ÿWPCD ûÿ2BVP Z¦Courier 10cpiðÿÿ‰?xxx,Úôxþ6X@É“8Ç;X@þþþþþþþÿþÿÿÿþÿÿþÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿHP LaserJet IIIHPLASIII.PRSÛx Œ @ɇÏ,\,ðy(x©X@#|xûÿ2B ÁÁ ÁÁ ÁÁ ÁÁ ÁÁ ÁÁUNITED STATES FOOD AND DRUG ADMINISTRATION ÁÁ ÁÁ ÁÁ ÁÁ ÁÁ ÁÁBIOLOGICAL RESPONSE MODIFIERS ADVISORY COMMITTEE ÁÁ ÁÁ ÁÁ ÁÁ ÁÁ ÁÁBethesda, Maryland ÁÁFriday, 13 July, 2001 ÁÁ ÁÁ ÁÁPARTICIPANTS: ÁÁDR. DANIEL SALOMON ÁÁChair ÁÁ ÁÁMS. GAIL DAPOLITO ÁÁExecutive Secretary ÁÁ ÁÁDR. STEVE BAUER ÁÁCenter for Biologics Evaluation and Research ÁÁ ÁÁDR. ESTUARDO AGUILAR-CORDOVA ÁÁHarvard University Medical School ÁÁ ÁÁDR. BETH HUTCHINS ÁÁCanji ÁÁ ÁÁDR. RICHARD SUBLETT ÁÁIntrogen Therapeutics ÁÁ ÁÁDR. PHYLLIS FLOMENBERG ÁÁThomas Jefferson University Medical School ÁÁ ÁÁDR. ED SAUSVILLE ÁÁNational Cancer Institute ÁÁ ÁÁDR. DAVID GAYLOR ÁÁSciences International ÁÁ ÁÁDR. MAHRENDA RAO ÁÁNational Institute on Aging ÁÁ ÁÁDR. BRUCE BLAZER ÁÁUniversity of Minnesota ÁÁ ÁÁDR. RICHARD CHAMPLIN ÁÁM.D. Anderson Cancer Center ÁÁ ÁÁDR. JOANNE KUTZBERG ÁÁDuke University ÁÁ ÁÁ ÁÁPARTICIPANTS (CONT'D): ÁÁMS. ALISON LAWSON ÁÁGenezyme Corporation ÁÁ ÁÁDR. DAN SALOMON ÁÁScripps Research Institute ÁÁ ÁÁDR. RICHARD MULLIGAN ÁÁHarvard Medical School ÁÁ ÁÁDR. GARY KETNER ÁÁJohns Hopkins University ÁÁ ÁÁDR. MARSHALL HOROWITZ ÁÁAlbert Einstein College of Medicine ÁÁ ÁÁDR. GENE ROSENTHAL ÁÁNational Institute of Health ÁÁ ÁÁDR. JOYCE FREY ÁÁDivision of Cell and Gene Therapy ÁÁ ÁÁDR. PHIL NOGUCHI ÁÁOffice of Therapeutics ÁÁ ÁÁDR. EUGENE ROSENTHAL ÁÁNational Institute of Health ÁÁ ÁÁMS. ALICE WOLFSON ÁÁConsumer Representative ÁÁ ÁÁ ÁÁ* * * * * ÁÁP R O C E E D I N G S ÁÁDR. SALOMON: Good morning everybody. We've fulfilled my criteria for beginning a meeting. Jay Siegel is seated. We're going to have to have a vote at the end of this meeting about what the criteria for the next meeting's going to be because Jay informed me that he wasn't going to be here, so -- ÁÁAnyway, I'd like to welcome all of you to Meeting Number 30 of the Biological Response Modifiers Advisory Committees. So, 30, that means we've had 30 meetings, so when did that start? When did -- ÁÁMR. SIEGEL: Probably ÀÀÀÀÀÀÀÀ. ÁÁDR. SALOMON: And we've had 30 meetings in or 12 years? I've been giving them a hard time for the last several meetings about not having more -- more titles, you know, good formal titles for these meetings. And, apparently, there was a title here. But it was taken off. Maybe we'll get to see it later. ÁÁAnyway, one of my pleasures this morning is -- in addition to welcoming you here to this meeting is to introduce four official new members of the BRMAC. Dr. Katherine High, who I believe is not here yet, but will join us a little bit later; Dr. Mahendra Rao, seated over there; Dr. Bruce Blazer; and Alison Lawton, seated to my left. And, again, the dedication and the amount of time and energy of the members of the BRMAC to the sort of process is something I think all of us involved in it really respect. ÁÁI also see Abbey Meyers down there, which is a pleasure, welcome back. We're never going to let you go, is that what it is? Abbey's been a member of this committee off-and-on for as long as I've been a member and it's a real pleasure to see her back, as well. ÁÁSo, I'd like to go ahead and get started. We have a one-day meeting. The next meeting of the BRMAC will be three days-long, which, I guess, is proof that no good deed goes unpunished. So just when you thought you were safe to come back to Washington, maybe not, but anyway, I think it will end up balancing out. ÁÁThis should be a very interesting meeting on adenovirus and return to my right-hand person, Gail. ÁÁMS. DAPOLITO: I'll read the Conflict of Interest Statement. This announcement is made part of the public record at this meeting of the Biological Response Modifiers Advisory Committee on July 13. Pursuant to the authority granted under the Committee charter, the director of FDA's Center for Biologics Evaluation and Research has appointed Ms. Abbey Meyers, Doctors Estuardo Aguilar-Cordova, David Gaylor, Marshall Horowitz, and Gary Ketner as temporary voting members. To determine if any conflicts of interest existed, the agency reviewed the submitted agenda and all financial interests reported by the meeting participants. ÁÁAs a result of this review, the following disclosures are being made. In accordance with 18 U.S.C. 208, doctors Katherine High, Estuardo Aguilar- Cordova, and Marshall Horowitz, have been granted a general matters waiver, which permits them to participate in the committee discussions. ÁÁDoctors Blazer, Champlin, Kurtzberg, Mulligan, Salomon, Sausville, and Ms. Meyers have associations with firms that could be affected by the committee discussions. However, in accordance with current statutes, it has been determined that none of these associations require the need for a waiver or an exclusion. Ms. Alison Lawton is the non-voting industry representative for the committee. She is employed by Genzyme. Genzyme has associations with various universities, investigators and research foundations that are involved in gene therapy. Ms. Lawton also has financial interests in several firms that could be affected by the committee discussions. ÁÁIn regard to FDA's invited guests, the agency has determined that the services of these guests are essential. The following interests are being made public to allow meeting participants to objectively evaluate any presentation and/or comments made by the guests. Dr. Beth Hutchins is employed by Canji, Inc., and has a financial interest in a firm that could be affected by the discussions. ÁÁDr. Eugene Rosenthal, is employed by the National Institutes of Health, Office of Biotechnology Activities. Dr. Rosenthal is substituting today for Dr. Amy Patterson, who is unable to attend. ÁÁDr. Richard Sublett is employed by Introgen Therapeutics, and has a financial interest in a firm that could be affected by the discussions. ÁÁIn the event that the discussions involve other products or firms not already on the agenda, for which FDA's participants have a financial interest, the participants are aware of the need to exclude themselves from such involvement and their exclusion will be noted for the public record. ÁÁWith respect to all other meeting participants, we ask, in the interest of fairness, that you state your name, affiliation and address any current or financial previous involvement with any firm whose product you wish to comment upon. A copy of the waiver's addressed in this announcement is available by written request under the Freedom of Information Act. Thank you, Dr. Salomon. ÁÁDR. SALOMON: Okay, going on, I'd like to -- I always try and start the meeting with every body going around and just kind of giving a quick blurb on who they are. I think this is particularly important that today, as we have a couple new members and we're also joined at the table by several additional experts, Dr. Ketner, Dr. Aguilar, Dr. Horowitz. And, again, we're very glad that you guys are here with us. And so why don't we start with Abbey, just tell us a little bit about yourself and we'll just go around. ÁÁMS. MEYERS: Yes, I'm President of the National Organization for Rare Disorders, which is known as NORD and we're the orphan-drug folks. We helped to pass the law and we monitor the development of orphan drugs. And I served on the Recombinant DNA Advisory Committee for many years, and the Biological Response Modifiers has me since then. ÁÁDR. SAUSVILLE: My name is Ed Sausville, and I'm the Associate Director for NCI's Developmental Therapeutics Program and we have an -- our major role is in the preclinincal evaluation and production of both drugs and biologics for use in clinical trials sponsored by NCI and by academic investigators with INDs. ÁÁDR. GAYLOR: I'm David Gaylor, I retired from the National Center for Toxicological Research of the FDA last year. I'm now working with a consulting firm, Sciences International. My area is biostatistics and risk assessment. ÁÁDR. RAO: My name is Mahendra Rao, I'm a Section Chief at the National Institute of Aging. My work is predominantly on stem cells, but I'm also interested in ÀÀÀÀÀÀÀÀ. ÁÁMR. BLAZER: My name is Bruce Blazer, in Bone-marrow Transplantation Division at the University of Minnesota and I'm involved in the immunobiology of bone-marrow transplantation and preclinincal and clinical studies. ÁÁDR. CHAMPLIN: Richard Champlin, I'm the Chairman of the Department of Blood and Marrow Transplantation at M.D. Anderson Cancer Center. ÁÁDR. KURTZBERG: Joanne Kurtzberg, I direct the Pediatric ÀÀÀÀÀÀÀÀ program at Duke University and work with umbilical cord blood transplantation. ÁÁMS. LAWTON: Alison Lawton, I'm the industry rep. I am senior Vice President of Regulatory Affairs for Genzyme Corporation and I chair the Cell and Gene Therapy Committee for the PhARMA Association. ÁÁDR. SALOMON: Dan Salomon, I'm in the Department of Molecular and Experimental Medicine at the Scripp's Research Institute, and my expertise and interest covers organ transplantation, gene therapy, tissue engineering and, more recently, angiogenesis. Thank you. ÁÁMS. DAPOLITO: Gail Dapolito, Center for Biologics, Executive Secretary for the Committee and Rosano Harvey, the Committee Management Specialist for the Committee. ÁÁDR. MULLIGAN: I'm Richard Mulligan from Harvard Medical School and we're involved in vector research and also stem cells. ÁÁDR. KETNER: I'm Gary Ketner, from the Department of Molecular Microbiology and Immunology at Johns Hopkins University School of Public Health -- excuse me, Johns Hopkins University, Bloomberg School of Public Health. I'm an adenovirus biologist. ÁÁDR. AGUILAR-CORDOVA: I'm Estuardo Aguilar- Cordova, I'm with the Harvard Gene Therapy Initiative, I'm involved in vector research and clinical trial applications of gene therapy. ÁÁDR. HOROWITZ: I'm Marshall Horowitz, I'm the Chairman of Microbiology and Immunology at the Albert Einstein College of Medicine and a member of their clinical and the pediatric infectious disease group. I work on adenoviruses, I work on their immunoregulatory genes. In a sense, I'm coming home, in a way, I served four years on the FDA Vaccines and Related Products. And at the time, which was about ten years ago, related products started to include gene therapy. It's nice to see the expertise in the room this morning. It was very hard to find it ten years ago, and I look forward to the discussion today. ÁÁDR. ROSENTHAL: I'm Gene Rosenthal, I'm a biotechnology program advisor in the Office Biotechnology Activities at NIH. I'm substituting today for Amy Patterson. One of the committees that we act as the support staff for is Recombinant DNA Advisory Committee. ÁÁDR. BAUER: My name is Steve Bauer, I'm a product reviewer in the Division of Cellular and Gene Therapies. ÁÁDR. FREY: Joyce Frey, Deputy Director of the Division of Cell and Gene Therapy. ÁÁDR. NOGUCHI: Phil Noguchi, Director of Cell and Gene Therapies in the Office of Therapeutics. ÁÁMR. SIEGEL: Jay Siegel, Director of Office of Therapeutics, CIBER (?), FDA. ÁÁDR. SALOMON: Well, please? ÁÁMS. WOLFSON: Alice Wolfson, I apologize for being late, I was driven to the wrong Holiday Inn. I'm the consumer representative. ÁÁDR. SALOMON: Been there, done that. And Katherine High hasn't joined us yet, so, that's okay. ÁÁWell, I want to say it's a pleasure to be with real professionals. All of you knew to push your button to speak, and then to turn it off, I'm impressed. That bodes well for the rest of the morning. Okay, Gail, are we ready to go? Okay, thank you all very much, and I think we're ready to go. ÁÁThe first speaker today is Steve Bauer, our own Steve Bauer, Office of Therapeutics Research and Review and he's going to talk about Adenovirus Vector Titer Measurements and RCA Levels. Steve and Joyce Frey-Concells, who are there were sort of the lead FDA people on putting together today and the questions involved, so they'll be very much involved in the rest of the discussions this morning with Phil and Jay. ÁÁDR. BAUER: It's a pleasure to talk to you folks this morning and I'd like to express my gratitude to all of you for being here and being willing to participate in this deliberation this morning about adenovirus gene therapy products. ÁÁMy purpose is three-fold: I want to talk, first of all, just to update everybody on an initiative from the gene therapy community to develop a wild type adenovirus reference material that we can use to better or improve our ability to measure the amount of adenovirus product, in terms of viral particles and infectious particles that we administer to people. ÁÁAnd the second purpose is to update the Committee and the gene therapy community and members of the public on recent changes in recommendations about some of the measurements of adenovirus products. ÁÁAnd the third is to introduce some of the questions related to those measurements and how they impact the clinical trials. ÁÁTo put that in perspective, I have here a chart that shows current active gene transfer INDs and I wanted to point out a few things about adenovirus. This is our second most common vector class, about 58 current active INDs in this area. The majority of them, shown in green are direct in vivo injection of the vector by a variety of routes. And then there are some, as well, that rely on ex vivo transduction. ÁÁAnd also, there's a variety of patients that are treated with these different vectors, so in order to put in perspective the kinds of patients who would be exposed to RCA and would be administered these products, I put this chart together to show you that the majority of patients are cancer patients. There are some coronary/vascular applications, some genetics and even some normals and then another thing that's relevant to some of the discussions we'll have later is how are these products actually administered, what is the rout of administration. And I've broken those out. So they go from injection directly into the lesion or the tumors. I mentioned ex vivo transduction of cells, but they're also intravascular, intraperitoneal, oral, directly into muscle and down the respiratory tract. ÁÁSo, there's a variety of routes of administration; a variety of kinds of patients. We'll be discussing later and I think it's an important issue: What is the status of the patients in terms of their ability to respond to either infectious virus or replication-defective adenovirus? A total of 58 current INDs, so that give you a perspective on what are the activity in this area. ÁÁSo, the first thing that I'm going to do is tell you about this development of an adenovirus reference material. When it comes to looking at an adenovirus preparation final product, we measure several important parameters of what is actually in that product. And I diagram this here. There are a non-infectious particles and, actually, this is a challenge to measure infectivity and Dr. Aguilar- Cordova will talk about some of the technical challenges that accompany that later. ÁÁAnd one of the important questions is what is our ability to measure infectious versus infectious particles, but I diagram here in yellow that actually the majority of particles as we measure them currently, or most people do, are non-infectious after purification. ÁÁThen there are infectious units, which will infect permissive cells they are replication defective. These are the ones that would actually deliver and produce the therapeutic trans gene. ÁÁAnother component, and this is the one that's of concern is the replication competent adenovirus, and with the most commonly used production methodology nowadays, this is an unavoidable consequence of manufacture. And there will be discussions later of how replication competent adenoviruses arise in the current production methods. But these would be infectious particles and could, potentially, could establish an infection in a recipient. ÁÁAnd then we also measure all these particles together and I'll discuss a little bit more how that's done. But we'll be talking about the technical challenges of these measurements and how we can improve them and the reference material, I think, will help the community at large to better measure these things. ÁÁThe vector particle measurement, which is looking at all the different particles is a physical/chemical measurement. The most common methodology is to lyse the particles, measure the amount of DNA and correlate that genome number with the particle number. We'll hear a little bit more about that later. ÁÁThe infectious activity is a biological assay and, therefore, inherently, more variable. There are a variety of ways to do this, including measuring plaques on a lawn of permissive cells. These are cells that will allow replication of the otherwise replication-defective particle. There are other ways through immunological detection of some of the virus proteins. ÁÁThen, finally, the replication competent adenovirus is a biological assay, where you measure the infection on cells that don't complement the vector, that are not permissive for replication of the vector, the intended product, but will reveal the presence of a recombinant replication competent adenovirus. ÁÁThese are all the measurements that I talked about are very important in terms of safety, efficacy and, also, in terms of an assessment of how well the production process is; how reproducible it is in manufacture of these vector products. ÁÁOf course the safety issue is pretty obvious, if you administer a replication competent virus, what are the consequences you're exposing to that? And we'll have a lot of discussion about that later. The other issue is that exposure to viral proteins can cause some toxicity, itself, so knowing how many viral particles that you're putting into the patient is an important parameter and if there's a sharp elbow in the curve between dose and toxicity, having precise measurements of this is important. ÁÁThe other thing is inefficacy. Of course, if there are a lot of nontransducing particles in a preparation, you're going to minimize your chance of actually having a benefit of expression of the intended gene. And as I mentioned before, this is a good way to assess a production process. ÁÁWell, all of these things, of course, have been recognized very well by the adenovirus gene therapy community for quite sometime. Starting in '93, adenovirus vectors were first used in cystic fibrosis protocols. And even back then some of the technical challenges of measuring reproducibly, reliably, and across studies were recognized and there were recommendations at that time that are reference standard for vector being produced. There was a long interim during which there was quite significant development in this field, but nobody had really stepped forward to produce this kind of standard. ÁÁIn 1999, with the unfortunate death of Jesse Gelsinger, there was a RAC safety symposium and there was a renewed call for standards. And there have been other calls for standards along the way, but I'm happy to say that this has gone forward rapidly since that point. ÁÁThe goals of producing a standard are to have more consistent, safer vectors. As it stands now, particle counts, itself, is subject to inter- assay variability, probably the best assays are 10 percent. Infectious units assay is even more variable, but there's nothing, really, to hang our hats on, in terms of comparing studies that are done in different preclinical settings, different manufacturers and even different clinical studies. So it's anticipated that our ability to relate measurements much more precisely to one another will really improve the safety profile and comparability of this vector product class. ÁÁIt also will help us to develop regulatory policy with more solid data, and we can make our recommendations based on measurements that we have more comparability between different clinical trials across those trials. ÁÁSo there's been formation of a group called the Adenovirus Reference Material Working Group. This is a partnership between government, industry, academia -- it's going forward under the coordination of the Williamsburg Bioprocessing Foundation. And there's a lot of good information on what the activities of this group are at this Web site. There's also some at the FDA website mentioned here. ÁÁBut a lot of the activities that were planned are going forward. This has been very gratifying to those of us who've been involved to see the spirit of cooperativity and volunteerism that has gone into this. A master cell bank, has already been donated; adenovirus wild-type virus stock has been donated; master viral seed stock has been made. The participants who will produce and formulate the bulk virus have been identified and the initial characterization and provisional titer of that preparation -- these participants have been identified; vialing, the people who will serve as repository and distribution, these people have already been identified. And there's an ongoing effort right now to solicit contributions to look at particle infectivity. So, it's very active. I think by the end of this year, we'll actually have a standard to distribute. ÁÁAnd Dr. Beth Hutchins, Estuardo Aguilar, myself are some of the people and Steph Simek, who've been involved in this. And I think it's really a nice contribution to the field. We're very happy and gratified it's going forward. And I think this will have a big impact on our ability to look at these products. ÁÁSo, now, I'm going to change to a discussion of or to a discussion of our recent changes and recommendations. And this is directly related to the information that we gathered from the March 6 letter that went out to all our sponsors of adenovirus gene therapy, as well as other gene therapies. ÁÁI'll discuss our change in recommendation on the particle to infectious unit ratio and the recommendation on RCA limits. ÁÁBut to put that in perspective, I just wanted to introduce this idea. There's a couple different ways to look at risks associated with different products and I borrowed this term from the people in the field of radiation safety. But I think this is the way we basically looked at exposure to replication competent adenovirus. And that is to say that we want it to be as low as reasonably achievable. ÁÁSo the recommendations that we've made were based on looking at the information in the March 6 letter responses and looking at the community's production data. We actually solicited data on all production lots, whether they were used or not, why they were not used in the clinic? Why they failed, in other words, a lot release. And based on that kind of information, we reformulated our recommendations. ÁÁAnd as the field has progressed, as people have gotten better and better at making virus or vector. And as our ability to measure that has gotten better, these recommendations have changed several times over the years. ÁÁThe other thing to discuss is that, as we currently use these recommendations, we apply them to all product lots, regardless of their intended clinical use. And I wanted to have discussion of the appropriateness of that. That is shifting paradigms to a more risk-based paradigm and the question is what kind of information do we have, what do we need if we want to reconsider this kind of approach, where we just say we want it -- the exposure to be as low as reasonably achievable. ÁÁSo these recommendations of the particle-to- infectious unit ratio, the reason for this is, as I showed you earlier, that one diagram of what's in a purified adenovirus particle. It's a mixture of noninfectious and infectious particles and the infectious ones can either be of a desired type that will deliver the therapeutic gene or replication competent. But in the past, we've asked that there be per infectious units, less than 100 total viral particles. So this is, basically, you could express this as saying you want at least 1 percent of your product to be able to transduce or deliver the gene. ÁÁThis previous recommendation, again, had been based on review of production lots in earlier days of gene therapy. The March 6 letter came in, that data was looked at and we decided in the spirit of trying to increase the efficacy or the ability to transduce to change this ratio. And the recommendation now is that per infectious unit there be less than 30 total viral particles. So another way to look at that is to say that you want more than 3.3 percent of your preparation to be infectious. ÁÁNow, recognize that infectivity assays are problematic, and we'll have more discussion of that later. Keep that in mind, but that's -- that was the rationale for this change. ÁÁThe other change was in the limit of how much RCA should be present in clinical lots. And the previous recommendation was that there be less than RCA and 10 to the 9th infectious units. And, as I've said several times, with recognizing that this is a problematic assay compared to the assay for vector particle, we've changed that. And the new recommendation is based on particle number, and we also dose on particle number, so I think this is also in harmony with that. I think there's better precision in this measurement and this way of expressing it. ÁÁSo the current recommendation is now there be less than 1 RCA and 3 times 10 to the 10th vector particles. And this was derived -- multiplying this number by that factor of 30, so that's how we came about getting that number. ÁÁFinally, I'd like to discuss application of this recommendation. As I said, this is currently recommended for all adenovirus vector lots regardless of the clinical use. And in a trial where there doses of 3 times 10 to the 13th vector particles, this recommendation would say that the potential exposure was up to 1,000 RCA. Now, you'll hear data later that there have been substantially higher potential exposures to RCA based on our previous recommendations and, as I said, this field has been going forward, our recommendations have been changing as the technology's changed and improved, but I think that it's important to realize that this would be the current level and in the past, the levels have been higher. ÁÁSo, if we wanted to think about changing from this overall type of risk minimization paradigm of as low as reasonably achievable to a risk-based recommendation, we need to consider what kind of information do we have and there's some literature on this. There's clinical experience with wild-type ad infection and Dr. Phyllis Flomenberg will give us a talk later about that. ÁÁAt the last BRMAC, Dr. Steven Channick gave us a nice talk about that. And I provided for the committee members the transcripts and slides from his presentation last time, as well. ÁÁThere's clinical experience with gene transfer studies and I've asked Beth Hutchins and Dick Sublett to come and tell us about that and to relate to us what the status of the patients was in terms of their immunity and their potential exposure to replication competent adenovirus. ÁÁThere's also some notable adverse events. I already mentioned the one, you know, the death, the unfortunate death of Jesse Gelsinger. And also an early notable adverse event in Ron Crystal's cystic fibrosis trial. And these still remain somewhat mysterious, but there have been questions whether or not they could related to RCA, but I think retrospective analysis of some of the lots using a standard will help us understand that sort of thing. ÁÁBut the other thing is, what information would we like to see. If we would like to say, for instance, that for certain routes of administration, certain patient populations and so on, we can try to make more a risk-based recommendation than this universal application of our current recommendations to all clinical lots. ÁÁAnd some, if you look at the data that came in with the March 6 letter, there are, actually, a wide variety of experience of how many clinical lots were made and how many had to be discarded because RCA. Some manufacturers had as high as 40 percent lot rejection, some had zero. And there's a lot of adenovirus vector biology that is involved in that. But if we have an ability, a better ability to quantify RCA and have some confidence that different clinical applications can have different standards, we could, perhaps, not be so wasteful in our adenovirus lot utilization. ÁÁSo, I wanted to just spend a few minutes talking about what Dr. Steve Channick's talk -- well, since he's not here, found his talk very informative, last time. But I mentioned that some of the things that we can use to look at risks of RCA exposure are what happens with the wild-type adenovirus infection and in different patient populations. And the previous information is focused -- the best studies, I think, are in the arena of bone-marrow transplantation. And in that setting, adenovirus, including types 2 and 5, which are used for gene therapy vectors can be a significant cause of morbidity and mortality in bone-marrow transplant. ÁÁAnd I think one of the take-home lessons is it's important to consider immune status in terms of infection or reactivation of a previous infection, and latency of viruses is an important issue. ÁÁSome of the lessons that have been learned by the field and that were communicated to us by Dr. Channick include, these: The neonatal adenovirus pneumonia is significant. There are sporadic, severe, localized outbreaks. If you look, these aren't really immunodeficient people, but they're immunonative. If you look at SCID population where there's a severe immunodeficiency, they're at high risk for adenovirus. If you look at a syndrome like DiGeorge, where there's immune defect, there have been case reports of fatal hepatic necrosis. ÁÁIn the realm of solid organ transplant, adenovirus infection of the transplanted organ has been recognized as a problem and perhaps been tied, recently, to rejection of cardiac transplants. The source of that could be reactivation in the recipient or the donor. ÁÁThen in AIDS patients, there's been some recognition that adenovirus can cause some problems. ÁÁSo in contrast to that, there has been a lot of clinical experience with adenovirus and we'll hear about that, as well, or with adenovirus gene therapy products with relatively good safety profile. ÁÁSo, with those kinds of considerations in mind, we're going to later discuss whether or not recommendations regarding acceptable levels of RCA should be the same for all different clinical uses. I explained these two different ways of looking at risk, either minimizing it to the greatest extent possibility or look at risk-based. And then, I think, I pointed out that the status of the patient in terms of their immunity is an important consideration. ÁÁThe other thing that's not on here, but I think is important, will be discussed later is route of administration. With the exception that we will discuss ex vivo transduce cells and whether or not RCA measurements should be done on those as well. ÁÁAnd then if we would like to move towards a different recommendation for RCA exposure, what kind of information should be gathering. And I think having a standard will help us look retrospectively at what's been given and correlate clinical outcomes there and we'll also be able to increase out precision and accuracy of future lots. So I don't think we necessarily have all the data that we want to do this now. That's a point of discussion, but I think we could think in terms of, you know, gathering it and what we need to do that. ÁÁAn with that, I'll conclude, thank you for your attention and I'll take any questions. ÁÁDR. SALOMON: Thank you, Steve. I had a couple questions and there may be some questions from the rest of the group. The first one is, are we going to come -- are we going to get some more information a little later in the day about going back to the standard reference set and exactly how that's going to be used? Because, if not, I think that would be something important to just to give a little more detail on. ÁÁDR. BAUER: Well, Dr. Aguilar will talk a little bit more about that in his discussion, so, if you haven't gotten the answer to whatever questions you have at that point, you know, bring it back up. ÁÁDR. SALOMON: A second thing is, are our speakers today going to also talk about -- there are some developments in the vectorology of the adeno, as well as some new strategies for producer cell lines, and other production techniques that could significantly alter, for example, the ration of replication-competent virus and also, hopefully, increase the efficiency of production of infectious viral units. And that also could have, I think, significant bearing on our -- ÁÁDR. BAUER: Absolutely -- ÁÁDR. SALOMON: Discussions. ÁÁDR. BAUER: Absolutely, there are developed and developing alternative ways to produce the vector that are much less prone to this recombination event, that ends up in RCA. And you'll hear from our speakers what that kind of recombination event is and how to avoid it. So, we'll get some information on that later. ÁÁDR. SALOMON: And the last question, I'm sort of playing a role now, as chair, but, you know, clearly the field has been under the shadow of the Gelsinger case, publicly, if not also within the experts involved. The obvious question here, of course, is when that's driven so much of the concern over the safety of the adenovirus in the last two years, until recently, I understand, that for legal reasons there were -- there's been not a lot of information that's been able to have been shared. At this point, are we going to have any information on the replication competent retroviral titer and particle titer, et cetera, of those preparations that were used in that case? ÁÁDR. BAUER: To my knowledge, the status hasn't changed and we're not able to disclose everything -- I don't know they'll -- any of my FDA colleagues have a different answer for that. ÁÁDR. SALOMON: Well, that may be something that other members on the committee may want to comment later but, again, sort of just playing a role his as chair, I would say that the discussions and the recommendations of the committee will all have to officially be taken in the context of if we don't know what exactly happened in that case, and I'm not, you know, saying that right now there aren't good reasons why we're not getting full disclosure on it -- I was just saying, without knowing that, at the same time all of us around the table knowing that, you know, there was a patient death involved, I think we have to make sure that officially, at least my view of this is, that that has to modify anything that we come to conclusion today. Any questions from anyone else, Mahendra? ÁÁDR. RAO: I had a quick question about the monitoring on the patient side. How do you distinguish between the activated virus versus donor virus that you've put in to the patient if the seratype is the same, I mean, from wild-type? ÁÁDR. BAUER: Yes, that's a good question. There are hallmarks, molecular hallmarks of recombinant virus that you can look at. And we have asked people if they have indication of a replication competent adenovirus arising in a patient to look at the molecular structure of that. So you could distinguish that between a reactivation of a latent virus that had already been there and a replication competent adenovirus. ÁÁDR. SAUSVILLE: So, there are two potential, as I see it, components to the issues that we will be discussing. One is a set of problems, such as the generation of replication competent that may occur at a given frequency, but affect a very special patient subset. And then there's the more general types of toxicity that may reflect the range of normal, quote/unquote "host responses" to adenovirus co- proteins, et cetera. ÁÁDoes the agency have a perspective as to which of these components is most important or should dominant the discussion, as it were, because I think they lead us in somewhat different directions. ÁÁDR. BAUER: I think they're hard to separate those out, and I think that both are important, but I hadn't really thought this through before, but thinking about the replication competent adenovirus as an established infection is probably the area that I would like most feedback on at this point. I think it is unclear and there, you know, there is some information on adverse events with -- in preclinical models saying that innate immunity and an immediate or very quick response to the adenovirus co- proteins is an important toxicity. But the relationship between that and the adverse events in the Gelsinger case are probably -- that -- since that was a very quick relatively fast event, that probably reflects the kind of response that would be innate immunity and not so much an infectious event, but I think we do want to discuss infectivity of adenovirus products. Does that help. ÁÁDR. SALOMON: Abbey? I'm just going to let Dr. Mulligan respond, because I think he's going to respond specifically on this, and then go to you. ÁÁDR. MULLIGAN: I would echo the importance of Ed's point. If you look at the Gelsinger case, I would think that it's not going to be because there is replication competent adenovirus that there is difficulties. But it may be there is just more adenovirus than a human has ever seen. And so, I'm struck by the issue on the focus on the replication competents as opposed to other things that track with real recombinant adenovirus particle. So, for instance, if there was proteins that are associated with the virus particle because of the cell type in which you propagated the vector, could that account for differences between the toxicities that you might see. This is exactly the kind of thing I would think in the Gelsinger case could be an issue. That is, you're not only putting in many, many particles but, you know, could it be that there's something in those particles, some contaminating host protein? ÁÁCertainly, I know about retroviruses and in the case of retroviruses, it's very, very clearcut depending on how you grow the retroviruses, you get different host proteins that incorporate in the code. Now it's a very different system than adeno, but I think that's an issue, perhaps, not for, you know, this conversation, but I think it's key. ÁÁWhile we're on the replication competent question, I'm curious whether you went into the very old literature that looks at the helper, the e1-like capacities of different cells, too? One of the questions I was thinking, as you were talking about how you calculate the amount of RCA is, you know, is there a possibility that there are human tissues uniquely, a pocket of pituitary cells or something, that actually have e1-like function and, therefore, you would have replication competent adenovirus vector in vivo, uniquely in those cells? And Marshall, you know, might have something to say about this. But that would be another way to look at this where all of a sudden the issue's not an issue of having a hundred or a thousand replication competent adenos, but 10 to the 6th or 10 to the 7th depending on what tissue you hit. So, I'd like to hear if there is something, you know, do we really know that that can't happen is there any precedent for it? ÁÁDR. KETNER: Yes, there is data. In the teratoma cell line, there are complementing e1a functions that allow the virus to grow. That's old data from -- it's quite old data -- state of the cells and I don't remember exactly which way it went, but one of the two states of differentiation of teratoma, human teratoma, did complement the e1a functions. ÁÁDR. MULLIGAN: That's interesting, from the tumor application, of course, you know, where, you know, if you had some undifferentiated strange tumor type and you were trying to direct infection, would it be possible that you would have replication, so it just calls into question, you know, how you'll actually measure and whether it is relevant to try to see whether there's pockets of tissues where you'll get replication competents. ÁÁDR. SALOMON: Well, I think that's very interesting. Part of what you're saying, though is not that replication competent adenovirus wouldn't be an important thing to measure, but you're actually suggesting it ought to be measured also after infusion of the viral vector, perhaps at several different times. And that's not something that was mentioned. ÁÁDR. MULLIGAN: More just the appreciation that their might be an interest in trying to push towards looking at this question of compensation. And the other issue is, whether or not, unfortunately that complicates potential regulatory guidelines where, maybe the status of the vector becomes much more important. Is it doubly, you know, how many different complementing events in the cell would you have to have to actually give something that would look like replication compromise. ÁÁDR. BAUER: Well, one thing, I can say is that you'll hear some data later about monitoring for virus shedding. Or in the peripheral blood and looking at what is there and how long it persists. So, the model that you're putting forth that you have complementation by e1-like activity in the cells, I think we would probably see that by that kind of monitoring. If you look at a virus that is increasing in titer or that there's a lot of shedding or unusual persistence, and you say that, then you could look at what is the structure of that virus. Is it what you put in that replicating for some reason? To my knowledge, there isn't any evidence that that sort of thing's going on, but I think that is something to keep in mind, and it might be affected by, as Dr. Horowitz said, different kinds of cancer. ÁÁDR. MULLIGAN: The other thing is that in the simplest case, the virus that is essentially replicating in particular tissue, when assayed in culture, would not be replication competent, so -- ÁÁDR. BAUER: Right. ÁÁDR. MULLIGAN: You might have a local effect of having replication competents, but you'd never really see it as such. ÁÁDR. BAUER: Right. ÁÁDR. SALOMON: Abbey, you've been put off a little bit, I'm sorry. ÁÁMS. MEYERS: Patient. If there were something that happened in the Gelsinger case that is important for the purpose of discussion today, and especially would help us in the public health mission that we're supposed to have, I don't understand why we can't talk about it. Now, if there's proprietary information, can we just meet in private without the public to discuss what we know? And if it's not proprietary and there's some other reason, what is the reason? ÁÁDR. NOGUCHI: Well, it is a good question, I think on a practical basis, Dr. Wilson has actually recently published or is in press a couple of more- detailed analyses of the events surrounding Mr. Gelsinger's death. In fairness, none of that really seems to shed any great light or insight as to what to really look for in terms of questions of whether it's RCA, whether it's an immune function, whether any of those are absolutely critical. I think the available data would suggest it's really a set of immune, cascading immune responses to the massive viral load that really is most closely associated with his death and the spiraling events that we saw. But that's really not anything that I think is any different than what was discussed almost a year and a- half ago. ÁÁSo I, your point is well taken, but I don't think we have anything that we feel would add substantially to our discussion here on that. That's part of the issue. The numbers of adverse events that are critical in terms of being much more than expected are relatively rare, and they're in a few different cases. And that's part of the difficulty that we're facing here in that we don't have enough information, even on those cases, in order to be able to compare. ÁÁI think that's one of the reasons that we have been focusing on creating a standard, which should have been available, probably, earlier, but now, by focusing on the standards, by focusing on all the different events that can occur and all the considerations, such as packaging of novel proteins, I think we can do a better job in the future. But right now, we just, unfortunately, don't have a lot more that either privately or publicly we can add on the Jesse Gelsinger case, vis-a-vis leading us to a new approach or a better approach. ÁÁMS. MEYERS: So, for the purpose of today's discussion, then, the Gelsinger case, there's nothing being held back that we don't know about? ÁÁDR. NOGUCHI: That's correct. ÁÁDR. SALOMON: Phil, one of the concerns I have is, if we go forward, let's say, and at the end and get into discussions of sort of a maximal viral particle load that could be administered in a clinical trial -- which is one of the questions, indirectly here -- do we know what Jesse Gelsinger got, in terms of number of particles? I know -- I've -- do we know what -- did we really know what he got? In other words, was there an error? Did he get more than we think? I mean, that's the kind of thing, I guess, that my initial comments were aimed at, that I just see that as coloring our discussions somewhat. Not necessarily implying that anyone's holding anything back, but just not know exactly what happened there is an issue for me. ÁÁDR. NOGUCHI: Well, that point is well taken and I don't have those figures right on hand, but it's clearly of the nature of close to 10 to the 13th particles, that's total particles and the transducing units was -- Steve, do you remember exactly what that way? It's an -- it's an excellent question, we just didn't prepare for the -- ÁÁDR. SALOMON: Right. ÁÁDR. NOGUCHI: I would like to suggest that a lot of these questions are of a more general nature and we'd like to make sure that we get all the prepared talks in here, some of the questions will be answered, many will not be, but that's part of what the more general discussion for the afternoon is about. ÁÁDR. BAUER: I just had one more comment in response to Dr. Sausville's earlier question and thinking about -- seeing the discussion that's going on now and where do we want to go? I would like to focus more on the relationship between, you know, the risk for RCA infectivity and exposure in patients, rather than toxicity. ÁÁI think the story that the viral proteins have toxicity, you know, is pretty well acknowledged and that the acute -- very acute toxicity is related to that dose. There might be a range of responses that we don't appreciate how to, you know, predict that when we first, you know, administer the product. But I think if we could focus -- it's -- it's hard to say these are totally separate, of course, but we can focus more on the infectivity risks. ÁÁDR. SALOMON: Last question, Dr. Kurtzberg. ÁÁDR. KURTZBERG: Is there any data regarding the state of the patient whose receiving the adenoviral vector, in terms of their innate immunity or previous exposure to wild-type adenovirus or whether or not they're shedding adenovirus at the time of the gene therapy? ÁÁDR. BAUER: When patients are enrolled, they're -- most of them, in the past, have been assessed for their immune status with regard to antibodies and to adenovirus and they're also looked at in terms of their -- whether or not they have a current infection, by adenovirus or other viruses, so those patients aren't -- these products aren't administered to those patients. So you don't want to give something to a patient where there's a helper virus already in the patient, for instance, so. ÁÁAnd you'll hear some data later about what was immune status and what was the outcome of the immune system response in certain patient populations, to the vectors. ÁÁDR. SALOMON: Thank you very much, Steve. ÁÁDR. BAUER: Mm-hmm. ÁÁDR. SALOMON: So just to put this into context, then, our first obligation as a committee by the end of this afternoon is going to be to respond to the FDA's staff's specific questions. And those are outlined in your book -- your panel. And that's -- that, at some time we're going to have to just focus on answering those questions because that's what the job of this committee is to do. ÁÁHowever, as is evident from the discussions that we just had, there are other issues and it's not necessarily meant to imply that those issues aren't as important in thinking about adenovirus. So, to the extent that we can separate questions that we specifically have to answer from more general discussions of topics that are not on the docket today, then that would be extra value for the -- for the FDA staff. ÁÁSo, at some point here, I don't find any particular problem with the idea that there may be alternative things to talk about. If you'll forgive me, ever-so-often focusing us down on that. I think we'll have a fairly reasonable day, in terms of time, but that's maybe famous last words, so -- ÁÁMR. SIEGEL: I just want to clarify on that, actually, we would much appreciate, and will benefit, time permitting from any discussions on any topics related to adenoviral safety. And I think what Phil was talking about in terms of general questions was, simply, that we think it might be useful to have questions for clarifications to the speaker-specific points after each speaker. But in terms of some of the general discussions we're getting into, it might be best to have all the speakers first, because I think they'll inform of those discussions. ÁÁDR. SALOMON: Yes, absolutely. Okay. Well, then, it's my pleasure to introduce a new friend. Dr. Estuardo Aguilar-Cordova from the Harvard Gene Therapy Initiative, whose going to talk to us about history and overview of adenoviral vectors. ÁÁDR. AGUILAR-CORDOVA: So, I was asked to just give a general overview of some of the adenoviral characteristics, which I think will be pertinent to the discussions and understanding. I know it will be very repetitious for some of you. ÁÁAlso, then, to give a little bit of the nuts and bolts of what it means when people talk about titers for replication competent detection. And, finally, to give you a little bit of background on the standard that is being developed by a working group with great participation by Beth Hutchins and Steve and Stephanie and others. ÁÁSo, historically, the adenoviruses were identified in the early 50s from an adenoid and they're an etiological agent for some of common cold- like symptoms and other inflammatory responses in various tissues, depending on the seratype, but they have not been identified as an agent of any tumerogenic potential in human tissues. ÁÁThey are a linear double-stranded DNA encapsulated in a protein shell. There's over 100 in the adenoviral group, two different characterized adenoviruses. And the wild-type has been used as a vaccine in military recruits, actually, of seratype 4N7 (?), and I think we'll hear a little bit more of that indicating some of the safety profile that is known in this type of viruses, at least when interaclly (?) delivered. It's important to understand a little bit about the characterization -- physical characterization of the adenoviruses and it's an icoshedral. And this is really quite critical: 13 percent DAN and 87 percent protein and that's how, back in the early sixties it was determined that 1-OD (?) of adenovirus is roughly equivalent to 1.1 times 10 to the 12th viral particles, by comparing the protein characterization and then the OD readings and so, and it has been borne out by other types of analysis in the future that in fact that does hold true and that's a fairly accurate quantitative measurement. The gene structure and organization of adenoviruses are -- has two ITRs and two origins of replication. The transcriptional units are 5 "early" genes; 2 "delayed early" genes and one major late transcript. What's important is that these are what has been primarily lated in the vector constructions and I'll go through that a little bit in a minute. ÁÁThey were able to be manipulated with vectors with two critical characteristics of the virus: That it can package up to 105 percent of it's genome size, so if one can take a little bit off, one can also put a little bit back in. And, also very important, it can be manipulated as a circular form, put back into the cell and it'll go into the linear transcriptional unit. ÁÁSo, the early antiviral vectors were considered replication deficient. One could package up to about 8 kilobases (?) of foreign DNA by doing E3 deletions and E1 deletions. It's relatively easy to produce it in high titer, it can infect a wide variety of tissues and it has high expression in non-replicating tissues. ÁÁSo there's been an evolution of adenoviral vectors and, again, this will be, perhaps, important in our discussion of what the importance may be of RCAs and such. Original vectors, one would stick a gene of interest in the E1 region and that would be replaced, and it could have various deletions in the E3 region. The E1 region, as I mentioned earlier, is a transcriptional activator that would then lead to expression of E3 for the other early transcripts and then, ultimately, to a cascade that would start the transcription of the late transcripts, most of which are structural proteins that would serve towards creating new viriants (?). ÁÁHowever, as mentioned by Dr. Mulligan a little bit ago, there are multiple cellular proteins that have been described that are able to basically pinch hit for the E1 functions, including IL6 is one of the critical ones. And that can, in fact, transactivate some of these early genes, which could then be toxic to the cell. We know that the E2 product and the E4 products, some of the E2 and some of the E4 products, are toxic to the cells and, in fact, there's an E3 product, which would be transactivatable because there are enough papa B and AP1 sites in that promoter, which could be upregulated and in different cell types. And there's a protein that are called the adenovirus death protein, which serves towards lysis. ÁÁSo there are multiple ways in which a vector not being replication competent can still cause significant tissue damage and it can be observed in almost all cell types in vitro, where, if we put sufficient concentrations of replication deficient vectors, we will observe cytopathic effect. ÁÁNow, the first-generation vectors that I've been mentioning here are E1 deleted and they can be either E3 deleted, partially deleted or not deleted at all. So-called second-generation vectors, went farther and deleted either the E2 or the E4 regions and transcomplimented in the production cell. I will show you some data that may or may not have any effect on their toxicity. ÁÁAnd more interesting, we get in the gray area which, what I'm calling here Generation X, I just -- because it doesn't really matter, but these are prototyped by the Onyx virus, which is, in fact an E1A positive cell, E1B minus, and that supposedly is effective only in P53 negative cells because the E1B would, then, bind the P53 under normal circumstances and absorb it out, so letting the cell go through the replication cycle. So, putatively, when there's a P53 minus cell, the E1B would not be necessary in that case and this virus would then replicate more efficiently in those cells and it can, again, be E3 plus or minus. ÁÁThis is very close to a wild-type virus and this is what's been used in all the Onyx trials. There are Generation X.1, which are where they put either the E1A or the E1B with tissue-specific promoters and then, theoretically, this would only replicate in the tissue where the promoter is active. But as we all know, many of these promoters can be leaky, especially when they are outside of the genome context. So, again, very gray zone as to whether these products are, in fact, replication competent adenoviruses and their measurement of what is an RCA in these would be a complicated issue. ÁÁAnd this, perhaps, should have been called Generation Y, because it's completely different. And these are the helper dependent vectors and the helper dependent vectors are those in which, basically all of the genome of the adenovirus has been taken out and only the ITRs and the packaging sequence remain. ÁÁSo, as I was mentioning in the sort of Second-generation vectors, some data that was by O'Neal, {IOn}et al{IOff}., in human gene therapy back in '98, and what we can see here, this is platelet count and it's been a fairly repetitive observation in vivo, in humans that there is at least a transient thrombocytopenia that's observed and in some cases a consumptive thrombocytopenia after adenoviral vector delivery. And we can see here that the deletion of an E4 or an E2 doesn't seem to generate much of an advantage with regard to the causation of thrombocytopenias. ÁÁAlso, when we see here the elevation of ALT, what we can see is this is 1 times 10 to the 11th, 1 times 10 to the 12th, all three of these experiments were done at 3 times 10 to the 12th and this is 1 times 10 to the 13th. Down at the bottom are doses, the effects were not noticeable, either with a first- generation or second-generation, only in the 3 times 10 to the 12th window was there a differential of any statistical significance, but once one gets to the 1 times 10 to the 13th, again, that significance is gone. ÁÁAnd what's also of interest to note is that the difference between almost no effect to a complete, very significant rise is only a ten-fold difference in as far as the quantity that was delivered to the -- to the animals. ÁÁThat's not to say that all of these things make absolutely no difference. Here we have what I call the Generation X.2, or the helper dependent vectors and what we see in the liver enzymes in this case. This is the first-generation vector, one can see as the dose increase, there's an increase in the liver enzymes that then drops down, so it's self- limiting. And that was not seen with this helper dependence, even at the same number of particles. ÁÁSo these are all equivalent particles, you can see the level of -- the duration of expression from this first-generation vectors was very low, whereas the duration of expression from this later helper dependent vectors was much longer lasting, this is 8 weeks in this case. And -- but, also, that's not to say that this is always the case. We have seen, in some of our monkey studies, where occasionally a monkey would develop immunity against the transgene and limit the effect at all. But this does start to address a little bit on the -- at least at these low doses of incoming viral proteins that maybe it's not the low-dose of incoming viral proteins that's causing that acute increase in liver enzymes, as here we have the same number of viral particles and we can see with the first-generation there is some more acute toxicity. It may have to do with the early gene expression that we see from these vectors. ÁÁSo, from this first part of it, we can see that adenoviruses can be converted into efficient transfer vehicles. They're not inherently dangerous as a class of vectors or vehicles. Not all adenoviral vectors have equivalent toxicity profiles, but some of the safety mechanisms that we might have heard or read about may not be that significant. The dose of vector is related to the toxicity of serve, that is -- that's very clear and reproducible, the more we give, the more toxicity we observe. ÁÁAnd I will now show you that standardization of dose specification is necessary and that it hasn't been there to date. ÁÁSo, characterization of viral vectors are two things, generally speaking: Purity and strength. And lack of contamination by advantageous agents, including replication competent virus as part of the purity issue. And the strength is the active concentration for toxicity and efficacy, just very loosely defined. And I'll go through some of the things that we do here. ÁÁPhysical determination, as I've mentioned is the most common method utilized at this point, is absorption measurements and, as I said, from the early 60s it's been known that 1OD 260 is about equivalent to 1.1 times 10 to the 12th viral particles. This has been confirmed by other methodologies, as well. So, that's a very quantitative measurement. ÁÁThe next is the biological determination, that is the infectious units. And we've heard a little bit from Dr. Bauer here about how, right now, we're content to get 3.3 percent activity, that's -- the other way of looking at that is that there's 96.7 percent nonactive in each one of these preps. ÁÁAnd how we determined this -- it has some physical characteristics and I'll go through some of these. But the key issue here is the likelihood of the vector and cell ever meeting in your detection system, so that you could ever see it. And there's some functional characteristics of the system whether the cell that you're testing it on has good receptors for the adenovirus and it's detected. Most people are using either HeLa A549s or 293 cells for detection systems, so I won't go through these. All of these are highly susceptible to adenoviral transduction. But I will go through some of the physical characteristics. ÁÁTypical titer set up has a culture dish, some cells at the bottom, adherent cells, and then one puts a mixture of virus on top of those cells. The collision between the virus and the detectors cells here is mandated by Brownian motion of the virus, a concentration gradient towards the cells or external forces, and I'll go through each of these. ÁÁBut Brownian motion is random and for small molecules within a liquid at this temperature, it's basically negligible. What that means is that, by Brownian motion, those viruses will never move from where they are, and these cells will mostly just see the viruses that are right here next to them, but they'll never see anything that's up here. ÁÁAnd with the -- four or five years ago some experiments to show this and that is by putting different volumes of the same virus dilution onto some plates, we could observe the number of positive -- this is betagal, so we could see little blue cells, and these are averages over many, many wells. And what you can see is that it doesn't matter how much volume we put on top of it, the number of positives of serve was the same, which is what we predicted by just physical characteristics of the system, which, if we translate that to titers, we can see there's a significant difference in calculated titer, so how one sets up the assay will have a significant impact of what titer is calculated. ÁÁWe then did some centrifugation experiments calculating the displacement of the virus, if we were to centrifuge the plate, and that was just calculated as a distance between the pressures, RCFs put in, the time and the sedimentation coefficient. And by doing that, what we can see is that now we've brought down by spinning at 90 minutes for 1,000 RCF, we calculated that the distance traveled by any viriant would be .4 centimeters and we knew the depth of each one of these wells. ÁÁSo, at 50 microliters per well, all of the vectors should have come down and the same is true at 100 microliters per well. But at 200 microliters per well, the depth was greater than the distance traveled. So we didn't expect all of the vectors to come down. And, in fact, that's what we saw. These are the static conditions and the vector particle-to- infectious unit ratio went from 22 to 82 under static conditions. ÁÁWhen we centrifuged the vectors, we observed that there was exactly a two-fold difference between 50 and 100 microliters, as expected and not a two-fold between 100 and 200, again, as predicted, the titers are much higher, as you can see, we're now at 1.5 times 10 to the 12th versus a low of 9.8 times 10 to the 10th, and if you remember where the toxicity thresholds, a ten-fold difference is very significant in whether there will be toxicity or not. And the viral particle-to-infectious unit now are between 5 and 9. ÁÁAnd I won't go through what all these calculations mean, but this is using the fixed law of diffusion and Stokes' equations and by that we can then calculate what is the maximum expected observed hits and what would be the true value behind that. And by doing that we come up with what we call the normalized and standard titer. ÁÁAnd if we calculate it whether using the static conditions or using the centrifuge conditions, we come to very much the same number of titers and we see that the true value of the viral particle-to-infectious unit ratio in that preparation was really closer to 1.3 to 2.4 and 1.3 to 1.9, very, very tight numbers. ÁÁAnd based on that, we conclude that the majority of the particles that are being produced in these high-quality productions are, in fact, infectious. And it is only a consequence of the detection system that we come up with the empty viral particles or non-infectious viral particles that are often talked about. ÁÁThere's one other little bit of information that's critical within this and that is that there's an original titer and there's a clinical titer. And it's important for people to keep track of the titer, not only at the point of production but, also, at the point of the clinical distribution. This was a little experiment, it was published back in "Nature Medicine," back in 1999. And what it showed was that when these vectors were being transported to the clinic in dry ice, even during a short period of time, the PH of the buffer that they were being transported in would drop, the virus would precipitate and one could lose seven logs of titer in a -- in a very short period of time, within hours. ÁÁSo that shows some of the problems that may be attached to quantifying these viruses, and also some -- why it might be very important and now then, why do we need to standardize and what is being done for the standardization? ÁÁWe need to standardize so that we have something to measure it against and we can all be talking about the same things. People talk about platforming units, CPE units, all kinds of different units and we in our laboratory, when I was back at Baylor, we had made a production and sent it to six antiviral laboratories to be titered, and there was a two-log differential between the lowest and the highest titer that we received back from the people that had experience in titering adenoviruses. ÁÁIt's very important to have something that really -- is meaningful. And when there's a threshold of toxicity between patients, and one group may base their starting dose or their continuing dose, based on another group's toxicity observed, if that has no correlation to each other's titer, then it really, it's very meaningless. So, it's crucial for managing the manufacturing process but, also crucial for maintaining consistent, quality controls and dose escalation studies. And, ultimately, of course, from the producer's standpoint, it will be critical for ever having a true product. ÁÁAnd, lastly, I'd like to say that the standard doesn't need to be perfect at this point. Since I've already pointed out to you how there is a lot of difficulties in where and how to measure these things. And that, hopefully, will continue to evolve, but if we have a fixed point, then everything else that's around it can be related to that fixed point. And if that fixed point is there as we talk about this square, we know that it's one down and one to the left, when we talk about this square, it's two right and two to the -- two down and two right. That, also let's us measure what's the difference between this point and this point. So that will be an equivalence so that we can all be talking about at least comparable units. ÁÁThis is the last slide and, unfortunately, one can't tell very much. But there's a moral in here. And this is just a quick note about talking of multiplicity of infection. It's a fairly meaningless term and we're all very used to it. It's a classic virology term, but as you can see here in these two diagrammatic wells, they're both a multiplicity of infection of one, we have the viruses here. I guess it would be .3, but this is much more likely to reach the cells than this one is. And, yet, they're the same MOI. And when one goes fishing, one numbers the number of Marlins probably a multiplicity of infection of 10 to the 16th and 1 and there's 10 to the 16th of them, it's very hard for me to ever find one of those so, I would say that using the terminology of multiplicity of infection is not very useful in characterizing these viruses. Thank you. ÁÁ(Applause) ÁÁDR. SALOMON: That was excellent, Estuardo. So, one of the things that just occurs to me is that I found it very striking that there was such a difference in the titer when one did something as simple as move the virus from a production facility to the bedside. And that's a concern, actually, we've had in our own laboratory with retroviral vectors, that many of them fall apart at rather small PH changes, and it sounds like the adenovirus is prone to the same thing. ÁÁSo how do we -- number one, have they figured out a way to address that issue, since that would be dramatic? And, secondly, related to that, how does that fit into how one would manage the standard, because it could be quite a problem, if at different times, you standard was changing because of the conditions. ÁÁDR. AGUILAR-CORDOVA: Yes, and so, the first answer is, yes, that the way to manage that has been figured out. And it's fairly simple, you know, one just doesn't -- one protects it from dropping the PH in this case, either by a physical barrier or different packaging, formulation makes a difference and such. And, you know, our industrial colleagues, in industry they have a lot of experience in this. And what they do is, they usually validate shipping conditions and they validate that the product at the end has the same characteristics as the product in the beginning. A lot of us in academia had that kind of experience in the past, so we weren't quite as aware of these things. ÁÁAnd for the standard, it has been done vialing and it was an issue that's come up in our many discussions -- the vialing system, the formulation of it and the shipping -- storage and shipping, to validate that it has the same characteristics. ÁÁDR. SALOMON: Dr. Mulligan. ÁÁDR. MULLIGAN: Could you walk us through the exact context where the standardization would be helpful? I can definitely see for detection of RCA how helpful having this would be, because you're looking at truly presumably the same thing, wild-type versus a wild-type. In the case where you have -- people have vectors that are different -- that are not growable on, you know, 293 cells, so doubly deleted or more sophisticated things. ÁÁI'm not sure I see how this would help guide standardizing the dose given to patients because, you know, the characteristics of the growth of these things in different cells is different. How would you see that this would be helpful for that application, which is a very important application to make sure that some people aren't giving, you know, 1,000 times more virus than we know to be safe? ÁÁDR. AGUILAR-CORDOVA: So, from the standard of particles, if one is working with an antiviral particle, a particle is a particle, so it doesn't matter how it grows within a cell. So it still will standardize the measurement of viral particles. In addition, to that, if one is working with what the majority of vectors are today, which are simple, that will grow in 293 cells and such, or whatever production cell line one is using, they will usually call a CPEF, a cytopathic effect on those cells. And so one does not need to be measuring, necessarily, whatever the transgene is but, rather, the physical effect on those cells. ÁÁAnd at least one has -- and the replication competent will cause the same issue. So, at least one has the ability to quantitate it with regards to something else. It's unlikely that one would have a vector that would be that much more efficient than a wild-type in causing the cytopathic effect. ÁÁNow, if you get farther towards, say, a helper dependent vector, in that case, the standard would probably only be useful in as far as the particle number. ÁÁDR. SALOMON: So, getting back -- the standard is going to be wild-type adeno? ÁÁDR. AGUILAR-CORDOVA: This first standard is going to be wild-type adeno. ÁÁDR. SALOMON: And it's going to be adeno-5? ÁÁDR. AGUILAR-CORDOVA: It is an adeno-5, yes. ÁÁDR. SALOMON: Right. So, how big a problem will that be for vectors based on adeno-2 or 4? ÁÁDR. AGUILAR-CORDOVA: For vectors based on adeno-2, it'll be almost no problem at all because they're very similar viruses. ÁÁDR. SALOMON: Right now, are the criteria for the clinical trials that are ongoing as specific as would be suggested by some of your discussions? Such as, when you do the titer counts, you have to centrifuge at so many RCF for so long in such a volume to do the sort of particle counts? I mean, how much standardization do we have today on that? And that may not be -- let me direct it first to you Estuardo, but then, perhaps, to the FDA staff. ÁÁDR. AGUILAR-CORDOVA: To my knowledge, there is no standard protocol and, in fact, there is this system, including the physical calculation which we term NAS titer, or normalized adjusted standardized titer, that was proposed a few years ago, but, at this point, I think that there is no standardized protocol, and that's why a lot of the numbers, even that were received in the March 6 letter, which is now, one has to do the best that one can to interpret those, but a lot of those numbers are just not comparable. ÁÁDR. SALOMON: Steve, Joyce, do you have comments? ÁÁDR. BAUER: Yes, I have a comment about that. It might be true that if you looked at different vectors from different people and, you know, gave each other your vectors, you'd get different titers, but we're also looking at the fact that it's one manufacturer whose done one set of preclinical studies and they're dosing -- in their preclinical and clinical studies based on their own measurements. So, at least within their study, they are, you know, achieving a fair degree of reproducibility. ÁÁThe question is, how would that compare across studies and that's an important thing for us to make some of these recommendations and considerations about toxicity but also, I think that the dose that you start with in the clinic is always based on a preclinical study with your particular vector preparation, so that the comment that Estuardo made about basing doses, I think that's true to the extent that you might look at other studies to base your starting dose in the preclinical model, but then you would use that dose and your own experience in a preclinincal with your own product to determine what your dose range in the patients will be. ÁÁDR. SALOMON: And we won't get into it today but, of course, preclinical models are either small rodents or large animal models. And I don't want to say the obvious, but for the record, you know, going from a dose in a nonhuman model to a human model has got it's difficulties. And most clinical trial designs are done as dose-escalation studies, particularly at this stage in gene delivery, so, you know, at some point here this dose issue is really -- is going to be critical. ÁÁDR. BAUER: I think, also, one of the benefits of having the standard is it will, in a way, bring people to use the same methodology, you were asking about that, but I think that's going to be an important outcome. ÁÁDR. SALOMON: I think that's great, yeah. That's why I wanted to make sure that we got enough discussion of that. Dr. Lawton and then Dr. Mulligan. ÁÁDR. LAWTON: I was actually following up on the comment that you just made and one of my questions was, from a practical perspective in the future, once the standard has been set and the methods have been set, are you assuming that everybody will, indeed, be working to that so that you have that measure, and that that will be a requirement that those methods are used so that you understand it across all of the different companies that are doing the work? ÁÁDR. BAUER: I think we haven't really seen what the outcome will be once we have a standard or -- let me just say one thing, we refer to it as a reference material, because the word standard in regulatory parlance has a specific meaning, so -- when the reference material is available, we haven't said that we're going to require everybody to do the assay the same way, but my belief is that, in effect, that's what will happen because there will be a titer assigned by a consortium of laboratories an agreed-upon standard and I think, given the kinds of consideration that have just been pointed out. I think people are going to have to use the same methodology, but if it -- if that doesn't work out, you know, we might go to saying, well, you need to do it this way so we have confidence that we can compare titers. ÁÁDR. SALOMON: Dr. Mulligan, do you have a comment? ÁÁDR. MULLIGAN: Yes, you raised this interesting question of the E1A plus E1B deficient vectors. How has the FDA looked at those in terms of RCA? I would have thought from what you see in papers that, you know, you can get virus titer, you know, in normal 293 cells in some of these. There's certainly some controversy about this but there's certainly a number of conditional adenovirus vectors where you will see, although it's reduced -- that is, things that have a preference for tumor cells or whatever, you will see some growth. How would you possibly get virus that passes even the old tests? ÁÁDR. BAUER: I think that's a good question. The approach has been that we realize these replication selective viruses, in fact, it's a selective, it's not an absolute barrier, as you pointed out, but the approach is to say we asked to develop better assays to measure the degree of selectivity. And the other thing is from the preclinincal and clinical studies, we do keep in mind that these vectors are a different class, in effect, and, therefore, ask for increased clinical monitoring and that sort of thing. So -- ÁÁDR. MULLIGAN: That's interesting because depending on how much you believe in how well they work, you might have a lot of -- ÁÁYou know, they're shooting in a lot of these virus particles. So, I mean, basically, they're as exempt from the normal standards for traditional vectors, is that right, in terms of helper. ÁÁDR. BAUER: We don't have a good handle on ways to measure it, that's true, but we respond to that by different clinical monitorings and different considerations for the preclinical studies. ÁÁDR. SALOMON: Jay, did you have a comment? ÁÁMR. SIEGEL: I have a question for Dr. Aguilar-Cordova. So, you show data suggesting that the sensitivity of an assay to infectious particles can vary vastly, depending on how the assay is done. As you know, we've set a limit at that 3 percent -- 3.3 percent of -- at least, of particles should be infectious, I guess that was tightened up from 1 percent. And that was based on what seems achievable in current experience. ÁÁIs it your conception that, actually, a much higher level is typically being achieved and that those numbers reflect insensitive assays for infectious particles? ÁÁDR. AGUILAR-CORDOVA: I can only speak to some of the data from our own productions and, in those cases, the majority of them are achieving that and better. Now, having said that -- ÁÁDR. SALOMON: The majority are achieving 3.3 percent or are you saying you're achieving -- because in one of your numbers, it looked like you were getting over 50 percent in one case? ÁÁDR. AGUILAR-CORDOVA: Well, not, let's see, I think I showed data that showed that the Vp-to-iu ratio was in the neighborhood of 1.9, 1.3 so up close to that, I think. And, again, I believe that some of that is just due to detection ability. But having said that, there are some constructs in which -- even with our own system and, as Steve mentioned, very reproducible system within house. Some constructs have a little higher Vp-to-iu ratio than others and we haven't quite figured out why yet? ÁÁDR. SALOMON: But, Estuardo, I also was confused by that, only in the context of what Steve had presented and what you presented. It's a ratio of 1.3 -- I ÀÀÀÀÀÀÀÀ mess up the math, particularly, when I'm up here not thinking straight, but 1.3 would be almost 75 percent, wouldn't it? ÁÁDR. AGUILAR-CORDOVA: It would be about 50 percent -- close to 50 percent. ÁÁDR. SALOMON: Yes, so -- ÁÁDR. AGUILAR-CORDOVA: 1-to-1 would be 50. ÁÁDR. SALOMON: So, that's a whole lot more than 3 percent. ÁÁDR. AGUILAR-CORDOVA: Right. ÁÁDR. SALOMON: So, I'm having trouble with -- ÁÁDR. AGUILAR-CORDOVA: Remember, though, the data that I showed you -- what it also described is that from the very same virus vector preparation, we were able to have titers, I believe it was from 9 times 10 to the 12th to -- I mean, 9 times 10 to the 10th, to 6 times 10 to the 12th, depending on how we set up the titer assay. And so, in one instance we're almost 100-to-1 and in the other we're at 1.3-to-1. It's the same vector preparation, it's just detected in a different way. And so, that just emphasizes the actual point, that it may not be an intrinsic characteristic of the vector preparation, but rather a consequence of your detection system. ÁÁDR. SALOMON: I can't think of any better way of articulating why were here today. Marshall. ÁÁDR. HOROWITZ: Do you take off the input virus after a period of absorption? ÁÁDR. AGUILAR-CORDOVA: We have tried that, yes. And under static conditions, we basically see no difference in titer, if the virus is taken out of the well and placed on a different set of wells 12 hours later and, I think it was 24 hours later, as well, and the titer did not change. Indicating that the majority of the virus was still in the supernatant and also indicating something about the stability of that virus. ÁÁDR. HOROWITZ: Yes, so what you did is the supernatant, replated, has the same titer the second time around as the first time around? ÁÁDR. AGUILAR-CORDOVA: Correct. ÁÁDR. HOROWITZ: Saying that the extraction is relatively low, because that's what I was going to suggest. And that's another way, although more tedious to really see what's there, until titer and exhaust, I should say, the supernatant. By repeating titers, you obviously realize there's more virus there than you're scoring the first time? ÁÁDR. AGUILAR-CORDOVA: Right. ÁÁDR. SALOMON: Dr. Sausville? ÁÁDR. SAUSVILLE: So, I think this discussion, though, illustrates a point that Alison Lawton was leading us to in the sense that, it's one thing to describe a reference standard, but it's -- but I think we're going to have to go a lot farther than that. There's going to have to be a, I think, a consensus, potentially, at some point as to how the assay's done, what the cell types, et cetera, so it's more than just a standard, because what this discussion is showing is that the concept of titer is really very context and assay dependent, and so it seems that without defining those things, you know that a standard is going to be relatively meaningless, actually. ÁÁMR. SIEGEL: On the other hand, it's clear that a standard is the first step toward evaluating what's the best assay. And, often, standards are sent to multiple labs for evaluation is where you determine what assays are sensitive and what are not. ÁÁDR. SAUSVILLE: Right, so I certainly take your point. And the standard is the first step. I just would certainly hope that we don't regard it as the last step. ÁÁDR. SALOMON: Yes, that's a key point, I think a couple of us have made it in different ways, so, that's great. I mean, we owe that to Estuardo for focusing us on that. Any other comments or questions at this point? I always, and I apologize to the audience, not having earlier said that we -- you're part of this discussion, as well, that's the purpose of these public meetings. ÁÁSo there is a microphone there. All I ask is that you identify yourself for the purposes of the record, but I would welcome any of you at anytime during this place to get up and make a comment. ÁÁDR. HUTCHINS: Hi, Beth Hutchins of Canji. As far as the reference material goes, the real purpose of this reference material is actually to define a unit. We're actually going to put the markings on the ruler and then later we'll figure out what's the best way to measure those -- measure -- how to use the ruler basically. And it goes back to what Estuardo was saying in terms of, we're creating, in essence, an arbitrary unit. The method is not so critical right now. In terms of the reference material, we are, in fact, including the diffusion calculation correction. That's being incorporated into getting that into the initial number for the reference material. ÁÁAnd down the line, then we can look at the issue of methodology. But we're creating a unit is what we're really doing. We're defining that unit right now, and that's why, in the end, laboratories, different companies, whatever, can use this reference material to validate their method. Their method may be different, but the unit will be defined and as long as they can report back units that mean the same thing, then we'll all be able to reference relative to each other regardless of the methodology that they're using. In the end, it may mean that people move towards very similar methods, but it doesn't force people to do that automatically. ÁÁThat's also true for not just the infectious unit, which varies a lot but, also, for the particle, the physical particle number itself. Because even though people do OD260 SDS methods, those methods, depending upon exactly how you do that, how you lyse the material and make the measurement vary also and I think the FDA has commented that they see variability even in that approach. ÁÁSo, even though you would think the physical measurement should be the same for everybody, in point of fact, a lot of us recognize that there are definite differences there, and so the reference material will also define the physical unit, which I think will also be very useful to the field. So I just want to point out those two issues that we're defining units, not so much how you measure the unit. ÁÁDR. SALOMON: Yes, good points -- still, however, you have to admit that the definition of a unit, then, would allow us if we set, then, standards based on units, that would be very useful. Yet, how one handles the reference material when it gets to the different labs in the field will also have a lot to do with how accurate the unit determinations are. ÁÁSo there's still going to have to be a significant amount of specification on exactly how the assay's done and that gets back to Dr. Sausville's point. ÁÁIs there another comment, and then we'll go on to the next talk? ÁÁMR. MURPHY: Chris Murphy with Genzyme. I just wanted to kind of clarify: The change in the particle to iu ratio that's being proposed -- is that going to fall in line with the addition and, I guess, correction with, you know, the calculation with Fick's Law? And the reason that I ask is, you know, currently it's not uncommon to get a particle-to-iu ratio using a Spearman-Carber calculation without correcting for the diffusion of virus up around 30 or 40. If I correct for diffusion of virus, I can bring my particle-to-iu down to 1-to-5, that sort of thing. I guess what I want to kind of verify is that is this change already going to be implemented now regardless of coming to a consensus on the titer assay? ÁÁDR. BAUER: I can respond to that. I think the methodology -- we have not, in the past told our sponsors the way that they need to do these measurements, but, of course we look at how they do the measurements. So, if you have a measurement and a system for calculating the infectious particles that, you know, we have looked at and accept, so I think, in effect, we will be applying those correction factors. I think that was it. ÁÁDR. SALOMON: Okay, then the next speaker I'd like to introduce is Dr. Beth Hutchins, Director of Process Sciences for Canji. The title of her talk is Replication Competent Adenovirus Assays and Clinical Data for rAd-p53 in Cancer Patients. ÁÁDR. HUTCHINS: Can everybody hear me? All right. I'm going to give a little bit of an overview of how RCA assays are typically done to just give you some better feel for, actually, the variation in that area and then talk, specifically, about adenovirus p53 vector and our methodology and then data from patients relating to RCA shedding and et cetera. ÁÁThere really are two sources for RCA: One is that it can be created during the actual construction of the vector and this can happen with the most common methodology and really it was the traditional methodology until more recently. And that is the large fragment method of recombinant adenovirus construction. And that's where the recombination takes place in the production cell line. And it does not matter what type of production cell line you're using. It doesn't matter if it's a 293 cell oops -- did something just cut out or, okay, or PER.C6 or any of the more truncated E1A complementing cell lines that are now becoming available -- you want to adjust something? In any case, if you allow the recombination event to occur in the production side and then try to select out viral plaques from that, what you'll find is whether you do serial plaquing or not, you don't eliminate multiple things in the construct. ÁÁThe newer E. coli recombination methods, where you do all your plasmid (?) recombination and then you select a then you select a single plasmid from the bacteria and use only that single plasmid to do your transduction into the production cell line can eliminate this as a source of RCA. Most of the constructs, though, that are in the clinic these days were still made by the old large fragment method of production or construction. ÁÁThe other source of RCA is -- during production that is recombination frequency between an overlap in the E1 region of the cell line versus the vector backbone itself. And the thing is, the frequency of that recombination isn't really known and it, actually, has to be estimated for every combination; that is for the overlaps that exist between your specific vector backbone and your specific production cell line. Now the newer cell lines try to address that by specifically eliminating this -- these types of overlaps. ÁÁIn the field today, RCA testing is a bioassay involving either one or two cell lines. The indicator cell line is most commonly the A549 cell line or a lung carcinoma. And detection can be by a variety of methods: Either cytopathic effect or immunostaining or PCR, as the end thing that you're measuring. When you rely on cytopathec effect, generally, the assays are set up with a confirmatory step to show that the specificity is not something that -- some infection event of some other virus that just happened to occur during the assay that is, in fact, an RCA. ÁÁThere's no guidance on, you know, which to do this and you'll find that there's a variety of methods out there. I think FDA can comment on that for the committee. ÁÁNonetheless, all of these assays are set up and qualified to be sensitive to 1 pfu or IU or infectious unit of RCA at some confidence level, hopefully, at the 95 percent confidence level. And, in essence, what you're doing is a plus/minus or quantal assay. And how you get quantitation then is based on the sample size that you test. So, you don't get out from the way these assays are typically run out is I have X-number of RCA in my material, what you get out is I have one or I have less than one in whatever amount I test. ÁÁAnd this goes through an example of this type of method of quantification. So, you test at different amounts of sample. 5 times 10 to the 8th; 1 times 10 to the 9th; 5 times 10 to the 9th, 1 times 10 to the 10th vector particles. And you get negative results, no positives detected at the two lower levels, but you get positive results at the two higher levels of particles assayed. And what you then estimate is that the amount of RCA is less than 1 in 10 to the 9th or greater than 1 in 5 times 10 to the 9th vector particles, which, if the clinical dose is 10 to the 12th vector particles, then it does contain somewhere between 200 and 1,000 pfu of RCA is what we estimate. There's no exact number that's coming out of this, it's a range. ÁÁPart of that is the 1 pfu detection is also, of course, limited by the distribution of virus in a sample. This is where the amounts of replicates of the RCA bioassay can become quite critical. This last calculation is corrected based on your handout, somehow my math was not that good that day. ÁÁNow, the construct and the data that I'm going to talk about more specifically relates to recombinant adenovirus that expresses the wild-type human p53 gene. The backbone of this vector is an E1a, E1b, protein IX deleted cassette with also a partial E3 deletion that completely inactivates the E3 expression. ÁÁThese deletions were created to allow the insertion of the p53 expression cassette in the E1 region and we purposely put the protein IX deletion in to decrease the overlap between the vector backbone and the 293 production cell line, which is the cell line that we use for production of this vector. That decreases it, actually, from about 1,000 base pairs down to 200 base pairs, just for background information. ÁÁWe had -- we've done pretty much of the routes of administration that Dr. Bauer showed earlier. The data that I'm going to talk about relates to these trials and these patients. We've done intratumoral injection, with 72 different subjects; intrahepatic artery infusion, with 50 different subjects, with doses up to 7.5 times 10 to the 13th particles, though most of -- a good majority of those patients got 2.5 times 10 to the 13th particles; and also by the intraperitoneal route, with 54 patients and the dose ranged in that case up to 7.5 times 5 to the 13th particles; and the majority of those patients got 7.5 times 10 to the 13th particles in a dose. ÁÁSo, the bioassay for RCA that Schering Plough (?) uses to test our production lots of the p53 adenovirus is, again, one of these quantal bioassays, and this one has a CPE readout. We use a two cell line method because the p53 gene has a particular effect on the A549 cells that we wanted to avoid and so the indicator cell line, though, is still the A549 cell line. It's a four-week -- typical four- week assay. We use PCR to confirm any positives so that we know that we are detecting RCA, if we do detect it. And it is sensitive to one pfu of RCA. ÁÁNow, it's sensitive to one pfu RCA, but because of plasone (?) distribution, when you want to validate the assay and show that you have 95 percent confidence level for detection of something, when you do triplicate tests of -- or you do an N of 3 on testing your material, we can, with 95 percent confidence, detect 4 pfu of the control spike. If we wanted to have 95 percent confidence in one pfu detection, the end jumps up quite dramatically and it's not practical to do that. And the only RCAs we've ever been able to detect, either through a process -- a validation study or that come up in actually assaying production lots are, really, the dl327 backbone, that is the p53 expression cassette's been kicked out, the E1 gene looks normal, the E1 region looks normal again, but it still has the E3 deletion, that is part of the parental backbone from which the vector was originally derived. ÁÁThe specification for the 58500 p53 adenovirus clinical product is less than 40 pfu of RCA in 7.5 times 10 to the 10th viral particles. Each batch is tested either in triplicate at 2 times 10 to the 9th particles, which is, actually, the way it's been done for the -- about the last 2 years. Or it's tested earlier the first years that we were doing this, it was tested and not at one but at three different levels. ÁÁBased on the assay confident levels for the triplicate test at just 2 times 10 to the 9th particle, if we find -- if there are no positives in -- of the three tests, or there's 1 of the 3 is positive, then the batch meets the specification. If two or three of the tests come up positive then the batch fails the specification because there would be greater than or equal to 40 pfu in that 7.5 times 10 to the 10th particle amount. ÁÁAnd this data just represents the summary of the validation data, looking at the confidence levels relating to this triplicate testing. And so, you can see that we have 95 or better than 95 percent confidence in testing -- we can detect the 40 pfu in 7.5 times 10 to the 10th or, because we're testing at 10 to the 9th for pfu detection limit. And that this confidence declines dramatically, which is why zero or 1 can meet the specification but more than that, obviously fails the specification. ÁÁNow, if you look at or as we look at our protection lots over a period of a number of year, what we find is that a certain percentage, roughly 10 percent consistently fail the specification. And -- which means that they would have greater than the amount of 40 pfu in the 7.5 times 10 to the 10th vector particles; and about half, based on testing at the three different levels come in right on this -- right where you end up in -- if you look at the 7.5 times 10 to the 13th vector particle dose, you would have about 4,000 RCA pfu maximally. And about, the other half would have ten-fold more than that. So at our highest dose levels, we could have put -- we could have dosed a patient in a single dose with as much as 40,000 pfu of RCA. Now, we can't say exactly how much we dosed with except that it's in a particular range, just because of the way the assays are set up, so you need to keep that in line. But it is, obviously, more than the numbers that Steve was talking about earlier. ÁÁNow, I'm also going to just briefly the type of methodologies that we were using to monitor the patients clinically and to look at various biological samples from them and these methods included both specific and nonspecific methods. So, one nonspecific method is an ELISA that detects the hexon antigen. This does not distinguish intact virus from just parts of the virus, and it also does not distinguish product from RCA or wild-type. ÁÁAnother type is an infectious assay on 293 cells. In this case, we're using fluorescence -- a flurocytometer and fluorescence to detect infectious adenovirus. Again, this does not distinguish the type of adenovirus that is an RCA a product or a wild-type infection, but it does say that it is an intact infectious virus and not just pieces of virus that are in the biological sample. ÁÁWe also had a variation of this assay where the sample was placed onto A549 cells. And, again, this detects infectious virus, but it only detects either RCA or wild-type and does not detect product because the product cannot replicate on an A549 cell. ÁÁAnd, finally, we had PCR assays that you could use those to specify what you were detecting, that is, you could specify that it was RCA or wild-type adenovirus or the product, but one caveat of that is that you're detecting DNA and it does not tell you that that DNA represents an in tact virian. So, these methods detect different things, but they do allow you to get certain valuable pieces of information, depending on the method you're using. ÁÁThe types of biological samples I'm talking about are either urine, feces, sputum, saliva, any variety of things and, also, several of the methods have been used to monitor serum samples, blood samples. ÁÁOf the studies that I mentioned earlier and the patient numbers, here listed by the various routes of administration and then the four different methods. These are the -- just looking at a variety of biological samples but just now looking at the issue of shedding. You'll see the only method that specifically you can look at RCA, these 2 methods, actually, but in this assay, we found, in 63 different samples we -- or from 63 different patients, we did not detect any shed RCA. This one positive here is product, not RCA. And these -- these were all product-related positives here. Now, you can see, also, it's a very low rate of shedding, but no RCA's been detected in any of the samples we've analyzed. ÁÁSo, what are the types of populations we're studying? This vector has actually been developed for -- was being developed for specific cancers. And so all the patients were cancer patients. All these subjects were initially selected to be antiadenovirus positive prior to entry onto the studies and, of course, not to have any evidence of an active adenovirus infection and what we did know that the neutralizing capacity of these antiadenovirus titers before they received any of the adenoviral p53 vector varied considerably in the population. But always, consistently, increased and increased dramatically with dosing with -- after administration of the vector. ÁÁAnd for the subjects were -- the very few subjects we had were adenoviral shedding was detected, we were not able to identify any adverse clinical sequeli. ÁÁThis is just to give you an idea of the type of information you can get from monitoring patients looking at their ÀÀÀÀÀÀÀÀ antibody response, this is now for patients enrolled in the ovarian interperitoneal administration study and this is a multiple dosing phase of -- portion of that phase 1 study, and in this case the patients we all receiving 2.5 times 10 to the 13th particles or 7.5 times 10 to the 13th particles, I think about half and half in this case. They also did, at the same time receive chemotherapy. ÁÁSo -- and the way the drug was dosed is there are five days in a row where the drug is administered each day at that does 7.5 times 10 to the 13th particles, and then 4 weeks later another cycle of 5 days of dosing occurs. So, this is referring to one cycle of dosing, another cycle of dosing and another cycle of dosing. And you can see that this is now average data, and I have to say that the arrow bars are not -- are fairly wide, but it gives you a sense that you do see some drop in the hemoral response but, of course, memory comes back and you see that rise again and it continues to go up and up and by the time you're out past cycle three in later studies we did up to five cycles, of course the neutralizing capacity is quite dramatic. ÁÁNow we do also have evidence, though, that we're still getting transgene delivery and expression at cycle three and we had very limited samples that told us the same thing for out to cycle five. So, even though there's this very large neutralizing hemoral capacity in the body, we were still getting delivery of our vector. ÁÁJust to get to the summary points, the bioassays that are used are quantal, but they are sensitive and they can reliably detect RCA, at least in replication deficient vector products. ÁÁThe more precise quantitation than this quantal method where you get, sort of, this range of RCA that you can estimate in your material is impractical in this bioassay mode, because of the amount of testing that was required. The amount of cell culture that would be required is what becomes impractical. ÁÁIf the desire is to have a specific quantitative number come out of a particular method, you would have to consider real time PCR as an option, looking for quantitating more specifically. No one's doing that right now, routinely, but that would be another way to look at the RCA issue. ÁÁWe were not able to detect any shedding of replication competent adenovirus in our clinical subjects. There was not PCR positive for RCA and no infectious assay-detected RCA. And where we did find shedding, there were no adverse clinical sequeli identified of the shedding's really very low amount and less than, you know, about maybe 1 percent by the infectious titer assay that was used to look at that. And I think that's basically it. So, I'd be happy to entertain questions from the committee. ÁÁDR. SALOMON: Thank you, Beth. I'm remiss in not having allowed the panel a break, so -- one option here, would be to do some questions, I think, while it's fresh in our mind and then have a short 15- minute break if that's okay with everybody. ÁÁI had a couple questions. Some of them may come up with others, so let me just ask one of them and then we'll see how things come along. One of the things I find very interesting is this whole idea of what are the molecular mechanisms that generate the recombinant -- the replication competent adenovirus through recombination in these models. So you have some evidence here that you gave that the RCAs were rising by replacing the p53 transgene cassette with E1, did I get that correct? ÁÁDR. HUTCHINS: Well, the RCAs that we've been able to detect their structure is all that the E1 - the p53 expression cassettes not there and the normal E1 region is there, but the rest of the viral backbone which has this E3 deletion, that's backbone on which the vector is based is the same. So, you don't have a substitution just at the place where, in fact, the large fragment recombination was supposed to be taking place. ÁÁDR. SALOMON: Right, so, again, I'm not an adenovirus -- so this could really be stupid, but what -- so what I'm trying to understand is when you constructed the vector, you did a number of different things; you deleted E1a, you deleted E1b, you deleted E3, and you made changes in protein IX to reduce the homology (?) right? ÁÁDR. HUTCHINS: Right, from a practical point of view, what's done is or what was done in this case is the dl327 adenovirus, which has the E3 deletion in it already was the starting point. And then the expression cassette was constructed through bacterial plasma technology. And so, you have sort of the front-end of the virus constructed in bacteria and you have what really is going to become the back-end of the virus be from a virus from the adenovirus dl327. That virus is digested using specific restriction enzymes, claw-1 (?) is most frequently used to get thee large fragment. And, basically, is what it does is you end up with the adenovirus but from the end of E1 on, and what you're doing is asking these things to recombine, now you're really, just the part of the virus that has the expression cassette and the front-end with the ITR to recombine in the producer cell line. ÁÁThis is the most traditional way that this is done, regardless of whether you have an Ad2 backbone and Ad5 backbone and what other elements you're putting together, that's what's traditionally done. It gets more complicated when people have additional deletions, like in the E4 region, so they change what the back-end is. ÁÁDR. SALOMON: So, I guess the question I'm asking is, in this process of engineering the vector -- I mean, part of my thinking here is, if we could figure out the nature of creation of recombinant adenovirus, I mean, replication competent adenovirus through these events, we could suggest that that be part of the criteria upon which one would accept a clinical vector in a trial. ÁÁSo, you made all these changes, you reduced it to 200 base pairs instead of 1,000, yet, you still got recombination with E1, so is that telling us that these are not -- ÁÁDR. HUTCHINS: We think it was there right in the beginning -- ÁÁDR. SALOMON: ÀÀÀÀÀÀÀÀ ÁÁDR. HUTCHINS: The virus material because of the way we did -- we created it, that it was always there and no matter how hard we tried to, basically, subclone it out by serial plaquing, we didn't -- we don't achieve that, at least not in ten rounds of serial plaquing. ÁÁDR. SALOMON: Okay. So there were -- there was already the backbones to create the replication competent adenovirus in the process of engineering the original transgene? ÁÁDR. HUTCHINS: That's our hypothesis for this particular vector, and we think it's a pretty common experience. Now, at the time that we were using that technology in the earlier nineties, we didn't really recognize that that could be a consequence of what we were doing. As we have more data and really begun to understand better what -- the consequences of what we were doing, we now moved to a method where we do everything in E. coli, select the specific thing that's the intact vector and that's what goes to create the intact viral -- the viriant, so -- and now that source, would be limited, and now the only way you could get RCA is through recombination of events that occurred during production. ÁÁDR. SALOMON: Okay, well, that would explain something Estuardo said, then, too. Good. Ed. ÁÁDR. SAUSVILLE: But to pursue that direction of thinking, we, in past meetings, have set out the general idea that vectors of certain sizes should be sequenced and there should be precise definition of what goes into a product. And we stated that for higher molecular weight, or higher sized genomes, larger genomes, there can be a point of ambiguity in relation to the gene of interest as opposed to the background. ÁÁSo, what this leads to is that, if recombination frequency is going to be the major determinant of this in the next generation, do we need to firm up the sequence definition, flanking the proposed target gene and try and use that for a basis of hedging our bets as to what the recombination frequency would be? I place that on the table. I mean, I'm not an adenovirus expert, either, but it seemed that might go part of the way to dealing with this. ÁÁDR. SALOMON: And that may be something that we want to take up right after lunch when we start going into the questions. But I certainly would welcome any comments. ÁÁDR. HUTCHINS: Well I have one, and we're one of the groups that fully sequenced our virus early on and submitted that information to the FDA and the -- there's limits of detection to what you're going to be able to see in there if you have a low-level variant when you're doing this sequencing, both strands, full length. To detect very small quantities of some other molecular variant, you would have to do a different type of analysis and study than you do when you're just trying to sequence the material and just say this is the sequence of my product. ÁÁSo, sequencing of your vector, the requirement that I think Dr. Sausville's referring to in terms of your earlier discussions and recommendations, is not, by itself, the answer to address this issue. ÁÁYou're talking about some other type of molecular analyses where you have to really look at and I hope the committee does not recommend this today, but you would have to look at, you know, a large number of clonal isolates of your vector from some kind of production lot or a number of production lots and then try to analyze those and that would be, you know -- one of the issues, I think, that's being raised today is, well, okay, we could maybe collect that type of data, but if the risk of that material in your product isn't very high to begin with, what is the necessity to do that. You know, what's the benefit that's gained from that. ÁÁDR. SALOMON: I would say that's exactly what you need to be telling us today. That's exactly the kind of reality check from people doing this that I want to hear and that the committee wants to hear and the FDA wants to hear. Dr. Ketner and then Dr. Siegel. ÁÁDR. KETNER: You mentioned that you constructed your E1 replacement to minimize the overlap between 293 cells and the recombinant. So, there are cell lines which are engineered to have no overlap, I mean, PER.C6, I think, is intended to be that way. I wonder whether you've had experience growing your viruses on PER.C6 and whether, in fact, the RCA level drops to zero, as in principle, perhaps it ought? ÁÁDR. HUTCHINS: Your vector backbone still has to match the PER.C6 cell line construction. So if I took my current p53 vector and grew that in PER.C6, assuming I even had -- okay, the way -- I could still, potentially, get recombination because my vector backbone does not, at least the way it's currently set up, does not currently match the way the PER.C6 construct is. So, what Crew Cell (?) will tell you is, if you want to use their PER.C6 system, your vector backbone -- they'll give you the -- or they'll offer to build it for you -- the specifics that you need to know so that you don't have that overlap. You actually need to know what the E1 region is in your producer cell line, so then you can make that match- up. ÁÁBut then, again, if you still do large- fragment recombination as the way to create that vector and you make it in a PER.C6 cell line, you're still going to end up with variant molecular variance in your material. You would still have to create that change, that matched backbone, outside of the producer cell system, such as through an E. coli bacterial system, and then put the selected thing in to not have RCA. And there are other cell lines that are designed the same way. ÁÁDR. KETNER: Given that, I mean, given that this is in principle a possibility, maybe we ought to at least discuss the notion that the RCA levels be reduced to, I mean, to zero by choice of appropriate sub strains. ÁÁDR. HUTCHINS: I would argue that that's the point of your discussion today, what the risk is and what is the necessity. ÁÁDR. KETNER: Yes. ÁÁDR. HUTCHINS: In countries where that's the requirement, they don't have clinical trials going on. ÁÁDR. SALOMON: Well, then that's -- that's good. I mean, we -- that's what we need to discuss, I agree. Dr. Siegel. ÁÁMR. SIEGEL: In discussing the sensitivity of your assay and the ability to use the virologic the viral culture for precise quantitation, you referred to issues of practicality and number of replicants. With the quantal assay, if you have a single hit, kinetics then -- with enough replicates if you, and you have a plasone distribution, you an get a rather precise quantitative in most systems using, you know, the, say, 96 cell plates or whatever there's not that much practical limitation to getting reasonably precise -- I wonder if you would speak, since I also don't work with adenoviruses, what are the limits? How many cultures can be done with how much effort? Is there a problem doing larger numbers? You said you did three replicates -- ÁÁDR. HUTCHINS: Yes. ÁÁMR. SIEGEL: Because more would be hard? ÁÁDR. HUTCHINS: Yes, typically, when people are testing near the level of the FDA guidance in that 10 to the 9th, 10 to the 10th particle range, because there are -- as Estuardo mentioned earlier, the more -- or the higher the concentrational particles that you're putting onto your culture, the more likely you'll see toxicities that will actually affect the assay. This is, you know, this is just a byproduct of using cell lines and high concentrations of the material. ÁÁSo, this means that in order to test, say I wanted -- decided that I wanted to test 10 to the 11th particles each time I did my assay. In order to -- and also still kept my pfu sensitivity to 1 -- and I would have to use a larger number of cells or more wells or roller bottles or flasks in order just to look at that 10 to the 11th particles, and now, if I want to do that with a certain percentage confidence level in my ability to detect that 1 pfu, I've got to do that same number of flasks x-times each time for each lot. ÁÁAnd, you know, I have to say that I'm very happy to be affiliated with a corporation where our lot sizes were quite large, but I know that many institutions, the lot sizes aren't that big, and so that -- that can be a real burden. ÁÁThe other thing is that because it's a quantal assay, the quantitation comes out of the amount you test, not -- because you're only checking one or none, basically, right? So, it's yes or no. And so the amount you test drives the quantitation in this bioassay method. So, that's the other reason why most people only test to about 10 to the 10th, because the amount of cell cultures that's involved. ÁÁAnd the other point that I should point out is that when you do this test, you don't just do the material by itself. You do the material by itself, the material with a spike, so, I mean, you already have these other controls built in that just in a single assay, you have other cell culture, not just the culture involved with the actual replicate numbers of the unspiked product that you're testing. ÁÁDR. RAO: We talked about a reference standard earlier. So if you were to use the reference standard here would you just use it as a spike of the wild-type instead of what you used, or would you have some other method in your system? Or would you use the reference standard? ÁÁDR. HUTCHINS: Well, I would not be using the reference material directly, because I don't want you to use up all the stuff we produce. What I would like people to do, and what we're going to be recommending is that people create an internal reference material that they tie to this material that they'll be able to use. But what the reference material will do is define the unit, the infectious unit, so now when I say I'm spiking or I'm making sure my RCA assay can detect one infectious unit of RCA, it'll be the unit's that's defined by this adenoreference material. And I'll make sure that my RCA assay is validated based on that -- the ability to detect that unit. So, does that clarify that point for you? ÁÁDR. SALOMON: Dr. High. ÁÁDR. HIGH: With respect to some of the questions that we are asked to answer this afternoon, I wonder if you could tell us something about this group of cancer patients that you presented data on? You have about 200 people with non ÀÀÀÀÀÀÀÀ and head, neck and colon cancer and ovarian cancer and so forth. Are there requirements that these are all people who are relatively early-stage? Do you know anything about their -- ÁÁDR. HUTCHINS: They were all -- ÁÁDR. HIGH: Immunologic status -- ÁÁDR. HUTCHINS: They were all relatively late stage, heavily pretreated, I mean, these were phase 1 studies, so this was definitely -- this was not a neoadjutant type of situation, this was, you know, they've already exhausted a lot of their other options. Nonetheless, they had to have relatively good performance status, so that we could distinguish disease-related effects from, presumably, product- related adverse effects. ÁÁSo, but they are, you know, treated -- on the other hand, we have evidence that at least many aspects of their immune capacity were quite well in tact. Hemural immunity as well as cellular immunity from cytokine profiles and some other data that I'm not presenting today. ÁÁDR. SALOMON: Marshall. ÁÁDR. HOROWITZ: Is it appropriate to ask about the 50 patients that had intrahepatic artery infusion with an average of 2.5 times 10 to the 13th -- ÁÁDR. HUTCHINS: Thirteenth. ÁÁDR. HOROWITZ: Times the 13th, yes. Would it be appropriate to ask about toxicity and chemokine and other measurements and expression in liver, is that expanding too much? ÁÁDR. HUTCHINS: It is appropriate to ask that -- we have reported, for instance in the December '99, I guess it was RAC Safety Symposium, we reported in some detail our safety data -- clinical data, and also preclinincal studies related to that and that route of administration and dosing. We, in fact, did see dose limiting toxicity. You notice that in our ovarian trials we heavily focused -- ended up with the 7.5 times 10 to the 13th particles. ÁÁNow that's by a different route of administration, that's tolerated quite well in patients. But at 7.5 times 10 to the 13th particles, by the intrahepatic arterial route, that is not tolerated in patients. But the effect was on -- was a cardiac toxicity, in fact, not a hepatic toxicity. And we went down then to being able to establish that 2.5 times 10 to the 13th viral particles was a safe dose. ÁÁOne of the things that -- we did not reinitiate our IHA protocols after the Gelsinger death, although we could have. We had permission to do so. But for a variety of reasons, including our own concern about how close were we to, you know -- the difference between 2.5 and 7.5 isn't that big. And while we felt very confident that we were defining our particle dosing consistently and, you know, accurately, we just felt that it wasn't worth the risk at the moment until we had more information. So we started to focus more on the ovarian IP route of administration after that. ÁÁDR. HOROWITZ: Did you measure chemokines circulating at that time? ÁÁDR. HUTCHINS: Actually, in the IHA studies we did and I'm hoping, very soon, that manuscript will actually be published, but I think Dr. Bob Warren has actually presented that type of data at several forums, including at the Recombinant DNA Advisory Committee. We do see cytokines IL6 (?) and such in both in serum as well as in tissue, we've looked at that quite extensively. I mean, there's definitely a local immune reaction in liver and there's a systemic immune reaction, but there's no evidence to us that RCA is related to that. I think that gets back to general -- ÁÁDR. HOROWITZ: Right -- ÁÁDR. HUTCHINS: Viral toxicity. ÁÁDR. HOROWITZ: These questions are not so much for RCA, but for the total dose -- ÁÁDR. HUTCHINS: Right. ÁÁDR. HOROWITZ: And if I may ask, were any liver biopsies done and evidence of expression of the transgene in the liver? ÁÁDR. HUTCHINS: Yes, yes. And there was, consistent -- once we got above a certain minimum dose, we had consistent transgene expression in both tumor and non-tumor samples from the liver. ÁÁDR. HOROWITZ: Thank you. ÁÁDR. SALOMON: One question that I had, let me see, is just a simple one, you know, from my calculations and these, again, are flawed when I'm up here, I'm not usually thinking as straight as I should be, but I think it means that about 50 percent of your production runs would not be acceptable if we adopt the new guidelines that FDA staff has suggested? ÁÁDR. HUTCHINS: Well, if you read the guidance, there's always the -- there's always the ability to collect data to support a different specification and then discuss that with the agency and determine if they agree that your data supports that different specification. That is, in fact, the type of thing that we did -- to have a different specification than the less-than 1 in 10 to the 9th that was in the original guidance. So, we have preclinical studies that support -- ÁÁDR. SALOMON: Well, my question though are you okay that this is going cut -- is this going to cut out 50 percent of your production runs, and is that okay? I mean, I have no idea what that means. ÁÁDR. HUTCHINS: If it becomes an absolute rule, yes, it would, but I -- I think the FDA needs to comment here on how they would apply that. ÁÁDR. SALOMON: I'm sorry, Alison, did I get that wrong? ÁÁDR. LAWTON: Can you just clarify that, Beth? If it becomes an absolute rule, my understanding is, yes, it would cut out 50 percent of the lots -- ÁÁDR. HUTCHINS: That's correct. ÁÁDR. LAWTON: And the question is, is that acceptable? ÁÁDR. HUTCHINS: No. ÁÁDR. LAWTON: Right, thank you. ÁÁDR. HUTCHINS: You need to understand we're pro -- we produced the vector -- ÁÁDR. SIEGEL: But -- ÁÁDR. HUTCHINS: At a fairly large scale. You're talking about vector runs where we had 10 to the 16th particles produced at a time, not that we would make purification batches on that same scale, but, I mean, and the RCA level would be consistent in that entire, you know, viral culture batch, so that would not be acceptable. ÁÁMR. SIEGEL: I think, though -- ÁÁDR. SALOMON: And that's what I was trying to get you to say. ÁÁMR. SIEGEL: There needs to be clarification. We've not proposed a rule, we're proposing a guidance. ÁÁDR. SALOMON: Right, exactly. And, I mean, I just wanted some feedback on, you know, what the field thought of that, you know, guidance before we get into it this afternoon. Dr. Flomenberg and then Dr. Sausville. ÁÁDR. FLOMENBERG: Phyllis Flomenberg, Thomas Jefferson University. You mentioned that you prescreened all of your patients for antibody to adenovirus, was that sera-type-specific antibody? ÁÁDR. HUTCHINS: The test was -- the assay was designed, you know, the adenovirus type 5, but how specific are -- we did not then go back and look at whether those antiadenovirus titers that we measured were only -- ÁÁDR. FLOMENBERG: So, it wasn't a neutralized assay? ÁÁDR. HUTCHINS: To ad-5 versus to any other adenovirus isotype. ÁÁDR. FLOMENBERG: So -- ÁÁDR. HUTCHINS: We wouldn't care, actually. That wasn't what we -- the question we were asking -- we were only looking at the ad-5, but to say that it was only -- those responses were specific to ad-5, I can't say that. ÁÁDR. FLOMENBERG: Was it a neutralizing assay? ÁÁDR. HUTCHINS: We did both kinds of assays, the requirement for entry did not discuss neutralizing capacity, just that they had to be seropositive. ÁÁDR. SALOMON: Dr. Sausville. ÁÁDR. SAUSVILLE: So a question that the introduction of the clinical trials data that you alluded to in the cancer patients and also in your response to the question about viral replication in both liver and tumor, do you routinely quantitate the expression of coxsackie adenovirus receptor in tumors and does that influence or could that influence the potential perception of infectability of a given lot -- given -- depending on the use of the material? ÁÁDR. HUTCHINS: We did look at that in a number of our Phase I studies. Actually, the IHA study and the ovarian IP study. In the IHA study, there was a relationship between CAR level and expression. In the IP study, there was not in tumor tissue. Why that is I can't say, but besides the fact that there's CAR and, of course, the secondary integren receptors are required for internalization, there are still undefined receptors that most of us believe exist and we don't know what they are, and so there can be other mechanisms for entry of adenovirus into tissues. ÁÁAnd just to correct a point: When I was talking about detecting expression of the p-53 gene, we don't believe that the vector was replicating in those tissues. You were saying, just to clarify -- ÁÁDR. SAUSVILLE: But, I guess it ultimately raises the question, when one uses a standardization procedure that's based, presumably, on a given report or cell type it has a certain receptor and entry mechanism. Is it clear that that's the same receptor and entry mechanism that might either me