Missouri 63167

Tel: 314-694-8396

Fax: 314-694-8562

james.d.astwood@monsanto.com

ALSO PRESENT:

MR. BOB LAKE

MS. JEANETTE GLOVER GLEW, CFSAN

MR. MICHAEL HANSON (Public Comment)

DR. JAMES MARYANSKI, CFSAN

DR. THOMAS CEBULA, CFSAN

C O N T E N T S

AGENDA ITEM PAGE

Welcome and Introductions 6

Acting FBS Chair

Conflict of Interest Statement 9

Welcome from FDA 13

Charge and Questions 23

Acting FBS Chair

FDA's Biotech Food Safety Assessment 29

Ms. Jeanette Glover Glew, CFSAN

Questions of Clarification 43

FBS Members

Codex Approach 74

Dr. James Maryanski, CFSAN

Questions of Clarification 92

FBS Members

FDA's Discussion Paper — Molecular

Characterization 103

Dr. Thomas Cebula, CFSAN

Questions of Clarification 122

FBS Members

Public Comment 153

Summary and Review of Charge and

Questions 172

FDA

Discussion 179

FBS Members

Response to Questions 194

FBS Members

Concluding Comments 334

Acting FBS Chair

P R O C E E D I N G S

CHAIRMAN BUSTA: The meeting is called to order. And welcome, all of the committee members, guests, and the members of FDA to the--this Biotechnology Subcommittee meeting of the Food Advisory Committee.

My name is Frank Busta. I'm a professor emeritus at the University of Minnesota. I'm a member of the, let's see, the term is full committee. I'm--that is not implied that I'm full of it, but some people would say it. And I've been asked to chair this meeting as an acting chair.

In front of you, you have an agenda, and, as you see, in addition to welcoming you, we'd like to have our introductions so that everybody knows who everyone is, and we have it on the record.

I believe that we're asked always to talk into the microphones because it is being recorded, and that gives us a record, as well as someone taking notes.

So, if we could start with introductions, telling us basically who you are and your--the

committee. Douglas, would you start?

DR. GURIAN-SHERMAN: Doug Gurian-Sherman. I'm with Center for Science in the Public Interest.

CHAIRMAN BUSTA: And you are a member of the--the regular member?

DR. GURIAN-SHERMAN: Yes, I'm a member of the subcommittee.

DR. QUALSET: I'm Cal Qualset, professor emeritus, University of California at Davis. I'm a member--a temporary member of the subcommittee.

DR. BENEDICT: Steve Benedict, University of Kansas, and I'm a temporary member.

DR. ARIAS: Jonathan Arias, University of Maryland. I'm a subcommittee member.

DR. FEDOROFF: Nina Fedoroff, Penn State University. I guess I'm a temporary member.

CHAIRMAN BUSTA: Abigail?

DR. SALYERS: Abigail Salyers, University of Illinois, and former President of the American Society for Microbiology. And I'm not sure what kind of member I am.

CHAIRMAN BUSTA: You are a member of the

subcommittee.

DR. SALYERS: Oh, okay.

DR. BUCHANAN: Bob Buchanan, University of California at Berkeley, and I believe I'm member.

DR. GONSALVES: Dennis Gonsalves. I'm at Pacific Basin and Agriculture Research Center, and I'm a temporary member.

DR. ASTWOOD: I'm Jim Astwood. I'm with Monsanto, and, I'm the industry liaison.

CHAIRMAN BUSTA: And Mike?

DR. WATSON: Mike Watson, the Food and Drug Administration.

CHAIRMAN BUSTA: And Mike does all the work.

All right. To get this organized and straight, I would like you to refer to our agenda. You see estimated times to accomplish what we need to do today on the agenda.

We hope that we can stay fairly close to that. I will try to be as firm a chair as possible so that we do not diminish our opportunities for discussion and summaries in the afternoon.

Mike, we're five minutes ahead of time, but I'd like--I don't mind that at all.

DR. WATSON: That's fine.

CHAIRMAN BUSTA: I will turn over the conflict of interest statement, et cetera, to Mike Watson.

DR. WATSON: Good morning, again. I'm Mike Watson. I'm the Acting Executive Secretary for the Biotechnology Subcommittee of the FDA Food Advisory Committee.

First, I would like to read the temporary voting member appointment and conflict of interest statements into the record.

By the authority granted under the Food Advisory Charter of July 2002, the following individuals have been appointed as temporary voting members by Joseph A. Levitt, Director of the Center for Food Safety and Applied Nutrition:

Dr. Steven Benedict;

Dr. Nina Fedoroff;

Dr. Dennis Gonsalves; and

Dr. Calvin Qualset.

With regard to the issue of conflict of interest, committee members, both permanent and temporary, were screened for interest in bioengineered food developers. As a result of this review, in accordance with 18 U.S.C., Section 208 (b)(3), Dr. Robert Buchanan and Dr. Dennis Gonsalves have been granted a particular matter of general applicability waiver that permits them to participate fully in matters at issue.

Copies of their waiver statements may be obtained by submitting a written request to the Agency's Freedom of Information Office, Room 12A-30 of the Parklawn Building.

We would also like to note that Dr. James Astwood is participating in this meeting as the acting industry representative, and is a non-voting participant.

With respect to all other participants, we ask that, in the interest of fairness, that they address any current or previous financial involvement with any firm that develops or sells bioengineered food crops.

Finally, FDA received a letter on Monday this week, September 22, 2003, from the Center for Science in the Public Interest, CSPI, related to the work of this subcommittee. This letter was signed by Dr. Gurian-Sherman, a consumer representative on this subcommittee, on behalf of himself and two other subcommittee members, Drs. Arias and Kapucinski.

The letter expresses concern about FDA's management of the subcommittee, selection of agenda topics, and other issues. FDA has provided copies of this letter to the subcommittee members, and copies are available to the public.

FDA appreciates receiving suggestions for enhancing the work of the subcommittee. Since this letter was submitted too late for FDA to fully consider its contents and suggestions for changes for this meeting, we intend to examine the letter carefully and respond following this meeting.

For today, we intend to focus on the agenda that we have prepared. Mr. Chairman, we note that one of the issues raised in the CSPI

letter dealt with future agenda items for the subcommittee.

We leave it to your discretion; but, if time permits, FDA will be interested to hear any suggestions that subcommittee members may have on future agenda items.

I'll turn it back over to Dr. Busta.

CHAIRMAN BUSTA: Well, thank you, Mike.

I'm not sure if we got a copy of that letter by e-mail before this meeting. I'm not sure how many of you are in a position to see your

e-mail en route. The--we'll have an opportunity to look at that in case you hadn't seen it earlier; and, as Mike indicated, we will do our best to get through the agenda. And we will place that at the end of the agenda for comment, and we can expand that for other comments beyond the letter as for future agenda items.

All right. Now, we're going to have a welcome from FDA. And I assume that's Bob Lake?

MR. LAKE: Yes.

CHAIRMAN BUSTA: They're really neglecting

you this morning, Bob. They didn't give you a name card.

MR. LAKE: Well, that's all right.

Let me--well, again, I am Bob Lake. I'm the Director of Regulations and Policy at the Center for Food Safety and Applied Nutrition. I report directly to Joe Levitt, and one of the things in my portfolio is policy issues relating to bioengineered foods. Jim Maryanski, Dr. Maryanski, who will be talking to you later, and who I think is--most of you know very well--has been my technical advisor for probably 15 years.

He is the biotech coordinator, but we established that position back in the late '80s; and his title and mine have changed a little bit over the years. But we have worked very closely together throughout that time.

I want to officially, on behalf of SISAM management, welcome you here to this meeting. I know everyone has very busy schedules, and it's certainly hard to find a time where we can get everyone.

But we very much appreciate all of you being here. You are important to us. We very much appreciate your time, your consideration; and will value whatever advice we get from you.

I thought I would take a little bit of a moment to just sort of update you on sort of what's been going on, or, in some cases, not going on. Just to give you a little bit of a backdrop, I'm going to try to do this very quickly.

As you all I think know, FDA is only one of the agencies in the Federal Government that is involved in issues that relate to bioengineered foods. The Department of Agriculture, most notably. The Animal, Plant, Health and Inspection Service has responsibility for, you know, the planning of bioengineered plants. EPA has responsibility for genetically engineered pesticides that are genetically engineered into new plant varieties. And then FDA, of course, has the responsibility for the safety of foods, and, you know, again, that's--we're talking about what happens to the person who eats the food.

Also, we have the responsibility for animal feeds, and so we have a concern about that, as well.

Since the StarLink episode of a few years ago, there's been a sense across the agencies that there has been a need for greater interaction. And, indeed, there have been numerous meetings among the agencies, these agencies, also others, from time to time, to deal with a couple of sort of cross-cutting issues. These, by the way, have generally been held under the auspices of the Office of Science and Technology Policy, OSTP.

One of these issues that we have grappled with in an interagency context has related to the potential for new plant varieties that are still in the developmental stage perhaps inadvertently contaminating traditional food crops. And the OSTP, after numerous interagency discussions, published a document setting out some tentative thoughts about that and invited public comment.

Following that, the--there was a recognition that there's a growing concern about

the use of food crop plants to produce pharmaceutical, industrial chemicals. We generally refer to these as pharm plants, spelled P-H-A-R-M, as a convenient shorthand for talking about these kinds of crops.

As some of you know, we did have an episode last year. Nothing got into food. But it sort of heightened the concerns within the government about the potential for contamination of food with non-food producing crops, such as corn used to produce pharmaceuticals.

Those discussions are ongoing. They have not come to closure yet. When they do, I expect that there will be something similar coming out of that to the OSTP publication that I mentioned a moment ago.

Of course, throughout all of these discussions, FDA is going to the meetings, participating in the meetings, with the recognition that our primary responsibility really is about the safety of the food; and we try to stay focused on that as we go through these discussions,

recognizing there are many other issues.

But the one that is primary importance to FDA, and that we like to stay primarily focused on is the safety of the food itself.

We--as you all know, we published a proposal to make our current consultation process mandatory. We have not taken final action on that. Any of you who read the proposal will know that we raised in that proposal an issue about whether we actually had the legal authority to make the process mandatory. That has continued to be an issue.

The other thing that has come along is the events of September 11th of 2001, and a need to focus a lot of attention on bioterrorism. I can tell you personally that that issue has largely taken over my life since I head the office that is responsible for developing FDA's regulations to implement that new law that was passed by Congress and signed President Bush June a year ago.

The other thing, of course, that we always face is that there are a large number of very

important issues before the agency at any given time, and we can't deal with all of them all at once.

When our Center Director, Joe Levitt, came back in 1998, one of the practices he instituted was a process, a priority-setting process, that results in something that we normally refer to as the Yellow Book, that articulates the Center's priorities.

And the A-list things are things we are going to do within the current, you know, fiscal year. B-list are things that are being worked on, and anything that's not on either of those two lists is not going to be done for a while.

We have--again, one of the other things that Mr. Levitt instituted was a process of seeking public input on what these priorities are. And we can't make everybody happy. We can't deal at any instant with anything--with everything that is deemed by even us to be important; but we do try to have this orderly process for setting priorities. And one of the things you will note is that the

mandatory proposal is not an A-list item.

The other thing that we are continuing to work on, but have not come to completion on yet, is allergenicity. This subcommittee considered some of those issues at your first meeting. We still intend to come back to the subcommittee with allergenicity issues, but we have not, you know, we were not ready to do that for this meeting.

We generally are trying to keep up with the science, though, we certainly see this subcommittee as helping us in that process. And, as you will hear more about a little later, we have, over the last four years, participated actively in the Codex Alimentarias effort at the international level to develop principles and guidelines for the assessment of bioengineered plants and also bioengineered microorganisms used in food production. And Dr. Maryanski will talk about some of that as it relates to the agenda before you today.

In the meantime, FDA is still following the guidance that we issued in 1992. We--it

continues to be our belief that all of the foods that are in the marketplace in the United States today have gone through that process; and, also, so far as we know, none of the foods that have successfully completed that consultation process is presenting any kind of public health problems to American citizens.

Nonetheless, we recognize that science changes. We want to assure that our guidance remains current. And that is one reason why

this--in fact the primary reason--why this subcommittee was created.

We want to use the very best science in all of the work we do, but I think particularly in this area where there's so much noise that is not about science. We want to be sure, and we understand people disagreeing on matters of policy et cetera. But we want to be sure that we are following the best science possible.

And with that in mind, we will be seeking your advice at this meeting, and in future meetings. And today, we're asking you to look at

some things on molecular biology--I can't even pronounce it, and I can assure you I know nothing about it, so I'm not going to say anything more about that other than to return the meeting back to the Chair, who will explain what it is we are asking for advice on today.

So with that, I will close.

Thank you very much. We are looking forward to your deliberations today, and to the advice you give us at the end of the day. And, if there's anything we can do to help, let us know. Thank you.

CHAIRMAN BUSTA: Thank you, Bob.

I know there--you have a busy schedule, and you'll be with us for--until the break at least?

MR. LAKE: Yes. I'll be here through the break; though, again, the other BT, that is, bioterrorism, as opposed to biotechnology, it causes to me leave at the break to go to some other meeting.

CHAIRMAN BUSTA: Okay. So, if we have

some questions or clarification for Bob before--

MR. LAKE: Yeah, and, in fact, I can stay through the break if anyone has anything they want to ask me privately; though, I will take any questions other than anybody might have now.

DR. GURIAN-SHERMAN: Dr. Gurian-Sherman. Bob, maybe you could clarify a couple of things: one is I think you mentioned that the mandatory process is not on the A-list. Is it on the B-list? You mentioned that things that are not on the

A-or B-list are probably not going to get looked at very closely.

The other is, do you have any kind of timeline, or rough timeline, on your progress on the allergenicity issue, as well?

MR. LAKE: Yes. Yeah, I guess I should have clarified. It not only is not on the A-list, but I've been--okay.

DR. WATSON: It's on the B-list for '03.

MR. LAKE: Okay. Well, it is still on the B-list. The '04 is still I think not finalized.

The--but there's a good chance that simply

because of everything going on, that it may not even be on the B-list for the next issue, though we have not yet issued the '04 plan.

With regard to the allergenicity, I don't know that I can give you a specific time frame, other than to note that that continues to be something that we consider to be very important; and we do, indeed, intent to come back to this subcommittee. But I'm not prepared to give you a time frame for that. I just don't know.

CHAIRMAN BUSTA: Thank you. We'll I'm on line now to talk about our charge, and questions that we have for today.

You all should have received in the mail a discussion paper and a draft guidelines of Codex; and this morning, in front of you, you have the charge and questions.

I'd like to go through this slowly and try and paraphrase it somewhat, but as you've heard this is an area that is of interest and activity with FDA.

And, as it states in the introduction, FDA

believes that it's important to the developers of products to characterize the genetic modifications introduced in those food plants.

Developers using molecular biological data to assess whether these new substances, intended or unintended, are likely to be expressed as a consequence of the inserted genetic material. There have been many advances in the field, as you've heard many times and are well aware. And FDA is seeking to determine whether any of these new advances would enhance FDA's food--FDA's safety assessment of bioengineered food plants.

So, our charge, specific charge for today is to consider the current FDA approach for the molecular characterization of bioengineered food plants, and to provide suggestions regarding additional information the FDA subcommittee believes would enhance the safety assessment. This is an assessment of the science.

Now, there's three items, issues that are listed below, and these are some. Our discussion and questions and issues are not limited to those,

but they are questions that have been put forth with our charge from FDA.

Number one: the molecular biology data provide information that assists in identifying new substances. Techniques like Northern or Western Blot have been useful in identifying newly expressed substances.

The question here is: to what extent does sequencing information contribute to the identification of newly expressed substances? And if sequencing information is important for the purpose of FDA's safety assessment, what sequence information should be reviewed? For example, the entire sequence of inserted genetic material or the sequence of the surrounding region of the plant genome et cetera.

If so, how does this information contribute to the safety assessment? The science behind that, and we've got a lot of experts here that should be able to give us some insight into that after we hear what FDA is doing.

Number two: current approaches to safety

assessment recommend certain kinds of molecular biology data. There are four that are listed

here--the number of insertion sites, number of gene copies inserted into each insertion site, the information on the organization of the DNA within the inserts, and the potential reading frames that could express unintended proteins.

The question here is: are there other data that would be useful to safety assessment? And, if so, what data, and how would safety assessment be enhanced?

The third item on the issues is: there have been many advances. Are there new advances that could be used to enhance the safety assessment? And, if so, what and how?

I'll open end, as we said, as I said at the beginning, these are some of the items. It's not limited to these, but these are considerations. From my vantage point, it seems like there's--that could keep us here for a week and a half. It's plenty open-ended. So, that is our charge.

The approach that I hope that we'll be

able to do is we will listen to the presentations on FDA's current food safety assessment and the Codex approach, and the discussion paper; and we'll have questions for clarification at the--after each of those presentations, but not significant debate at that point, or extensive discussion; but questions of clarification.

And then, we will have an opportunity for public comment. Has any public comment been submitted? One? Okay.

And then, we will, again, look at this issue and have some real opportunity for discussion and evaluation, getting into greater depth that need from the participants of the experts that are here. And then, in the middle of the afternoon, I would like to see us go round, one by one, and succinctly--I love that word--succinctly summarize our thoughts in response to the charge and the issues.

At that point, and, hopefully, we're on time or even a little ahead of time, we then--there is a little bit of agenda on the back--concluding

remarks. But then I would like to address the letter and future guidance, future agenda items, at that point.

But we will have addressed our charge, and then we can move on to the future.

That's all right? We'll move on.

Well, I'm quite proud that we're running about 10 minutes ahead of time, and since no one else has his timing agenda, no one else is sort of coming 10 minutes late, so I'm very pleased.

So, if our next presenter, Jeanette Glover Glew is ready to go, we will--oh, Mike Watson has a statement.

DR. WATSON: I'd just like to ask if there's any questions for Mr. Lake, please do them before the break so they can be part of the record?

CHAIRMAN BUSTA: Oh. All right. So we have it on--

DR. WATSON: On the public record.

CHAIRMAN BUSTA: All right. You have copies of the presentation, PowerPoint, so that you only have to take notes and not write what's on the

slides.

MS. GLOVER GLEW: I'm going to give a little caveat up front, as they showed me how to work this device, but I was not very successful in the rehearsal earlier. So, we'll see how we do right now.

Good morning. As I said, I'm Jeanette Glover-Glew. I work in the Office of Food Additive Safety in the Center for Food Safety and Applied Nutrition; and I'm going to be talking to you, in general terms, about our policy for evaluating bioengineered foods. I will also touch on parts of our molecular biology review for these types of foods. However, I'll just do a little foreshadowing here, and let you know that Dr. Tom Cebula will be going into some more detail about the type of molecular biology evaluation we do, a little bit later this morning.

Do I have to point to that? To give you a quick outline of what I'm going to cover in my presentation, I'll start out talking about the Federal Government's regulatory framework, then

talk specifically about FDA's policies and procedures. And I'll wrap up talking about some recent initiatives we've made in light of public input we've received.

In 1986, the Federal Government proposed a coordinated framework for regulating genetically engineered foods. There were three primary players: the U.S. Department of Agriculture, the Environmental Protection Agency, and the Food and Drug Administration.

Two of the primary principles based on this coordinated framework were that it would be the product that would be evaluated, not the process of its development; and that the products would be evaluated under existing frameworks. The United States Department of Agriculture evaluates agricultural food safety, doing field trials; and EPA evaluates the food safety and environmental safety of products that have, well,

pesticidal--insecticides inserted; in other words, if BT proteins have been inserted into a crop that is EPA's responsibility to evaluate.

That brings us to FDA. Our statutory authority derives primarily from the Federal Food, Drug, and Cosmetic Act. I'll be calling that the "Act" through the rest of my talk.

Under the Act, we're responsible for the food safety and proper labeling of all foods and food substances, except for meat and poultry products, which are evaluated by USDA.

That means that we have oversight over cereals, fruits, vegetables, plant by-products, such as starch and oil, milk, seafood, and other substances added to food, such as flavorings and preservatives.

There are two provisions of the Act that are particularly important in ensuring food safety. One of them is our post-market adulteration provision in Section 402. This is our primary legal tool for regulating the safety of whole foods. It means that if a food enters the marketplace, and we find it to be adulterated, for example, if it had a level of toxicant that was outside of the normal range, or if it had an

unlabeled food allergen, then we could take action to remove that product from the market using our post-market adulteration authority.

We also have the Section 409 authority, which means that an additive that is going to be added to food must undergo pre-market review and approval by FDA before it can enter the marketplace, that is, unless the product has been demonstrated to be generally recognized as safe.

If we see a product that is not substantially similar to products that have been consumed in food, then we could ask that it come in under our pre-marked approval authority.

I'll next talk to you about FDA's policies and procedures.

In 1992, FDA published a policy stating how we believed the foods derived from new plant varieties should be regulated. It was developed for genetically engineered foods, but it described how we would apply our authority under the Act to ensure the safety of all foods under our authority.

It was, as I said, derived primarily for

developers who, at that time, were using recombinant DNA technology in crop production. And those are primarily the users that have operated under that policy statement.

Let me go over some of the basic principles in the policy statement.

First, we consider the nature of the food, not primarily the method of development. Any food, including genetically engineered foods that enter the market place, must meet the same stringent safety standards as convention foods. And those traditional counterparts are the foods that we compare the bioengineered food to.

Some other basic principles in the policy statement. We recommended guidance on particular scientific and regulatory issues, and we also recommended a voluntary consultation with FDA. The purpose of that consultation was to make sure that any questions that might arise of a scientific or regulatory nature would be resolved before the developer went to market.

The core of the guidance document was a

series of flow charts related to the safety and the nutritional values of genetically engineered foods. You could see if you were a developer, you would walk yourself through this flow chart, and you would--could end up eventually in one of two boxes. The box at the bottom says no concerns, and the boxes off to the side say consultant FDA. There is also in the flow charts a third box that says basically don't even think about going in that direction; but that's not on this particular slide.

What happened, practically speaking, is most developers, when they go through the flow charts, end up in the no concerns box. However, because of our desire to encourage the voluntary consultation process, we believe that all of the people have come in to us to demonstrate what they think are the issues that, once they go through our flow charts, and to make sure we agree that they fall into the no-concerns box.

The frequency and level of consultation that we carry on with the developers is really dependent upon the complexity and the novelty of

the crop and the insert.

We encourage them to come in early and often. This is an iterative process. We believe that, by coming in early, it benefits both the developer and FDA. We're aware of what's being done in terms of the research, and they're aware of the kinds of issues that we might find of concern.

We--in our 1990 policy, we did not publish a suite of testing regimes that people could go and say, okay, I've got a tomato or I've got this type of insert, therefore, I need to do these types of testing.

We wanted this to be done on a

case-by-case basis. And so what happens is the company comes in. They say, okay, this is the information we've developed in the lab or in the greenhouse or in our field trials. And, we say, okay, well, based upon what you're demonstrating to us, we have these kinds of questions. And the company will use the input FDA gives them to go back to their lab or the greenhouse or the field trial, and develop the data to answer those

questions.

So we have an iterative process, where FDA has credible input about the kind of testing regime. It's done on a case-by-case basis.

Prior to commercialization, the notifier will come in to us with a final data package that summarizes the information they've developed, and FDA will respond to them with a letter saying either we have--still have remaining questions or we have no further questions.

The approach in these consultations is multidisciplinary. We have teams of individuals who review these submissions. They include molecular biologists, several of which are here today, microbiologists, chemists, nutritionists, food scientists of various kinds; and we cover a broad spectrum that's provided, everything from the agronomic characteristics to the compositional characteristics of the crop.

In this next slide, I've just presented the information in this way to emphasize that we look at both at the intended changes and the

unintended changes to the crop. In the intended changes, of course, we're going to be looking at information about the trait--desired trait that was inserted and to the kinds of alterations that might make. And we look at information about the background of the crop, and the exposure.

However, we're also looking for unintended changes, because this tells us whether or not cryptic pathways have been turned on. We'll look and see whether or not there's been changes in known toxicants or nutrients to see if there's something that lets us know that a genetic pathway has been interrupted by the insertion, and that there's something that we need to be paying special attention to. And another way that we do that is to look at the genetic stability over time.

I've added these next four slides just to go over real quickly because it gives you the elements of the safety evaluation, including what we look at in terms of the molecular biology evaluation. But Dr. Cebula is going to be going into some more detail later on; so this is a just a

skim to prepare your minds for what we're going to be talking about.

In the host plant, we'll look at the taxonomy, history of safe use, the presence of any naturally occurring harmful constituents, and important nutrients. If we were going to be looking at an orange, we'd expect Vitamin C to be there.

In the donor organism, we're going to be looking at similar characteristics, but we're also going to be looking through--whether or not any baggage has been picked up due to passage through microbial host. And, of course, most importantly for the donor organism, we're going to want to know about the identity and function of the introduced material.

For that introduced material, we're going to look at concentration, so we know what dietary exposure is.

We're going to look at the potential for it to be an allergen or a toxin. I already mentioned that we're going to be looking at

similarity to other substances in the food supply. Talked about looking for changes in metabolic pathways.

We're going to find out if there's

post-relational modification, such as glycosylation of a protein. And something that I'll throw out here. Other countries have been concerned about the presence of antibiotic resistance marker genes. And we also evaluate those.

And this comes to the information that is more the basis of what we're going to be talking about and getting input from you on today is the inserted genetic material. And here, we're going to look at the method of transformation, what the regulatory sequences are, the promoters, the terminators, how those are working, whether there's the chance of open reading frames that might read through into the indigenous genome, how many inserts they are, where they are, if they, you know, flipped or they inserted into a site that may create problems; and then we're going to be looking at Mendelian inheritance, so whether or not it's

genetically stable.

I'm going to switch gears here, and just for a few minutes talk about some of the recent initiatives that we've taken as a result of a public outreach.

We know that the biotechnology field is rapidly changing. That's why you're here. That's why we want your advice. But also, in 1999, we decided to hold a series of three public meetings. And we said, we're going to look at our 1992 policy to see if it needs to be modified in any way. And, as a result of soliciting that input, we received over 50,000 comments. Excuse me a moment.

We were pleased to find that, as a result of those comments, that we really hadn't received new data to question the safety of bioengineered foods that are on the market today. However, we heard concern from public, academics, public policy groups about the future, of whether or not the products that were going to be coming down the research pipeline were such that we had either the regulatory or scientific framework and background

to address the safety of these products.

So, as a result of that solicited input, we have taken several initiatives. One of them is you guys. That's why we're here today. We established the Food Biotech Subcommittee, and we hope to be using you for this purpose quite a bit in the near future.

We have committed to support research on allergenicity, and we have combined with USDA and EPA to support a study by the National Academy of Science on potential unintended effects in bioengineered foods.

Before I wind up, I'm just going to give you a snapshot of the types of products we've seen. The corn, cotton, and soybean are the primary commodities that we've seen in. We've seen other minor crops. The kinds of traits that have been introduced have been primarily agronomic traits that are of interest to the farmer, either pest resistance or herbicide tolerance. Though we expect, in the second and third generation of crops, to see more complex physical traits, such as

salinity, tolerance, or drought resistance; and also traits that would be of interest to the consumer, such as increased nutritional value.

The goal of us at CFSAN and here today in talking with you is to make sure that we're prepared to meet our scientific and regulatory responsibilities in the future. So we're really looking forward to you looking at the charge and the questions and advising us as to where we can to specifically in the arena of molecular biology techniques.

And I just wanted to let you know that I've given you our website. If you want to look at any of the documents that I've mentioned today, either our 1992 policy statement or the proposed rule that Mr. Lake mentioned, or a list of the products we've developed, including our scientific memos evaluating those products, you can go to the website and find that information.

But I think we're running ahead of schedule, so I have time to answer any questions you might have before we go on to the next speaker

at the break I guess.

DR. GURIAN-SHERMAN: Jeanette, I had two questions, clarification. One is concerning the website. That doesn't have, as far as I know, any of the more detailed aspects of the consultations, in other words, the data summaries, which I found very useful to look at to kind to get a better idea of what is actually submitted and what you're actually doing with it.

MS. GLOVER GLEW: That. That's a--

DR. GURIAN-SHERMAN: So one question is: if committee members want to see any of those, could they be made available and how could that be done?

And the second question is: If you could clarify a little bit kind of the case-by-case approach. How do you determine kind of what the baseline minimum assessment should be? And to put that in context, for instance, when I looked at some of those studies, just as one example, two of the BT crops, looked at whether phytate levels and known anti-nutrient in the corn had changed, two

did not. You know, so how do you determine what's enough? Which anti-nutrient should be looked at? Which allergens, you know, et cetera, under the current system?

MS. GLOVER GLEW: Okay. Try and remove the first question before I get to the second.

The first question has to do with the fact that I mentioned that at the end of the consultation process, we receive a data package that summarizes the material, and this is what we use when we are completing our evaluation.

That information is not currently available on our website. It is available by Freedom of Information Act request, which is how Dr. Gurian-Sherman got it.

We--I did not go into all of the efforts that we have committed to make as a result of the public meetings. I wanted to focus primarily on the scientific nature of some of the concerns that were expressed, particularly the ones that you can address today in the molecular biology discussion.

However, we have made some commitments to

try and have increased transparency. Those are something that we have to balance with our obligations to protect confidential business information. But we have--our ongoing discussions. I honestly don't know, Doug, whether or not the Food Advisory Committee can get copies of those documents outside the FOI process, and perhaps one of the people on the FAC Committee can help us with that a little bit later.

The second question: how do we get a baseline? That's a good question, and the thing is it's like mini-targets as it moves. And I'll use your phytate example; is that when some of the earlier consultations, we would talk about the areas of concern--the nutrients, or the toxicants, or the anti-nutrients--and the company would go away; and they would develop that information. And, this is in addition to general principles, which we outlined in the 1992 policy, about looking for allergenicity, examining whether it's a potential for it to be a toxin.

So we have the general principles, and

then we have some of the more specific give and take on the case-by-case basis. And the answer about the phytate is that it does--the target moves.

As we've been involved with Codex, as we've been involved with our European counterparts, particularly Dr. William Price, who is in the audience today, has been participating in committees that are laying out some of the known important nutrients, anti-nutrients, and toxicants; and we're able to now use these kinds of materials to give clearer, better, more particular information.

So, you can look at earlier submissions, and you'll find that they're missing information that perhaps are in more current submissions. And that's because, as the science advances, as we are more aware of the types of information that are important for us to look at, we would be

giving--tailoring our information.

Does that answer your question?

DR. GURIAN-SHERMAN: Yeah, I guess the only follow up I would have is, if you're

developing some of this internal guidance or guidelines, is there any plan on making them publicly available? Are they just still in the development stage?

I think in 2000, NAS recommended that you do that with, you know, come up with a database of known toxicants and anti-nutrients. So, you know, is that something that's going to be made public?

MS. GLOVER GLEW: We're working internally on guidance. It suffers from the same resource problems that Mr. Lake referred to earlier, but one of the things that, if everything works out in terms of our resources, and how productive we think this would be for industry, we might use the information that we develop today for a molecular biology guidance.

The compositional guidance we have held off a little bit on because we're waiting to see what the NAS says about the unintended effects. So, guidance documents are something that we constantly have it in the back of our head in order to give guidance to industry.

But, as we've just talked about the phytate issue, we weren't--we didn't know that that was something that we probably be asking for perhaps early in the process. We know now. So, if we have guidance and it's hard and fast, that leaves us a little less flexibility in order to look at these on a case-by-case basis. So there's advantages and disadvantages to having written guidance.

CHAIRMAN BUSTA: Abigail.

DR. SALYERS: I have a question for you, and for Mr. Lake, both. It has to do with priorities.

If I were prioritizing issues on your A-, B-, C-lists, I would put bioengineered foods on the E-list, and I'd put bioterrorism on the F-list. You know, as in forget it, you know.

And I would put on the A-list things like probiotics and functional foods that have been largely un--not looked at; and also just good old early detection of microbial, unintentional microbial contamination of foods. And so that we don't have to pull things back after they've

appeared on grocery shelves.

And so my question is, and I understand why you've had to put bioengineered foods and bioterrorism on the A-list, but so I'm going to

be--this question is kind of pessimistic. I know you're stuck with that because of political considerations, but are you thinking about ways that you might use whatever money you're pouring into those areas to also address some of these other issues, the real, which in my opinion, the real public health issues, the things that sicken and kill people at still somewhat horrifying level; that there are ways to divert some of this attention and finances into doing something in those areas.

MS. GLOVER GLEW: I'm going to let Mr. Lake answer the part about dealing with substances that might kill people.

DR. SALYERS: I'm not, I'm not--this is not a hostile question.

MS. GLOVER GLEW: I'll answer something that perhaps is underlying your question, and you

may not have intended this to underlie your question, is that we have for bioengineered foods a process that is working, and that we believe is protective of the public health.

So, there are many ways that we would like to expand our abilities to be responsive to industry in giving them guidance about molecular biology. That's why we're here. That's why we're interacting with the National Academy of Science. There are things that we would like to do to make certain that our regulatory framework and our scientific background is what we need in order to evaluate these things. And I think that we have that, and we'll continue to have that in terms of Center priorities.

DR. SALYERS: Well, let me just give you an example. This is one of these things that happens to me accidentally. It makes me even more unpopular than I am already.

But I happen to think probiotics is a good idea, but I found out, sort of accidentally, that most of the strains of bacteria that are used for

probiotics are resistant to vancomycin, which is one of our less antibiotics for treatment of certain types of diseases.

And so I said to them, to these people, I said, have you--now are you talking about markered antibiotic resistance genes in transgenic plants, which are for, you know, effective against antibiotics that basically are not a problem any more; but here's some vancomycin resistance in these probiotic strains.

And so I got this storm of anger out of that. And I said, but you really ought to ask yourself is this just some metabolic thing or some, you know, could possibly pose a public health risk. And that, as far as I could tell, was not under investigation anywhere.

And I just wonder how many more examples of that kind are out there in the probiotics, the functional foods area; and then, of course, we have the problem of is there some way that we could try to detect contamination of foods before the food actually gets on the grocery shelves. So that's

the reason I was asking this question.

CHAIRMAN BUSTA: And I'm being very good on chairing this, and letting you ask these questions before we lose Bob. But that wasn't a terribly expansive question on clarification. Bob.

MR. LAKE: Nonetheless, I think it was a good question, so let me comment to the extent that I can.

Though we try to do priorities in a way that we think is rational from our perspective, you know, there are a lot of forces out there, including congressional mandates that are, you know, in the case of bioterrorism, very explicit.

And one of the things that's driving my life now is that Congress, and the law that I mentioned earlier, the new bioterrorism law has a couple of regulations, a couple of requirements, that will go into effect on December 12th of this year--one on registration and one on prior notification on imported foods--that will go into effect whether FDA has regulations on the street or not. However, we have been mandated to write

regulations, and it is our strong belief, and I think the belief of everybody we're hearing from in other governments and the industry and whatnot, that, if we fail to have regulations on the street, that there will be chaos; that everybody needs those regulations in order to implement this law in an orderly way.

So, whether we or anyone else, you know, necessarily would have chosen those as being a top priority, it is certainly clear that the events of September 11, 2001 have altered the state of play in a very significant way for FDA and for, frankly, all of the Federal Government.

But also, I think aside from the very real issues, which I personally believe need to be addressed in the area of bioterrorism, I think we're facing a situation with the registration and prior notice activities, and I will be going

here--from here to a meeting about those because we really need to publish those by October the 10th if we are to have an adequate lead time. And we've had a huge effort ongoing, and hopefully we'll meet

that date.

So that's one thing that's kind of out of our control. But also, it's just one of the realities that, you know, the world sometimes gives us things that we didn't expect.

With regard to--when you talk about probiotics, I assume you're tying that back to genetic engineering. And I guess my general comment about that is that we try to keep up with products that are actually entering the marketplace. And that, of necessity, I think sometimes means maybe we're lagging a bit on things that are under development that are not quite close to being marketed.

I should, however, note a couple of other things with regard to genetic engineering. We do have, you know, on the pharmaceutical side of the house, a lot of activity that has been ongoing for a number of years. And I'm not, you know, qualified to speak about all of that. But just keep in mind that bioengineered products are not just about foods; they're about other things. And

the Agency has people heavily involved in many of those areas.

The other thing is, just to comment on your issue of methodology. That is something that is receiving attention here and internationally, as well. But, you know, what we are really focusing on here is our job of evaluating the safety of in particular new plant varieties, and being sure that we're doing that in the right way. And that's why we're seeking the advice that we're seeking here.

DR. SALYERS: Well, the reason I asked that question, which our esteemed Chairman obviously thinks is not germane to the issue at hand, and I think you sort of don't either.

But my point is that, if we're going to talk about technology for evaluating safety, fortunately because of the new types of technology that are available, many of the things that we would talk about here today have broader implications; and maybe we should be aware of those possible broader implications; that there are broader uses. And that was the only reason I

raised the issue.

MR. LAKE: Well, no, I think, you know, I mean you're raising an important point, and the Chair has added to the agenda, if I'm understanding it, the issue of what are topics that this subcommittee perhaps might address in the future; and maybe, at that time, you could, you know, put some of those thoughts on the table. And we're certainly anxious to hear from all of you on your thoughts, and we'll certainly consider them.

By the way, in my thing earlier, I did fail to mention the one thing that Jeanette did briefly touch on. And after I'm gone, if you have any further questions, Dr. Maryanski can address them; and that's the National Academy of Sciences study on unintended or unexpected effects. And, as Jeanette pointed out, that is being co-sponsored by USDA and EPA, as well. And we are certainly looking forward to that input. Thank you.

CHAIRMAN BUSTA: Anne?

DR. KAPUSCINSKI: Thank you. Thank you, Jeanette. That was a really nice, clear overview

presentation.

My question has to do with when you talk about that you are using the language we look at, for example, when you were showing the slide about intended modifications and unintended modifications, I'd like to just get a better sense of how the look at process really works. When--if you're going into a consultation with a developer, are you primarily going to look at what that developer on their own has though is necessary to do to look for both the effects of the intended modifications and possible unintended modifications? Or, to what extent are you--and are you going to basically look at what they provide to you as they think is the information and the issues they should be asking about. Or, to what extent are you going in with a set of things that you think should be asked about, given their specific case. And then, I guess, as a follow up to that question, how do you envision that might change after the National Academy report on unintended modifications comes out? And I'm sort of focusing

on the unintended modifications, because that's clearly, I mean, from my perspective, that's sort of the main bug-a-boo in this area. And, personally, I'm interested in, you know, helping the agency come up a way that those unintended modification issues can be addressed in a way that's not going to completely paralyze the Agency. I don't think it's realistic for the Agency to be the one that actually carries out the risk studies and all the assessments. I think the burden of that has to be on the developers. But I'm interested in helping the Agency figure out what's a good balance so that, you know, there is some kind of independent--if it's not a standard checklist, which may not be always useful, at least some kind of independent process of sort of going through, an eliciting of, you know, have we thought of this. Have we thought of A, B, C, and D to make sure that the developer has done an adequately thorough job.

MS. GLOVER GLEW: Okay. Excellent. For a while there, I was going, yes, yes, no, yes, to

your various questions. But since I've lost track of which are yes and nos, I'll try to answer it in a more global sense.

As I mentioned earlier, there are specific principles and flow charts in our 1992 guidance document that walk the developer through the types of considerations they should be evaluating toxicity, allergenicity, is there a marker gene, is the marker gene expressed for an antibiotic that, you know, is valuable clinically. So we have principles in general in our 1992 policy.

Then, when the company comes in for the early part of their consultation, this iterative process, we develop collaboratively what they're going to be doing. They usually will bring us some early data, and we'll look at it. And in terms of unintended effects, we may be looking at a Western Blot, and see that there's a protein that, you know, doesn't belong. It's the wrong shape, size, and it's showing up on--and then we'll be going, what's that? And why is it there? And what's your explanation for that?

And so, they'll go, okay, well, we, you know, sequenced this little piece, and we found out that this happened. So--it--we have--so what

I'm--I guess I'm trying to say, we have principles that people follow using our 1992 policy, and our 1996 consultation procedures, which are also on the web. Then we have an iterative case-by-case discussion; it's a chat that we have with these people. They tell us what they're doing. We tell them the kinds of regulatory and scientific questions we think arise based upon what they're doing. And they will go out and develop the data that answer those questions, come back to us with a data package, and we'll say, okay, these are the questions we asked. Did they address them? Are more questions raised by the information that they developed? Do we need to go back again?

And so it's a collaborative, iterative process based upon established principles, plus a case-by-case interaction.

And--has that answered at least a piece of your question?

DR. KAPUSCINSKI: Yeah, the second piece I guess is I was just wondering how you envision the, you know, the recommendations you get from the NAS? Do you envision that, at this point, changing this process in some ways and?

MS. GLOVER GLEW: I just don't have a good picture in my crystal ball about that. We're real excited about the potentialities for what we can get, and how we might use that. We're hoping that they will do something similar to what you're doing today, and say, okay, the technology has changed enough that we can recommend that you ask for this type of information or that they develop this type of data, because this will provide valuable information. But I'm not sure what's going to happen with that. But we're looking forward to it.

CHAIRMAN BUSTA: Nina?

DR. FEDOROFF: My question is partly historical and partly contemporary. It has always puzzled me that the EPA was approving or not approving StarLink corn based on the potential for allergenicity of the cryonide seed. That seems to

be your bailiwick, because it's environmental.

MS. GLOVER GLEW: Your confusion is understandable. I didn't spend a lot of time at the very beginning talking about the regulatory framework because I wanted to speed through into FDA's part. EPA is responsible for evaluating both the environmental and the human food safety for pesticidal proteins, for plant incorporated protectants. And so this is mostly BT products. It's the way the framework of the law is divided up. We will look at the product and see if there are any considerations in terms of compositional changes that might require, like, different labeling. So we do evaluate the product from FDA's perspective, only we love the fact that on StarLink, we can do this huge burden shift. It wasn't our responsibility.

DR. FEDOROFF: Pardon?

MS. GLOVER GLEW: It wasn't our responsibility.

DR. FEDOROFF: Yeah, I noticed, though--on the other hand.

MS. GLOVER GLEW: Because of—

DR. FEDOROFF: On the other hand, is there any effort to make the criteria uniform? Is there any effort to reconcile your criteria and theirs?

MS. GLOVER GLEW: Absolutely. Mr. Lake mentioned the fact that we have collaborative efforts that are ongoing. We are right in the middle of discussions with EPA and USDA about what our requirements--I say requirements loosely since it's a voluntary program--but what types of information we look at in the data packages that we get. Mr. Watson, Dr. Watson is working on that right. He developed this table, and we pay very close attention to the kinds of information that EPA developed in terms of allergenicity detection for the StarLink, because that's something that's real pertinent to us. Those of you who were on the previous meetings, subcommittee meeting, you know, heard about what we're doing about allergenicity, and the kinds of considerations we're taking. A lot of that came from our interaction with our counterparts at EPA and their experience with

StarLink.

CHAIRMAN BUSTA: Bob?

DR. BUCHANAN: A general question: how many new conventional foods and genetically engineered foods do you approve per year approximately?

MS. GLOVER GLEW: I'm sorry. I didn't.

DR. BUCHANAN: How many new conventional foods, for example, a new type of nut that existed somewhere, and it's sought to have approval in the U.S., and how many genetically engineered foods do you approve per year?

MS. GLOVER GLEW: Per year. Okay. We have since 1994, when we first started receiving these biotech notification files, or BNFs, evaluate a little over than 50. If you're talking about do we have some brand new novel, never-before-seen food on the marketplace, I don't think that you could characterize it that way. It's mostly people who have taken maize, soybean, cotton, and they have added an herbicide tolerant, or a pest-resistant chain to it. So it's something that, if

you drove by the field at 55 miles an hour, you'd look out there and say, that's cotton.

So--just--an entirely novel crop isn't something that we see. We see mostly modifications to existing commodity crops.

DR. BUCHANAN: So there's no new macadamia nut or something like that?

MS. GLOVER GLEW: No, not at this point in time.

DR. BUCHANAN: Thanks.

MS. GLOVER GLEW: That could be coming down the research pipeline. Who knows?

DR. FEDOROFF: I think your question was whether your approve, when someone brings in a new whole food from a different part of the world that has not been marketed here before, wasn't that the nature of your question?

DR. BUCHANAN: Right. Yes, right.

MS. GLOVER GLEW: Okay, so it's like--

DR. FEDOROFF: I misunderstood your question. Do a kiwi fruit--

MS. GLOVER GLEW: --If I showed up from

New Zealand.

DR. BUCHANAN: Yeah, kiwi is a great example.

MS. GLOVER GLEW: Okay. If--I mentioned that our post-market adulteration provisions were the primary legal tool for evaluating whole foods. If something came on the market today, and we didn't have any information indicating that it

was--had toxicant, or wasn't allergic, we probably would have no need to take action; and so we would not utilize our post-market authority to take action against that crop. However, if something came on the marketplace that was just totally novel for the American food supply, we have our authority under the Act, totally aside from the bioengineered food process.

DR. BUCHANAN: Yeah, that was the--

CHAIRMAN BUSTA: That's separate from the biotechnology activity. So I'm not sure that--

MS. GLOVER GLEW: Yes.

CHAIRMAN BUSTA: I'm trying to keep us obviously focused--

MS. GLOVER GLEW: Well, I'm trying to explain is that we still use our authority under the Act for bioengineered foods. If we needed to see something that was adulterated or, you know, if we saw that something needed to come through the petition process, we would do that for a bioengineered food. So the Act applies uniformly. We have the same safety standards for all foods the FDA evaluates.

CHAIRMAN BUSTA: First of all, before I call on Doug again, are there others who have not asked a question that want to--clarification questions?

Or, last call for Bob Lake.

Doug?

DR. GURIAN-SHERMAN: Yeah, I want to make two points of clarification. They're not actually questions, but they're both points of clarification. In terms of new foods, I know of at least one that's come on the market in the last few years in the U.S. It's called Kworn. It's made with fusarium finatum. It's a soil fusarium. It

was never the food supply before; was marked in Europe, and did go through the grasp process at FDA that did have some toxicity studies, animal studies, et cetera that that was subjected to. That was, again, grasp studies, voluntary.

To Anne' question, I would just like to venture, and I know the committee isn't really privy to this, but when we did our study looking at FDA, we FOIAed not only the actual data packages that FDA looks at, but also the communications between the developers and FDA. And I won't say that it's impossible that I missed anything, but only found about six examples where FDA, on record, asked the company, let's say for more data or for clarifications. And three of those went unanswered or were not, we felt, adequately answered. So, there certainly is give and take. Maybe some of it isn't on the record. It should be. But there's not, from our perspective of what we saw, a lot of information to indicate that FDA is pushing the companies for more data.

CHAIRMAN BUSTA: Doug, to keep us--

DR. GURIAN-SHERMAN: So that's—

CHAIRMAN BUSTA: On line that's good discussion comment, but not in the presentations. We have this afternoon to get into discussion. We're still in the clarification--questions for clarification. Dennis?

DR. GONSALVES: Well, I just was going to add, you know, the question on the whether FDA asked questions that had an impact. Well, when we were deregulating the transgenic papaya, we actually were moving two lines through the process. And when we hit the FDA, another line had some other inserts in it, and because of the questioning that we needed more data, basically we withdrew that line, because we did not have enough information. So the point here is that--the experience that I had was that the questions that FDA did ask had an impact on our actions to withdraw one of the papaya lines.

MS. GLOVER GLEW: Just a quick remark. I'm sorry to interrupt. I just want to say that FDA, along with all the other regulatory agencies,

occasionally struggles with the do we want to know this or do we need to know this. And occasionally, we do err on the want to know, but primarily believe that when we ask questions, it's in the need to know area. And if a company or developer did not provide us with the information that we needed to know to evaluate the safety of that product, we would not give them a letter saying that we had no more questions.

CHAIRMAN BUSTA: Any more clarification. James?

DR. ASTWOOD: Thank you. I have one question, and it kind of relates to Dr. Fedoroff's earlier question. Would you agree that, although it appears complex that both the EPA and the FDA share jurisdiction over plant incorporate protectants, or BTs in particular, that when you look at the history of the regulation of pesticides and the experience that the EPA has with food-feed environmental safety of chemical pesticides, and then the subsequent history that they have had regulating and evaluating the safety of microbial

pesticides, including BTs, that it makes sense, in fact, for the FDA and the EPA to collaborate and share responsibility for the evaluation of those kinds of products.

MS. GLOVER GLEW: Yes, I think the way the roles are laid out right now has been working very successfully.

CHAIRMAN BUSTA: I have one last question. If the use of biotechnology has developed to, let's say, double a specific nutrient--

MS. GLOVER GLEW: I'm sorry. I'm not hearing you.

CHAIRMAN BUSTA: If one uses a biotechnology to double a nutrient, a given vitamin or whatever, this is arbitrary, would that be in the same consideration?

MS. GLOVER GLEW: As I just mentioned briefly is we believe that some of the products that are going to be coming down the research pipeline are going to be products that are targeted to consumers who want to have increased levels of, you know, a fancy nutrient or something. And

certainly a product that was bioengineered and had an alteration in its nutrients would--we would request that they come to FDA during our--are part of our consultation process. And if there was a substantial change in nutrients, we would probably have to consider whether or not the product would need different labeling or a different common unusual name.

CHAIRMAN BUSTA: We're at break time. I just would like to--before we come back say that what I'm trying to do is keep us on focus. Our charge is to consider the current FDA approach to the molecular characterization of bioengineered foods and make suggestions for additional information that will be needed. We'll try to stay with questions of clarification for the next three presenters. And then we can get, and after public comment, get into depth of discussion and where we want to--what we want to tear apart and put together. Let's take 15 minutes. Let's start promptly at 10:14 on my watch.

[Whereupon, the meeting went back on the

record at 10:15 a.m.]

CHAIRMAN BUSTA: May we reconvene, please. The next presentation is on the Codex Approach, by Dr. James Maryanski. I have been negligent in my chairing to remind you all--I've just been calling you by your first name, and it's appropriate that individuals give their name before you make your comments or you ask your questions so that it's in the official record. I think the stenographer, being that no one has the same first name, we're all right. But if would be appropriate that everyone give their name as they speak. And with my voice, I rarely have to do that. Most people, whenever I talk, people know who I am. I'm--I guess I've stalled long enough. We're about ready to--

DR. FEDOROFF: Almost everyone's back.

CHAIRMAN BUSTA: Almost everybody's back.

DR. FEDOROFF: Not quite. She's still getting coffee.

CHAIRMAN BUSTA: Yeah, and we will have Dr. Anne Kapuscinski read her name into the record

when she returns, because she wasn't here on our initial round. But she is here for the record.

Dr. Maryanski?

DR. MARYANSKI: Okay. Thank you, Mr. Chairman. Good morning, ladies and gentlemen.

I have the pleasure to tell you a real success story in biotechnology this morning. You know, we hear so much discussion that's often seems either difficult or even contentious; and you often hear of differences between our country and other countries around the world on biotechnology.

I want to tell you a really different story, a story where everybody actually came to agreement at the end of the day. So this is really a pleasure to be able to tell you about the new guidelines that have been developed on the--in the international community for assessing the safety of foods developed by biotechnology, and one which we at the Food and Drug Administration have participated in.

But let me first tell you who Codex is. Many of you, I'm sure, know, but, in case you don't

know, the Codex Alimentarias Commission is the body that has been established under the U.N. system by the parent organizations--the World Health Organization, the Food and Agriculture Organization. Codex Alimentarias is Latin. It means food code. And, as such, it is a collection of internationally adopted food standards and guidelines, and those are intended to ensure consumers' health, as well as to facilitate free trade. And, so, this is a body that has been in existence since 1962. There are about 169 member countries, so this organization has global representation around the world in both developed and developing countries, as well as NGO representatives that include public interest groups, as well as industry. And so it's a very broad organization.

Woops. Wrong way. Sorry. In 1999, the Codex decided that it would take on some work in the area of food biotechnology, and it established a temporary committee. Codex has lots of committees that work on various things. But it

also occasionally establishes temporary committees for specific purposes; and these are often called ad hoc task forces; and, in this case, with the long title of the Ad Hoc Intergovernmental Task Forces on Foods Derived from Modern Biotechnology was established by the Commission in 1999. It was given a four-year charter, and Japan was selected as the host country to facilitate this committee. And the Japanese Ministry of Health, Labor, and Welfare was the part of the Japanese government that ran the secretariat, essentially helped organize the meetings for this.

The task force was chaired by Dr. Ushikora, who is world renowned virologist, and is fluent in both English and French, and Japanese, of course; and did a superb job of negotiating through various complex issues during discussions. And I think everyone involved recognized that the Government of Japan deserves a great deal of credit for the success of this work.

Japan did host four meetings of the Task Force that just concluded this past March in

Yokohama. And the recommendations that came out of that work then were presented to the Commission at its meeting this summer.

The Task Force was given, in terms of reference by the Commission, which you see here, very general terms, consideration for the elaboration of standards, guidelines, or other principles for foods from biotechnology. So a very broad mandate. The Task Force, at its first meeting, in looking at this mandate, felt that there are many things in the word of food biotechnology, and it would have to somehow identify work that it could accomplish within four years so that it would have something finished at the end of that time; and do that in a way--on a priority basis. And they felt that since foods derived from crops were already in the marketplace that that should be the top priority; and that they would focus their work on developing guidelines for assessing the safety of those foods for human consumption.

They also felt that foods developed using

organisms modified using recombinant DNA, using microorganisms modified with recombinant DNA techniques, would also be on the frontier and of a priority.

Animals, of course, were well understood to be in research, but they felt they were not quite ready to discuss animals yet at this time.

So that was the order of work that was set out by the Task Force.

And the work that was accomplished I'm really showing you here the final product, and then I'll tell you what it means. But the Task Force, within its four years of work, developed three documents.

One was a--the principles for risk analysis you see here. That's really an umbrella document that talks about how risk assessment and risk analysis can be carried out within the existing framework of the Codex. And it's thought to be sort of an umbrella for the guideline documents, but the real crux of the matter, the important work, are the two guideline documents.

The guideline on foods derived from plants, and the guideline on foods derived from microorganisms. Both of these guidelines have an appendix to them as well that deals with assessing the possibility for allergenicity for new proteins that would be in the foods derived from the plants or microorganisms. And so those also are important documents that go with and are part of these guidelines documents.

These documents were presented to the Codex full body at its meeting this summer, and were adopted unchanged. And they are available, at least in their draft form, on the Codex website. They will be published simply taking off the draft. So they are the texts that--I believe we have given you the texts for the guidelines for foods derived from plants, because that's the one we'll be focusing on for the discussion today.

Now, I want to talk about one of the principles that the Codex felt was important. The Codex system has a number of defined terms--risk analysis, risk assessment, risk management, and so

forth--that are all well understood withing the food safety system and Codex to deal with things like pesticides and food additives and contaminants in food, essentially single chemical substances. And often times, certain of those substances will have some hazard associated with them and how to manage that in terms of tolerance setting, for example, for pesticides. But it was recognized early in the discussion that we're dealing with food crops when we're talking about foods derived from crops, and trying to establish guidelines for safety assessment. But these are crops we're well familiar with--corn, potatoes, soybeans, and so forth--that are modified in some way, using recombinant DNA techniques; and that often when we go through the process of evaluating these, we don't find a hazard. In fact, that has been in the case in the 50-some varieties has looked at.

And so we felt that there really was a new term that was important here. And so we established the term "safety assessment" as a way of having something to focus on that really takes

the emphasis on making sure that this food is as safe as other foods, but doesn't necessarily find a hazard or a risk. So, if a hazard is identified, and, of course, there's always the possibility that that could be the case, then that hazard, then, would be treated as it is in the Codex system. You wold have to decide does this food have to be removed from the marketplace, or are there conditions by which the food can be marketed, recognizing the nature of the hazard; for example, establishing a tolerance for a pesticide as a way to manage a risk. And so, if there's a hazard identified, then the system would work just the Codex systems normally operates. But the focus that we're going to talk about is really on safety assessment.

So safety assessment, then, is something that the Task Force felt was important, recognizing that it does fit within the Codex framework. But it's--this safety assessment process is really a comparative process. It is one we recognize that the standard that we have are the crops that are

already out there. The foods that we have accepted in the marketplace are the gold standard. That's what we consider to be safe, recognizing that no food is perfectly safe or necessarily safe for all individuals all of the time. But, in fact, the process, then, is to compare the new food with what has gone before, with its counterpart. And the purpose of that is to try to identify whether there are any differences in the new food derived from this newly modified plant that would have any effect on health; and where those differences are identified, to make sure that those differences are also safe for the consumer.

So the idea is to look at what's new; make sure that that's safe for the consumer; make sure that the food is still what it's expected to be, because this is food that we have consumed and have a history of consuming. And this comparison should take into account both the intended changes that have been made in the plant, as well as, to the best that one can, to look for any unintended changes that may have occurred; and make sure that

if they have occurred and can be identified, that they are also safe.

This is what we call substantial equivalence, a big, long term that is often confusing, but a very simple concept. We don't decide at the end of the day that the food is substantially equivalent. That's not the decision. This is the tool at the beginning of comparing what's new; what's different about this new product, if anything, and making sure those differences are safe. That's what we mean by substantial equivalence.

The Codex has adopted a number of definitions so that they would have terms to use in the documents that would be understood by all the countries. Modern biotechnology was adopted as the general term, and that was adopted because it had already been defined in the Cartagena Biosafety Protocol, so there was a large agreement on the definition by many countries, and so it was felt that there was no need to create a different definition and that would work for the purpose of

this work of elaborating food safety guidelines.

The Codex also did something that I think was very useful. They avoided all the terms that cause everybody a great deal of angst, such as genetic engineering, genetic modification, GMOs, and other kinds of terms that would be difficult to get consensus. And, so, they simply referred to recombinant DNA plant or recombinant DNA microorganism where they needed to refer to the modified plant.

And then they did, of course, design--or define conventional counterpart because we have a comparator, they wanted to make clear what we mean by the comparator. The comparator is a food or a component of food for which there is a safe history of use. And the Task Force recognized that, in some case, there might have to be more than comparator used in a particular assessment.

But those are the definitions. I don't want to spend much time on those, just to make you aware of them.

And I'll quickly just go through some of

the elements of food safety assessment that are laid out in the Codex guidelines. These will be no surprise to you, but it is, of course, asking what the plant is. What's known about the plant in terms of its use for producing food? How has it been modified? What are the new techniques that have been used?

The Codex guidelines use a method of first describing what the developer intends to do, and then going on to characterize what was actually done. So you will see paragraphs that start description of molecular characterization, for example, or genetic modification rather; and then characterization of the same thing. That's the reason for the difference. The idea is, well, they intended to do this. This is what they started to do. Now, what did they actually do? What do you find actually occurred. So that's how the document is set up.

In terms of safety assessment, of course, use for substances that would be in the food are really one of the major keys to safety assessment,

and are those substances likely to have any impact on public health, including toxicity or allergenicity. As I've said, there is a separate guidance that deals with allergenicity. I think that today that annex represents the best consensus that there is internationally on how to approach the question, is a new protein likely to be an allergen.

The compositional analyses are part of addressing unintended effects, a part of answering the question, is the food still what we expect it to be?

Evaluation of metabolites: if there are new substances of food that break down into other metabolites, obviously one wants to be sure that any other substances are also safe.

Effects on food processing may be taken into account. In some cases, of course, such as a refined oil, there may be no protein in the food. Or there may be other cases where the processing may concentrate things in the processing of the food.

Nutritional modifications. Obviously, if the food is modified in a nutritional sense, it has to be looked at in terms of the overall diet of what people eat, not necessarily just the modification of this particular food, but how does that affect the diet and what are its implications; and other considerations, such as the use of selectable marker genes and so forth.

Now, to get more to the topic of today, the guidelines do set out, in a number of paragraphs, first, as I said, describing what the developer has done in terms of a genetic modification, what is the host plant that's being modified, what are all the materials that are contributing to that modification, in terms of the source of the organism; the genetic material that's obtained from those organisms and how that's combined and introduced into the plant; what is introduced; what its size is; what its function is and so forth; and then going to what actually happened in the plant, characterizing the plant in terms of--at the molecular level, what are the

genes that have been inserted; how many sites; how many copies.

Using sequence data to identify open reading frames, for example, so that one can get a sense of what likely substances are--can be expected, and will those be present in the food.

The Codex discuss the sequencing of the inserted region and the border sequences, as well. And they also recognize that there are other ways to get at the question of expressed substances, looking at transcripts and expression products as a way of keying in to what might be new substances that would be present in the food.

So I think in terms of looking at new substances, the molecular data being used to both establish what is the function of this substance; what's the phenotype of the plant; what is the level of expression. In other words, one of the common food safety questions is always how much do we eat. So are these metabolic enzymes or are these substances that are seed-storage proteins. What will this mean in terms of actual consumption.

In some cases, we can think back to the old flavor saver tomato. We may be actually not adding genes, but modifying the genes.

[Reporter notes that main tape deck stopped here due to a sticky tape.]

inclined to modification. So that is here in the guidelines, as well. And there are other things that are taken into account when this genetic material has been inserted. Have there been rearrangements in the DNA, and is there any consequence of those in terms of safety for health.

Post-translational modifications of the proteins that may have occurred as a result of changes in protein sequencing and so forth. And is this material inserted in a way that it will be heritable, stable, over several generations.

So I think that these are things that are not of any surprise to you.

So what is the goal of these guidelines? First of all, that the food should be safe; that it should not, when it's used as it's intended to use, it should be as safe as other foods on the market.

And that's important. We're not trying to show that this food is absolutely safe or guarantee that the food is safe. We're comparing it to foods that are already on the market. We're not accepting foods that would be of a lower standard with respect to safety, so they should be as safe as their conventional counterparts, taking into account any changes in dietary patterns that may occur, if there have been nutritional modifications. And the idea of these safety standards within the Codex system is then to allow risk managers to have information that if there are hazards identified or other considerations that have to be taken into account, that they have this information that is based on sound science that they can use then to make their decisions about allowing these products to enter the marketplace.

So this is--these are the new guidelines that have been adopted in the Codex system, and represent an international guideline and, basically, yardstick for countries to look at to see whether a food has been evaluated in a manner

that has been accepted in the international community. One of the things that these guidelines do note is that while this is all focused on recombinant DNA and was designed to answer questions about recombinant DNA, no one knows any other way to evaluate the safety of a whole food if we were asked to do so. So if someone came to FDA and said, well, we didn't use recombinant DNA, we used some other method, how should we go about evaluating the safety of this food.

We would point to similar guidelines. Our own or these guidelines, because we don't know any different way. Some of the questions might be different, just as some of the questions might be different for recombinant DNA product that might present different characteristics than have been laid out in this general framework. But while this focus is on our DNA, you should really think of this as way to evaluate the safety of a whole food using a scientific approach.

Thank you very much.

CHAIRMAN BUSTA: Thank you. Abigail?

Abigail Salyers. Say your name, and then we'll have it in the record.

DR. SALYERS: Abigail Salyers. Okay. So my question is I don't know if any industries are getting ready to do this, but, in theory at least, one could produce foods that are safer. Take the peanut, for example, if you could produce a peanut that didn't make peanut allergic people sick, and you used recombinant DNA to do that, would proving that that is a--I mean, how would you evaluate something like that?

DR. MARYANSKI: I would actually--

DR. SALYERS: Casava also has been brought up as a possible case--

DR. MARYANSKI: Right.

DR. SALYERS: Where they might use recombinant DNA to--

DR. MARYANSKI: No, and actually, of course, the Japanese did some early work using antiscents, trying to suppress the allergenic proteins in rice, and, you know, it was partially successful. The problem with allergenicity is that

there are a number of proteins within any given food that are allergens, and some are more allergenic to some individuals than others. Dr. Astwood is much more of an expert in this, and Dr. Jones, who works with me, is also much more knowledgeable about this. But I think the potential is there for--

DR. SALYERS: I'm just asking in terms of the Codex, do you think that the basic principles there would apply to cases--

DR. MARYANSKI: Well, I think that one of the things that Codex recognized is that these guidelines were set up to assess the safety of food. They--we--they can't really address that question in the sense that you would have to look at clinically, probably, whether the food still can cause clinical reactions. In other words, have you eliminated all of the allergenic potential, or at least sufficient allergenic potential that that food is going to be safe for consumers. That's going to be a difficult standard to meet, at least as far as we know technically now. No reason not

to try, of course.

CHAIRMAN BUSTA: Dr. Benedict?

DR. BENEDICT: Steve Benedict. The previous question actually raised another thing. Isn't it true, though--this isn't my question. Isn't it true, though, that if someone did attempt to eliminate an allergen, it would still have to go, the product would still have to go through the Codex for changes in safety and unintended effects and other sorts of things? Exclusive of what you talk about proving no longer allergenicity, you still have to take it through all the recombinant questions; would you not?

DR. MARYANSKI: Yes. I think if you're asking would--for the product to be acceptable in various countries, and it's, you know, modified using these techniques, I think that today at least most countries would expect that it would also

be--you know, these other issues would also be taken into account.

DR. BENEDICT: Yes. So actually my question is fairly trivial; and that is at the

meeting, did anyone discuss the range of sequence, border sequences up and downstream? How many KB that people thought should be sequenced in order to determine that you're really not near an opening reading frame you could influence?

DR. MARYANSKI: I think that certainly one of the questions that was raised was whether there could be any fusion proteins. In other words, is the insert done in a way that there's not

read-through into the plant DNA; and if there is, then can you identify that. But there certainly wasn't any boundaries given in terms of how far out one should look. And, quite honestly, some countries were interested in that more for methods of detection than actual, you know, safety considerations. So there were various reasons why people are interested in border sequences. So, but our focus here, of course, is on the safety assessments. So, but the issue of whether that insert leads to any fusion proteins or read-through into the plant genome that would express different proteins than what the plant normally expresses, I

think is the sort of fundamental question about the region surrounding the insert.

DR. BENEDICT: Yeah, one of the things I recall that we sort of discussed in 1994 was insertion into a gene, which you would pick up with just a short distance, inactivation of a host gene. But the other possibility is that if you're bringing in regulatory sequences, the possibility of some sort of downstream promotion, having not so much to do with fusion, but just is there an opening reading frame nearby that, for some esoteric and maybe unpredicted reason, you can, in fact, activate or inactivate an adjacent gene. And so that was why I asking how many KB downstream you should go to say, well, there's a low probability we're going to influence a host gene in its expression.

DR. MARYANSKI: Well, I think that's a very good question for you to come back with after Dr. Cebula gives his talk.

CHAIRMAN BUSTA: Other questions of clarification?

DR. ARIAS: Jonathan Arias. James, I'm--

CHAIRMAN BUSTA: Could you? Why don't you pull the mike in front of you?

DR. ARIAS: Is that better? Okay. I'm concerned about the issue of substantial equivalence in regard to characterizing the meat food as being no more unsafe perhaps than conventional derivatives. Does this include

long-term studies, or just acute, sort of toxicology biochemistry?

DR. MARYANSKI: No, we have taken

long-term effects into account at FDA, and in the Codex, because long-term effects are, of course, a question and a difficult one. But we know a lot about the foods that we eat. There are also a lot of things about foods we don't know. I don't know what consuming carrots over my life time does for me, other than my mother told me it was good for my eyesight. But whether it has any other effect on my health over consuming that, or any food does, we know very little about long-term effects of most of the foods that we consume beyond the nutritional

kinds of things that are, you know, we know about saturated fat and so forth. But we don't really know much about consuming foods to begin with.

What we do know about foods is that they do contain certain nutrients, certain toxicants, and so forth, and that we have accepted the foods that we have in spite of the toxicants that certain crops have; that at the levels those occur, we've accepted those as safe. What we believe about the foods that we've seen so far with bioengineering is that the modifications that we see are fairly small modifications of the genome. And most of these, of course, produce metabolic proteins in the case of the proteins that FDA looks at or the BT proteins or viral proteins that EPA has been seeing. And we know a lot about protein toxicity, and we know that the proteins that we're seeing do not lead to

long-term effects. So we know at least the introduced proteins do not lead to long-term effects. And we're not aware of any reason why these foods would produce long-term effects any differently than foods developed by other methods

of plant breeding.

DR. ARIAS: Thank you.

CHAIRMAN BUSTA: Are there other questions. Dr. Benedict?

DR. BENEDICT: Since there's open air--

CHAIRMAN BUSTA: It's Steven Benedict.

DR. BENEDICT: Steve Benedict, again. One of the things that, as I read through this, clearly, you're dealing heavily with allergenicity, and the thing that came to mind, and I'm an immunologist, and I'm not sure I have the definition correct, even though I am an immunologist; but, as we all know, there are billions of epitopes going through the system daily, and their toleringenized [ph]. And so we've been tolerant, immunologically speaking, to all of those. And the lack of tolerance may lead to an allergenic response. But lack of tolerance might not necessarily be allergenic, but it might, in fact, lead to something that was immunogenic, and I'm not sure that there--this is where I'm not sure that there's a distinction between the two, because

I haven't thought it through for more than five minutes. But the possibility seems to me that that checking for immunogenicity exclusive of allerogenicity [sic] might be something that someone should think about other than me. And I'm wondering whether that came up at the Codex discussions.

DR. MARYANSKI: Yeah. I would say yes that it did in the sense that there was certainly a recognition that there is more to allergenicity and immunogenicity than IGE responses. I think that what was felt to be important is that we do our best to address the issue of the IGE kind of response because we know that certain proteins are allergens through that mechanism, and that the idea of transferring genes from one source to another, and obviously the expression of those proteins, raises the question of whether there could be an IGE-type response.

I think that your question is one that's much more appropriate when we get to the issue of allergenicity as opposed to the molecular subject

today. But I just want to recognize that, yes, there is--we don't just simply put blinders on; that we only look at IGE. But these guidelines are focused on IGE because of the need to address the safety of the new proteins.

DR. SALYERS: Just a brief question. Abigail Salyers. The--we have people who are today who are living a lot longer than in the past, and we also have people who are alive today, like cystic fibrosis patients, who were not alive in the past, and so did the Codex consider the fact that we've got these new populations of people, especially I'm thinking especially about the geriatric population, where our previous experience may not be as good a guide as we tend to think it is as to effects that foods, even conventionally used foods, might have.

DR. MARYANSKI: Yeah. I'm becoming much more interested in the geriatric population.

[Laughter.]

DR. SALYERS: So am I.

DR. MARYANSKI: Actually, the Codex didn't

address the geriatric population specifically, but it did discuss the fact that there are basically subgroups within the population that one needs to take into account, including the elderly, pregnant women, infants and so forth. And so, yes, while not addressing that in detail, the guidelines do generally recognize that in thinking about the safety of a new product, you do have to think about subgroups in the population as well as the population as a whole.

CHAIRMAN BUSTA: Thank you very much.

DR. MARYANSKI: Thank you.

CHAIRMAN BUSTA: We'll continue on. Dr. Thomas Cebula on FDA's approach.

DR. CEBULA: Thank you. I'm Tom Cebula. I'm the director of the Office of Applied Research and Safety Assessment. I'm also the lead scientist for molecular biology for the Center, and those that know me in the room know that I never give a scripted talk. I'm taking out a crib sheet because I was taking notes.

The first thing I'd really like to say

about the process that FDA uses is that it's an evolutionary one. And, since joining the Center, I've interacted with Jim Maryanski, Bob Lake, others in this room since about 1987, so, like Jim, I'm also worried about the geriatric population.

In that process, we saw a number of people come to us with a lot of ideas about what they were going to do, and I think Dennis made a very valuable point when, early in a consult, some advice might be given, the company goes back, and we don't hear anything more about it. So there is a lot of oral history perhaps that should have been captured, and I--the point that you make earlier about how do we see those numbers. Early in the process, it was an evolutionary one. I think Jeanette made a very good point about after a series of discussions, clearly, we're trying to make that a very much more transparent process to really capture how many times the sort of advice that's given from this side out has influenced.

The other point I'd like to make before going into my talk: I think Abigail Salyers made a

very credible point about connecting dots in the scientific community. All science is interconnected and interwoven. Jim Maryanski talked about R-DNA and the subject for transgenic plants. Clearly, we've taken cues from R-DNA in microorganisms. I serve as one of the experts for WHO for the microorganisms in R-DNA. But Abigail's point about probiotics. I'd just like to extend probiotics, counter terrorism, and our budgetary process called OMB. I do believe that what we have heard is that there is a real need to really take some of the counter terrorism dollars and move them into other areas of food safety.

I believe something that FDA is doing, and our laboratories within CFSAN are actively engaged in. My office, for example, has the mission to ensure, to do research that ensures food

safety--microbiological food safety--and we also have a team of toxicologists.

Post 9-11, that mission changed somewhat, and it became food security and food safety. It's a very valuable point, because the research that

we're doing for food safety could be immediately applied to food security. And, of course, those dollars that are coming in, we're using as a twofer. So if we're developing methods for the intentional contamination, those serve for methodologies for accidental. So, again, we're trying to be very proactive, recognizing that it's a zero-sum game. There's only a certain number of budgetary dollars out there; and if it's applied to counter terrorism, we have to get the most bang for the buck.

The point about Vancomycin is a very important one. And I'm just going to say that probiotics in the genomic era, as people have looked to sequence total genomes, people have looked at probiotic organisms; FDA, in contributing to the genomics project, has often emphasized it is nice to know that a bacillus cereus that is being used a probiotic, we can know the entire genome. But isn't it more important to know about a bacillus cereus that causes disease. So, again, FDA is trying to move in the direction of saying if

we're going to use those dollars, let's get some information about food safety from that.

I just want to make sure that I've covered some of those points.

The final thing I would say is that since 1987, we have served in a way as internal consultants for some of the process that you've been hearing. We are primarily the laboratory portion of the Center, and since our hands are the wettest on the techniques, we're often called in to say how would you evaluate this. So there's a very responsible team, and we interact with that responsible team to offer some advice and also do the consult--also do a review, as well. So there is, if you might say, there is some partial duplication, but it's a very important duplication, I believe.

With that said, I'm really going to take some of the slides that Jeanette gave and Jim gave and kind of put our wrinkle on it.

Jeanette pointed out that really we're asking for companies to come in to provide the

documentation that assures us that there's a reasonable certainty of no harm. But I think the important caveat is this is the real opportunity to find out what's in the pipeline and what's coming down in the future so that we can be proactive rather than reactive.

We neither prescribe or proscribe specific tests up front. We're there to hear what industry has to offer, and somebody said, well, how much input do we have. Clearly, the input is we are going to offer advice, but we're going to be focused on the characteristics of the food product, as you heard Jim say, as you heard Jeanette say.

We are there for scientific evaluation. Our intent is really obviously the food safety question. But, as scientists in this room, as you heard around the room, often we get off on a tangent. We talk about the strengths, and that should say limits of a particular technique that might be used, because this helps us in future deliberations. But our primary focus is food safety, and that's the important thing to remember

for our evaluation here.

Again, borrowed from Jeanette, we do take a multidisciplinary approach, but I want to point out that the duplication is very important, because we're going to be really focused on the molecular analysis and really talking about the genetics of a particular manipulation or a particular plant.

We often get clues from the chemical and nutritional analysis about stability and genetic stability, so we're looking at pretty much the same thing that a chemist might be looking at, a nutritionist, but perhaps from a different slant.

Clearly, as you've heard, we want to know about the identity and source of introduced genetic materials because that will tell us a lot about what are the potential food safety concerns. For example, in a sidebar during the break, we were talking about transferring genetic material from one microorganism to another. I point out that the nomenclature has changed over the years, and something that is streptococcus a few years ago and may have been considered commensal is now

enterococcus and is a pathogen. So we definitely ask people to identify the organism, but to trace the nomenclature of that organism to ensure that we have taken into consideration that nomenclature changes haven't affected our assessment.

And, again, the intended changes of the composition of the food--clearly, we're always concerned about the intended effect, but, as you heard from all of the discussions, clearly, in evaluating food safety, it is often the unintended effects--allergenicity, toxins, and I should add the toxicants that Jeanette mentioned--because in a manipulation, as you heard, if you happen to insert and get downstream expression, you might get the expression of a toxicant. We want to know that.

The allergenicity, as you've heard, will be a separate discussion, but I would point out in our evaluation, since we are seeing sequenced data, the bioinformatics tools are there to say what sites could be potentially glycosylated; and these days, with the glycosylation kits, we're often seeing the data for whether those sites are,

indeed, glycosylated.

The nutrient levels. Again, you've heard why we look at that, but I would just like to take a digression, and say when we talk about natural foods, and we're looking at a nutritional status of anything be it a nutrient and this is also true for a toxicant, you would expect a normal distribution from the natural varieties over time. When you're looking at a R-DNA plant, since the manipulations are among one variety, we would expect that distribution to be much tighter, and the sort of data that we see says that. And so, if we're looking for toxicant values, which might also be this value, where this is low to high, clearly the data that we have been seeing up until now is that can fall on the low side of the toxicant. So, when we're considering an R-DNA plant versus a natural variety, we find that all of the data, thus far, seems to follow the strength; that you're getting nutrients that you expect, with a narrow distribution, and lower toxicant values. And I would say that's a fairly general way of stating

the sort of data that we have been looking at.

And finally, something I neglected to say that our laboratories are very interested in DNA repair and effects on the emergence of antibiotic resistance, and the penetrants of virulent straits. We are very interested in evaluating antibiotic resistance. And, again, from an evolutionary perspective of the process, as you are all aware, the primary marker that was used early was antibiotic resistance, and we now see a drift away from antibiotic resistance.

Now stealing some of Jim Maryanski's slides and the Codex slides, I would really like to emphasize some points, but I'm not going to reiterate all the points. When we're describing the genetic modification, the sort of data that we have been seeing is people are showing you the nucleotide sequence of the material that they're starting with.

They're telling you about the promoter set they're using. They're telling you about the terminator set they're using. That's a very

important piece of data, because that's direct information, and now you can use indirect methods; and there's a host of indirect methods that could be used to evaluate the construct once it's inside the plant. And that's a very important point as we consider do we need sequence data and how much sequence data do we need.

What I'm establishing here--if you know the sequence, you then can do a restriction map and ask does--is the restriction map consistent with the sequence that was delivered to the plant. You can look at restriction maps. You can look at expression. Northerns. We actually--or expressions of the protein, Westerns. Now, clearly, there's a distinguishing factor here. If you look at expression, Western analysis of the protein, you're usually looking at the intended effect. The Northerns, however, give you a lot of information, because, if there are other transcripts made, that's something that we would be looking for as an unintended effect. So unless it is stated that the intended effect is, any Northern

expression that would be aberrant, we would be questioning what is the significance of this RNA, this extra RNA.

Again, I've already made the point about marker genes and the nutritional modification, so I'm just going to skip this slide and talk about the transformation process that Jim alluded to.

Again, when you are talking about a ballistic method of random insertion, we might have a number of things that we're looking at. You know, how many sites does this randomly insert. If we're talking about site-specific recombination, if it is a site-specific recombination, it's a more tailored type of mechanism. So, again, in making the point that it is a case-by-case analysis, it's important that if you prescribe the tests too soon, scientifically some of those tests would not make sense or some of the questions would not make sense. So we're very interested in the DNA to be introduced--the genes, the markers, the regulatory sequences. But Steve Benedict mentioned a point about regulatory sequences and sequencing, and I

really would like to make the following points right here.

We are consistent, I think, with Codex, we're talking about size, identity, and location. If we're saying we know that there is material inserted into a genome, say, the tomato genome. We are also consistent in saying we would like to know orientation of that insert. However, the location down to the nucleotide is something that we personally don't ask for. And the reason is, and I should state it up front: all of the effects that we're looking at are SIS proximal because you can sequence 10 KB, 20 KB versus a trans effect from another chromosome, you're going to miss it anyway. So I would really view them as a--up until now, we have said that we are looking at SIS proximal effects and the integrity of the genome in the intended and unintended effects the researcher has inserted into a gene.

As far as the data that we've been seeing, basically, there have been times over the years, I can remember where people have provided inside-out

PCRs as to tell us some conference. Okay so I will leave it at that to say that the location to us is that it is inserted and that it is stably inserted.

The number of sites and organization. Clearly, we all know that when the material goes in, that you have duplication. Okay, so at the same site, you can have many inserts lined up or you can have many inserts lined up at different sites in the chromosome. One is more important in our evaluation than another. If you have multiple copies at different locations, the chance for recombination, reassortment, rearrangement, and ultimately effects on genetic stability are much more of a concern than if you have a single insert. So that's why talk about wanting to know how many inserts and how are they located. And, of course, as Jim emphasized, we are definitely interested in any open reading frame and insertions the possibility of fusion proteins.

The important thing here I wanted to emphasize un-translated RNA. Jim mentioned one possibility. If the intent is to make antiscents

RNA, it's not going be translated. We want to know about that. But there's another--it seems every journal you pick up at least there are two articles in each of the journals talking about un-translated small RNAs. That is not our intent. We're not asking for people to tell us about every small RNA that's produced in the plant, because I don't know what those data would tell us. But I leave that to you. That is a very burgeoning field right now, the un-translated small RNAs that act as regulatory molecules.

And, again, the intended effect and heritable stability. I've mentioned a couple of ways we measure genetic stability or we assess genetic stability. One of them is the Mendelian inheritance. Clearly, if the insert is unstable and is hopping, the mosaicism rather than the Mendelian inheritance pattern would be seen. So that's a very important thing. But some of the other data that we've been talking about, the expression data et cetera, are also measures of stability also.

And, again, these should read strengths and limits rather than limitations, because from the perspective of an R-DNA, it's important to realize that one or only a few genes are being incorporated into the plant. There's a limited number of manipulations. One can talk about unintended effects, and we'll get to that in a moment.

But it's also important to remember that, as genes have been introduced by R-DNA techniques, those genes will become the docking points for traditional breeding, and those sites, then, will be now the conventional traditional breeding that we're all used to. So, again, as people talk about propagation of R-DNA, the R-DNA will be propagated by traditional breeding also.

So the questions that I took from this were what tests are necessary and are there tests that should be required. I think those are all the questions in your minds, and you will be discussing this afternoon.

I'd really like to point out that Mendel

did a nice experiment many, many years ago, talking about round and wrinkled peas. And that experiment was reproduced in the molecular era about a decade ago or so. And basically, if you do traditional breeding, and you get round or wrinkled peas, and then subject it to some proteomics, you find that 62 of the 636 protein spots are qualitatively different. The R-locus, of course, now we know affects sugar content, lipid content, storage protein composition. But the round pea is safe to eat, and so is the wrinkled pea. Okay.

The way I would like to look at this is as we were musing about this in Paris at a meeting, I went off to the Monet exhibit at the Musee Marmite, and there's some wonderful water lilies, canvases that just dominate. For those that have seen, this is definitely Monet. Everybody knows the blues, the greens, the purples. This one happens to be the blues and the greens and the whites, but this is also Monet. It looks more like a Jackson Pollard. And this is about 1917. The other one was 1880. Okay. But they're Monet, and I submit

to you that they're safe to consume.

[Laughter.]

It's also important that Gottlieb and de Vienne in '88 did an experiment also that found that if you take the AF mutants, you'll see that tendrils can replace the leaflets. So it's something that you wouldn't want to see in a plant. Weeding this is not a good characteristic, but, basically, the protein profiles were absolutely identical. So, again, the tests would not reveal a difference, whether it was a food safety issue or not. I'd leave that for discussion.

I said I'd return to unintended effects, because, again, I think we have to recognize that, as the techniques get better and better, pleiotropy will be a universal trait of any mutation. We know how much cross-talk is going on in the cell, and within an organ, and within an organism so that there will be always cross-talk anytime a mutation is introduced; and now I'm talking about a point mutation, be it an insert of a couple of genes or a point mutation, a base substitution.

So, again, not to argue that pattern recognition can't work, okay. We are believers and users of technologies like genomics, like DNA arrays, like proteomics. But, again, I'll borrow from the impressionists and say that pattern recognition works Seurat proved that points of paint on a canvas can be recognized. It can look like a lighthouse, or it can look like an afternoon on the island of the Grand Jatte.

However, in a regulatory setting, I want you to concentrate on the dog, because that dog is very docile, at least captured at this moment. So when we are doing expression profiles, if that dog changes to green, red, polka dot but remains docile, it is not a food safety question. If that dog happens to bit that individual in front, it becomes a food safety question.

I submit to you that we're at a stage to recognize that expression arrays do, indeed, work, but it's not at a point where we can find and isolate the genes that define safety. And so, as people are talking, I would suggest that we talk

about the new technologies as they relate to food safety. And with that, I'll stop.

CHAIRMAN BUSTA: I just want to make a comment. You didn't need your entire 55 minutes.

DR. CEBULA: That was intentional. I really wanted this to be a conversation, and, as I had pointed out, I would--

CHAIRMAN BUSTA: Dr. Salyers.

DR. SALYERS: I have a question.

CHAIRMAN BUSTA: Please identify yourself.

DR. SALYERS: Oh, Abigail--Abigail Salyers. I have a question about the extent to which or how you see your group's position vis a vis advising industry about what it should do. I'm struck, as I talk to various, especially the small biotech companies, but even the bigger--people at the bigger companies like Monsanto is the extent to which they have a very narrow range of expertise. And even in the big companies, where they might have a broader range of expertise, the components don't necessarily communicate freely with each other.

You're a world renowned scientist with a lot of contacts, and I'm sure that you've--so you're the sort of person I wish we had more of in industry, but we don't. And I'm sure you've assembled a group that is--has outstanding representation of the various sciences; and that you are probably more like to the many people in industry to become aware of changes in the basic science technology. So my question is one would hope that the FDA somehow could serve in an advisory capacity to industry; and yet, of course, you see the other side of it, which is the conflict of interest side I guess. But how do you see your role as an advisor of people in industry?

DR. CEBULA: I believe the--when I said that we were serving a consulting role, I meant within the context of FDA.

DR. SALYERS: No, I understand that.

DR. CEBULA: Oh, okay.

DR. SALYERS: So I'm just asking you to expand on the outreach part of it.

DR. CEBULA: In the outreach, I believe,

as I said since '87 or thereabouts, everybody time somebody has decided to talk to Jim or Bob Lake, Jim Maryanski has graciously said, well, why don't we get Tom involved. He may have some ideas. When we're out at meetings, we are scientists; and, clearly, if there's a technique out there, you know, we're doing that right with pyrosequencing. Pyrosequencing is a very valuable technique. Some people haven't heard about it. We're using it. We're saying how about this. So it is scientist to scientist. We're trying to exchange information and the technologies as we see it, and we're welcoming the input the opposite way. We're getting fantastic input from academia and from industry; and we're working collegially I think--is the important thing. And if I might say something about the whole process, again, over time, let's start when we first consulted, if we said, do you have data. The first response was why. I believe now it is, you know, developed to a point where people are saying, well, these are the sort of data, and this is--there's scientific reasons for

the data. So, I believe there is a trust in its finest sense of the word, a mutual trust, that questions aren't being asked for it would be nice to know, but as was stressed in the two talks here, there is a need to know because it may impact a safety decision.

And so I think there's been a good rapport built up over time.

DR. FEDOROFF: Nina Fedoroff. I was very struck by your cogently pointing out that you can have many differences, and the foods are still safe. I think if people had suggested that we would knock out starch genes and get sweet corn, somebody would haven't objected to the religious grounds or ideologically grounds.

In any event, the other point that you made that I'm struck by, and that leads to my question, is that, so far, your assessments have indicated that the distributions are narrower and that none of these unintended effects have surfaced. My question is: how many instances do you have to assess before you can evolve your

guidelines to say, okay, this is not a major problem, we don't have to collect all of these data.

DR. CEBULA: Well, I think one thing you've heard from the two speakers before me, we are clearly waiting to hear what the National Academy has to say. We're awaiting what the input of your group will say. But, clearly, I wanted to point out there will be unintended effects. The question is, will those unintended effects affect the quality or food safety of--

DR. FEDOROFF: Sure.

DR. CEBULA: Okay.

DR. FEDOROFF: But the other thing is that there are--in--in-those of us who work on little experimental plants have been making insertions left, right, sideways, hundreds of thousands of them. And the information that comes back is that plenty few of them make a huge difference. In fact, people struggle to identify insertions that change how the plant grows, flowers, reproduces. It's a very small fraction of them. And the

ability of insertion--even when people set out to activate genes by pointing a promoter out, that doesn't even work much of the time. It does work occasionally, but not--certainly not as frequently as you'd expect. And it seems to me that there's a large volume of information, not just new techniques, but accumulating information from actually insertions done for mutagenic purposes that one could bring to bear to increase the amount of information that you have available to you in making judgments about how frequently there are unintended consequences.

DR. CEBULA: Those data, indeed, would be very helpful. I think when we were talking about CalGene, I think that you were on the committee then, and I was giving a talk. And I--you said the problem is that the negative experiments are seldom, or can't be published, and we're saying that we need that database, whether that database is in the published literature or available, we would like to see the number of cases so that that can really help us.

DR. FEDOROFF: That's a real challenge because people frequently don't--you know, they put a lot of effort into identifying mutants. They don't put a lot of effort into quantifying the frequency of those mutants.

CHAIRMAN BUSTA: Doug.

DR. GURIAN-SHERMAN: Doug Gurian-Sherman. I think that whole issue brings up a discussion that we should, you know, probably take up at some point. But, with time limited, I wanted to focus on a whole other issue.

You mentioned several instances of types of data that you look for, so you--you know, you talked about stability, and, you know, differences in nutrient levels, and all that kind of stuff. One concern that I have that maybe you could address is, what the level or type of analysis that you consider to be acceptable, because, again, and I have to put this in context of the studies that I looked at that you folks looked at, and often found there was limited methodology described, limited statistical analysis. You know, stability is one

thing brought up. Half the cases we looked at--a lot of them did look at Mendelian analysis. Probably half of them didn't do any statistical, like the chi square analysis to tell you--I mean, you're not going to get a perfect four to one ratio in most cases, so you have to have some kind of analysis that tells you how much of a deviation you get, and then try to figure out what that means.

So, I mean, I think a whole other issue the quality of the data, and the expected level of analysis, and how--I mean, you know, one standard could be comparable to what goes into a peer review journal. To be frank, I don't think that, on average, that was the case. So how do you decide what the quality is that's acceptable?

DR. CEBULA: I think it's very important to recognize that I said several times that one way of measuring genetic stability is Mendelian inheritance. As I said, if it weren't Mendelian inheritance, there would be a large deviation. However, if those were the sole data that we were looking at, I would say yes, you would have to do a

very sophisticated statistical treatment to unequivocally say it's Mendelian inheritance. But there's a body of other evidence that says, the material is still where it is because on the second generation, we've done, you know, a map that has provided us other data that supports it's sitting in the same site that it started out with. It's also being inherited at almost, or, I mean, I'm taking your word. You know, I don't know which data you are speaking to. But if it's almost Mendelian, and it's sitting at the same place, it's likely to be Mendelian rather than some runaway jumping gene. And that's so--you use the different techniques. I would rather see the company offer three different techniques to verify it rather than working on the statistics to come up unequivocally to come up with a Mendelian inheritance.

DR. GURIAN-SHERMAN: I think that's a good point, but, again, you know, addressing that issue: you know, when we looked at this, some companies did do things like expression analysis for a bunch of plants that, you know, the F-3 generation or F-4

generation, but many of them didn't. And similarly, you know, some companies did bioassays, but at levels of expression that might be, you know, 20 times over the lethal dose, so variations that could be fairly substantial wouldn't even necessarily show up.

So, I think, you know, what it comes down to often is a lot of times the devil is in the details, as we'd all expect, you know, in empirical science. And I'll just raise it and leave it as a concern that there need to be sufficient analytic standards, redundancy if you want in tests, but then they need to be there. And I didn't find that that kind of redundancy was there in many cases.

CHAIRMAN BUSTA: Jim.

DR. ASTWOOD: This is Jim Astwood. Dr. Cebula, thank you for the thoughtful presentation and interesting art. In thinking about it, I was wondering if you could comment on, what seems to be in this presentation, an analysis at the genotype level, you know, fundamentally, if you think about it as a geneticist. What importance should be

placed on the phenotype, the agronomic evaluations, the compositional analyses, which can be thought of as biochemical sampling--these other factors. How do you, as a molecular biologist, how do you sort of think about the weight of those different aspects?

DR. CEBULA: Well, the agronomic issue I believe is away from my bailiwick. I'm expecting a company to make a decision if there's, I hate to say it, but there's a different bottom line. If it's going to be developed and marketed, there has to be one bottom line.

The molecular analysis, by the time somebody has made the decision that they're going to bring something forward, I believe they've already done some of the routine analysis that impinges upon molecular biology; and that is the genetic stability, the ease of manipulation, and when I say stability, I mean it on two different levels: the internal rearrangements that could occur, the possibility of methylation and silencing--all of those have been done by the time

they came in. So, in that sense, by the agronomic consideration that a company might do, I feel that a lot of experiments have been done that we haven't seen the data for. But, again, the data that we're analyzing supports the notion that you have a stable insert; and it supports the notion that it is constantly being expressed.

I presume, again as Nina Fedoroff said, those data are to come by, but within the notebooks, they are probably a lot of failures to come forth simply because a methylation pattern decided to silence a gene or something. But those are the data that we don't see, so we can't give statistics on those. But, by the time we're seeing the data--

DR. KAPUSCINSKI: Thank you, Dr. Cebula.

CHAIRMAN BUSTA: Identify yourself.

DR. KAPUSCINSKI: Anne Kapuscinski. I also appreciate your presentation, and, by the way, I'm a lover of that art, and I was born in Paris, so it kind of got under my skin a little bit.

But I have a question. It seems to me

that the question about these genomic and proteomic methods for the committee should really be, and I assume, therefore, for the FDA, should really be how could using those help to improve sort of strategic searching for the few cases where something unanticipated might cause a real food safety problem, rather than asking about those techniques alone.

So, I would like to get your feedback on that. What I'm really getting at is I think I want to ask sort of what Dr. Astwood asked but a little bit differently. Isn't it better for us to think about how can we direct the use of these techniques in combination with information at the level of biochemistry, and then at the level of the whole plant, phenotype, et cetera, rather than ask to only look at those techniques alone. I think if we just look at those techniques alone, we're kind of barking up the wrong tree. And I fully appreciate, you know, your concern that the status of the technology right now is you could generate tons of data about transcripts and expressed proteins and

not really know what it means. But I think it seems to me what we need to do is figure out how can we, knowing that there's that limit, how can we direct the Agency, and maybe even direct recommendations for research, to improve the use of that technology in a very strategic, focused way so that it could add to the toolbox that you have right now. Because it seems like the bottom line challenge in this area is that, yes, the large majority of the time, there isn't going to be a problem. We just want to make sure that that doesn't make us complacent so that it becomes hard to proactively detect the few cases where there will be problem. And we all know that coming down in the future, this technology is becoming more and more powerful, and there's going to be chances to make much larger scale changes as we get into gene stacking and genes that we know will have major effects on composition because you actually want them to have major effects. So, that definitely raises the question that they might also have some major unanticipated effects. So, we have to be

looking at the future. So, what's your reaction to what I'm saying?

DR. CEBULA: I think these technologies are wonderful technologies to do just what you said, search and discovery. If I could read into your question and say, what could you do to help, the first message that I would do--ask you to send is to ensure that any of the technologies being used have been validated, because there are a number of these expression chips out there, and you read the same literature that I do. Recently, one of the major manufacturers of a particular chip, there was an article in Nature, called "When the Chips are Down." The company in their quality control missed the fact that they put the wrong strand, and, even then, 30 percent of their sequences were aberrant sequences. And this was sent out, and people are using them, and flooding the literature with array papers.

DR. KAPUSCINSKI: Trash.

DR. CEBULA: The rush to public a new technique has often showed us that there are lot of

holes in the technique. So, I would say that what you could do is make sure that we don't damn a technology because we got started with the technology prematurely.

So, for expression, there are some very good examples. David Bottstein teamed up with Stanford to come up with a lymphoma chip, very predictive. But to get to using a lymphoma chip, they had to go back and show that the repository for the sequences were cluttered with about 24 percent of the sequences that had the wrong sequence.

So, everybody's using these chips, and it takes a couple of scientists to say, well, let's not rush to publish. We'll just go ahead and find out what's going on first.

So, my first message would be could you carry back and say, let's go slowly but very together on making sure that the chips we're going to use to evaluate biotechnology, natural foods, or any other is the best chip to make sure that it's right. We don't repeat mistakes. I mean, we saw

the same thing with the emerging PCR technology. Everything gets exponential the first few years, and that's where the mistakes are buried. I'm sorry.

CHAIRMAN BUSTA: Dr. Fedoroff.

DR. FEDOROFF: But I think that's missing the point. In order to make a lymphoma, and you've made the point earlier, in order to make a lymphomas chip, it's not only important to have accurate sequences, but it's important to have the lymphomas.

DR. CEBULA: That's right.

DR. FEDOROFF: Okay. And that's really the whole point--is that we're--we need to avoid focusing on the technology when we don't know what we're measuring. There is a huge range of permissible compositions that are perfectly good foods, okay? We seem to slip off that over and over and over again. And I think it's extremely important because some of the notorious incidents that have hit the press worldwide have been based on the lack of recognition that the very processes

of transformation, the culturing of cells, which has been used to generate variation, does generate variation. And, so, people attribute variation to somehow this terrible recombinant DNA when it has to do with tissue culture.

But it's not an unexplored territory. And so the question is one of connecting the limits of variation with permissible--and I think that we're going about it the wrong way. Instead of saying what is--characterizing the entire range of variation for every component of foods, we need to collect the information on what has proven to be problematic health wise--glycoalkaloids in potatoes, sugar binding proteins, which are lectins, which are under characterized as yet in both plants and animals.

But it's not an infinite variety. One doesn't have to look at everything.

DR. KAPUSCINSKI: That was why I was asking about, to kind to get your input, on how your thinking about potentially directing the use of these new technologies for strategic searching

for the things that might go wrong, sort of playing off of what Nina is saying. So, I just wanted to get a sense of whether you, as a molecular biologist, are thinking in those terms and whether you realize that or to what extent you're thinking on drawing on information from the biochemistry and other levels of the phenotype to guide, then, how you would use, for example, a DNA chip.

CHAIRMAN BUSTA: That the first comment there was Dr. Fedoroff. The second one was Dr. Kapuscinski.

DR. KAPUSCINSKI: Sorry.

CHAIRMAN BUSTA: We'll let you respond, and then we've got two more over here.

DR. CEBULA: I apologize. Nina, thank you for bringing me back. You're absolutely right. That is the major point.

The one thing that I think would be very helpful, and this is something that we've discussed internally, is that narrow distribution versus the broad distribution is, in a way, comparing apples and oranges, because that is the gamut of varieties

versus one variety.

DR. FEDOROFF: But only for one component.

DR. CEBULA: No, I know. And no, no. And what I meant to imply is if we are going to use the new technologies, it might be important to really do counterpart, the natural counterpart to, the variety from whence it came compared to the transgenic and then ask, because I've looked at some data where clearly there are differences in expression, but I don't know what it means; because I don't know what the variety used, how that compared to the manipulated variety.

So I think first of all we have to do some fundamental genetics. If you're going to approach it, approach the new technology, take the new technology and combine it with some good genetics. That's all I'm saying. And, there, I would say that we would like to see, you know, again,

data--if you're doing it strategically, as you're suggestion, I would say that any gene that you see from an array would be verified by another technique, a more sensitive technique.

To say, if you have a knockout, do you get the expression that you expect or the

non-expression. It's what's going on in the field right now in array technology everywhere else. So, if it's going to be applied, and I'm not saying I'm recommending it, I'm saying if it's going to be applied, it should be applied in a very structured way to follow the sound scientific principles.

CHAIRMAN BUSTA: First Jonathan, then Calvin.

DR. ARIAS: Jonathan Arias. Thank you for the very interesting and informative presentation. I had two points, or questions, actually, to address. One is in regards to the overall view of risk assessment and safety.

And it seems to me that one of the difficulties intrinsic in these more computational biological approaches of genomics and such is ultimately how to interpret all of this data. Genes will go up. Genes will go down. And are those changes in ways--or confer changes in the plant that would be significant in terms of health.

And I think that's a real challenge, that maybe the answer isn't quite known at present.

The second point is related to more global issues, and that is, the input of safety analysis. You mentioned that, and I commend, that the FDA has a consultative process with developers, and, as part of that, wants to get heads up about new products in the pipeline. Likewise, does the FDA also communicate what they would perceive as recommendations to minimize risk intrinsic in the state-of-the-art technologies, for example, the use of strong viral promoters that can activate genes at distances might be, for instance, less preferred than using more tissue of targeted promoter systems. Obviously, if a product is expressed in non-food components, it may be less of a concern than if it's generically expressed in the plant, including in the food component. So, is that dialogue also two ways?

DR. CEBULA: I'll answer the second first. I believe it is. I think over the years, there have been a number of times when, in a very

interactive discussion, FDA scientists have made some suggestions, and that has led to a shift in the thinking of what was going to be used. There have been cases where we suggested an experiment, and I keep on saying it was an evolutionary process; so these are earlier times, where the very technique, there was a concern that there may be read-through through a border sequence, and we suggested some experiments that ultimately turned out to be a very nice publication for the company. I noticed that in the acknowledgment that FDA scientists weren't acknowledged.

[Laughter.]

You know those things happen, but it was a good experiment. It was a nice experiment. So, I mean, I think it happens, but again it's part of that consultative nature: that somebody comes forward and shows you the data. I think I said in the sidebar, I almost treated it as an academic, you know, show me what you've done in the lab lately. And over the first few consultations, it was why do you want to know this. And, over time,

people said, I did this experiment but. And, you know, that's how you find how people are thinking. You know, we really--it was more collegial, and people were opening up to say we were trying this. It didn't work. We were trying this. It did work. That's the process that I'm trying to define, and I'm probably not doing a good job of it. But it is a learning process on both parts.

DR. ARIAS: I think we recognize that in very few cases are we going to find clearly demonstrable risks associated with a particular transgenic plant. I mean, those would be probably be quite evident in the range of studies that are done to date. And it's the more subtle ones that we would obviously be--considered as problematic. Thus, using new to change the overall level of risk or probability of risk would seem to be also intrinsic in this process. And I gather that's what I'm hearing, as well.

DR. CEBULA: I would just like to return to your first question of combinatorial. I actually should have put a slide in here, because I

usually keep the yeast genetics--the yeast array up there. Brown, out at Stanford, now has looked at 20 different strains of yeast, using a chip that has about 2,000 pieces of DNA down there, so 2,000 reporting groups for each experiment. And he's done it under seven different environmental conditions. So, you have 2,000 to the 20 times 7, 2,000 to the 140th power, bits of information coming in. And so that's why I'm saying, let's be careful what we ask for because that information cannot be sorted through. And we will always see genes go up and genes go down, but the question, again, remains is it a food safety concern.

CHAIRMAN BUSTA: Dr. Qualset. You've been very patient. I appreciate that.

DR. QUALSET: This is Cal Qualset. I've sort of forgotten what I was going to talk about--

[Laughter.]

Because I'm intrigued with this discussion. I'm a old-fashioned plant breeder. And I see us drifting quite a ways from the 1992 policy; that is, product versus process issue.

What I understand is that FDA will receive a product--is that not on?

DR. FEDOROFF: I think it's not close enough to you.

DR. QUALSET: Okay. Anyway, I think that we are drifting away from the discussion of approval of a product versus discussion of all the academic issues of process. So, I think we need to, as a committee or a group, needs to slow down on this and decide what is important to know about a product. And it's a--the developer is already obligated to product a product that is stable; that demonstrates the functionality of a gene or segments being introduced; and to show how they established the stability, the genetic stability as well as the performance stability, over environments in which a plant will be grown.

So I think it will be nice to know about these various kinds of molecular things that can happen with genes jumping around and all that. But the issue is, is the product produced and is it stable. We cannot understand every genetic thing

that could happen. But if the product is as advertised, and it does not have any risks with human food safety as a whole food or as a processed food, I think that we are thinking steps further than are necessary for the evaluation process. So that's just my comment. I'm throwing that out as a point for detailed discussion I think by the people who understand both product marketing, development; and they understand whether they're on the right track.

So, I like the idea that there's consultative process with FDA, so they can kind of get a clue of where the product--what the targets are for a new product. But the developers really have to provide the information to support the claims. So, I hope we can discuss that further.

CHAIRMAN BUSTA: Bob. Identify yourself.

DR. BUCHANAN: Bob Buchanan. I was intrigued by your discussion of Mendel and the proteomic analysis. Of the 62 proteins that differ, did any, would any raise eyebrows as to whether the product might be a problem? We know

the peas are safe. Do any of the protein changes look suspicious?

DR. CEBULA: Well, I gave you some classes. Some affected sugar content. Some affected sugar storage. But none of those changes said, you know, there's an unknown toxicant, cryptic toxicant peas that came up. That answer is--yeah.

DR. BUCHANAN: That's good to know.

DR. CEBULA: Right.

DR. GURIAN-SHERMAN: Just--did they actually look for--

CHAIRMAN BUSTA: Ah, yeah—

DR. GURIAN-SHERMAN: Doug Gurian Sherman. I'm sorry. Just for clarification in that study, did they look for those kinds of changes? I mean, we assume they're safe, and they're probably safe, you know. But in terms of specific new anti-nutrients, you know, toxicants, were those looked at in that study or they looked at just general classes, like protein, starches, et cetera?

DR. CEBULA: What they did was they

resolved over 600 proteins on a two-D gel, found approximately 10 percent changed, and then went about to characterize all 10 percent qualitative changes, and showed that, you know, it was sugar storage. It was lipid storage. So, they would have--if there were a toxicant, they would have had an unknown protein, but it didn't show up in their analysis. Sixty-two of them were fully characterized. So.

DR. GURIAN-SHERMAN: Well that's--just again to follow up. Doug Gurian-Sherman. I mean, most--many products, especially legumes do have toxicants and anti-nutrients, and they're often removed through the process, and through cooking, they're not--for instance, they're not heat stable. So, you know, if there's a change in levels, some of them could theoretically have been, you know, toxicants and anti-nutrients. I guess what I was getting at was were they trying to identify whether those proteins were, you know, trypsin-inhibitors, or amylase-inhibitors, or, you know, whatever things that--maybe it wouldn't be problematic

anyway, because they may be removed through processing; but others that if the level was increased--I mean, some allergens are also, you know, some of these types of proteins that might have an effect, and that's what I was getting at.

DR. CEBULA: I really don't know if the wrinkled peas in that experiment were consumed by the investigators. I am sure they went a lot of blood, sweat, and tears to get it published in Genetics, but that was the standard.

DR. FEDOROFF: What does conventional alleles work? That wasn't recombinant DNA anything.

DR. CEBULA: It wasn't. No, no.

DR. FEDOROFF: It was just round peas and wrinkled peas.

DR. CEBULA: It was a natural. Right.

CHAIRMAN BUSTA: You know the conversation is rolling toward lunch. Any other questions for clarification? The speakers will be around this afternoon, correct, as we go to discussion for additional expansion if we need it?

DR. CEBULA: I will be on a conference call, but I said I would be available.

CHAIRMAN BUSTA: Okay.

DR. CEBULA: I will be on a conference--

CHAIRMAN BUSTA: But, if we have to, we can drag you in; right?

DR. CEBULA: Okay.

CHAIRMAN BUSTA: Any--immediately after lunch, we will have the public comment. Is that one? Is that correct? And we will have that public comment--it's generally set for 10 minutes, and then we will go directly into the summary and discussion. Anne?

DR. KAPUSCINSKI: I was reminded that you need to read my name.

CHAIRMAN BUSTA: I was going to do that right after lunch, but do it right now. We have the time.

DR. KAPUSCINSKI: Okay. So, my name is Anne Kapuscinski, and I'm from the University of Minnesota; and I'm on the subcommittee, and I'm really here.

CHAIRMAN BUSTA: And she showed up immediately after we went through that activity this morning.

DR. KAPUSCINSKI: Twice.

CHAIRMAN BUSTA: Twice.

DR. KAPUSCINSKI: That's because I have to get some good coffee. It's my one weakness. One of my weaknesses.

CHAIRMAN BUSTA: Thank you all. I'd like to thank all the speakers this morning--and for your tenacity and the group. I look forward to an exciting discussion after the public comment this afternoon.

[Whereupon, the meeting went back on the record at 1:05 p.m.]

CHAIRMAN BUSTA: We have one public comment, and Mike Watson will introduce it.

UNIDENTIFIED SPEAKER: We have one public comment from Dr. Michael Hanson, from Consumers' Union, and he asked for 10 minutes to present to the committee.

CHAIRMAN BUSTA: Yes, please. Use the

microphone, so it's recorded.

DR. HANSON: Thank you. My name is Dr. Michael Hanson. I work for Consumers' Union. They're the people that publish Consumer Reports magazine. And we've been very interested in genetic engineering issues; have actually been commenting on all the--ever since 1992, on the FDA's various bio tech proposals. And I should also point out that Consumers' Union is part of the international network called Consumers International, which is composed of 270 consumer organizations in 117 countries. And we have observer status at the U.N., and we've also been very actively involved in the whole Codex process. We went to all the meetings of the Task Force on Biotechnology. And we would actually agree with Dr. Maryanski that that was a very positive process.

What I'd like to do here today is very quickly just make a few comments about molecular characterization. And these are the few points that I want to make.

First, we think that it's a very good idea that the FDA is now talking about requiring a lot of molecular characterization data, as is laid in these Codex documents. But we have a basic question, and that is, whether this will be mandatory and not voluntary, and also to what extent data is going to be required post transformation. Because, in what you were told earlier today, it made it seem like the FDA has sort of been requiring this data all along. But that's not really true. That was a new requirement that came in with this 2001 policy. And I'll tell you about three years ago, I was on a State Department Committee, this Trans Atlantic Policy Project, and it was focusing on biotechnology; and the idea was to try to get the U.S. and Canada and the European Union to work toward harmonization. There was a lot of debate within the subcommittee, because, first, they wanted to look at full biotechnology, like a--and they argued about should we require, should we look at a human data package or environmental issues. And they couldn't decide

on anything, so the one thing that they decided on in this working group was to focus just on molecular characterization. And what happened, this was I think in 2000, late '99, early 2000; when we finally met, data had come in from both the U.S. and Canada saying here is the molecular characterization data that's required. And the European Union, because I guess the data is different in different countries, they actually never submitted anything. So, this subcommittee has only met twice. We still haven't gotten any data from the European Union. But it should be pointed out that the data that was submitted by the U.S. and Canada, the molecular characterization that they said they were requiring, was functionally complete maps of the vectors and plasmids and what you were going to insert. But post insertion, it was just information to show that the transformation had been complete; that is, that the gene was in there and was active. But there was no requirement in terms of the data that they were showing us to do this--these nice maps of

how many insertion sites, what the structure is at each insertion site. So, that's what caused us some concern, and we really think that all that information needs to be in there. I would also point out that, in the 2001 pre-market biotech notification proposed rule, it says in this paper that was handed out today that it's just an extension of the 1992 policy. But it really isn't because there's, in the Federal Register Notice, they said for the first time they will be requiring data, and the data that they--part of the reason that they said that they were going to be requiring data is they said that the unintended effects that come from the random insertion, they called it insertional mutagenesis, because that can differ between different transformation events, the Agency said that they would require data on each separate transformation event, even if the vectors and plasmids used were the same, and the genetic background used was the same; that they would require separate data.

We think this is a very positive thing,

but we wonder whether the FDA is actually going to move forward with those rules, because I should point out that the molecular characterization data, for example, for the BT crops that EPA used in their safety assessment, they actually bridged, they allowed companies to use data from separate transformation events and have it go collectively.

And--so, we think that the data that should be required is that there should be complete molecular characterization of each line with respect to identity, stability, and unintended positional and pleiotropic effects. And the agency--the components of a complete molecular characterization should, for molecular identity, would include for each transgenic or transformed line the total number of inserts; the location of each insert, whether organelle, chloroplast, mitochondria, et cetera, or chromosomal; the exact chromosomal position of each insert; the structure of each insert, whether duplicated, deleted, rearranged, et cetera; the complete genetic map of each insert, including all the elements--the coding

region, non-coding regions, marker gene, promoters, enhancers, enterons, leader sequences, terminators, T-DNA borders, plasmid sequences, linkers, et cetera, including any truncated incomplete sequences; also the complete nucleotide base sequence of each insert. And we also suggested that the base sequence of at least 10 kilobase pairs of flanking post-genomic DNA on either side o the insert, including changes in methylation patterns. The reason for this is because the prominent use of hyper-promoters, such as the CAMV-35S promoter.

And the other main point I wanted to make is for determining molecular stability, we think there needs to be data on both functional and structural stability. And the functional stability should be the level of expression remains constant over time and over successive generations. And I would point out that just requiring Mendelian data is not good enough. For example, with the BT crops, there's a high dose strategy, and the high dose is defined as 25 times the LD-99. And so, if all you're doing, in some of the data that was submitted to the EPA, they would just say to show that something's stable, you just had to show that the trait still worked; that is, that it still had activity against these insects. Well, if you're talking about a high does being 25 times the LD-99, you could have variability, and one trait could be 20 times lower and only have 5 times the LD-99 of the cryoprotein in there. But if you did a simple bioassay with caterpillars, you would find that 99 percent of them still die. That is, you couldn't see the variability there.

So, that's why we think that for structural stability that the FDA would need data on the physical location of the insert in the genome, as well as on the structure of the insert throughout the lifetime of the plant and over successive generations, say, three to five generations. And, particularly, to look at that structural stability, we think there needs to be data on the flanking sequences, so you can see whether it's moving around inside the plant or not,

because we noticed that those kind of data were not submitted for the majority of crops that have gone through this process. Of the 54 crops that--54 things that have gone through the FDA's process, I believe only two of them have gone through since 2001.

And finally, for the other form of stability, we think that the FDA--I'm sorry--to look at the tests for unintended positional effects, we believe the FDA could carefully look at methylation patterns of genes in the flanking host genome DNA. And that's, again, 10,000 or 10 kilobase pairs upstream and downstream of the insertion site.

So, in sum, we think this is a very positive movement that the FDA is proposing, but we do have the question of whether these data, which they say are going to be required through the PBN, since we hear that that's going to fall off the radar screen, not be on category list A or B, whether the FDA will be requiring all these data that they said they will in the future. Thank you.

CHAIRMAN BUSTA: Thank you. Any burning questions? Just one.

DR. SALYERS: I just want to make a comment about this. Monsanto must love you, because if that grocery list is actually enacted as obligatory, it's going to make it even harder for small companies to get into the biotech; and for developing countries who follow our lead, even though they're not bound by our requirements to develop their own products.

So, I just think that we need to be careful about--it's always appealing to add on another item to the long list of existing ones. But the question is, is there a reason to do it? And are we doing anything but further increasing the power of the big companies in this area?

DR. HANSON: Well, my response to that is I think that all we're asking for is the kind of data that has now been agreed upon internationally; that are actually--it's in paragraphs 30 to 33 of this Codex document that was agreed upon. And the reason that this is important is, written into the

GAT agreement, which set up the World Trade Organizations, if there is disputes between countries, they'll look, written into the GAT agreement is, indeed, Codex Alimentarias is considered the neutral, scientific standard. So the concern we have is that some other countries could pass laws saying that they require all the data that is mentioned in here, and then they could turn around, and since this data is not required in the U.S., they could exclude products coming from the U.S., saying that they don't meet this standard.

So, we're not suggesting that further data should be added. We're asking that the FDA require the level of data that has now been agreed upon internationally so that there is not a problem with trade, potential trade impacts with products that come from the U.S. Because you should know, most of the world did agree that these safety assessments should be required. The U.S. does not require that at this point. So, that's why--we're not asking for extra data that will overly burden

developing countries. It's just the data that has been agreed upon globally.

CHAIRMAN BUSTA: Now, it's Dr. Salyers.

DR. SALYERS: Abigail Salyers. Sorry, I didn't hear--

CHAIRMAN BUSTA: Dr. Qualset. Go ahead.

DR. QUALSET: Okay, it's Cal Qualset speaking. It seems that if the developers are developing a product for U.S. markets, then why would they be required to get data that isn't essential to the evaluation of the efficacy of the product in terms of food safety? If the developers want to go to international markets, they can follow the international standards. But why make everybody fall into this, whether it's a small university, a small company, or large company? I think it should be clear that we have to evaluate for the product safety, and not necessarily all the ins and outs of its development.

DR. HANSON: Well, but I think a response to that is the main crops are engineered right now, corn and soybeans and cotton, are actually used a

lot in international commerce and are submitted--if this is only something used within the U.S., then that could be a different story.

DR. QUALSET: What I'm thinking is that there are people considering doing modifications to minor crops that may not ever get into the export stream, and they can't afford to go through all of this for a crop that will be sold to small farmers, for example.

CHAIRMAN BUSTA: Dr. Fedoroff.

DR. FEDOROFF: Nina Fedoroff. I guess my question is, could you say there's data from the literature that would suggest that this information that you think should be made mandatory is predictive of--is useful in predicting unintended consequences that affect food quality?

DR. HANSON: I can get back to you on that. There is--all I can say is that there was a lot of data that was talked about at these Codex meetings, and part of the problem is you have to know what you're dealing with before you can start to figure out what the safety implications are,

right? And so, with the unintended effects, since that's actually a new field, and there was a huge discussion at Codex. There's quite a number of paragraphs on that. They're talking about these metabolomics and the use of genome arrays. And I think that's useful, but I agree with the other scientists here. The basic problem is how do you take, once you've identified all those differences, how do you then figure out what the safety implications of them are. But before you can figure them out, you have to know what they are.

CHAIRMAN BUSTA: Dr.

DR. FEDOROFF: Yeah, but my question to you was how does accumulating all of these sequences and methylation patterns help you identify what the safety--that the food quality implications are?

DR. HANSON: Well, I can point to studies in the scientific literature with yeast and actually with tobacco where there have been unexpected effects that have led to toxins that were not previously present in the plants or to

things that have an--would have an adverse effect on DNA or other things, I mean.

DR. FEDOROFF: That's what I'm asking for. Could you cite those references?

DR. HANSON: Yeah, I can get those. One of them is Enos and Eurata. I mean, we can talk afterwards. I mean, I do have a list of those various studies that I can get to you--

DR. FEDOROFF: Please do.

DR. HANSON: And I will submit them.

CHAIRMAN BUSTA: Dr. Gurian?

DR. GURIAN-SHERMAN: Yeah, I guess one issue in terms of--

CHAIRMAN BUSTA: Introduce yourself. No, introduce yourself.

DR. GURIAN-SHERMAN: Oh, I'm sorry. Doug Gurian-Sherman. I think the economic issues kind of go beyond, you know, what we're supposed to be talking about here, and I don't think we're qualified to really even talk about them, because there's--you know, we could get into all kinds of sidetracking to discuss those. For instance, you

know, for pesticides, where there's a minor crop issue to getting pesticides for minor crops comparable to this with biotech, there's a set-up system that is not as effective as it could be but could be stronger called IO4 that probably a lot of your are familiar with to help small companies and minor crops get through the system. So I think we're just getting into areas that are way beyond what we need to be talking about here, which is the safety issue.

CHAIRMAN BUSTA: We'll take one more question. Go ahead.

DR. KAPUSCINSKI: I'd like to ask a question.

CHAIRMAN BUSTA: Would you give your name?

DR. KAPUSCINSKI: Anne Kapuscinski. It might be actually be better to ask this of Dr. Maryanski and of Dr. Hanson. I'm assuming when I got this Codex document that these guidelines on paper here are the outcome of a very rigorous and intensive scientific discussion, and that there is science backing these up. So I'd just like to get

a sense of whether that's a reasonable assumption.

In other words, when I read, you know, Item 30 in this, "in order to provide clear understanding of the impact on the composition and safety of foods derived from recombinant DNA plants, a comprehensive molecular and biochemical characterization of the genetic modification be carried out." And then when I read the stuff below it, I mean it looked to me like a lot of thought had gone into figuring out what should be considered and what is not that crucial to consider. And I think it would be useful if our committee doesn't try to second guess another scientific process that was, you know, broadly viewed as being scientifically based. So I'd just like to get a sense from Dr. Hanson and Dr. Maryanski about that.

DR. HANSON: Yeah, I'll point out first and Dr. Maryanski say anything else he wants. But as part of the Codex process, what they did is the first year there was a request for five questions on safety. And they asked that an expert

consultation be set up, and one was set up and happened in June of 2000. And so, a big technical report came out of that. They then at the end of that technical report said, well, there's actually the question of allergenicity. There is new data which suggest that the present methods aren't good enough. And so they suggested holding another expert consultation on allergenicity. That was held in January of 2001, and came up with an excellent report and recommendations. And then there was a third joint expert consultation that happened in October of 2001 that looked at safety assessment for foods derived--produced using recombinant DNA microorganisms. And those you

can--all of those expert consultations, the papers that fed into them, and the reports, they're all up on Codex's and FAO's and WHO's website. And yes, there was a huge amount of scientific consultation that went into those, and that ultimately is showing up in this Codex document. So I would say look at the results of those three expert consultations.

CHAIRMAN BUSTA: Dr. Maryanski, anything to add to that?

Dr. Maryanski will respond when he gets to the summary statement.

DR. ASTWOOD: Thank you. This is Jim Astwood. I have a question for Michael. When I read the discussion paper from the FDA, they basically concluded that, in the FDA's opinion, that their proposals and guidelines were--and practices--were consistent with the Codex guidelines. And listening to your list of suggestions, it was not completely clear to me which suggestions you had that were above and beyond the FDA practice. I did detect two, and I wanted to be sure I had those two correct. And if there are others, I'd like to know them. The one I detected was your suggestion that the methylation pattern should be determined. And the second was you suggested that there should be 10 KB of flanking sequence determined. Were there others?

DR. HANSON: No, those are basically the two. Although I have a basic question. Since a few

months ago, Dr. Lester Crawford got up on the Hill and basically said that they thought that the old, the 1992 policy, is fine, and they don't really need this new one. So I don't know what the status of this 2001 PBN is. If this is now going to be the new thing at the FDA, and all these data will be required, then those are the only two modifications that we would have to that. But I think it's still an open question since that hasn't been finalized, and we had heard 12 months ago that it was going to be finalized in 18 months, and now it appears to have moved off, and who knows if it's going to be finalized. So I don't know what the FDA is going to be actually whether they're going to be requiring this. If it's true, then we applaud that, and, if it's consistent with Codex, then that's fine. But that's the one question I have to the FDA.

CHAIRMAN BUSTA: I think, or it sounds like we're moving right into the discussion ahead of the discussion. So thank you, Dr. Hanson.

DR. HANSON: Thank you.

CHAIRMAN BUSTA: And we will now have the summary and review of the charges and questions by Dr. Maryanski.

DR. MARYANSKI: Oh, I'm sorry. Just a moment. Mike just needs to make a statement.

DR. WATSON: Dr. Hanson was the only person who asked for public comment in time either before the meeting or on a sign-up sheet outside. So we're just going to move to the next agenda item at this point.

DR. MARYANSKI: Jim Maryanski from FDA. Thank you. I just will take a couple of minutes here to provide a little bit of clarification. And I would like to thank Dr. Kapuscinski for raising that question about the information in the Codex, and what supported it, and Dr. Hanson's response because I realize that's something I need to add to my presentation are the expert consultations that FAO and WHO conducted that were--that served as background material for the consultations. And those were very important; and, as Dr. Hanson pointed out, there were a number of them. There

was one in 2000 that dealt with the question of overall safety assessment and the application of substantial equivalence. And then there was a session on allergenicity. And there was a session on the use of microorganisms to produce foods that Dr. Cebula and I actually did by video conference from Washington in the middle of the night for a week because of 9-11. But those consultations were part of the information. Now countries also brought forward information on their own part, based on their own experience, so they were not the only information that was behind the Codex guidelines.

But the basic answer to your question is, at least from my perspective is I do believe that those guidelines are based on what the Codex member countries that participated in developing, understand, were the best science available at the time.

I would also like to point out that the Codex guidelines are just that. They are guidelines. The Codex guidelines are not binding

on any country, so they're not requirements. And they're guidelines in the sense that you should use them as they're appropriate. If certain information is applicable, it should be used. If, in some particular case, it's not, it would not. I mean, that's how--that's the intent of guidelines. It's to provide a framework, to provide a general direction, always recognizing that some questions may be applicable and sometimes there may be questions that aren't within the guidelines that need to be added because there's some difference about the question at hand. So we always want to think of these not as a checklist that you go down, but as a tool that you use in the process.

And in terms of requirements, whether the data is required or not, in our view at FDA, is not the question for this subcommittee, because what we're asking you to think about is if we're assessing the safety of a food, and, in this case, we're looking at specifically the use of molecular biology data, what data can contribute to answering the question about safety. So whether it's

required by FDA or somebody else, or not required, is not the question. What scientifically will help us do our job. That is the question that we're asking you to think about in the context of how one can use molecular data to support safety assessment.

Now, the question was raised earlier of how much is enough, which, of course, we're always asked. How much is enough, which is a difficult question. But that's why we've established this safety standard that we're not trying to prove absolute safety. We're trying to get at the question is this new variety as safe as what went before it. Now, we don't expect the molecular biology piece of the safety assessment that you're talking about today to fully resolve that, because we're going to be taking into account all of the information that we have in front of us in reaching that final decision. But in thinking about how much is enough, how is this going to help put the pieces of the puzzle together so that ultimately you have that level of standard is I think what

we're asking you to think about.

The other thing we take into account, and it goes to some of the discussion this morning about the wrinkled pea, and the fact that you can see all kinds of changes in the plant, in the food, and they're not always relevant to safety. We, of course, ask companies to compare the new, modified variety with its parent or some comparable counterpart in terms of doing the analysis at various steps. But that's not the end of the game, because they often see differences between the modified and the parent plant at what the statistician would tell you is the 95 percent confidence limit.

What we do ask, though, what's important is are those changes within the range that's typically of that plant, when that plant is produced commercially, has been accepted as a commercial food product. That's what's important, because we've recognized that genetic background and differences in weather patterns and growth and stages of growth, and all of these things affect

the composition of the food in different ways. So we have to ask the question is there a change that's outside the pattern that's been accepted as safe. So that's how we really look at the data in terms of food composition, which, of course, again is not the question we're posing for you today, which brings me to the study of the Academy.

DR. FEDOROFF: You can start singing.

[Laughter.]

DR. MARYANSKI: Which brings me to the study of--yes, I'm sorry I don't have that kind of voice at all. But we have a study with the Academy on unintended effects, and that's a much broader study than what we're asking you to think about today. They are there thinking about the overall question of long-term effects and unintended effects on human health as a result of genetic modification. So today, to come back to what we're asking you to think about is this part of the safety assessment that deals with the molecular biology data, and how that data can help us better understand questions related to the safety of the

food as part of the overall safety assessment process. And that would include, for example, how can this data be used to generate--to identify new substances that may be in the food, because one of the questions we'll ask is if there are new substances in the food, are they safe. The obvious question. But you may also be thinking about the fact of gene insertion in the genome, and at the molecular level. Is there data that will help us understand? To what extent can we get at the question of other effects in addition to the introduced substances that may be important to safety assessment.

So, we're not asking you to take on the whole issue of unintended effects, but really it is how can molecular biology help in this process. That's what we've asked you to do today. I'm just looking at the Chairman.

CHAIRMAN BUSTA: Dr. Fedoroff.

DR. FEDOROFF: You could help us.

CHAIRMAN BUSTA: Dr. Nina Fedoroff.

DR. FEDOROFF: Nina Fedoroff. Sorry. You

could help us a lot by telling us how these different kinds of data have helped you so far, if they have helped you. In other words, you've collected, if not all of this--these different flavors of data, some of them; right?

DR. MARYANSKI: Mm-hmm.

DR. FEDOROFF: Have, in any case, they helped you assess the food safety?

DR. MARYANSKI: Well, I think I can give you an example I think that Dr. Cebula referred to this morning. Early on, you know, that--there was a basic mantra that what you used the TI insertion process; that the information--the region between the borders is what goes into the plant, and the region outside does not. And when we actually asked CalGene to look closer--you may recall from the flavor-saver meeting, when they actually did the PCR outside the region, where there is an antibiotic resistant marker on the backbone of the plasma, they found that actually an insignificant proportion of transformation events that material from the backbone--in other words, the biology was

not as precise as just the borders. It was--were occasions when that material would be transferred. And, so, then, companies then did begin to look at the inserted material to be sure that it was introducing the material that was intended.

So this information can be very helpful in that sense, because that allowed us then to eliminate the question of having to look at that other antibiotic resistant marker; and, you know, did that need to be considered.

And, so, there are certainly examples where this helps. We found it very helpful, for example, using some of the blotting techniques to identify the presence of transcripts, as you heard this morning, to find out are they the transcripts that are expected or are we also seeing other transcripts, and then have to sort of ask the question of are those meaningful or not. Or Westerns to get at the question of how much protein or what proteins are being produced.

So, there certainly is information at the molecular level. And one of the things I have told

people in many talks around the world is that, you know, we often hear about plants being developed by R-DNA as new plants, a new technology, and these new things are coming. But, in fact, it's the molecular biology techniques that have given us much more powerful tools in terms of safety assessment. In other words, being able to understand the changes that have been made at the molecular level so that we can identify the new substances in the food, so we can decided whether or not they pose any safety problem.

I think, and certainly if any our other colleagues here, you know, feel there's another point, they can put up their hand. But I think that's the crux of, you know, what we've been seeing in the molecular biology data.

Now, I would add to that we have also asked companies to look at the stability. And the stability, as I understand it has been to address the question of unintended effects down the line, as these plants are--you know, will be bred, of course, by conventional means and further

developing the cult of ours, the idea being if the initial insertion event has been done in a way that that's stable in the chromosome, that once that plant then is put into the normal breeding process and, as breeders do, and bred into a number of different lines, that any changes that will then take place afterwards will not be due to that insertion event, but will be the normal kinds of changes that take place through the other methods of plant breeding.

I'd suppose one of the other things we've looked for is how many insertion sites there are, just to get at the question of what's the likelihood of recombination events occurring, as well.

So, that's--so, yes, we have found some of this. We felt some of this information has been useful. But, again, you know, we're not here to tell you what's useful. We're really asking you, from your perspective, what do you think will continue to help us.

DR. FEDOROFF: I guess what I'm hearing is

that you found that the T-DNA transformation or whatever transformation is messier than you thought it was. But what I didn't hear you say is how any of those things have actually helped you identify a food hazard, a problem. Okay? A real problem that you said, okay, like the potatoes with--high

glyco--high solanine potatoes. Okay, that's a real problem. You don't want it to get out in the marketplace. And I guess what I'm asking is, in all of the cases that you have analyzed, did this information allow you to identify some such problem coming up the pike?

DR. MARYANSKI: It kind of gets to the question we're often asked of how many times did we say no to a company. How many times did we, you know, reject a submission because there was some problem with it.

DR. FEDOROFF: Well, there's probably--

DR. MARYANSKI: And we really haven't done that, and the reason is because the guidelines and these questions that we set out really help the companies to, you know, basically make sure that

the products that they have don't raise any of these questions.

So, I think it's not so much a matter of FDA finding the problem, but everyone now understanding what the questions are so that the problems can be avoided, and generally are avoided. I mean, we did have the one example of the Brazil nut in soybean that, of course, turned out to be an allergen. But, otherwise, no, I don't know of any cases where we have identified a problem.

DR. FEDOROFF: A problem.

DR. MARYANSKI: Yes.

DR. FEDOROFF: By using those particular molecular markers.

DR. MARYANSKI: Right.

DR. FEDOROFF: Okay.

CHAIRMAN BUSTA: Dr. Arias.

DR. ARIAS: Jonathan Arias. I think perhaps the answer to that question was given earlier by the number of applications that have been submitted so far to the Agency, which I believe were about 50 over the last several years

for safety assessments of transgenic plants. Most likely, there are just too few examples to get a sense of the probabilistic issues of safety, and I assume companies and developers as good stewards are already screening those that show potential problems in terms of agronomic traits right off the board. So, you wouldn't likely see examples of those.

I do want to also point out there is, of course, one highly contentious example in the public and the scientific literature in regards to the safety of genetically modified foods that would, I think, impact our discussion: that is, of course, the study by I think it's Rowers and Hustive in the Lancet a number of years ago. The conclusions of that study, of course, have been highly debated, and I'm not going to redress them here. But they do raise the potential issue, and I think that's what we're really dealing with

today--is the potentialities for risk, given that certain processes may be mutagenic or altered the metabolic profiles of plants. And there are

also--the other fact is that there's clearly an end point to this, which is predicated on time, money, and other issues in relationship to the public safety concern.

And, so, I think those have to be factored into this, as well; that the discussion hopefully will focus on the fact that what are the potential risks that might be likely given certain circumstances implicit in genetic engineering.

CHAIRMAN BUSTA: Dr. Gurian.

DR. GURIAN-SHERMAN: Doug Gurian-Sherman. I guess I'd have to ask a further question in terms of the kind of data that we're being asked to look at today, and further on Nina's question is, of the submissions that you've actually looked at, what percentage of them actually looked at the

specific--some of the specific kinds of data that, let's say, Michael Hanson mentioned or that are mentioned in the Codex, such as potential for fusion proteins, potential insertion mutagenesis into an open reading frame--that kind of thing. Because, again, my awareness is that, at most, a

very small minority of the submissions actually had that kind of data. So, I'm not sure that there's a very--much of a data set on those specific questions. And I think, you know, to go a little further, I think the general potential issues are just that--are we inserting something into a gene that may have an effect. And, again, it's a potential risk, and I think part of what we have to think about is what, you know, what kind of potential risk is there from an insertion mutagenesis, or for activating several upstream or downstream genes. It's a, you know, that's the potential risk that I think we have to address is how much should be done; how much is it worth, you know, pursuing.

CHAIRMAN BUSTA: Jim.

DR. ASTWOOD: Jim Astwood. I have a comment or an observation based on Dr. Maryanski's response that of the 50 or so petitions that have been reviewed by the FDA, certainly there have been no cases of safety issues being raised on the basis of molecular information. And he indicated that

that's not surprising because there are a lot of screening activities done by companies, and surely that's the case.

But as a matter of safety, I would say that, as a practical matter, the molecular characterization has led to very few decisions, at least in my company, about moving a product forward or not forward to the FDA. And the reason for that is that those unexpected effects, those pleiotropic effects that Dr. Gurian has mentioned are readily apparent when you do agronomic evaluations and compositional analyses by chemical evaluations, nutritional studies in animals. That is the way you pick up the unexpected effects. And so, it's not surprising that nothing has come to the FDA. Likewise, companies really are not using that as a criteria from a safety assessment point of view. And that's--that gets to a comment about whether we need--whether we're operating at a phenotype level or a genotype level, and is the characterization of the insert a matter of responsible information or is it truly a decisive study in terms of is the

product safe or not safe. I think it can be decisive, but it's--so far, it has not been our general experience that it has been a decisive study or a pivotal study.

CHAIRMAN BUSTA: Dr. Kapuscinski.

DR. KAPUSCINSKI: Yeah. Dr. Maryanski, I have a question that I think--

CHAIRMAN BUSTA: This is Anne Kapuscinski.

DR. KAPUSCINSKI: Anne Kapuscinski. We have a--I think actually will follow from what Dr. Astwood was just saying. I wanted to ask a really elementary question. When you're asking us in our charge about molecular biology data, are you using molecular biology in the narrowest sense to just asking us for data at the level of DNA and RNA? Or are you including the expression, expressed protein? Because I guess the way I'm looking at this is that I would want to start with some level of the phenotype of the organism, ideally, knowing what's useful about the proteins expressed, and then higher level organismal traits, people often refer to those as agronomic traits; and then maybe

use that to guide me. If I see something that raises a red flag, to guide me and to try and figure out if I need to get more information to figure out if it's really a risk. And maybe, at that point, some of the DNA- and RNA-level information might be useful. And I'm saying might because in some cases, it might not be that useful. I seems like down the road, if we got the point that we knew that every time we noticed a certain class of proteins being expressed, that it gave us a certain kind of signal on a DNA chip, then, in the future, you could maybe just rely on the DNA chip data early on in your production of a engineered plant to quickly identify that. But we're not really at that point yet, because we haven't yet been able to do those kinds of correlations. So, I'm trying to get a better sense of really what you're charging us to do, because I feel that it's not going to be very useful for us to limit our discussion only to DNA and RNA information. It feels to me like that's just looking at that stuff in a vacuum. When, really,

what's going to drive whether the thing is safe or not is the actual phenotypic consequence of changes in the DNA that's being transcribed and actually then what happens with those RNA transcripts.

So, how narrow or broad would you like our discussion to be of your charge?

DR. MARYANSKI: Well, I think we don't want to--

DR. KAPUSCINSKI: You see what I'm getting at?

DR. MARYANSKI: Well, we--what we would like to do is have you focus on the part of the safety assessment that can be gleaned from molecular data. So, in other words, we're not asking you to take on all the pieces of safety assessment and evaluation. For example, we're certainly interested in what happens at the gene level and the expression of message in protein. We feel confident about evaluating the safety of substances once we know they're components of food. If a protein is identified as it's going to be in food, then we feel that we have the capability

today to do that. That's not the question that we're posing to you of how do you ultimately, then, determine if that protein is safe, for example.

So, yes, the expression of the message, the expression of the protein, there's certainly some things that we would hope that you would talk about. But, you know, I don't want to place limits on what you, this subcommittee, might decide are issues that you think are important for FDA to think about from data that can be gleaned at the molecular level. I would prefer to leave that to your deliberations.

We've put down a set of questions as a way of giving you some of the things we've been thinking about. They are not intended to be things that we absolutely need answers to or the only things that you should think about. But this is intended to be guidance so as to get you started at least in an area that we have some question about.

CHAIRMAN BUSTA: You know we could keep you up there for the next two hours, and see how--

DR. MARYANSKI: No. No.

CHAIRMAN BUSTA: To see how resilient you are.

DR. MARYANSKI: I appreciate the opportunity, Mr. Chairman. I think I'm done. Thank you.

CHAIRMAN BUSTA: But what I'd like to propose is that we go around the entire group--it's been a normal group technique of each one expressing some of their thoughts. This is not the final discussion, but limiting it to no more than maybe about five minutes of where you see us at this point. How you would suggest we address our charge, and your thoughts on some of the issues that we've been given. That--knowing how time slips, that will probably take us to at least a bio break. And then we can come back and do a conclusion-type discussion of one-by-one, sometimes called the vote, of where you see the committee's comments should be going. If it appears that there's a significant consensus, at that point, we can try and summarize then. If it looks like it's a number of individual comments, they will go in as

that.

Is that acceptable for an approach?

All right. I've always had a problem. Those of you that know Meyers-Briggs, I'm at the

E-level, at about 99 percentile on E. And, so if you lose me every once and while because of the laughter next store, these external inputs really distract me. I'm trying to focus real hard, but I just want to go over to see what's funny.

[Laughter.]

We can start in one way. I've started one there, one time there. Jim, would you like to start?

DR. ASTWOOD: Gosh. I can start. I mean, I've made a couple of comments. I think that in reacting to the issues that's presented in the charge from the FDA, I would say that certainly I would agree that the FDA's proposals and practices are consistent with the Codex guidelines. One of the things that I continue to struggle with in the safety assessment realm is really this question of how impactful molecular characterization is to the

safety assessment. We can talk about theoretical fusion proteins. We can talk about theoretical rogue transcripts, a lot of things. What I tend to come back to is the overall risk assessment and all the other tools that we bring to bear. And whether or not that we lose sight of those tools when we are thinking about molecular characterization. I mean, specifically, Dr. Arias brought up the Pustai example. And for those who aren't too familiar with the Pustai example, Dr. Pustai and colleagues in Scotland evaluated a number of transgenic potato lines that were transgenic for lectins. And they showed, and there's some debate about this, but superficially they showed that these potatoes were, in fact, toxic to rats. Toxicologists debate that point, but let's accept that as a fact; that they were, for the sake of argument.

Can one predict on the basis of the molecular characterization of those potatoes that those potatoes would have been toxic? Probably not. Except to say that a lectin was expressed in potatoes, and lectins are known to regulate gene

expression of plants; and, so, perhaps it could have affected other genes. It turns out, in that case, that those potatoes were not compositionally equivalent to traditional potatoes. And, so, in this case, those potatoes would have failed a key test in the Codex guidelines or the FDA process, which is that those potatoes fall outside the range that one would expect for the composition and nutritional profile for potatoes. And that would raise concerns that would have to be addressed in the FDA process or in the Codex process. You could not have predicted that knowing the molecular arrangement of the lectin gene in the potato plants. And, so, for examples that we know about from the literature and from our own experience as producers of this technology, those tend to be highly pivotal studies versus the molecular characterizations.

So, when we think about the broader debate, there's probably no end to the amount of characterization one can do. One has to make a decision about how much risk you leave on the table

in terms of the compositional analytes you measure. How much molecular characterization you do. How many animal feeding studies you do. How many toxicology studies you do. You're always leaving a little bit of risk on the table; and the key question, therefore, is when do you meet the standard of reasonable certainty of no harm for the FDA. And I think the FDA has done a very good job of describing scientifically backed internationally supported lists of criteria for molecular characterization that I think is rational. But someone needs to articulate exactly I think, or in some more precise ways, when you will need to do actually more than that. It's not clear to me that you do.

And, so, it's a question to ponder. It's probably bigger than this one meeting, but I'm sure as we move forward, we'll have opportunities to ponder it in the future.

CHAIRMAN BUSTA: Dr. Gonsalves.

DR. GONSALVES: In analyzing the papaya, we actually maybe have some unique experience in

that we went to the FDA guidelines back in 1996; and we actually have also--are moving through the guidelines for Japan, which, more or less, follows these Codex rules. And so, we've had to analyze in much more detail the papaya than we--now than we did in 1996. And my general conclusion is that more or less what we knew about--what we knew in 1996 based on previous history of viral coat protein and so forth--what we analyzed in getting more data did really not add much more to what we knew back in 1996. And it's not surprising to me. Like either the question of allergenicity, I know there was a report saying that this papaya ring spot may be allergenic and so forth. But with all the history of eating virus-infected papaya for years, and we deliberately instantly inoculated millions of trees with a mild strain of the virus and sold those papaya commercially with no impact on allergenicity. I was almost certain that our transgenic papaya was not going to be allergenic.

Now, should I have, at that time back in 1996, sequenced the--analyzed the amino acids and

then checked the susceptibility of the co-protein to pepsin or other stuff. Well, that remains debatable. We did not do it. But I think the preponderance of data was such that we really did not have to do it.

Now, getting back to another tack, I think there's obvious areas in which we have to pay attention to alkaloids and so forth. But I also think that sufficient data have come forth that there are areas in transgenic plants--for example, viral coat protein resistance. Well, that is a major thing to many people that produce crops. It appears to me that in that kind of case, it might be a lot of the data that have been accumulated, we may be able to recommend just a few maybe key tests or analysis of the protein that might suffice so that you don't have to go through a lot of all of these other tests.

Now, with a lot of the other stuff, I would think you definitely want to go through that, because they are definitive, well-documented things that could happen.

CHAIRMAN BUSTA: A compliment. Two people have automatic timers. They were in four minutes, four minutes apiece. I'm impressed.

Dr. Kapuscinski.

DR. KAPUSCINSKI: Thank you. I would actually. I would appreciate having more time to reflect on this, like, you know, and speak again after hearing everyone. But my thoughts right now are the following: first of all, when resources and time are limited I guess I'd rather see more effort into characterizing the proteins than in to the DNA and the RNA, because I think that alone is not--if we focus only on that level, I'm just really worried about losing the forest for the trees, if we focus only on that level.

And it seems like the big challenge is looking for pleiotropic effects. That's where our level of understanding, whether you're a quantitative geneticist and have done traditional breeding or a molecular geneticist, that's where our level of understanding is really very poor. And I don't see right now how only doing stuff at

the DNA level is going to really increase our chances of detecting pleiotropic effects. So, if someone comes up with a way that combining those methods with some protein-level analysis and even higher level phenotype analysis that might hint at a pleiotropic effect, if they find that using DNA methods could help in screening for that, then that would be great. But I'm personally not aware of some kind of screening like that right now that's helpful. So, and I don't have a really detailed advice about what FDA could do to do a better job of looking for pleiotropic effects, but I think that would be a really good topic for this committee to maybe examine in more detail. And, again, you don't want to look for pleiotropic effects just ad infinitum and for ever. You would want to be very targeted and strategic in searching for that by using the kinds of ideas that Dr. Gonsalves just talked about. There are clearly certain kinds of organisms, such as potatoes and tomatoes, or we know already there are some metabolic pathways that might cause some problems

if they get upregulated, for example. So, especially if a developer is using a construct that does not have tissue-specific expression, you know, then that might raise the desire to look more proactively for pleiotropic effects. And, if they're putting it in a kind of plant that there is some history of knowing that there are some chemicals, biochemicals in that plant that might cause a health risk. So, in other words, you'd want to combine things like that. So I guess I'm more interested in taking a more integrated approach to it.

My final thought, just to reemphasize something I said earlier in the question-and-answer period, is I could see that maybe five years from now if there's some well targeted research that's conducted to help this, that you might reach a point where you could find some DNA signals that would be strongly positively correlated with evidence of something at the protein level that is a problem. So then it--maybe five years from now, you could use the DNA-level screening as a first

cut kind of early on way of testing. So it wouldn't be if a DNA itself is the problem, but it's positively correlated with some levels of changes in protein expression or even secondary, no probably not secondary metabolites, but changes in the levels of expression or certain genes getting turned on that, you know, turn out to pose problems. But I don't have a sense that we're at that point today.

CHAIRMAN BUSTA: Three out of three. Timing is great. Dr. Buchanan.

DR. BUCHANAN: Yes, I'm willing to follow that perhaps in a little different way. First of all, there are far greater experts in DNA, the area of DNA stability than I am, and DNA rearrangements. But I have, in the past few years, our group has embarked on studies with proteomics, and I'm very impressed with the power of proteomics. I hope we're not in the formative stage that half of what we've found turn out to be artifacts, as referred to earlier.

But we've discovered things that would

have taken years before the advent of proteomics. That's why I was interested in the Mendel, the pea experiment, with proteomics.

And, I think as time goes on, we will have maps of major crops called proteome, as we now have for arabidopsis. There's a group in France that's developed that, and it's a very nice website with many of the proteins. We are developing the same in collaboration with the USDA laboratory in Albany for the endosperm, the wheat endosperm. And one can see with great precision when a protein is different. And I think with time, whether the time is right for this now, whether the technology is at that stage, but one can compare the genetically engineered plant with the null segregant, the protein content of the organ or tissue of question.

And I think one can see the differences, as we've seen with the Mendel peas. And 62 proteins is not a great number now to identify with the common methods.

And I think one might pick up interesting candidates, and it's not a great expense for most

companies, either, or will not be.

And then one, I would have to think it about it more, but one could go further and look at subsets of proteins like gylcosylated proteins, populations within the null segregant and within the genotype of interest. And I think with this, it would be reassuring without undue labor and undue costs to convince one of the safety of the particular product. But, as I say, I think with the--in the future, this is going to become very common, and we're going to have proteome maps of certainly the major food crops.

CHAIRMAN BUSTA: Dr. Salyers.

DR. SALYERS: Well, basically, I don't have to take much time because I'm going to--this is Abigail Salyers. I'm going to underscore what Dr. Kapuscinski said. I think that it's important to--at least as I interpret--as interpreted through me, I think it's very important that there is some flexibility in this, in the FDA's response to new products that come through the pipeline. I think it's a very bad idea to have a shopping list that

is applied uniformly to everyone for the simple reason that there are some cases in which there are problems that might be missed that way. And, for example, we know the cases--all of these come from traditional practices, where we have seen people produce foods that actually made it to the supermarket that were toxic at some level--have been. The alkaloids have been the problem.

So, it makes sense in certain foods where we know that that is a potential problem to look at the alkaloid levels in anything that people mess around with genetically. And, so, I try--I guess there are those of us who trust the FDA, and those of us who don't.

I trust the FDA, their intentions and the abilities of their scientists; and, therefore, I'm willing to accord them a certain amount of flexibility. I think it's a big mistake to--I think we're fooling ourselves if we talk about taking DNA sequence information at our current level of understanding and extrapolating too far from that, because I don't think that we right now

have the information that would allow us to do that. And, so, you get yourself the illusion of safety when actually it's not there.

So, I argue for an approach that is one that has a certain amount of commonsense, based on knowledge about the individual type of plant and the way it will be consumed, the uses of which it will be produced; and structures a regulatory program for that particular areas. And I wish we could build that kind of flexibility into anything that is done rather than having certain set lists of things that are applied uniformly to every example.

CHAIRMAN BUSTA: Thank you. Dr.

Gurian-Sherman.

DR. GURIAN-SHERMAN: Yeah, I'd like to build on a couple of the comments here. I think the comment that Jim Astwood made was a good one that, you know, these things all have to be looked at in a broader context of the overall risk assessment, and I think Dennis also addressed that. There's differences between types of genetically

engineered organisms and probably the inherent risk in them.

And I guess I would also agree with Abigail that there needs to be some flexibility to address those differences.

I guess the concern I have is that, in part, that I'd like to see more of a system that does both, that tries to, where possible, and where there is an issue with the types of tests and the protocols, that there is some guidance, some specific guidance given, but in the context of broader assessment of the risk of the plant. I mean we certainly do know that some of the protocols--there's been experimental data, and I don't think the characterization issue is necessarily the best one to address this--but where differences in protocols make differences--result in differences, practical differences in the result, and probably in genetic engineering the best one, at this point, is with pepsin digestion, which is part of the allergenicity assessment, where an FDA scientist varied the conditions of the

experimental protocols, and came up with very different results in terms of how you might interpret the allergenicity of a GE protein based on its stability.

So, I think it is important that there's some--in some cases, some uniformity.

On the specific issue here, I guess I'm going to go back a little bit on what I said before in terms of economics. The--there's a lot we don't know, and it's not predictive about sequence data. But, on the other hand, we're developing more and more, through genomics, databases of proteins that have, to some extent, known functions. And I think that has to be weighed with the, again, in part how difficult is it to get this data, and I guess I agree, although on a slightly different issue, with Bob that some of this data is not that difficult to get. The sequence--DNA sequence, which then can be translated in the protein sequence of the inserted DNA in the plant is not very difficult or expensive to acquire. Now, is it going to tell us everything? No. There's a lot of things it's not

going to tell us, and there's always the risk that when somebody sees a difference, that they're--you know, that they're going to interpret that as being a significant difference in terms of safety, which will often not be the case. I think we have to have a little more confidence that when we have that kind of data, that we get what we can out of it in terms of looking at homologies with allergen epitopes or looking at homology in toxin databases or allergy databases to see if potential changes through the transformation of that or insertional mutagenesis or fusion proteins do, you know, tell us anything. But again, I think that has to be weighed with, against, you know, how difficult is it to get this data. And I do share the concern that, you know, this kind of analysis could go on forever, ad infinitum, and I'm certainly not pushing that. But I think in this particular case, you know, some of the types of data we're talking about--getting Northern blots, you know, to look at expression of flanking sequences, that kind of thing is not, you know, that difficult; and may,

may, in some cases, reveal some useful information.

So, again, I think, you know, the point that's been brought up before. You have--it does make sense to weigh how difficult is this data, and then how useful is it.

I guess the final point I'd want to make is that, you know, Nina had a good point earlier I think that, you know, there is often a lot of data floating out there that can be useful if we try to gather it together. I haven't seen enough. I'd like to see more of what we might be able to find out with some of this data. In other words, how predictive are secondary protein structure models. Or, you know, that if you see a difference in the structure of--the sequence and structure of the protein how much can we, you know, derive from that.

So, I think we're also, you know, working a little bit blind here, and working at the last minute trying to digest a lot. And I think it would be useful to try to pull together some of that data before we make any kind of substantive

decision on this, or recommendation.

CHAIRMAN BUSTA: Dr. Qualset.

DR. QUALSET: Yeah, I'm Cal Qualset, the token plant writer, to emphasize my point.

The issue I am hearing is what I would call mostly dealing with development stage information. It's molecular data that's useful in the development of the product, and maybe not so useful after it's into the evaluation stage for FDA.

That doesn't mean it's not important, because I believe it's useful in developing the guidelines and the protocols that the developers are using to know that they should be looking for certain inserts and that sort of thing. So

that--what I am hearing is that I'm not sure that that information can be translated into safety and risk.

Now, Bob talks about proteomics, and once we know what a protein does and structurally, that it might be related to some undesirable effect, then we can use that information. But we need that

kind of trait-based linkage to the molecular structure.

So, what is safe and what is not safe? What do you know about that question when you are doing molecular characterization. So that's a question.

I'd like to think we are talking about consumer oriented evaluation of safety, food safety, and that we should be looking at, very carefully, the ways the product can be used, whether it's in traditional food preparations or unusual food preparations. And some of the products may be targeted to specific components of the population.

If that's the case, then we need some post-release identity issues, and there may be, Bill, some molecular characterizations there that can be very quick and prove that you, in fact, are looking at the variety as it was released. Labeling it comes up in that context, however.

So, those are the issues I think we have to think about if we're talking about molecular,

value molecular characterization in evaluation of food safety. I think the connection is tenuous in many respects, I think.

CHAIRMAN BUSTA: Thank you. Dr. Benedict.

DR. BENEDICT: Steve Benedict. I have sort of in front of me some fairly hastily scribbled airplane notes that, as I was reading the documents on the way out here, and they turn out to be for me, anyway. And I want to agree with almost everybody--what they said. It seems to me that genomic evaluation has a couple of values, but they're fairly limited. And one thing I don't think can help us is any kind of array technology. I really can't see how the effort and expense we have to go through to get that analysis is going to tell us anything at this point. And what everyone has said is, yeah, ten years from now, we might be able to know some things.

I also agree that the real relevance here with respect to safety is something a toxin. Is something an allergen, or is something doing something odd that we can't predict. And, as a

couple of people have said, except for things that I don't understand, like, non-proteinaceous materials, which have to be done by compositional analysis. The proteins that we're putting in, we're going to know going in, based on databases is this an allergen, is this a toxin, or is it related to them, and so that will require considerations. But if it doesn't do any of those things, then we have to ask questions about changes in the proteome. And I agree with Dr. Buchanan, this is a powerful way to go.

And, so, I think that we could end up with an iterative process, a decision tree where FDA asks before the projects begins even, is your gene of interest an allergen or a toxin or a related sequence, 80 base pairs or 80 amino acids, whatever that thing is. If it isn't, well, go thou forth because you don't have a problem yet.

And, so, then we ask questions based on the proteome. If you put your protein into a cell, a tissue culture cell just to find out will the presence of that protein change the proteome in any

way. And, again, you're only talking--you're not talking about looking at wrinkled versus turgid peas. You're really asking--you've got an organism. You're changing one protein in it. So, does that have secondary effects on other proteins. You can ask that quickly with a two-D gel, and if you don't see anything, okay, now you go engineer your organism. And, so, you ask these questions back and forth, and none of this, I think, is particularly expensive with the possible exception of tissue culture, which I don't understand in plants. And, so, looking--following the changes in the protein seems to me the place to go, and there are always places to bail out.

Then when you get to the point of analyzing the vector, that's already been said. The real question for me there is how strong is your promoter, and are you doing, as Dr. Arias said, have you selected some tissue-specific promoters like we do in animals. And if you can do that, then that helps you also. And so, this can be part of a list. And then, once you're

transfecting--I'm one of the early 1994 proponents of learning about upstream and downstream sequences. But that was before we could look at proteomes and determine what's going on. And so, again, if you have no problems, then we don't need to I think ask many of the questions. The relevant question with the genome to me is, how many insertion sites are there, because that gives you the chance of homologous cross-over, and that could mess up your plant, big time, I guess.

So, if you ask that question, and the answer is you don't have that, then again, you could go forward. It seems to me that we don't to apply everything all at once. We can apply them at the stepwise, logical fashion, but always based on what proteins are being expressed, not necessarily what data we can get. But what's possibly going to cause the problem.

And I think I have one other point. Give me a second. How am I doing?

CHAIRMAN BUSTA: You're doing fine.

DR. BENEDICT: Oh, you're so gracious.

CHAIRMAN BUSTA: Everybody's doing so fine, I can hardly stand it.

[Laughter.]

DR. BENEDICT: So to sum it up, I think we have to concentrate on the proteins or some compositional analysis that tells us that something has been changed; and if, as we go through the steps nothing has been changed, and if FDA works with the producers all along, then it could be a very painless process. And Dr. Gonsalves brought up a question at lunch, and that is what about the small guy who maybe can't afford to do this. There can be a cut-off, an economic cut-off, the size of your company, and why not have FDA, as he suggests, do a simple two-D gel and analyze it with--I can't use a company name--with a series of database, and send it back to them. Help them out.

I think you can do it, and I don't think it's such a big deal.

CHAIRMAN BUSTA: Okay. Dr. Arias.

DR. ARIAS: Jonathan Arias. It's wonderful being at the end of a long line of

distinguished speakers because I can be truly brief and say I concur. I just will add a few additional points.

I think, really, all of the comments that have been made are right on the mark, and, in particular, we need to recognize a couple of things that is that the state of the art is intrinsically flawed. Most of the techniques, even in measuring changes in gene expression, may not predict deleterious effects for human health. That's the state of the art.

Yet, we still have to deal with the risk management aspects of that. And so it to be prudent, clearly some of the specifics that have been outlined in the draft guidelines that we've received for this subcommittee make good sense.

Looking at changes in the gene itself, the transgene, its insertion site, and any attendant effects, like translocations of the genome in which it's inserted, would be prudent, because, clearly, we want to avoid any untoward effects that are not anticipated.

And, so, I would say that the state of the art still precludes many advanced analytical techniques, which have been raised by various members of this subcommittee, such as functional proteomics, which I do believe, in the future, will probably provide enormous insight, yet currently lack the predictive power to tell us about what, again, are the risks that changes in protein expression would engender to the consumer.

Given that, I would like to propose a slightly different tact in this, and that is, a proactive one taken by the FDA and the developers. And some of my points, perhaps my thunder has already been stolen by various subcommittee members, and I welcome that; but I think some of the points that I'm about to make seem very prudent indeed.

As we know, in certain cases, expression of the recombinant protein in the food or edible portion of the plant raises red flags in and of itself. As the state of the art in plant gene expression has advanced enormously over the years,

as evidenced by numerous publications, the use of tissue-specific promoters, as Dr. Benedict mentioned, would seem a very good prospect for redirecting expression in those cases where it can be done of a recombinant protein to those tissues that it needs to be targeted to, preferably those that may be outside of the human food chain.

However, we also recognize that coming through the pipeline, no doubt, are modifications which will affect the sort of nutriceutical portion of the plant, the edible portion of the plant. And, in those cases, clearly a different level of analysis may have to be done. So, we could see a decision tree that would readily distinguish, just on the basis of the promoters that drive expression of the product, different types of decision trees. And, so, that would give a different type of balance entirely.

The second point is related, in a sense, to the first, or as a consequence of that; and that is that most of the promoters that are currently used to drive expression of recombinant proteins

inplants are strain viral enhancers, which drive very high expression often of the recombinant protein and do so typically in a somewhat constitutive manner throughout the plant. Again, more targeted approaches I think would be prudent for the future, would alleviate concerns by consumers, as well as scientists to review the expression patterns of these, and would at least take up, perhaps in many cases, the problem out of the food chain, which is the strict focus of this committee.

So, I would also encourage us to think proactively in terms of what types of technologies are in hand now that could be recommended to developers in the future to help them in creating a better, safer product that would conceivably raise less problems and less red flags. Thank you.

CHAIRMAN BUSTA: Dr. Fedoroff.

DR. FEDOROFF: Last, but not least. I am an expert in genome rearrangements, having worked on transposed lines. T-DNAs don't move. And some of the problems we're taking, we're suggesting are

the purview of FDA are really the problem of the developer. It's a problem for the developer if there are genomic rearrangements and recombination events, which, by the way, are quite rare in plants; that is, ectopic recombination events between homologous sequences located on different chromosomes. If that messes up the chromosomes, the developer will see that and throw that away long before that plants gets to the point at which it's ready to come here.

So, I think that already the amount of information that is being required about the inserts is probably more than we really need given that they have so little predictive value. I can tell you--I probably couldn't quantify it. I could quantify it, more or less, from people who have put together insertional databases in arabidopsis. But the bottom line there is that rather few of the insertions affect the way the plant develops and reproduces, which is what plants do.

So, the probability that there will be a huge accumulation of insertions that happen to

affect just that part that we use as food is very low. I think there are other places we should be looking. I think at the biochemical level, we know what the problem compounds, whether they're proteins, alkaloids, lipids, or secondary metabolites, we know a lot about that for different kinds of foods. That's where we should be looking. Sometimes it's at the--at the level of proteins, but it certainly isn't always at the level of proteins. So, the tests should be relevant to what's known about the history of toxic compounds or allergenic compounds in the food.

Okay. Something that Dr. Gonsalves I thought would bring up but didn't bring up is asking for information that's relevant to the particular modification. The RNA-based inactivation processes minimize the effect of the protein. So, if you have a viral coat protein gene, basically what you're doing is shutting down the translation of that gene, because you're destroying the RNA. We didn't know that 10 years ago. We know that today. You will see on the web

lots of commentaries about how terribly allergenic the--what is it--the papaya ring spot viral coat protein is. But the fact of the matter is that the mechanism by which the coat protein gene confers immunity to the papaya minimizes the production of that coat protein gene. And, by the way, the evidence that it's allergenic is so thin. I've just been reviewing some of those papers, and it's simply approximally [sic] eight amino acid homology between it and another coat protein. And someone was commenting, I think it was James at lunch, that we now have a very large database of allergenic proteins, and the probability that two

homologous--two proteins will cross react that are not homologous, with the same IGE, is pretty small at this point.

So, those are the kinds of data we should be incorporating into our assessments. But I think at the level of DNA characterization, there maybe isn't much to be gained. There's not any indication that I've heard from anybody that you can even point to one instance where the particular

site of insertion passed through all of the tests that the developers put the plant through, whether it's back crossing or simply growing out, and then suddenly a problem was revealed when one looked at the food.

The probability that you will uncover food issues at the DNA level is much smaller, it seems to me, than looking at either specific proteins or the biochemical composition of the food.

So, I would strongly urge us not to extent the amount of molecular evidence that we're asking for at the moment, but to ask that the kinds of data that the FDA requires be relevant to the kinds of compounds that have proven problematic in those foods.

And my final comment-I'm probably way over my time--

CHAIRMAN BUSTA: That's all right. But you're doing a nice job of summarizing. Keep going.

DR. FEDOROFF: Okay, the one thing that I wanted to add to Dr. Astwood's summary of the

Pustai problem is that it's really important for people who have an animal background and a plant background to come together on these issues. I think that's what went wrong in that case, because what he saw was nutritionally different potatoes, never mind that rats didn't like them, but. And he concluded that it was something about the recombinant DNA that was different because his control was potato plus lectin, okay. And that was the wrong conclusion--what--that--but--probably. I mean, you wouldn't know that unless you went back and did the experiments, which are hard experiments. And that is, to ask why kind of variability comes out of tissue culture. And that's something that many animal biologists don't know. But in plants, most of the processes that we use currently, not all of them, but most of them for introducing genes, involve a de-differentiation of the plant tissue into something that doesn't even look like a plant, and then regeneration of that. Now, that changes methylation patterns. That changes things, both transiently and stably.

All the different kinds of stable genetic mutations--single-base changes, insertions, deletions, translocations--have all surfaced in those regenerated cells. That's one of the first things that happens in a breeding program. You throw them--and that's actually been used. It's called somaclonal variation. It has been used to evoke more variability, but then you have to sort through the garbage and throw away the garbage.

So, the point is that you can't simply transform a gene into a tissue culture cell, analyze the protein composition, because that isn't necessarily what you're going to get out of it.

One, and the plant that you get out to begin with, if you begin to backcross it, you will get rid of the other junk you've accumulated, which we've actually never worried about when we used radiation to induce variation. That is very much more less specific, and introduces a lot more damage. And, yet, the whole breeding process, and its multiple generation--I don't think it's

been--it's probably reduced to--how

many generations do you typically do now?

DR. ASTWOOD: I'm sorry.

DR. FEDOROFF: In your breeding programs?

DR. ASTWOOD: Before going to the marketplace?

DR. FEDOROFF: Correct.

DR. ASTWOOD: It could be as many as 10.

DR. FEDOROFF: Okay. So, it's not really much less than was classically used.

There's--there's ten generations of backcrossing, which gets rid of other junk, okay, that might have happened.

And, finally, rearrangements, is the basic stuff of plant change. Even closely related varieties differ from each other by insertions, deletions, rearrangements. That's a constant process. So, the notion that we have a genome that can't be disturbed is an illusion. It's an illusion in people. It's an illusion in plants. The phenotype is what's important, whether you're looking at the biochemistry or the way the plant grows and reproduces.

End of lecture.

CHAIRMAN BUSTA: As the ultimate novice here, compared to the rest of this panel, in

my--not sufficiently alert or listening to the singing next door, I'm not sure what it was. But there seems like there's a great deal of consistency here, a great deal of agreement. Did I miss something?

Gee, it's going to be sort of dull.

[Laughter.]

DR. FEDOROFF: You can make a difference.

CHAIRMAN BUSTA: As I was listening through this, I'm hearing that as we looked at our charge, and as some of the things have, saying scientifically what will help FDA do their job, what I'm hearing is that they shouldn't put a lot more effort into the genetic assessments sequencing, et cetera, than currently going on. And what I heard was that no changes in the genetics may not really indicate--may not identify a possible food safety issue if it might exist, but that an awful lot of food safety issues that might occur in the biotechnology never get very far

because they disappear in the development; that the development takes care--rules out essentially all of that.

DR. FEDOROFF: It depends on the plant.

CHAIRMAN BUSTA: Depends on the plant.

DR. FEDOROFF: Some are crossed, and some are not.

CHAIRMAN BUSTA: Okay. The changes in proteins, and I heard proteomics a lot, and these new techniques get awfully exciting. Now, I heard that we're on the edge of that yet, and we still got a lot to learn. We may not--don't want to get the chip upside down.

DR. BUCHANAN: That hasn't happened.

CHAIRMAN BUSTA: That hasn't happened yet. But did I hear that if you detected a change in the proteins in some way, it wouldn't necessarily be allergenic; it wouldn't necessarily be a food safety issue. But if there weren't any changes in the protein array, if you would, that we're fairly sure that there isn't a potential food safety issue.

DR. KAPUSCINSKI: Oh, I think we need to talk about that a little more.

DR. ARIAS: Are we talking about the complement of plant proteins encoated by the genome, in addition to changes in the recombinant protein or apart from?

DR. KAPUSCINSKI: I think in addition to. I mean, and that's part of why I was stressing pleiotropic effects. You know, or things such as the promoter ending up driving the expression of another gene that's somewhere downstream. And, you know, rather than trying to deal with that by figuring out how to look at it at the DNA level, which is fraught with problems. My understanding is that through the kinds of techniques you're using, if it really did drive the expression of a protein that either is not expressed otherwise or at much lower levels that your methodologies have the potential to reveal that.

CHAIRMAN BUSTA: In sequence that was

Arias--

DR. ARIAS: Arias, again. My concern is

there isn't enough information yet, except in a well few well characterized conditions or states to identify a change in the expression of an endogenous protein with a specific effect. In fact, you know the metabolomics of plants is still in its infancy, even in the model plant arabidopsis has a lot to go before a full characterization occurs. So, perhaps that's something that when we get to that level, it can be done.

But I'm concerned about the issue of

post-translation changes in the plant of the recombinant protein as a more significant issue; and that this is something I think that can and should be perhaps be readily addressed by developers because of the potential for changes in plant due to post-translation modifications that might conceivably alter the biochemical properties of the recombinant protein. So, in most cases, when recombinant proteins are expressed and tested, they're done so in heterologous systems, like yeast or in bacteria, frequently.

And, in that sense, their bioactivity

might differ significantly from that expressed in the plant because of intrinsically different metabolic reactions that lead to post-translational modifications in different plant cells and tissues.

So, I think a complete characterization at the biochemical level of the recombinant protein in plant is probably just as important as looking at changes in gene insertion and mutagenesis events, because those changes in post-translational effects could, at least, act as a red flag for thinking about what the bioactivity properties might be. And this is certainly not beyond the realm of current technology, such as mass spectroscopy, where proteins that have been purified can be completely sequenced in many cases, and their

post-translational modifications can be analyzed.

So, I think that it would behoove us to think specifically as well about those effects as they relate to bioactivity.

CHAIRMAN BUSTA: Dr. Astwood.

DR. ASTWOOD: This is Jim Astwood. I wanted to reinforce and agree completely with what

Dr. Arias just said about the importance of full characterization of the protein as it's expressed and manifested in the plant. Although it's basically out of the scope from the questions we ask here, conventionally what developers do is, in fact, a complete characterization of the protein as produced in plants. They typically purify from the plants, sequence it, look for post-translational modifications, the good glycosylation, if present, look at bioactivity, if it's an insecticidal protein, look at enzyme activity and enzyme kinetics, if it's an enzyme, and carefully compare that to the e-coli material, which is typically used in safety studies. So there's usually a significant investment in the characterization of the actual transgenic protein. And so I think that's an extremely important point. But I also think that is what we are doing.

DR. ARIAS: I'm sorry just one additional comment. I believe that is true, but I think that the realm of analysis still needs to go that extra step. I know glycosylation is looked. I don't

believe phosphorylation, for example, is typically done so, although I may be wrong on this. And there are other post-translational modification states, as well as proteins, that may be equally relevant. Given the high analytical capabilities of mass spec these days, it seems like this would be a technology that is already available for applications for these types of things. And it's also highly sensitive.

CHAIRMAN BUSTA: One, two, three.

DR. BUCHANAN: Bob Buchanan. Yes, I just wanted to add one minor point, and that is I think proteomics has the capability to pinpoint unintended effects. For example, if we see new proteins, we don't know what all of them are, we don't know what all of the secondary products, or natural product pathways are, which might form toxicants. We know many of them, and, with time, we'll know most of them.

But I think we can identify potential problems. I think it's very unlikely that they'll arise. But the capability to identify them is

present or soon will be.

CHAIRMAN BUSTA: Dr. Gurian?

DR. GURIAN-SHERMAN: Doug Gurian-Sherman. Yeah, I wanted to reiterate Dr. Arias' point about the importance of post-translational modifications. I think his laying out the testing kind of protocols and using bacterial surrogates is very accurate. And, but I'm not, again, you know, I hate to keep harping on this point, but I have looked at a number of the actual studies. And there's a lot of variation in the extent of analysis of post-translational effects. And that was one of the issues that came up, for instance, with StarLink, which was a major protein, and the committee of allergists and immunologists and others that looked at StarLink concluded that they were uncomfortable about Aventis is the company that produced that crop. Their analysis of the glycosylation state or the cryo 9 C protein. So, there's an example where, you know, there as a group of scientific experts that have a lot of experience with that, and glycosylation can be

importance for a number of reasons, especially immunological reasons. It has, you know, there is some association between glycosylation and allergenicity, among other things.

So, again, you know, I think all the specific points about the issues that we're dealing with and discussing are important. But, again, you know, I have to come back to the point about how studies are actually done; what's actually looked at in practice; or what's at least what's submitted to the agencies--maybe, you know, companies are doing more than what they submit--and the bearing that that has on the rigor of the results.

So, it's a little bit different issue. But I--you know, I think it's still very important that that be considered more carefully.

CHAIRMAN BUSTA: Dr. Benedict?

DR. BENEDICT: Steve Benedict. Just a couple of loose comments. One is that with a properly done two-D gel, and other associated things, you can pick up post-translational modifications pretty readily. And I come back to

this. It's not that I would advocate that the

two-D gel could tell us everything that's going on. They could pick up the difference between the parent plant and the progeny plant. And that's all I was sort of suggesting, which was to echo what Dr. Buchanan said; and, that is, if you find the difference, then you go to mass spec, you find out what that protein is, and you ask yourself, is this likely to be a problem. It will be a problem either because it's a toxin or an allergen, or it will be a problem because it's misdirecting some metabolic pathway that you can now learn about because you know what the protein is. And so that's why I'm pounding on this a little harder.

We can tell phosphorylation differences with a Western blot and the right antibody. That's trivial almost.

The other thing for Dr. Fedoroff is, when I mention plant tissue culture, the sort of statement I didn't make clearly was that my argument was that if you have a tissue culture, and you put a protein in there, if there is a change in

the proteome there--admittedly, you can't predict anything about the plant, you can't predict anything else, but you can ask the question whether that protein that you've introduced can alter gene expression levels, can alter the pathway of some expression pathway. And if the answer is no, then you have good information that that protein itself introduced into the plant that it's going to be in is not by some way likely to soak up some metabolic pathway protein and cause a shift in gene expression. That was sort of the question that I was trying to address.

CHAIRMAN BUSTA: Dr. Fedoroff.

DR. FEDOROFF: That's actually a harder question to answer than you might think because it's not impossible, but, basically, it's a given that different tissue culture lines have different modifications. It'll shut down genes methylation at exchange. Lots of junk happens, okay? So, in order to ask whether it's that protein that you introduced that's doing whatever it is that you picked up with your proteomics, you really have to

use one of these site-specific recombination systems so you can excise the particular gene and say, okay, is it still there.

So, that's doable, but, again, the criterion of the plant growing it back, and then asking the questions after you've cleaned up your background, which is basically what breeders have always done, is more relevant to the food safety issues than asking whether tissue culture messes up stuff.

DR. BENEDICT: But I think I'm still not being clear. I really didn't care, in my question, what happened to the plant at that point. The question is, is the introduced protein, all by itself, regardless of future phenotypic changes, is that protein capable of doing something to the metabolic milieu that causes a rearrangement of gene expression.

DR. FEDOROFF: And what I'm trying to tell you is that it's harder to tell, to distinguish that from random changes that happen as a result of tissue cultures, than you might think, because the

tissue culture itself is evolving. So, if you culture cells for long enough, you can't even regenerate plants from them.

DR. BENEDICT: Really, we're only talking about a few hours, you know.

DR. FEDOROFF: Well, no.

DR. BENEDICT: A day or so.

DR. FEDOROFF: We're not talking about a few hours. This is getting very technical, but we're not talking about a few hours.

CHAIRMAN BUSTA: I'm going to take two more, and then we're going to take a

biological--three more--and then we'll take a biological break.

DR. ASTWOOD: Jim Astwood. I wanted to pick up on the proteomics conversation, because it's a fascinating one, and share with you a bit of my own experience. I have a Ph.D. student who actually is at a proteomics conference tomorrow presenting data from arabidopsis, Twelve Different Varieties of Arabidopsis, with a View to Can You Detect Differences from One Variety to the Other. And, of course, you can. Of the 850 proteins

identified, over 300 vary from one accession really of arabidopsis to another. And, so you start getting a sense of the variability in the proteome. And that really gets to Dr. Maryanski's comment early. When we do compositional analyses, it's actually insufficient in a risk assessment context to do a head to head comparison between the

so-called isoline and the actual transgenic product precisely for the reasons that Dr. Fedoroff mentioned, which is that, because of somaclonal variation and the actual underlying genetics of most crops, you can very rarely find a true genetically thing to do the comparisons from. So, you actually have to do a comparison to a population, and so you have to understand what's the range in the population before you can actually do these kinds of experiments. And so, at a technical level, I think we're at the frontier, and that there's some opportunity that proteomics could be of value. I'm not sure when that will be, but, as we talk about applying proteomics, it seems to that before we apply such technology, in addition

to all the ground work we got to lay in terms of understanding it, we also need to ask ourselves, what incremental public health protection are we actually achieving by the application of that technology or a different technology or a new kind of toxicology study. Usually, in a regulatory context when you apply a new methodology, it's because you're attempting to address a risk that your current assessment does not address. So, as we contemplate those technologies, I think we need to understand what risks we are not currently addressing by the range of studies that we're already using. Until we know what those are, I think it's premature to really talk seriously about applying proteomics or any other kind of technology until we understand that safety context. So, it is two things. There's when is the technology really ready to use, and then what is the public health benefit of using the technology. And for me, proteomics isn't there yet, but I think it's an attractive area to work, as Bob mentioned and others. And, in fact, we're working on it, and I

know others are.

So, it's a tough area.

CHAIRMAN BUSTA: Dr. Gurian-Sherman.

DR. GURIAN-SHERMAN: Yeah, I think getting back to--those are some good points that Jim makes that, you know, may be something that requires further development to be able--and I think there's been--this point has been made a number of times to interpret changes that are very likely to be seen. I think addressing both, you know, Nina--Nina's point and Dr. Benedict's point that it seems to me that what we're concerned about is the risk of the plant, you know, both the genetic material, its products, and the overall--the whole plant.

And there would probably be--certainly be more appropriate to look at it after whatever backcrossing has been done for the last, you know, maybe whatever is typical--one centimorgan of DNA, flanking DNA, that will remain after backcrossing. But, be that as it may, it's still, even at that point, whatever risk still remains from both the insert, any linked mutations that might still be

there is really, you know, ultimately the point. And can we get at that, and how much marginal risk is there. And that, I think, goes to Jim's point about what is the purpose or what would be the value of that. And I think, again, it goes back to the whole issue of pleiotropic effects and how much we do and don't know or what is predictable and not predictable about the safety of the plant and changes that can occur in the plant that may have deleterious effects. I think, again, you know, echoing what Ann said earlier, that's where there's a certain amount of lack of understanding. You know, without--I'm reluctant to even say this because I don't want to sound like I'm looking for that infinite risk, which is really not the point I want to make. But, certainly, if we can improve the safety of the crop, looking at a genetically engineered crop, by new knowledge, and I go back to what Jim said earlier, you know, he--we eat carrots all our life and maybe there's some incremental risk from eating carrots. We don't want to be paranoid about eating carrots. But, as we do find

out new issues that can improve the safety of crops, I think, you know, even if we did not know them before, it's relevant to consider them, and there may be, you know, issues that we learn about pleiotropic effects that can have relevance.

And I'll just bring up two, and there punitive cases, so, you know, again, I'm bringing this up in the context of, you know, not real well established science. But there have been recent examples with corn, for instance, where some studies have been done with rodents, and I think it was ovarian, human ovarian cells, that discovered a new compound; and I don't remember the chemical compound. But it seems to have some estrogenic effect. Well, that wasn't known five years ago. How important is it? I don't know. It may be very irrelevant.

There have been some recent compounds in wheat proteins that have been associated with the development of Type-II diabetes.

The point I'm getting at is there's still a lot we don't know, and that by, you know,

improving our analysis of pleiotropic effects, we may be able to improve the safety of these crops, because there are compounds that may be expressed at very low levels now that we don't know much about that could--that we shouldn't. I'm not suggesting that we should be paranoid about those or not--or overly concerned about those. But those are the kinds of things that proteomic analysis, you know, can ultimately get it. So, what I'm trying to do is answer Jim's question. I don't think we're quite there yet. But there are, you know, questions that this technology could be relevant for. And, so, you know, I think it is something that needs to go forward and continue to be developed.

CHAIRMAN BUSTA: Dr. Buchanan. You can conclude this portion of the--

DR. BUCHANAN: Oh, that's a rare opportunity. Yes, I just wanted to add one point to what Jim, and when I described what we were doing, I said we had compared the nulsegregant with the transgenic; and I think that's the best

control. That's the--it's gone through all the tissue culture treatment. It's just that the gene at point has segregated out.

And when we have compared our plants in that way, we've looked at seeds, we don't see differences other than the gene of interest, not great differences, anyway. We see very interesting developmental differences because of the protein we're looking at. But if we look at the chromosomal product, it's kind of like the twins. It's hard to tell one from the other. Go for it. Identical twin.

CHAIRMAN BUSTA: A true geneticist. Yes. Dr. Fedoroff.

DR. FEDOROFF: This came up at lunch, and I think it's a point worth making. Corn breeders look at agronomic traits. They never worried about the corn and the edibility of the corn. They never analyzed it, never looked at it. And as far as I know, they didn't do much damage.

UNIDENTIFIED SPEAKER: They didn't eat it.

DR. FEDOROFF: Huh?

UNIDENTIFIED SPEAKER: They didn't eat it.

CHAIRMAN BUSTA: I--we will take a

15-minute break, and then we're going to come back, and we're going to try and come up with some conclusions and recommendations of this discussion. And then we will, after that summary, look at some of the future and the letter that Trudy generated, and look at other recommendations and future agenda items.

[Whereupon, the meeting went back on the record at 3:23 p.m.]

CHAIRMAN BUSTA: Our next challenge of the afternoon is to come up with a summary of what we've discussed and what we feel we want to conclude on our charge.

And I think as I was listening to this apparent consensus to me, that when we're talking about the molecular characterization of bioengineered food plants, and suggestions regarding additional information, I tended to hear that the genetic characterization that's being done right now is sufficient; that it's--for the

purposes of specifically food safety and our charge of scientifically how can we help FDA scientifically do their job, that identifying any more genetic information, any more sequencing, any more information than is already there would not enhance the assessment of food safety. Okay, I've got two hands. So, that--at least we started the conclusion.

DR. GURIAN-SHERMAN: Okay. Doug

Gurian-Sherman. I'm not entirely clear on what we may have consensus about, but I think I dissent from that. And so, I just want to be clear about what I'm dissenting from. I basically would agree with the Codex analysis. I certainly would also agree that, you know, potential unintended effects, such as changes in known toxicants, allergens, anti-nutrients, are probably, you know, more important on a risk assessment basis than the characterization of the gene. But I also have to put that in the context of, again, how difficult it is to acquire the sequence of the transgene and possibly some of the flanking sequences, at least

whether there's an open reading frame that's been interrupted. So--and I think while, you know, there are certainly a lot of limits in terms of how valuable that data is or can be, it can be of value. And, again, I just want to give one potential example. Small changes in--we know--most changes that may occur with the insertion of a gene are going to be deleterious. They are probably more likely than not to significantly reduce the effect, the intended effect, of the gene and gene product. And it's going to be the minority of the cases where there's going to be some change that causes some safety issue. But I think that has to be weighed against, again, the cost and the ease of getting that data and some of the kinds of things that can happen. We do know there's consensus sequences for, at least for M-glycosylation. I don't know if we know enough about O-glycosylation, the sequences involved.

But some of these sequences are fairly small. Allergen epitopes can be, you know, four to five, six amino acids. So, sometimes small changes

can be meaningful. Most of the time, these changes are not. But if we have the data, we can eliminate those issues where they're non-issues. If there seems to be an issue, then there can be further exploration. So, I guess what I'm saying is I agree with the Codex assessment, and I think that has to be, again, put in the context of what FDA is actually getting. And, from what I've seen, they rarely, if ever, get the sequence of the inserted gene in the plant or expression analysis of surrounding flanking regions, open reading frames or a sequence of those flanking regions. They may get it sometimes, but that's not the norm from what I have seen. So I think I need to be more clear on what there may or may not be consensus about.

CHAIRMAN BUSTA: Thank. Kapuscinski.

DR. KAPUSCINSKI: Yeah, I, during the break, looked again at the specific points on

our--in our questions and realized that one of the bullets under number two I wasn't sure I understood clearly. And that was a bullet about information on the organization of the DNA within the inserts.

So, I was assuming that that meant that FDA is currently recommending, when a developer comes to them, that the developer provide information on the sequence, the complete sequence of the inserted construct. And, during the break, I ask Jim, and he seemed to indicate, Jim Maryanski seemed to indicate that they haven't specifically asked that from the past. So, if that's really true, then the one thing I would recommend is that that information would be useful because it seems like, today, it's fairly easy to get that data; and it's at least a first cut. If you think about taking the kind of iterative approach that Dr. Benedict was recommending earlier, and I really endorse that kind of approach, the sequence data would be a first cut at whether there are any predicted amino acid changes could then raise a flag or they might not raise a flag. You do have to then ask the next question: are any of those amino acid changes problematic? But this gives you at least a first cut at knowing whether you need to even ask that.

So, I want to make sure that the people

realize that the comments I made earlier when we went around the table, I was focusing my comments on the issue of should FDA go beyond this list to, you know, go into that whole universe of all this DNA array stuff and genomic stuff; and that's what I was sort of wanting, suggesting we sort of hold back on, and place the effort more at the biochemical and protein level. But that was with my assuming that they're presently encouraging developers to present complete data on the sequence of the inserted constructs.

CHAIRMAN BUSTA: And how does--without me looking, what is Codex say on that?

DR. KAPUSCINSKI: I think Codex recommends doing that.

UNIDENTIFIED SPEAKER: They recommend the sequence.

CHAIRMAN BUSTA: The full sequence?

DR. BENEDICT: I think so.

DR. KAPUSCINSKI: Let's see--

DR. BENEDICT: I shouldn't say that, nodding because I think so, no, because I know so.

DR. KAPUSCINSKI: Well, it says the characterization--item 31 big "A" says the characterization and description of the inserted genetic materials. It's not really clear about--

DR. GURIAN-SHERMAN: "C" says the organization of the inserted genetic--this is Doug Gurian-Sherman. The organization of the inserted genetic material, at the insertion site, including copy number and sequence data--

DR. KAPUSCINSKI: And sequence data.

DR. GURIAN-SHERMAN: And sequence data of the inserted material and the surrounding regions sufficiently to identify, but then it also says or where more appropriate other information, such as analysis of transcripts, expression products, to identify any new substances that may be in the food. And then it goes on to "D." So, yeah, and I would agree. If you have expression data that's more appropriate or can be shown to be more appropriate--but it seems to encourage the sequence data of the insert.

CHAIRMAN BUSTA: But what I also heard was

that there was a lot of agreement with the decision tree approach of looking at a number of other things as well as the genetic composition.

DR. KAPUSCINSKI: Yeah, my sense is that everyone's agreeing.

CHAIRMAN BUSTA: Right. Yeah. Right. I guess I should have started with that, but, to me, that results in not really expanding what FDA is currently doing on the gene work.

DR. KAPUSCINSKI: Well, but I think the devil is in the details. It would depend on what kind of information they're actually requesting or encouraging on the organization of the DNA within the inserts. If it's compatible with what's laid out here in the Codex guidelines, then fine. And if that is one of the steps on the decision tree, that's great. But if it's not compatible with what's in the Codex, then that's--then I'm not sure I would agree.

CHAIRMAN BUSTA: What would you like to say there? I'm just--

DR. SALYERS: Abigail Salyers. Just very,

very brief. What I heard, and what I think is something that is very strongly to be recommended is that the full characterization of the gene product--I mean, I think the sequence data may be nice, but I think it's an illusion to think that's going to give you the information you really want. And, so, that's--Steve, you may have been the person who brought that up or Bob, I've forgotten who. But that the protein product, if there is a protein product of the gene, needs to be very thoroughly characterized. I think that's being done or my assumption is that's being done already. Because that's where you'll see the glycosylation, phosphorylation and other, you know, any abnormalities that might have occurred during the cloning process.

CHAIRMAN BUSTA: Dr. Arias.

DR. ARIAS: To add on to the scope of the discussion relevant to the Codex document, I wanted to just get some clarification on the statement about--under 31 C, where sequence data of the inserted material and of the surrounding region

sufficient to identify any substances expressed as a consequence of the inserted material. There's another issue that is also relevant to the flanking region of the insertion of T-DNA, and I discovered that doing a literature search last week in preparation for this meeting, in which I found several papers that imply that T-DNA insertion itself, in addition to potentially being mutagenic by inserting into intergenic promoter regions or coding sequences can also induce fairly large, in this case up to 40 kilobase region, translocations between chromosomes. And, in the particular example cited here in Plant Physiology 2001, by Tax and Vernon, they show, using arabidopsis daliana as the model, that a significant region of one chromosome, five, was induced to translocate next to the transgene in response to the insertional event on a different chromosome. So, my question is--I have no idea about the frequency of these events. This is a deported case. They were a couple of other papers that have suggested also translocation events may occur in response to T-DNA

insertion. So, I'm wondering whether or not it would be prudent to consider at least having some marker genes that are comparative in terms of their chromosomal distribution in the progenitor lines versus the transgenic lines to at least confirm that, you know, large scale translocation events may have occurred as an untoward consequence of the T-DNA insertion. I have no idea whether this specific case or others might pose a risk for safety or health. But it certainly would be a concern if large regions of chromosomes are relocating to new sites, which may result, as we all know, in a number of cases that have been document in animal and plant literature in dramatic changes in phenotype.

So, I just raise for discussion, and I have no particular suggestion, whether or not it would be advantageous to consider including some marker analysis to show that genes that are identified flanking the T-DNA insertion events are in the appropriate chromosomal location.

CHAIRMAN BUSTA: Dr. Fedoroff.

DR. FEDOROFF: This is a problem with T-DNAs that's been known for a long time. Not only do they insert in multiple places, but they can induce translocations.

But, again, is it a food safety issue? Or are we just disturbed by it, and, therefore, asking for data? If you want to know that the same sequences--that the sequences that were originally together are at each end of the T-DNA, all you have to do is amplify and sequence the two flanking sequences. They should be together in the original DNA and not.

But that's a problem for the developer. Unless you can point to a case where it's a food safety issue, I think it's going to be more of a problem. It's not the case that translocations and various rearrangements don't affect plant growth and development, but if they have a major effect, you're going to throw it out in the field. Or it's going to mess up your backcrosses.

CHAIRMAN BUSTA: Dr. Gonsalves. Gurian and three--one, two, three.

DR. GONSALVES: Yeah, it's Dennis Gonsalves. As I listen to a lot of this, I think we're in danger of becoming too academic. You know, it's easy to say, if you put markers, and you're going to change the non-antibiotic resistance markers, but you know to go through, and I've been through this, so I know, to go through the process of deregulation and AVIS, EPA, and FDA, and all what we have to go through is not a trivial thing. And it takes time.

So, I think that the decisions that we make could theoretically have tremendous impact on whether products even go through. And this is why I really think we have to look at food safety issues, and I heard, I thought I heard, that Dr. Fedoroff said that in many of all the things that have been studied, very few insertion events make a difference. And I see, Dr. Sherman's concern about, you know, he want more systematic information; and I really think we might want to come back and just look at the charges and questions that were asked here and maybe just try

to answer some of these questions; that one was: to what extent does sequencing information contribute to the identification of newly expressed substances? If sequence information is important for the purpose of FDA assessment, what sequence information, should be reviewed, for example, border sequences?

You know, I think if we address some of these things more specifically, we might be able to address some of these charges, but, you know, we can go on and on and talk about all these possibilities that might happen.

DR. ARIAS: I'd like to interject and say this would fall under the domain of information on the organization of the DNA within the inserts. This is an organizational issue at the chromosomal level. And I'm not saying that this has any implications for food safety, but this is a mutagenic event. It will also result in altered junctions, presumably between the translocated chromosome into the new site. Does this engender concern? That's my question.

DR. GONSALVES: Yeah, for food safety.

DR. ARIAS: For food safety, exactly.

DR. GONSALVES: Right. Yeah.

CHAIRMAN BUSTA: Dr. Gurian-Sherman.

DR. GURIAN-SHERMAN: Doug Gurian-Sherman. I guess there's a couple of points, and I definitely with Abigail's point that the protein is ultimately, and the phenotype after that, is more important than the gene. But I think a thorough characterization of the protein would include the protein sequence. And, again, I have to go back to the--a few examples that we do have where there's been characterization of the protein. I think some of these points have been brought up. There have been problems that have been identified with the characterization of the bacterial surrogate compared to the inserted DNA or inserted consequent protein, where there's a protein involved. And I think we just have to be careful about what a thorough characterization of the protein really is, and certainly sequence is a good part, I think, a good part of that. I mean, either, you know,

certainly the protein sequence would ultimately be probably better than the DNA sequence, the mature protein sequence. But you can also obviously translate the DNA sequence, with some caveats.

So, you know, again I think that that's right, but--and, again, I think the devil's in the details about what's actually done, and how it's done, and how effective it is. You know, the point was made several times before: bioassays are sometimes done. But how those bioassays are done are going to determine the sensitivity of the expression data that you're looking at.

So, I think that really, again, whether it's the protein or the DNA, it's not that big a deal and the data can be useful in terms of getting at some of the structure and function relationships of the protein and the plant.

CHAIRMAN BUSTA: Dr. Astwood.

DR. ASTWOOD: This is Jim Astwood. Back to the question of consensus, Anne and others have suggested that we may have consensus that the FDA approach is consistent with Codex. And the one

sort of unresolved issue was whether that also implied that the DNA insert itself was sequenced. It's conceivable that you can have enough information without the sequence, but certainly the general interpretation internationally would likely be that you need to have the DNA sequence of the insert.

From a U.S. perspective, what you would need to judge is whether you need to meet that international standard for all products. I can envision that, for a small academic project, that might not be necessary at all to establish the safety of the product. And I'm not sure that the FDA is explicitly suggesting that you have the DNA insert sequence or not. From the organization that I represent, we always do. For a large corporation, this is routine, because it meets international standards. But you may need a different standard locally to meet the local needs, and, so, I kind of put that on the table.

As for the suggestion that we need to also characterize and fully understand the extent, if

any, of potential chromosomal rearrangements, that is actually a fairly large question. And going back to first principles, does the knowledge that there is a chromosomal rearrangement in and of itself become criteria for the safety, or do we have this battery of other tests that really gets at that question through finitive compositional analysis and other things. If that

rearrangement--it's not that a rearrangement is a priority, a concern. It is rearrangements that cause the phenotype to be different. And we have a lot of ways of measuring the phenotypic properties with respect to food safety. So, I would be reluctant to support a criteria that said you couldn't have rearrangements. In fact, we know most inserts do create rearrangements.

So, it's not clear that the knowledge and the specific knowledge of how large it is, and what genes were deleted actually informs you about the safety of the product.

And that has been a question mark, a question that has been raised and debated

internationally many, many times. And you'll see that's actually not in the Codex document.

What is also suggested in the Codex document is the need to have some sort of flanking sequence DNA, DNA sequence from the flanks, which kind of gets at the question of did you create that novel fusion protein at the junction of the insert. So, I'm not sure if the FDA intended that also to be implicit in what they're proposing, and in their argument that the discussion paper is consistent with Codex, but if it is, I suspect we would have consensus there, too. But if it's not, then we probably need to have more discussion about why.

Sorry, I missed three topics into that. I apologize, Mr. Chairman.

CHAIRMAN BUSTA: Very good. That moves us along.

DR. ARIAS: May I make comment. Jonathan Arias, again. It seems that the differences I noted between the committee members can fall into sort of two categories: the obvious differences in the transgenic coded product itself, such as

splicing and untoward effects, and the unintentional, more speculative nature, of what insertion does, and I can well appreciate how a number of us have different levels of concern for this. I think, at least from my standpoint, it's only to raise these issues as perhaps preliminary devices by which developers might consider ways of reducing their own risk. I would think it would be in the developers best interest, when they make 200--thank you, Nina--independent transgenic lines of the particular recombinant DNA that they would do some preliminary screening to look at things like large-scale translocations, insertional events that obviously result in changes in the plant phenotype, et cetera, et cetera. And so these are more, from my standpoint, perhaps suggestions by which the developers should start thinking about how to minimize the ultimate risk so that when they come down to their application to the FDA, they've already assumed that many of these things are no longer an issue; that we're not looking at a plant whose genetics are not characterized or are really

up in the air. And I do think, though, however, that chromosomal translocation should be looked at rather carefully by developers in their process, because we know that they can result in significant potential for mutagenesis that may or may not result in obvious or perhaps subtle effects that we can't currently characterize, but that, nonetheless, would represent a potential risk.

So, I, again, mention I brought this up only to help advance the issues of risk safety and for no other reason.

CHAIRMAN BUSTA: I, as Chair, I like Dr. Gonsalves' suggestion that to try in the next half hour to wrap this up; that we look at the issues and get some, you know, objective--it's like the objective quiz at the end of the lecture. Get some consensus, if we can, to these issues. And I see number three as a place where the--one of the recommendations on proteomics, ischemics, and come in. Yes, Dr. Fedoroff.

DR. FEDOROFF: I think one of the most important things to add is the one that I added

based on the recent work on the mechanism of inactivation of viral coat proteins; and that is that the advancement that has happened in the last 10 years is the recognition that viral coat protein genes work at a post-transcriptional level. So that kind of data really should be collected, because the implication is immediately that the protein level is very low, and that means you have to worry less about allergenicity of the protein, if you have a mechanism that doesn't increase, but massively decreases the expression level.

CHAIRMAN BUSTA: Did everybody follow what she said?

DR. ARIAS: Yeah, I think that is a very good point, and there's other mitigating parts to the risk analysis that deal with these issues; so, you know, if you have a protein that's being expressed in food and being consumed in the level of the transgene, it is expressed at a similar level or lower. That's a mitigating issue to FDA. So, again, you know, we're not looking at these things in a vacuum. And I think that's a very good

point. If there's no protein being expressed, then that's a whole risk issue that is basically eliminated.

DR. FEDOROFF: Right.

CHAIRMAN BUSTA: Dr. Gonsalves. And I think that's a good point. I didn't bring it up about, you know, the mechanism, but for viral resistance, which there's many cases of that, it essentially come down to post-transcriptional gene silencing. And even now, we can induce resistance with a

200-base pair segment of the coat protein that is

non-translatable. Well, do we have to go through all of these other aspects on the food safety issue. That's a point I was getting at, where, you know, there's certain kinds of things that you do that you probably don't need to look into detail in that aspect. But there's other aspects of alkaloids and other things that you have to look into much more detail because the dangers are much more evident than, let's say, a non-transmittable viral coat protein.

DR. ARIAS: Jonathan Arias. That's a very good point, Dr. Gonsalves, and it reminds me of a comment made by Dr. Fedoroff earlier about looking at the various metabolites, the proteins that are expressed in the plant as really an indicator of potential risk and safety. But one of the limitations that I see with that approach is that we really don't what a lot of the proteins and metabolites are of many agronomic plants. Certainly, some of the cereal grain crops have been well studied, but I think FDA envisions that there'll be many new plants. In using transgenic technologies, I think saw a slide earlier today that suggested that there, in fact, is quite a wide range in the pipeline than these highly studied crops. And their metabolic and biochemical properties, I surmise, are probably much lesser well known.

So, our ability to predict risk is only going to be as good as our knowledge base, which is rather poor in the case, I suspect, of cantaloupe, for instance, or kiwi in terms of what could be

potential secondary metabolites.

So, just one point, and then I'd like to just add my two cents on number three about new technological advances. As I indicated earlier, I'd like to really look at this as a proactive process for recommendations for developers in the future to enhance their safety of their products.

I mentioned earlier several examples. Using tissue-specific promoters is one. Another, which I'd like to also put on the record, is to use homologous recombination, either through a variety of constructs, such as the Creelocks system that's been developed, which might minimize some of the potential risk associated with the random mutagenesis or insertional events that are engendered by use of T-DNA itself. And so, as the technologies are evolving, and have matured, I think the developers should certainly be encouraged to try to integrate these to lower the potential risks of the products that will ultimately be seen by FDA.

CHAIRMAN BUSTA: As long as we're on

number three. Let's deal with it. We've gotten one on the viral coat, and we've gotten this one that I wouldn't try to reiterate, anyway. And we will keep talking--

DR. ARIAS: Targeted gene expression.

CHAIRMAN BUSTA: Yeah, targeted gene expression.

DR. KAPUSCINSKI: And I think that the discussion before the--this is Anne

Kapuscinski--the discussion before the break that especially Dr. Buchanan was saying about proteomics and the fact that it's--we're getting close to being able to use that well would also fit in--

CHAIRMAN BUSTA: Number three.

DR. KAPUSCINSKI: Number three.

CHAIRMAN BUSTA: That the proteomics and the--that the decision tree aspect seems like it's appropriate for number three to rethink that approach of looking at the whole variety of situations that Dr. Benedict talked about. I think you're the one who's had the decision tree; right?

DR. BENEDICT: Yeah. Excuse me. Yeah,

I--what I was hoping to convey is that we don't have to apply everything to every situation, and that if we invoke the decision tree, there will times when we'll need genetic information. There will be times when we--most times we won't. But what we most likely might need would be the proteomics information. And, although I'm sensitive to whatever has said, that we're not ready, I don't care in the sense that we weren't ready to do the genome sequence based on sequencing 300 bases a day, and then money came, and now we can sequence a lot more. And my point would be that if we as a group say the way you need to go is proteomics, find a way to do it, then that gives strength to FDA to go somewhere and say, look, we got all these people telling us we've got to ramp up proteomics. That was all I wanted to say.

CHAIRMAN BUSTA: And number three--

DR. BENEDICT: That's number three.

CHAIRMAN BUSTA: Yeah. And did--do we have agreement with that statement?

UNIDENTIFIED SPEAKER 1: Yes.

UNIDENTIFIED SPEAKER 2: No.

UNIDENTIFIED SPEAKER 3: No.

UNIDENTIFIED SPEAKER 4: No.

DR. FEDOROFF: I agree with that statement.

DR. ARIAS: Eventually, the statement is whether research and development should be done in the field of proteomics to facilitate it in the future; is that correct?

DR. FEDOROFF: Yeah, I agree with that.

DR. SALYERS: We're talking about food safety.

DR. FEDOROFF: Well, for food safety purposes.

DR. ARIAS: Yeah, I don't think it's--

DR. FEDOROFF: I mean, I assume that's what she meant as an additional tool to help identify food safety issues. I mean, it seems to me like that is possibly one of the most useful ways to be able to anticipate ideally pretty early on in the development process pleiotropic effects that you really didn't have any way of predicting

otherwise. And I think the more we can encourage this kind of searching and also proactive approach to design transgenic plants that won't raise safety problems upstream, the better everybody is off. And I think it's appropriate for our subcommittee to be thinking about that, not only thinking about the point at which the developer comes to the FDA. The ideal is by the time they come to the FDA, they've actually resolved those safety issues. And it seems like proteomics, encouraging research and development in proteomics, would be extremely helpful for that. So, I strongly support that, with the understanding that it's to help guide food safety assessment, not just proteomics for the sake of proteomics.

CHAIRMAN BUSTA: Is that all right?

DR. SALYERS: No, it's not all right with me, but he had his hand up.

DR. ASTWOOD: Well--this is Jim Astwood. I don't disagree that there's a high degree of value in exploring the potential utility of proteomics as it could apply to safety assessment,

food safety of genetically modified foods. In fact, I have a research program in that, so I'm very interested in it and support that.

It is a different thing to recommend back to the FDA that they should be looking for proteomics data in dossiers that we submit today. So, as long as we're very clear about what it is that we're recommending, my comfort level goes up.

DR. SALYERS: I agree with that.

DR. BENEDICT: I thought this one was--

DR. ASTWOOD: I understand that, and--

DR. QUALSET: I think you said the future. That was--

DR. ASTWOOD: But I--you might not even pin it to proteomics. It could be--it could turn out that after a couple years of research, we find out that metabolomics is really the business end of pleiotropy, and that's a much more powerful and sensitive to detect things that we're missing today. So, so--

CHAIRMAN BUSTA: Is it all right to expand that?

DR. ASTWOOD: So, you need to be a

little--if you're going to be prospective and create some language-- and I suspect the Chairman is tasked with creating that language--but create language that induces that flexibility and that prospective nature. And all the suggestions that Professor Arias has made are very good suggestions as things for us to evaluate and consider, et cetera.

CHAIRMAN BUSTA: So we're talking about future being prepared.

DR. FEDOROFF: Right. But we're asking what new advances could be used to enhance the safety assessment. And I think the issues is that the connections between changes in protein composition have not been correlated with food safety today. Nor is the variability that's tolerable in perfectly safe foods known. So, as long as there's--it's a little bit, again, looking, you know, using the technique that you have in hand, but I would not support its invocation today. I would support a recommendation that says that

linkages between food safety and protein constitution should be explored with the most contemporary tools. But I would also include other metabolites, not just proteins.

CHAIRMAN BUSTA: Is that--are we restoring the role? Is that all right?

DR. KAPUSCINSKI: This is Anne Kapuscinski. I'm okay with that, and I think really I understood Dr. Benedict, and what I was supporting is encouraging research and development in this area so that we're at a better position maybe two years from now or five years from now to, you know, maybe revisit this and see if there are certain aspects of those technologies, the proteomics and the metabolomics, that, you know, would be particularly useful. But if we don't start encouraging that that work be done as research and development; it can be done in academia; the research and development can be done in industry, but for it to be useful, it's got to end up in the public scientific literature somewhere. And I agree with Nina, research needs

to be done to link that kind of data to food safety data. But we have to get going on that; otherwise, we'll be having the same discussion five years from now.

DR. FEDOROFF: Actually, I think there's more activity in that area than you can give either companies or researchers credit. There's a--

DR. KAPUSCINSKI: Yeah, and I realize that, but I'm trying to respond to as subcommittee member how--what we should say about item three.

CHAIRMAN BUSTA: Are you doing anything in FDA right now in that area?

DR. CEBULA: They don't do research.

[Laughter.]

DR. KAPUSCINSKI: Let's ask doctor, let's ask the--what the--lab guy.

CHAIRMAN BUSTA: He's back.

DR. CEBULA: Within FDA, we do not have an active program in proteomics directed at this.

CHAIRMAN BUSTA: Okay. The answer was

that--it's right now there's not an active program in the areas that we are talking about. So, the

recommendation would not be redundant or what they might be doing. Finally, you are the most patient person--

DR. QUALSET: I don't have that much to say. I think a consensus we could come to is that we have already is the decision tree framework that's in the 1992 thing, and we saw it again this morning. But that leads to the specificity of the recommendations for food safety. For example, if you knock out the allergenicity of peanuts, it would be a very bad thing if that was a repressed gene that got de-repressed, and people were eating it, and they were having reactions. So, what is your protocol? The developer could use, the evaluators could use, to ensure that that is a stable action, that the repression if that what is it or whatever the mechanism is, that that is a stable one that we can confidently put it on the shelf. So, as we go through the series, the

co-proteins and virus resistance in the flavor enhancement, vitamin enhancement, amino acid enhancement, all those--if you have an enhanced

lysine, for example, and that character is lost, nobody gets sick, probably. So there's different levels of risk based on the trait and the gene. And I think that's where I think we should come to some organizational sense of how would FDA proceed. What are the potential types of genes and traits, and then develop the assessment based on this framework for our decision tree.

CHAIRMAN BUSTA: That was Dr. Qualset.

DR. QUALSET: I'm sorry.

CHAIRMAN BUSTA: No, I should have called that at the beginning.

Could that statement be added as something that FDA should be doing, is putting that sequence of assessment together?

DR. QUALSET: Well, it follows to me that it's just a restatement of their decision tree idea. So I just think that maybe put some flesh on it, and illustrate with some ideas of potential value-added traits, and toxicities and that sort of thing.

CHAIRMAN BUSTA: Are there--as long as

we're trying to beat number three, are there other items in number three that should be recapped from our earlier comments?

It doesn't stop us from adding. Let's back up to number one, and if we ask the question, to what extent does sequencing information contribute to the identification of newly expressed substances. I'm going to get beat down on this, but what I heard was when it comes to the food safety issue, and if I could add in there newly expressed substances that influence food safety, I heard that it doesn't do us much good. Is that correct? Is that what I heard?

DR. KAPUSCINSKI: If you stop only at that, it doesn't do much good.

CHAIRMAN BUSTA: Right.

DR. KAPUSCINSKI: This is Anne Kapuscinski, again. But if you're going to follow the iterative decision tree approach that Dr. Benedict talked about and that I understand FDA espouses its using, it only makes sense that this would be a starting point. Now, you don't

necessarily have to require everybody to do this. If some parties want to immediately go to the detailed compositional analysis, let them do that, and that would be another way to get at it. But some people might decide it's more strategic and perhaps even more cost effective to start with sequencing the inserted DNA construct, and if they find some dramatic changes in the sequence due to rearrangement or whatever, that then gives them a hint as to whether they should do some additional looking. If they don't find something, it also tells them maybe that there's a bunch of further testing that they don't need to do.

So, I don't see any reason to discount that, and I think, in some cases, it might very useful. But with the understanding that none of these techniques is very useful standing by itself. That's why you do need a decision tree kind of approach to any kind of risk assessment, and it's also why you want it to be iterative. What you want to try to do is gather information from different sources, but do that in a smart,

strategic way. So that's where the decision tree comes in, because you don't want to have everybody to do all of them. You want to start with a couple leading questions, as Dr. Qualset was saying, then use that to guide you--and okay, I'm going to gather this type of info. Based on what you learned from that, then decide if you should gather another type of info. Then you look at the two types of info, and you decide if you have to revisit something or not.

CHAIRMAN BUSTA: Is that a satisfactory answer to number one's last, the remainder of that question--if sequencing information is important for the purpose of FDA's food safety assessment, what sequence information should be reviewed, and, if so, how does this information contribute to the safety assessment. What you're saying is, if you see something, well, it's a starting point--

DR. KAPUSCINSKI: That's basically what I'm saying.

DR. BENEDICT: Yeah, Steve Benedict. If circumstances warrant it, based on the decision

tree, you may need to know flanking sequences. You may need to know the sequence of the input series of genes. There will be occasions where it won't be necessary, so the answer to this is not a yes or no. It's a when. And the--

DR. ARIAS: Jonathan Arias. It's clear that because of our relative ignorance about events that affect gene expression is fairly profound, we're not going to be able to predict in many cases what the sequence information that we derive from an insert is going to entail in terms of risk and safety assessments. As far as I know, there are no quantitative indices for those. But, again, to err on the side of prudence, it would seem to me that because this is a mutational event, the insertion of T-DNA, that at least knowing where it is would be a good potential for identifying the subsequent probability of any risk. Not asking that question and addressing it is only opening us up to the potential errors of ignorance, when we didn't have to be ignorant. And, so, knowing the identity of the gene and its insertion would at least give us a

level of comfort as to whether any additional information needs to be derived. And, so, I would agree with Anne that it's a preliminary device to determine whether there is the potential for any further risk in a risk assessment tree.

CHAIRMAN BUSTA: So we're really answering number two here; right?

DR. ARIAS: My sense it's number one, because there is no answer to that question, because there isn't enough data as far as I know--

CHAIRMAN BUSTA: For number one.

DR. ARIAS: For number one.

DR. ASTWOOD: Well, except for the parenthetical question, which is--this is Jim Astwood. The parenthetical question was do you need to have the entire sequence of the inserted genetic material, and it's almost a yes/no question. And I daresay it seems to be yes most of the time. And Anne's suggestion that perhaps

or--no I think Dr. Benedict is suggesting there may be times when you might not need it, but it's hard to define when they might be.

DR. GURIAN-SHERMAN: Doug Gurian-Sherman. I would, you know, basically agree with Jim that it should be part of the analysis, unless there's some good reason not to; and there may be in specific cases good reasons not to.

The way it's used, I mean, typically what's done now, my understanding is the sequences of the gene, and usually now before the insertion of them, is used to search databases, allergen databases, toxin databases to look for matches, for homology.

Clearly, if there's a change during the insertion, a small, you know, percent, you know, the chances are fairly small, but that change may show you some homology that you wouldn't see prior to the insertion event, where there's no change in the gene. So, again, you know, I've got to go back to the point that the chances probably in any given case that there's going to be a change, that the change is going to have a health effect are pretty small. But it's pretty easy to do these informatics, and it's pretty easy to do the

sequencing; and why leave ourselves, as Dr. Arias put it, in that state of ignorance, when it's fairly easy to do.

There may be a change, and you may not see any change in, you know, in your database searches or anything else that raises a red flag, and then you say that these changes are--don't apparently have any meaning. But that's how they're generally used now. I mean, that's how the sequences are generally used is to do homology searches.

CHAIRMAN BUSTA: Dr. Fedoroff.

DR. FEDOROFF: I think maybe we could bring this discussion to a close by acknowledging that if we strictly answer the question, the answer is no. But we should do it anyway? Just in case.

DR. KAPUSCINSKI: I don't totally agree with that. This is Anne Kapuscinski. I think the challenge we have in this food safety of recombinant DNA products is that we're primarily going to be dealing with rare events that could have big consequences; and that's the most tricky aspect of the universe of risk assessment. But the

very nature of risk assessment is that it's very hard to have total yes or no answers. So, I'd rather word it sort of in the way that other commentors were wording, which is that, in some cases, this information could be very useful. It's fairly easy now to get that information, so unless there's a good reason for your specific case to argue that it's not useful information, it's probably a good idea to get this information, and then go on from there.

CHAIRMAN BUSTA: Okay. I do want to close up in a couple or few minutes so that we can discuss the comments and the letter with review, and also future agenda items; and guaranteed at five o'clock, we are done, because that was guaranteed to adjourn at five because some of you are--have to get the METRO or something. Some of us are catching planes.

Let's go to number two, and see if we can conclude that one.

Current approaches. This is--now, what I was hearing before was that if we do what Codex

says we should do, everything is copasetic.

DR. KAPUSCINSKI: We should do everything that's what?

CHAIRMAN BUSTA: If we follow the Codex operation, everything is fine. Is that what I heard before?

DR. ARIAS: Yeah. Jonathan Arias, again. Yeah, I agree. I just want to again raise the issue of whether or not under the subheading of information on the organization of the DNA within the inserts, whether any mapping data on potential translocation events should be included as part of that characterization. I would welcome other comments if anyone feels one way or the other about it.

CHAIRMAN BUSTA: Let's make sure--here, again, and presumably Codex is food safety. That's--and you keep saying that. And the--and that this is addressed at food safety.

DR. GURIAN-SHERMAN: Can I venture something on Jonathan's point? You know, I think he makes a very good point. I think it also points

up part of the problem of trying in this forum kind of on the spur of the moment to consider fairly complicated issues, new issues and come to some conclusion about it without time to consider and look at it. And I would suggest that on that issue--I mean, I'm not prepared to say how important it is. I would follow, you know, Dr. Arias' lead on that.

But as a matter of kind of procedure, maybe it's something that we should consider further as an addendum on to this at more length, where we have more time to consider how important it is to the risk assessment, what's the

frequency--you know those kinds of things, which we, I mean, I have not had a chance to think about before this meeting. And, again, I think it gets to the point of having time prior to the meeting to be able to consider these issues thoroughly, which we haven't done.

CHAIRMAN BUSTA: Dr. Benedict.

DR. BENEDICT: Steve Benedict. Jonathan, doesn't number four address your question a bit,

because if you're going to look at potential reading frames, you'll sequence around, and you'll know essentially what the environment of the gene is?

DR. ARIAS: Perhaps. Actually, it depends on how the translocation event occurs, and what distance from the actual T-DNA insertion event it has brought in. I don't know specific examples cited, but the allogenic recombination event that was characterized appeared to imply that a large region of one chromosome was brought into a region adjacent to I think it was a hundred base pairs from the T-DNA left border. So, again, if you sequenced, for instance, both directions, you might be expected to pick that up. That's correct.

CHAIRMAN BUSTA: Dr. Fedoroff.

DR. FEDOROFF: Again, I will tend to resist this. There are lots and lots of rearrangements that distinguish even closely related organisms, be they animals or plants, and that make no difference in the phenotype. So, I think the phenotype has to be the primary focus of

your examination. If it's a plant that is genetically propagated, it's going to be backcrossed. If it doesn't, if it's selfed and the arrangement is highly deleterious, it's not going to survive in the breeding program. If it's somatically propagated, some of our most valuable food plants, are somatic sports that are carefully maintained, and if you propagate them genetically, they fall apart. So, I think it has to be the food safety issue that has to be primary. The genetic rearrangement underlying it is not a good indicator of either the benefit, the upsides or the downsides. It's--the whole evolution of plant genomes is one of scrambling.

DR. ARIAS: I'll have my say on this. Jonathan Arias, again. I agree entirely with Dr. Fedoroff that mutations can be both, you know, beneficial to human agriculture, but we also know examples where they're not. I think one of them was cited earlier about traditional genetics resulting I think it was in higher psoralen content of celery, which went to market, I'd point out,

before it was recalled.

So, we do know that there are examples that are both beneficial and not. Again, here, not having any quantitative indices for risk assessment of any of these parameters, I think it behooves us to err on the side of prudence and conservation.

CHAIRMAN BUSTA: Dr. Benedict.

DR. BENEDICT: Steve Benedict. When, Dr. Fedoroff, when you talk about the backcrossing and stuff weeding out the bad ones, I fully understand that. But is it a circumstance where the insertion merely activates expression of an allergen, and you could learn that by sequencing a little bit. Would your process of genetics and backcrossing point this out?

DR. FEDOROFF: If I were worried about an allergen, I would look for the allergen, not look at the DNA. The insertional activation--okay, so that in arabidopsis, we have special T-DNA constructs that are designed for activation. The general T-DNA constructs don't activate genes. The activation constructs rarely activate genes. But looking for ancillary rearrangements, which is what the conversation is about, is--seems to me to have a much, much lower probability of identifying, for example, a change in psoralen level than looking at psoralen levels. I--what I keep trying to come back to is that--is--we keep shifting to say let's do this because this is what we can do right now. Okay. And there are lots of things that we can do biochemically and looking for specific allergens using serum from allergic patients that are much more relevant to diagnosing the presence of allergens than looking at rearrangements that move whole pieces of chromosomes that, in fact, are bloody difficult to even identify sometimes. So, the question is using the tool that is most likely to identify the problem that you want to identify.

DR. GURIAN-SHERMAN: I think to kind of rephrase--this is Doug Gurian-Sherman--rephrase that issue, and it has been brought up before, is where the uncertainties lie in terms of unintended effects and how much we do and don't know about potential deleterious, you know, genes and their products in the plant. And, so, where we have a targeted means of addressing one of those issues, such as where we have serum banks for known allergens, that does make, you know, often more sense, although not a trivial issue. We don't have those serum banks right now. They should be developed. But the problem is where we don't know and how much effort should be put into those areas where we don't know what the changes are or what their importance are. So, I mean, I would agree with Dr. Fedoroff that where we do know to look for something, it makes more sense to look at the specifics. But that still leaves open the question of, if you have a large change in the genome that might affect many genes, doesn't it make some sense to have some awareness of that.

And I just want to add one thing: you know, it's certainly the agronomic traits that might be affected are--may, you know, screen that out before. But one question that was brought up in at least one of the FDA studies that was never answered was--where agronomic traits were looked at--and FDA asked this themselves in one of the BT cases: can you correlate the agronomic traits with health safety traits? And that was never answered. And, so I mean I think looking at agronomic traits by themselves is not necessarily going to get at some of the changes in the plants that might have health effects. So, you know, again, the problem I'm having with this particular is not having a lot--had a lot of time to consider it, but it certainly seems like something that we should consider further, because I see the potential there for, you know, fairly large rearrangements to affect the number of genes, you know, at once.

CHAIRMAN BUSTA: Dr. Salyers.

DR. SALYERS: Abigail Salyers.

The--something about this that bothers me is--and, I--ordinarily, I would say, you know, more information is better than less information. But, in this case, suppose you did sequencing out at the ends. Now, if you sequence within the gene, and you saw that there had been changes in that gene, then that would motivate you to take a really close look at the gene product. If you see, if you get DNA sequence to go to the trouble to get DNA sequence from the sides, and you see that there might have been a rearrangement, what are you going to do with that information? Does that mean you're going--you're not going to go forward with the product? I don't think so. I just think it's information that if any--that you--there will be time spent to get it. And then there would be a temptation to say, oh, well, there's a rearrangement, and so, you know, we don't want you to go forward with that product. And I just think it's kind of useless information for purposes of food safety.

DR. ARIAS: Jonathan Arias. I'd just like to respond to that.

It's an excellent question, and I think the answer is embedded in the discussion that we were having before; and that, typically, a large number of independent transgenic lines are generated containing the recombinant DNA of interest. And those are then, we know, down to a few that are going to be fully characterized. I sort of envision this as a very early analytical tool for the developers to use, to determine which of those several hundred perhaps--perhaps Dr. Astwood could tell exactly how many lines might be generated. It's a secret. Thousands perhaps. But the point is that it is not a final analytical tool of when the product is developed. This is something that would be done to screen out potential risk very upstream in the process, at fairly low cost, I would surmise.

CHAIRMAN BUSTA: Dr. Astwood.

DR. ASTWOOD: I mean, think that's an attractive concept, too. This is Jim Astwood. How do you codify that in a recommendation back to the FDA, who wants to give us specific guidance in terms of what they require in a dossier? That's the struggle. I mean, you can imagine that that might be useful a lot of times or not. But to acquire it in every single case is a matter of the data requirements for the food safety assessment, I tend to agree with Dr. Salyers, that as a--it's not clear to me how a risk assessor would use that information. That's the bottom line.

CHAIRMAN BUSTA: Doctor.

DR. SALYERS: Abigail Salyers.

I'm beginning to realize that we're talking about two kinds of food safety. There's safety of the food producer from lawsuits, and there's safety of the consumer. And, so, this probably goes in the first--as the first kind of food safety rather than safety of the consumer. And so I think that our charge here--

DR. FEDOROFF: Consumer--

DR. SALYERS: Safety of the consumer. And let the companies worry about--

DR. FEDOROFF: They have to sell it the second, you know--

CHAIRMAN BUSTA: Dr. Gonsalves.

DR. GONSALVES: You know, I'm glad I developed the papaya back when I did.

[Laughter.]

Because, you know, all these discussions about developers, you're acing out these academic people that, you know, everybody say, you got to do this, this, this right, really organized. Well, you know, a lot of the best things that come out came out because you're doing some experiment. And I think the big lack in products in coming out is because it has almost become the purview of big companies. You know, it's easy to talk about this and this, but some labs are just not suited to do that. And if they need to do that, then they probably won't pursue it. And I think if you make it--if you only think about these developers, these big people who got specialized, they're going to do this and this and that, I think you're tending to miss these innovations that really make a difference. And I can tell you the papaya is one of them. We had only 15 transgenic lines, because during those days, that's all we could do to transform. And I would think we're unusual. I think some of the best innovations to really help agriculture comes from these things that are not well planned; and if you make it such that

if--because of lack of foresight, where you didn't follow these protocols; and, therefore, your product is out the window. I really think we may miss something.

CHAIRMAN BUSTA: Dr. Qualset.

DR. QUALSET: Yeah, I just thinking we're talking about two sides here: the development and the evaluation for the consumers. And we've written a helluva of a textbook for the development side of things, about how to--what you need to measure and what kind of information will be to have, and all this. But let's get back to the issue of how do you judge whether a product is safe to--for the farmers to grow and the processors and the consumers to buy it and cook it up. So I think we've got to get back on track with the--what are the things you can measure and ask about in the stability issue, for example. There's stability in the--of the gene. There's stability of performance, and all those. I think that we can put all this nice-to-know stuff on the table, because it will help you understand, as a developer, what has happened, what does your thing look like. But if there's two inserts instead of one, does that affect the phenotype. If it does affect the phenotype, what happens if one is lost? Have you lost a significant amount of the phenotype.

So, those are issues. I think we need to structure this looking from the backside a little bit as to what the consumers might see when this product goes to market, if it does, and why would you not let it go. And I think the issue that Dennis has raised is outstanding. I think that a lot of developments can be done relatively simply. They might need to do a little clean up analysis toward the end, but if the stable--if the product is stable and useful, you can go from there.

CHAIRMAN BUSTA: How would you like to answer number two?

DR. KAPUSCINSKI: I think we've answered it. I mean, the only thing we've really debated was the chromosome translocations. That's what generated the debate.

CHAIRMAN BUSTA: Okay.

DR. KAPUSCINSKI: So--

CHAIRMAN BUSTA: We're happy—

DR. KAPUSCINSKI: You know, people were agreement on the other stuff. And, you know, I've been fairly silent on this particular debate because I don't know much about this, and I guess I feel the same as Dr. Gurian-Sherman: I'd like to have a little more time to think about it.

CHAIRMAN BUSTA: On the last--on the issue.

DR. KAPUSCINSKI: Yeah, and to read up on it. Because, you know, I appreciate all the different points that have been made--

CHAIRMAN BUSTA: Not the remainder of this, but this last issue is what we're talking about?

DR. KAPUSCINSKI: Right.

CHAIRMAN BUSTA: Okay, I'm just trying

to--as you see, I'm trying to close, get a little bit of closure here.

DR. KAPUSCINSKI: Right. I would support that.

CHAIRMAN BUSTA: Dr. Fedoroff.

DR. FEDOROFF: Yeah, Nina Fedoroff. I think that trying to revise one through four is a losing proposition. I think that's in the Codex, and that's pretty much whether people eventually change that and back of it or not is not for us to decide. But I think that I would strongly resist adding great characterization of rearrangements for all the reasons that I discussed before.

DR. ARIAS: Jonathan Arias. I'm sorry. I just think that this falls within the domain of the particular one through four comments as I mentioned earlier. And it's not really a separate issue. It's a definitional one, on information on the organization of the DNA within the inserts.

DR. FEDOROFF: Within the inserts.

DR. ARIAS: That's--that's--well, within the inserts; that's right.

And I think that this extends in terms of the analysis of the insertional site.

DR. GURIAN-SHERMAN: I have a question. Jonathan, would you be content if we consider this further? I mean, do you feel like we need to come to a decision about this in an hour?

DR. ARIAS: No, I don't. No, I just feel that this is something that should be considered at some point. It does need to be done today.

DR. GURIAN-SHERMAN: I mean, I would absolute support that there should be on the--I'm just making a suggestion--

DR. ARIAS: But this must--that this should be on the agenda.

CHAIRMAN BUSTA: To try to tie this up--

DR. KAPUSCINSKI: Yeah, and I would support that.

CHAIRMAN BUSTA: I got--to try and tie this up then, we feel that this characterization that is consistent with the Codex description is something that is going to be done, if I can sum--

DR. FEDOROFF: The train has left the station.

CHAIRMAN BUSTA: Yeah. And what that other data that might be useful to safety assessment that is something that we need some time to think about, like the topic resistance--

DR. ARIAS: Mapping the chromosomal positions.

CHAIRMAN BUSTA: Right. Or maybe some others. So the last part of other data is still an outstanding question that we're probably not in a position to answer at this moment. Does that really--

DR. GURIAN-SHERMAN: Can I--I would just like to make one point about the whole context of this that we really haven't considered, which is on page 50, paragraphs 20 and 21. There's two parts of that that I think, you know, are important and reflect on the specifics, which, you know, one that the design of the experiments, the assessment is done in accordance with sound and scientific concepts and principles. That's kind of vague, but I think, you know, hopefully we know what they're getting at. But then, in the next paragraph, "in light of the best available scientific knowledge," and I think that kind of goes to what Abigail was saying about, you know, for instance, the protein, you know, the best characterization of at least the protein that's inserted needs to be done. And it's the best characterization, and part of that is the sequence of the protein. I mean, that's--so I think that this gives some important context that yeah, there may be some flexibility built into this, but it needs to reflect back on what's the state of the science now that can give us answers that are, again, reflect on risk assessment and safety. So, I think that language is helpful.

DR. ASTWOOD: Yeah. I got a quick question, and it relates to that is in terms of our answer to other data, I'm not sure if it was under two or three, did we chose to include, Mr. Chairman, the concept of a robust characterization of the protein as expressed in the plant? Would that be an add-on? It's very closely related to insert characterization. Anyway, is that something--I think everyone is agreeing that that's what we should we do, but it seems out of scope, but even though it's out of scope, we could chose to have it in scope.

DR. FEDOROFF: Maybe that goes up as three?

CHAIRMAN BUSTA: Yeah, I thought that was in three.

DR. FEDOROFF: Yeah, I think we talked about number three.

DR. ASTWOOD: Yeah.

DR. FEDOROFF: And it seemed like that was something we all did agree upon.

DR. ASTWOOD: I agree. Exactly.

CHAIRMAN BUSTA: Yeah.

DR. ASTWOOD: Okay.

CHAIRMAN BUSTA: I thought that was in number three. I--I--Dr. Fedoroff.

DR. FEDOROFF: Well, I didn't want to--I was going to jump to three, but you think you want to close out two.

CHAIRMAN BUSTA: If you are--

DR. ASTWOOD: Three's down.

DR. FEDOROFF: At the beginning.

CHAIRMAN BUSTA: Okay. We got number two is closed. Number three, we've put down a number of issues that I hope they are recorded, because I didn't write them down.

DR. FEDOROFF: But did somebody? I mean--

DR. KAPUSCINSKI: Yeah, I mean, we--if I remember correctly.

CHAIRMAN BUSTA: Right we had three or four issues that we named off in number three.

DR. FEDOROFF: I think that the major ways that you can expand the decision tree are to increase the characterization. We now know what we didn't know ten years ago that some genes, some from the processes that are used while they protect plants against disease and pests are--work at the level of RNA, destroying the RNA, that should increase the safety assessment--the probability that it's a safe food because and decrease the need for allergenicity assessment. If the process knocks down the amount of protein, you don't have to worry so much about it. Do you see what I'm saying?

DR. GURIAN-SHERMAN: Mm-hmm.

DR. FEDOROFF: In other words, it's a mechanism that destroys the RNA, messenger RNA, so it doesn't get translated, so when you overexpress that coat protein gene, you're not actually overexpressing the coat protein, you are triggering a mechanism that destroys the messenger.

CHAIRMAN BUSTA: And what are we to do with that?

DR. GURIAN-SHERMAN: Well, Mr. Chairman, I have perhaps that may help. If you look at the decision trees in the '92 policy, there are sections in there that directly relate to exposure of the protein in terms of aggregate human consumption. And, so, I believe what Dr. Fedoroff is saying that if you have a gene for which you can show that consumption is actually lower as a result of the expression of that transgene, then that can be taken into account in your risk assessment, and it would be a positive risk assessment.

DR. ASTWOOD: I have one comment.

DR. FEDOROFF: And the flip side. Excuse me, let me just finish. And the flip side of that is if your target, if what you're trying to accomplish is to inactivate the gene, then you have to show, in a sense, in other words, the required data would be that it is a stable--it is genetically stable, that ability to inactivate, particularly in the case of a known allergen, for example. If you're going to sell the food as a allergen free food, then the need to show that as a food safety issue is much, much greater than the need to show stability of a trait if it's an add-on trait. That's a problem for the company, in a sense. But when it's--when the objective is to decrease the production of a known allergen or toxin, then it becomes more important to show that stability, that genetic stability.

DR. GURIAN-SHERMAN: I had one relevant comment, I think, to this, which--I think the concept is a good concept. There are some exceptions where we're in a little bit of trouble, and allergenicity is one of them, because, and this we went through ad nauseam on the StarLink issue, where the amount of protein needed to sensitize someone, nobody can agree on what that is at this point. So, it's hard to set that low limit. I mean, I agree with it in principle, but we can't say that 20 parts per million is okay. You know, 10 parts per million--I mean parts per million is okay. Twenty parts per million is not. So, unfortunately, with allergenicity, I mean, I think you can say that if there's no expression of the protein, we're all right. But with that particular issue, with allergenicity, we're in a little bit of a hole.

CHAIRMAN BUSTA: Right. But it is a consideration in that, and allergenicity is a different session.

DR. ASTWOOD: This is Jim Astwood. You could actually genericize the concept to say that for some categories of products, or for some straits, demonstration of expression stability of the transgene is much more relevant to the risk assessment than for others; and that will be on a case-by-case basis.

CHAIRMAN BUSTA: I see shaking heads.

DR. GURIAN-SHERMAN: I'm sorry, Jim. Could you say that--say it one more time.

CHAIRMAN BUSTA: Dr. Kapuscinski.

DR. KAPUSCINSKI: Anne Kapuscinski. I just want to make sure we remembered what we'd agree on about three earlier, because we actually did--addressing that question first, and then we went to one and two.

CHAIRMAN BUSTA: I think some of these are--

DR. KAPUSCINSKI: So, I want to make sure we haven't erased any of that.

CHAIRMAN BUSTA: No, and it shouldn't be.

DR. KAPUSCINSKI: Okay.

CHAIRMAN BUSTA: And hopefully in the minutes, under number three, when we said we were at number three, and it should be taped--

DR. KAPUSCINSKI: Okay, because to--my recollection was we all agreed that it's a good idea to encourage research and development on proteomics and metabolomics, and we're still doing that; so that--and especially on the correlation of the data from that with food safety--

CHAIRMAN BUSTA: Right. In relation to food safety.

DR. KAPUSCINSKI: Questions. And really encourage that because that would help--

CHAIRMAN BUSTA: Down the road.

DR. KAPUSCINSKI: So, just--as separate from the conversation we're having right now, but I want to make sure that it didn't get eclipsed.

CHAIRMAN BUSTA: Right. It's a

different--it was a different issue.

I am going to move on to the issue of the letter that three of you sent in. There were some items there that we can take now. We only have about I would say 10 to 15 minutes to discuss items that you have in that letter by the whole committee. I'd like to keep comments to a couple minutes per person so everyone has a chance to comment, and include in this discussion of future agenda items. But before we do that, I'd like an idea of just your quick feel of a one-day meeting, which this is a first experience, with a one-day meeting. We've had day and a half or day and

two-thirds meetings, finishing at 3:00 p.m. for a two-day meeting. Do you have a quick opinion on the tolerance to this one-day meeting, or do you like a day and a half, or do you like a couple of days. Any feelings or opinions?

DR. GURIAN-SHERMAN: I think there's a couple qualifiers that make a difference here. You know, I think we accomplished a lot in the end. It's a pretty limited, a fairly limited topic in the whole spectrum of risk assessment. So if you have a very limited topic, it may be enough.

The other piece of that is to have, you know, which is one of the points in the letter, to have adequate time prior to the meeting to, for any of us to consider the issues. You know, do whatever of background reading research we need to do, so when we come to the meeting, you know, we're more prepared. And the third point is the length of an individual meeting versus the frequency of meetings, because, obviously, you know, that's going to influence it. So, you know, a one-day meeting may be sufficient for a small topic where we have adequate--

CHAIRMAN BUSTA: Focus. As long as we

can--and it takes a great deal of effort--

DR. GURIAN-SHERMAN: And adequate time--

CHAIRMAN BUSTA: To stay on the focus.

DR. GURIAN-SHERMAN: Right. And adequate time. And if we are meeting frequently enough to address the issues that the committee needs to address.

CHAIRMAN BUSTA: Anne.

DR. KAPUSCINSKI: You know, I would just add to that sometimes we don't have--I think for me one of the key issues is to have a little more regular communication. And, you know, I'm overworked, so it's not necessarily that I'm looking for more work. But it's easy--if we have a little more frequent communication, even if it just means that there's some sort of update via e-mail to us, once every three months of sort of where things stand, update can be just a couple paragraphs, that just helps to kind of know where we are, and that would also make more effective when we do meet. And it might turn out that in some cases, we can have a teleconference meeting, depending on the subject at hand. We don't necessarily have to always meet in person. So more frequent communication might be able to be done without the extra cost of always bringing us together, face-to-face. I wouldn't want teleconferences to replace face-to-face meetings in totality, but it could be a useful add-on. I've served on several other committees, especially NRC committees, National Research Council Committees, where that's been used very effectively.

CHAIRMAN BUSTA: No, I don't know the mechanics of the official FDA advisory committees. You know, as you see, everything was recorded. Everything is very, very official, and so I'm--and we can look into that. We may be able to do that, but I don't know if there are constraints that other places, like NRC or whatever, would not have.

DR. KAPUSCINSKI: I mean, there could still be speaker phones in a public room, where anyone else could come in and listen, and we could even hear people give public comment to us.

DR. BENEDICT: The experience that I've had from--with this committee was that if it was a meeting of this nature, it must be public, and people have to have a right to come and speak. When we had teleconferences, they were with working groups that were subsets of the food advisory committee. And we could--we had a charge to do significant scientific agreement of a lot of data dealing with particular approval, and then we could have essentially a out-of-public session. But then we had to report it publicly. So there are a lot of really strange, not strange, there some quite well prescribed reasons.

CHAIRMAN BUSTA: What I've heard is number one, the length of the meeting is very dependent upon the topic and the breadth in the amount of time. It's not a matter of whether you need the evening to think through the meeting; it's more the breadth of the topic; is that correct? Is that what I'm hearing?

DR. KAPUSCINSKI: But add on to that, that we would really appreciate getting the briefing materials as far in advance as possible.

CHAIRMAN BUSTA: Yeah. Yeah, I was just trying to dispatch the length of the meeting. We'll get back. Yes, Dr. Qualset.

DR. QUALSET: This is Cal Qualset speaking. Today we've gone through some stuff that nominally will be in a report form. If we were here tomorrow morning for two hours, we could edit that, work through it, and decide if that's what we--that's the product of our meeting. That's the advantage of the overnight that you can get some thinking about it and then consensus when you--if we're asked to do a report. And I would favor a couple hours in the next day. And that, from coming from the West Coast, it's just a little easier to get home on the same day.

CHAIRMAN BUSTA: I think you can get home by midnight on your time. Okay on the time. That's good enough. Now how about the comments in the letter? There were three items there that I highlighted but other comments.

DR. GURIAN-SHERMAN: Well, I think, you know this goes again to something, you know, Abigail had mentioned earlier about the flexibility versus specifics. And so the first issue was on guidance and guidelines.

I think, you know, the rigidity is a downside, but you can build in still flexibility into guidelines and still have some specifics by giving caveats and saying, if, you know, other methods can be justified, then go ahead.

But the reason that, you know, that that specific was in there was because we were starting to consider some specific protocols, or if not protocols at least performance standards, such as should we have the sequence of the insert. And when were considering allergenicity, we were talking about specific at least kind of performance standards. How should those tests, you know, be done for allergenicity. And I think there's two benefits to that if you can build in some flexibility as well. One it gives some of the predictability that Dennis is talking about having a problem with: that if a researcher doesn't know what to expect or if the ground going to be constantly changing, and that's a very

difficult--if you know what you're up against ahead of time, and if it's reasonable, you can take that into account.

But it's the same for the companies as well, I would think; that to have some independently determined either performance standards or specific, you know, testing standards where they can be developed would be useful to everybody to know what, you know, the ground, the foundation and the ground, the playing field is. So that was kind of the first point that I think needs to be considered. And part of that, again, is attached to the second point that, you know, again this goes to a study that, you know, that we did that probably most people here, or a number of the people, may haven't seen. But the concern that I have is that we can set up what we consider to be reasonable parameters to be looked at in risk assessment. But I don't think that it's a good idea to leave it there and not consider if those are being done according to good scientific practice and that the data that the Agency is looking at are good.

And, so, that kind of goes to both one and two; and what I would suggest, and I don't--you know whether we look at my report of not, I don't really care. But I think two things that I found very useful in doing that report that--you know, that I think it could be very useful for the committee is one, well, the main thing is to actually look at what FDA evaluates in some cases and how they made their decisions and how they acted on them, because, again, you can come up with all the kind of general suggestions or not, but if the process for evaluating those is not adequate, then you're going to end with a product that's not great. And I think, you know, there was a fair amount of some discussion between, you know, myself and some of the FDA people, where we just kind of disagreed about the kind of data that they're seeing and what they are seeing, what they're asking for.

So, I'm kind of just throwing that out as a possibility. I have 14 of these studies. They have been cleared through the Freedom of Information Act, so they can be made available.

That's one possibility, and then the third one goes to some of these procedural things that we've talking about: you know, getting information, meeting adequately enough, and I think we've already discussed.

CHAIRMAN BUSTA: Other comments now on these topics by other members of the committee. Anne?

DR. KAPUSCINSKI: Well, just a minor one regarding the third topic about process. You know, I think one of the things that had frustrated me was just not knowing what the outcome was of the last meeting that the committee had; and, you know, I just want to make sure that when I put effort into this, that its' useful so it's constructive. And, you know, just hearing from Bob Lake today that, yes, they're still going to do something about allergenicity. They do plan to come back to us with recommended guidance, but they just haven't gotten to that. Even that helped, because without having had any communication at all about it in the last year, I just didn't know if maybe it had fallen into a Black Hole. Maybe it would something we recommended that they didn't really want to deal with or what. So, that's part of why I was suggesting earlier, even if the staff could once every three months briefly check in with us, and just tell us, this is where we're at. That would, at least, give me the sense that our subcommittee is actually of some value.

CHAIRMAN BUSTA: So major and ongoing communications.

DR. KAPUSCINSKI: And it doesn't have to be onerous. I can brief, but just to let us know what's going on.

CHAIRMAN BUSTA: The format of the

meeting--these meetings generally, I guess all of them that I've been at, have been a presentation sort of bringing the committee or subcommittee up to speed on FEA is doing, the various specialists; it depends on what topic it is. Presentations. Essentially a show-and-tell, and then a discussion. Is the format useful or would you rather just get the write-ups and come in and go right, brute force, into discussion--discussion questions?

DR. ARIAS: I just have a suggestion here. It would be nice to know in sufficient advance the topic of the subcommittee meeting so that perhaps we, as subcommittee members, could provide input to the FDA on potential technical expertise that could be presented at the meeting as part of the process of discussion that would incorporate, you know, sort of an education as well as a point of discussion. The discussions today were very useful, but, for instance, I don't think at the two meetings that I've been involved, we've had a detailed presentation by expertise, for instance, in genetic engineering of plants about, you know, the molecular basis of the process and sort of a summary of the literature to get a sense of the potential risks or not. Most of these have been brought, I would say, relatively ad hoc to the committee or through our experiences.

I think it would be useful, just for input purposes, if the FDA would like to hear about potential topics for discussion and presenters, that--this might be a good opportunity.

CHAIRMAN BUSTA: Other comments on--

DR. FEDOROFF: There's a recent report released in July. Can you get the latest report? The big--

DR. KAPUSCINSKI: Big U.K.--the GM

plants--

DR. FEDOROFF: That's available on the web, and it's a very up to date assessment of the procedures, the data on horizontal transfer, everything--

CHAIRMAN BUSTA: So what you would be recommending there, consistent here, is that we would be alerted to that ahead of the meetings, so you could go into it or, even to the point of bringing in someone to summarize for the people who aren't up to speed.

DR. FEDOROFF: Yeah. People have summarized. There's lots out there that you can use to catch up on.

DR. ARIAS: I think particularly the risk aspects of this--safety risk assessment.

DR. FEDOROFF: It's a British report is very, very comprehensive--

CHAIRMAN BUSTA: Anne.

DR. KAPUSCINSKI: Anne Kapuscinski. I think the generic point is that it's hard for us to feel prepared as best as possible if we get the charge very late in the game. So, and I can appreciate that the staff is really busy, and they got a lot of pressure on them. And I can see why they would be trying to refine the charge at the, you know, right up to the day before. But even if they could send us a draft charge so that we have a sense of what they're thinking about, you know, a couple weeks ahead of time. I mean, ideally, it would be about a month ahead of time. But at least a couple weeks ahead of time. Then, and if they've got suggestions of things that we could read. That--we, obviously, don't them to necessarily to photocopy all the relevant materials, but if they want to recommend five or six things that we look up, great. And plus, if we get that charge ahead of time, each of us is reading different aspects of the literature isn't on top of different things, and we can go and, you know, collate the information that we think will help us refresh our minds and sort of prepare us for the meeting. Kind of get our head around what the key issue is going to be. So, if we can get it earlier, even if it's draft form and subject to change, that, at least, gets us a sense of what the Agency is thinking about, and what it is they would like us to talk about at the meeting.

CHAIRMAN BUSTA: I don't know some of the official mechanics. You know, I'm a academic. We're not very regulated, especially when you're an emeritus--then you're really not regulated. So I'm not sure of the mechanics. Whether there's a problem with sending out a draft that then it's official, even though it's a draft. You know, it can go to the news media, and even though it didn't surface, and it was a draft, it might still

be--then, I don't know those situations.

DR. KAPUSCINSKI: Well, I mean if that is a problem, then at least get the final one to us, you know, at a minimum two weeks beforehand.

CHAIRMAN BUSTA: Last call for any comments.

DR. GONSALVES: My brief comment is that, from a format, your prior question was the format of this meeting today. I thought it was very productive. It was extremely helpful to hear from Bob Lake and what the FDA activities are, priorities are, and, of course, the other presenters I thought laid the context and landscape of the issues for us very nicely. So I'm very appreciative to the FDA for that.

DR. ASTWOOD: Yes, I'd say the updates are very important.

CHAIRMAN BUSTA: Well, I would like to thank all of the committee, subcommittee, and appreciate all your effort, your intensity, your putting up with me as the ultimate, whatever, acting, the acting FBS Chair. I would like to thank the staff who are sitting back there, the FDA group, for all the preparation, and we appreciate all that you've done, and you've heard the comments. We appreciated that presentations. Have a good trip home.

[Whereupon, at 5:05 p.m., the meeting

adjourned.]

POST-ADJOURNMENT DISCUSSION

DR. GURIAN-SHERMAN: I would just I guess like to say in closing, what were trying to get at with this letter, I mean, and entering into the public record was a consideration of the issues we were bringing up.

So, I don't want to belabor it now. I mean, we're running out of time. We're out of time.

But I think, you know, what I would suggest is that a couple of those issues be considered further, and I don't know how we would get at that or whether there's any consensus about that.

But maybe, you know, one way to get at that is if there's further thought or discussion about the points made in the letter that somehow we could consider that as a committee. We'd be able to dismiss it and say, those, as a committee, we don't agree with those points, or if maybe we should pursue them further.

But I didn't want to let it just die here with just making a couple of points, and then forgetting about it forever.

CHAIRMAN BUSTA: Well, I think Bob Lake said, you know, FDA does not have enough.

[End of tape recording.]