• Decrease font size
  • Return font size to normal
  • Increase font size
U.S. Department of Health and Human Services

Vaccines, Blood & Biologics

  • Print
  • Share
  • E-mail

Transcript for Public Workshop on Cell and Gene Therapy Clinical Trials in Pediatric Populations

1

UNITED STATES FOOD AND DRUG ADMINISTRATION (FDA)
PUBLIC WORKSHOP ON CELL AND GENE THERAPY
CLINICAL TRIALS IN PEDIATRIC POPULATIONS

Bethesda, Maryland

Tuesday, November 2, 2010

2

PARTICIPANTS:

KAREN MIDTHUN, MD
Director, Center for Biologics Evaluation and
Research (CBER), FDA

CELIA WITTEN, PhD, MD
Director, Office of Cellular, Tissue and Gene Therapies (OCTGT), FDA

WILSON BRYAN, MD
Clinical Evaluation Branch Chief, OCTGT, FDA

STEVEN JOFFE, MD, MPH
Assistant Professor of Pediatrics, Harvard Medical School

ROBERT SKIP NELSON, MD, PhD
Pediatric Ethicist, Office of Pediatric
Therapeutics, Office of the Commissioner (OC), FDA

GEORGE Q. DALEY, MD, PhD
Harvard Stem Cell Institute
Children's Hospital Boston
Harvard Medical School
Professor of Hematology and Director, Stem Cell
Transplantation Program, HHMI/Children’s Hospital
Boston, and Professor of Biological
Chemistry/Molecular Pharmacology and of
Pediatrics, Harvard Medical School

HELEN HESLOP, MD, MB, ChB
Professor, Department of Pediatrics and Medicine,
Baylor College of Medicine

CHRISTOPHER BREUER, MD
Associate Professor of Surgery (Pediatrics) and of
Pediatrics and Director of Tissue Engineering,
Yale School of Medicine

3

PARTICIPANTS (CONT'D):

DONALD KOHN, MD
Professor, Microbiology, Immunology & Molecular
Genetics and Pediatrics, University of California at Los Angeles

RONALD CRYSTAL, MD
Professor of Genetic Medicine and Chairman,
Department of Genetic Medicine, Weill Cornell
Medical College

THERESA CHEN, PhD
Pharmacology/Toxicology Reviewer, OCTGT, FDA

DAVID MAYBEE, MD
Medical Officer, OCTGT, FDA

JEFFREY BOTKIN, MD, MPH
Professor of Pediatrics, Department of Pediatrics,
Adjunct Professor of Medicine, Department of
Internal Medicine, University of Utah School of
Medicine and Associate Vice President for Research
Integrity, University of Utah

ANGELICA WALDEN, MBA
Patient Advocate, Department of Quality
Management, Medical College of Georgia and
Institutional Review Board Member, Medical College
of Georgia

JEFFREY KAHN, PhD, MPH
Director, Center for Bioethics, University of
Minnesota and Professor, Department of Medicine,
University of Minnesota School of Medicine

SUSAN KORNETSKY, MPH
Director of Clinical Research Compliance,
Children’s Hospital of Boston

4

PARTICIPANTS (CONT'D):

SCOTT DENNE, MD, FAAP
Professor, Department of Pediatrics, Indiana
University School of Medicine

*********************************
5

PROCEEDINGS
(8:00 a.m.)

DR. MIDTHUN: We'll give it half a minute and then we'll get going. Great. Good morning and welcome. My name is Karen Midthun. I'm the director of the Center for Biologics and I'd like to welcome all of you to this conference this morning. Especially, I would like to start by thanking our co-sponsors. They include the Office of Good Clinical Practice and the Office of Pediatric Therapeutics and the Office of the Commissioner; the NIH Office of Biotechnology Activities, which is also graciously providing a webcast for this conference; the American Academy of Pediatrics; the American Society of Gene Cell Therapy; the International Society for Stem Cell Research and the Public Responsibility in Medicine and Research. And so you can see that it's really been a very strong collaboration in putting on this conference and we really thank all of our co-sponsors for working with us on this. In addition, we would like to

6

recognize our many colleagues who worked to develop the agenda, the questions, and who will be speakers and panelists today. Why are we focusing on cell and gene therapies with regard to pediatric populations and studies in these individuals? I think we recognize that gene and cell therapies offer great promise. They are products that may restore function and modify the nature of the disease; however, they also present unique safety concerns and there are limitations to existing methods to predict safety or monitor patient safety with these products. I think we also know that there are special considerations when it comes to pediatric populations. They are a vulnerable population and as a result we have regulations which you oftentimes refer to as Subpart D, which you'll hear more about later today. But these are in place to ensure that IRBs provide additional safeguards for studies in pediatric populations. And there are particular considerations also when it comes to some of these products. For

7

example, it may not be possible to study these products in adults because the disease may be unique to pediatric populations, and in those instances you're relying on information from preclinical studies to go into a first-in-children study. I think that today we're going to hear a number of different things and the goals of the workshop are to review the roles of the FDA and the IRBs in oversight of pediatric trials, especially regarding 21 CFR 50 Subpart D to gather information from the clinical and research communities, including the IRBs, clinical trial sponsors and investigators, and the public and other stakeholders involved in cell and gene therapy are best practices in support of IRB review of pediatric clinical trials. Also, to gather information from the clinical research communities regarding best practices to identify, minimize, and describe risks prior to initiating pediatric cell or gene therapy clinical trials. Also, to gather information from the community

8

regarding best practices to identify and describe the potential benefit of a pediatric cell or gene therapy clinical trial, and to gather information from the community regarding best practices on obtaining consent and assent from guardians and pediatric subjects. And so with this backdrop, I now turn the conference over to Dr. Celia Witten, who is the Director of the Office of Cellular, Tissue and Gene Therapies in the Center for Biologics and she'll take you to the next step. Thank you so much and welcome. DR. WITTEN: Thank you, Dr. Midthun. So clinical trials in cell and gene therapy in pediatric populations posts special challenges, and Dr. Midthun has described some of the challenges and concerns in the area of pediatric clinical studies. Nonetheless, it's important to study promising investigational agents in children with proper scientific and ethical assessment of the study. One thing to note is the protection of the rights and welfare of children in clinical

9

studies is a shared responsibility between FDA clinical investigators, IRBs, and parents. The challenges are shared but the perspectives are different for each group, and it's these perspectives we'll be hearing about today. FDA concerns include minimization of risk, safety assessment, and an assessment of the appropriate population for first-in-man trials. FDA also asks for information on informed consent and the prospect of direct benefit for pediatric studies in these areas. As you will hear, FDA has also performed an assessment of sample IRB practices in the area of pediatric trials. For pediatric studies, there are regulations termed the Subpart D regulations that specifically provide safeguards for children in clinical studies. Background on FDA review of cell and gene therapy studies, the Subpart D regulations in pediatric clinical study research ethics will be provided in the first set of presentations. Researchers have similar issues to address as are covered in FDA review but from a

10

researcher's perspective. For a first-in-man study, when are the bench and animal data adequate to evaluate the risks and potential benefit? What is the appropriate population for first-in-man study? How can the study be designed to meet ethical standards but also contain the assessment controls or other features to allow for an interpretable and informative study that will help with the next step of development? How can information be communicated to parents, and how can therapeutic misconception be avoided? We will hear about some of the approaches taken to minimize this. IRBs share the responsibility to minimize risks and make sure that risks are reasonable in relation to anticipated benefits. IRBs are required to assess and categorize risks in accordance with Subpart D, and depending on the risk, perform an assessment of the prospect for direct benefit. Most IRBs approve clinical studies of cell and gene therapy products in the pediatric population under the criteria of grea

11

than minimal risk but prospect of direct benefit. How do they make these determinations? IRBs also review informed consent and assent materials and procedures. IRBs need to evaluate the patient population, the risk of study procedures performed outside of administration of the product. Ensuring the adequacy of scientific review is heightened in the area of cell and gene therapy trials for a number of reasons, including the complexity of these trials and often a lack of ample data on the product prior to introducing it to children. Parents of children who are research subjects in one of these trials share these concerns on a personal level. How can they be sure they're making the right decision or the best decision for their child? We will be discussing these and other questions today. We appreciate the thoughtful input of the speakers, panel members, and the public at this meeting. I'd like to just briefly give you an overview of the agenda for the day. There are three sessions. Session one will include general

12

background on FDA review, pediatric ethical issues related to product development, Subpart D, and pediatric research ethics. We'll have time for questions following each presentation but there will be no formal panel discussion. Session two starts with presentations from researchers in cell and gene therapy areas. There will be a panel discussion following the presentations. Some possible discussion topics for this session are in your folder. Lunch is on your own. There will be information at the registration tables about lunch options, both walking distance from here and also in the hotel. Session three will have presentations from IRB members, the current ethicists serving on the NIH Recombinant Advisory Committee, and a parent advocate. Following these presentations there will be a second panel discussion with a focus on the discussion topics listed in your folder of session three topics. We are looking forward to an informative day. Now I would like to introduce our first

13

speaker, Dr. Wilson Bryan. Dr. Bryan is the clinical branch chief in the Division of Clinical Evaluation of Pharmacology Toxicology in the Office of Cellular, Tissue and Gene Therapies. He's a neurologist and neuromuscular specialist who graduated from the University of Chicago, Pritzger School of Medicine. His postgraduate training includes an internship in internal medicine at Emory Grady Memorial Hospital and a neurology residency at UT Southwestern Medical Center, Parkland Memorial and a neuromuscular fellowship at Tufts New England. His academic interests include clinical trial design and neuromuscular disorders. DR. BRYAN: Thank you Celia. I'm going to talk today about the regulations and some challenges in the FDA IND review, particularly of pediatric studies. I'd like to start by introducing FDA's Office of Cellular, Tissue and Gene Therapies. The director you just met, Celia Witten, and the deputy director, Dr. Stephanie Simek; Patrick Riggins serves as director for

14

Regulatory Project Management; the Division of Cellular and Gene Therapies is led by Drs. Raj Puri and Kimberly Benton; Ellen Lazarus leads the Division of Human Tissues; and Mercedes Serabian and I serve as branch chiefs for Pharmacology/Toxicology, and Clinical Evaluation. In OCTGT, we now have 113 active pediatric trial protocols, including 19 first-in-man protocols. This graph shows that that number has been steadily increasing over the years and this last -- does not include the last three months of the year, so we'll expect to have a record number this year. Those 113 protocols are divided approximately equally between oncology indications and general internal medicine indications. Approximately 60 percent are cell therapies, and approximately 27 percent are gene therapies, and others are some combination of variance on that. I'd like to outline the therapeutic process, particularly the objectives of the therapeutic process, say a few words about IND

15

review and clinical holds, and with regard to pediatric studies, the Subpart D of the Code of Federal Regulations and some regulatory challenges with regard to pediatric studies. The objective of therapeutic development is to provide evidence that drugs, including biologics such as cell and gene therapies, are safe and effective for a specific indication. The process of therapeutic development should start with some knowledge or understanding of the disease. Based on that knowledge, there are generated hypotheses regarding interventions that could have an effect on the disease process. Those hypotheses lead to drug discovery where academic investigators or pharmaceutical companies will do in vitro studies looking at different candidate molecules to see which ones seem to have the best likelihood of being successful and should be taken forward into nonclinical studies. The objectives of those nonclinical studies include assessments of toxicity, biodistribution, carcinogenicity, and often we ask for proof of

16

principle studies in the animal models. The data from these nonclinical studies is essential to guide the design, including dosing, the population, and the monitoring for the subsequent phase one studies. Then the phase one studies have as their objectives assessment of safety, tolerability, maximum tolerated dose, pharmacokinetics and -- pharmacokinetics for small molecules, of course -- and if feasible, assessments of activity and efficacy. The data from phase one is used to guide the dosing and monitoring of the subsequent phase two studies, and then the phase two objectives include trying to determine the dose, route, regimen, population, endpoints, estimated magnitude of effect, and using this information, to guide the design of subsequent confirmatory phase three studies. And of course, the phase three study objectives are to provide the evidence of efficacy and safety to support a marketing application, either a new drug application, NDA, or biologics licensing application, BLA.

17

So therapeutic development should be an orderly rational process where each step of development builds on the previous step and the ultimate likelihood of success depends on how well each step is done. When a sponsor wants to study a new drug or biologic in humans, the sponsor submits an investigational new drug application, IND. At the FDA, this IND is looked at by a review team that includes a project manager, a reviewer from our CMC division -- the chemistry manufacturing controls -- a reviewer who look at the nonclinical pharmacology toxicology issues, the animal studies, a clinical reviewer who looks at the proposed protocol. And depending on the stage of drug development and the individual protocol, others may be involved, including often a statistician, sometimes an epidemiologist, site inspectors, patient representatives, and there are many others involved. These are just some of the folks at the FDA who may be on the review team for IND. The objective of the FDA review in all phases of the investigation is to assure the

18

safety and the rights of the subjects. Sometimes the FDA decides that the IND needs to go on clinical hold. A clinical hold is an order issued by the FDA to the sponsor to delay a proposed clinical investigation or to suspend an ongoing investigation. The regulations state that the FDA may place a proposed or ongoing phase one investigation on clinical hold if it finds that human subjects are or would be exposed to an unreasonable and significant risk of illness or injury or if the FDA finds that the clinical investigators named in the IND are not qualified. I will say that we actually seldom use this particular hold issue. But sometimes we find that the investigative brochure is misleading or erroneous or incomplete. And the FDA may find that the IND does not contain sufficient information for us to assess the risk to subjects of the proposed studies. The most common clinical hold issues, particularly for phase one and phase two studies, are that human subjects are or would be exposed to

19

an unreasonable and significant risk or that the IND does not contain sufficient information for us to assess the risks. So early in clinical development, the most common hold issues focus on safety. One issue is what constitutes an unreasonable risk. And the assessment of whether a risk is reasonable includes consideration of available in vitro, animal, and clinical data; considers the objectives of the study, the study population such as the disease severity, the prognosis, the availability of treatment alternatives. For example, risks that are acceptable in a population with a late stage malignancy, patients who may have no alternative therapy available. The risks that are acceptable in that population might be unacceptable in a pre-symptomatic population with a genetic defect. So the population has to be considered. Phase one review considerations include the overall design of the study. Particularly we consider is there proof of concept data. Is there

20

sufficient evidence of potential efficacy to justify the risks? With regard to the study population do the potential benefits justify the risks for the study subjects in the experimental arm and in the control arm if there's a control arm? Would a different study population have lower risk and still achieve the study objectives? Sometimes sponsors include stopping rules and/or a data safety monitoring board to ensure that appropriate controls are in place to stop the study if the adverse events suggest that it is an unreasonable risk. With regard to dose, are the proposed starting dose and the proposed maximum dose both justified by the available nonclinical and/or clinical data? Is the rate of dose escalation acceptable considering the available nonclinical and clinical data? And is the regimen appropriate for the stage of drug development? All of these are phase one review considerations. FDA also considers the monitoring plan, the intervals, both the intra-cohort intervals and the inter-cohort

21

intervals. The intra-cohort intervals refers tothe time between sequential subjects within thesame cohort; the inter-cohort interval is the duration for monitoring a single cohort, a group of subjects, prior to beginning enrollment in the subsequent cohort. The importance of these intervals was highlighted by the unfortunate experience with TGN 1412. TGN 1412 was studied in a first-in-man study in 2006 in England. And this study enrolled normal volunteers, enrolled eight normal volunteers. Six were enrolled to receive TGN 1412 and two were randomized to receive placebo. The intra- cohort interval for these subjects was a matter of minutes so that all eight subjects sequentially treated had received the treatment, either TGN 1412 or placebo, within a few hours. And within minutes after the last of those eight subjects received their dosing, the first subject developed headaches and fever and other symptoms. And all six subjects who received TGN 1412 developed cytokine release syndrome and had to be

22

hospitalized. In retrospect, one of the criticisms of this trial design was that the intra-cohort interval was not sufficient for monitoring the first subject before treating the second subject. To minimize risks, it's important to have these intervals be judicious. So one question for the review is are these intervals justified by the available nonclinical and clinical data? What are the appropriate intervals when there is concern about long-term toxicity? With regard to the monitoring procedures themselves, are the monitoring procedures sufficiently sensitive and frequent todetect adverse events? Do those monitoring procedures have unacceptable risks such as if there's a biopsy procedure or if some monitoring procedure would require anesthesia. What duration of follow up is necessary when primary toxicity may be long term as in some cell and gene therapy studies? We do have available a guidance for industry on gene therapy clinical trials that talks about observing subjects for delayed averse

23

events. So the review considerations with regard to monitoring and dose and overall study design relate both to pediatric studies and actually to adult studies as well, but for children there are additional safeguards which are described in the Code of Federal Regulations Subpart D. And Dr. Nelson will provide more details about Subpart D in his presentation. But I will mention that Subpart D, Section 50.52 describes the regulation of clinical investigations involving greater than minimal risk but presenting the prospect of direct benefit to individual subjects. OCTGT believes that most cell and gene therapy trials have more than a minor increase of minimal risk, and to provide evidence of the prospect of direct benefit under 50.52, OCTGT often asks IND sponsors to provide proof-of-concept data from nonclinical and/or previous human studies. If a cell or gene therapy study is not approvable under Section 50.52, the study might be approvable under Section 50.54; however, OCTGT

24

does not refer the study for consideration under 50.54. And again, Dr. Nelson will provide more details about these sections. We do encourage IRBs that are considering referring a clinical investigation under 50.54 first to discuss with the sponsor whether there are appropriate modifications to the protocol that would allow the clinical investigation to be approved under another provision of Subpart D. IRBs should send referrals under 21.50.54 of clinical investigations to FDA's Office of Pediatric Therapeutics which coordinates the review. Some challenges. With regard to pediatric studies, some challenges for the sponsor and for us as regulators, how to minimize the risks while maintaining a prospect of direct benefit and an acceptable risk-benefit- ratio. Considerations. What constitutes sufficient evidence of a prospect of direct benefit, such as proof of concept data? Should nonclinical proof of concept studies be replicated

25

by independent groups? What is the appropriate study population? Is there an adult population that would be sufficiently informative with an acceptable risk-benefit ratio? What study procedures are acceptable? Consider the risks of the procedure. The benefit, if any of the procedure to the subject and the value of the resulting data. For pediatric studies, OCTGT asked the sponsor to describe the following: How the study meets the requirements of Subpart D and why the study of children is scientifically necessary. Considerations for the FDA: Can the study risks be sufficiently minimized so that it would be appropriate for a pediatric study to be a first-in-man study for a new experimental cell or gene therapy. And when adults are studied first before proceeding with the study in children, what is the appropriate number of adults for the study? What is the appropriate duration of monitoring of adult subjects before proceeding in children? And

26

what data are necessary to provide sufficient evidence of safety and/or proof of concept before proceeding in children? These are questions that we hope to get some insights from this workshop on with regard to how IRBs function. That is, do some IRBs defer assessment of the scientific and/or human subject protection issues to other entities, such as the FDA or the NIH's Recombinant DNA Advisory Committee? And how do IRBs determine whether early stage gene or cell transfer studies have a prospect of direct benefit? If you have questions regarding my presentation, you can reach me at this e-mail address, but if you have questions regarding the regulatory process and a specific IND, I recommend that you contact Dr. Riggins, who is our head of the Regulatory Management Branch. And I do want to mention that I have the privilege of working with an outstanding group of physicians who serve as reviewers for the clinical protocols and they are listed here. And I will stop there.

27

DR. WITTEN: Thank you. So are there -- DR. BRYAN: I'll take easy questions. DR. WITTEN: Yeah, we'll take some easy questions now. If there are questions from the audience, there's a mic there. SPEAKER: Hello. Is this on? Yes. I have a question regarding DSMBs. Is there a requirement that these kinds of studies have DSMBs? DR. BRYAN: The requirements are that the subjects be protected and that the risk be reasonable. There is no specific requirement that every study have a DSMB, that every study have stopping rules, but it is important that someone is watching the adverse events to be sure that if there becomes evidence of an unreasonable risk, that the study is stopped. So, sometimes the FDA will ask for DSMB, sometimes will ask for stopping rules, sometimes will ask for both. But there's no specific requirement that every study have a DSMB. SPEAKER: Yes, thank you. That was an

28

informative presentation. Your last slide was interesting. If I could ask a specific question on that. In that you state that when adults are studied before proceeding with a study in children, the FDA reviews a number of other considerations. And I'm curious if FDA has at all considered looking at the definition of a pediatric population. Has there been consideration, for example, of hormonal status to determine whether the patient is pediatric or if it's an adult. And then further, if there's a concept of looking at an age range. So rather than comparing an adult to a pediatric, what about using the, if you will, the pubescent or patient population that is bridging between the adult and pediatric, if you will. DR. BRYAN: And when you say bridging population, for example -- SPEAKER: For example, patient population, say, to 18. Have you looked at that or have sponsors considered that as a way of

29

finding a road to, particularly if there's a younger patient population which may be at higher risk and the adult may not actually model properly, would that -- would that status be useful to determine appropriate patient selection? DR. BRYAN: Right. So I think some of those questions will be addressed in Dr. Nelson's presentation later this morning. Or Dr. Nelson -- let me ask Dr. Nelson to address that question then. DR. NELSON: The problem, I mean, from a scientific perspective there may be some validity there. The problem is that a child is specifically defined as someone who is not of a legal age to consent. And so the procedures that you're likely proposing in these trials are not ones that they could consent to on their own until their either 18 or 19 depending on the jurisdiction. And so basically, Subpart D would apply to that population independent of the scientific merit of the transitional population issue.

30

SPEAKER: Just one quick follow on then, Dr. Nelson. So you use a legal definition then for that pediatric is generally the review? DR. NELSON: The definition, strictly speaking, would defer to the local jurisdiction. So if, in fact, in your state adolescents had the legal right to consent to gene transfer and stem cell research, which I doubt, that would certainly be possible. I know of no jurisdiction that that would be the case. Usually it's things like access to contraceptives, treatment for mental disorders or sexually transmitted diseases, those kinds of things. SPEAKER: Thank you. SPEAKER: My question is about appropriate animal models, whether FDA has guidelines. And if so, if the guidelines are made available to IRBs for the assessment when they do their review. DR. BRYAN: Go ahead. DR. WITTEN: I think that is a challenging question. And we do have guidance,

31

you know, on a number of specific disease areas but not across the board about animal models. And, you know, and pediatrics, per se. I think that would be challenging to do. What we look at when we review is we have a pharm tox group that reviews the animal data for safety and for, you know, evidence of or possibility of evidence of benefit. But I think that would be a good question for, you know, the IRBs as to how they make that determination also. So we'll be hearing about that in the afternoon. SPEAKER: Thank you. If I can just follow up, knowing the level of expertise on the IRB at the medical school where I teach, I am wondering whether in these first-in-human trials the FDA upon request would be willing to provide some of the analysis that it has used in making that assessment. DR. WITTEN: That's not, well, right now that's not part of what we do but I think one of the things we're hoping to get out of this workshop is to hear some suggestions about what

32

might improve the process. So, you know, that might be, you know, one worth discussing. I mean, we are hoping to hear suggestions about what might improve the process. I don't know that we could do that per se but there might be some ways to help address some of the challenges that we could consider doing. SPEAKER: Thank you. SPEAKER: You mentioned in your slides that when the FDA puts a study on clinical hold it may be for reasons relating to human subjects being exposed to unreasonable risks or clinical investigators named on the IND are not qualified and some other bullet points. Would FDA make that information available to IRBs so that they'll know what FDA's concerns were that resulted in the clinical hold? DR. WITTEN: Again, I think that, you know, is something -- we'd be interested in hearing what would be helpful to IRBs. Certainly, when we correspond with the sponsors who have submitted an application if we put something on

33

hold, they receive a letter and the letter outlines the concerns. And the IRBs could certainly request those letters from the sponsor. That's something they could do right now. I think a lot of them do. And those outline in detail why the studies are on hold and what the further information is that we want. Okay. DR. BRYAN: But as I said, it is unusual that we use that specific hold issue. DR. WITTEN: Oh, I'm sorry, I didn't realize you were asking specifically about investigators not being qualified; I thought it was in general. Yes, that's unusual. Okay. I think we'll go onto the next presentation. Thank you, Dr. Wilson. And now I'd like to introduce Dr. Steven Joffe. Dr. Joffe is assistant professor of pediatrics at Harvard Medical School and a pediatric hematologist oncologist at the Dana Farber Cancer Institute. He attended Harvard College, received his medical degree from University of California at San Francisco and his

34

public health degree from U.S. Berkeley. He trained in pediatrics at UCSF and undertook fellowship training in pediatric hematology oncology at DFCI, Dana Farber Children's Institute Children's Hospital in Boston, and his clinical work is in the area of stem cell transplantation in children. He serves as the Dana Farber Cancer Institute Hospital ethicist. Welcome, Dr. Joffe. DR. JOFFE: Thank you, Dr. Witten. So, I'm not from the FDA, obviously, so I'm not going to speak from an agency perspective but just try to review some of the general ethical issues and ethical considerations in pediatric clinical research with some particular attention to the kinds of studies that we're talking about today. So just one disclosure. I am a paid member of a data safety monitoring board for the Genzyme Corporation, not directly relevant to anything I'm going to talk about today but in the interest of complete disclosure. So what I'd like to do is develop with you a systematic approach to evaluating the ethics

35

of a research study involving human subjects generally. And then to spend some time trying to understand the particular ethical challenges of pediatric research. And we'll focus this in particular on the context of early phase trials involving novel interventions, the kind of things that we're focused on in this session. So just as a case to sort of get us started and focus our attention -- and by the way, I don't mention this case in any way to suggest that it's a problematic study, just a challenging study to design and conduct. This is a study that was published in JAMA a couple of years ago titled -- the article was titled "Autologous Non-myeloablative Hematopoietic Stem Cell Transplantation in Newly Diagnosed Type One Diabetes Mellitus." And I'll tell you about the details of the study on the next few slides. This involved subjects who were eligible for the study if they were ages 12 to 35, so going down into the pediatric age range, although not young children. To be eligible, adolescents or

36

young adults had to have a diagnosis of Type One diabetes mellitus made during the previous six weeks and had to have no prior episode of diabetic ketoacidosis. Actually, that was not true. When they opened the protocol, the first subject did have diabetic ketoacidosis and there was no sign of any beneficial effect from the study, so they altered the study for the second and subsequent subjects so that patients with DK were not eligible. An example of close monitoring. There were two phases of the study. The first one involved stem cell mobilization. Remember, this is autologous hematopoietic stem cells. This was done with cyclophosphamide, a high dose of a chemotherapeutic, two grams per meter squared, followed by granulocyte colony stimulating factor to help mobilize hematopoietic stem cells into the peripheral blood. And then leukapheresis where peripheral blood was circulated through a machine and the white cells were taken off and separated for lateral reinfusion.

37

The second phase involved immunosuppression of the subject and then rescue with the stem cells. The immunosuppression was done with more cyclophosphamide, 200 milligrams per kilogram. This is actually about six grams per meter squared, so you can see a higher dose than in the first phase. And then anti-thymocyte globulin, antibodies against T cells. This is rabbit derived antibodies 4.5 milligrams per kilogram and then reinfusion of the stem cells. This is a fairly standard stem cell transplant sort of regime that would be used clinically, for example, for severe aplastic anemia in a patient who is getting a match sibling donor allogeneic bone marrow but used in this context for severe immunosuppression of the subject in the hopes of killing off the cells that are -- the immune cellsthat are causing the diabetes. This protocol obviously has many risks. These include allergic reactions, nausea and vomiting, infectious complications, acute infectious bacterial and fungal infections during

38

the period of neutropenia and opportunistic infections because of the immunosuppression, painful mouth sores because of the high doses of chemotherapy, bladder bleeding because of the cyclophosphamide. Some risk of long-term infertility with these doses of cyclophosphamide. It's certainly not a sure thing and it's particularly an issue for males. And then finally, an important omission for this slide, probably in the best centers -- and this, by the way, was done in Brazil at a very good center -- risks of death probably on the order of a couple of percent. So real measurable risk of death from the high doses of chemotherapy and immunosuppression used here. The protocol endpoints, the primary endpoints were safety endpoints, transplant mortality and transplant morbidity. What's the adverse effect profile of this regimen? And then a surrogate efficacy endpoint of changes in exogenous insulin requirements. Everybody who went on the protocol was insulin dependent at the

39

time that they went on the protocol and the question was what happened to their insulin requirements over time. There are many scientific challenges, and you've just heard about a way of thinking about those in designing this study, articulating the rationale for this study. What are the preclinical data that support going ahead? What are the prior human studies that support doing this? And here they analogize from lesser immunosuppression used to try to modulate the course of diabetes and also from stem cell transplantation regimens used for other severe autoimmune diseases. Who's the appropriate study population? Again, from a scientific point of view. What exactly should this intervention look like? What trial design should be used? Should we use some sort of dose escalation format? Should there be controls? Is there any role for randomization, placebos, et cetera? This was a single arm study without dose escalation. And then what are t

40

correct endpoints? The safety endpoints and the efficacy endpoints and it's worth breaking efficacy endpoints down into surrogate endpoints and then real clinical endpoints that would convey evidence of clinical benefit. There are also many ethical challenges in designing this study. How do we fairly select the study participants? Who belongs in this study? Are there populations for whom there would be a much lower chance of benefit? In which case you might want to exclude the populations for whom there might be excessive risks or burdens, in which case you might want to exclude them on those grounds. How do we assess, minimize, and justify risk? That's been a big part of the conversation so far. Appropriately identifying the benefits from being in the study. This is the first time at least in the published literature this has been done. So what do we say -- how do we estimate the benefits that may come from this protocol for those who take part given that there's no close

41

analogous experience? How do we ensure informed consent for those who are taking part? And also, how do we negotiate tensions between our roles as physicians and investigators? All of these participants were new onset diabetics who were looking for medical care and there they come into a center where they're offered this research option. And so the people who are doing the offering are both physicians and investigators and how do they negotiate the tensions between those roles? There are also many challenges that are specific to the pediatric parts of this protocol. The vulnerability of children, in this case adolescents, and I mean vulnerability in a number of senses. Physical vulnerability, psychological vulnerability, and a concept very important to pediatric ethicists, decisional vulnerability, the concept that they are not making their own decisions, somebody else is ultimately making the consenting decision for them. Are there any unique pediatric risks here? Physical risks that

42

are unique to the pediatric population? Neurocognitive risks? Perhaps not so much in this protocol but in many protocols.What's the ceiling of permissible net risks? By which I mean the excess of risk over benefit that would be permissible in a pediatric study and whose values count in weighing those risks and benefits? Is it the values of the reviewers at the FDA or the values of the IRB? The values of the parents? The values of the kids if they're old enough to begin to weigh in? The challenges of obtaining permission -- in this case from the proxies, which would typically be the parents. The question of what role the child should play in the decisions or the concept of assent. I'll come back to that. And then the issue of responsibility to the child over time. For example, if we were going to be following these children long term, what role does their consent begin to play as they get up to be the age of majority wherever they are? So to help us think systematically about

43

these kinds of issues, I want to offer you a framework that is drawn both from the Belmont Report and then from an important article by Zeke Emanuel and colleagues. And it's now about 10 years old in JAMA. And I want to start by just saying let's be clear about what we're talking about when we are talking about clinical research. The Belmont Report, which you're all familiar with, published in 1979, started out by defining the distinction between research and practice or research and clinical care. A very important distinction to make when we're talking about applying research-specific standards of review and ethical judgment. And they put the locus of this distinction in the intent of the person or the group that's doing the activity. So research is that class of activities that's designed to develop or contribute to generalizable knowledge.Research is those things that are about creating generalizable knowledge that will be carried forward to future populations.

44

Practice, on the other hand, is interventions that are designed solely to enhance the well being of an individual, patient, or client. And again, so the intent of the clinician who is doing the activity is what defines something as practice. And then once we define something as research, how do we judge that it is an ethical project that should be allowed to go forward? And this is where the Emanuel, et al., framework comes in, and these authors said that there are seven requirements for ethical research and that they should be applied in order. And so the first of these is the idea that research must have social or scientific value. This really comes down to is this an important research question. Is this an important problem to try to tackle? Now, sometimes that's a very applied kind of problem, and the example that we started with trying to find ways to alter the course of nuance of Type One diabetes, it's clearly a very important practical issue and I don't think anybody would

45

argue with that. Sometimes we're' talking about trying to answer scientific questions that don't have immediate clinical application or near-term clinical application but have the potential for adding to the stock of scientific knowledge and that we hope and expect will be beneficial down the line. And that's acceptable, too. But the question is does the question that the investigators are seeking to ask, is it important for them? The second requirement is scientific validity. This gets to the design of the study. Are the measures appropriate? Are the interventions appropriate? Are the statistical analysis plans appropriate? These sorts of questions. So that assuming that we've decided that the question is an important one, that we actually think that there is a high probability that we'll answer it with confidence, precision, and validity. The third requirement is the issue of

46

fair subject selection. This is the idea that we don't want to inappropriately burden certain individuals or certain classes of individuals, that there's obviously a historical legacy of certain groups of individuals being excessively burdened and we want to make sure that that doesn't happen. At the same time we want to be sure that the risks -- the benefits of research and the benefits of the knowledge that flow out of research are fairly distributed across populations. So we're thinking here both about fair distribution of burdens and also fair distribution of benefits.The fourth criteria in this favorable risk-benefit ratio. This is a challenging one, obviously. We've already gotten to some of the challenges here. The idea is that the benefits of a clinical trial or any research study or research procedure must justify the risks of that study. The problem is sometimes the risk-benefit ratio for the participant is clearly unfavorable, either for a study as a whole or for a particular

47

procedure within a study. So if you think about a research-related biopsy, something that's not going to be clinically informative, that will carry some risk to the individual. No clinical benefits to the individual. So if that's all we were thinking about we would say, well, you can't do that because the scales would weigh on the side of risk. So what we have to do is begin to weigh benefits to society on the scale along with any benefits that might accrue to the participant, and then that makes it acceptable to approve the protocol. And this is a very difficult weighing to do because you might say, well, these benefits and risks are incommensurable. Some apply only to the subject, others apply to society at large. How do we weigh those against each other? It's a difficult thing for investigators, subjects, reviewers, anybody, IRBs, to do. At the same time, we're told that we have to do this. This comes from the FDA regulations governing the criteria that IRBs must apply in

48

approving research. And in order to approve a study, an IRB must find that risk to subjects are reasonable in relation to the anticipated benefits of any two subjects and the importance of the knowledge that may be expected to result. So regulations tell us that we must be willing to apply these social benefits on the scale. The fifth requirement is independent review. This is a procedural requirement. This is typically enacted by -- through the means of IRB review. The sixth requirement is informed consent. And an important concept related to informed consent is the idea of the therapeutic misconception. This has been defined by many groups, but my favorite definition comes from the National Bioethics Advisory Commission from about 10 years ago, which said that the therapeutic misconception is the belief that the purpose of a clinical trial is to benefit the individual patient, rather than to gather data for the purpose of contributing to scientific knowledge.

49

Think back to the definition of research that we talked about in discussing the Belmont Report. The NBAC went on to say it's not a misconception to believe that participants probably will receive good clinical care during research. These two things are not inconsistent with each there. They can't -- it's not that they can't live together; but it is a misconception to believe that the purpose of trials is to administer treatment rather than to conduct research. So we have to keep reminding ourselves the purpose of what we're doing is to learn something. The purpose of what we're doing is to answer the protocol question. And finally, the idea that we must have respect for potential and enrolled subjects. This would entail things like when the results of a study are known, being willing to go back to our study population and offer them the results of the study if they would like to know those results. It would include things like keeping promises we've made, maintaining confidentiality, sharing

50

information, being very careful and open about conflicts of interest or financial relationships, those sorts of things. Now, everything I've said so far is about ethical challenges or systematic ethical approach to research in general. What about early phase pediatric research? And here I want to draw your attention to three things in particular. The first is the concept of therapeutic orphanhood. I'll say more about this. Second is how we think about risk-benefit assessment and justification in pediatrics. And the third is how we make decisions about children's participation in research. So what do I mean about therapeutic orphanhood? This is something that's been remarked on for at least 25 years, recognizing that there's less research or historically there was less willingness to include children in research and therefore, less known about appropriate treatments for children. Think about the prevalence of off label prescribing for kids

51

because of the number of agents that are using kids but not actually labeled for pediatric indications or that don't have pediatric safety and efficacy pharmacokinetic, et cetera, data. So this comes from the American Academy of Pediatrics Committee on Drugs' recent statement which argued that the performance of research studies to evaluate drugs in children is critical for determining the safety and efficacy of medications in children. I would broaden that beyond medications. Without this type of research, medication use in children will be limited to extrapolation from adult studies or off label use for indications that have not been studied in children, thereby putting children at increased risk of adverse effects. Without proper drug studies, children may not benefit from, and may even be harmed by drugs that are available to adults. Also, certain disorders affect children primarily necessitating drug testing on appropriately aged subjects. It is morally

52

imperative, therefore, to formally study drugs in children so that they can enjoy appropriate access to existing and new therapeutic agents. So the idea here, there's a moral imperative to actually do pediatric research so that we have safe and effective treatments available for children and we don't neglect them as a class. What about risk-benefit assessment for research? And here I'm going to focus particularly on studies that involve more than minimal risk in particular. So the first question that has to be asked by a reviewer, by an IRB, etcetera, is, is there a prospect of direct benefit from participation in this study? And if the answer to that is yes, then the next question -- and this, by the way, comes from FDA or alternately OHRP regulations on review of pediatric subjects, what you've already heard referred to as Subpart D. There are a couple of risk-benefit type determinations that must be made. First, the determination that the risk is justified by the

53

anticipated benefit to the subjects. And second, that the risk-benefit ratio is at least as favorable as that of alternatives. So one has to think about what else is available to this study population and how does the risk-benefit profile of what we propose in the study compare to the risk-benefit profile of those alternatives? And if those questions are answered in the affirmative, there's a couple of other criteria related to proxy permission, assent. But if those are answered in the affirmative, then the study can proceed with appropriate IRB oversight. If the answer to the question of prospect of direct benefit is no, then studies, in order to be approved by the IRB, must be -- must meet additional criteria. These include the fact that the risk is limited to a minor increase over minimal risk. The term minor increase is not actually defined in regulation. And then there are several other conditions that must be met. And if these conditions are met then a study can be approved under the appropriate subpart and can

54

go ahead with appropriate local IRB approval. If neither of these criteria can be met, and if the study is judged not to meet the criteria on the slide, then it can be referred for what you've already heard referred to as 50.54 for review. And there are pathways to approval of that study outside of this framework but those are used sparingly. What about decision-making for children? So children's enrollment in research requires, as a rule, permission of the legally authorized representative of that child, which in most cases is going to be a parent or parents. There are very rare exceptions to this requirement. And an important concept is that proxies have fewer degrees of freedom to decide for their charges than competent adults have for themselves. So I could decide -- let's think about a study that offered no prospect of direct benefit, a physiology study but did involve substantial risks. If that study were offered to me as apotential subject, I might be able to decide to be

55

in the study because I can say I'm willing to put myself at risk for the benefit of other people as an adult. On the other hand, if that same study were offered to my child, I might not be able to give permission for my child to be in the study because the idea is that parents can't expose their children or legal decision makers can't expose their children to studies that offer more than a certain ceiling of risk. So the degrees of freedom that a proxy has are much more limited than the degrees of freedom that a competent adult has to decide for herself or himself. And then where appropriate, assent is required. Now, assent is defined in regulation as a child's affirmative agreement -- these are obviously my bold letters -- to participate in a clinical investigation. Near failure to object may not be construed as assent. Now, we'll talk more about assent but I just want you to notice for a moment how demanding that standard is, that the child must ffirmatively agree I want to be in this study.

56

And if we really read the letter of the regulation, unless the child does that, if assent is required, he or she can't be in that study, even if the parent would like them to be in the study. So the regulatory algorithm here involves a presumption that the IRB will require the child's agreement, the child's assent, agreement to participate in research. But there's two pathways by which the IRB can rebut that presumption. The first of them is if the IRB judges that the study offers a prospect of direct benefit to the child, not just any prospect of direct benefit. I'll qualify this in a moment. Or alternately, the IRB judges that some or all children lack the capacity to make this assent determination. So what about this direct benefit exception? Well, assent is not required if the IRB finds that "the clinical investigation holds out a prospect of direct benefit that is important to the health or well-being of the children and is

57

available only in the context of the clinical investigation. So there's, as you can see, a lot of room for judgment in applying this standard. Right? What makes something important to the health or well-being of the children? And I think perhaps even more challenging, what does it mean that something is only available in the context of the clinical investigation? Or that the benefits are only available in the context of clinical investigation? And how broadly can we apply that? Certainly, we don't mean all studies that offer a prospect of direct benefit or the regulations wouldn't have presumptively required assent for studies with a prospect of direct benefit, but how wide or narrow is the subset of studies for which we can say there's a prospect of direct benefit and we will waive the assent requirement? The second pathway to waiver, of course, is the pathway of capacity. Here, the regulatory language says the assent of the children is not a necessary condition for proceeding if the IRB determines that the capability of some or all of

58

the children is so limited that they cannot be reasonably consulted. So what does it mean that capacity is so limited that they can't reasonably be consulted? A great deal of room for judgment. There is some guidance in determining whether children are capable of providing assent. The IRB must take into account the ages, maturity, and psychological state of the children involved. So some things to think about in deciding about applying this assent standard. Nevertheless, I would contend that there's a great deal of room for flexibility and judgment and more negatively you might say inconsistency on my IRBs in applying this standard. What do we know developmentally? Well, I would contend not enough. There's not enough conceptual clarity about what the standards should be in determining whether somebody has the developmental capability to give assent, and I think that there's a great need for more high quality empirical research in this area. There have been some influential groups that have

59

recommended that assent to research be obtained from children seven and older, and you often hear the age seven quoted. The National Commission in the 1970s recommended seven as the presumptive age of assent. This was importantly not put into federal regulation. There is no age of assent in federal regulation. There are some local laws that would require age of assent at seven for some kinds of studies, in particular California legislature for certain kinds of studies has a requirement that assent be obtained from children who are seven or older. However, I would suggest based on research quoted here and others that it's likely that few children who are under age nine to 10 -- I wouldn't say no children but few children -- can comprehend complex studies. One thing I want you to note is that there's a vast difference between the capability to understand a very simple study,a onetime blood draw, answering a few questions, that sort of thing, and the kind of complex, biological studies that we're talking about today.

60

At the other end of the scale, under optimal circumstances, I think the evidence suggests that many teens who are age 14 and above can probably actually meet adult standards of comprehension. And there we don't even need to think about capacity to assent because the capacity to consent -- not the legal right to consent but the capacity to consent is probably there. And so the transition time or the gray zone where we struggle appears to be somewhere in the age range of 9 to 14. What do we do with these kids? What role should they play in decisions? Do they get a veto over their parents if they don't want to be in a study but their parents would like them to be in the study? Those are some of the hard and unsolved questions. So to summarize, pediatric research, and I would say in particular that research involving very novel interventions, raises many ethical challenges. These include the sealing of permissible net risk, determination of when there is a prospect of direct benefit, the validity and

61

acceptability of proxy permission, and a particularly challenging one, the possibility of child assent for these kinds of studies. And even more than in other types of research, these kinds of studies really demand a rigorous and systematic approach to ethical assessment. Thank you for your attention. (Applause) DR. WITTEN: We have time for a couple questions from the audience if there are any. SPEAKER: Could you expand a bit on your sense of the ethics of parental permission? One conventional standard is the parental decision-making in general should be done on the basis of the best interest of the child. And some have viewed that as in the well known or infamous Grimes versus Kennedy-Krieger case as excluding from parental consideration societal benefit or benefit to the next generation of pediatric patients. Could you comment on that issue? DR. JOFFE: Well, you know, if we held to that standard we would really not be able to

62

enroll children in any research that involved more than minimal risk without the prospect of direct benefit, or ultimately, studies -- in any research that children were involved in, any risk would have to be outjustified or outweighed by the prospect of direct benefit. And as you know, a lot of research that's done with children could not be done under those sorts of very rigorous standards. I think the federal regulations, which do give a couple of degrees of freedom more to approved studies is a much more reasonable approach. I think -- and there are other aspects of life where children -- where adults can give permission to their children to do things that may not have direct benefit to those children but may entail some risk. Children involved in charity events, for example. We wouldn't want children to be, you know, climbing Mount Everest for charity but we allow them to do walks for charity where they're benefitting other people and walking five kilometers may entail some risk, even just things

63

like crossing the street. So I think in much of life, children -- parents are allowed to expose their children to some risk for benefits that aren't necessarily directly for those children. And I don't see why we should have a different standard here. If we had more time we could talk about the McCormick-Ramsey debates back in the 1970s where Paul Ramsey, a prominent theologian, said any child involvement in research must be justified by the benefits to that child. And Richard McCormick, another prominent theologian, said children should be -- there's a presumption that children should be involved in research that entails a very modest risk because they ought to want the benefits that come out of that for other children. And that's obviously a long debate that we won't be able to get into now but I think it actually captures the essence of your question. So, thank you. It's a great question. Yeah. SPEAKER: Maybe my question is the same. Could you provide any other examples of direct

64

benefit? By direct benefit, do you mean directbenefit to the population of children or direct benefit to the children in the study? Or -- DR. JOFFE: So, interpreting the regulatory language, what we mean by a prospect of direct benefit is a prospect of direct benefit to those children who will be in the study. And we can even talk about what counts as a direct benefit versus an ancillary benefit or an indirect benefit. And I think with direct benefits we're really talking about medical, clinical, or psychological, or other types of sort of benefits that are entailed by the mechanism of whatever it is that's being studied in the trial. So, for example, an incentive, a child who was offered, you know, $50 to participate in some study that involved a couple of hours in a psychology lab. That $50 incentive would not count as a benefit, at least a direct benefit to that child for regulatory purposes. Does that answer your question? SPEAKER: Maybe I can talk to you later.

65

It's still confusing to me because, as you pointed out correctly, the purpose of the trial is the research. It's not to have any particular benefit to the subjects. And yet, it must have a direct benefit before we can go forward. DR. JOFFE: So I think we can distinguish between the purpose of the trial, which is to answer our research question, the benefits of which will accrue to other populations in the future and the possibility -- the question of whether those who take part in the trial do or don't have a prospect of benefit from taking part. So a recent phase one study reported in the New England Journal for lung cancer, the purpose of that study was to find out about the tolerability dose, et cetera, pharmacokinetics of that agent. The fact is, quite a number of people who took part in that study did, in fact, benefit. So did -- even though the purpose of that study was a safety purpose, did it offer a prospect of direct benefit? In that case, adult subjects. I think the answer was yes

66

SPEAKER: Thank you. Dr. Joffe, thank you again for yet another thoughtful presentation. You know, Dr. Emanuel and his group have expanded the seven bullets of what makes clinical research ethical, to include community. To make it eight. And I wonder if you might comment on the involvement of the community in these discussions because, you know, whether it's central commercial IRB that's providing oversight or the central review models that CPSAs are considering today, it seems to me that the local IRBs are probably not in a position to provide the type of oversight of the nature that you're talking about that would be truly beneficial across the board. So I wonder if you might share your perspectives on what you think about how community involvement, community discussions ought to take place with regard to the regulatory folks regarding review of this type of research. DR. JOFFE: So that's a question that could occupy half a day. And actually, I hope

67

that it can be put on the agenda for one of the panels this afternoon because I think it's worth a bit of a roundtable discussion. I think the first question one has to ask is who is the community that we are talking about? And so if there is a defined insular kind of a community, then the notion of really consulting with that community, involving them in the research, having them be really a partner in taking that research forward I think is critical. I have a supine Native American case that, you know, came to our attention in the last year is a really critical example of that where a group of Native Americans were -- their tissues were used for purposes well beyond their initial consent or at least what they believed to be their initial consent, and in ways that really, I think, violated both the letter and spirit of the relationship with the tribe. So that's an example of a case where there is a well- defined community, well-defined community representatives, and one can engage them in discussion and work --

68

and move forward together with them. I think for the -- for clinical trial-type situations, it's much harder often to define who the community is that one goes to. Is it the geographic community in which one's institution works? Is it representative advocacy groups that have an interest in what goes forward with those sorts of trials? So the first question is who is the community with whom one engages? If one can define that community, I think it is important at least to hear the perspective of representatives of that community to bring them in to offer them the opportunity to be involved in commenting on studies and in having an impact on what goes forward, what doesn't, and how this is. For many of the diseases that we're talking about today there really are orphan diseases. There are well organized advocacy groups. There's really a great deal of expertise I expect that many people in the room actually are here representing groups on behalf of particular conditions. And I would want those sorts of representatives at the table.

69

You know, the HIV experience taught us that back in the 1980s, how important it was for successful research and for scientific -- high quality research to bring knowledgeable advocates into the table and engage them in discussions. DR. WITTEN: Okay, thank you. Thanks again. DR. JOFFE: Thank you. DR. WITTEN: We're going to move on to our last speaker for the last session, which is Dr. Robert Nelson. Dr. Nelson is currently the pediatric ethicist lead medical officer in the Office of Pediatric Therapeutics in the Office of the Commissioner at the U.S. Food and DrugAdministration. After receiving his M.D. from Yale University, Dr. Nelson trained in pediatrics at Massachusetts General Hospital and Neonatology and Pediatric Clinical Care at the University of California, San Francisco. He has a Master of Divinity from Yale Divinity School, and a Ph.D. in the study of religion from

70

Harvard University. He has long-standing experience and interest in pediatric ethics, ethical research, and related issues. And I'd like to welcome him here to give this talk on the scientific and ethical path forward in pediatric product development. DR. NELSON: Thank you, Celia. And as you'll see, many of the themes that Steve touched on I'll be adding, if you will, a further specification. And I will say as an ethicist, even though I'll be presenting a lot of regulatory language, I don't want you to be misled that this is only about the regulations and not in a sense trying to flush out with more concreteness the sort of ethical frameworks that Steve has been presenting. And I will be commenting on some of the questions that were asked, particularly the prospect of direct benefit. I've got a couple of slides about how one might begin to approach that concept. Steve started with his -- notice the question from the American Academy of Pediat

71

about the moral obligation to conduct pediatric research. Over the past 15 years we have evolved from a view that we must protect children from research to a view that we must protect children through research. And clinicians have a professional obligation -- one could argue also a moral obligation -- to ensure that there are adequate data to support the safe and effective use of drugs and biological products in infants, children, and adolescents. The critical need for pediatric research on drugs and biologic products reinforces our responsibility to assure that children are only enrolled in research that is both scientifically necessary and ethically sound. Children are widely considered to be vulnerable persons who, as research participants require additional or special protections beyond those afforded to competent adult persons. Now, these protections have been framed in what we call Subpart D -- 21 CFR Subpart D. And I want to just walk briefly through the

72

structure of that and I will then begin to unpack that in a little bit more detail through the rest of my presentation. So there are four categories under Subpart D under which research can go forward that is considered both -- that is considered scientifically sound. The first is research not involving greater than minimal risk. And I'll give you the definition of that for your information in the next slide. The second is research that is greater than minimal risk presenting the prospect of direct benefit. You've seen the criteria that Steve presented in his slide. I will show you that again. The third category is greater than minimal risk, no prospect of direct benefit, but likely to yield generalizable knowledge about the disorder or condition. This is where the category of minor increase over minimal risk comes in which was mentioned by Steve in his slide showing you the prospect of direct benefit, and then the

73

decision tree, yes or no. And then the fourth, which I'll talk about briefly that's also been mentioned, is the category where the research is not otherwise approvable by a local IRB under these other three categories but is thought to be worth doing. It presents a reasonable opportunity to understand, prevent, or alleviate a serious problem affecting the health and welfare of children, and the IRB feels it can be conducted ethically, it can be referred for federal panel review, and then potentially move forward after that review. And I'll talk a briefly about that process. And then I won't be going into much detail around parental permission and child assent. Steve presented I think a quick summary of that. But, of course, to remind people that that's the other requirement that's required for research, and I give you the citations that you can find within 21 CFR 50 for those different categories. Now, minimal risk is the anchor, if you

74

will, to these categories. And here's thedefinition. The probability and magnitude of harm or discomfort anticipated in the research are not greater in and of themselves than those ordinarily encountered in daily life or during the performance of routine physical or psychological evaluations or tests. Now, I might just say as an aside, if you look back at the National Commission, they came up with this definition to try and capture the range of decisions that Steve alluded to that parents might make where there's no obvious direct benefit as we understand direct benefit, but capture the kinds of everyday decisions that parents make around child participation in different activities. And so the National Commission tried to present this definition as a way of, if you will, framing the scope of what I like to think of as scrupulous parent parental decision-making. In other words, what are the situations under which parents are making reasonable decisions? And to give you an example, what some

75

might argue could potentially be thought of as not reasonable decisions, ABC News last week gave a story on ATM -- ATV use, all-terrain vehicle use and, you know, 55 deaths around the country from ATV use with kids down to the age of seven. I'm sure some parents think that's reasonable. Obviously, some are participating in that and allowing their child to participate but the, if you will, draft of the news report was that that was thought not to be a reasonable approach for parental decision-making even though that's being done in daily life. Now, the additional protections of children, as Steven mentioned, we think about risk being balanced against both benefits and knowledge in the broad context, but in pediatrics when we specify that there's a limit to the risks that we could expose children absent direct benefit, which is where the language of minimal risk or minor increase comes, or the notion that the risks must be justified by anticipated benefit. So basically -- and this laser is not

76

great, is it? It gives you -- all right, we'll forget that. So basically, we have additional protections for children in research that fall into these two categories. If there's no direct benefit, the risk needs to be capped; if there is direct benefit, we need to evaluate the balance of risk and benefit and then set that in the context of available alternatives. And I'll be walking through some of the implications of this. Now, for the basic ethical framework, I think there's three principles we can identify that rest behind Subpart D. The first is that children should only be enrolled in a clinical trial if the scientific and/or public health objectives cannot be met through enrolling subjects who can provide informed consent personally, adults. And I'll walk through this. The second is the absence of prospect of direct therapeutic benefit to the children enrolled in the clinical trial. As I mentioned, the risks need to be capped to a low risk or to

77

what we would otherwise consider minimal risk or a minor increase of a minimal risk. And then the third principle is children should not be placed at a disadvantage. After being enrolled in a clinical trial, either through exposure to excessive risks or unreasonable risks or by failing to get necessary health care. I'm going to walk through these three principles in the context of the kinds of trials we're going to talk to today. The first is scientific necessity that I'll start with. So, the practical application of this principle is it impacts on the timing and the type of studies that are required. You heard Wilson walk through the kinds of considerations that can be given for phase one and phase two and phase three trials. Behind that is this notion of trying to decide what really is the scientific path forward to be able to bring a product from the bench, if you will, to the bedside and then into the clinic. The goal here is a public health benefit, and again, equitable selection. The

78

notion of equitable selection behind this is that we only choose children if we have to. If you look back at the National Commission's discussion, they were specifically applying equitable selection in pediatrics to the notion of using individuals who can consent first and working their way down, as opposed to just limiting it to gender and ethnic representation. Now, one caveat or rebuttal to this is a life threatening disease absent suitable alternatives, and so the notion that you would necessarily delay investigations in children is something that I'll speak to later. But that wouldn't necessarily be the conclusion that I would draw. So as you think of different pathways to pediatric licensure, there's really three general ways you can get something developed. The first is if there's a pediatric indication alone and there is no adult indication. And the challenge here, as you heard Wilson mention, is developing sufficient preclinical data to support the initiation of pediatric clinical trials, whether

79

developing some safety data perhaps use in adult populations for a different indication or developing a sufficient proof of concept around direct benefit to allow one to go forward. If you do have a pediatric and adult indication, the goal is concurrent licensure. If we think that licensure -- that off label use is inappropriate, our goal should be concurrent licensure. There may be times that that's not possible but there are other approaches to get there. One would be sequential development, either linear or staggered. What do I mean by that? Linear would be you do all the adults first, then all the kids. Staggered would be you've got your phase one adult and then phase two adult and you start phase one pediatric when you got your phase two, et cetera. And then the other would be parallel development. There may, in fact, be times where going forward in both children and adults are appropriate. But the issue still stands you need sufficient proof of concept around prospect of direct benefit to allow

80

that to move forward. And so as I mentioned, this is the challenge. The challenge is to develop sufficient data to basically move forward, either to demonstrate that the risks are low, which I think in this field, gene transfer and stem cells, is a bar, if you will, that has not been met as yet. Or to establish that there's sufficient prospect of direct benefit to allow one to move forward. So let's talk about this notion of limiting therapeutic risk. So a minor increase over minimal risk just briefly, these are the regulatory categories. What's key in here is that the study needs to be done in children that have a disorder or condition. Not only is there no definition of a minor increase over minimal risk other than if you look back at the National Commission saying it's slightly more than minimal risk, which we've debated how to interpret minimal risk, but they had this notion of disorder or condition, which is also not defined. Now, what's key here is the National

81

Commission considered minimal risk a fairly strict definition and they limit it to healthy children. Our regulations do not. That was admitted in the writing of the regulations, but most ethicists and at least every federal panel I'm aware of since the National Commission has recommended that minimal risk be interpreted against the life of healthy children. So it's a fairly strict standard. As you heard in Wilson's presentation, administration of experimental drugs or biological products is neither normal nor routine and it's not minimal risk. And therefore, the interventions or procedures must be limited to low risk if there's no direct benefit and to be done in children with a disorder or a condition. There is no definition of disorder or condition. There's one that's been proposed. This is the Institute of Medicine recommendation. This was developed and modified slightly by the Secretary's Advisory Committee on Human Research Protections as a recommendation. The key concept here is you

82

either have a disease, which is fairly straightforward, or you may be at-risk for a disease. And then finally, just to point out that there is harmonization at the international level at this idea. This is ICH GCP guidelines, which again talk about foreseeable risks being low and that it should be done in children with a disorder or a condition. So I would argue at this point that our Subpart D, as well as internationalized ICH GCP guidelines are harmonized on that point. But I think where most of the action is in this field at the moment is in first-in-children's studies and this issue of sufficient prospect of direct benefit to justify the risks of a clinical trial. And so let's focus on that for the moment. So I would argue that this category where we talk about the risks being justified by anticipated benefit to subjects and that the relationship of this anticipated benefit is at least as favorable as alternatives follows from this more general ethical obligation that children

83

shouldn't be disadvantaged by going into a clinical trial compared to what would be available to them outside of that clinical trial. And the challenge here is in first-in-children studies, how do we get there? Can we infer a sufficient prospect of direct benefit from animal studies alone to justify a first-in-children clinical trial? You won't be surprised; the answer will be sometimes yes, sometimes no. It depends on the details. So what is a prospect of direct benefit? And here are some thoughts to consider. So the first idea is that it's a direct benefit if it's my benefit and not your benefit. So if you're talking something that directly benefits me, it's not something that's going to directly benefit you. So we generally don't consider knowledge to the community to be a direct benefit. Now, one caveat for those working in Europe, they do use the language of direct benefit to the group. The reason they've done that is because there was a prior -- a prior legislation passed by the

84

European Parliament back in the 1990s that specified you can only enroll people in research that offers a direct benefit. And so in order to do what we call non-beneficial research, they started to develop language of direct benefit to the group. So don't get confused by that fact. They use the language direct benefit to the group to talk about what we call indirect and direct benefit to the individual about what we call direct. So there's agreement with different language. So the second is that it results from research interventions being studied and not from other clinical interventions. I'll talk about component analysis, but if we didn't have this notion, I could basically take a risky intervention and then pile in a lot of health care. So when we're talking about direct benefit, we're limiting that to the intervention that we're currently discussing and not to the entire protocol, although it's theoretically possible that every procedure in a protocol offers direct

85

benefit, and therefore one could say the entire protocol does. But you need to be very careful in using that. We then often use the word clinical to modify benefit to indicate that the direct benefit relates to the health of the enrolled subject, but the problem there is, I mean, Steve mentioned that we don't consider money to be considered a direct benefit. Strictly speaking, it is a directed benefit. We just don't consider it appropriate to balance money against the risk of research because, again, we could then just pay more and more and more and justify the risk absent any notion of direct benefit to the individual from a health perspective. So it would directly benefit me if every one of you gave me $20 after the end of this talk. (Laughter) It would. I'm not advocating that, but it would. Now, the other thing is that the direct benefit is based on the structure of the intervention. So you all can judge independently of my intention, whether the act that I'm

86

proposing will directly benefit that individual. So it's really based on the data that we have at hand. It's not a psychological state of mind. So that's a very important thing, that we can judge that benefit based on the data and not based on what I, as the investigator, profess is my intention. Now, the second point is the evidence for prospect of direct benefit must be weaker than the evidence for efficacy because if it wasn't it would be a circular definition and we would end up not being able to do any research because the threshold you would need for evidence would be the same threshold that you would need to prove that it works and that would just be impossible. So somewhere we're looking for an evidentiary standard that's less than the efficacy standard that we would at the result of the trial. So we might base it on surrogate endpoints but I'm using it here not as a surrogate that's a proven surrogate of efficacy. So it should be clear on my use here.

87

You know, certainly there could be evidence linking that surrogate to clinical efficacy which would be the gold standard, but there may not. But you need sufficient empirical evidence of prospect of direct benefit to justify exposure to the risks. And this is a complex evaluation. If you think clinically about how you make these evaluations, you're looking at the options, you're looking at the importance of that benefit, you're looking at the possibility of avoiding greater harm from the disease. And if you actually look back at the National Commission's discussion of this category of prospect of direct benefit, they drew an analogy to clinical decision-making and to the kinds of risk-benefit assessments that go into clinical decision-making, and they carried that analogy over into the research setting. Again, we'll have to be careful about therapeutic misconception and the like in reinforcing that, but the logic of evaluating risk and assessing the prospect of direct benefit was what they were referring to.

88

And so whether the experimental intervention offers the prospect of direct benefit first of all is also a separate question from whether there is sufficient probability, magnitude and type to justify the risks of the intervention given the overall clinical context. So it's a complex judgment. So one proposal that -- I can't say that this is an FDA policy. It's something that I and others within the agency have thought about is this notion of a sliding threshold. In other words, at what point do you have sufficient data to think you can move forward? And that is going to be set within the severity of the disease. So you think about something like structure. So if I can show a structural change, would that be sufficient? Well, I mean, it may or may not be sufficient, it depends on the condition that you've got. You may want to have some evidence of functional change. So if you'd been able to say myelinate a nerve sheath, maybe you need to show

89

that a nerve impulse goes from point A to point B. So that would begin to show function. Function could be based on mechanism of action. There's different ways of doing that. Targets, receptors, biomarkers, metabolic products, transgenic technology. I'm just offering these as examples. It's not limited to this, but that's where you begin to get some functional evidence of prospect of direct benefit. And then you go into clinical disease models. And there are some animal models that are very good replicates, if you will, of the human disease condition, either based on surrogate endpoints or the gold standard would be a clinical endpoint and the application of that might be the FDA's animal rule where you could actually have something marketed. And labeled and marketed just based on animal studies alone. An example there would be (inaudible) intra-inhaled inhalational anthrax, which was done in non-human primates for obvious reasons. Now, the other thing to keep in mind is

90

dose. When we think of animal studies and we think of dosing, we often think of establishing a no observed adverse effect level. But on the other hand, if we're trying to optimize that balance of risk and potential benefit, that might not be the right dose to start the pediatric trial at. So this notion of a maximum recommended starting dose I think needs to be taken into consideration where you might be going higher, recognizing that at the same time you could be potentially increasing some risk, you're also trying to optimize that prospect of direct benefit because this standard applies to the first child you enroll in that trial, not the last. Whereas adults, we can start down very low and we can work our way up. Pediatrics, we need to think that that first child has a sufficient prospect to justify the risk. And so that's the challenge, picking the dose as well. Now, I mentioned component analysis just to mention it again. Research protocols may combine nontherapeutic interventions with other

91

interventions, either that offer prospect of direct benefit or would be considered part of necessary health care. Oncology is an excellent example of that. Their protocols often include what they would otherwise receive for that condition, and there may be a very narrow research question and maybe simply as, you know, how do you get the T cells out of the stem cell product in some way whereas everything else is standard of care but it's all included in the research. So the risk of that experimental intervention as I mentioned must be justified from the prospect of benefit from that same intervention. And this is called component analysis. Now, one possible exception of this, although I don't have an example in mind, would be if you have a possible monitoring procedure that is made necessary only because of the benefit that's offered from the intervention that you're giving. It may be that could potentially exceed a minor increase over the minimal risk but the problem is if that's the balancing you're doing,

92

the prospect of direct benefit from an intervention needs to be sufficient to cover then the risk of your monitoring procedure. So if someone wants to say proposed biopsies that are invasive, for the purpose of simply monitoring therapeutic effect that offer no direct benefit, then you've got to feel that the prospect of direct benefit to what you're giving them in the first place would justify that risk. And that may be a bar at this point that's too high to meet. So what's the impact of this on placebo controls and sham procedures just to toss this on the table for discussion? First of all, a placebo does not offer a prospect of direct benefit. The risk of placebo is generally minimal if chosen well. There may be placebos that are chosen where that risk is not minimal but hopefully the risk would be minimal if chosen well. Generally, under those circumstances then the risk of a placebo-control group relative -- related to the risk of harm is from not receiving something that

93

you would otherwise have received. And so the notion here is then that risk of what you would otherwise not have received would then be restricted to no more than a minor increase over minimal risk. Now, I don't have examples in mind at this point for placebo use in gene transfer or stem cell products. In the drug area there might be approaches like randomized withdrawal designs for hypertension, randomized withdrawal designs for monoclonal antibodies for the treatment of juvenile idiopathic arthritis and other kinds of examples that could meet this with the inclusion of a placebo control. But where this might come in is the issue of sham procedures, which is often brought up, certainly in the adult world for these. These obviously do not offer prospect of direct benefit either, and thus the sham procedure can be no more than a minor increase over minimal risk, which is a standard which in pediatrics might preclude some of the more basic sham procedures that have been

94

used in the evaluation of these kinds of products in the adult experience. So let me briefly talk about referral for review. This was mentioned. These are the categories -- again, if an IRB can't approve it under those three categories, it can refer it. But first of all, it has to think it's good science and it has to think it's ethical to do. This is not a category where the IRB would say we an't approve it and where the investigator says then I want you to appeal to the federal government. The answer is the IRB still has to think it's good science and it's ethical. If they don't, they should just disapprove it. And then we hold a federal panel and hopefully move forward from there. The Pediatric Advisory Committee and the Pediatric Ethics Subcommittee is chartered to be able to do these reviews. You have the slides so I'm not going to go through these in the interest of time. And then the required findings again are that this is both scientifically possible. Now, that federal

95

panel could decide that it fits under the other three categories. Or it could decide it could move forward under 50.54 provided that it's good science, good ethics, and that there's appropriate permission and assent. And these are the references for guidance that you can get, both through the FDA guidance, and this is a harmonized process with OHRP. So let me -- Celia had mentioned briefly as part of our interest in this that we did a BiMo, which is, I guess, biological monitoring oversight. DR. WITTEN: Bioresearch monitoring. DR. NELSON: Bioresearch monitoring is what it stands for. And I'm just going to present a little information about this. This is obviously confidential so I'm not presenting confidential information. It's fairly high level sort of aggregate data, but we inspected -- we, meaning the staff, not I, myself, inspected 24 IRBs which represented 45 INDs or IDEs and over 52 protocols. We looked at 33 consent and assent

96

forms and the associated IRB minutes for that. Now, most of the studies were approved under 21 CFR 50.52, which is prospect of direct benefit. I will say the minutes often lack much robust discussion of the details of that. Usually just say what the category was. The consent forms generally had detailed discussion of the procedures and risks, did not promise benefit, and adequately differentiated research from clinical care. A couple of observations though is the benefits were usually framed as disease improvement without information as probability. So things like we hope that this intervention might improve your health. And so this language regarding hope was common. So we hope that this treatment will improve your condition but cannot be sure, which is a true statement, I think, but I'll raise in the next slide a question about this. The other is the research versus clinical care. In some studies, for example, the only research intervention was a cell sorting prior to

97

a bone marrow transplant of enriched or depleted subpopulation, but yet the consent document included all of the other information that was part of the standard of care for that transplant, was inncluded in the consent documents. And I think it raises two interesting questions for -- well, three interesting questions that could be a topic for discussion. What about the use of this language of hope in early phase consent forms? Does it accurately convey the likelihood of benefit to parents of children with serious medical conditions? And should information about procedures that are unmodified by the research, such as bone marrow transplant preparative regimens be included in the consent forms? And does this sort of obscure the research question that's at stake. And to what extent do either of these approaches exacerbate the potential for therapeutic misconception? I agree with Steve's definition. You know, the notion that you can somehow think that the protocol you're going into is designed for your benefit,

98

not that you hope there to be some benefit; that's a separate issue, but that it's designed to give you benefit specifically like it would be if you went to your personal physician. So again, I've covered what I think are the three principles behind the basic ethical framework, scientific necessity and appropriate balance of risk and potential benefit. And I look forward to taking any questions. Thanks. (Applause) DR. WITTEN: Okay. We have time for two or three questions before we go to the next session. SPEAKER: Can you hear me? DR. NELSON: Yes. SPEAKER: So the question that neither of you addressed -- neither you or Dr. Joffe -- is the issue of direct benefit in phase one trials of chemotherapeutic agents or small molecule drugs that are now very commonly being tested in pediatric populations. I mean, we never offer phase one studies with the intent of benefitting

99

the individual; we always say that it's to discover the dose or the route or whatever or the pharmacokinetics. How do you balance the approval of these studies which are generally approved under 50.52 in terms of prospect of direct benefit when for the most part there's no prospect of direct benefit, especially if the child is getting the lowest dose in a dose escalation phase one study? DR. NELSON: Well, as I mentioned, first of all, the dose needs to potentially be higher. But let me give you, you know, the devil is in the details. You look at the intervention and decide does the data support the possibility of it as a prospect of direct benefit? In my view, and this just is an aside, before I joined the FDA full-time, I was at Children's Hospital Philadelphia and the University of Pennsylvania, so the example is one I was familiar with. But take Leber's amaurosis and the gene transfer trials for the treatment of Leber's amaurosis. The dog data there in my view, and I

100

was, you know, as full disclosure, I was not involved in CBER and the evaluation of this in any way, but I supported the clinical team going before the RAC to argue based on the animal data alone. There was sufficient evidence of prospect of direct benefit to go directly into children. Now, they eventually decided to do three adults and three children in that but, you know, it really comes down to the data. Yes, you can do that. In oncology, for example, you might start phase one pediatric trial at 80 percent of the MTD of the adult dose. You won't start down at a low dose. So it's all about the dose. It's all about the data. It's all about the population you've selected. Often in pediatrics it'll be a population where there may be no other option. Severity of disease. So all of those considerations go into structuring it around prospect of direct benefit. So do I think it can offer prospect of direct benefit? Yes. Do I think we should promise it? No. Do I think we

101

should tell people that this will be -- offer a prospect of direct benefit? No. But can IRBs approve it under that if the data supports it? Yes. DR. WITTEN: Okay. Two more questions. SPEAKER: Yeah, good morning. I want to thank you for a lucid and scientifically informed discussion. I have one reservation and it's in your introductory slide in paragraph two of which you say, "We have evolved from a view that research regulation must be protective." I haven't evolved from that view. I think neither has FDA and your point is overstated. I don't minimize the importance of research. I don't minimize the importance of protection. DR. WITTEN: Yes, the woman in front. SPEAKER: Thank you for that talk. You've been discussing about the importance of calculating risk-benefit in the context of available alternatives. You know, what crosses my mind is in rare disease, unlike maybe oncology, we don't have textbook protocols and ways of doing

102

things and it's very varied across institutions and geography. How do you view available alternatives when you're looking at risk-benefit in these types of situations? DR. NELSON: Well, in any given context that's going to be a hard question to answer. Let me just make a couple of general comments. I mean, first of all, the challenge is what's the evidence base because simply because something might be provided doesn't necessarily establish it as an evidence-based alternative. And so that would be part of the challenge. And whether or not you would set that in the context of what would be considered innovative interventions that don't have an evidence-based but yet are being offered as part of clinical care I think would be a complex questions. But you're right. There is a lot of variability around that and it would have to be set within that context. DR. WITTEN: Okay. Last question, and then we have to move on to the next session. SPEAKER: Just in terms of scale, I know

103

most trials fall under 50.52. But I was curious about what percent of trials go onto 50.54? What's the outcome of that? What type of trials are they? Any comment on the 50.54 pathway? DR. NELSON: Yeah, briefly on that. I mean, there's no trials that I'm aware of that have come out of the stem cell and gene transfer community that have gone to a 50.54. I guess maybe with the exception, there was a sibling arm of a bone marrow transplant trial that was, I think CDER-regulated because it involved GCSF stimulation. The sibling arm went to a 50.54 because there was felt to be greater than a minor increase over minimum risk and no direct benefit to the siblings who were the donors. And because it was not defined as routine standard of care, but they were the objects, if you will, of a research question, the usual approach of saying that the donor was just being treated as clinical standard of care could not apply. So that example is the only one that comes close. Most of them that have been

104

referred, at least recently, have fallen into the use of healthy children as normal controls in stimulation studies of hormonal response, for example, where it was felt to be minor increase over minimal risk potentially given the risk profile, but the children did not have a disorder or a condition. And so those are most of the categories. Since 2003 when this was set up, I think there have been five. I could be off by one, but I think there's been five. So it's not used that often. So. DR. WITTEN: Okay. I'd like to thank Dr. Nelson again. And I just will mention that the transcripts for those meetings that he's referred to are actually on the web, on the Pediatric Advisory Committee website. So if you want more information you can look there. Now I'd like to introduce Dr. Daley, who is going to moderate the second session. Dr. Daley is the Samuel Lux professor of Hematology and director of Stem Cell Transplantation Program at Children's Hospital in Boston. He's also

105

professor in the Department of Biological Chemistry and Molecular Pharmacology at Harvard Medical School and investigator of the Howard Hughes Medical Institute, associate director at the Children's Stem Cell Program, a member of the Executive Committee of the Harvard Stem Cell Institute, and past president of the International Society for Stem Cell Research. His extensive research contributions and other contributions are described in the sketch in your folder, so I'm not going to repeat them here. Thank you, Dr. Daley. DR. DALEY: Thank you. Terrific. We have an excellent session scheduled with four leading speakers, four leading investigators in the area of cell, tissue and gene therapy. This is, I think, arguably one of the very, very exciting areas of science right now. After literally a couple of decades of laying the groundwork in gene therapy, we are now really on the cusp of having the ability to really recount the potential benefits because we're actually seeing -- we're seeing efficacy in a large number

106

of patients. And yet we're also still acknowledging and understanding the risks which are real. So this offers, I think, some very interesting prospects for discussion this morning. I think we also are riding the wave of tremendous enthusiasm around this new area of stem cell biology, which again once again raises complications because it certainly highlights the concerns for therapeutic misconceptions in the context of clinical trials. The speakers we have this morning are all well versed in these issues, and what we've asked them to do is obviously use their own science to illustrate the issues but to not necessarily focus exclusively on the science but rather how their experiences contribute to our ability to think in this area about the new prospects for clinical trials. So without further ado, let me introduce Helen Heslop, who will be our first speaker. Dr. Heslop is widely acknowledged to be a world leader in the biology and clinical translation of cell-based therapies. She's a leader in using

107

antigen-specific cytotoxic T cells for adoptive immunotherapy both in the context of malignancy and in restoring antiviral immunity post-transplant. Dr. Heslop is a professor in the Center for Cell and Gene Therapy at the Baylor College of Medicine, and I hope she's here. Ah, she is. Terrific. And we'll have presentations followed by short question and answer for each speaker, and then we will have an extensive panel discussion. We will also have a break in the middle of this. Great. DR. HESLOP: Thank you very much, George. So at our center we have 43 cell therapy studies that we've implemented in the last decade on 25 INDs. Some of these studies have targeted only adults or only children if they've been targeting disease that only occur in either population, but the majority of these studies have targeted both populations. And that's because our general philosophy has been that there should be equal access to research studies for both children

108

and adults. Our pediatric patients are treated at Texas Children's Hospital, which is a large pediatric hospital that serves a very diverse population in Houston, and children treated there have care by specialist physicians, nurses, and other support services such as child life specialists and social workers in a child-centered environment. I'd like to illustrate some of the issues by talking about two studies. One study is a diseased neuroblastoma that occurs predominantly in children, and this is a study where we were treating patients with high risk relapse disease with autologous T lymphocytes to recognize GD2 supported by an NCI program project grant. And the second approach that targeted both children and adults post-transplant where they received multi-virus-specific CTLs on a study sponsored by the NHLBI Specialized Centers of Cell Therapy Program. So the neuroblastoma study was developed

109

about a decade undergo and underwent its initial review in 2003 and 2004. And the rationality for this approach was that we were targeting children with relapse disease where there was no standard therapy and where there was a very poor survival. And this is a slide shown at the RAC review where you can see the survival after relapse in months in the patient population that we were targeting. The rationality for exploring immunotherapy approaches in this context was that this tumor has potential unique antigens that could be targeted and clinical results with monoclonal antibodies had shown activity. So the approach that we used was to transduce these cells with chimeric antigen receptors targeting the GD2 antigen expressed on the tumor cells. And this slide shows how chimeric antigen receptors are generated by making a construct that includes the antigen recognition domain of a monoclonal antibody with a signal transduction moiety, in this case CETA. And this chimeric antigen receptor could be transferred to T cells, which

110

would then acquire specificity for the GD2 antigen on the tumor cells. The primary objective of this study was to evaluate safety in these patients with high risk relapse neuroblastoma and our secondary nobjective was to evaluate the persistence in activity of these CAR-activated T cells. And we were comparing two populations of T cells -- activated T cells and EBV-specific T cells. And so we were generating these two T cell populations for each patient and transducing them with a CAR with the same specificity but which could be distinguished by a noncoding sequence which allowed us to discriminate the two populations on PCR. So subjects received both T cell populations and we could track each T cell population over time. Now, when we developed this study, the question as obviously was a clinical study justified? And we thought it was for the following reasons. Firstly, we had longstanding safety and efficacy data with EBV CTLs in humans

111

from studies that we did previously at St. Jude Children's Research Hospital where we transferred these cells to patients post-transplant and patients with Hodgkin's Disease. There was clinical experience with monoclonal antibodies targeting GD2, and we had preclinical safety data with the GD2 chimeric EBV CTLs which was predominantly in vitro data. And there was also preclinical safety data from other groups with other sorts of neuro and chimeric T cells. So during the IRB and RAC review in 2003, one of the major issues that came up was the risk of ancestral mutagenesis. And this was because this study was reviewed shortly after the first reports of T cell lymphoproliferation in the gamma chain SCID studies in France. So there was considerable discussion with both review bodies of differences between targeting stem cells in marrow and mature T cells of differences in the trans gene and the risk profile. We had data on our own clinical experience with gene modified T cells for 33 patients that we treated over the previous

112

decade, and there was a recent publication from Chiara Bonini where she summarized safety data in patients who had received T cells transduced with the NGFR. And another consideration was that these were patients with advanced relapse disease. Another issue that came up during the RAC review was our procedures for both consent and assent. Most children with neuroblastoma, even at the stage of relapse, are going to be under the age of 10. So with this cell therapy study, one issue is that consent is a prolonged process because the study has two components. We first have a procurement component where we get consent to manufacture the cells, and then after manufacturing, which can take two to three months, we have a separate consent for the infusion component. And that time for manufacture also gives us a lot of time to discuss the risks and benefits of the study with the patient and child. And often subjects and their parents are in communication with study staff during this time with questions about the protocol.

113

In discussion with the RAC, we also decided that we would use a child life specialist to act as an advocate for the patient. And we chose child life specialists because these are available on all the services at our hospital, they have experience with developmental issues in children, and they're also very used to acting as child advocates because they do that for routine clinical care of children in our hospital. Now, one other issue that came up during the FDA review in 2003 was that our clinical reviewer requested we report infusion reactions of grade two or greater, rather than the usual grade three to four, and also requested reporting of musculoskeletal pain. The reason for these requests was not clear to us at the time, but subsequently when the results of the GD2 antibody study were reported it became clear. And I think this just illustrates that the FDA sometimes has information that the investigators don't have because of their access to results in other trials. So on this study we treated a total of

114

19 patients. It was initially a dose escalation study and then we modified it to treat additional patients at the lower dose level to obtain information on whether there was a difference depending on the dose level infused. There was a split between males and females. The median age was seven but we had a range of 2 to 20 and one patient was obviously able to give their own consent as an adult. There were a total of 22 infusions of 44 lines and that's because three patients received an additional infusion after approval because they had a partial response. And we saw no dose limiting toxicities in this study. At the time of the infusion, eight of the patients had no evaluable disease because although they had relapsed they had attained a remission with other therapy. Four had active limited disease, three of whom had solitary bone lesions and one bone marrow disease, and seven had very bulky disease. Looking just at the best response in the 11 patients with active disease, three had a complete remission, and there was one

115

involving bone marrow disease and two isolated bone lesions. One had a partial response, one had stable disease, and two additional patients had tumor necrosis but which didn't reach criteria for assigning it to be PR or CR. So the conclusion of the study which was comparing these two T cell populations transduced with chimeric antigen receptors was that we saw no severe or dose-limiting toxicities. We saw clinical responses, including resolution of active disease with an overall response rate of 47 percent in patients with active disease. And in following patients longer term, these first generation transduced T cells have been detected for an extended period of time up to a couple of years, although at low dose levels. So our future directions with this approach are to look at alternative ways to increase the in vivo expansion by giving the cells more closely after chemotherapy or with cytosine infusions and Crystal Lewis, who is leading these studies, is currently developing a phase one study

116

in both EBV seropositive and seronegative patients with neuroblastoma to evaluate activated T cells after chemotherapy. And we're also looking at translating to other tumor types that also expressed GD2 such as melanoma. Now, this is a study where the patients received T cells transduced with retroviral vectors so they require 15 year long-term follow up. And at present we have nine patients on this study who are between one and 15 years, with the longest patient just coming up for their 5-year follow up next month. And so far there have been no adverse events attributable to gene therapy. This long-term follow up can be challenging because you're following the patients at the stage when your grant funding has ended and you no longer have the resources to contact them long term. And my longest experience with such follow up has been in the very first gene transfer study I did which was protocol number 38 that was initiated in 1993. And this was undertaken at my previous institution, St. Jude Children's Research

117

Hospital, where we administered EBV-specific CTLs to high risk recipients after mismatched or unrelated donor bone marrow transplant. And in the first 26 patients we marked these cells so we could track their long-term patient persistence. And following these 26 patients, 10 have died, 8 in the first 2 years of follow up of other infections or relapse of their primary malignancy. In long-term follow up, the patients who have died between 2 and 15 years have actually died of trauma in motor vehicle accidents. One patient after about 10 years decided they no longer wished to be followed up on the study, but the other patients have continued follow up. Sixteen are alive, and in March next year the last patient will complete 15 years of follow up. I think we've been able to follow these patients long term for two reasons. Firstly, because they're transplant patients and it's a general part of the transplant culture that you follow patients long term, but also because St. Jude has really been highly committed to following

118

these patients long term and has devoted clinical research resources to enable this to be done. And these patients have moved around a lot and traveled to other countries. One of them currently lives in Europe, so it is challenging to follow them long term. Looking at the long-term effects, there have been no long term effects attributable to the development (inaudible) cells. Three patients have developed tumors -- one a benign thyroid tumor, one a skin tumor, and one a low-grade neural sheath tumor. We were able to get tissue on two of these tumors to show that there was no detection of the marker gene. The other one was challenging because it was a patient who as in Europe and it was impossible to get sample from another country. Looking at this risk, I think it's similar to other recipients of allogeneic transplant for relapse leukemia who have received total body irradiation containing regimens. So I'd now like to move to the second study which was a study of multi-virus-specific

119

CTLs. The patients undergoing allogeneic hematopoietic stem cell transplant who are at high risk of infection with one of three viruses -- CMV, EBV, or adenovirus. And we had two studies, one of tri-virus-specific CTLs and one of bi-virus targeting EBV and adeno. Now, both adults and children are at risk of these infections post-transplant, but there's a slightly different risk profile, and in particular, there's a much higher risk or adenovirus infection in children compared to adults undergoing transplantation. When these protocols underwent IRB review, the main concern was actually with the amount of blood that we required from the donor to generate these lines which is 50 to 100 CCs and there was a concern that with the requirement of no more than 4 CCs per kilogram this might be problematic in very small children who were donors. And because of this concern, we modified the study so where the donor was under 12 kilograms they would not be eligible for this protocol.

120

The other concern that came up during review was might these lines potentially be alloreactive. And this had also come up in our initial studies with EBV-specific CTLs that started about 15 years previously. And there's no really good animal model to assess this possibility, and there was also no in vitro data at this time. So we initially developed a release criteria of looking at killing of recipient PHA blasts, and this had to be less than 10 percent for the line to be released for infusion to the patient. Of note, more recently in the last year a publication has come out from a group in Holland which showed when they looked at virus-specific CTL lines in vitro, they could find extensive cross reactivity with HLA molecules. So because of this, we went back and looked at all the 153 donor derived lines we had given on our studies of virus specific CTLs post-transplant and we could identify that 28 of those had some degree of mismatch because the donor was haploidentical and

121

in 43 additional lines it was an unrelated donor with a mismatch at one or more antigens. Looking at the consequences of infusing these lines there was no de novo GVHD and we saw reactivations to grade one or two at an incidence of 13 of 153 overall and the incidence was not different in the recipients of mismatched CTLs. So to see if our CTLs also had the same in vitro reactivity with the type of assay that the Dutch group had developed, we collaborated with John Barrett and Josh Melenhorst at NHLBI who had a similar assay, and they also showed looking at four of our lines that had not caused GVHD that there was activity in these in vitro assays. So I think this illustrates that in vitro assays do not always predict in vivo reactivity. And so preclinical studies are not always predictive of the consequences of in vivo administration. So the way that we generate these multi-virus- specific CTLs is an approach developed by Ann Lein in our group, and we initially adhere monocytes overnight. We then

122

transduce these cells with an Ad535 vector which encodes the immunodominant pp65 antigen of CMV. So for the first stimulation the responding T cells are seeing adnoviral and CMV antigens. For subsequent stimulations we see EBV LCLs, also genetically modified with the same adnoviral vector. So the responding T cells are now also seeing EBV antigens. And the resulting product of multi-virus specific CTL can be administered to recipients after all the QA/QC review is completed. Now, we submitted this protocol to the RAC because we were using gene transfer to our antigen presenting cells and we felt we couldn't exclude the possibility that we might transfer some of these cells, although we had done validation runs to show that this was a very low probability. But actually, I think this illustrates that you need to read regulations very carefully because although we couldn't exclude the fact that we might be transferring antigen presenting cells, we weren't deliberately

123

transferring gene-modified cells. So this protocol didn't strictly need RAC review. In looking at consent process on this protocol, we had a similar procedure for consent/assent in that we would initially discuss the study at the time of the procurement of the cells from the donor pre-transplant. At this stage we also procure some cells from the recipient to act as the PHA blast targets in our release assays. We then re-discuss the protocol when the patient is eligible post-transplant. So again, it's a prolonged process over two to three months. And for children we try and make sure that the discussion is age appropriate and try and include an advocate who in this case may also be a social worker that the child knows well because these are our own patients receiving transplant. Looking at the clinical outcome of these multivirus-specific CTL studies, they've enrolled a total of 35 patients. And if these nine patients had reactivated EBV, either before or just after receiving the cells and all nine had a

124

decrease in viral load as their EBV-specific commune response increased post-infusion. Seven of eight patients with CMV reactivation cleared the virus after receiving these cells. The eighty patient had a strain that was also resistant to all pharmacologic therapy, and eight of eight patients with adenovirus infection in the stool or the blood cleared the virus after CTL infusion. The most dramatic response was seen in this patient who was actually not on the study because they were treated with progressive adenovirus pneumonia and were on the ventilator -- on an oscillator at the time they received the cells, which is normally a very poor prognostic feature in our patient population post- transplant. This patient was treated as emergency use after FDA approval, and as you can see, had a very nice response clearing the infection and remains well over four years later. So our conclusions of these multi-virus-specific CTL studies that we saw expansion of CTL-specific for latent viruses EBV

125

and CMV, and we saw antiviral and antitumor effects for all three viruses. The future direction for this approach that we're looking at process development to simplify the manufacturer of the donor- specific CTLs, but this study also led to evaluation in a multicenter NHLBI study of off-the-shelf CTLs where we gave the most closely matched CTL to recipients. And this study in the first 26 patients has a response rate of partial or complete response as admitted by viral load of 84 percent. So we are proceeding to a Type C meeting with the FDA in December 2010 to discuss how we can further develop this approach. In moving to the multicenter trial there were some additional issues with regulatory review. This study, because it was an NHLBI-sponsored study had to be reviewed by the NHLBI, PRC, and DRC, in addition to the other local review entities. So at our institution it needed to go through 10 levels of review, which was completed in 6 months. I think partially because our regulatory group is very

126

experienced and also because our IRB and IBC is very familiar with these types of protocols. When it was sent to the other sites for the PRC, IRB, and IBC review, it took a different amount of time at different sites which I think might reflect the familiarity of their IRBs with these types of studies ranging from three months to over a year at MD Anderson. So in summary, I think that with cell therapy studies, even in phase one, studies have shown benefit to follow on some of the discussion in the previous session, so we feel that children should have access to these studies. I think the risk of the underlying disease needs to be considered in making decisions about studies. I think with cell therapy studies, one advantage is consent can be a prolonged process because of the time taken for manufacture which allows time for decision-making. I think from my experience in these sorts of studies, parents are often highly informed and I think it is helpful to have a child life specialist or a social worker not involved

127

with the research team who can act as a child advocate. Thank you. (Applause) DR. DALEY: We'll take a few questions. We'll take a few questions for Helen, if you'd come up to the microphones, please. SPEAKER: So Helen, I'm interested in your last point with the advocates, either a social worker or a child life -- did that come internally from your group or was that the RAC or the FDA that made the recommendation? DR. HESLOP: That was actually from the RAC recommendations when we were doing the neuroblastoma study. Diana Wara was the chair of the RAC at that stage, and really in the sort of process of RAC review initially by e-mail and then in the public review, we devised that as having a means of having a child advocate. SPEAKER: And can you tell us a little bit more about how well that worked and what kind of feedback you got from the people that were involved in doing that? DR. HESLOP: Well, I think the child

128

life specialists enjoyed doing it. They were from the sort of tumor team so they weren't involved with the research team and cell and gene therapy and they would sit in the consent meetings and make an appraisal really based on the developmental age of the child, that the child's interests were being followed. SPEAKER: So they were involved in the formal consent process or was it external? DR. DALEY: Just as observers or actually part of the provision of information in the context of informed consent? DR. HESLOP: They were involved in the consent conference and they would also sometimes talk to the child to provide information. SPEAKER: Thanks. DR. HESLOP: At an age appropriate level. UNIDENTIFIED SPEAKER: Yes. Just a quick question on your slide about the generation of tri-virus specific to CTL. You indicated that you had QC/QA release. I’m curious if you could

129

talk at a high level what your release criteria were to ensure that the patient was actually receiving a product that was biologically active. DR. HESLOP: Sorry. I didn't quite hear the question. SPEAKER: If you could speak at a high level or as much detail as you wish on the release criteria that you had which ensured that the patients actually received an active CTL. DR. HESLOP: Okay. So the release criteria are predominantly safety criteria, so we have to obviously have negative bacterial, fungal, mycoplasma, endotoxin. We have to confirm identity by HLA typing the line and make sure it's identical to the donor. And we also have exclusion criteria. If we have our antigen-presenting cells at a level greater than two percent in the final line and then for allogeneic donors we have a criteria of killing a percipient PHA blasts, we measure the activity of the line against the three viruses. We don't, because it's still a phase one study, have a

130

criteria that we use because this is exploratory to try and define what the criteria would be for a later phase study. We obviously wouldn't infuse a line that had no activity discernable in testing. But we've infused lines that have had relativelylow activity, yet have had very good activity in vivo. So it can be hard to ascertain which clones are going to expand on contact with antigen in vivo. SPEAKER: Just one quick follow-on if I might. The last point you were making actually is what I was driving toward. I think that's an excellent point, and that is how we actually establish a way to get a release criteria to ensure that the biological active being delivered, that does have the activity. And then how it's correlated with (inaudible). DR. HESLOP: So potency is something that you definitely need a validated release criteria for in later phase studies, and I'm sure the FDA speakers can expand on that. But in phase one you're often developing what is the

131

appropriate potency assay. Having said that though, we obviously wouldn't infuse a line with no activity. DR. DALEY: Could you just expand a bit on the issue of the jurisdiction of the RAC over your creation of the multi-virus CTLs? So you're saying that obviously the cells that are being infused aren't necessarily gene -- I mean, when you're in this context where the cells aren't being gene modified but the stimulators are, how do you -- how do you -- DR. HESLOP: And that's a lesson that I guess is personally a little embarrassing because I was on the RAC and I sat through many orientation sessions and I've seen that slide of what needs to go to the RAC many times. And I hadn't really read it the way it's meant to be read. I'd read it as that if there was any possibility that you may be transferring gene modified cells. And when I read it now it's quite obvious that it's deliberate transfer. But that's just a lesson.

132

DR. DALEY: Who's decision is it? You subjected yourself to that. I mean, who actually is responsible? DR. HESLOP: Well, I think if you submit a protocol to the RAC, they will review it even if it may not necessarily fall under their purview. And when I talked to them later and we found this out, I think they thought that we were experienced enough that we wanted it reviewed even though it doesn't necessarily fall under the criteria. So that's just a lesson of always read instructions very carefully. On the other hand though, our IBC did like the fact that it was reviewed, and now that we're not having them reviewed by the RAC, these sorts of studies, it's making it more challenging for them. DR. DALEY: Okay. Any other questions? Thank you. Thank you very much. While we're getting the other set up let me just introduce Chris Breuer, who is that fabulous combination of surgeon and scientist, who

133

has been working in this very exciting area of cardiovascular tissue engineering. He's a surgeon in the cardiovascular tissue engineering laboratory and director of the Microsurgery Core Facility at Yale's Program in Vascular Biology and Therapeutic. DR. BREUER: Okay. Good morning. It's a pleasure to be here today. Today I'd like to talk to you a little bit about my work creating, developing, and trying to translate a tissue-engineered vascular graft for use in children. Congenital heart disease affects nearly one percent of all live births. Despite significant advances in the treatment of congenital heart disease, it remains a leading cause of death in the newborn period. Single ventricle anomalies represent a diverse group of structural malformations that result in the formation of a single ventricle. Without surgery, this patient population is associated with a 70 percent mortality in the first year of life, while

134

survival to adulthood is very unusual. Fortunately, due to significant advances in the field of congenital heart surgery, patients born with single ventricle anomalies now live much longer lives with improved quality of life also. In order to enable this, surgeons have developed an operation in which the surgeon essentially rearranges the plumbing in such a way that the single ventricle pumps oxygenated blood to the body, oxygen is delivered and then the deoxygenated blood is returned directly to the pulmonary artery where the blood can then passively go through the lungs, become oxygenated, and start the process over again. Unfortunately, this operation requires the use of either synthetic conduits or patches that are used by the surgeon to perform the operation. And the use of synthetic conduits or patches is a significant cause of morbidity and mortality. It's a significant source of thromboembolic complications, the leading complication after these sorts of operations.

135

Additionally, these conduits have poor durability due to stenosis or ectopic calcification, and additionally, they lack growth potential which is very important in the pediatric population. Occasionally, a child can be born in which primary repair is possible in which the surgeon can directly anastomose two blood vessels in order to perform the operation. And this subgroup of patients do extraordinarily well. Unfortunately, it only represents about one percent of all patients that come to surgery. We've used the excellent results of this patient population as the impetus for our research. We try to create autologous blood vessels using an individual's own cells. We'll seed an individual's own cells onto a biodegradable tubular scaffold. The scaffold provides sites for cell attachment and space for tissue growth. As the tissue forms, the scaffolding degrades so that ultimately you're left with an autologous tube. And our hypothesis has always been that these tissue-engineered

136

vascular grafts would behave more like autologous blood vessels and less like synthetic vascular grafts. This is our scaffold. We start with polyglycolic acid fibers which are woven into a tube and then coated with a 50/50 copolymer of polycaprolactrone and polylactic acid. These compounds degrade by hydrolysis and lose biomechanical integrity over about eight weeks while total fiber degradation takes about six months. This is our cell source. We use autologous bone marrow derived mononuclear cells. It might seem unusual that we'd try and make a blood vessel from bone marrow cells. Well, originally we didn't. In the beginning we made a blood vessel from blood vessel cells. We'd perform a biopsy of a blood vessel and then separate out the different cell types, the endothelial cells, the smooth muscle cells, and the fibroblasts, expand these cells in culture, seed them onto our scaffold, and allow some time

137

for tissue formation. This technique actually worked quite well but it was very time consuming, typically requiring two to three months in order to make a blood vessel in use in an operation which dramatically decreased the clinical utility. Additionally, over time we discovered that sick people have sick cells, and sometimes the patients that would most benefit from one of these tissue-engineered vascular grafts could not have one because we could not passage the cells and culture. So we began to explore other cell types and found that these bone marrow-derived mononuclear cells, which are obtained by performing a bone marrow aspiration and then separating the cells using density centrifugation and Ficoll are available in such abundance as to preclude the need for cell passage and culture. In other words, these cells can be obtained in a single setting, seeded onto the scaffold, and we'll have a tissue-engineered vascular graft in dramatically less time.

138

We seed our cells onto the scaffold using vacuum seeding and then incubate the seeded scaffold and autologous serum, which is also obtained from the bone marrow aspirate for about two hours in order to allow time for cell attachment prior to implantation. Thus, we can create one of our tissue-engineered vascular grafts in a short period of time. We studied our vascular grafts in a variety of animal models in order to determine how they work and how the tissue forms. Here we see some of our experiments using a juvenile lamb model. The lamb is an excellent model for studying the use of biologic vascular grafts using congenital heart surgery because lambs have a very exaggerated form of calcification. And one of the leading causes of graft failure and biologic grafts using congenital heart disease is ectopic calcification. In this study we implanted tissue-engineered vascular grafts into juvenile lambs and then followed them serially over time

139

using MRI. Here we see tissue-engineered vascular graft six months after implantation. During this time there's been adequate time for total scaffold degradation and neovessel formation. Here we see a corresponding MRI demonstrating a widely (inaudible) graft. If we look at the tissue a little more closely, we can see that the tissue-engineered vascular graft resembles the native inferior vena cava. It's got a mono layer of endothelial cells surrounded by concentric layers of smooth muscle cells and it's embedded in an extracellular matrix that's rich in both collagen and elastin. Here we see two images of the tissue engineered vascular graft. The red image demonstrates an MRI picture of the tissue engineered vascular graft approximately one month after implantation, while the superimposed green image shows the same vascular graft six months after implantation. Note how the green image is larger than the red image suggesting that the tissue-engineered vascular graft increases in

140

size. And when we compared the increase in size to the surrounding native vessels, we note that it was proportional, suggesting that this is growth and not simply aneurismal dilatation. In 2001, Toshi Shinoka, the individual with whom I developed the basic technique for creating a tissue- engineered vascular graft reported the first clinical application of the use of a tissue engineered patch in a child undergoing congenital heart surgery. Here we see an angiogram of another patient approximately six months after implantation of a tissue- engineered vascular graft for use in a Fontan operation. And here we see a 3-D CT of yet another patient who had a tissue-engineered vascular graft, and this picture is taken three years after implantation. This table summarizes our early one-year follow up on 25 patients who had undergone a modified Fontan operation using the tissue-engineered vascular graft. There was no graft-related mortality, all 25 tissue- engineered

141

vascular grafts were patent and intact, and there was one graft-related complication, a partial mural thrombosis. Here we see a CT of that patient demonstrating the tissue-engineered vascular graft with clot on the wall. The patient was treated with Coumadin and had complete resolution of the blood clot. This table summarizes our most recent five- to seven-year follow up on the same cohort of patients. Again, there's no graft-related mortality. All the tissue- engineered vascular grafts were patent and intact. There were no new thromboembolic complications. However, four patients developed some evidence of stenosis. All four of these patients were successfully treated with angioplasty or angioplasty and stinting. The indications for angioplasty in pediatric surgery are not clearly defined. When a PTFE or standard -- excuse me, when a PTFE or standard synthetic graft becomes stenotic or narrowed, it's either angioplastied or replaced surgically. Each operation is associated with its

142

own morbidity and mortality. This raises the obvious question what is the graft-related complication rate for synthetic vascular grafts for use in Fontan surgery. Most complications arising from the use of graphs used for Fontan surgery are silent and will not be identified unless specifically looked for using a variety of imaging modalities, such as MRI, CT scan, or angiography. There are only two papers in the literature that serially monitor patients undergoing Fontan surgery with the same imaging modality. Both papers noted approximately a 20 to 30 percent narrowing in synthetic vascular grafts within the first six months after implantation; however, the narrowing tended to plateau with more significant narrowing only occurring in a small fraction of the patients. However, there's a clear lack of quality data describing the safety and efficacy of currently used synthetic grafts, which suggests in some ways it's the blind leading the blind. If we're going to develop an improved vascular graft

143

for use in congenital heart surgery, we're going to need to better study the safety and efficacy of synthetic vascular grafts in addition to studying the safety and efficacy of tissue-engineered vascular grafts. This will slow down the process but it's a necessary step. The other elephant in the room relates to the variability of outcomes between individual surgeons and individual institutions. Surgical outcomes are both surgeon-dependent and institution-dependent. Control of these variables is critical in the design of any trial evaluating the use of vascular grafts in congenital heart surgery. How do you bring a project from the bench to the clinic? How do you get -- how do you fund a project like this one? Where do you begin? When I began this process, I assumed that it was a streamlined set process. It wasn't exactly the case. In the drug industry, this chasm between basic science and clinical practice is referred to as the valley of death.

144

I'm not sure what the right answer is, but I can certainly tell you what I did. From the standpoint of where you begin, it's obvious you're either going to start with the IRB or the FDA. I had more familiarity with our IRB and so I approached our IRB and discussed this project. And after lengthy discussions and conversations, we, together, came up with the design of our initial project. And this design of the initial project that I had developed in collaboration with our IRB actually formed the cornerstone of my FDA application. And it was a useful point of departure for our discussions. Funding. Funding is clearly a critical issue. From my perspective, I find it challenging to work with industry in that typically there are timelines applied. And research certainly doesn't follow a timeline. So if I do use industrial funding, it's never with an associated timeline. I've been fortunate to benefit from a lot of philanthropic funding particularly from the Doris Duke Charitable Foundation, which has been very

145

enabling. When considering a first-in-man study or early clinical feasibility study, safety issues have to outweigh any potential benefit in looking at a particular project. Our primary consideration in selecting the first application for investigative use of the tissue-engineered vascular graft had to do with optimizing safety. We had selected to move forward in a slow and deliberate fashion. We recognize the inherent risks of translating any new technology and feel that adapting a walk before you run policy is both safe and prudent, especially in children. How do you weigh the relative risks of a new device versus the various options that are available? We weighed the potential risks and benefits of a new technology and tried to design a study that presented the least safety risks and we're willing to sacrifice efficacy for this. There are clearly other surgical indications in which a tissue-engineered vascular graft could provide more bang for the buck or have more

146

improvement in efficacy. But we strategically selected the Fontan operation for the reason that we thought it was the safest. And it was the safest because it was a large caliber conduit, it was a high flow system, and it was a low pressure system. So there was the least risk for catastrophic graft failure in the form of aneurismal dilatation or in graft thrombosis. And so we tried to create a balance that favored safety over efficacy. Here's a study design. It's a single institutional study looking at the safety of the tissue engineered vascular graft. It's very small. It's only six patients. And we're adapting a staggered enrollment where we enroll one patient, wait for six months, make sure things are going better before we enroll any additional patients. There's already been some talk about consent. As a surgeon, I obtain consent on a daily basis, and the key is to tell the truth, the whole truth, and nothing but the truth. Consent

147

cannot be obtained in a single setting but, in fact, takes quite a bit of time in order to do it properly. Specifically, for our trial we've adapted a couple of additional things for consent. We have both parents sign the consent form, and additionally, we've created a consent team for the final consent process, and that team is comprised of myself, the surgeon, the cardiologist, and an outside reviewer from our IRB committee. We've also taken advantage of the orphan product line. We have applied for and received humanitarian use device designation and our plans long-term are to use the HDE pathway. This pathway is efficacious in a variety of ways. Primarily, it will enable us to study safety instead of efficacy, but it's also advantageous and once an HUD has been successfully obtained you can begin to charge for the product that you're trying to investigate. And while charge is limited to cost, this is very enabling, especially within academia. So as we're poised to take the leap of faith and begin the process of bringing

148

tissue-engineered vascular graft from the bench to the clinic of the United States, I realize that in many ways my project is really just beginning. I think of two papers, and I think about these papers every single day. The first was a New England Journal of Medicine article by my colleague, Dr. Shinoka, describing the first clinical use of a tissue-engineered vascular graft. And I can't tell you what a wonderful feeling it is to see a project come to clinical fruition and be used to help patients. The second paper is from the European Journal of Cardiothoracic Surgery. And these were some of our contemporaries who developed a tissue-engineered heart valve that was created using a decellularized pig heart valve as the scaffolding onto which autologous cells were seeded. And they had some spectacularly successful preclinical work. However, when they initiated their clinical trial, they implanted six patients. Four died unexpectedly, and two had their tissue-engineered heart valves removed

149

emergently. And a post hoc analysis demonstrated that the pig heart valves had been inadequately decellularized, highlighting the incredible importance of quality control and quality assurance. I'm reminded that the road to perdition is paved with good intentions. So as I move forward with this project, I'm going to strive to achieve the same balance in translating the technology to the clinic in trying to weigh -- appropriately weigh the benefits compared to the risks and perform a well designed, adequately controlled, and properly regulated investigation. Thank you. (Applause) DR. DALEY: Well, the experience with the heart valves is certainly a sobering lesson. Questions for Dr. Breuer? SPEAKER: I'm curious about your use of an HDE rather than orphan drug route given that this is a cellular product and whether you chose that route or what feedback you got from the FDA on that.

150

DR. BREUER: We're a combination product and that's why we selected the HDE. We're actually -- our application is an IDE, instead of an IND. SPEAKER: And I'm just surprised by that given that it is a cellular product because the FDA seems to be moving in the other direction. So were you actually given a choice or were you -- DR. BREUER: This is our plans for the future. SPEAKER: Okay, thank you. SPEAKER: Just an informational point. Based on congressional legislation trying to stimulate device research, actually for HDEs in the pediatric application of the HDEs, you can actually -- you're not limited to cost. You could potentially charge a profit for that portion of the development. There is a guidance document that exists on that you might want to take a look at. DR. BREUER: Thank you. DR. DALEY: Any other questions? Okay,

151

thank you, Dr. Breuer. Okay, so let's take a break. We have a little bit more of a break. We've got until 11 o'clock and we invite you back at 11 o'clock to hear from Drs. Kohn and Crystal. Thank you. (Recess) DR. DALEY: Welcome back to the second part of the session. We're going to hear from Don Kohn who is one of the pioneers in applications of gene transfer in primary immune deficiency. He's a professor at the University of California at Los Angeles. DR. KOHN: I'm going to talk about the three immune deficiencies listed there at the bottom. ADA- deficient SCID will be most of the talk. I'll talk a little bit about X-linked SCID and what's been done in trials in Europe and tell you about a trial that's just opening up now as a second generation. And I'll have just a little bit at the end about some of the work we've done in HIV which we're no longer doing as you'll see and some of the ethical cul-de-sacs we wind up in.

152

Adenosine deaminase deficiency is one of the deficiency is one of the causes of SCID. It's about 15 percent of cases so that there are probably 20 kids born worldwide a year with this disease. It is a very rare disease. This was the first genetic form of human SCID where the biochemical and genetic bases were determined back in the early 1970s serendipitously. It was found that a fraction of SCID patients were missing this enzyme and it was cloned in the mid-1980s allowing consideration for gene therapy once the gene was available. These patients have profound pan-lymphopenia due to a unique biochemical lymphotoxicity from accumulated adenine metabolytes. ADA is shown in the middle of that upper figure. It deaminates adenosine and in the absence of that the metabolites on the left build up and poison off the T cells. I will tell that there are multiple therapeutic options for this disease and our view of those has changed over time so that how we look at the risks and benefits and the potential of

153

gene therapy changes as we think about the alternatives. The current therapeutic options include an enzyme replacement therapy where there is a polyethylene glycol conjugated ADA that is an FDA approved orphan drug for about 20 years and I'll show you some results with that. Or there is allogeneic hematopoietic stem cell transplant, and abbreviations I'll be using include MSD is matched sibling donor, my brother or sister who'd ideally a perfect match, a matched unrelated donor and that would be an adult or a cord blood that's HLA closely matched or haploidentical which is a half-matched typically parent so that transplants can be done with those sources of stem cells or what I'll talk about, autologous stem cells with gene transfer. We're not allowed to call it gene therapy. If we look at these three alternatives over time, in 1990, when gene therapy was first being proposed for this disease, PEG-ADA had just recently been approved by the FDA as an orphan drug and what we knew at the time was that it has

154

short-term value. For allogeneic transplants we knew already that with a sibling there was a greater than 90 percent chance of having a good outcome for a transplant for ADA SCID, whereas with haploidentical transplants from a parent, less than 50 percent of the patients were doing well. In gene therapy at the time all we knew in the laboratory data in T cell lines and primary peripheral blood lymphocytes you could correct the defect and with that data the first trial was started at the NIH. Ten years later when our second trial that I'll talk to you about started we knew more. We knew, one, that PEG-ADA is very expensive. It's about a quarter of a million dollars a year per patient year after year. But it's clearly life sustaining. By that point there were many patients who had been on it for more than a decade who were doing well clinically. In the field of allogeneic stem cell transplant, unrelated donors and cord bloods were starting to take off as a new alternative therapy for transplants. In the gene therapy field we

155

knew that they could get in vitro correction of T cells and hematopoietic stem cells, but the clinical trials that had been done in the 1990s showed minimal benefits. Where we are today is I think our view of all three of these has changed. We still know that ADA is expensive. In fact, when we look back now at these kids who have received it for 10 to 20 years, their immune function is only moderately good. It's really subnormal in the laboratory, but for most of the patients it's been sufficient. Access becomes a problem with such an expensive medication and as these children are now becoming young adults because of this therapy, they're having problems finding financial support for the drug. A recent review with the figures at the lower left looked at patients on PEG-ADA with an N of 80 and there was about an 80 percent survival over two decades so that it's not perfect. There is some continual attrition of patients as their immunity wanes or they've had opportunistic infections like EBV lymphomas. For

156

allogeneic transplant, again that same review on blood on the bottom right shows a survival curve, so matched sibling donors was 87 percent, matched unrelated donors was 67 percent and haploidentical was 43 percent. As I'll show you, in the field of gene therapy there's been real progress over the last decade so that now with more than 30 patients treated worldwide with the current approach, there has been 100 percent survival using a very reduced intensity conditioning chemotherapy. Immune reconstitution has been 67 to 80 percent in the different studies. One of the big risks that we worry about is insertional oncogenesis and that has not been seen in the ADA SCID patients. The initial trials for ADA SCID as I said started in the 1990s using retroviral vectors to transfer the normal gene to either T cells or CD34-enriched hematopoietic stem cells from bone marrow or cord blood. The results were minimal in terms of efficacy. T cells with the normal gene were produced, but at levels too low to restore immunity and the subjects all continued to receive

157

enzyme therapy which while keeping them healthy may have blunted the selective advantage so that the specific reason for studying SCID that a small number of corrected cells could have a selective advantage to repopulate the immune system was blunted by this and this was one of the early ethical issues that was raised in thinking about this disease. Minimal clinical benefits were seen. I'm going to tell you about the trial that we've been doing in slow motion for the last 10 years, cutting back and forth with what's gone on in Europe which is continuing to lead ahead of us and we're continuing to try to catch them. In the late 1990s, we developed a trial of gene therapy based on our preclinical studies to test the safety and efficacy in 10 patients targeting hematopoietic CD34 cells from cord blood of neonates or bone marrow. The process to isolate the bone marrow and stem cells, add the normal gene and give it back to the patients, follow them for 2 years of active, follow-up and then another

158

13 when they've reached the total of 15. As it was developed at the time because we knew about this blunting effect for the enzyme, the protocol said that if the patients had more than 1 cell in 1,000 with the gene at 1 year with some evidence of some effect, we could then withdraw PEG-ADA and look for a clinical response. There was no pretranslant conditioning given, although giving some chemotherapy before giving back the cells may increase the engraftment that there was risk to it so that again from the ethical discussions it was felt that that wasn't allowable at that point. It took us 2 years to open up the trial going through the regulatory reviews, vector production and qualifying the sites and in 2001 and 2002 we treated four subjects. These were children between 4 and 16 years of age who had been on PEG-ADA for 3-1/2 to 10 years. The outcome of this first phase was that there were no adverse effects, the patients currently still remain well on PEG- ADA with no effect from the gene transfer. The two younger subjects, a

159

4-year-old and a 5-year-old, continue to have persistence of some gene marked cells now even 8 years later, but at a very low level. As for the two older subjects, there as a 15-year-old and a 20-year-old, only had transient marking and even no persistence of gene marked cells beyond that point. As we completed those first four patients in the study, results came out from a study in Milan where they had achieved efficacy in gene therapy for ADA. They reported initial results for two ADA deficient SCID infants treated by a very similar retroviral mediated gene transfer to bone marrow cells, but they did two things differently. One is that they gave them a nonmyeloablative or not a full wipe out the marrow, but a less than a dose of busulfan chemotherapy. They used 4 milligrams per kilo. Normally for a transplant fully you'd give 16 milligrams per kilo plus two or three other drugs. These subjects were from countries that couldn't afford enzyme therapy so, therefore, ethically

160

they weren't in the box of whether to take them off the drug and they didn't get ADA enzyme therapy. They reported immune restoration in these patients with increased number and function of T and B cells and the patients remained well without other therapy. Based on that at that time in the summer of 2002, we modified our protocol to follow this approach, to give some busulfan and withhold PEG-ADA. We're about ready to send that amendment floating through the regulatory maze when results of leukoproliferation were reported from the X-SCID trials in Europe. The first report came in September 2002 of a child who had been treated with retroviral mediated gene transfer for SCID 3 years earlier developing a leukemia like illness and at that point all the U.S. Clinical trials involving gene transfer to stem cells were put on hold including ours. Over the subsequent months there were many discussions to explore the mechanisms and methods to detect and minimize these risks. The U.S. ADA gene therapy trial was on

161

and off clinical hold for about 2 years while this was all being sorted out. During this time the follow-up and outcome from the ADA SCID patients treated in Milan, but this point I think they had treated five patients, became known with clear-cut clinical efficacy and no serious adverse events. While the numbers were low, 5 to 10 in the different studies, it seemed like the risk was different for X-SCID where now the numbers are 5 out of 20 develop leukemia, whereas in ADA now it's 0 out of 30. In discussions with the FDA, we discussed how to modify our protocol to allow it to reopen and we developed a plan for monitoring for leukoproliferative complications primarily clinical monitoring through history, physical exams and CBCs looking for increase in white cell populations, and then using lineage amplified media to PCR to detect clonal proliferation if the gene mark was above 1 percent. We also defined an upper limit of cell dose because the first two X-SCID patients had gotten higher cell doses. We

162

also were asked to limit enrollment to subjects 6 months or older because the first two X-SCID patients had been 3 and 4 months of age and age was thought to be a risk factor so that this then would exclude treating neonates with the cord blood arm of our study. Then we took all of this and tried to put it into our informed consent document to describe in lay language the mechanisms and relative risks of insertional oncogenesis as we understood them in 2004 so that that is obviously a bit of a challenge to take something this complex and poorly understood and put it into seventh-grade English. The results in Milan continued to look really outstanding so they published in "The New England Journal" last years the results of their first 10 patients and they now treated as far as I know 15 total. For the first 10 they reported follow-up on 1.8 to 8 years. Eight of their patients remain off enzyme therapy and that's that 80 percent event-free survival. The gene marking in the myeloid cells which reflects the level of

163

gene marking in the bone marrow stem cells is 3 to 5 percent and in the peripheral blood lymphocytes that would have the T cells it's 50 to 80 percent. Nine of the 10 patients had normalization of T cell function and 5 are no longer requiring intravenous gamma globulin replacement and are responding to vaccines. They had some relatively minor adverse events like you would see in any transplant protocol with prolonged neutropenia in two patients, central venous catheter infections in two, EBV reactivation in one, a hypertensive crisis in one and autoimmune hepatitis. The conclusion was that gene therapy combined with reduced intensity conditioning is safe and effect treatment for SCID in patients with ADA deficiency. Our study has gone on with all these various holds so that in 2005 I told you about the four patients on the left and 2001 treated without chemotherapy. In 2005, we reopened the study with the amendment to give busulfan and not to give PEG-ADA and six subjects were treated between 2005 and 2009. At the completion of that we opened a

164

Phase II trial and in 2009 we've done our fourth patient 2 weeks ago. This is the timeline of what we've gone through. I won't go through this in detail, but just to say blue is regulatory review or getting ready, open is when we're treating patients with the Ns and black is time on hold. Over the course of 9 years or so we were on hold for about 3 years, we were under regulatory review for about 3 years and we were open for business for about 3 years, just to make that it's difficult to continue to do trials with funding that has a very finite time period when the timeline for doing the trial is obviously not fully controlled. The are the results from our study. This is just one piece of data showing the lymphocyte counts. Normal for children these ages might be above 1,000 so that everyone shown here is less than normal. The blue line is the patients from the first phase of our trial who remain on enzyme so that they continue to have lowish but stable T cell numbers over the course

165

of follow-up, where the patients in red and green which are the second part of the first trial in the Phase II trial, when they go off enzyme their lymphocyte counts go down putting them at risk for viral infections until they come back up, but after 1 to 2 years they're back up where they were on enzyme therapy without receiving additional enzyme therapy. What we've learned from this study is that adding busulfan and withdrawing PEG-ADA leads to greater numbers of gene-marked cells which correlates with greater levels of endogenous enzyme activity and clinically beneficial immune reconstitution, but in fact we had to add risk to get efficacy so that we had to take them off their enzyme and give them some chemotherapy to have it work. Obviously that creates an ethical dilemma in that how do you jump off that cliff to use the metaphor that was used before. In fact, what we learned looking back now at the nine patients we've treated with this modified approach, only those under 5 years of age have benefited so that

166

we've had three or four patients who were in the older age group and they didn't. This likely reflects superior thymic capacity in infancy compared to older children and possibly also the stem cell content and quality in younger children so that if only older children were enrolled we would have failed to identify the clinical benefits that we've seen. This slide has the world's experience as far as I know it for gene transfer ADA deficient SCID in the Milan study and in the London study and then our trials done with the NIH. If you look across the bottom of 30 patients enrolled, 21 out of 30 are off enzyme and 100 percent survived so that when the gene therapy has not worked on the older patients they have gone back on enzyme therapy. And the overall disease-free survival at 67 percent, in our study if we just look at those under 5 years of age we also have about an 80 percent response rate. Looking forward we'd like to make the response rate even higher than 80 percent so that

167

when we think about how we can do that, we know that lentiviral vectors which have been developed more recently are more effective than retroviral vectors for introducing genes into hematopoietic stem cells. They allow you to do this in a shorter time of cell culture so that this could in theory lead to more rapid and robust immune reconstitution. And lentiviral vectors in theory may have a lower risk for causing insertional oncogenesis than the current retroviral vectors and even though that hasn't been a problem in the ADA trials so far, it certainly is a risk and we'd like to lower that further. We're in the process of trying to develop the next clinical trial to be done jointly in the U.S. and the U.K. to assess lentiviral vectors for ADA deficient SCID and this is a cartoon of the vector that we're using that has the viral control and it's removed and runs off a cellular promoter. We are now faced with how do you take something that has pretty good preclinical data and will be more effective and probably safer, and how do you get that funded and

168

approved and that's where we are with this trial right now. To summarize the ADA SCID, back in 1990, when things first started the prospects for benefit were unknown and the major risk concern I think at that time was concern about a replication competent retrovirus causing an infection in the patient. It was a new technology and that was the worry. In fact, looking back now that has not been a problem with the vectors in the type of testing that's been done. In 2000, the prospects for benefits were still unknown and the major risk or concern was the lack of efficacy, that is, if this isn't working why are you continuing to do it. Now, in fact, in 2010, it's been proven to work in multiple trials and the major risk or concern that emerged in the 2000s was something that hadn't really been high on the list which is insertional oncogenesis in the X-SCID trials. Right now with gamma retroviral vectors the efficacy is in the range of unrelated donor transplants and enzyme replacement therapy and

169

possibly improvements in limiting enrollment to infants and/or using lentiviral vectors will make it equivalent to match sibling donor transplants without the risks of graft versus host disease. Let me switch now and tell you about another disease that I referred to a bit which is the X-linked form of SCID. There is a receptor in lymphoid cells which is shown here so that the receptors for all of these different cytokines have multiple components and the change shown in green is the common cytokine receptor gamma chain and that's what's on the X chromosome that's defective in X-link SCID so that missing that receptor subchain makes the lymphocytes unable to see all those important cytokines. As the immune system develops you fail to develop on the left side T cells, on the right side NK cells. B cells in the middle are produced, but are nonfunctional so that the patients have SCID because of that defect. There were two trials performed for this disease in Europe, one in Paris and one in London,

170

that used basic retroviral vectors that I've talked about where have that gene, the IL2 receptor gamma gene in a retroviral vector and they put it into the bone marrow of the patients. In this case they did not need to give chemotherapy because the selective advantage was expected to be so strong. In fact, there was immune reconstitution in most of the patients, but 5 out of 20 developed T lymphoproliferative disease as I mentioned. Four of these patients have gone into remission and have sustained good immunity. One of the patients died despite multiple treatments of leukemia. This was the first observation of this vector related insertional oncogenesis in humans. It hadn't been in the preclinical studies that had been done. This is the outcome of those two trials added to together in terms of SCID-free, disease-free survival. Despite the complications, it still has an 83 percent survival of patients which is equivalent or better than the alternative of an unrelated donor transplant or haploidentical

171

transplant. The outcomes for SCID with a T cell depleted transplant for example, various studies have reported 50 to 75 percent survival and all those continued to get better. About half the patients especially with haploidentical transplants failed to produce productive antibodies. There was a small incidence of EBV-induced leukoproliferative disease and variable degrees of graft versus host disease. With gene transfer as I showed you for X-SCID, 83 percent of the patients are well and in ADA at 80 percent at least of the patients have immune reconstitution. Certainly at this point it's too early to know which is the better therapy, allogeneic or autologous gene transfer. A new trial is being launched in Europe and the U.S. to try and capitalize on these results which is the hypothesis that a safer vector with equivalent efficacy should lead to better outcomes. This is a vector designed by Christopher Baum that, again, like that lentiviral vector I showed, has the retroviral LTRs indicated by those

172

blue triangles deleted and use the cellular promoter as the internal promoter to drive the gene. In preclinical data as shown in the bottom right, this type of vector has a much lower risk of causing outgrowth of mouse stem cells in culture. On the left those higher dots are the old types of vectors showing the frequency of clones growing in culture and on the right is this new vector showing that at least in this assay it doesn't score for the ability. At least with this preclinical assay this new vector appears to be lower risk for causing insertional oncogenesis. Multi-institutional trials are being developed to look at this vector and the trials are being done somewhat separately because of jurisdiction issues in different countries. The trial will be done at the University College London by Adrian Thrasher and at Hopital Necker in Paris, and then in the U.S. Sponsored by David Williams at Children's Hospital at Boston. The vector, as I said, was developed by Christopher Baum. The plan is to enroll 20

173

patients with X-SCID, 11 at the European sites, 9 in the U.S. trial with 3 from each of the 3 U.S. sites. The criteria to be eligible is you have to have a diagnosis of X-linked SCID and not have a matched sibling donor which is still held as the gold standard. Then it got complicated going through the regulatory reviews and each committee it went through tweaked what should be the mandatory regulatory or eligibility requirements. At the RAC meeting one SCID expert brought in as an outside commenter said that all patients under 3-1/2 months of age should have a haploidentical nonablated transplant and therefore shouldn't have gene therapy offered to them. The RAC recommended that in fact only children older than 3-1/2 months of age should be enrolled in the gene therapy study and that if they don't have an unrelated donor transplant, but if they were very sick and too sick to undergo a haplo transplant then they were eligible. This puts a bias in this trial that only sicker patients will be available. The European studies did not adopt this

174

trial so because of this multiple regulatory review the trials have now diverged a bit. As I said, the RAC recommended infants under 3-1/2 months can only be enrolled if they have a specific infection. The alternative proposed by RAC, haploidentical transplants without ablation, leaves the majority of the patients with incomplete immune reconstitution which is an outcome inferior to the two independent gene therapy trials. As I said, this severely limits the ability to enroll patients and especially now that there is neonatal screening for SCID, many infants are going to be diagnosed in the first weeks of life which is probably the best time for them to have a treatment and the outcome at the different sites will differ. The last little bit I want to talk about is a couple of slides on gene therapy for HIV-AIDS. Our group worked on gene therapy trying to develop ways to immunize cells or patients to be resistant to viruses with several trials. We performed two investigative trials of gene

175

therapies in retroviral vectors into bone marrow cells, one in 1996 and one in 1998. In these trials we didn't give any pre-transplant conditioning, gamma retroviral vectors were used and we got very low to undetectable engraftment of gene marked cells in the patients. The next logical approach based on what was learned from the SCID studies was a trial using a lentiviral vector and nonmyeloablative conditioning to get better engraftment of gene corrected cells. We then went through the process of trying to get funded and approved for this. We got a planning grant from NIH and published our preclinical studies. But when we got to the next stage to get a clinical trial award it was not funded due to ethical concerns about potential risks to children who now had heart therapy that certainly was greatly improving their clinical condition, the various risks I've talked about, with the unknown prospects of benefits. Actually at that point we folded our point and are no longer working on HIV. This raises a dilemma. The capacity for

176

the adult human thymus to educate new T cells from gene- modified stem cells is unknown. Adults may have minimal clinical immune reconstitution from bone marrow gene therapy and even with successful engraftment they may not be able to make new T cells because their thymus is middle aged or beyond. Gene therapy for HIV in fact may only be effective in younger children with the potential for active thymopoiesis and de novo production of T cells. Trials performed only in adults with negative results may preclude assessment in children who may uniquely benefit so that we've painted ourselves into a corner here. A few last comments. Cell and gene therapy for pediatric disorders which are of relatively low prevalence, i.e., not a big market share making it hard to get companies interested, are mostly being developed at academic medical centers and the very long timeline to obtain multiple awards for funding the successive stages and the redundant regulatory reviews present nearly insurmountable hurdles to progress. I hav

177

a couple of recommendations, none of which are novel. One is it would be good if the granting agencies could develop integrated milestone-driven funding mechanisms for the stages to translate from preclinical at the bench to the bedside, pharm/tox clinical regulatory studies, product production and then the trial. Each one how is a separate grant and each one takes you off in different directions. You could make it through the first two and not get funded for the third. The other comment is thorough review from multiple perspectives is certainly beneficial. Many of you are in the room so that I will say all my interactions with regulatory entities have been positive and productive. I'm not trying to tamper with the jury. Multiple thorough reviews by entities with major overlap in jurisdiction like the informed consent wording adds complexity and a great deal of time. The consent documents getting longer and longer may ultimately not benefit the informed consent

178

process. Multi-institutional and multinational review of protocols leads to divergent evolution as I talked about in the X-SCID trial with nonharmonized protocols in fact weakening the scientific strength of the study. A recommendation would be to harmonize reviews when possible like IRBs at multiple institutions maybe if there is some mechanism for them to talk or things in parallel rather than in series and coordinating the review process to accelerate progress and improve quality of shared data. I'll stop there and thank all the people in my group who have done the trials that I talked about and also thanks for all the funding. Again, I also have to thank the Doris Duke Foundation. I got a grant for them to start this trial and they allowed us to stretch it out as we were on hold, and we're currently working on a grant from the FDA's Orphan Products Development Group and they have also allowed us some flexibility in funding as it has taken longer to do the trails than we

179

had hoped. Thank you. DR. DALEY: Some very interesting lessons. Are there questions? SPEAKER: Thank you, Don. That was a very interesting talk. As you know, one of the important issues in these types of trials is monitoring for clonal dominance. That was seen in the X-SCID trials prior to the development of lymphoproliferative disease and there's a question of it having occurred in the thalassemia trial. Would you comment on what measures you're using in the new SCID-X I Trial to monitor for clonal dominance? Is it a stopping role? In your experience what are the best techniques to follow this in patients? DR. KOHN: That's a very good question. Allow me to plug. There will be a workshop put on by the RAC on December 9 and 10 that's going to address exactly all those questions. I think it's not known what are best practices both for the preclinical assessment as well as the clinical monitoring and how do you look for it. All the

180

patients, as far as I know, who were diagnosed with lymphoproliferative disease were done so on clinical grounds. When they presented they had elevated white counts and organomegaly and only looking backwards could you see the emergence of the dominant clone. In the thalassemia trial where there has been a dominant clone, that was seen prospectively because it's clinically asymptomatic or in fact clinically beneficial. I think the current standard is to monitor clinically CBCs, clinical exams and then to do LAM-PCR and look for emergence of a clone. One of the things we've seen in our early patients where we got very little marking was we had oligoclonal marking also, but it's not proliferative, it's just there are only a few clones marked with the gene so you have to distinguish between benign oligoclonality and progressive monoclonal dominance. SPEAKER: Thanks for a terrific talk. Maybe the question is not directed at the right person, but I want to take this opportunity to put

181

it on the agenda maybe for discussion later today. The ADA story leads me to wonder not the question of when a trial should be started or when it's appropriate to take a trial to children, the question that we're spending most of our time on, but when is something ready for consideration of actual FDA approval for marketing or routine use in the clinic outside the experimental setting? I'm hoping that at some point today we can hear something from folks at the FDA about when these sorts of technologies might be ready for those sorts of approval discussions. DR. KOHN: I think that's a very relevant question. Coming out of an academic setting and for very rare orphan diseases I think once we get beyond funding these by hypothesis-driven grants and once it does become therapy, how we sustain what's the business model I think is a real challenge. SPEAKER: One of the things that's striking to me is your last slide where you talked about I think your existential angst with the

182

regulatory and review process. Could you spend a little bit more time on that? You've spent enough time on it already, but it seems to me that this is where a lot of the problems were in trying to harmonize and trying to get multiple institutional review boards and trying to get the regulators all singing from the same song sheet, and I was wondering if you had any recommendations as well. I think this is indicative of the kind of situation that not only you but a number of us are finding. DR. KOHN: I've been talking about the woes of academia, but it's certainly also the same for biotech companies that have limited capital, timeline demands, also facing regulations. I don't have great easy answers. As I said, I sincerely mean that each regulatory review is very well intended and they're really trying to use their expertise to do the best for patients and that's what's important, it's just that everyone does it somewhat in a vacuum. I don't know how to get the different groups talking to each other

183

because it is a jurisdiction issue. You can't have the FDA telling an IRB what to do because the IRBs are local authorities. I don't know a great way. I think maybe the national IRB approach may help a little bit if there's going to be a multi-center study to have a single entity or some coordinated entity to look at it rather than three different institutions or five different institutions, IRBs, each look at it. Maybe that will help although that may get away from the community standard idea for the IRBs so that I don't have great solutions. DR. DALEY: Terrific. Thank you, Don. Our last speaker of the session before the panel discussion is Ron Crystal who has had a very long and distinguished career in gene transfer studies. He's currently professor and chairman of the Department of Genetic Medicine at Weill Medical College, Cornell, and he's going to be talking about gene transfer for CNS indications. DR. CRYSTAL: Thank you, George. What

184

I'd like to review with you is some of the science, our clinical experience and then some of the ethical issues and how we've dealt with them regarding a CNS gene therapy study for a rare pediatric fatal disorder. The disorder is the late infantile form of late infantile neuronal ceroid lipofuscinosis, referred to in jargon as LINCL or Batten's disease. There are several groups of different genes that cause similar symptoms. This is an autosomal recessive disease. There are about 200 to 600 cases worldwide. The disease onset is ages 2 to 4. Children have cognitive impairment, visual failure, seizures, deteriorating motor development and vegetative state and death by ages 8 to 12. There is one of the children in the top picture and below is that child's MRI. All the black that you see is cerebral spinal fluid and not brain because as neurons drop out the brain shrinks and the ventricles enlarge. This is caused by mutations in the CLN2 gene which is a classic lysosomal storage disease. The lysosomes are the

185

-

Transcript for Public Workshop on Cell and Gene Therapy Clinical Trials in Pediatric Populations

-