1 UNITED STATES OF AMERICA DEPARTMENT OF HEALTH AND HUMAN SERVICES PUBLIC HEALTH SERVICE FOOD AND DRUG ADMINISTRATION CENTER FOR BIOLOGICS EVALUATION AND RESEARCH AND NATIONAL INSTITUTES OF HEALTH NATIONAL INSTITUTE OF ALLERGY AND INFECTIOUS DISEASES WARREN GRANT MAGNUSON CLINICAL CENTER NIH GRANULOCYTES FOR TRANSFUSION: RESEARCH AND CLINICAL EXPERIENCE WORKSHOP Friday, September 11, 1998 The workshop took place in the Jack Masur Auditorium, National Institutes of Health, 9000 Rockville Pike, Bethesda, MD 20892, at 8:00 a.m., Liana Harvath, Ph.D., Chair, presiding. PRESENT: KATHRYN C. ZOON, Ph.D., Director, CBER JOHN GALLIN, M.D., Director, WGMCC LIANA HARVATH, Ph.D., Chair DAVID STRONCEK, M.D., Moderator DOUGLAS ADKINS, M.D., Speaker DANIEL AMBRUSO, M.D., Speaker DAVID DALE, M.D., Speaker JOSE-LUIS DIAZ, Ph.D., Speaker THOMAS LANE, M.D., Speaker SUSAN LEITMAN, M.D., Speaker CONRAD LILES, M.D., Ph.D., Speaker THOMAS PRICE, M.D., Speaker ALSO PRESENT: JAY EPSTEIN, M.D. 2 AGENDA Welcome/Introduction Kathryn C.Zoon, Ph.D., Director CBER/FDA . . . 3 John Gallin, M.D., Director, WGMCC/NIAID . . . 5 SESSION I Moderator Liana Harvath, Ph.D. . . . . . . . . 7 Historical Perspective & Clinical Trial Considerations David Dale, M.D. . . . . . . . . . . . . . . . 9 Cytokine Administration to Normal Granulocyte Donors Thomas Price, M.D. . . . . . . . . . . . . . . 47 Functional Properties of Granulocytes from Normal Donors after G-CSF Administration Daniel Ambruso, M.D. . . . . . . . . . . . . . 71 Cytokine Mobilized Granulocytes: Assessment of Efficacy Susan Leitman, M.D. . . . . . . . . . . . . . . 91 Panel Discussion . . . . . . . . . . . . . . . 120 SESSION II Moderator David Stroncek, M.D. . . . . . . . 154 Product Evaluation David Adkins, M.D. . . . . . . . . . . . . . 154 Storage Considerations Thomas Lane, M.D. . . . . . . . . . . . . . . 175 In vitro Assays Predictive of Product Function Conrad Liles, M.D., Ph.D. . . . . . . . . . . 197 Panel Discussion . . . . . . . . . . . . . . . 216 Presentation of Submitted Abstracts David Stroncek, M.D. . . . . . . . . . . . . 229 Dr. Diaz . . . . . . . . . . . . . . . . . . 242 3 1 P-R-O-C-E-E-D-I-N-G-S 2 (8:08 a.m.) 3 CHAIRPERSON HARVATH: Good morning. I 4 think we will get started in the spirit of trying to 5 keep on time. On behalf of the Steering Committee for 6 this workshop, it is a pleasure to introduce two very 7 distinguished individuals, Director for the Center of 8 Biologics Evaluation and Research, Dr. Kathryn Zoon, 9 who will give welcoming remarks, and Dr. John Gallin, 10 who in this audience really needs no introduction 11 since his career in the granulocyte field has made 12 enormous contributions to this field. Dr. Gallin will 13 be speaking on behalf as the Director of the Clinical 14 Center as well as making remarks on behalf of NIAID. 15 Dr. Zoon? 16 DR. ZOON: Good morning. It is a pleasure 17 to be here and to open this important workshop with 18 Dr. Gallin. John and I have known each other many 19 years, in our interferon days and working on various 20 activities of interferon gamma and granulocytes. I 21 think there is lots of work to still be done and I am 22 very anxious to hear the science today and understand 23 where we are with the technology and where we need to 24 go tomorrow. 25 It is a pleasure to welcome you on behalf 4 1 of the Center for Biologics Evaluation and Research, 2 which is one of the FDA centers that has oversight of 3 blood and blood products. This workshop today, I 4 believe, will really help us focus on some very 5 important issues. The findings that the 6 administration of cytokine, such as granulocyte colony 7 stimulating factor and granulocyte macrophage colony 8 stimulating factors to normal volunteers results in 9 the peripheral mobilization of high concentrations of 10 granulocytes have renewed the interest in the 11 collection of granulocytes for transfusion. I think 12 all of us have become very familiar in the literature 13 with both the effects of G and GM-CSF on this. And 14 while there was a lot of interest in granulocytes for 15 transfusions -- this peaked back in the 1970's -- 16 there is now a renewed interest because of our new 17 tools. We are very interested in exploring both the 18 efficacy parameters as well as the safety parameters 19 associated with this. 20 We are very interested, and of course 21 there are many others interested, in the scientific 22 and clinical experience with cytokine mobilized 23 granulocyte transfusion products and the effects of 24 the cytokine administration on normal donors. This is 25 clearly important with regard to the safety of the 5 1 patients and the donors, but also important in the 2 quality of the products, in this case the granulocytes 3 looking at their functional capabilities. 4 Our colleagues at the NIH are here today 5 to hear what areas in this field need to be further 6 explored and supported in the research area, and we 7 really are very appreciative of doing this workshop 8 today in a collaborative effort, because many of the 9 fruits of the science and the research that do arise 10 come to FDA for our review and evaluation, and it is 11 very important that the Agency be very active in the 12 science and understanding the science so that we can 13 do the very best job at facilitating the review and 14 access of these important products. 15 This is the second workshop this week. We 16 had one yesterday on hematopoietic stem progenitor 17 cells and this one today on granulocytes for 18 transfusion, and we really appreciate the attendance 19 today. We feel this is a very important area. We 20 feel very strongly that we need the best scientific 21 data in which to move forward, and I wish you a very 22 good meeting and look forward to the fruits of this 23 workshop. Thank you very much. 24 DR. GALLIN: Well, normally I don't know 25 that I would get up and be willing to share in two 6 1 introductions to one meeting, but for this meeting I 2 couldn't resist. And the reason is because of my own 3 longstanding personal affection for the granulocytes, 4 which are by my way of thinking the most beautiful 5 cells in the body, and also because of the importance 6 of granulocyte transfusions. So on behalf of both the 7 Warren Grant Magnuson Clinical Center and all the 8 staff who work here as well as the National Institute 9 of Allergy and Infectious diseases, both of which are 10 places that I work, we welcome you here. 11 Now my personal interest really goes back 12 to the use of granulocyte transfusion in patients with 13 granulocyte defects, particularly patients with 14 chronic granulomatous disease and a few patients that 15 we have seen in this building with neutrophil specific 16 granule deficiency. And over the last 25 years, we 17 have been convinced, truly on anecdotal evidence, that 18 there are some patients in whom granulocyte 19 transfusions made a difference in helping them get 20 over life-threatening infections. But, of course, 21 there is no proof that they work, and that has always 22 been an issue. And we have also been worried that 23 maybe we were actually doing something bad or 24 potentially bad. Perhaps we always worried that maybe 25 we would be precipitating ARDS in patients with 7 1 chronic granulomatous disease or that we would be 2 alloimmunizing the patients to an extent that future 3 bone marrow transplants or gene therapy might be 4 compromised. 5 Nonetheless, we continued to use them on 6 occasion. So I personally really look forward to the 7 results of your deliberations today with the hope that 8 you will come to some conclusions in terms of not only 9 the correct approaches for mobilizing these cells and 10 harvesting them and storing them, but also for 11 designing some clinical trials that will answer some 12 of the definitive questions. So have a great 13 symposium. 14 CHAIRPERSON HARVATH: This is a great 15 pleasure for me to be involved in helping to organize 16 such a workshop, because my own area of research for 17 the past 25 years also involves those beautiful cells 18 of the body, the granulocytes. 19 In looking through our sort of cryptic 20 files of workshops at the FDA, I came upon some 21 transcripts of a workshop held in October of 1980 in 22 Natcher Auditorium, and it was called the Conference 23 on Leukopheresis Donor Safety. And that was actually 24 the last conference that FDA was involved in where 25 there was a discussion of the collection of 8 1 granulocytes and leukocyte products, and the 2 discussion at that time focused on donor safety issues 3 when donors are given hetastarch or when they are 4 given steroids for mobilization. 5 There were a number of participants in 6 that workshop who were actually serving on the 7 Steering Committee. So what I would like to do is to 8 in the next slide -- what I had is a list of people. 9 There should be two slides in that carousel. The 10 second slide -- there we go, thank you -- is to 11 acknowledge and thank the members of the Steering 12 Committee. All of these people have distinguished 13 themselves in the granulocyte field, and I would like 14 to go through this list. Dr. Daniel Ambruso from the 15 University of Colorado and Bonfils Blood Center will 16 be speaking to us; Dr. David Dale from the University 17 of Washington; Dr. John Gallin; Dr. Jeffrey 18 McCullough, from the University of Minnesota, who 19 unfortunately can't be here but who made major 20 suggestions for the organization of this workshop; Dr. 21 George Nemo from the Heart, Lung and Blood Institute; 22 Dr. Daniel Rotrosen, who worked in this building with 23 Dr. Gallin and Dr. Malik in neutrophil research; Dr. 24 Ron Strauss from the University of Iowa, who is one of 25 the pioneers in this field, who unfortunately can't be 9 1 here because he is giving a talk on granulocyte 2 transfusions at the Pediatric Oncology Meetings today 3 in Chicago; Dr. David Stroncek, who now works here at 4 the NIH Clinical Center, and you will hear some of his 5 work this afternoon; and finally, our colleague, 6 Joseph Wilczek, who has served as the program 7 coordinator in taking care of the laborious details 8 that go into putting a conference like this together. 9 It is a great pleasure for me to have the 10 opportunity to introduce the speakers in the morning 11 session. Dr. David Dale began his interest in 12 granulocytes almost 30 years ago in this very 13 building, where he was working as a clinical 14 researcher, and he is now a professor of medicine and 15 has had a very distinguished career in education and 16 research in this field. Dr. Dale is going to talk 17 about the historical perspective and clinical trial 18 considerations for granulocytes for transfusion. 19 DR. DALE: Well, thank you, Liana. It is 20 nice to be here. I would say nice to be here again. 21 I think I last spoke about this topic in this room 25 22 years ago, so it does bring back a lot of memories. 23 If I can have the first slide, which is 24 just a title slide. I am going to talk this morning 25 about really three things. I am going to talk about 10 1 history. That is always a dangerous thing to do. I 2 will mention a lot of names. I may not mention 3 everyone because of the time, but there have been many 4 people involved over the years in studying this topic, 5 many who have done other things in their lives and you 6 will recognize them. I am going to talk a bit about 7 the effects of G-CSF in particular on neutrophil 8 formation and function, again a part of the background 9 for our overall discussions today. And then finally 10 talk about recent history, and that is the data that 11 has led up, I think, in many ways. And then finally 12 to mention a little bit about clinical trial 13 considerations. 14 Our focus today is on the neutrophil, 15 which my predecessors here this morning have described 16 as beautiful, and they really are interesting cells. 17 Most oncologists see them on a laboratory slip as a 18 number. Hematologists may see them on a blood smear. 19 But they are truly interesting and beautiful cells. 20 Perhaps made more interesting and beautiful if you see 21 them in a diagram like this, which shows some of the 22 features of a neutrophil. I won't dwell upon this 23 today, but suffice it to say that there are many 24 features of these cells that are regulated very 25 tightly, and the cells that circulate in the 11 1 circulation are not always the same. Infection, the 2 administration of hematopoietic growth factors, 3 diseases, many factors can change the characteristics 4 of these cells. The key features though are the 5 surface of the cell, where the cell has receptors 6 which allow it to interact with its environment, the 7 granules of the cells, which are critical for the cell 8 function, and then, of course, the nucleus, which 9 allows us so easily to recognize a neutrophil in a 10 blood smear. 11 In terms of the formation of neutrophils, 12 as a background statement I think it is very important 13 to think about the kinetics. Because the dynamics of 14 how the body produces and how neutrophils are 15 distributed in the blood and their ultimate fate have 16 so much to do with the development of this field, both 17 in the past and in the future. Neutrophils are formed 18 from hematopoietic stem cells and the steps of 19 proliferation and differentiation and then maturation 20 are very unique for these cells. The most unique 21 feature compared to other blood cells is the storage 22 in the marrow of a substantial portion of the body's 23 total supply of the cells. How much is it? Well, it 24 is probably 10 times the circulating supply or perhaps 25 more depending upon where you draw the line. But 12 1 suffice it to say it is a very large reserve, a 2 reserve that is not there for monocytes or eosinophils 3 or red cells or other of the hematopoietic cells. And 4 it is really this reserve supply of neutrophils in the 5 marrow that are critical for the development of the 6 whole concept of collecting neutrophils from normal 7 donors for transfusion. Neutrophils, as you will 8 recall well, have a short blood lifespan, and in fact 9 almost all of their function are in the tissues. And 10 although clinically over many years we have related 11 susceptibility to infection to the number in the 12 blood, it is actually the total body supply and the 13 ability to deliver these cells to the tissues which is 14 critical for the outcome in terms of the problem we 15 are talking about. 16 And then finally as background, I will 17 mention the process for the killing of organisms, 18 represented by this pink cigar here, by a neutrophil 19 is a complex process that we have unraveled in 20 research supported here and in much done here as well 21 as elsewhere around the world, to clarify the various 22 processes involved in the dumping of myeloperoxidase 23 into the phagocytic vacuole and the involvement of 24 oxygen and oxygen derivatives in the actual killing of 25 these organisms. And it is this event, actually the 13 1 ability to kill organisms, that is why we are all 2 healthy enough to be here today. So it is vitally 3 important not just for sick people but for people who 4 regard themselves as healthy. 5 Now critical finally for the understanding 6 of this topic is the short lifespan or the rapid 7 turnover of neutrophils. This is illustrated here 8 just by comparison with other blood cells. Red cells 9 having a lifespan of roughly three months or a 10 turnover rate of 1 percent per day, platelets roughly 11 a tenth of that in terms of their lifespan, and ten 12 times as rapid turnover, but neutrophils are among the 13 most rapidly turning over cells in the body. In many 14 audiences I have described it as just think of the 15 fact that in your blood today, you have a whole fresh 16 supply of neutrophils from what you had yesterday. 17 You have replaced all of those cells. And if you go 18 just a few days without a new supply, of course you 19 are in trouble, as we all know. 20 Now in terms of thinking of history, this 21 is actually an old slide, I think one that I showed 22 here in a slight variation 25 years ago. The problem 23 of neutropenia has been one that has been recognized 24 for almost the whole 20th Century. The problem was 25 actually recognized soon after the turn of the 14 1 century, with neutrophils being counted in blood 2 smears over the few years before but really not very 3 long. But the treatment of neutropenia has 4 languished. This is basically our 1970's approach, 5 that is, the use of antibiotics, and we still depend 6 upon them to deal with the problem of neutropenia. 7 We had then in the 1970's a variety of agents to try 8 to increase neutrophil production, I would say all 9 very, very weak. There were efforts to try to treat 10 neutropenia with other factors such as removing the 11 spleen in chronic states, but of course this was not 12 applicable to acute states as occurs after 13 chemotherapy. And finally then, as now, a key concept 14 in the treatment of neutropenia was an alternate 15 supply, that is, to transfuse the cells. 16 In terms of the history of this idea, that 17 is, the history of transfusing neutrophils to treat 18 neutropenia, most reviewers of the topic would point 19 to a study done in 1934, the height of the Depression, 20 a study involving injection of buffy coat cells into 21 patients, intramuscular injections, which were 22 undoubtedly very painful and basically had no effect 23 on the patients. But it was a part of a desperate 24 approach in the pre-antibiotic era to doing something 25 about the problem of neutropenia. 15 1 Actually, the first important 2 investigation in this area was conducted here by Dr. 3 George Brecher and associates, working with Gene 4 Cronkite, who was then at the Naval Medical Center, 5 and that study, I think, was a landmark in terms of 6 the building of the basic physiological concepts 7 underlying this field. What Dr. Brecher did -- I 8 don't know, somebody in the room may know him -- I 9 remember him from when I first came here, a really 10 great man. What he did were studies where donors, 11 that is, dogs which had been made aplastic with 12 radiation -- the donors were injected with turpentine 13 to try to increase their counts and then the 14 recipients were irradiated to aplasia, and then cells 15 were separated and it was shown that cells could be 16 accumulated at the site of inflammation. In Brecher's 17 original studies, he showed that some cells could get 18 there and they could circulate and that some effects 19 could be seen. 20 What happened historically from that point 21 was really very gradual. But a key event again here 22 in the early 1960's was the development of the concept 23 of transfusing cells from CML donors. Studies that 24 were, I would say, led by Jay Freireich but involved 25 a number of people here and subsequently elsewhere, 16 1 showed that you could take CML cells, donors were 2 patients who were untreated and recipients were 3 patients with leukemia usually, and you could get the 4 CML cells to circulate in those persons and the more 5 mature of the CML cells would actually migrate to a 6 site of inflammation. More dramatically in those 7 early studies it was shown that using a very simple 8 index that the patients became afebrile. Dr. 9 Freireich often talked about the fact that patients 10 treated with CML cells showed clinical evidence of 11 improvement very rapidly. Those studies were greeted 12 with great enthusiasm. I would say the enthusiasm was 13 tempered in time as treatments for CML improved, and 14 also it was recognized that you could have the CML 15 cells engraft, you could transfer infection with CML 16 cells, and that there were a variety of complications 17 that were associated with this approach to therapy. 18 Actually, the next period of development 19 centered on the development of the cell separator, 20 work that was supported and performed here by a series 21 of investigators. I would say Dr. Seymour Perry, who 22 many of you may remember worked here for many years 23 studying granulocyte and leukocyte kinetics, was the 24 real father of the NIH efforts in this regard. But 25 there were a series of investigations performed here 17 1 by Dean Buchner, Bob Epstein, Bob Graul, and then 2 myself over a period of years which really tried to 3 lay a physiological framework for the advancement of 4 neutrophil transfusion. This is a picture from one of 5 the first papers about this topic of the centrifuge 6 for collecting cells. And we all know that have been 7 involved in blood banking how centrally important the 8 development of this centrifuge concept for separating 9 cells on the basis of their density has been. So it 10 was great research done by the IBM in collaboration 11 with the NIH, and it did lead to the capacity to 12 procure large amounts of cells. 13 This is a very brief slide about Dean 14 Buchner's work, studies which were originally done in 15 dogs, and showed that you could collect of the order 16 of 24 billion white cells if you kept a dog on this 17 centrifuge long enough, and that you could then 18 actually transfuse these cells and see them circulate, 19 sort of reproducing the work that George Brecher had 20 done, but showing it with larger numbers of cells in 21 the same species, but showing that you could, in fact, 22 get very good increments if you used enough cells. 23 And importantly, they showed in this very early study 24 that cells collected with a centrifuge would 25 circulate. And subsequent to this work, Bob Epstein, 18 1 who worked in Seattle with Reg Clift and Don Thomas 2 went on to show that you could use this same model and 3 then in the irradiated dog injected with E.coli to 4 develop bacteremia, you could in fact show an effect 5 on an experimental infection in terms of the clearance 6 of the infection. 7 Now when I joined and began working with 8 Bob Graul, he was then actually in that era just 9 beginning clinical trials here of the use of 10 neutrophil transfusions for patients with sepsis. 11 These were pioneering studies, studies that were done 12 not in a rigorous controlled trial but studies 13 published in the New England Journal, which clearly 14 suggested that this was a promising area for 15 application. The original studies were done primarily 16 but not exclusively with centrifuge collected cells. 17 The studies that I was involved in here 18 with Bob and Herb Reynolds and a number of other 19 investigators involved dogs, again irradiated to 20 produce neutropenia, and then injecting those dogs 21 with Pseudomonas aeruginosa intratracheally to cause 22 a localized pneumonia, and then treating the dogs in 23 a randomized controlled rigidly monitored study where 24 some dogs were supported with platelets only and the 25 others were supported with platelets with 19 1 granulocytes. 2 Just to make it a little more interesting 3 and colorful, I can comment very briefly upon things 4 I remember about the trial. As I was walking around 5 here yesterday, I was remembering I used to keep an 6 old bicycle outside the cafeteria downstairs to ride 7 back and forth to the animal quarters to administer 8 all the treatments to these dogs. And a key reason 9 for our success in this trial is that the NIH then had 10 a farm in Poolesville. I think it is all developed 11 now, but we had a farm out there with great big 12 foxhounds who were the donors, and in our clinical 13 trial or our preclinical trial, we used very small 14 beagles as the recipients. So that allowed us the 15 advantage of despite the number of cells we could 16 collect, we could see and measure and do a lot about 17 neutrophil increments. 18 This is the picture of the lung of a dog 19 injected unilaterally with pseudomonas and developing 20 a characteristic hemorrhagic pneumonia in an animal 21 with a very low neutrophil count. The key observation 22 we reported in these studies in the Journal of 23 Clinical Investigation in 1974 is in fact that if you 24 looked at controls versus transfused animals, you 25 could clear the Pseudomonas of the specific type we 20 1 had injected from the lung by a series of 2 transfusions. And we showed in our randomized trial 3 that you could improve survival. A number of other 4 things came from the study, but it was the place that 5 I personally became convinced if you had enough cells 6 that were functionally intact that you could use this 7 approach to treating severe infections. 8 We also studied at this period the 9 interaction of antibiotics and neutrophils and derived 10 certain conclusions about that, and I think that is 11 another important consideration that will come up when 12 any clinical trial is now considered, that is, which 13 antibiotics are best and how to approach the 14 antibiotic neutrophil interaction. 15 In terms of what happened then is a nice 16 illustration, I think, of the circuitous path of 17 clinical research. This is a picture of a 1970's 18 filter that we used and was used widely then to 19 collect neutrophils by filtration leukopheresis. Many 20 of you will remember this. It basically depends upon 21 the property of neutrophils to stick to anything 22 almost, and that is in fact the way that they are 23 selectively recruited to a site of inflammation, and 24 this basic stickiness of the cells was how they could 25 be collected in much larger numbers. Many of the 21 1 studies performed in the 1970's involved the use of 2 this filtration system as a way of trying to increase 3 cell numbers. 4 And what happened is that in the 5 development, it was learned that you could improve 6 collections if you used starch to accelerate red cell 7 sedimentation and if you used steroids to raise the 8 counts in the blood. But what really helped in terms 9 of the numbers was the use of a filter to get lots of 10 cells. This is data from that era showing a 11 comparison of how many cells you could obtain at best 12 with a centrifuge and how many more you could collect 13 by filtration leukopheresis. These are probably 14 conservative differences. That is to say if it were 15 three to five times as many that that would be the 16 expected. The problems that occurred were that 17 although the efficiency of collection was more, the 18 cells were damaged in the process of their collection. 19 And in fact probably in the process of activating 20 cells, we would say in modern terms the release of 21 cytokines from neutrophils, you often saw febrile 22 transfusion reactions in response to the 23 administration of filter adherence collected cells. 24 Nevertheless, until around 1980, this technique was 25 widely used. 22 1 There were a number of studies in this era 2 that sort of took on, if you will, the filtration 3 leukopheresis cells and looked at what was wrong with 4 them. This is a paper by Tom Price, who is speaking 5 next, about this, looking at the difference in the 6 disappearance rates of various types of neutrophils, 7 cells from phlebotomy, cells collected by 8 centrifugation, intermittent flow centrifugation, and 9 then cells collected by filtration leukopheresis. And 10 basically the cells were -- it was found that the 11 cells were damaged enough in the collection process 12 that they wouldn't circulate. And although there was 13 suggestive evidence that they might be useful for 14 therapy, the evidence was never very strong. And 15 particularly because of the transfusion reactions, 16 this process of collecting cells by filtration 17 leukopheresis fell out of vogue. 18 Now there have been many summaries of the 19 studies that were done in the 1970's and early 1980's 20 looking at the benefit and the use of granulocytes 21 based on various trials. This is a slide borrowed 22 from the summary work by Ron Strauss outlining what he 23 would consider the best of these trials, dating from 24 Bob Graul's trial that I mention here as the first of 25 these, up until a trial performed at UCLA in 1982. So 23 1 it spans a 10-year period. Those of you interested in 2 clinical trials will say aren't these pretty small 3 numbers, and they really are. In fact, the number of 4 patients in the treated groups are sufficiently small 5 that part of the problem with these trials was just 6 basically their size. Another problem is that cells 7 in these trials were collected by various techniques, 8 and in general the trials that showed the best results 9 involved the transfusion of the largest numbers of 10 cells. 11 Another of the problems in these trials is 12 that the patients weren't always the same. That is 13 reflected probably most easily here if you look at the 14 percent survival of the control groups. Because if 15 the comparison group did well, it was very difficult 16 to imagine that you would show a benefit of the 17 treatment. So, for instance, in a study like the last 18 study with a 72 percent survival rate in the control 19 groups, the fact that the transfused group did more 20 poorly, these numbers are probably not different, but 21 this is so high it is hard to imagine that this trial 22 would have shown a benefit. Suffice it to say the 23 clinical trials were not sufficiently convincing that 24 although there were people who spoke enthusiastically 25 about this topic for a number of years, clinicians in 24 1 general dropped this idea because of the results of 2 these randomized trials and the difficulty in seeing 3 the benefit to a patient of a single transfusion. 4 If you look in more detail at these trials 5 though by specific types of infection, you can see 6 that for certain kinds of infections it appears that 7 patients really did do better. And I would say that 8 Ron Strauss, who wrote this report, were he here would 9 say that he believes that what these trials showed is 10 a benefit, but that the trials were not sufficiently 11 convincing to convince everyone of that. 12 Probably the most positive and most recent 13 trial of note is the trial that was performed by Mitch 14 Cairo and associates, a trial that involved an 15 interesting comparison. These were in children, so 16 you had the advantage of small recipient/big donors, 17 but you also had a comparison group, that is, the 18 comparison of neutrophil transfusion versus a control 19 which involved gamma globulin injections in neonates. 20 What this trial showed is that there was a very 21 significant benefit in neonates, but critics of the 22 trial have said that the two groups in the study were 23 not really comparable and that the methods of 24 randomization lacked the rigor to make this study 25 really a definitive study for the treatment of 25 1 neonates with sepsis, and in fact it is for this 2 reason probably that this trial was never really 3 accepted as convincing evidence or the practice of 4 neutrophil transfusion in neonates, despite these very 5 striking results, has never been widely introduced. 6 Now paralleling these efforts by many 7 people to develop neutrophil transfusion was another 8 development that is very important for the reason that 9 we are here today. And that is the development of our 10 understanding of the regulation of granulocyte 11 production and the use of the colony stimulating 12 factors in patients and in normal subjects to try to 13 raise the neutrophil count. And because it is so 14 basic to our discussions, I thought I would review 15 this background information with you as well as a part 16 of my history talk. 17 Many of you will recall that in the mid- 18 1960's, it was learned that you could take bone marrow 19 cells in a petri dish and with a tissue culture media 20 and some source for the stimulating factor, the cells 21 would grow and form colonies. Again reflecting back, 22 it was during my years here that this technique came 23 along, and it was a very exciting development with 24 Paul Carbone and Clarence Brown. We did the original 25 colony assays here at the NIH showing that you could 26 1 grow cells because this led to so many important 2 developments in the whole field of hematology, 3 oncology, and transfusion medicine. 4 Suffice it to say the field has come a 5 long way. And I would like to make just a couple of 6 important points related to this slide about the 7 various growth factors involved in regulating 8 leukocyte production. The key concept is that early 9 cells have lots of surface receptors affecting the 10 formation of cells, but the late stage of development, 11 or more specifically, the regulation of the number of 12 circulating neutrophils is governed by a single 13 factor, G-CSF. That is to say that an animal made 14 deficient in G-CSF does not have a normal circulating 15 count. And although they make neutrophil precursors, 16 they don't mount a normal neutrophil response usually 17 or in response to infection. So the G-CSF, one of the 18 reasons that we are talking about it today is that it 19 is the natural regulator of the level of circulating 20 neutrophils in the same way that erythropoietin is the 21 natural regulator of the circulating level of red 22 cells and thrombopoietin is the natural regulator of 23 the level of circulating platelets. 24 Now G-CSF as a drug was introduced in the 25 late 1980's, and many of you know a great deal about 27 1 it. The gene was originally isolated from a bladder 2 cell line, a cell line that was from a patient with 3 neutrophilia. In fact, many malignant cells 4 overproduce G-CSF and served as the original source 5 for the material that was used in developing the basic 6 structure of this protein and understanding its 7 genetic regulation. 8 The pharmacological effects of injecting 9 G-CSF are now well characterized and most of you know 10 then. That is, if you inject this drug, you can 11 quickly achieve levels that are far higher than you 12 normally achieve with infection or stress or a variety 13 of other natural stimuli. The drug is prepared and is 14 easily administered and in fact has relatively few 15 side effects. This will come out further as the 16 conference proceeds. 17 In terms of how the colony stimulating 18 factors work, just a few key points. One is we heard 19 a lot yesterday about the use of G-CSF, and you could 20 say that for GM-CSF too. That is, they are agents 21 which mobilize the earliest hematopoietic cells from 22 the marrow to the blood. The details of exactly how 23 that works are still not know. But it is a dramatic 24 effect, an effect that was totally unexpected when 25 these agents were originally studied. 28 1 G-CSF is particularly powerful also to 2 stimulate the flow of cells down the pathway or as I 3 usually describe it, to squeeze down the production 4 time for neutrophils from early cells to mature cells 5 in the marrow. So like an accordion, you squeeze down 6 this time dimension and you get more cells and you get 7 them quicker. And finally, because you have more 8 cells -- well, G-CSF also releases the mature cells 9 from the marrow to the blood and then finally because 10 you have more cells in the circulation, it allows for 11 the possibility of a larger inflammatory response. So 12 you can see that this natural stimulus, that is, it 13 arises in infection, or as the drug might be used to 14 stimulate neutrophil production has a multiplicity of 15 effects. And you can imagine that in the development 16 of this agent that there have been many potential 17 clinical applications. We are talking today just 18 about one of them. 19 We began in Seattle to try to investigate 20 and to build this picture further, now about 8 years 21 ago, about 1990. Our original studies were a trial 22 that we did to try to compare the effects of G-CSF in 23 young and elderly subjects. Our original idea was to 24 try to study aging. That is, we wondered if there was 25 an impairment of the proliferative capacity of cells 29 1 as people get older. So we compared healthy young 2 people and healthy elderly people and we had a regimen 3 which involved G-CSF daily for two weeks and then a 4 whole variety of measurements. I am not going to 5 dwell upon many of these, but just to show you some 6 highlights from studies that we have done then over 7 the last eight years. 8 This was the original curve showing that 9 normal people have a very stable neutrophil count with 10 no injection. If you give a small dose of this drug, 11 you get a small effect. If you give a larger dose -- 12 this is 30 mcg total dose per day and this is 300 mcg 13 total dose per day injected once in the morning 14 subcutaneously measuring morning counts, and what you 15 see are these characteristic patterns. And if we had 16 gone up higher in the dose, there probably is a 17 plateau, but no one has ever really measured how high 18 that plateau may be in terms of driving neutrophil 19 production with this drug. But you can see with a 20 dose of 300 mcg daily, you can get to a plateau count 21 of roughly 25,000 in healthy people fairly quickly. 22 And in fact it is this rapid increase which 23 distinguishes the effects of G-CSF from GM-CSF when 24 administered to normal subjects. GM-CSF causes a much 25 more gradual rise in the count over a more protracted 30 1 period. 2 If you look at the cells that are produced 3 in this kind of a setting, show on the left-hand side 4 are normal neutrophils and the same person's blood 5 looked at 5 days later after daily administration of 6 G-CSF. It is a gradual transition. But what you see 7 is the production of bigger, bluer cells with a 8 somewhat less mature nucleus. And if you look in 9 greater detail, you can see a number of other 10 interesting features of the cells. These are 11 sectioned electron micrographs which show normal 12 neutrophils and cells from a person treated for five 13 days with G-CSF. And as you can easily see, the cells 14 are bigger. If you look more carefully at the cells, 15 you can see the surface of the cells are smoother. 16 You can see in these cells bits of endoplasmic 17 reticulum or what would be referred to in a laboratory 18 as delivery bodies. You can see differences in a 19 variety of things, including probably the average size 20 of the granules. Suffice it to say the stimulus 21 changes many aspects of the cell morphology, but in 22 general produces cells that are younger looking. 23 If you look by scanning electron 24 microscopy what you see are if this is a normal 25 neutrophil with its rugged surface -- this is a 31 1 neutrophil from someone treated for 5 days with G-CSF. 2 There is more redundancy to the membrane of the cell, 3 and you can see lots of these cells that look like 4 this with scanning electron microscopy. And it is a 5 reflection of the fact that as neutrophils mature, 6 like as people mature in general, they shrink and the 7 membrane shrinks around them. So these changes are 8 not totally unpredictable in terms of general cell 9 biology, but they are rather dramatic to look at. I 10 have often described these cells as looking like 11 someone running down the hall with their white coat 12 flapping behind them. And there are probably many 13 features of how the cell functions that are slightly 14 different for these cells versus these cells, but in 15 general the cells have the same basic function. 16 Tom Price and I did studies in this era of 17 investigation looking at how much does G-CSF stimulate 18 the flow of cells from the marrow to the blood. And 19 this is kind of a classic study, something I learned 20 to do here from Seymour Perry. Studies which show if 21 you injected tritiated thymidine and look at the 22 yellow curve here, this is the normal emergence time 23 for a neutrophil from the marrow. That is, you label 24 with tritiated thymidine the cells at the last stage 25 of cell division, and then watch for those cells to 32 1 appear in the blood. There is normally a lag time of 2 about six days. This has been known since the 1950's. 3 If you inject G-CSF, what you can show with our 4 studies is that you shift this curve to the left. The 5 30 mcg curve shifted it this much and the 300 mcg 6 curve shifted it this much. That is, you reduce the 7 post mitotic transit time for the neutrophil from 8 roughly 6 days to 3 days or reduce it by 50 percent. 9 That is a big change, and so much of what you see 10 reflects this pushing of the cells down the pathway 11 and getting them into the blood sooner, younger, and 12 looking as I just showed you. 13 If you look at a schematic of what happens 14 when you give G-CSF to a normal person for a period of 15 time, you go from a schematic that looks like this 16 with each of these bars representing a cell between 17 divisions, and the number of divisions reflected by 18 the number of forks along the road. You can see that 19 if you give G-CSF, either as a drug or if people 20 produce it in response to infection, you amplify the 21 number of cells produced, and you do it in a shorter 22 period of time by reducing primarily the G-zero phase 23 of cell development. 24 If you look at the function of these cells 25 -- we have done a number of studies and this is 33 1 terribly relevant to the development of this field. 2 If you look at neutrophil function after 3 administration of G-CSF, what you see depends upon 4 when you look. Because you are looking at a dynamic 5 circumstance. Now this is a graph from work we 6 published a few years ago, work with Bob Allen. Other 7 people have done this general line of work. But if 8 you look at the three groups of subjects from this 9 trial we did -- control, 30, and 300 mcg treatment -- 10 and if you use a low stimulus like PMA in a low 11 concentration, what you observe is that cells from the 12 blood of a person treated with G-CSF are primed but 13 they are not actually stimulated or activated by the 14 treatment. And that is to say if you take the cells 15 from the blood and you expose them to a low dose or a 16 low amount of this stimulus, you really see no effect 17 of treatment. However, if you use a high dose of PMA 18 or some other agonist, what you can show is that there 19 is a time-dependent change in the response of the 20 cells to the stimulus. Now these colored bars at the 21 bottom are just a reproduction of the data I showed 22 you a moment ago about emergence time. The purple is 23 the shortest emergence time, which refers to the 24 highest peak here in terms of a change in the PMA 25 response of the cells. And that is to say if you give 34 1 a higher dose of G-CSF, you get cells produced in the 2 blood which are, again, not activated, but they are 3 primed to a greater degree to make a bigger response 4 to an agonist like PMA. 5 Now why does this happen? I think a basic 6 underlying concept is that if you administer a 7 powerful cytokine like G-CSF, you actually alter many 8 aspects of the cells. And we believe this occurs 9 because of effects on coordinated gene expression. 10 Not only are you inducing cell division, but you are 11 actually inducing the enzymes that are packaged in the 12 granules to be different than they would be normally. 13 And we believe that this reflects, in fact, a 14 plasticity in the production of neutrophils that 15 occurs with infections and that is simulated by growth 16 factor administration, so that the effects of 17 treatment, like infection, are actually to produce 18 cells that are more effective than normal cells would 19 be in adaptation of the host, as we have learned over 20 many years occurs in tuberculosis and in other kinds 21 of infectious diseases. 22 Now if you look at the cell surface. I 23 showed you pictures -- if you look at the cell surface 24 and look at the various properties of cells, a number 25 of investigators have shown this. What you see, 35 1 again, in terms of surface properties of neutrophils 2 depends on when you look. If you look at one day of 3 G-CSF treatment, actually the circumstance for most 4 treatment and then transfusion studies, you find 5 relatively little change. If you look at longer 6 periods of time, though, you can see effects on 7 various markers for neutrophil adherence and for 8 function. And shown down here, for instance, in this 9 corner is the substantial enhancement of the 10 expression of CD14, a binding moiety for endotoxin 11 that is induced on neutrophils by G-CSF treatment. 12 Probably greater than any of these is the effect on 13 the expression of the high affinity receptor for IgG 14 on neutrophils, which is greatly induced by G-CSF 15 treatment. And it is conceptually important in terms 16 of the internalization of bacteria by neutrophils and 17 their killing of the organisms. But the full benefit 18 of that effect is not really known. 19 In terms of some effects of these changes, 20 though, there is an interesting experiment that my 21 colleague Conrad Liles did and published just last 22 year. This is looking at the killing of fungal 23 organisms, a focus of interest in neutrophil 24 transfusion therapy, and looking at the effects of G- 25 CSF treatment of a normal person or potentially a 36 1 normal donor and looking at the capacity of the cells 2 to kill this class of organisms. And what this shows 3 is the purple bars being the controls and neutrophil 4 killing then for three classes of organisms -- 5 Candida, Aspergillus, and Rhizopus. And what this 6 shows is if you look at neutrophils from a normal 7 person after five days of G-CSF, there is really no 8 effect on the killing of Candida, which are relatively 9 easily killed by a neutrophil. However, for 10 Aspergillus and Rhizopus, you can show in this kind of 11 a model the induction of an enhanced capacity to kill 12 these organisms. These experiments were done with 13 spores. We are currently doing experiments in Seattle 14 now looking at the hyphae forms of these organisms. 15 But suffice it to say that there is considerable 16 evidence to say that you can use cytokines not only to 17 enhance the number of cells the body produces but also 18 the functional capacities for critical functions like 19 this of these cells. 20 This is sort of a summary of what I have 21 told you. G-CSF in this setting and what is relevant 22 is it increases production by accelerating release of 23 cells leading to the shift of band neutrophils into 24 the blood and other slightly immature cells. The 25 cells that are released are primed for an enhanced 37 1 metabolic response. If you use the right organisms, 2 you can show that you have enhanced microbicidal 3 activity, and actually there are a number of other 4 changes that occur, most of which are changes which 5 would enhance the body's response in an infection. 6 Now one of the most interesting effects of 7 cytokines on neutrophil production that is important 8 in the development of our concepts today is the effect 9 on cell viability. This is a panel from work that 10 Conrad Liles did a few years ago actually looking at 11 just taking a test tube of blood, if you will, or 12 isolated neutrophils and looking at how long those 13 cells survive in vitro. What this shows is normally 14 the blue line neutrophils poop out, right? You know 15 that. If you leave a tube of blood in your pocket and 16 forget to do a count today and test it tomorrow, the 17 count is lower. Neutrophils die by the process of 18 apoposis, their natural process of death, and you can 19 show this in the laboratory very nicely that they fall 20 off over time. This has been a central issue in the 21 conceptualization of how you would ever supply 22 neutrophils for transfusion therapy because they don't 23 last very long. Well, what has been shown by a number 24 of investigators now is addition of G-CSF and GM-CSF 25 or interferon gamma, these are all agents which tend 38 1 to prolong the in vitro survival of neutrophils. And 2 as I will show you in a moment, they also promote the 3 in vivo survival of these cells. 4 Now we took these ideas first to the 5 clinic in about 1993 in work that was performed at the 6 Puget Sound Blood Center in Seattle and at the 7 Hutchinson Cancer Center, and actually there were a 8 number of other groups around the world who were 9 interested at the same time -- a group of 10 investigators here and in several other centers 11 particularly, including the M.D. Anderson Hospital in 12 House. The basic idea that we investigated in Seattle 13 was the concept of providing neutrophil support for a 14 person after bone marrow transplantation to keep their 15 counts from going low. And because we were concerned 16 about alloimmunization and other problems, we used the 17 actual marrow donor, an unfortunate circumstance where 18 we had some twins and syngeneic individuals, so that 19 we could try to optimize neutrophil support through a 20 period of neutropenia using cells collected from G-CSF 21 stimulated donors. 22 This is a summary of the basic plan. We 23 had seven people. We used G-CSF at 5 mcg per kilogram 24 per day. The people donated an average of, as you can 25 see, just over 7 times. We collected cells by 39 1 centrifugation and used starch to accelerate red cell 2 sedimentation. And because it was being done 3 repeatedly, these volunteers became patients in a 4 sense as they had a subclavian catheter implanted. 5 The controls were historical controls of other people 6 not given G-CSF. These were the rather dramatic 7 results of this trial, which was published in Blood in 8 1993. Bill Bensinger is the senior author. 9 What the trial showed is compared to no G- 10 CSF, that the number of cells that were collected were 11 roughly tenfold higher. And more importantly, the 12 increments in the blood of the recipients were almost 13 tenfold higher, with counts measured 24 hours after 14 the transfusion. Now if you are familiar with this 15 field, you know that for many years, you could 16 transfuse lots of cells but you couldn't count them. 17 In fact, it was interesting to review some old papers. 18 If you look at determining hematopoietic recovery 19 after transplantation even though you are giving 20 granulocytes, it wasn't difficult in the old days 21 because you could transfuse the cells. There were 22 never any in the circulation, so you could still see 23 when the marrow recovered. 24 Actually in this trial and in I would say 25 other work since then, what you find is you get enough 40 1 of an increment with transfusing cells from a G-CSF 2 stimulated donor that it makes it hard to recognize 3 when recovery occurs. As you can see here, we found 4 in this trial transfusing roughly 40 billion 5 neutrophils a day that we got a median increment at 24 6 hours approximately of 570, and a mean increment of 7 nearly 1,000. 8 Now we weren't satisfied. And actually 9 the following summer, using a medical student for a 10 graduate honors project, we conducted a randomized 11 trial of giving G-CSF with and with dexamethasone to 12 see if we could use these two agents together to get 13 the counts even higher. Now being interested in this 14 field for a long time, I was skeptical that it would 15 make any difference, but I thought it was worth a try. 16 This is the schedule we used, chosen somewhat 17 arbitrarily. We used the dose of G-CSF 300 mcg that 18 we had used before or twice this amount with and 19 without 8 mg of dexamethasone. This was administered 20 subcutaneously and this orally simultaneously, and all 21 we did was to do blood counts over the next 24 hours. 22 But what you can see is if you look at the 12-hour 23 points, that the addition of dexamethasone to G-CSF 24 substantially increased the levels of the counts. Or 25 as you can see also quite easily, we went in normal 41 1 people from a count of 4,000 to 40,000 in 12 hours and 2 they barely knew it. Most of what they noticed was 3 the effects of the dexamethasone. 4 So we were very impressed at how much we 5 could raise the counts. It is still not known why 6 this occurs. My bet -- but there is on data to prove 7 it -- is that the steroids actually effect the 8 capacity of the cells to be mobilized with the G-CSF. 9 Probably some effect on receptor or post-receptor 10 mechanisms of the cells. Suffice it to say it is a 11 big effect and again it has potentially a large effect 12 in planning or conducting a clinical trial. 13 What we did was to go on and conduct some 14 studies using this combination of drugs, collecting 15 cells, and making measurements. This work is sort of 16 barely history, but it is published in August of this 17 year in The Journal of Transfusion. Here are a 18 picture of the cells collected in this way, nice- 19 looking but young-looking cells. This is just giving 20 G-CSF and dexamethasone and collecting cells 12 hours 21 later. And Dr. Price in a moment will talk more about 22 this. 23 If you look at the general effects on 24 these cells, and Dr. Liles this afternoon will 25 describe this in greater detail, but you can collect 42 1 75 to 100 billions cells now with this technique, 2 which have normal functions, slight changes in their 3 immunophenotype, and again, as I mentioned, an effect 4 on the survival of the cells. What we have shown with 5 isotope labeling studies is in fact that going from a 6 normal half-life of around 8 to 10 hours for 7 neutrophils, the neutrophils collected in this fashion 8 have a blood half-life of roughly 20 hours. So they 9 have a long survival, as I showed you in vitro, and 10 the calculated production rates or turnover rates 11 would be, of course, very large for large increases in 12 the cells with a long survival in the blood. 13 Now just as a transition to what I am 14 going to say about conducting a clinical trial, it is 15 very important to know that there is enough data now 16 to say what happens if you transfuse cells like this 17 into patients. We have done some studies in Seattle 18 transfusing cells from people treated with G-CSF and 19 dexamethasone to patients who are marrow transplant 20 patients with serious bacterial and fungal infections. 21 This is just to highlight this information, but what 22 it shows is if you transfuse a person with almost no 23 neutrophils, you can bring their count to near normal 24 with a transfusion and you can then go up another 25 notch if you give a second transfusion the next day. 43 1 That is, you can normalize the neutrophil level of a 2 person with no neutrophils, something that heretofore 3 was never possible. 4 It is based upon that I feel and I think 5 several people here feel like it is time for a real 6 reconsideration of this idea or a consideration of a 7 clinical trial to evaluate neutrophil transfusion 8 again. 20 years have gone by since this was really 9 undertaken, and there have been a lot of changes in 10 many aspects of medicine which makes this justified. 11 The best choice, although challenging to do, is a 12 randomized control trial using therapeutic transfusion 13 rather than prophylactic transfusion. The biggest 14 problem with this is alloimmunization if you give 15 cells early. So late after transplantation when you 16 really need them, the patient might have a smaller 17 response. And also logistically this is a huge 18 undertaking. 19 In general, the focus of a trial should be 20 on patients who really need it. And those patients 21 these days in marrow transplant centers and in most 22 intensive cancer centers are on difficult to treat 23 organisms, particularly yeast and molds and some 24 resistant bacteria. We believe in contrast to earlier 25 eras that the best way to proceed would be with 44 1 cooperation, a multi-center trial. I showed you 2 reports of small, relatively inconclusive trials, and 3 part of the problem was how they were conducted. And 4 finally, you need to do them with some standard 5 approach to the patient care, the other aspects of 6 patient care, in order to be certain about what you 7 see. 8 In terms of trial design then, the best 9 ideas are to use people who have marrows that are 10 expected not to recover quickly, that is, they have 11 received aggressive chemotherapy or transplant. We 12 believe that this is an applicable approach to 13 patients with neutropenia at present. There may be 14 other ideas, but that should be the focus. And as I 15 mentioned, fungal infections and people preferentially 16 chosen to be not demonstrated to be alloimmunized 17 before transfusion support is given based upon much 18 evidence that you can alloimmunize somebody and not 19 get a response. And the basic idea in a randomized 20 trial should compare if patients have fungal 21 infections or bacterial infections that they receive 22 a standard therapy, a standard therapy plus 23 neutrophils procured in what I would say is a modern 24 way. 25 Now there are many issues, and what I said 45 1 may make it sound simple, but I assure you it is not. 2 That is, we are still studying issues around 3 mobilization strategies. How much difference does it 4 make about which drugs and how much you give and when 5 you give them. There are issues still about the 6 quality of cells. I have said that cytokines affect 7 cell formation and function, and that diversity of 8 effects needs to be considered in terms of the actual 9 trial design. There are lots of issues related to 10 donor willingness and safety. It is amazing in this 11 country the diversity in terms of the willingness of 12 people to give blood. Reasons that are very complex. 13 And if you think about another layer of complexity, 14 that is accepting the idea of being treated with a 15 drug before you give blood, you can imagine that there 16 are many aspects of this to be considered if a trial 17 is to be conducted well and conducted safely. There 18 are many issues, some of which will come out today, 19 about recipient benefits and risk, and then there is 20 the question of having in a trial design good, clear, 21 acceptable evidence of therapeutic efficacy. 22 In terms of how this should be done, I 23 believe and I think many do that neutrophils are there 24 to kill microbes. And so a trial should be designed 25 primarily to show an effect on microbes, that is, the 46 1 clearance of infection. There are many secondary 2 endpoints, though, that are important. Important 3 socially, important for patient well-being, important 4 for the people who pay for medicine in this country. 5 So that one has to decide a trial with enough other 6 information in it so that you can apply the results of 7 a trial in the clinical and economic settings where we 8 practice. 9 My final point in terms of general 10 comments about a clinical trial comparing the past 11 with the present is that we live in a very dynamic 12 world. Look at the paper today, right? And in this 13 particular field, we live in a world that is changing 14 rapidly. I have said enough, I bet, to convince some 15 of you, if you weren't already convinced, if you went 16 home tonight and it was your mother or father or 17 sister or brother and I offered you this, you would 18 say, of course. And if I said you might not get it in 19 a randomized trial, you would say, oh no, I don't want 20 to participate. Because times have changed. This 21 approach does show considerable promise. And I think 22 that the window of opportunity in clinical trials in 23 this field, like in other fields, is relatively small. 24 Because knowing as I do people around the world who 25 work in this particular area, many people are taking 47 1 this approach and taking it without really firm 2 evidence that it is a clinical benefit. So the time 3 for a clinical trial is relatively brief. 4 I am going to stop at that point. Again, 5 it is great to be back at the NIH and thank you very 6 much. 7 CHAIRPERSON HARVATH: That was really 8 great. Thank you very much. It is a pleasure to 9 introduce the next speaker, Dr. Thomas Price, who is 10 also a Professor of Medicine at the University of 11 Washington and the Director of Puget Sound Blood 12 Center. He is going to speak to us about his 13 experience of cytokine administration to normal 14 granulocyte donors and some other really great 15 information I think you are going to add as well. 16 Thank you. 17 DR. PRICE: Thanks, Liana. Thanks to you 18 and the organizers for inviting me here. If I could 19 have the first slide. What I am going to do today is 20 to share with you our experience, which is an ongoing 21 experience, with a trial of neutrophil transfusion 22 that we are doing in collaboration with the people at 23 the Fred Hutchinson Cancer Research Center involving 24 patients that are undergoing bone marrow 25 transplantation. And I would just like to tell you 48 1 kind of where we are with this. 2 This slide you have seen before. It is to 3 kind of remind you of the mobilization that we are 4 talking about here. And as David said, we looked at 5 these five different mobilization regimens here. The 6 dotted lines are the ones that included dexamethasone 7 in addition to the G-CSF, and as might be obvious from 8 looking at this slide, in the trial that I am about to 9 show you, we picked the one that we thought was going 10 to give us the most granulocytes. So that is why the 11 study that you will see and that I am going to talk 12 about now involves the 600 mcg dose of G-CSF and also 13 dexamethasone. Now whether it will turn out that 14 there is really that much difference between these top 15 two, we won't be able to say as a result of what I am 16 going to tell you. 17 The other thing to note, of course, is 18 that the timing in here is to suggest that doing this 19 12 hours before you collect the neutrophils would 20 probably be the best time to do this. So this is what 21 we aimed for, but keeping in mind with the logistics 22 of when donors can actually show up and when you can 23 actually collect the cells. As you see, we don't 24 always hit the 12 hours right on the mark. 25 The basic design of this ongoing study is 49 1 shown in this slide here. This is sort of a Phase 2 I/II study, as I said a collaborative study between 3 the Blood Center -- this is the Puget Sound Blood 4 Center, it is not a misprint for peripheral blood stem 5 cells here -- and the Hutchinson Cancer Center. One 6 of the wrinkles on this thing is that the design here 7 is to use community donors. Now most of the studies 8 that have been reported using G-CSF stimulated 9 granulocyte donors have been when the donor has been 10 a family member or friend of the patient, a fairly 11 captive person that you can lasso and do this to. The 12 idea here was to see if we could supply granulocytes 13 to patients as they needed them using community 14 donors. These are donors that are just ordinary blood 15 donors who have volunteered to be, for the most part, 16 platelet donors or to be apheresis donors for patients 17 that they don't even know. Could we involve them in 18 such a process? 19 As I said, the dose mobilization was to 20 give them 600 mcg of G-CSF and 8 mg of dexamethasone. 21 This was done as close as possible to the 12 hours 22 prior to the collection procedure. The collection 23 procedure itself was routine. We used a COBE spectra 24 machine. We used the high molecular weight 25 hydroxyethyl starch, the hetastarch, as the red cell 50 1 sedimenting agent, and we processed 10 liters of blood 2 for these collections. 3 Now the patients were all patients on the 4 bone marrow transplant ward and the Seattle transplant 5 ward at the Hutchinson Cancer Center. Most of these, 6 as you will see, were patients who had already 7 received a transplant, although there were a few 8 patients in there who were pre-transplant. They are 9 all neutropenic. The idea was to limit this to people 10 who had 100 neutrophils or less. And they were people 11 with documented fungal or resistant bacterial 12 infections. 13 Now the original plan or the goal of this 14 thing were these three things listed here. One of the 15 things we wanted to do was to evaluate the feasibility 16 of using community donors. We started out by just 17 calling some pheresis donor up and saying how about 18 coming in and getting a shot of G-CSF. We had no idea 19 how easy it was going to be to convince people to do 20 this and whether we could basically supply with any 21 kind of regulatory these components. We also wanted 22 to see what we actually could get in neutrophil yields 23 by using this sort of a mobilization strategy. The 24 slide before was, as David said, just some normal guys 25 that we gave these things to and did some blood counts 51 1 on. The proof of the pudding was what we would 2 actually get when we hooked them up to a machine. 3 And finally, we wanted to determine what 4 the hematologic effects were going to be in the 5 recipients. What was going to happen to them in terms 6 of their counts and in terms of where these cells 7 went. It would have been nice to make some sort of 8 clinical determination of whether this was efficacious 9 in the recipients, and we were going to look at that. 10 But right from the beginning we knew we weren't going 11 to have enough patients probably to really make a 12 determination that was convincing of clinical 13 efficacy. 14 Let me turn a little bit to the community 15 donor recruitment business. The way this worked was 16 we have got this pool of 4,000 or 5,000 people who 17 have signed up to be pheresis donors. For the most 18 part, these people are platelet donors, but they are 19 also subject to being called for a granulocyte 20 collection which traditionally has involved taking 21 some prednisone as a stimulating agent. The idea was 22 that this regular donor list was that the donor would 23 be contacted by the regular pheresis scheduler, 24 somebody that they would almost know because they talk 25 with them fairly frequently, and be informed that we 52 1 had a study going that we thought was likely to be 2 able to improve the product, but it did involve them 3 taking a drug and would they be interested in 4 participating. If they were, then they were scheduled 5 for a donation, but they were also then put in contact 6 with the study nurse coordinating this study who went 7 over the study in detail with them and basically went 8 through the informed consent procedures as to what 9 this was all about. 10 Then what happens is the 12-hour ahead of 11 time visit, the donor comes in to one of the Blood 12 Center's fixed sites. We have five sites scattered 13 around the Seattle area and the donor could come to 14 any one of these sites 12 hours before. Now as a 15 practical matter what this meant is that we tried to 16 shoot for 12 hours, but as a practical matter it was 17 somewhere between 8 and 16 hours ahead of the 18 scheduled leukopheresis. What this means is that you 19 can't just do a leukopheresis at any time. The 20 leukopheresis had to be either scheduled at the crack 21 of dawn so that the donor could come in the evening 22 before at a reasonable hour -- come in at dinner time, 23 you know 7:00 or 8:00 at night, and then we could do 24 an early morning collection. Or alternatively, the 25 donor could come in at the crack of dawn for the pre- 53 1 visit and have the blood drawn at 7:00 in the morning 2 and then we would be able to do a leukopheresis 3 procedure in the 5:00 in the afternoon sort of range. 4 But it does sort of limit you because you can't do a 5 collection procedure at 1:00 in the afternoon because 6 there is no right timing for the donor to come in for 7 the ahead of time visit. 8 Anyway, when they do come in, the consent 9 form is signed. They go through a preliminary donor 10 screening, just to make sure that there are no 11 surprises there and 12 hours later we are not going to 12 find out that the guy had hepatitis last year. We 13 also draw blood for the routine things. We draw blood 14 for a CBC and also for the ordinary testing -- for the 15 ABO, the Rh antibody screen and for infectious disease 16 testing. And then, of course, the G-CSF is 17 administered subcutaneously and the dexamethasone is 18 given for the donor to take. 19 Now since this is an FDA sponsored 20 conference, I did want to make one what I think is 21 important point about the infectious disease testing. 22 What we did in this study is considered that the 23 testing that we did on the sample obtained 12 hours 24 prior to the leukopheresis, we considered this to be 25 the testing of record for the collection. Now for 54 1 those of you who are in blood banking, this is not the 2 usual FDA approach to things. The usual FDA approach 3 is to say that you have got to draw the blood sample 4 that you are going to do for the testing actually at 5 the time that you are collecting the blood component. 6 Now I think it is important, though, that we be 7 allowed to do this as I have described it here and 8 count this as the testing of record because of 9 basically what David was saying about the storage 10 capabilities of these cells. With the current 11 techniques, the neutrophil integrity is likely to be 12 compromised if we store it waiting for these tests. 13 With the more sophisticated testing -- I mean, stuff 14 that we have done in the past basically has shown that 15 if you store cells for 24 hours, in terms of the 16 cell's ability to localize to an inflammatory site, 17 the cell loses about 75 percent of its activity at 24 18 hours. So it is very important, at least now, to give 19 these cells as soon as possible after collection and 20 we can't really wait until all the testing is done. 21 And this is getting worse because the time for testing 22 keeps getting longer. When PCR comes along, it is 23 even going to be worse. Also, it is not -- you know, 24 the testing lab is not necessarily right next door 25 these days to the place that you are drawing this 55 1 blood. It may have to be sent off to a different 2 contract place that is doing the testing. So this 3 timing gets worse and these things do not store well. 4 So I guess I am making the pitch that should the FDA 5 decide they want to write some rules for this, that it 6 would be very important to be allowed to have the 7 testing of record be this sample that we draw the day 8 before, let's say. 9 Now it may turn out, to be optimistic 10 perhaps, that one of the effects of G-CSF will be that 11 we will be able to store these cells better. But 12 there is really no data on that in vivo yet. So that 13 is just totally an unknown right now. 14 Okay, well what happened? We had as part 15 of this -- what I am going to tell you about so far is 16 our experience with 19 patients that we were trying to 17 provide granulocyte support for. If we started at the 18 time that these patients were identified up until the 19 time that granulocytes weren't needed any more, there 20 were 233 slots that had everything gone swimmingly we 21 would have had a collection for each of these slots. 22 As I said, we have about 4,500 donors in our pool that 23 we have available to call. One of the things that 24 surprised us, I guess, was that when we did contact 25 these donors and said how would you like to do 56 1 something kind of experimental and get a subcutaneous 2 injection of a drug, about two-thirds of them said 3 fine, I will be right in. Now because of various 4 logistic things and part of having to deal with this 5 timing that I am telling you about and the time slots 6 that had to be available and the donor had to be able 7 to fit, in fact we only succeeded about 75 percent of 8 the time in getting somebody actually when we wanted 9 them. That was more of a logistic problem than it was 10 a problem of not being able to find a donor. 11 What I am going to report to you now is 12 the results of 175 collections with this stimulation 13 here, this 600 of G-CSF and 8 of dexamethasone. A 14 little bit about donor side effects. You have heard 15 many times and those of you who were here yesterday 16 heard again the story of donor side effects from G- 17 CSF. The experience we have had at the Blood Center 18 has been similar to all of this. Most of the donors 19 experienced some side effects from this pre- 20 stimulation. Mild to moderate in the vast majority of 21 donors. With these 175 donors, 40 percent experienced 22 some sort of bone pain, 30 percent headache, 30 23 percent insomnia, probably more due to the 24 dexamethasone than to the G-CSF. About a third of 25 these patients or a quarter to a third of the patients 57 1 had no side effect at all. 2 Now the severity of these side effects I 3 think you can judge by the fact that of the donors 4 that donated, when we asked them later how big of a 5 deal this was, 98 percent of them said that they would 6 be more than willing to come back and do it again. 7 What was the experience in how much we 8 actually got. The donors neutrophil count at the time 9 prior to getting the G-CSF on the sample that we drew 10 12 hours ahead of time was normal. It averaged 3,700 11 with this sort of range. The time interval between 12 getting G-CSF and the beginning of the collection 13 averaged 13 hours. As you can see here, it varied 14 with an extreme for 5 and 23 hours. The donor 15 neutrophil count right before the collection was 16 almost 31,000, varying here between 14,000 and 56,000. 17 This is the neutrophil count now and not the white 18 count. And the number of neutrophils that we got 19 averaged 82 billion. It ranged between 24 billion and 20 144 billion. Now just to remind you of the numbers 21 that David showed you before, the traditional 22 neutrophil yield that is obtained by modern cell 23 separators using corticosteroids alone as a stimulus 24 is usually in the 20 to 30 billion range. So you can 25 see here that this now is two to three times the yield 58 1 that you normally get without using G-CSF. 2 Now this slide here shows you the 3 relationship between what the donor's neutrophil count 4 is right before the pheresis procedure and what we get 5 in terms of the yield going all the way up to -- I 6 can't really read that but it looks like 160 billion 7 cells there. This, again, is neutrophils and not 8 white blood cells. I think you can see that there is 9 a little bit of scatter here, but there is clearly a 10 pretty linear relationship between what the donor's 11 neutrophil count is and what you get out of this. It 12 nicely extrapolates down to zero. What you see here 13 actually as the small dots are the 175 dots 14 representing the donors that I am telling you about. 15 The heavier dots there are another roughly 20 16 collections that we have done where we used only 600 17 mcg of G-CSF and did not give the donor dexamethasone. 18 I think you can see that you basically get a lower 19 neutrophil count and you get less yield. Remember 20 that the average neutrophil count in the 175 was about 21 30,000. The average for the G-CSF alone is about 22 22,000. The average yield is 82 billion for the 23 combination and about 56 billion for the G-CSF alone. 24 But the interesting point is that the G-CSF alone 25 dots, the big dots, appear to be in the same continuum 59 1 as the other dots. It is just a matter of how high 2 you get the count. So it looks like this is kind of 3 a validation of the idea that the higher you get the 4 count, the more cells you are going to get. It is 5 sort of a no-brainer. But if we could figure out a 6 way to routinely get the cell count up to 60,000, we 7 might be able to get a lot more cells. We might get 8 the average yield up to 160 billion. 9 Now who are the recipients here? These 10 are the 19 patients that we gave these cells to. 15 11 of them had had a bone marrow transplantation and 4 of 12 them, as I said before, were pre-transplant. 16 of 13 these patients had a fungal infection, 8 fungemia and 14 8 an invasive infection. Most of these are 15 Aspergillus infections, either pulmonary or sinus 16 infections. And 4 of the patients had resistant 17 bacteremias. You can see that this adds up to 20, 18 which means that one of the patients had two 19 infections. 20 Another item which I think is important to 21 note about these guys is that in general this 22 population was not an alloimmunized population. We 23 did HLA antibody screens at the beginning of the 24 transfusion support on all these folks and the screen 25 was negative in 15 of these 19 patients. There was a 60 1 little positivity, but not very much. The PRAs were 2 less than 8 percent in 4 of them. So none of these 3 patients was highly alloimmunized to begin with and 4 most of them had no evidence of alloimmunization. 5 What did we see in the patients in terms 6 of the hematologic results? There was an average of 7 8.6 transfusions per patient. It ranged from 1 to 25 8 transfusions. You have already seen that the average 9 dose delivered was 82 billion cells. Now this is what 10 happened to the patient's neutrophil count. I have 11 listed two things here. One is the one-hour 12 increment, that is, comparing the neutrophil count one 13 hour after the transfusion with the count immediately 14 prior to the transfusion. You can see here that the 15 average was about 2,600. And as David mentioned, this 16 is in marked contrast to the usual experience with 17 granulocyte transfusions where one didn't see any 18 increment and we always used to say, well, that is 19 because they are all doing what they are supposed to 20 do and going to the site of infection. But in this 21 situation, you actually do see a substantial increment 22 in the neutrophils. And you can see that this varies 23 from one guy who actually had a negative increment who 24 had a fairly high count to begin with pre-transfusion 25 to a very high neutrophil increment. 61 1 The other thing to note is that these 2 cells stick around. They do drop off as the day goes 3 on, but if you do a count the next morning, the 4 average count in these people was 2,600. This is just 5 a coincidence that these happen to be the same number 6 -- varying anywhere from nothing to 15,000. So that 7 you can see on the average we are taking patients who 8 begin severely neutropenic and we can, again on 9 average, convert them from somebody severely 10 neutropenic to somebody who has a sustained neutrophil 11 count which is normal or near normal. Now I will also 12 take you back to the other slide David showed you of 13 the sequential days and the sawtooth sort of thing 14 where if you would actually pull out this next AM 15 count, of course you start out at zero in these 16 patients and after the first -- it goes up for one day 17 and comes back down again but not quite back down to 18 where it started from, and then on day two you get it 19 up a little higher and you can sawtooth this thing up. 20 So that the general experience is after a few days, 21 the patient is running a neutrophil count that is 22 often in the normal range. 23 Now you will note some low numbers here, 24 though. Two of these patients got no increments. So 25 this is the average, but a couple of them got no 62 1 increment at all. 2 Now the next thing here is what we are 3 calling the buccal neutrophil count. This is kind of 4 a crude way of determining whether these cells are 5 capable of leaving the circulation and getting to the 6 tissue sites where it is important that they do their 7 work. And what we do to this thing is have the 8 patient take 24 cc of saline into their mouth, swish 9 it around, spit it into a can, and then by staining 10 the cells and counting them, we can count how many 11 neutrophils are in the guy's spit. What we find is 12 that when we did this before we started the 13 transfusion support, basically this is in millions, 14 the average is .01 with this sort of a range here. 15 Post-transfusion -- and generally we made these 16 measurements the next morning after the transfusion -- 17 you can see that on average there were about a half a 18 million cells in there with this sort of a range. Now 19 if we were -- just to give you an idea of the normal 20 numbers -- if we were to do this same little test in 21 everybody in the room here, the average amount of 22 neutrophils in a normal person is about a half a 23 million. So these guys on average were going from 24 nothing up to where the amount of neutrophils in their 25 mouthwash was normal. So these cells are capable -- 63 1 they don't just circulate, they are capable of 2 extravascular migration and getting supposedly to 3 where they are supposed to go. 4 Now again you can see that there is a 5 range here. In fact, there turned out to be a 6 correlation between this. The guys that got no 7 increment in the blood also got no increment in the 8 buccal neutrophils. This is what you would expect, I 9 guess, but it sort of validates that maybe we really 10 are measuring here something that means something. 11 What happened in terms of the side 12 effects? I think it has been mentioned here earlier 13 that one of the concerns of giving much larger doses 14 of neutrophils, particularly neutrophils that have 15 been primed by G-CSF, the early concern was that this 16 might give an awful -- this might sort of exaggerate 17 transfusion reactions and might exaggerate in 18 particular pulmonary transfusion reactions and be a 19 dangerous thing to do. So we are looking here then at 20 the 175 collections, but only 165 of those ended up 21 being transfused. If we look at the traditional sorts 22 of things here like chills and fever, you can see that 23 in about 7 percent of the transfusions, one saw that 24 patients got chills and some fever. This meant that 25 in these 19 patients about a third of them at one time 64 1 or another in one or more of their transfusions had 2 this experience. These were mostly mild to moderate, 3 in fact. And actually what tended to happen was these 4 patients would then on subsequent transfusions be 5 premedicated with Tylenol or something like this and 6 most of the time they did not recur. So they tended 7 to be things that were easily handled by the usual 8 pre-medications you might give. There were an unusual 9 number or a low number of itching-hives type 10 reactions. 11 Now the other thing we do since we were 12 particularly concerned about the pulmonary reactions, 13 is we measured oxygen saturation by oximetry prior to 14 the transfusion and after the transfusion. You can 15 see here that the baseline oxygen saturation was about 16 95 percent. It varied between 61 and 100 percent. 17 Some of these patients, particularly the guys with 18 pulmonary Aspergillus, might not start out with a 19 normal oxygen saturation. On average, the change was 20 basically not existent. But if you looked at these 21 individual things, of these 165 transfusions, the 22 oxygen saturation decreased by 4 percent or more in 23 only 11 of the transfusions and by 6 or more in only 24 three of the transfusions, and in those three it 25 actually ended up below 90. The important clinical 65 1 point to make is that in no transfusion of these 165 2 was it ever the clinician's view that something had 3 happened. These were just kind of measurements that 4 were made, but there was no obvious pulmonary 5 deterioration that was attributed to the transfusion. 6 So the fear that suddenly we could have set ourselves 7 up for a real dangerous transfusion reaction doesn't 8 appear to have happened. It is important to remember, 9 though, that these patients are not highly 10 alloimmunized people. It is not that kind of set of 11 patients. 12 Now a little bit about the HLA 13 compatibility. And you can tell by the fact that it 14 is a cheaper looking slide that this is preliminary 15 data. What we did is we obtained serum samples on all 16 of these patients before we gave the first transfusion 17 and then weekly thereafter until the patient was off- 18 study. We also, every time a donor came in, obtained 19 lymphocytes from that donor and froze them. So that 20 after we are done with the patient, we can 21 retrospectively come back and in sort of a batch run 22 all those serum samples against all those lymphocytes 23 and basically do a lymphocytotoxic cross-match for 24 each transfusion that was given. We have those 25 results available on 12 of the 19 patients so far. 66 1 What happened was if we looked at all these serial 2 blood samples, there appeared to be an antibody to one 3 or more of the donors in 4 of these 19 patients. In 4 other words -- well, I should say 12 -- the people 5 that are finished. In 8 of these guys, it was clean. 6 The lymphocytotoxic cross-match was negative in 7 everything. But in 4 of them, there was a reaction to 8 one or more of the donors, and it turned out to be 14 9 potentially incompatible transfusions. Now I say 10 potentially because some of these things were 11 situations in which a late serum sample might show a 12 reaction to an early donor, but whether or not the 13 antibody was there when that donor was given, that 14 might not have happened. But for the purposes of this 15 analysis, I would have to assume that if it ever 16 happened, it might have been there at the time of the 17 transfusion. 18 Well, if you look at these 14 potentially 19 incompatibles, what you find is that of the 14, none 20 of them was associated with chills and fever in the 21 patient. In one of them, there was somewhat of a 22 decrease in the oxygen saturation. The average 23 neutrophil increment at one hour was 2,300 and the 24 next A.M. count was 2,200. So that if you will 25 remember the overall group, these were both 2,600. 67 1 Just on the service of it very preliminarily, it 2 doesn't look like these HLA antibodies that showed up 3 during the course of this really had any influence on 4 the transfusion reactions or the hematologic results 5 of the transfusion. 6 Well, how about what finally happened to 7 the patients. The reasons for discontinuing the 8 transfusions are listed here. In 7 of these 19 9 patients, we stopped because the patient's neutrophil 10 count was high on its own or the patient had grafted. 11 In 3 of them it was stopped because the infection 12 appeared to be gone. And in 9 of them it was stopped 13 either in 6 because the clinical situation was 14 determined by the clinician to be futile and support 15 was withdrawn, and of course the ultimate futile 16 situation when the patient died. 17 If you take all of these patients, 9 of 18 the 16 survived until engraftment and 8 of the 19 19 cleared the infection. If we sort this out by the 20 kinds of infection, about half of the patients with 21 the fungal infections, whether it was either fungemia 22 or Aspergillus, survived until engraftment. That was 23 true of everybody with bacteremia. In terms of 24 clearing the infection, about half of the fungemia 25 patients cleared the infection, nobody with 68 1 Aspergillus was thought to have cleared the infection 2 and all of the bacteremias were thought to clear the 3 infection. 4 Now what to make of this. This is one of 5 these things that you can read as the glass is half 6 full or the glass is half empty, I think. I would say 7 that the general impression of the clinicians on the 8 ward was to be impressed that this was probably useful 9 therapy. There were a number of these patients that 10 anecdotally were people that they thought normally 11 would have done very badly who ended up clearing the 12 infection or surviving longer than they thought they 13 otherwise would have, but that is obviously just a 14 clinical impression anecdote style and may or may not 15 hold up. 16 So in summary so far, I think we can say 17 that perhaps surprisingly that community apheresis 18 donors are fairly easily recruited for G-CSF 19 stimulation, that such stimulation in normal donors 20 results in marked neutrophilia and greatly increased 21 neutrophil yields, that when you transfuse these 22 concentrates into patients, this can result in normal 23 or near normal neutrophil counts in the recipients 24 with migration of transfused cells to extravascular 25 sites, and that although the clinical impressions are 69 1 sometimes impressive and the clinicians can be 2 convinced that they are really being useful, I think 3 we can't really say that based on these numbers and 4 that we really are going to need control trials to 5 assess the clinical efficacy. Thanks very much. 6 CHAIRPERSON HARVATH: Thank you very much. 7 That was very interesting data. Before taking the 8 break, Dr. Epstein, who is the Office of Blood 9 Director in the Center for Biologics was sitting next 10 to me and we were talking about the question that you 11 had asked of the FDA. So before he has to leave, I 12 would like Dr. Epstein to address your question and 13 then we will take our break and then we will assemble 14 a panel after the last two speakers of the morning 15 session. 16 DR. EPSTEIN: Thank you very much, Liana. 17 Just a brief comment. The current regulations require 18 the donor sample to be obtained on the day of 19 collection, and we can interpret that broadly to be 20 within 24 hours, certainly encompassing 12 hours. 21 There is no requirement that the sample tested be 22 integral to collection, although clearly we prefer 23 that. Additionally, if it proves to be infeasible to 24 test within 24 hours, the regulations provide for the 25 possibility of exceptions to the regulations, and you 70 1 simply have to request an exemption. It is under 21 2 C.F.R. 640.120. So I think that there really isn't a 3 regulatory obstacle to doing what is scientifically 4 and medically appropriate. We just have to be in the 5 right dialogue. 6 CHAIRPERSON HARVATH: Okay. On the panel 7 discussion, the presentations this morning also raised 8 some very interesting questions that I think would be 9 good to ask of the panel and get all of your feedback 10 on for the experiences you have had. One of the 11 things I think that will be helpful to us will be your 12 collective experience on whether this is going to be 13 a major obstacle and what your experience is with the 14 testing of the products. Also, there was early report 15 in the literature that perhaps some cytokines may 16 alter some of the test results, and I know that we 17 have heard this, and I think it would be very 18 interesting to pose that question to those of you who 19 have been collecting these products and actually 20 performing the routine tests on your donors. So it 21 would be something very interesting to hear of all the 22 speakers. 23 I would like to give everyone an 24 opportunity to take a 15 minute break and we will come 25 back here to begin the second part of the morning 71 1 session at 10:00. 2 (Whereupon, at 9:44 a.m. off the record 3 until 10:10 a.m.) 4 CHAIRPERSON HARVATH: We are going to try 5 and get started. Our next speaker, Dr. Dan Ambruso, 6 is a Professor of Pediatrics and Associate Professor 7 of Pathology at the University of Colorado Health 8 Sciences Center, and he is the Associate Medical 9 Director of Bonfils Blood Center. It is a pleasure 10 for me to introduce Dan, and he is going to talk to 11 you about his experience regarding the functional 12 properties of granulocytes that he studied from donors 13 after G-CSF administration. Dan? 14 DR. AMBRUSO: Thank you, Liana. It is a 15 pleasure to be here this morning. I am happy to be 16 involved in this workshop. I am going to present to 17 you some information that we have on normal volunteers 18 who received G-CSF, and I will say at the outset that 19 Tom Price has talked about response that his donors 20 had with a single dose of G-CSF, and I am going to 21 talk about a project that we were involved with where 22 our patients received five doses of G-CSF and we 23 looked before and after the administration and tried 24 to focus on neutrophil function. 25 This seems initially a little bit farther 72 1 away from the practical aspects of blood donors for 2 granulocytes where a single dose might be more 3 practical. On the other hand, I am sure all of you 4 are aware of the fact, and this was brought up at the 5 meeting yesterday and it has been brought out in the 6 literature recently that there may be a number of 7 paradigms including multiple dose administration of G- 8 CSF to not only collect stem cells but also 9 granulocyte support for these patients. So I think 10 this information has relevance to granulocyte 11 collections. 12 The objective of this talk, as I stated, 13 is to review detailed studies of neutrophil functional 14 capacity obtained during G-CSF administration, 15 multiple dose administration. I would also, if there 16 is some time at the end, present some preliminary 17 findings of functional capacity of neutrophils stored 18 in the presence or absence of G-CSF. That was part of 19 the study as well. And then I would comment on areas 20 that we think need further study. 21 The previous two speakers covered this 22 part of the talk and I don't need to go into this in 23 detail. Everyone knows that the effects of G-CSF 24 include increased numbers of mature neutrophils 25 enhanced, and I put that with a question mark because 73 1 I am in agreement and I will show you data to 2 substantiate this, that neutrophils are really 3 different. But the concept of a neutrophil which is 4 red hot and angry and ready to explode and kill 5 anything in its path is probably not what we get when 6 we mobilize and we treat patients with G-CSF. And one 7 of the other interesting and important effects of G- 8 CSF is prolongation of time to apoptosis and its 9 effect on program cell death. 10 Our clinical protocol is summarized here. 11 We had healthy adult volunteers and in subsequent 12 slides I will call these patients. There were 9 males 13 and 5 females. We administered G-CSF at a dose of 10 14 mcg per kilo subcutaneously for 7 days. Some of these 15 patients were part of a control trial for the ACTG 16 stem cell mobilization study. We looked and took 17 peripheral samples or samples of peripheral blood 18 before the first dose and after the fifth dose of G- 19 CSF. In the studies, when you look at the data, day 20 0 is the first day. So we actually sampled before the 21 day 0 dose, and day 4 is when they received the fifth 22 dose of G-CSF. In 8 of these subjects, we completed 23 granulocyte collections after the fifth dose. The 24 collections were completed using the COBE Spectra with 25 hetastarch and the granulocyte products that we 74 1 obtained were stored in the absence or presence of an 2 added G-CSF 25 nanograms per ml at 22 to 24 degrees 3 Centigrade in a stationary state. Samples were 4 removed from these products at 24 and in some cases 48 5 hours of storage for analysis. 6 I will start out with some of the adverse 7 events. To summarize, as has been other people's 8 experience, there were very few adverse events in this 9 administration. All of our patients had a mild 10 headache and bone pain. We would rate it as grade 1 11 to 2. One patient dropped out of the protocol at the 12 end of the fifth day or fifth dose in apheresis, but 13 all of the rest had much milder problems. It usually 14 peaked by day three of administration and usually 15 treated well with either ibuprofen or acetaminophen 16 with resolution of the symptoms. All of the symptoms 17 completely resolved within 24 to 48 hours of 18 discontinuing G-CSF. 19 Just a few words about the quantitative 20 response. As with other studies that have looked at 21 mobilization of neutrophils with G-CSF, we saw a 22 marked increase in the leukocyte count from a mean of 23 4,870 per microliter up to almost 32,000, an almost 24 tenfold increase in the absolute neutrophil count, and 25 a marked increase in the percentage of band forms in 75 1 these individuals. 2 Now I show this slide as a prelude to 3 reviewing the functional data on the neutrophils. 4 Most of you don't need a lecture on neutrophil 5 function, but I do this to give you a perspective and 6 a focus on how we organized our function studies. As 7 you know, neutrophils travel in the laminar flow of 8 the blood stream until they identify an area of 9 inflammation, exhibiting a rolling behavior at first 10 and then finally sit down with firm adherence, 11 diapedese through the endothelial barrier and move 12 towards the area of infection or inflammation. Once 13 they get there, they ingest the microorganisms, which 14 does two things. Associated with ingestion is 15 activation of the respiratory burst, the neutrophil 16 NADPH to oxidase enzyme system, which is responsible 17 for initiation of production of oxygen radicals and is 18 associated with oxygen dependent killing. In 19 addition, there is a variety of contents in the 20 granules which then are released into the 21 phagolysosome which affect oxygen independent killing. 22 So we essentially focused into two general 23 groups, those kinds of function and biochemical 24 parameters that are related to adhesion and motility, 25 and those that are related to microbicidal activity 76 1 and degranulation or the status of the granules. 2 I present this also because again of the 3 concern that over the past 10 years or so, somehow we 4 have the concept that the neutrophil that is produced 5 under the influence of G-CSF, this is not only for 6 donors but certainly for patients who get G-CSF, is a 7 neutrophil that is charged and ready to go. And my 8 concern is, and this certainly has been presented by 9 Dr. Price in the last talk -- my concern is that in 10 fact if this were so, we would be in big trouble. 11 When the neutrophil is able to get to the site of 12 infection and to eliminate the organisms, that is one 13 issue. If the neutrophil is charged on the 14 endothelial surface and is activated on the 15 endothelial surface, you get excessive inflammation 16 and probably you get -- this is responsible for a lot 17 of multi-organ failure syndromes which we see 18 certainly in the lung and perhaps other organs. So my 19 concern about this has always been that if the 20 neutrophils are so charged, we are going to be putting 21 patients at risk for these multi-organ failure 22 syndromes. Certainly that is not borne up in the 23 patients who have gotten granulocyte transfusions, but 24 perhaps the patients who have received granulocyte 25 transfusions or G-CSF for clinical indications. 77 1 The first thing we did was to look at 2 chemotaxis. This was done in a modified Boydian 3 chamber. The results are summarized here. I hope 4 that is in focus for all of you. What you can see is 5 that we looked at non-directed migration towards 6 buffer and we looked at migration with zymosan 7 activated serum, which is essentially C5A. The open 8 bars are the 0 values and the closed bars are day 4 9 values for controls and for the patients, that is, the 10 volunteers who received G-CSF. And what you can see 11 here is a marked decrease in the motility in this 12 Boydian chamber assay in response to zymosan activated 13 serum. There is a mild effect, although this isn't 14 statistically significant, in terms of directed 15 migration. So there seems to be in the neutrophils 16 that are circulating after the fifth dose of G-CSF 17 administration, there appears to be a decrease in cell 18 motility. 19 We looked at two other parameters that are 20 related to cell motility and might be a reason for the 21 reduced motility. One is to look at the expression of 22 CD11B, which is one of the major adhesion proteins for 23 the neutrophil. And we looked at the expression of 24 CD11B in response to 4 ball ester at the concentration 25 noted here, FMLP, which is a bacterial tripeptide and 78 1 platelet activating factor. This slide summarizes the 2 results for day 0 patients and controls and day 4 3 patients and controls. And what you can see is that 4 expressed as a ratio, the stimulated up-regulation of 5 CD11B was no different in the treated patients. In 6 addition, what I don't have here was the baseline 7 expression of CD11B, which was not increased in the 8 patients on day 4 of treatment. So we could not blame 9 the decrease in motility on a change in perhaps an up- 10 regulation in CD11B. 11 One of the other biochemical correlates 12 related to motility is F-actin assembly and one can 13 measure this with a dye MBD felacydin. One of the 14 other questions we asked was whether there was 15 something that was different in the modal apparatus of 16 the cells. So again we looked at the two groups, the 17 controls and patients on day 0 and day 4, and this is 18 a lot of data. The important thing is to look here. 19 This is again an expression of mean channel 20 fluorescence in unstimulated cells or cells that are 21 treated with 10-7 molar FMLP. And what you can see is 22 a decrease in F-actin assembly that is statistically 23 significant and we think probably practically 24 significant too. 25 So in fact one of the things that we were 79 1 able to find then was a decrease in motility and an 2 associated decrease in F-actin. One of the other 3 things we wanted to look at was channels in cytosolic 4 calcium. This ends up being very important in 5 motility and ingestion and in most receptor mediated 6 functions of the cell. In these experiments what we 7 did was to load the cells with a fluorescent 8 indicator, binding calcium Endol-1, and we treated the 9 cells with FMLP. This is a plot of the results for 10 one patient. The neutrophils that were collected on 11 day 0 were assayed and the calcium flux was followed 12 in response to FMLP and the same patient on day 4. 13 What you can see on day 4 is a marked increase in the 14 total flux in the cells. The onset and initial rate 15 are the same for day 0 and day 4, but this increase is 16 two to three-fold. If we look at all of the patients 17 now, and this slide summarizes results for all 18 patients, looking at cytosolic calcium, and what I 19 have here is the response of cytosolic calcium to FMLP 20 10-7 molar and also platelet activating factor. There 21 are two columns for each stimulus. The first column 22 reflects the baseline level of calcium and the second 23 the peak calcium flux. This is in micromolar 24 concentration. What you can see -- the important 25 thing -- again, a lot of numbers -- the important 80 1 thing is to look at the second and fourth column at 2 the bottom. This is the results for day 4. One can 3 see a marked increase in all of the patients. This is 4 a two- to three-fold increase of cytosolic calcium. 5 This enhancement in calcium flux in the 6 cells in response to the specific stimuli, we are 7 still not sure exactly what that means and the 8 importance and relevance of this to the chemotaxis and 9 perhaps other activities is not clear at this point. 10 So the next group or classification of 11 studies that we did is spectracidal activity. This is 12 a standard bacteriocidal assay. In these studies, 13 this summarizes studies for the normal controls and 14 patients on day 0 which are included in the dots and 15 in the squares are patients on day 4. In this assay, 16 there is a 1 to 1 ratio of bacteria to cells, and the 17 bacteria is Staph aureus. This is done in the 18 presence of 10 percent normal pooled serum. What you 19 can see is killing that is equivalent to control in 20 the patient on day 4. Perhaps there is a suggestion 21 of attenuation of killing, but this is not 22 statistically significant. 23 We looked at this in a little bit more 24 detail because neutrophils when they first get to an 25 area of inflammation have probably a lot more bacteria 81 1 to phagocytose and kill than just one each. In 2 addition, one can see in a variety of patient 3 disorders a mild killing defect. So we wanted to 4 stress the system and we did a killing assay with a 5 ratio of 10 bacteria to 1 neutrophil. This is 6 summarized on the next slide. What you can see is 7 that at 30 minutes and at 90 minutes, there is a 8 statistically different and I think practically 9 different percent killing in this assay. So perhaps 10 there is a mild defect, and I would underscore that -- 11 I would say a mild defect in killing and it may or may 12 not be significant. 13 We looked in detail at the respiratory 14 burst and the oxidase activity. And I am going to go 15 through in the next two slides looking at the 16 respiratory burst measured as cytochrome C reduction 17 in response to a variety of different agonists. The 18 first one we used was FMLP, a chemotractant which at 19 a little bit higher dose than used in chemotaxis will 20 activate the oxidase. On this plot you see the 21 control and the patients on day 0 and day 4. What you 22 can see here is an increase in the FMLP response in 23 patients on day 4 compared to controls. This seems to 24 parallel what was found in vitro and also other 25 studies that have been done, studies that Dr. Dale 82 1 presented earlier this morning. 2 When we look at a variety of other 3 agonists to try to define or get a complete picture of 4 the oxidase, we see some divergence in the results. 5 These are results for host cell superoxide activity 6 production with 4 ball ester. As we can see here, 7 this is day 0 control and patient and this is day 4 8 control and patient. We see a marked depression in 9 the PMA response. If you look at another stimulation 10 sequence, and in this sequence we try to look at 11 priming of the cells and we essentially prime the 12 cells or incubate the cells with platelet activating 13 factor for three minutes and then come back and look 14 at the response to FMLP. This usually gives us kind 15 of the maximum respiratory burst. This is even a 16 stronger set of agonists than PMA or most other 17 agonists that you can use to look at the respiratory 18 burst. What you can see is again a decrease or an 19 attenuation of the production of superoxide in the 20 intact cells with this stimulus. 21 Now if you look at a third stimulus, which 22 is opsonized zymosan, which is essentially a 23 phagocytic stimulus which is coated with complement 24 and so ingestion is most likely by complement and C3BI 25 receptors, what one sees is no difference in the 83 1 respiratory burst. This is probably a more 2 physiologic stimulus. So you see a divergent set of 3 reactivity that shadow or characterize the respiratory 4 burst. Some are increased, some are low, and some are 5 normal. 6 We looked very carefully at oxidase 7 components and this would be -- I am not going to show 8 you all the data, but this would be of interest to 9 individuals who are looking at or are interested in 10 the oxidase itself. What we found on these cells when 11 we looked at subcellular fractions, we found increased 12 amounts of cytochrome B558 in the plasma membrane and 13 normal contents of the cytosolic oxidase components, 14 the P47-phox, the P67-phox and the P40-phox. So the 15 oxidase itself seems to be intact, but we seem to have 16 to certain kinds of stimuli a decreased response, 17 which would suggest that it is other systems perhaps 18 than the structural oxidase proteins that are 19 affected. 20 One of the other things that we did was to 21 look at granular marker proteins in order just to 22 define the status of granules in these cells. As Dr. 23 Dale suggested earlier, these cells look different. 24 If you do EMs, you have a sense looking at the EMs 25 that the granules are not quite the same and the 84 1 granule compartments are not the same. So we looked 2 at this and looked at alkaline phosphatase, which is 3 increase, which everybody would expect and which has 4 been really well documented as an effect of G-CSF on 5 neutrophils. Myeloperoxidase seems to be normal. The 6 specific granule marker, lactoferrin, seems to be 7 decreased. The question that is raised by this data 8 as to whether there is a defect in the specific 9 granules themselves and their production, when we 10 looked at cytochrome B, the content of cytochrome B 11 was actually normal to increased. So this needs to be 12 looked at a little bit more carefully. I am not sure 13 at this point that we can say that there is a decrease 14 in specific granules, but there appears to be a 15 decrease in specific granule content in some proteins. 16 And that may certainly have some functional impact on 17 the cell. 18 One interesting side note, and that is 19 that we have saved now cell lysates and subcellular 20 fragments of plasma in the membrane and granules, is 21 to look at also to save RNA, and the question is what 22 does RNA do to some of the genes and some of the 23 proteins in the cell. The interesting observation 24 that we made is that when one looks at the cells that 25 are collected on day 4, there is a much larger amount 85 1 of RNA which can be extracted. This is roughly the 2 RNA content for 108 cells. This is control day 0 and 3 patient day 0 and control day 4 and patient day 4. 4 You see almost a two-fold increase in RNA. In fact, 5 when you store cells for 24 hours, you double the RNA 6 again in looking at how much you can extract from the 7 cells. This is very interesting. We are not sure 8 what the significance of that is, but I think it is 9 going to be an important clue to some of the defects 10 that we are finding. 11 This is just kind of a mental break. I 12 wanted to talk a little bit about apoptosis, because 13 this is another area that we evaluated in this study 14 with patients. These tests are done by looking 15 morphologically. We take cells at the sampling times 16 and we isolate them and we put them into culture with 17 RPMI and fetal calf serum. Then at different times 18 after that, up to 48 hours, we take little samples out 19 and we evaluate them for the extent of apoptosis. Of 20 course these are very labor intensive studies and 21 people go crazy as they stay up in the lab and try to 22 get these things done. So our argument was, was this 23 a-pop-tosis or apop-tosis. I mean, you get pretty 24 crazy when you are doing these kinds of experiments. 25 Dr. Levy, who is from Dublin, Ireland, was a fellow in 86 1 my lab who was involved with most of the studies that 2 you are seeing here. He had a different feeling and 3 he would come out with his thick Irish accent and say, 4 no, it is not any of those, it is really O'potosis. 5 Let me show you an example. This was a 6 technique in which you use a double stain, preputium 7 iodine and acridine orange. These are both non- 8 apoptotic cells. You don't see any of the nuclear 9 changes related to apoptosis. This is a dead cell and 10 this is a live cell. This is a live apoptotic 11 neutrophil. So on the basis of these morphologic 12 features, we would characterize during the culture the 13 percentage of cells that were apoptotic, and we could 14 generate a graph, if you will. We called it the LT50. 15 It probably should be the AT, the apoptosis time 50, 16 but the time to 50 percent apoptosis. That is what is 17 graphed here. What you can see is the dark bars are 18 the patients at day 0 and day 4 with the controls. 19 What you can see is that before G-CSF administration 20 to the patients or in the control group, we see a time 21 to 50 percent apoptosis, live apoptosis, is somewhere 22 around 17 hours. If you look at patients on day 4 23 after the fifth dose of G-CSF, what you see is the 24 time to 50 percent apoptosis is prolonged to about 34 25 hours, it is doubled. 87 1 In other studies to the in vitro system, 2 we added G-CSF to a dose of 25 nanograms per ml. What 3 we saw is that these cells could respond further by 4 prolonging their apoptosis. You will notice that this 5 curve is now a different curve starting out at 30 6 hours and what we see is a prolongation of the control 7 and patient day 0 and control day 4 cells to somewhere 8 around 42 to 45 hours. And you see the patient day 4, 9 which had received 5 doses of G-CSF, is prolonged even 10 further. So the cells have, at least in vivo 11 circulating that we can take out and culture and look 12 at apoptosis, have something going on that prolongs 13 the process to apoptosis and that these cells can be 14 further manipulated by adding G-CSF in vitro to 15 prolong that time to apototic death even further. So 16 this is actually the good news about G-CSF. This is 17 not doctor-assisted suicide of the cells, but in fact 18 the reverse, that is, we can help prolong the lifespan 19 of the cells. That has, I think, some implications 20 perhaps for storage. 21 So in summary, administration of G-CSF for 22 five consecutive days is well tolerated with, I think, 23 minor problems and adverse events. Although I would 24 echo Tom's comments about how are you going to get 25 normal donors to accept any kind of discomfort which 88 1 may be significant. And certainly we can see an 2 enhanced number of mature neutrophils. I haven't 3 shown you this, but I will present a summary slide in 4 a minute where you certainly get robust huge numbers 5 of granulocytes that you can collect by apheresis. 6 Interestingly, the time of apoptosis is delayed and 7 that the neutrophils themselves that are mobilized 8 under the influence of G-CSF continuous administration 9 for five days present a divergent pattern of 10 functional characteristics but overall the function is 11 not markedly enhanced. And we would further summarize 12 that we think the effect of prolonged administration 13 of G-CSF and the advantage to patients themselves may 14 be related more to numbers and the effect on apoptosis 15 and perhaps survival and not so much the enhanced 16 functional characteristics. 17 Let me just summarize some of the results 18 we have in storage. We can perhaps talk about this 19 this afternoon in the discussion section and the 20 poster session. But as I said, there were 8 products 21 that we stored with this in the presence and absence 22 of G-CSF. What we noticed is that looking at 23 superoxide and intact neutrophils to all the stimuli 24 that I mentioned before didn't seem to change much 25 during the first 24 hours of storage and then it 89 1 deteriorated gradually by 48 hours. It was about half 2 of what you saw that I had presented with the time 0 3 studies. Most of the collected neutrophils remained 4 viable, that is, greater than 98 percent viability of 5 the neutrophils in storage, and were not apototic at 6 that time. Their apototic rate was probably no 7 different than what we found for the neutrophils that 8 were collected right at the time or just before we had 9 done the granulocyte collections, and that is the 34 10 hour time. So there seems to be, at least for 24 11 hours and it may be longer, this is something we need 12 to look at a little bit more carefully, a viability 13 and postponement of apoptosis. Chemotaxis, although 14 it was deficient -- as I showed you, the day 4 data 15 was deficient -- in storage this didn't get any worse, 16 at least for the first 24 hours of storage. And when 17 we looked at all these things and the addition of 25 18 nanograms per ml of G-CSF, we really didn't change 19 these characteristics of the neutrophils. That is 20 probably most likely that we didn't add enough and 21 that may need to be looked at a little bit more 22 carefully. But there may not be really any additional 23 effect that these cells have. 24 So what are considerations for future 25 studies? First of all, I think more work needs to be 90 1 done in defining standardized schemes for mobilization 2 and collection. There may be reasons -- certainly one 3 can collect granulocytes after a dose of G-CSF, but 4 there may be reasons to look at multiple doses of G- 5 CSF. There may be reasons to perhaps think about 6 other cytokines as well. So this needs to be defined 7 a little bit more clearly. It would be worthwhile to 8 develop the optimal storage conditions for neutrophils 9 to try to support normal function and survival for a 10 little bit longer. Right now what we are doing, as 11 Dr. Price has suggested -- and we have a difficult 12 time in the blood bank because we get our donors in 13 several days early and use that sample to allow the 14 physicians to get the blood, that is, we can release 15 it with that. We also have to do the processing on 16 that sample that we collect. So this is really 17 problematic. Obviously in terms of providing a 18 product that has no storage time, if we could develop 19 some, even for 48 hours, it would be very helpful in 20 doing the collections. We really need to expand and 21 develop techniques for evaluating in vivo function of 22 transfused granulocytes. To extend in vitro studies 23 to evaluate this and to look at what their functions 24 are in the patients is going to be very, very 25 important. And, of course, clinical trials to 91 1 document their efficacy, toxicity, and cost 2 effectiveness are going to be important to complete in 3 order to really revitalize this type of blood 4 component. So with that I will stop. 5 CHAIRPERSON HARVATH: Thanks, Dan. The 6 next speaker is Dr. Susan Leitman. Dr. Leitman is the 7 Chief of the Clinical Services Section of the 8 Department of Transfusion Medicine here at the NIH. 9 She is going to talk to us about her experience with 10 G-CSF mobilized granulocytes. 11 DR. LEITMAN: Thank you, Liana. And thank 12 you for inviting me to speak at today's conference. 13 This is a slide I made to entitle a talk at another 14 conference on this topic, and I found that it applied 15 well to the issues that we are bringing before the FDA 16 perhaps in consideration of licensure of this product. 17 And with growth factor mobilized granulocytes, is this 18 an exciting or stimulating -- no pun intended -- new 19 component or are we stuck with the same old problems. 20 I would like to remind all of us, as if we 21 needed to be reminded, that despite nearly three 22 decades of clinical experience and dozens of 23 publications of observational or controlled studies, 24 the FDA does not recognize granulocyte concentrates as 25 an approved blood component. And from the very nice 92 1 review we heard by Dr. Dale this morning, there are 2 very good reasons for that non-recognition. This is 3 my version of Ron Strauss's review from the Blood 1993 4 article reviewed by Dr. Dale this morning. I want to 5 point out that the reason we can't find efficacy 6 across all of the seven prospective, some randomized 7 and some non-randomized studies, is in large part due 8 to the choice of the subjects for the study, and thus 9 in studies designed to determine clinical efficacy you 10 have to choose patients in whom the mortality is 11 estimated to be substantially above 60 percent. In 12 those studies in which the survival was 60 percent or 13 greater, no efficacy could be demonstrated. It was 14 only when the mortality was quite high that efficacy 15 could be demonstrated here, with mortality of percent 16 surviving of 26, 15, and 36 percent. So in designing 17 prospective trials, we have to choose the right 18 population to study and, as Dr. Dale already stated, 19 have sufficient numbers to power the study or 20 analysis. 21 I will talk just for a moment about dose 22 because that has been covered very well so far this 23 morning. To remind you that granulocyte apheresis 24 yields without any donor preparation are in the range 25 of .3 to .5 times 1010 cells. With the addition of 93 1 hydroxyethyl starch, this is high molecular weight 2 starch, that increases or that doubles to .5 to .9 3 times 1010. When one uses some combination and some 4 regimen of steroid administration plus starch, that 5 again doubles to 1 to 2 times 1010. I am going to 6 diverge for a moment and say that granulocyte 7 apheresis took a giant step backwards in the mid to 8 late 1980's with the introduction of a new form of 9 starch, pentastarch. Pentastarch is a less highly 10 substituted amylopectin backbone, the same backbone 11 that is in hetastarch, with an average molecular 12 weight of 264 rather than 480,000, a significant 13 reduction in the number of hydroxyethyl groups per 14 glucose residue. What made pentastarch very 15 attractive to blood bankers was the safety for the 16 donor in that the 24-hour urinary excretion is much 17 higher with pentastarch than hetastarch and the 18 overall survival in blood is only 96 hours as opposed 19 to 17 to 26 weeks with some residual hetastarch 20 remaining in a donor's body for substantial periods of 21 time. So the entire blood banking field sort of moved 22 to pentastarch rather than hetastarch in the late 23 1980's. But if you look at the studies of 24 pentastarch, the efficacy of granulocyte apheresis or 25 the efficiency of granulocyte apheresis and the yields 94 1 were never prospectively compared to hetastarch. A 2 fellow in our lab, Dr. John Lee, who is now with the 3 FDA, did a very nice set of studies in the early 4 1990's where he looked at the comparison of 5 pentastarch and hetastarch. Just to remind you, the 6 granulocyte collection efficiency with apheresis 7 devices, the GCE, varies directly in proportion to the 8 donor's erythrocyte sedimentation rate. The more 9 quickly the red cells sediment, the better the 10 separation in the granulocyte layer and the more 11 efficiently the machine can collect them. So with 12 increasing donor sedimentation rates, there is an 13 increasing granulocyte collection efficiency. What 14 was not known at that time or not clearly defined was 15 that hetastarch quadruples the donor's sedimentation 16 rate in vivo and pentastarch increases it by one and 17 a half to two-fold. 18 When John Lee prospectively compared in 72 19 apheresis donors a granulocyte apheresis procedure 20 using pentastarch and three months later a granulocyte 21 apheresis procedure using hetastarch and looked at the 22 granulocyte collection efficiencies, if they were the 23 same the line of identity would be here shown by this 24 dash blue line, and they were not the same. The 25 granulocyte collection efficiency with hetastarch was 95 1 substantially and significantly better than with 2 pentastarch in all but three donors. If you look at 3 the yield, not only the GCE, you see again here is the 4 line of identity and with hetastarch the yields were 5 always, except for three donors, substantially greater 6 than with pentastarch. I will summarize this 7 numerically on the next slide. So there is a 60 to 70 8 percent increase in granulocyte yields times 1010. 9 This is before G-CSF. These are all non-G-CSF 10 mobilized donors. From 1.4 to 2.3 times 1010 in these 11 72 paired collections. Collection efficiency 12 increases from 33 to 58 percent. With the publication 13 of this study, I believe that most centers have 14 returned to hetastarch. 15 At about the same time that Dr. Dale was 16 giving his five college students various doses of G- 17 CSF, we were giving 20 healthy apheresis donors under 18 protocol three varying mobilization regimens. The 19 donors underwent three leukopheresis procedures each 20 separated by at least four weeks. The three 21 preparative regimens were nearly identical to what you 22 have heard this morning -- dexamethasone 8 mg orally 23 was given 12 hours prior to donation. We would tell 24 our donors to take the dexamethasone about one hour 25 after dinner, which is about 8 p.m., and that is 12 96 1 hours before they come to our apheresis center at 7:30 2 to 8:30 in the morning. We give them the next dose of 3 dexamethasone to take home with them at the time of 4 apheresis so that they don't have to come to a 5 pharmacy to get the tablets. G-CSF we administered at 6 a dose, a per kilogram dose of 5 mcg per kilogram, and 7 we used Amgen's form of filgrastim, and we gave it 8 subcutaneously between 16 to 24 hours prior to 9 donation. And then in the third arm, they received 10 both types of preparative drugs. 11 G-CSF comes commercially or is available 12 commercially in two size vials, a 300 mcg vial and a 13 600 mcg vial, 300 per ml. So it is one ml in the 14 first vial and 2 ml in the second vial, which is 15 exactly why in the Seattle study they used either 300 16 or 600. Apparently the Government on GSA schedule 17 gets a vial that contains 480 mcg, which is partly why 18 we use 600 mcg per kilo, because that does not exceed 19 one 480 mcg vial. So this is going to be a little bit 20 different in comparing our study to the Seattle 21 studies, but probably not substantially so. 22 All our granulocyte procedures are 23 performed using a Baxter CS-3000 plus apheresis 24 device, continuous flow, two access sites needed. 25 These are the parameters of leukopheresis. The two 97 1 instrument chambers who blood flow rate is between 50 2 and 60 ml per minute. Our endpoint volume 3 traditionally in the last 15 years of collecting 4 granulocytes has been 7 liters. So for this study, we 5 did not change that. You will note the Seattle 6 protocol was 10 liters and other protocols vary from 7 7 to 12 liters processed. The anticoagulant is sodium 8 citrate. The sedimenting agent is 6 percent 9 hetastarch. 10 If you read the operating manual for 11 performing this procedure on the CS-3000, it tells one 12 to set the interface offset at 15. No one does that. 13 In studies done 10 years ago along with the engineer 14 who developed this device, Mr. Herb Cullis, it was 15 found that an interface offset setting of 33 yields 16 optimal efficiency of the procedure, and so most of us 17 have been using an IO of 33 for the past decade, not 18 what is in the operating manual. 19 These are the results of the three 20 different regimens, dexa, G-CSF, and D+G. This is the 21 peripheral blood polymorphonuclear count immediately 22 prior to apheresis, and you can see that the addition 23 of G-CSF or D+G increases the white cell 24 polymorphonuclear cell count in the donor by 3.5 to 25 4.5 fold, very similar to what you saw this morning, 98 1 from 6,000 to 21,000 to 29,000. Similarly, the 2 product content increases 2.5 to 3.5 fold from 2.5 3 times 1010 with our traditional dexamethasone alone 4 arm to 5 to 7.2 times 1010 with a combination of both. 5 All of these comparisons are statistically significant 6 at the .05 level for every comparison within groups. 7 Addition of dexamethasone to G-CSF alone 8 resulted in a 43 percent increase in the granulocyte 9 yield in the product. We also looked at granulocyte 10 collection efficiency and our usual efficiencies, as 11 you have seen in the previous slide, are in the range 12 of the low 60 percent. Somewhat to our surprise, we 13 found that the efficiency dropped by 10 percent when 14 we added G to this regimen. When we spoke with the 15 engineer who had designed this device, he told us that 16 the machine was designed for donor counts of 10,000, 17 not for total white counts of 30,000 to 35,000, and 18 that was necessary probably was a complicated change 19 to the interval between spillovers and the duration of 20 spillovers, for those of you used to this machine, and 21 that we would eventually do that. But as a start, we 22 simply increased the interface offset setting to 45, 23 and I don't have those numbers right here because they 24 haven't changed. Simply increasing the depth of 25 penetration into the buffy coat layer as this machine 99 1 collects the cells did not significantly increase the 2 GCE. So I think further work needs to be done with 3 this particular device in maximizing collection 4 efficiencies. 5 One day, we just happened to have three of 6 these products quite by accident in the processing 7 area of the transfusion medicine department, and they 8 looked so distinctly different that we took the 9 opportunity to take a picture of them. This is the 10 traditional dexamethasone stimulated product. It 11 looks redder because the buffy coat layer is less 12 thick. This is G-CSF alone with a thicker buffy coat, 13 and this is the combination of G-CSF plus 14 dexamethasone. This is just sedimentation on the 15 counter top over the course of the six to eight hours 16 between the end of collection and the time of 17 transfusion. You can see the buffy coat layer 18 sediment out. In our institution, we start apheresis 19 in all such donors on this protocol at 8:00 to 8:30 in 20 the morning. The product comes off the machine by 21 10:30 in the morning, and we do same day transfusion 22 transmitted infectious disease testing starting at 23 9:00 in the TTV laboratory. So we do same day testing 24 although we are one of the very, very few institutions 25 that I think can continue to do that today. So our 100 1 products are transfused at 8:00 p.m. that day. They 2 spend about 10 hours on the shelf and we saw the 3 sedimentation within those 10 hours. 4 What is the effect on the product of the 5 three different regimens other than in the 6 polymorphonuclear leukocyte count? We always process 7 the same volume. This is the volume without the 8 anticoagulant added. So this is actually actual true 9 blood volume processed. The machine is set to 7 10 liters but about 500 ml of that is anticoagulant. The 11 product volume is set by the operator. We 12 traditionally set it to be about 240 ml. The platelet 13 content was identical across all three collection 14 regimens and the red cell content was identical across 15 all three collection regimens. Just to remind you, 16 the mean red cell content is about 30 ml of packed 17 cells, so a cross match is always necessary between 18 donor and recipient. 19 This is a summary that I made a while ago 20 and it is not updated to reflect the newer Seattle 21 data, which we heard this morning and which came out 22 in August. But it looks at the four published studies 23 at that time of granulocyte apheresis yields following 24 G-CSF containing regimens. The dose varied from 200 25 mcg per day to 5 mcg per kilo. Steroid use was 101 1 variable. Starch use was variable, pentastarch or 2 hetastarch. Volume processed was variable. And the 3 yields at that time were in the range of 4 times 1010 4 or 40 times 109, except for our experience in which it 5 was double and the current Seattle experience in which 6 it is right about this number -- 82 is what we heard 7 this morning. I think the difference here was in the 8 steroid use. In these initial studies, we were always 9 using steroids in our clinical study in products that 10 were getting administered to patients and we were 11 always using hetastarch, whereas the type of starch 12 used and steroid use varied. 13 I show this slide because it wasn't quite 14 presented in this way by the earlier speakers today. 15 This is the adverse reactions to mobilization regimens 16 in donors who have undergone all there preparative 17 regimens. With dexamethasone alone, we have an n of 18 38. And 44 percent of donors have symptoms and they 19 are almost all universally related to hyperactivity -- 20 insomnia, feeling wired, feeling restless, waking up 21 at night several times. And a small proportion had 22 flushing and this was commonly delayed. So they would 23 call us from work later that day to say their 24 colleagues told them their face and their ears were 25 bright red. It could also even happen the next day 102 1 after. So 56 percent of such donors on dexa alone did 2 not have symptoms. With G-CSF alone, the same number, 3 n equals 38. We had the usual 39 percent with bone 4 pain, 26, a quarter, with headache, 24 percent with 5 that sense of wiredness, and also 5 percent with 6 flushing. Only 32 percent did not have any symptoms 7 at all. With the combination, the instance of bone 8 pain is a little bit higher. Headache is about the 9 same. Insomnia is the same with dexamethasone alone. 10 Some nausea. I forgot to mention the fatigue. 10 11 percent have fatigue whenever you give G-CSF and 12 flushing, so that only 28 percent did not have 13 symptoms. 14 We have enrolled 120 donors on this study, 15 similar to what Dr. Price described this morning. We 16 took them from our pedigreed platelet pheresis donor 17 population. And as he told you this morning, when you 18 approach these intensely altruistic individuals with 19 yet another regimen or a new product that may increase 20 the potential for survival for critically ill patients 21 with cancer and other serious illnesses, they are most 22 eager to cooperate and be, if you will, on the cutting 23 edge of transfusion medicine practice. Of all the 24 reactions I described on the last slide, 15 percent 25 were judged by the nurses or myself in asking the 103 1 donors these questions to be severe and interacting 2 with everyday activities of the donor. 8 of 76 of our 3 first donors or 10 percent -- and this continues, it 4 is now 10 out of 100 -- have requested discontinuation 5 of G-CSF mobilized collections. The most common 6 reason was they didn't like feeling as if they had the 7 flu once a month -- or as one donor puts it, I am 8 tired of aging from age 40 to age 80 overnight once a 9 month. The other common reason was the inconvenience 10 of coming to the blood bank twice, once the day before 11 the injection and once to donate. Donor reactions 12 were stereotypic. Mild reactions tended to become 13 milder with further donations. An initial 10 percent 14 of donors had no symptoms on subsequent G+D 15 mobilization although they had had symptoms on their 16 first occasion. 17 Now I would like to talk for the remaining 18 time on patient outcomes. The results in the first 19 three patients that we transfused with G-CSF mobilized 20 products were so dramatic that as my colleague Harvey 21 Alta says, you should make slides of things that work 22 before you find that they don't work. Make your 23 slides quickly. So we made these of the first three 24 patients. Our very first patient was a 55 kilo female 25 with T cell large granulolymphocytic leukemia who had 104 1 a sigmoid phlegmon due to diverticulitis. She had a 2 bacterial process. She was extremely ill and toxic, 3 persistent rigors, chills, fevers, unresponsive to 4 antibiotics. The second patient was a larger male, 5 130 kilo, day 10 post a T cell depleted marrow 6 allograft for myeloma with a systemic Aspergillus 7 flavum infection. The third patient was a similar day 8 7 post T cell depleted marrow allograft for CML with 9 a systemic fusarium infection. 10 Let me go back before I do that. What was 11 dramatic in all these patients was the almost 12 immediate response to the administration of these 13 cells. This patient became afebrile for the first 14 time the day of the granulocyte administration and 15 remained afebrile until she actually recovered her own 16 white count. Both of these patients were showering 17 skin with new systemic fungal lesions on a daily or 18 more often than daily basis. One could watch the new 19 lesions develop. New lesions stopped developing with 20 the first granulocyte transfusion and old lesions 21 resolved with subsequent granulocyte transfusions. 22 This is the saw-tooth pattern you see when you look at 23 the increment in ANC, absolute neutrophil count, with 24 each subsequent granulocyte transfusion. These are 25 the first two patients, who as I said had very 105 1 dramatic responses. The first got a total of four 2 granulocyte transfusions. Her absolute neutrophil 3 count was 0, increased to 2, and as has been said by 4 speakers this morning, remained elevated for the next 5 8 hours, which we had never seen before. So the next 6 day she gets another increment of 2,000 and goes up to 7 4,000. By the next day, she gets another increment 8 and goes up to 6,000. We do not collect on the 9 weekends unless it is a very serious patient problem. 10 So we did not collect on this day and she promptly 11 dropped her neutrophil count at 30 hours. We gave her 12 one more transfusion and then you can see her own 13 cells recover. 14 This is the recipients of the T cell 15 depleted marrow allograft. They increment to 1,000 16 and stays there for 8 hours and increments to 3,500. 17 Then the weekend occurs and we wanted to see if he had 18 recovered his own counts. He had not. The same saw- 19 tooth pattern, the weekend, and then his own counts 20 finally returned. 21 We looked at the one-hour post transfusion 22 increment in the ANC level as a function of the number 23 of granulocytes transfused for these first three 24 patients and there was a direct correlation, which was 25 highly significant, with the increasing counts as a 106 1 function of the number of granulocytes transfused. 2 This graph is perhaps less steep because there were 3 only four points. 4 How did they do overall, though? They 5 clearly responded in terms of their infection in the 6 short-term. The first patient is in complete 7 remission and back to her everyday activities now. 8 Her diverticular phlegmon was removed surgically after 9 her own white count recovered. The second patient 10 died two months after the course of granulocyte 11 transfusions due to multi-organ system failure as a 12 complication of bone marrow transplant. Aspergillus 13 was present at autopsy but was not thought to be 14 contributory to his death. The third patient 15 stabilized, eventually was discharged from the 16 hospital and is still alive now with chronic graft 17 versus host disease. Serial one-month skin biopsies 18 were obtained and at day 30 and 60, he still had 19 fusarium in his skin, but at day 100, the fusarium was 20 gone. So it takes a long time to clear invasive 21 filamentous fungus infection. 22 Here is our summary of the 14 total 23 neutropenic recipients of G-CSF mobilized granulocyte 24 transfusions to date at the NIH Clinical Center. You 25 heard about the 19 patients at the Fred Hutch, and the 107 1 data are very, very similar. These 14 individuals 2 received a total of 135 granulocytes that had been 3 mobilized with G-CSF plus dexamethasone. In some 4 cases I have the total number of G-CSF mobilized 5 granulocyte transfusions in parenthesis because the 6 total reflects both dexa and G-CSF mobilized products. 7 We have given this product to 4 patients with aplastic 8 anemia. Three of them had invasive fungus infections, 9 Aspergillus or fusarium, and one had a strep 10 pneumonia. All four patients initially improved. 11 These two stabilized to an impressive degree. The 12 fusarium resolved completely. That was the previous 13 patient. And the strep infection or the pneumonia 14 promptly improved. However, in the first two 15 patients, eventually since they didn't recover their 16 own counts, they had ANCs of close to zero and the 17 disease progressed. In our experience here, we give 18 immunomodulatory therapy with ATG and cyclosporin or 19 cyclophosphamide. It takes 6 weeks to see an increase 20 in the ANC with about 65 to 75 percent of all patients 21 responding, and 6 weeks of an ANC of zero, even in the 22 presence of granulocyte transfusions, is very 23 problematic. So only one was discharged from the 24 hospital. 25 In the allo-peripheral blood stem cell 108 1 transplant patients, there were four, three had 2 disseminated fungal infections and one had an RSV 3 pneumonia, which we didn't know at the time. We 4 thought he had a fungal infection or a fungal 5 pneumonia. These two patients -- this patient was 6 started when he was nearly in extremis. The 7 Aspergillus was progressive and he rapidly died. This 8 patient continued to do poorly despite granulocyte 9 transfusions until we realized he had RSV as his main 10 process and not fungus. These two patients improved. 11 I discussed them on an earlier slide. One died of 12 multi-organ system failure unrelated to fusarium. One 13 is doing well. This patient with a non-Hodgkin's 14 lymphoma and two patients with NHL had a vancomycin 15 resistent enterococcus and multi-organ system failure 16 and was an extremis when we started granulocyte 17 transfusions, had progressive disease and died. 18 Another young girl with pulmonary Aspergillus in the 19 setting of HIV infection and non-Hodgkin's lymphoma 20 had a very nice response to 10 granulocyte 21 transfusions. And we have two patients with LGL 22 lymphoproliferative disorder, both of whose 23 infections, one bacterial and one candida, resolved. 24 And a breast cancer patient with a pseudomonas ulcer 25 of a myocutaneous reconstruction flap within two days 109 1 worth of transfusions, her own count had returned and 2 she improved. I am not sure we can say much about 3 this one patient in blast crisis at CML who had a 4 presumed fungal pneumonia who also improved, although 5 he eventually died of other complications. 6 So there were 9 of 14 patients with 7 invasive fungal infections. Overall, 11 of 14 8 improved and it was not surprising that these two 9 patients in extremis and the one patient with RSV 10 pneumonia did not improve. But the overall hospital 11 discharge was slightly less than half as you just saw 12 from the Seattle experience, and that has been the MD 13 Anderson experience as well. 14 These are essentially 14 anecdotes. And 15 what we have heard this morning for what we really 16 need is a randomized prospective trial so that they 17 become more than just anecdotes. 18 We have also in this institution treated 19 a number of patients with chronic granulomatous 20 disease of childhood, an inherited disorder where 21 granulocytes cannot make phagocyte oxidase and 22 membrane bound oxidase, and they can't handle various 23 types of fungal and bacterial infections and are 24 subject to recurrent life-threatening infections with 25 organisms like Serratia, Pseudomonas, Candida, 110 1 Aspergillus, and nocardia. We have treated a total of 2 10 such patients with 220 transfusions in the past two 3 years. Again, this is a set of 10 anecdotes. And in 4 the absence of a prospective trial, there is not that 5 much that can be said except that 9 out of the 10 had 6 resolution of infection and 9 out of the 10 were 7 discharged from the hospital. And whenever you see 8 such an excellent response, it makes you think that 9 perhaps the granulocytes were involved in this 10 excellent response. And what makes me have some 11 confidence in saying that was that in one patient here 12 and one patient here, they only received two 13 granulocyte transfusions before an anamnestic response 14 in their HLA alloantibodies became clear and they had 15 very significant pulmonary transfusion reactions. We 16 stopped the transfusions after two and three 17 transfusions, but they still resolved their fungal 18 pneumonias and their bacterial pneumonias with 19 excellent antimicrobial therapy. So again, in the 20 absence of a trial, it is hard to say exactly what the 21 role of granulocytes, even G-CSF mobilized 22 granulocytes, is in patients with CGD, although there 23 is some controversy among the clinical care staff, Dr. 24 Malik and colleagues here at the Clinical Center, 25 about the role of granulocytes. 111 1 This is all 14 of our neutropenic 2 patients. These are the white cell increments. This 3 is the incrementing count, not the absolute count in 4 that patient -- but the increment following 5 transfusions of G-CSF mobilized granulocytes. On the 6 x axis are hours after transfusion, and please notice 7 this is not a linear scale. There is pre-transfusion, 8 1 hour, 4 hours, 8 hours, 24 hours, and 30 hours. And 9 there aren't points at each of these time intervals. 10 There is always a point at 1 hour, 8 hours, and 24, 11 but not always at 4. The orange line is 10 patients 12 who did not have either HLA alloimmunization or 13 splenomegaly. The 1 hour post increment was 1,900, 14 very close to the 2,600 that you heard this morning 15 from the Seattle group. And 4 hours later in most 16 patients, that count was slightly higher. It was 17 2,000. Suggesting that one-hour post-transfusion, 18 there may be some sequestration in organs such as the 19 spleen and the peak increment is not seen until 4 20 hours. There is still an increment at 8 hours, a 21 substantial increment over the pre-transfusion count. 22 That persists at 24 hours and still persists at 30 23 hours to a very low level. If you try and calculate 24 an in vivo or biologic half life, half of 2,000, you 25 get about 20 hours, which is exactly what Dr. Dale 112 1 reported from his study of radio-labeled cells, 2 autologous cells, in study participants. 3 I separated out the three patients with 4 splenomegaly, two with non-Hodgkin's lymphoma and one 5 with CML because they had a markedly different 6 response. The 1 hour post increment was in the 300 7 range, went down to the high 200 range, and then 8 slowly decreased and then went down to 0 at 30 hours. 9 And then there was one patient who we did not know had 10 HLA allosensitization until the first transfusion was 11 given. He also did not respond as did the mean group 12 without HLA allosensitization. But interestingly, he 13 had some increment. It wasn't 0. That is an n of 1, 14 so I can't say too much about the amount of increment 15 you can expect in an allosensitized recipient. He had 16 a multi-specific positive lymphocytotoxicity screen. 17 So this was a very impressive alloimmunization in 18 vitro. This individual was intubated and had 19 monitoring of every possible pulmonary parameter in 20 the ICU, and we looked carefully at whether there was 21 02 desaturation, increased need for positive index 22 pressure, decreased compliance of the lung, increased 23 temperature, chills, et cetera, related to the 24 transfusion of HLA incompatible very large numbers of 25 granulocytes and we did not see it in this n of 1. 113 1 There was one patient who received a total 2 of 11 granulocyte transfusions early in the course of 3 our work with G-CSF mobilized products, and on only 4 five of those occasions were we able to find a donor 5 enrolled in the protocol that could give G-CSF 6 mobilized product. On the other six occasions, she 7 received a dexa mobilized product from donors already 8 participating in our apheresis program that were used 9 to taking dexamethasone. So I could compare in one 10 study subject the response to G-CSF dexa stimulated 11 product versus dexa alone, and there is a marked and 12 significant difference as you would expect. This 13 happened to be the patient that had the highest 14 increment of all of our patients to granulocyte 15 transfusions. This is the increment, not the absolute 16 count. It was 6,500 one hour post-transfusion, rose 17 to 8,400 four hours post-transfusion, and dropped to 18 a high 2,000 level 8 hours later and at 24 hours was 19 still significantly above her baseline of almost zero 20 as opposed to the response to dexamethasone, where it 21 was a little bit above 1,000 at one hour, and you can 22 see the trend down here. 23 Alloimmunization, as has been stated, is 24 a major risk of granulocyte transfusions. In previous 25 publications, the instance of transfusion reactions is 114 1 90 percent in individuals getting granulocytes who 2 have preexisting alloimmunization and 11 percent of 3 those who do not. In very elegant studies done almost 4 20 years ago by Jan Dutcher and Charlie Schiffer where 5 they radio-labeled allogeneic white cells with indium 6 111 and transfused them, they saw increased pulmonary 7 retention of cells and decreased trafficking to sites 8 of infection in individuals who had preexisting HLA 9 alloantibodies. 20 of 20 successfully migrated to 10 sites of infection without HLA alloimmunization versus 11 3 of 14. 12 We looked at the NIH retrospectively at 13 CGD recipients of multiple courses of granulocyte 14 transfusion therapy in the era where we used 15 dexamethasone alone mobilized products. And of 18 16 patients that we looked at, 14 of 18 had developed HLA 17 alloantibodies during the course of these 18 transfusions. So in some populations, and I suspect 19 that would be CGD and also aplastics, the instance of 20 alloimmunization is extremely high, on the order of 80 21 percent. Whereas in bone marrow transplant 22 recipients, it is probably much less because of the 23 state of suppression of their immune system. 24 We looked more carefully at patients with 25 CDG that we transfused. This is the first six that 115 1 received G-CSF mobilized product. One with 2 Aspergillus bacterial pneumonia, diffuse nocardia, an 3 unspecified fungal pneumonia, bacterial pneumonia, and 4 staph hepatic abscesses. Our protocol said we would 5 not give these cells to patients with preexistent HLA 6 alloimmunization, but the clinicians taking care of 7 these patients were so impressed with the response to 8 granulocyte transfusions in patients without HLA 9 alloantibodies that they prevailed upon us to make a 10 deviation to our standard operating procedure. So in 11 three of these patients, there were preexistent HLA 12 alloantibodies and in one there was prior 13 alloimmunization, but it was not initially evident on 14 the screen although it became evident later. 15 Interestingly, not all patients with alloimmunization 16 had pulmonary adverse reactions. This could be 17 pulmonary infiltrates, fever, chills, dyspnea, or 02 18 desaturation. None of the patients without 19 alloantibodies had reactions. Three of the patients 20 with alloantibodies did and one did not. 02 21 desaturation was seen in two of the four with 22 allosensitization and was not seen in the other two. 23 So it wasn't uniform here either. An indium labeled 24 allogeneic white cell trafficking scan was done in the 25 individual with diffuse nocardia. She had 116 1 disseminated skin lesions which should have been easy 2 to see on the scan and all she had was pulmonary 3 retention with no traffic. So in this patient with 4 alloimmunization, again the cells did not go to sites 5 of infection. 6 Very nicely with the help of Dr. Fleischer 7 and Dr. Malik here at the NIH, a flow cytometric study 8 using a DHR dihydrorodamine stain was done. 9 Dihydrorodamine fluoresces inside neutrophils when the 10 cells undergo a respiratory burst, which is detected 11 by the flow cytometer. Such patients with CGD don't 12 have granulocytes that can undergo respiratory bursts 13 and the DHR is zero percent of cells in the wild type 14 state. Following transfusion, the percent of cells 15 that are phagocyte oxidase positive rises to anywhere 16 from 6 to 64 percent, starts at 0, and was always 17 greater than 1 percent in all individuals who did not 18 have HLA alloimmunization. It was less than 1 percent 19 -- substantially less -- and this may be some 20 background activity or background noise, in patients 21 who did have alloimmunization. This high of 54 22 percent was the first transfusion given to this 23 individual before the HLA alloantibodies became 24 evident in this serum. This number actually decreased 25 to less than 1 percent after the alloimmunization 117 1 occurred. 2 I would like to end, as Liana has asked us 3 to, with considerations for the focus of future 4 research studies. I have decided these foci into 5 donor collection methods, components, and recipient 6 considerations. As referred to numerous times both 7 yesterday and today, we need continued long-term 8 follow-up of healthy recipients of G-CSF to see if 9 there are any long-term complications. There do not 10 appear to be with over 1,000 or several thousand 11 individuals receiving not only a single dose but five 12 consecutive daily doses of G-CSF. The other 13 consideration is how often we can do this in a 14 volunteer donor for granulocytes and not for stem 15 cells. FDA allows us to do cytopheresis 24 times a 16 year in a normal donor. Our policy at the NIH 17 transfusion medicine department is to allow donors on 18 this G-CSF mobilized donation protocol to donate no 19 more often than monthly, but that is an arbitrary 20 restriction. Should there be a difference -- should 21 there be a different restriction for healthy donors 22 getting G-CSF mobilized products than for routine 23 cytopheresis donors? 24 Collection methods -- certainly with the 25 CS-3000, enhancements and modifications are needed to 118 1 the apheresis devices to increase the efficiency of 2 collection and those studies are in progress. 3 Components, as we just heard from Dr. Ambruso, it is 4 very difficult if not impossible for most centers to 5 collect, test, and transfuse on the same day or even 6 to collect, test, and transfuse within 24 hours. So 7 evaluation of storage solutions and conditions if 8 ongoing in several institutions is critical to allow 9 this kind of component to be made available at 10 multiple blood centers. 11 It appears that we have defined a 12 component. The component definition is becoming very 13 crisp. What we found in a retrospective analysis was 14 that 75 percent of our components contain greater than 15 5 times 1010 granulocytes and 90 percent contain 16 greater than 4 times 1010 granulocytes in an analysis 17 of about 200 G+ dexa mobilized components. So you can 18 define a minimum number. You can define the apheresis 19 procedures and the donor preparative regimens to yield 20 that product. So the product definition again is not 21 so much a problem. What is a problem is determining 22 efficacy. 23 So studies of recipients -- and we have 24 heard this before from other individuals -- a multi- 25 center, randomized, prospectively controlled study, 119 1 and I think that this will be so expensive and complex 2 that there probably won't be more than one such study 3 which is being organized out of the infectious disease 4 department at the Fred Hutch, or at least there are 5 discussions of it right now, to identify patients most 6 likely to obtain survival benefit. 7 Eligibility criteria, as we have heard, 8 are deep-seated filamentous fungal infections in the 9 hematopoietic transplant setting, perhaps in patients 10 with severe aplastic anemia, and also life-threatening 11 bacterial infections where we assess the patient as 12 having a greater than 70 or 80 or 90 percent chance of 13 mortality with best available antimicrobial therapy, 14 and in a separate population of patients who are not 15 neutropenic, CDG patients with fungal or bacterial 16 infections. 17 In terms of assessment of efficacy, we 18 have designed such a trial for patients with aplastic 19 anemia in this institution, which is a referral center 20 for SAA. Our statistician reviewing our trial told us 21 that our primary outcome should not be resolution of 22 infection but should be survival at three months when 23 she looked at the data we had on patient outcomes in 24 the first 14 patients treated, and that is different 25 than what we heard this morning from Seattle where 120 1 resolution of infection was a primary outcome. You 2 don't judge platelet transfusions or red cell 3 transfusions by whether the patient leaves the 4 hospital. So this would be sort of a new paradigm. 5 And I am not sure which one of these is best, and I 6 think the statisticians, whether it is hospital 7 discharge or long term survival at three and six 8 months or resolution of infection -- the statisticians 9 should have more discussion on what is the appropriate 10 primary endpoint. The other endpoints can be 11 evaluated in a logistic regression. But you have to 12 define the primary endpoint more clearly. 13 I would like to stop there with the 14 exception of my last and of course the most important 15 slide, my acknowledgements to Mr. Jaime Oblitas, 16 Virginia Morgan, and Sandy Bangham and the outstanding 17 staff of the NIH Apheresis Center for allowing these 18 studies to take place. 19 CHAIRPERSON HARVATH: Would the speakers 20 from this morning's two sessions please assemble and 21 we will give the audience time to put questions 22 together. 23 DR. ADKINS: Adkins from St. Louis. As a 24 patient, if I had a less than 10 or 20 percent chance 25 of surviving from a progressive infection while 121 1 neutropenic, I guess I would have a problem being 2 asked to be randomized to receive or not receive a 3 granulocyte transfusion in a study such as this. So 4 I guess I would ask the response of the speakers how 5 they would feel as a physician who is trying to 6 counsel patients for a proposed randomized trial of 7 therapeutic granulocyte transfusions. If they feel 8 that is appropriate or how they would go about trying 9 to convince people to participate on a trial like 10 this. 11 DR. LEITMAN: That is certainly one of the 12 most difficult questions. Our aplastic anemia trial 13 is not yet running. We have had in the last two 14 months several patients with deep-seated filamentous 15 fungal infections who we knew the mortality would be 16 100 percent with conventional therapy. The attendings 17 that month were asked whether they would have 18 randomized those patients if the trial were active and 19 they said, oh, I don't think so. But then when we 20 look at our outcome data, all such patients have died 21 with granulocyte transfusions, 3 out of 3. Only the 22 bacterial infection patient survived. All three had 23 improvement in their infections over the first two to 24 three weeks. But as one waits that six weeks or eight 25 or twelve to recovery of their own counts, it becomes 122 1 increasingly difficult to support with granulocyte 2 transfusions and then alloimmunization is likely to 3 occur as well. So given the data, though small 4 numbers of patients here and at other sites, I think 5 that that is the way to do the study, to randomize. 6 DR. ADKINS: I think each of you shared 7 your enthusiasm for this area and actually have been 8 very positive about the outcomes you have observed. 9 I am all for randomized trials to try to prove whether 10 or not these things work. I guess I am questioning is 11 this the right setting as a therapeutic maneuver. In 12 my own talk I will discuss kind of a strategy we have 13 taken as kind of a prophylactic maneuver. I think 14 that that is another way to look at efficacy analyses 15 with granulocyte transfusions, and perhaps a more 16 acceptable way from a patient perspective. So, again, 17 if all of us, let's say, were going to join up and do 18 a Phase III trial, I think it is very important that 19 we are all convinced that we can comfortably go to a 20 patient and say these are your two options. I don't 21 know which one of these provides the best benefit and 22 I am very comfortable in recommending you to receive 23 or not receive by flip of the coin a granulocyte 24 transfusion. So I just would encourage us all to 25 think about this as we leave today. 123 1 DR. PRICE: Well, I might comment. I 2 mean, I agree with you that it is going to be a 3 problem. I personally -- I don't think I would have 4 a problem randomizing. I don't know whether this 5 stuff works or not. But the more we talk about this 6 too, the more there are going to be enough people 7 around who are going to have a problem with that 8 randomization, and I think we are hearing about that 9 now and seeing it. One of the possibilities that we 10 have at least batted around a little bit would be 11 whether in a randomized trial there could be an out 12 and some sort of a deal that says if you are in the 13 control arm and you start going down the tubes that 14 there is an escape clause and you can switch over. So 15 that is another possibility. 16 One of the problems that we have faced in 17 terms of considering a prophylactic trial is that if 18 you look at the fraction of the patients that get 19 these kinds of infections -- I mean, maybe it is 10 20 percent of the patients or 5 or 10 percent of the 21 patients that get one of these kind of infections that 22 we end up treating, that we are going to have to be 23 giving 10 or 20 people granulocyte transfusions for 24 the one we would have given therapeutically. And when 25 you start looking at how much that is going to cost 124 1 and from a blood center point of view how many 2 granulocyte donors we are going to have to find every 3 day, that is another real problem that comes into it 4 when you are considering the prophylactic. 5 DR. ADKINS: Maybe I should just stay up 6 here. But I guess, Tom, if this is historically what 7 has happened in the past with granulocyte 8 transfusions, if I am randomized to not receive 9 granulocytes and I have got a fungal infection and it 10 gets worse after that randomization, the tendency is 11 going to be to take care of that patient and to then 12 go on to later give them those granulocytes "off 13 study". And then if your analysis is to determine 14 survival from a statistical standpoint, you are never 15 going to be able to prove that if you allow people to 16 "be rescued" for want of a better word. So this is a 17 very challenging area to try to prove efficacy if you 18 focus this as a therapeutic maneuver. So, again, I 19 think you really have to think about this very 20 carefully based on the historical trials that have 21 been done with randomized trials. We know that these 22 are practical problems that we are going to run into, 23 and how are we going to deal with them. How are we 24 going to manage then and how are we going to account 25 for that statistically? That is not going to be an 125 1 easy -- there is no easy answer to that and there may 2 be no answer to that, which is I think the point I am 3 trying to make. 4 DR. PRICE: I agree. 5 DR. DALE: I'd just make a brief comment. 6 In the randomized trials that Ron Strauss reviewed, 7 the survival with transfusions was roughly 11 percent 8 for invasive molds. So historically, there is really 9 not much evidence. You would have to base evidence of 10 treatment benefit on new results like were described 11 today. And they are not certainly clear cut at this 12 point to know the benefits. So I feel like from an 13 ethical standpoint, it is a reasonable thing to do, 14 particularly before we and others encourage the more 15 widespread application of this very expensive and 16 resource consumptive technology. 17 DR. TORLINI: Hi, Sergio Torlini Inova 18 Fairfax Hospital. Since the granulocytes go to a 19 population that is almost 100 percent immune 20 suppressed and therefore at risk of graft versus host 21 disease and therefore the products have to be 22 radiated, I would like to know if anybody has any data 23 as far as the radiation on granulocyte function, I 24 mean anything recent, and also on the dose of 25 radiation on that. 126 1 DR. DALE: I will comment briefly. We 2 haven't studied that in Seattle. At MD Anderson, 3 there are some recent studies that do suggest some 4 modest degree of cell injury with radiation. That 5 may, in fact, be overcome by the addition of other 6 cytokines, the cytokine being gamma interferon, for 7 example, having a protective effect. It is a murky 8 area, though. And basically I think what has been 9 done universally is to irradiate cells assuming they 10 are not damaged. So it is something that needs to be 11 better studied. 12 DR. LEITMAN: There is a very nice study 13 by Kasberg, Lur, and colleagues in Switzerland 14 published in Blood in 1993 where I can't remember the 15 dose, it was 2,500 or 3,000 centigrade, were given to 16 the product and before and after radiation very 17 careful studies of migration and chemotaxis and 18 fungicidal and on the respiratory burst activity. 19 There were 5 or 6 parameters looked at and there was 20 no difference pre and post to radiation. 21 DR. DIAZ: Jose-Luis Diaz from Mitre 22 Pharmaceuticals in La Jolla. I have a simple nuts and 23 bolts question for Dr. Ambruso, which is how did you 24 measure apoptosis in the cells that you were 25 measuring? 127 1 DR. AMBRUSO: How did we measure it? 2 DR. DIAZ: Yes. 3 DR. AMBRUSO: We looked at morphologic 4 criteria. That slide that I showed you showed a 5 distinct change in the morphologic appearance of the 6 nuclei. And we looked at percent of live apototic 7 cells. 8 DR. DIAZ: Right. And the other thing was 9 in some of the other functional assays, how did you 10 distinguish the response you were getting -- how did 11 you determine that the response you were getting was 12 from a neutrophil and not from something like a 13 monocyte, for example. Obviously monocytes can burst, 14 et cetera. 15 DR. AMBRUSO: Sure. Our preparations are 16 95 to 98 percent neutrophils, and that was checked 17 each time that the cells were separated. 18 DR. DIAZ: Oh, I see. So you took the 19 samples and then did a separation? 20 DR. AMBRUSO: Right. There was a further 21 separation. 22 DR. DIAZ: Okay. Thanks so much. 23 AUDIENCE MEMBER: I would like to comment 24 on the respiratory distress that you saw, Susan, with 25 some of your granulocyte preparations. We see a lot 128 1 of respiratory distress after transfusion and only the 2 anti HLA2 agglutinating antibodies of the HLA 3 antibodies apparently cause respiratory distress. Now 4 the antigranulocyte antibodies for the granulocyte 5 antigen specifically also cause respiratory distress, 6 and we find that we have many, many more 7 antigranulocyte antibodies causing respiratory 8 distress than HLAs. But in HLA territory, if you can 9 avoid giving anti-HLA2's -- giving HLA2's to the 10 people who might have anti-HLA2's, you might avoid 11 that small group. But they look to be only about a 12 fifth or a tenth of the respiratory distress cases 13 that we have before us. So I wonder if anybody else 14 has any different experience with that. We have been 15 very interested in that because it is a very serious 16 complication. 17 DR. LEITMAN: In the retrospective 18 analysis of pulmonary complications in CDG recipients 19 of granulocytes that was published by Dave Stroncek a 20 year ago with the CDG patients here, the pulmonary 21 reactions were seen with HLA alloantibodies alone in 22 the absence of anti-neutrophil antibodies, and all 23 patients were screened for presence of both anti- 24 neutrophil and anti-HLA. I am almost certain that it 25 wasn't only HLA A2. Most of the recipients had multi- 129 1 specific antibody screens. I am not sure we can pull 2 out the A2. But I don't think it was only HLA A2. 3 DR. PRICE: Becky, are you referring to 4 granulocyte things only or is this trolley type stuff 5 you are talking about? 6 AUDIENCE MEMBER: This is trolley type 7 stuff. Because we have seen it with the anti- 8 granulocyte. We have seen it anti-HLA2, but with 9 other kinds of anti-HLA. I would certainly be willing 10 to learn. I need to go read David's paper again 11 obviously. 12 DR. AMBRUSO: Pulmonary reactions related 13 to transfusions is something that our laboratory has 14 been interested in in the past few years and it 15 certainly occurs for reasons other than antibodies. 16 And these patients, aside from the granulocyte issue, 17 certainly are receiving other products and have other 18 illness-related problems that certainly could 19 predispose them to that. I am surprised that, in 20 fact, there aren't more and it is something, more 21 reactions, and may not necessarily be related to the 22 antibodies or the granulocytes themselves. But there 23 are other compounds that have been indicated -- 24 cytokines, lipids that were intensely interested and 25 that certainly have an etiologic role. So this gets 130 1 very murky when we start talking about pulmonary 2 reactions and there are lots of ways that it can 3 happen. 4 DR. KLEIN: Harvey Klein, NIH. Is anyone 5 looking at or concerned with the issue of CMV 6 transmission when patients who are immunosupressed get 7 these large numbers of granulocytes over long periods 8 of time? 9 DR. PRICE: Well, I mean the other people 10 can answer this. Our routine, of course, is for any 11 recipient who is CMV negative to provide donors that 12 are CMV negative. Whether a CMV positive donor 13 getting CMV positive stuff is going to get some other 14 strain and do him some damage is not something that we 15 have addressed. 16 DR. AMBRUSO: We have routinely used the 17 same or adopted the same approach and that is either 18 using sero negative or using leuko-reduced. But most 19 of what we are doing still is sero negative -- CMP 20 sero negative for patients that would fall into that 21 category. And I assume we would continue. 22 AUDIENCE MEMBER: Could I ask anybody, 23 what would be the upper level of neutrophils that you 24 would exclude a person from this proposed trial? 25 DR. DALE: Do you mean what -- 131 1 AUDIENCE MEMBER: Yes, what degree of 2 neutropenia do you think that you would think 3 neutrophil transfusions would have an impact? 4 DR. DALE: That is a good question. The 5 traditional or historic level of 500 is a level that 6 is cut off as a pretty high level. The risk of 7 infection is the severity of infection is considerably 8 more if you use 200. You could argue for using even 9 a higher threshold based upon studies of functional 10 deficiencies of the neutrophils produced after 11 transplantation in the early phases of hematopoietic 12 recovery. I don't think that any choice would be less 13 than arbitrary, though. It would fall somewhere 14 probably between 200 and 1,000. 15 AUDIENCE MEMBER: Could I just ask another 16 question? If you were treating a fungal infection, 17 would you disqualify people for this trial on 18 steroids? 19 DR. DALE: That is another very good 20 question. In the post-transplant period when people 21 are on steroids, that is going to be an important 22 factor in their infection. And it has been thought 23 that it would be useful to transfuse people with 24 higher counts because again the endogenous cells are 25 perhaps ineffective. That is mostly speculation, 132 1 though. I don't think anyone -- there is not a right 2 answer. So that whatever trial were conducted would 3 be some consensus. But certainly people I have talked 4 with have suggested that for fungal infections after 5 transplant for people on steroids, you should consider 6 people with higher counts than 200 or 500. 7 AUDIENCE MEMBER: But wouldn't they 8 suffer, then, the same paralysis that is going on in 9 vivo? 10 DR. DALE: Yes and no, and that depends 11 upon whether the effects of steroids or 12 immunosuppressive drugs have their effects on 13 developing cells or developed transfused cells. And 14 I suspect it is the former, not the latter. But, 15 again, another good research question. 16 DR. STRONCEK: Dave Stroncek, NIH. A 17 couple of comments. One is I know the average data 18 looks very good for the increments and neutrophil 19 counts after transfusions and patient outcomes. But 20 I have seem some of these same patients anecdotally 21 and I don't think the average data -- I don't think 22 the whole picture is quite as rosy as the average data 23 tends to show. That said, I think when you consider 24 clinical trials, I have a similar concern about the 25 patients studied as Dr. Adkins but for a different 133 1 reason. I think bacterial infections granulocyte 2 transfusions will work great for. But the number of 3 patients we see for that have bacterial infections not 4 responding to antibiotics is really very rare. The 5 biggest problem seems to be patients with fungal 6 infections. And I have a concern that the granulocyte 7 transfusions may be effective in increasing white 8 count, but they still might not be effective in 9 resolving fungal infections, or at least not effective 10 enough to see in a reasonable trial. So for that 11 reason, if you set up a trial that just looks at 12 treating fungal infections in neutropenic patients, it 13 may fail. So it may be worthwhile to try it in a 14 different patient population, maybe in the 15 prophylactic setting. 16 And the other comment too is I know in the 17 past we had to look at really patient survival as an 18 outcome because that was the only outcome measure we 19 had. You couldn't measure increase in granulocyte 20 counts. But I think that needs to be relooked at in 21 further studies. A number of the studies on cytokines 22 that have been given to patients with chronic 23 neutropenia and neutropenic patients getting 24 chemotherapy have shown shorter days of neutropenia 25 and shorter days of hospital stay and decreased 134 1 febrile incidence, but I don't think they have been 2 held to the same standard to show that there is less 3 infection -- less fungal and bacterial infection. So 4 I just caution that we don't set ourselves up to fail 5 because we are being more strict with the criteria to 6 say granulocyte transfusions don't work. If you do it 7 because we are more strict than any other standard we 8 hold for other blood products or other drugs. 9 DR. HENDERSON: I am Theresa Henderson 10 from Georgetown University and I have a couple of nuts 11 and bolts questions for Dr. Ambruso. I just wanted to 12 clarify that the cells that you were testing were not 13 the apheresed product but had been stored and cleaned 14 up by fical or something like that? 15 DR. AMBRUSO: Right. The actual data that 16 I have showed was cells from the -- they were drawn 17 from the patient before any G-CSF had been given. And 18 approximately one hour after the dose of G-CSF on day 19 4. So that is the fifth dose of G-CSF. Subsequent to 20 that, the patients underwent granulocyte collections 21 and I didn't show the data but put it in a summary 22 slide that we looked at additional studies in those 23 granulocytes. And we really didn't find much 24 difference in terms of the function for the first 24 25 hours. I had 24 hours of storage. 135 1 DR. HENDERSON: Oh, that is good to know. 2 And you said that you looked at C-11B and saw no 3 significant changes. I wondered if you looked at 4 Selectin-62? 5 DR. AMBRUSO: We did not. 6 DR. HENDERSON: And you also touched on 7 the cytosolic calcium increase. I am sorry, I don't 8 understand -- I am new at this, so I don't really 9 understand the significance of looking at that. Could 10 you enlighten me? 11 DR. AMBRUSO: We did that in part because 12 we wanted to look at a biochemical parameter that we 13 thought was related. I mean initially we did it and 14 were excited related to the chemotaxis. But obviously 15 calcium and signaling related to calcium is something 16 that is significant to all functions of the 17 neutrophil. We also did it in response to some 18 studies in another patient group. There is a group of 19 patients who have neutropenia and neutrophil 20 dysfunction that are associated with glycogenesis 1B. 21 These individuals, when you put them on G-CSF, their 22 neutrophil counts are restored. They are not 23 neutropenic. They still have a chemotactic defect and 24 their cytosolic response to calcium, which was 25 aberrant to begin with without treatment with G-CSF 136 1 normalized. So we wanted to look at a group of 2 control individuals given G-CSF to see if that affect 3 was consistent. It is consistent. I don't know if it 4 has any relationship to the defects or anything that 5 was shown in the function of the cells in normal 6 patients given G-CSF. 7 DR. HENDERSON: Thank you. 8 DR. DIAZ: Just a quick point. About L- 9 Selectin. We did some studies on isolated neutrophils 10 looking at 62L, which is marked for L-Selectin, and 11 what we found was that in isolating neutrophils, the 12 L-Selectin dies off slowly and after about 24 hours, 13 there is only about 50 percent. And then by about 48 14 hours, there is only about 20 percent left. If at any 15 of these points you actually activate the cell by 16 giving it a stimulus, it disappears within an hour or 17 very quickly. 18 DR. LEITMAN: I have a question for the 19 FDA. You can now define a product much better than 20 you could before. You can now define a measurable 21 outcome in a transfusion recipient. You can even set 22 criteria for assessing that outcome, an increment in 23 the granulocyte count at certain times that is 24 sustained for a certain length of time. So your 25 definition of the component of how you treat the donor 137 1 and what you measure immediately after transfusion in 2 the recipient is quite well defined. Would the FDA, 3 in considering licensure of this product, have to see, 4 want to see, insist on seeing the results of a 5 randomized clinical trial? And what would be the 6 endpoints that they would look at of that trial? 7 CHAIRPERSON HARVATH: Susan always asks me 8 these tough questions. That is why we are holding the 9 conference. We wanted to hear from you what you feel 10 is an appropriate approach because we have heard that 11 this is very expensive. We also hear that normal 12 donors may be asked to be receiving cytokine 13 repetitively. And as you know, we have to constantly 14 look at risk benefit in terms of public health issues. 15 I agree with everything you have said. I think that 16 you have made enormous strides in defining a component 17 and in terms of the cell biology. I think that that 18 is the easier part of all of this. One of the things 19 we wanted to hear from this workshop and yesterday 20 were your concerns about the exposure of normals and 21 the effects long term. So I don't really have an 22 answer for your question, Susan. I think what we 23 wanted to do was hear what you were all comfortable 24 doing. We know some folks have talked about 25 collectively doing a randomized trial. We know the 138 1 NIH is interested in hearing what the interest is in 2 that and in determining whether there would be support 3 to try and do that. We also know that is going to be 4 very expensive to be done and I think you said it very 5 accurately this morning that it probably would be done 6 once and hopefully done so that one gets a clearer 7 answer. 8 So I honestly don't have an answer. I 9 can't speak on behalf of the Agency. As you know, I 10 am interested in the cell biology and also I think we 11 have an obligation to ask investigators who are 12 collecting these products how comfortable they are 13 with giving their normal pedigree blood donors 14 cytokines and giving them cytokines perhaps on more 15 than one occasion. 16 One question I have for all of you along 17 that line -- and I am sorry I don't answer your 18 question but that is my non-answer. One question I 19 have for you is I have heard all of you say that you 20 are working with your pedigree donors. So you really 21 have a very clear picture of their medical histories 22 and their hematologic picture. When you do a complete 23 blood count on someone, are you including a 24 differential in that and how would you feel about 25 giving G-CSF or some other cytokine to someone where 139 1 you only had a blood count and not necessarily were 2 resting assured that they had a normal differential? 3 Do any of you have concerns that there might be an 4 individual perhaps with perhaps a pre-leukemic state 5 that could possibly be missed? That might not be your 6 normal pedigree donor population that you have 7 followed over the years and have a clear health 8 picture on. But something we have kind of wondered 9 about and we don't have the answer to it is what kind 10 of -- would you include a differential in your workup 11 initially in a new donor situation? And that is for 12 everyone. 13 DR. PRICE: Well, we do a differential, 14 but by the time we get the answer back, the G has 15 already been given to the donor. Like you say, these 16 are donors we know. But a guy who is going to develop 17 leukemia will do it at point X and that may happen. 18 I think -- I mean, my read on this is that there have 19 been a lot of people looking at this and the evidence 20 of giving one shot of G even to somebody who is 21 developing leukemia that it is going to do him any 22 harm would be about zero. 23 DR. AMBRUSO: We, like you, are doing 24 counts with automated differentials on our granulocyte 25 donors as well as we are doing counts on platelet 140 1 donors. That is something that we worry about and I 2 wonder if you would be more concerned or more likely 3 to run into problems with very high counts -- just 4 with the white count and looking at that carefully 5 than with the differential. I am not sure that we 6 collect a lot of reasonable data. The parameter that 7 should be different is looking at counts that exceed 8 or are above or below certain levels in terms of the 9 risk for leukemia or other problems. 10 DR. LEITMAN: We automatically do a CBC 11 with an automated differential on all our donors 12 including platelet donors, and part of the reason is 13 to look at their platelet counts to qualify them for 14 the next platelet pheresis donation. Because the 15 granulocyte donor one month later is most likely to be 16 donating platelets and not granulocytes again. And 17 that is the reason I think most centers if not all 18 have to do a pre-pheresis count. Like Tom, our count 19 is obtained the day of pheresis in the current 20 iteration of the study so that we get the count after 21 they have been given the stimulant. Does the count 22 have to include a differential? It is almost a moot 23 point since the automated instruments provide the 24 differential. But I don't see in the absence of a 25 trial the critical necessity for a differential. 141 1 DR. DALE: I would just comment too that 2 this safety issue, that is, a single dose must be an 3 extremely small risk. Because the substance that is 4 used as a drug is really very close to the natural 5 hormone and a surge of this magnitude probably happens 6 many times in a person's lifetime. 7 AUDIENCE MEMBER: I have a question. What 8 are the legalities about giving G-CSF to an unrelated 9 donor? I know that in our institution we would have 10 to go through IRB to get this approved. Is that 11 pretty much the way it stands for everybody right now? 12 DR. DALE: We have -- we do it with IRB 13 approval and informed consent. 14 DR. LEITMAN: I think in this 15 institution, all G-CSF given to normal donors on 16 numerous protocols, not only ours, is done with IRB 17 informed consent. 18 AUDIENCE MEMBER: All right. And I guess 19 a follow-up question is what exactly -- how do you 20 modify your informed consent form for a G-CSF? I 21 mean, how extensive is it? Several pages? Just a 22 paragraph? 23 DR. LEITMAN: It is four pages. 24 AUDIENCE MEMBER: A paragraph. Thank you. 25 We were thinking of producing a movie to go with it to 142 1 show. 2 DR. PRICE: And I think that although 3 maybe everybody up here does it in IRB, I think there 4 are places around who don't do it in IRB. I still 5 think they have a consent form and explain to the 6 donor what is going on. But I figure that there is 7 enough information around that as long as they are 8 keeping track of things that it is not really a 9 research thing. I think that is not undefensible. 10 Did I say that right? Too many negatives? 11 AUDIENCE MEMBER: Having an almost 20-year 12 history in apheresis collections and a nursing 13 background, I have to preface this by saying that I am 14 a donor advocate and having been a nurse and being 15 trained to try and help people get better, I have a 16 real internal conflict going on with asking volunteer 17 donors to take a drug that we don't have long-term 18 studies as to show what happens with these donors with 19 exposure to the drug. And then tied in with that is 20 the fact that there really -- if I am hearing 21 correctly what I have heard today, the efficacy of 22 what we are trying to collect really has not been 23 established. And I am real worried and terribly 24 protective of the donors and what we ask of them. I 25 think also as a personal observation of donors that we 143 1 have worked with in the past, yes, the donors will 2 come back after many circumstances and I would be 3 willing to wager that because they are a captive 4 audience, they are a very dedicated conscientious 5 group of people that would probably -- and I 6 acknowledge that I have no data to support this fact 7 other than a gut feeling -- but I think the donors 8 would probably come back no matter what we asked of 9 them, and where do we draw the line between what we 10 ask of our donors and what we are trying to accomplish 11 in our patients? 12 DR. DALE: I would just comment. I think 13 that is a very good statement and in fact provides 14 part of the rationale for a randomized trial. It is 15 not to do something with donors which is not 16 convincingly proven to be of value. Protect their 17 interest as well. On the other hand, I think being 18 open with people and honest about the potential 19 benefit and the known risk that the donor population 20 can make an informed choice. As adults, some will say 21 yes and some will say no and that is okay. 22 DR. AMBRUSO: I think you bring up an 23 important point and I think we really do need to know 24 long term effects. I guess I am not concerned about 25 what happens with one dose of G-CSF and then the donor 144 1 goes merrily on their way. People who are involved in 2 providing blood products know that you are relying on 3 the bone marrows of a very few individuals to support 4 the rest of the community. And it is really not a 5 single donation per year that we are concerned about. 6 It is the donation of something of someone who might 7 undergo this procedure many times in a year and that 8 is completely undefined. As you were saying this, I 9 am reminded also that we have many, many dedicated 10 platelet donors. And you know, we don't know long- 11 term what happens to somebody who is having their 12 platelets collected many, many times a year -- 23 or 13 26 times a year. You know, we don't have long term 14 data on that, but we still go ahead and do those times 15 of collection. So this is a muddy area, but I think 16 we really do -- with G-CSF mobilization, we do need to 17 get some long-term data, particularly on the donors 18 who are going to be dedicated and are going to donate 19 more than once a year. If they are going to be 20 denoting once a month, I think there is some long term 21 information that we need on them. 22 DR. LEITMAN: The NIH has been collecting 23 granulocytes for transfusion for greater than 15 24 years, maybe for 20 years. Since the Michler studies 25 in the 1970's about starch and steroid to optimize 145 1 collection efficiency, donors have routinely been 2 given dexamethasone plus starch. Neither of those 3 medications are licensed for use in obtaining 4 granulocytes because granulocytes aren't a licensed 5 product. So the past two decades in this field has 6 been characterized by giving donors drugs that do have 7 adverse effects. 45 percent of donors getting dexa 8 don't have a good night sleep. They have nightmares 9 and insomnia. I see nothing in the past decade of a 10 history of administering G-CSF to normal, healthy 11 individuals that increases my concern that G-CSF has 12 any long term, adverse consequences above and beyond 13 what we know about dexamethasone and starch. Acutely, 14 certainly, it causes more discomfort, which is why I 15 think you can expect a 10 percent dropout in 16 individuals participating in this, which is fine. 17 AUDIENCE MEMBER: I would hope -- this may 18 be rather unpopular, but I would hope that the FDA 19 would like to see or would want a randomized trial, an 20 appropriate trial. I think one could ask the question 21 that if you cannot demonstrate efficacy, number one 22 does it matter and number two, who do you decide to 23 give it to? Are you going to give it to a huge number 24 of patients who may not need it or who don't need it 25 or where you can't demonstrate efficacy? So I think 146 1 to study it and to determine how you ought to use it, 2 you really need an appropriately designed trial. It 3 may be too late to do that and this is a plea, 4 perhaps, fir people to begin to do appropriately 5 designed trials very early in the course of the 6 development of a new component. Many years ago -- 7 some of us are old enough to be more interested in 8 history and perhaps unrecorded history. 9 Everybody knew that fibrinogen given to a 10 woman bleeding from low fibrinogen after a pregnancy 11 that the fibrinogen was effective until a major 12 university did a comparison trial of fibrinogen versus 13 no fibrinogen and found that the increase in 14 fibrinogen was as rapid without it as it was with it. 15 That was a pretty well defined derivative. It could 16 be measured and you gave a dose that you knew about. 17 So I think being able to define a component does not 18 necessarily mean that it is going to be effective and 19 useful. 20 DR. SHAPRIO: Arell Shapiro from Life 21 Source. From the data presented, I didn't get a very 22 good understanding of how the patients are treated. 23 Is it -- you know, once it is ordered, is it daily? 24 I mean, Dr. Leitman, in your presentation it was daily 25 except for the weekends. Is that how -- it is just 147 1 ongoing until the person either recovers back their 2 white count? I mean, what are the endpoints and when 3 do people give up? 4 DR. LEITMAN: The call comes to the 5 consult service to consider granulocyte transfusions 6 for a patient. One of our fellows immediately within 7 hours sees the patient and performs a full evaluation 8 and it is not a 100 percent approval. The Transfusion 9 Medicine Department takes a very active role in 10 deciding and looking at everything involved with that 11 patient as to whether this is a good thing to do. 12 Once we start, it is daily, omitting in general 13 weekends but depends on how critical the patient's 14 status is and we do have the potential to collect on 15 Saturday and Sunday and holidays, which we do. 16 You raise a very good point. In some 17 individuals, the increment after the first two 18 transfusions is so high and so sustained one can ask 19 whether it is necessary to continue daily transfusions 20 especially in small kilo recipients such as children 21 or small sized adults, and in such individuals on such 22 cases with discussion with the clinical staff we have 23 moved to every other day transfusions and the 24 neutrophil count does not diminish generally to zero. 25 AUDIENCE MEMBER: So basically you just do 148 1 it by following the patient and see what their 2 response is. Have you thought about doubling up on a 3 Friday so that you could sustain them over the weekend 4 -- give them a double dose? 5 DR. LEITMAN: The data on storage for 6 longer than 24 hours is problematic. So we do it 7 every day. 8 AUDIENCE MEMBER: No, I mean infuse both 9 on Friday. 10 DR. LEITMAN: Since the increment is in 11 the 2,000 range with the single transfusion, that is 12 enough for us. We have not done double doses. 13 DR. PRICE: Our general approach is 14 similar. Once we start, we try to do it daily. We 15 even do it on the weekends. The endpoints are a 16 little bit muddier. Part of this comes back to the 17 question that was asked earlier of what is a good 18 neutrophil count to have. 19 And once you get it up to 2,000 are you 20 okay or is somebody with a fungal infection, would 21 they rather have a count of 8,000 than they would of 22 3,000? And our feasibility study didn't really 23 control that and it was kind of up to the clinicians. 24 But that would have to be very carefully laid out for 25 a real trial. 149 1 AUDIENCE MEMBER: We have donors who are 2 willing to take G-CSF for many days in a row. Have 3 you looked at seeing if you can collect granulocytes 4 from your community pools on a daily basis for say 5 four or five days and would this be of any benefit? 6 DR. PRICE: We haven't. Part of this is 7 that I -- I mean, I think you are in two different 8 ball games when you are talking about a one-shot thing 9 and a multi-thing as Dan was talking about. I mean, 10 you run into are the cells the same after five days, 11 say of G, as they are 12 hours later. You also run 12 into issues of cell separation efficiencies if the 13 cells aren't really different. 14 But I think at least from my point of 15 view, that wasn't something I was willing to ask a 16 regular pheresis donor to do. We have -- as you know, 17 Scott, on our early studies we did that with the bone 18 marrow donors from the Hutch. And when we did that, 19 there were several of those donors that did have some 20 problems getting daily starch. 21 We had to stop collections because people 22 got fluid overload and got bad headaches and things 23 like this. It didn't happen all the time, but there 24 was enough of that that for our community donors it 25 seemed to me that it was sort of above and beyond the 150 1 call of duty just to do it once. 2 AUDIENCE MEMBER: That is the reason why 3 we use pentastarch in our repheresis unit because of 4 the multiple collections that they will be undergoing 5 over -- up to 12 days is what we allow at our center. 6 DR. LEITMAN: I just want to agree with 7 Tom Price. There is a limit beyond which I will not 8 ask a normal healthy community volunteer to do 9 something even under the auspices of trial, and I 10 think one granulocyte aphoresis collection is all I 11 would ask them to do rather than serial. We have 12 occasionally, when we have been unable to get a donor 13 in over the holidays or whatever, I have had a donor 14 donate two weeks after their last donation. But that 15 is a very rare event, once a month is what I feel 16 comfortable asking an individual to do. It is really 17 an entire day dedicated to apheresis. They don't feel 18 that well afterwards. Their work is affected and 19 their home life is affected. I think that is as much 20 as I would ask. 21 DR. PRICE: And another issue that comes 22 up there, and I don't know if you are still doing 23 this, Scott -- but when we were doing the family 24 donors, those guys had central lines and that is also 25 something I am not too interested in having a regular 151 1 community donor have to have. 2 DR. CONFER: Dennis Confer, National 3 Marrow Donor Program. On the issue of G-CSF safety, 4 I hope it is sort of a rhetorical question, but the 5 question is, is there really anybody among us who 6 would give G-CSF to a normal donor if in fact we 7 thought there was a reasonable chance that it would 8 cause some late hematologic effect? If we thought 9 that there was some reasonable chance that in fact 10 these donors would develop leukemia at a higher rate 11 or if we thought that there was some reasonable chance 12 that in fact these donors might develop aplastic 13 anemia at old age, I think none of us would give G-CSF 14 or any other hematopoietic growth factor to a normal 15 donor. I think in the same time, I am convinced in 16 making preparations to give G-CSF to volunteer 17 hematopoietic stem cell donors, I am confident that 18 this drug has no long-term effects. But I think to 19 demonstrate that, the best way is to collect the long- 20 term follow-up data. 21 And we will make plans to follow these 22 donors for as long as we can. And it is interesting 23 because it is something that we talked about 24 yesterday, that we haven't done it with bone marrow 25 donors. That is truly a deficiency. We also need to 152 1 follow the routine bone marrow donors for as long as 2 we can to prove that that in fact doesn't cause late 3 effects, which again we are confident or we feel 4 reasonably confident that it doesn't. 5 But the other thing to keep in mind is 6 that I can assure you that people who get G-CSF will 7 develop leukemia. Because normal people develop 8 leukemia who have never been exposed to G-CSF. And 9 the real question is not how many cases or whether 10 people develop leukemia, the cases is how many cases 11 develop and is that different than what would have 12 been expected among a normal control population. So 13 the data we collect really has to be very 14 comprehensive. 15 And it has to be compared to an 16 appropriate control population in order to determine 17 whether the incidence of leukemia is, in fact, 18 excessive. Because it will occur. And I know from 19 experience that it has occurred in bone marrow donors, 20 both before and following bone marrow donation. So it 21 is a matter of how much, and that is just going to 22 take a long time to answer that question. 23 DR. PRICE: Yes, I think -- because what 24 we are doing is the mortality of our community donors 25 eventually is 100 percent. 153 1 CHAIRPERSON HARVATH: With that concluding 2 statement, I would like to thank all of you very much 3 for I think a very informative session this morning. 4 We will convene in an hour. 5 Thank you. 6 (Whereupon, at 12:05 p.m. the workshop 7 recessed for lunch to reconvene this same day at 1:10 8 p.m.) 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 154 1 A-F-T-E-R-N-O-O-N S-E-S-S-I-O-N 2 (1:10 p.m.) 3 DR. STRONCEK: I am Dave Stroncek and I 4 am going to moderate this session. We have three 5 presentations and then after that we will have a panel 6 discussion. Following the panel discussion, we only 7 have three abstracts which will be presented and then 8 a little further discussion. So we anticipate we will 9 probably get done a little early today. But the first 10 lecture today will be by Dr. Douglas Adkins. Dr. 11 Adkins is Assistant Professor of Medicine at 12 Washington University School of Medicine, and he is 13 the medical director of their National Marrow Donor 14 Transplant Collection Program. And he will talk to us 15 today about granulocyte product evaluation. 16 DR. ADKINS: Okay. I'd like to thank 17 Liana Harvath for inviting me to discuss our data from 18 St. Louis on granulocyte transfusions that are 19 mobilized with G-CSF. Could I have the first slide? 20 Liana has asked me to discuss product evaluation in 21 this area, and I will focus on cell dose and leukocyte 22 compatibility. Because at least at this point in time 23 in my opinion, these are perhaps the two most 24 important issues in this area today. 25 I don't have any specific recommendations 155 1 about either of these two issues, but I would suggest 2 or put a vote in to provide additional research 3 support to determine efficacy parameters as they 4 relate to product evaluation. 5 In my opinion, granulocyte transfusions 6 may be the only strategy which has the ability to 7 eliminate severe neutropenia after high dose therapy, 8 which is obviously the reason to pursue this area. It 9 is important to ask what are reasonable goals to 10 achieve with granulocyte transfusions. Certainly it 11 would be nice to show that these products reduce 12 febrile days, antibiotic requirements, reduce the 13 occurrence of documented infections, and to 14 successfully treat documented infections as we have 15 discussed. Perhaps in high risk populations, one 16 might be able to demonstrate reduced mortality. 17 I think as you have already seen this 18 morning, it can be very difficult to demonstrate these 19 efficacy parameters in clinical trials. There are 20 many causes, for instance, of febrile days beyond just 21 an infection. Transfusions can cause that, 22 antibiotics, et cetera. So it introduces a lot of 23 confounding variables in trying to set up a well- 24 designed clinical trial in this area. 25 Why use granulocyte transfusions to 156 1 prevent neutropenic infection related complications? 2 Well, certainly there are preclinical models, as you 3 know, that demonstrate a clinical benefit in terms of 4 improve survival with granulocyte transfusions. In 5 addition, there are two meta-analyses of randomized 6 human trials that also demonstrated a clinical 7 benefit, but only if certain conditions were met. 8 This is data from two decades ago published by Fred 9 Applebaum demonstrating the critical importance of 10 component cell dose on efficacy. In this case, which 11 is a model of neutropenic dogs with Pseudomonas 12 bacteremia, this data demonstrated the steep dose 13 response curve that they observed in this trial. And 14 if one was able to infuse products containing more 15 than 2 times 108 per kilogram, uniform survival was 16 observed. And interestingly, the same threshold dose 17 would correlate with significant increments in ANC. 18 This kind of data suggests to me that it is important 19 to demonstrate significant increments in the ANC to go 20 on to demonstrate measures of efficacy. And 21 interestingly, in a 7 kilogram human, this threshold 22 cell dose correlates with about 1.4 times 1010 23 granulocytes. 24 As I mentioned, two meta-analyses of 25 randomized trials of granulocyte transfusions in 157 1 humans have demonstrated benefit. In the analyses of 2 prophylactic granulocyte transfusions, granulocyte 3 transfusions that contain an adequate dose of 4 leukocyte compatible components resulted in decreased 5 relative risk of infection, mortality, and death from 6 infection. In the analysis of therapeutic granulocyte 7 transfusions, these products resulted in improved 8 survival if an adequate dose was administered and if 9 they were given to patients with inherently low risk 10 or low likelihood of survival. 11 The combination of these trials suggest 12 that cell dose is a very important determinant of 13 efficacy and perhaps leukocyte compatibility as well. 14 These are trials of non-G-CSF mobilized granulocyte 15 transfusions. 16 Historically, limitations of granulocyte 17 transfusions have primarily been issues revolving 18 around low cell dose and the presence of leukocyte 19 incompatibility. Both of these factors are probably 20 the primary determinants or the primary cause of poor 21 and nonsustained increments in the ANC with 22 granulocyte transfusions not mobilized with G-CSF. 23 Another important problem in prior trials was the 24 frequent occurrence of febrile transfusion reactions, 25 a problem which confounded the analysis of efficacy. 158 1 It is my hypothesis and perhaps that of others in this 2 room that significant and sustained post-transfusion 3 increments in the ANC may be a key pre-condition for 4 demonstrating reproducible improvements in clinical 5 outcomes with granulocyte transfusions. So that 6 became the focus initially of our research in this 7 area. 8 Let's look at ways of improving component 9 yield or component cell dose. Most people in the last 10 few years have been focusing on trying to increase 11 donor ANC, and certainly as Tom Price has shown, that 12 is probably the most important determinant of 13 granulocyte component yield. And so with growth 14 factors such as G-CSF, we can clearly increase the 15 donor ANC pre-collection. There are apheresis 16 factors, though, as Susan Leitman has eluded to which 17 we need to work on to try to improve component yields 18 with apheresis. I will talk about our data with 19 varying the interface offset and showing how that 20 influences component collection efficiency and yield. 21 This is a table that I put together which 22 shows you the component granulocyte dose based on 23 choice of mobilizing agent. And as has been discussed 24 earlier, with granulocytes collected at steady state 25 and without a red cell sedimenting agent, one can only 159 1 collect about .1 to .3 times 1010 cells. With 2 corticosteroids and hetastarch, the average collection 3 contains about 2.3 times 1010 cells but no better than 4 3 in general. With G-CSF, it has been our observation 5 that one can increase component cell doses by a factor 6 of several fold, number one, and number two, the 7 larger the dose of G-CSF that we have given to normal 8 donors, the greater the component yields, as you can 9 see here. We have already seen data from Seattle and 10 also from this institution showing that there is an 11 added benefit of giving Decadron along with G-CSF and 12 resulting component yields. I am aware of at least 13 one study that has used GM-CSF to collect 14 granulocytes, but that was an abstracted report which 15 did not comment on component cell doses. 16 As I mentioned, although most people in 17 the last few years have focused on ways of increasing 18 donor white count as a strategy to increase component 19 yields, we have also looked at the effect of altering 20 apheresis parameters. One that we have looked at is 21 the interface offset setting. As Susan mentioned, the 22 machine -- using the Baxter device, the machines 23 default setting is 15. And to my knowledge, there are 24 actually quite a few centers that do use that 25 particular setting, even today. So we did this study 160 1 where we collected granulocytes using three different 2 interface offset settings varying from 15 to 25 to 35, 3 and we determined the effect on granulocyte collection 4 efficiency and granulocyte yield. These were all 7 5 liter pheresis and donors receiving G-CSF at 5 mcg per 6 kilogram along with hetastarch. 7 So as you can see, as you increase the 8 interface offset setting, you improve granulocyte 9 collection efficiency from 40 to 60 percent. And this 10 results in improved yield as you can see here. So our 11 practice now is to use an IO setting of 35 instead of 12 our old practice of 15. 13 This is data we recently published 14 demonstrating again, as others have, that significant 15 and sustained increments in the recipient ANC occurs 16 with transfusion of G-CSF mobilized HLA matched 17 granulocyte components. Our initial clinical model 18 was to collect granulocytes from HLA matched sibling 19 bone marrow donors who were receiving G-CSF and 20 transfuse these products as prophylaxis against 21 infection. The advantages of this model was that we 22 chose allogeneic bone marrow transplant patients who 23 had an expected interval of severe neutropenia during 24 which most indeed developed complications of that 25 problem. There was also an available HLA matched or 161 1 leukocyte compatible granulocyte donor, and that 2 person would be motivated to undergo frequent 3 granulocyte collections for their sibling. In my 4 opinion, prevention of infection may be an objective 5 more easily reached with granulocyte transfusions than 6 successful treatment of established infections. Many 7 times these are associated with multi-organ failure or 8 confounding causes for infection which complicates the 9 interpretation of efficacy trials. 10 The limitations of this model is that the 11 donor has to be ABO compatible and the donor must be 12 HLA matched. And obviously the donor would have to 13 agree to undergo additional time commitment to 14 participate and perhaps somewhat greater risk over 15 just bone marrow collection. 16 The objectives of this study were to 17 carefully document the kinetics of the recipient ANC 18 with each granulocyte transfusion. The donor 19 underwent bone marrow collection on day zero, 20 transplant day zero, and then received G-CSF daily for 21 five days at 5 mcg per kilogram. They then underwent 22 alternating day granulocyte collections on days 1, 3, 23 and 5. The recipient underwent marrow infusion on day 24 zero and then received G-CSF daily from day plus 1 25 until neutrophil engraftment. Granulocyte 162 1 transfusions were given fresh on days 1, 3, and 5, and 2 then we carefully determined ANCs pre and post- 3 transfusion as you can see here. 4 This is a table demonstrating for you the 5 increments that we observed after transfusion of these 6 granulocyte products on days 1, 3, and 5. You can see 7 here that again these are increments. This is the 8 mean one hour and mean peak increments that we 9 observed, and they are quite substantial with the mean 10 peak increment being up to 11,095 cells per 11 microliter. Interestingly, the peak increment 12 typically occurred about 8 to 12 hours after the 13 granulocyte transfusion, not at one hour as you might 14 intuitively expect. 15 This table demonstrates for you that the 16 increments were sustained with the time after the 17 transfusion in which the mean ANC was above baseline 18 being at least 25 hours. So we ask the question, 19 where do these granulocytes go once transfused. So we 20 took samples of the granulocyte components collected 21 on day +5. We labeled them with indium and then 22 transfused them into the allogeneic bone marrow 23 transplant patient and we monitored scintographic 24 scans serially. This is a scan obtained within four 25 hours of granulocyte transfusion demonstrating very 163 1 intense uptake in the lungs. 24 hours later, you can 2 see that the lung uptake has essentially dissipated, 3 with intense uptake now in the liver and spleen and 4 the marrow. So it looks like these granulocytes 5 initially, once infused, immediately track for the 6 most part to the lungs and then are probably gradually 7 liberated over many hours resulting in that peak ANC 8 that I mentioned to you earlier. 9 We also wanted to know, based on these 10 kinds of assessments, were these granulocytes 11 functional, that is, do they localize to sites of 12 inflammation after transfusion. Again, this is 13 another indium scan of G-CSF mobilized and HLA matched 14 granulocytes. This is a scan obtained within four 15 hours of infusion demonstrating predominantly lung 16 uptake and spleen. This is a scan obtained 24 hours 17 later. Again, the lung uptake has dissipated quite a 18 bit. But now you begin to see an area of uptake here 19 in the cecum and ascending colon. 48 hours later, 20 this is a very intense area of uptake now, as you can 21 see outlining the ascending colon and cecum. This is 22 a patient who had diarrhea and colitis after their 23 preparative regime. We have similar scans 24 demonstrating this kind of uptake in the mouth in 25 patients with mucositis. So these kinds of studies 164 1 along with studies from David Dale and Tom Price's 2 group would support that these are functional 3 granulocytes. 4 We have just completed a fairly large 5 study of granulocyte transfusions in the allogeneic 6 peripheral blood stem cell setting. The reason we 7 chose to move from bone marrow to peripheral blood was 8 really a practical one. In allogeneic bone marrow 9 transplantation, the duration of neutropenia is 10 substantially longer. And since we are using a single 11 granulocyte donor, who is the stem cell donor, it is 12 very difficult to expect them to do 8 collections over 13 two weeks or three weeks. So from a practical 14 perspective, it was better to pursue the allogeneic 15 peripheral blood stem cell setting where the duration 16 of neutropenia was about a week shorter. 17 So in this study, we accrued concurrently 18 two cohorts, cohort A and cohort B. Cohort A received 19 prophylactic granulocyte transfusions on days 3 and 6, 20 and these were components that were collected from the 21 stem cell donor who received G-CSF to mobilize these 22 granulocytes. Again, they were obviously HLA matched. 23 Cohort B did not receive granulocyte transfusions. 24 This was a biologic randomization determined by the 25 availability or not of an ABO compatible, HLA matched 165 1 related donor. And the donor, as I mentioned, of the 2 stem cells was also the donor of the granulocytes. I 3 have data to share with you on 13 of cohort A. We have 4 19 actually in cohort A. And 11 of the 51 we have in 5 cohort B. The patients in this study received the 6 same preparative regimen, the same graft versus host 7 disease prophylaxis and the same supportive care, 8 including the way we initiated and stopped 9 antibiotics. 10 This demonstrates for you that in cohort 11 A that received granulocyte transfusions, we did see 12 significant and sustained increments in the ANC after 13 the granulocyte transfusions given on days 3 and 6, as 14 you can see here. 15 We then looked at the ANC one day 16 following granulocyte transfusions in cohort A and on 17 the same day in cohort B, and we found that the 18 absolute ANC -- the lowest ANC we observed on that day 19 was significantly higher in cohort A that received 20 granulocyte transfusions. We then tried to determine 21 an efficacy with this preventive therapy, and we have 22 looked at numbers of days of IV antibiotics from day 23 zero until neutrophil engraftment and have observed 24 that the numbers of days of IV antibiotics in that 25 interval is about 4.5 days less in the cohort that 166 1 received granulocyte transfusions. One might argue 2 that the proportion of patients in cohort B receiving 3 antibiotics on day zero was greater, and to some 4 extent that was true because we looked at the numbers 5 of days of IV antibiotics in the interval from 6 starting the preparative regimen to day -1, and found 7 that cohort B had received more antibiotics, but only 8 by two days, which in my analysis wouldn't account for 9 this 4.5 day difference we see after day zero. 10 We do not -- although the absolute numbers 11 of febrile days was greater in cohort B compared to 12 cohort A, this was not statistically significantly 13 different. That may be more difficult to prove in 14 this limited number of patients given the relatively 15 small numbers of febrile days you see here. So to my 16 knowledge, this is the first -- this is preliminary 17 data, but it is the first data that I am aware of that 18 demonstrates a potential clinical benefit of giving G- 19 CSF mobilized HLA matched granulocyte transfusions to 20 such patients. 21 With that in mind, we were interested in 22 knowing the potential importance of leukocyte 23 compatibility with G-CSF mobilized granulocytes. I 24 think one could arguably ask, since we are able to 25 collect granulocytes with such a huge cell dose today 167 1 using growth factors such as G-CSF, can that massive 2 cell dose overcome the problem of leukocyte 3 incompatibility? So that was the question of this 4 trial. 5 This is a study again with the objective 6 being to determine the kinetics of the recipient ANC 7 after transfusion of prophylactic granulocyte 8 components that are mobilized with G-CSF, but these 9 components may either be leukocyte compatible or 10 incompatible. The donor was a first degree relative 11 of the recipient, received G-CSF daily -- or actually, 12 I am sorry, four doses on transplant days 1, 3, 5, and 13 7, and the dose of G-CSF we chose in this trial was 10 14 mcg per kilogram. Granulocyte collections were 15 performed on the evening of day one and the mornings 16 of day 4, 6, and 8. The recipients were all 17 autologous stem cell transplant patients who had 18 reasonably adequate stem cell products as defined here 19 based on CD34 numbers. The recipients received G-CSF 20 daily from day zero until neutrophil engraftment, and 21 then received fresh granulocyte transfusions early 22 morning of day 2 and in the afternoons of day 4, 6, 23 and 8. And then we carefully documented the kinetics 24 of the ANC of the recipient of these granulocyte 25 transfusions as you can see here. 168 1 Pre-study, we assessed leukocyte 2 compatibility between donor and recipient based on HLA 3 and B typing of both, based on an leukoagglutination 4 cross match and measures of HLA antibodies using 5 lymphocytotoxicity assay. We have accrued 25 6 donor/recipient pairs in this study, which we have 7 closed. I only have data at the moment on six of 8 these people, which I have shown for you here. 9 This is data showing the granulocyte 10 component cell dose times 1010 for each day of 11 transplant. This is the average cell dose. This shows 12 you the increment in the average ANC at post-infusion 13 hours 1, 4, 8, 12, 24, 36, and 48. I think that the 14 increments that we observed on day +1 were reasonably 15 good, but the increments that we observed on 16 transplant days 4, 6, and 8 in my opinion were 17 inferior to our prior results. And that occurred in 18 spite of transfusing larger numbers of granulocytes. 19 I show for you here the results of the 20 leukocyte compatibility test. Five of six of the 21 donor/recipient pairs were not HLA and B matched. 22 Lymphocytotoxic antibodies were detectable in four of 23 six patients. And the leukoagglutination cross match 24 was zero or negative in all six pairs. So this 25 preliminary data suggests to me that leukocyte 169 1 compatibility may be an important determinant of 2 neutrophil increments after transfusion of G-CSF 3 mobilized granulocyte products, and it certainly at 4 least provides an important clue that we really need 5 to look at this area more carefully. 6 This is just a table comparing this data, 7 the current data that I just discussed of autologous 8 transplant patients who received predominantly 9 leukocyte incompatible granulocytes. The component 10 cell dose was 8.6 times 1010, and the maximum mean 11 peak ANC increment occurring was 796, and that was on 12 only the day +2 transfusion. As you recall, the 13 increments after that were substantially lower. If 14 you compare that to our data that we published last 15 year in Transfusion, these are allogeneic bone marrow 16 transplant patients who received HLA matched 17 granulocytes, again mobilized with G-CSF. In spite of 18 transfusing a lower component cell dose than this 19 current data, the ANC increments were substantially 20 greater. 21 This brings to me the importance of 22 considering doing red cell reduction of granulocyte 23 components. Why? When selecting only leukocyte 24 compatible donors, the requirement of ABO 25 compatibility reduces the pool of potential 170 1 granulocyte donors. And the data from the last study 2 suggested that we probably need to select leukocyte 3 compatible donors. The objective of this study was to 4 decrease the component of packed red cell volume to 5 under 5 mls, a guideline established by the AABB, 6 which we think will reasonably insure against 7 hemolytic transfusion reaction in recipients of 8 granulocyte components that are collected from ABO 9 incompatible donors. And this is data we recently 10 published in the Journal of Clinical Apheresis. 11 The trial design involved apheresis of 12 granulocytes with hetastarch, and then after 13 collection we performed gravity sedimentation of the 14 component for 60 minutes, and then we transferred the 15 red cell poor fraction to a sterile docked transfer 16 bag utilizing a plasma expressor. The residual red 17 cells were retained in the collection bag and were 18 defined as the red cell rich fraction. And then we 19 documented cell numbers and packed red cell volumes 20 with each component or fraction. This data shows you 21 that without manipulation of the component, the 22 average granulocyte component contains a packed red 23 cell volume of about 25 mls, as Susan Leitman had 24 discussed. 25 However, after component manipulation 171 1 using gravity sedimentation ex vivo, the red cell poor 2 fraction did contain substantially lower packed red 3 cell volume, 6.3 ml on average. We observed that in 4 these manipulated components, 40 percent of the red 5 cell poor fractions contained under 5 ml packed red 6 cell volume. Unfortunately, as I point out here, 7 there is some cell loss. 20 percent of the 8 granulocytes are lost to the red cell rich fraction 9 using this technique. 10 So ex vivo, a hetastarch sedimentation as 11 we performed in this study did reduce the number of 12 red cells from the granulocyte components, but most 13 red cell poor fractions still contained more than 5 ml 14 packed red cell volume and thus were not acceptable 15 for transfusion into ABO incompatible recipients. We 16 are currently looking at extending the duration of 17 sedimentation to 90 and to 120 minutes. At 120 18 minutes, we have been able to uniformly reduce the 19 packed red cell volume to under 5 ml with this 20 technique. 21 So I have thrown a slide up here for you 22 to suggest an ideal trial design for prophylactic 23 granulocyte transfusion support of transplant 24 recipients. The principle being the following: daily 25 prophylactic granulocyte transfusions from onset of 172 1 neutropenia to recovery from neutropenia. We would 2 like to choose leukocyte compatible donors receiving 3 a large dose of G-CSF. If you recall an earlier 4 slide, this dose of G-CSF resulted in granulocyte 5 components with 20 times 1010 granulocytes. So there 6 is a reason for that, which I will get to. And then 7 the donors undergo granulocyte collections on 8 transplant days 2, 4, 6, and 8. The apheresis uses 9 the higher IO setting of 35. Hetastarch is the red 10 cell sedimenting agent. And then conceptually, from 11 a practical perspective, it would be nice to split 12 each granulocyte component, freshly transfusing half 13 and trying to store overnight the other half, 14 realizing that each component will contain roughly 10 15 times 1010 granulocytes. It is probably feasible to 16 do that. The recipients would include transplant 17 patients who were receiving adequate stem cell 18 products, again as I have defined here, and also 19 receive post-transplant G-CSF. And it would be nice 20 to have some measure of choosing non-alloimmunized 21 patients. 22 To conclude, first of all G-CSF improves 23 the cell yield of granulocyte components collected 24 from normal donors. Secondly, in allogeneic bone 25 marrow transplant patients, transfusion of G-CSF 173 1 mobilized HLA matched prophylactic granulocyte 2 components resulted in significant and sustained 3 increments in the ANC, localized to sites of 4 inflammation for up to two days post-transfusion, did 5 not cause febrile reactions -- and I didn't mention 6 this, but in our HLA matched granulocyte donor 7 scenario, we have not observed febrile transfusion 8 reactions. These components also resulted in 9 significant increments in the platelet counts and 10 reduced platelet transfusion requirements in this 11 cohort of patients. And as many of you know, these 12 granulocyte components contain 2 to 4 times 1011 13 platelets, which is equivalent to almost a unit of 14 single donor platelets. In our experience in this 15 patient cohort, we have been able to reduce platelet 16 transfusion requirements in half with granulocyte 17 transfusions. Third, antibiotic utilization was 18 reduced in allogeneic PBSC transplant patients 19 transfused with G-CSF mobilized HLA matched 20 prophylactic granulocyte components on transplant days 21 3 and 6. Fourth, the preliminary data suggests that 22 leukocyte compatibility was an important determinant 23 of ANC increments after transfusion of G-CSF mobilized 24 granulocyte components. And fifth, we really don't 25 know the optimal component cell dose and we really 174 1 don't know the optimal frequency of transfusion of 2 these products today. I think we need to learn that. 3 And finally, there is preliminary data 4 from Phase II studies suggesting potential clinical 5 efficacy of G-CSF mobilized HLA matched prophylactic 6 granulocyte transfusions. I would suggest that we 7 give some consideration for pursuing a Phase III trial 8 at some point to really test this hypothesis. 9 I would like to recognize all of these 10 individuals at Washington University for supporting 11 these trials. I would also like to recognize Barnes 12 Jewish Hospital, which also provided some financial 13 support for these studies. Amgen has also been very 14 kind in helping in doing these studies. And I would 15 like to also recognize Gary Spitzer, who provided for 16 me the initial encouragement to pursue these clinical 17 trials back in 1992 and 1993. 18 Liana asked that I comment on what areas 19 that I would recommend that we would pursue in terms 20 of support of research in this area. I think that we 21 really need to understand better the importance of 22 leukocyte compatibility and incompatibility with G-CSF 23 mobilized granulocyte transfusions. So I think that 24 we really need to pursue that issue better and we need 25 to define that issue better before we pursue a Phase 175 1 III trial. Secondly, the cell dose is really unclear. 2 What is the cell dose we should be using in a Phase 3 III trial? I don't think we understand that issue. 4 How often should we give these granulocyte components. 5 I think that we need to know answers to these 6 questions before we do a Phase III trial. Red cell 7 reduction may be an important area to pursue in 8 research. Indeed if leukocyte incompatibility 9 adversely affects outcomes, red cell reduction will be 10 an important practical issue in order to expand the 11 available donors that we could choose from. And 12 finally, granulocyte storage. Again, from a practical 13 perspective, we all would wish that we could store 14 granulocytes. It is just, I think, an area that is 15 probably worthy of considering. So I will just close 16 with that point. Thank you. 17 DR. STRONCEK: Thank you for that very 18 clear and insightful presentation. He ended with the 19 right thing to talk about studying next storage. Tom 20 Lane will now discuss storage considerations of 21 granulocytes. Dr. Lane is Professor of Pathology at 22 the University of California, San Diego. He is the 23 Medical Director of their transfusion service and 24 their stem cell laboratory. 25 DR. LANE: Thank you, David. And thank 176 1 you, Liana, for inviting me to this conference. I 2 have learned a lot and it has been very interesting. 3 It is always a pleasure to talk about granulocyte 4 storage. I will, of course, indicate that many of the 5 people in this room contributed to the studies that I 6 am going to summarize. Give me the first slide, 7 please. Were I to actually go through all the data -- 8 the first two slides are actually Dr. Harvath's. So 9 if you will move on to the third slide in that 10 carousel, that starts mine. 11 So were I to actually go through all the 12 data, we would be here all afternoon. So I am going 13 to summarize some data that has been collected over 14 the years regarding granulocyte storage. I think the 15 previous speaker has already answered this question 16 for us, why should we store granulocytes or why should 17 we know about the storage of granulocytes? And the 18 answer to that is for one thing, as Jeff McCullough 19 has said for years, granulocytes are inevitably stored 20 for at least some period of time prior to transfusion. 21 This relates to variables such as off-site harvesting, 22 the testing requirements as Dr. Price talked about, 23 who basically summarized my entire talk by saying that 24 there are some defects associated with granulocyte 25 storage, transportation issues, and patient 177 1 considerations. The patient may be receiving other 2 forms of therapy that at least some people may 3 consider incompatible with getting granulocytes at the 4 same time, such as amphotericin. Whether or not that 5 does make a difference. Or the patient may just not 6 be able to get the granulocytes when they are ready in 7 the transfusion service. And, of course, there are 8 those of us who actually still do research on 9 neutrophil function and storage becomes an important 10 part of that. 11 But certainly for a clinical trial in 12 which it may be of interest to store one portion of a 13 granulocyte preparation, obviously you need to know 14 whether or not they work. So I thought I would 15 summarize this as others have, and I will go through 16 this quickly. What the critical granulocyte functions 17 are. Obviously, granulocytes need to stay in 18 circulation for a period of time. They need to be 19 capable of a certain amount of adherence but not too 20 much adherence until they encounter an activated 21 endothelium. They need to recognize the activated 22 endothelium, a chemotactic gradient as indicated here 23 by these little dots. They need to then migrate 24 through a chemotactic gradient towards this happy 25 bacterium that unbeknownst to him has been opsonized 178 1 by antibody and complement. They then need to engulf 2 the bacterium and finally kill it. So these are 3 critical granulocyte functions that must be 4 accomplished. 5 Well, what do we know about the clinical 6 efficacy of stored granulocytes. In fact, there are 7 no studies out there that will attest to the clinical 8 efficacy of storage of granulocytes. If you look at 9 the seven studies that people refer to, those that 10 were reviewed by Dr. Strauss, and look in the methods 11 sections regarding storage, you will find that the 12 granulocytes were transfused either immediately or 13 within 4 hours or it is not specified. So at this 14 time, there are no clinical studies that will attest 15 to the efficacy of stored granulocytes. So that has 16 left us then with looking at surrogate markers of the 17 efficacy of the function of stored granulocytes. 18 These can be easily broken into two general 19 categories, in vivo studies and ex vivo studies. The 20 former include granulocyte recovery, kinetics and 21 survival, and distribution, either measured by 22 isotopic techniques, migration into the buccal cavity 23 as Dr. Price mentioned, or a more classical skin 24 window chamber studies of a variety of different 25 types. Or what have been measured more frequently 179 1 because they are simply easier and don't involve human 2 studies or human manipulation, I should say, are 3 looking at the cell numbers after storage, adherent 4 function, chemotaxis function, and the other functions 5 that I mentioned that are critical for granulocytes to 6 do what they are supposed to do, which is to kill 7 invading microorganisms. 8 Well, if you look at all these potential 9 surrogate markers, which ones are really relevant to 10 clinical efficacy? This is actually a fairly 11 difficult question to answer. Let me back up for a 12 moment and say I think clearly the number of 13 granulocytes that need to circulate in vivo has been 14 studied. There are classical studies by Bodie and co- 15 workers in the late 1960's suggested that we need in 16 the range of about 500 per microliter. Now that may 17 be a facile measurement of something more complicated 18 such as the total granulocyte storage pool, but there 19 is at least evidence to suggest that if the 20 circulating granulocyte level is above 500 that people 21 are less susceptible to infection if they are 22 functioning normally. 23 So apart from the cell number, what else 24 do we know about how ex vivo function relates to the 25 susceptibility for infection? At least -- I don't 180 1 think you can come to any firm conclusions, but you 2 can get clues from, if you will, experiments of 3 nature. The two that I have listed here are leukocyte 4 adhesion protein deficiency of the beta 2 integrins 5 and chronic granulomatous disease. This defect 6 relates primarily to the adherence of neutrophils, 7 LAD, and CDG, as you all know, relates to the failure 8 of granulocytes to generate toxic oxygen radicals. 9 Now if you then separate patients with LAD 10 into severe, moderate and mild and then look at 11 various functions that are correlated with these 12 clinical defects, you can come up with at least some 13 generalities. And likewise with CDG. And rather than 14 going through all this, I have summarized that on the 15 next slide. All this data put together suggests that 16 if skin window migration, primarily generated through 17 looking at LAD deficient patients, is less than 80 18 percent of normal, this is associated with at least 19 mild defects in resistance to infection. Likewise, in 20 vitro chemotaxis defects, less than 70 percent of what 21 passes for normal -- and anyone who has done this 22 knows that this can be quite variable -- have been 23 associated with infection. Adherence less than 50 24 percent has been associated with infection. 25 Phagocytic activity of less than 40 percent, microbial 181 1 killing of less than 25 percent, and oxygen radicals, 2 surprisingly you need very little in the way of oxygen 3 radical generation to sustain normal microbial killing 4 function. 5 So this gives us a clue then as to what 6 kinds of surrogate markers are going to be useful to 7 study when you look at granulocyte storage. Again, 8 commenting on the fact that while perhaps the best 9 studies are those related to in vivo function, that is 10 to say do the cells migrate or localize to sites of 11 infection, recognizing that those are difficult to do, 12 most people are going to look at least initially at ex 13 vivo functions. And then the next question you want 14 to ask is well looking at ex vivo functions based on 15 all of the relevant past experience, how do these 16 functions fall out? And nearly all studies agree that 17 chemotaxis, the migration of neutrophils, is the 18 single most sensitive function during storage. It is 19 the function that seems to have the earliest defects 20 and is most sensitive to granulocyte manipulations in 21 storage. And that seems to be followed by changes in 22 adherence and microbial killing, followed in turn by 23 changes in phagocytosis and oxygen radical generation. 24 So the next studies that I am going to summarize will 25 focus to a large extent on chemotactic function. 182 1 Because arguably if the cells won't migrate, then it 2 doesn't matter if they have 70 percent or 100 percent 3 oxygen radical activity. 4 So what are the important factors we need 5 to look at when looking at stored neutrophils? What 6 factors may affect the quality of stored neutrophils. 7 Obviously the donors presumably will have normal 8 neutrophil function and that is usually ascertained by 9 whether or not they have a history of infection. 10 Otherwise, they wouldn't be normal donors. Obviously 11 related to the use of G-CSF, donor treatment with G- 12 CSF or steroids. Collection techniques as has been 13 summarized by Dr. Dale. There are differences in the 14 efficacy of granulocytes based on whether they are 15 collected using centrifugal techniques or filtration 16 leukopheresis techniques. The concentration of the 17 neutrophils in the component itself -- and this is 18 perhaps the single most important thing I am going to 19 say today. We are going to have to be careful in the 20 modern age regarding the concentration of neutrophils 21 in the bag as regards their storage. And others have 22 shown that the concentration of platelets too may 23 affect the storagability of granulocytes. 24 Physical parameters of storage including 25 temperature, agitation, the type of container, and 183 1 metabolic parameters such as the anticoagulant used, 2 the pH of the medium, which relates of course to the 3 concentration of neutrophils, and the amount of 4 glucose in the protein have been shown to affect 5 granulocyte storage. And finally, the presumed 6 effectiveness or lack thereof of preservatives such as 7 growth factors. A letter to the editor regarding the 8 use of gels to protect granulocyte function. 9 I may, in the interest of time, skip this 10 slide since Dr. Dale has already reviewed the fact 11 that cells collected by filtration leukopheresis can 12 be shown up front to have diminished function and this 13 simply illustrates that while cells collected by 14 centrifugation leukopheresis at least after collection 15 are normal -- and by the way, everything that I am 16 going to talk about will be regarding donors not 17 stimulated by G-CSF. So while cells collected by 18 centrifugation leukopheresis are relatively normal, 19 those by filtration are not and survive in storage 20 very poorly. 21 What about storage conditions and how this 22 relates to function of stored neutrophils? Well, very 23 early studies by McCullough and co-workers have shown 24 that different kinds of bags may affect storage. PVC 25 bags appear to be the best of those looked at, at 184 1 least at that time, and are certainly better than 2 storing neutrophils in tubes. More recently, TOTM, 3 and I can't remember right now what this stands for, 4 or the CLX bags have been shown to provide improved 5 granulocyte storage compared to other types. 6 McCullough and co-workers looked at a variety of 7 different kinds of anticoagulants and found some 8 differences which I will show in a moment. 9 Sedimenting agents have been largely shown not to 10 affect granulocyte storage. I will talk about 11 temperature, agitation, neutrophil count. Platelets 12 I have already mentioned. Glasser and colleagues 13 showed some years ago that platelets -- the presence 14 of platelets diminishes granulocyte function after 15 storage as a continuous variable, largely through the 16 diminution, they found, of glucose content. Likewise, 17 glucose content of the storage medium somewhere 18 between 50 and 1,000 mg per deciliter appears to be 19 optimal. Glasser and colleagues also showed that 20 there is a requirement for optimal storage for protein 21 and found that 1 percent albumin or plasma were 22 equally effective, but not IgG. Other studies have 23 shown a pH optimum in the range of 7 to 7.5 in 24 preserving chemotactic function. I am going to just 25 show you a little bit of that data now. 185 1 This is a study by McCullough and co- 2 workers published quite a long time ago which showed 3 for granulocytes stored at 4 degrees Centigrade in 4 those days a slight advantage at 24 hours to cells 5 stored in ACD or CPD anticoagulant compared to heparin 6 or ion exchange. Not a great deal of advantage but a 7 little bit. 8 For pH, McCullough and co-workers also 9 looked at this as have others. Shown here is the 10 initial pH of a storage medium now 24 hours at room 11 temperature and the resultant ATP. This is 12 chemiluminescence, the measure of toxic oxygen radical 13 generation and chemotaxis using, I believe, a Boydian 14 chamber technique. They found, focusing on 15 chemotactic function, that a pH range between 7 and 16 7.5 was optimal. Either side of the chemiluminescence 17 dropped off radically. This illustrates that 18 chemiluminescence or the ability to generate toxic 19 oxygen radicals in response to a phagocytic stimulus 20 was somewhat more resistant to changes in pH and ATP 21 showed decrements in pH at anything other than about 22 7.5. 23 This slide summarizes two studies relating 24 to the effect of agitation. It is interesting that 25 there are differences. The first one by McCullough 186 1 back in 1978 suggested that granulocytes stored at 2 room temperature for 24 or 48 hours had a greater 3 defect in chemotaxis if they were stored agitated, and 4 Mary Clay was kind enough to remind me that this was 5 using horizontal agitation, and these changes were 6 more pronounced after 48 hours of storage. So there is 7 an advantage to not agitating neutrophils according to 8 these studies. But approximately 9 years later, some 9 Japanese workers published in Transfusion just the 10 opposite result using a somewhat different bag but the 11 same kind of agitation, that is to say horizontal. 12 They found chemotactic function better preserved in 13 cells that had been agitated as opposed to left 14 stationary. Now it is difficult -- looking at these 15 two studies, it is difficult to make any sense out of 16 this since most of the other factors relating to the 17 granulocyte storage were relatively equal. What is of 18 interest is that in the later study, the cells that 19 were stored in a stationary fashion had a very, very 20 marked defect in chemotaxis after 24 hours, much 21 greater than most other workers have found, and this 22 leads one to believe possibly that there may have been 23 something else going on in this study which didn't 24 permit these cells to function as well. I just point 25 this out to indicate that maybe this is something we 187 1 need to look at again in view of this controversial 2 data. 3 A great deal of work has focused on the 4 temperature at which we ought to store granulocytes, 5 and this is a study again by one of the leaders in 6 this field, McCullough and co-workers, looking at room 7 temperature versus 6 degree storage for 8 or 24 hours 8 and looking at in vivo recovery and survival and 9 localization, this percent skin window migration. To 10 summarize this, these investigators found after only 11 8 hours of storage at either room temperature or 6 12 degrees, a significant benefit to room temperature 13 storage in terms of overall granulocyte recovery, and 14 this is percent recovery, compared with 6 degrees. 15 And likewise after 24 hours of storage, again an 16 advantage to room temperature over 6 degrees storage. 17 The half-life measurements of granulocytes were a 18 little bit more difficult to interpret, but again 19 suggested an advantage to room temperature storage. 20 But these investigators found, again, a marked 21 advantage to room temperature storage compared to 6 22 degrees in looking at migration of the cells into skin 23 windows. 24 Some studies we did later suggested a 25 possibility for the changes in granulocyte recovery 188 1 comparing 6 degrees to room temperature storage. We 2 looked at granulocyte adherence to endothelial 3 monolayers after 24 and 48 hour storage of 4 granulocytes at these two temperatures and found that 5 cells stored at 6 degrees were somewhat hyperadherent, 6 whereas those stored at room temperature for this 7 period of time had relatively normal adherence 8 function. 9 We also, as have others, found a 10 significant benefit to room temperature storage 11 compared to 6 degrees in terms of chemotactic function 12 and this is distance migrated shown on the ordinate 13 scale at 24 and 48 hours, room temperature versus 6 14 degrees, and this is random migration. 15 So this slide then summarizes a number of 16 studies looking at room temperature versus 4 degrees 17 Centigrade storage in granulocytes. Again, 18 highlighting the differences, there are differences in 19 adhesive function of cells stored at room temperature 20 versus 4 to 5 degrees, improvements in chemotactic 21 function or I should say less of a decrement in 22 chemotactic function and less of a decrement in the 23 recovery after transfusion, and less of a decrement in 24 skin window migration. So these and other studies led 25 most people to the conclusion that room temperature 189 1 storage for these unstimulated donors was superior to 2 6 degrees. 3 Someone asked the question today about the 4 effects of irradiation. I have summarized here I 5 think it is 8 different studies that have over the 6 years looked at the effect of irradiation on 7 granulocyte function. And I think we can summarize 8 these by saying that over the range of irradiation, 9 these are in gray, used to prevent GVHD, there really 10 is no consistent effect on granulocyte function, 11 either in cells that are collected freshly, as most of 12 these were, or after storage for 24 or 48 hours here 13 in two studies in which a wide variety of functions 14 were observed. There is only one study that suggested 15 that 50 Gray irradiation might decrease the nitroblue 16 tetrozolium generation in granulocytes. And this is 17 using a semi-quantitative technique. So I think the 18 great weight of evidence suggests that irradiation 19 does not affect these cells. Again, this hasn't been 20 looked at in G-CSF stimulated donors, at least not 21 published to my knowledge. 22 So using the "best of current techniques", 23 what can we expect for granulocytes that have been 24 stored for 24 or 48 hours. That means room 25 temperature storage, not agitated. Well, we can 190 1 expect about 50 percent recovery, which is not very 2 different from normal -- I am sorry, 50 percent 3 decrement in recovery compared with normal, which is 4 decreased, whereas the survival of the cells will be 5 close to normal. In vitro, that is to say in the bag, 6 we can expect to recover most of the cells. Most of 7 them will be there, between 99 and 88 percent. We can 8 expect up to 20 percent decrement in in vitro 9 chemotaxis and perhaps 10 percent decrements in 10 microbial killing. So all of this data suggests then 11 that these cells ought to function relatively normally 12 once they are transfused. At 48 hours of storage that 13 may not be the case. 14 The next and final thing that I want to 15 talk to, and I am going to try not to go too much over 16 time here, is I think important to the modern 17 situation. Because we did some studies some years ago 18 looking at the effect of granulocyte concentration, 19 that is to say the number of granulocytes or their 20 concentration in the bag, on the subsequent function 21 of those cells. We looked at three different cell 22 concentrations, 2, 5, and 7 times 107 per ml. We 23 stored the cells at room temperature for 24 or 48 24 hours in autologous plasma unagitated at room 25 temperature. We found progressive decrements in ATP, 191 1 in glucose, and marked changes in the pH of the 2 surrounding medium. We found that -- well, let me 3 summarize this. We found basically that you could 4 prevent the changes in pH by adding 15 millimolar 5 bicarbonate. Let's start with this box over here or 6 this panel. Shown here are cells stored at -- let's 7 see, it doesn't say it on here -- at 8 times 107 per 8 ml, either in the presence of no additives, of glucose 9 alone, of bicarbonate, or glucose and bicarbonate. 10 And you can show that bicarbonate preserves the 11 starting pH of these cells. Glucose will preserve the 12 glucose content, but as shown here will not preserve 13 the pH. And that likewise cells stored in the 14 presence of bicarbonate, either with or without 15 glucose, maintained their content of ATP, at least at 16 relatively normal amounts. 17 Now how does this relate to their 18 function. In another study, we looked once again at 19 ATP content and the pH of cells stored at 2 or 8 times 20 107 neutrophils per ml, again in the presence of 21 bicarbonate or without it. I think if we focus on 22 this panel, these are cells stored at 8 times 107 per 23 ml in the absence of bicarbonate. You can see that 24 their chemotactic function now using a Boydian chamber 25 technique drops off rather remarkably. But in the 192 1 presence of bicarbonate -- unfortunately I don't have 2 the key here, so I am having to remember these old 3 studies -- in the presence of bicarbonate, you can 4 preserve chemotactic function at least for 24 hours, 5 even at cells stored at this high a concentration. 6 Likewise, you can preserve pH and ATP as I showed 7 before. So the point of all this is that in the cells 8 that are being generated and the granulocyte 9 concentrates that are being generated today, if you 10 translate the doses that you are giving in the range 11 of 40 billion or so versus the volumes into which they 12 are being collected, you are exceeding even what we 13 studied here by two or three-fold in terms of overall 14 cell concentrations. So if we found marked defects in 15 pH maintenance and ATP maintenance and chemotaxis in 16 cells stored at 8 times 107, then cells stored at 20 17 times 107 are probably going to be much worse off than 18 we have here. 19 So I would hasten to add that we really do 20 need to study how these cells will function after 21 storage. And perhaps I would recommend then that if 22 one is going to try to store these cells, that you 23 need to try to maintain the pH of the surrounding 24 medium. That seemed to be accomplished by using 15 25 millimolar bicarbonate, at least in these studies. 193 1 But I would not speculate whether that would be enough 2 in cells stored at two or three times this 3 concentration. I think that is an area that really 4 needs to be looked at. And as an abstract to be 5 presented later on shows, in fact the changes we found 6 in pH here certainly are found in cells in stimulated 7 donors. 8 So finally, I want to summarize here what 9 everybody already knows. Over the past several years 10 there have been a number of studies that suggest that 11 certain cytokines can, in fact, prolong the storage 12 life of neutrophils. Most of these studies were 13 performed in cells stored at 37 degrees in tissue 14 culture flasks or plates and looked only at so-called 15 viability or trypan blue dye exclusion. So all one 16 knew was that these cells were surviving longer. But 17 more recently, Rex and co-workers have published in 18 Transfusion some more interesting results. Again, 19 these granulocytes were stored in culture dishes at 37 20 degrees, but they looked at some more relevant 21 functions, in this case granulocyte chemotaxis to 22 FMLP, and found that when you store cells in this 23 fashion for 20 hours, there is a marked decrement in 24 chemotaxis that is preserved or prevented by the 25 addition of G-CSF and gamma interferon. They also 194 1 looked at superoxide anion generation at fresh and 2 after storage, and again found a decrement with 3 storage which was again prevented by G-CSF and gamma 4 interferon. They looked at bacterial killing and 5 found perhaps some improvement with G-CSF and gamma 6 interferon in this decrement. Now this is percent 7 surviving Candida. This is not really a convincing 8 difference. 9 And then they also provided, as have 10 others, an explanation for why these decrements in 11 function were prevented by G-CSF. This is a slightly 12 different organization here. We are looking at cells 13 stored at 20 hours at 37 degrees as a percent of 14 control, either unirradiated or irradiated with in 15 this case I guess it is 5 Grays -- maybe that should 16 be 50. They found, looking at apoptosis, that cells 17 stored in this fashion with or without irradiation 18 underwent marked apoptotic changes, as others have 19 reported, but that the addition of gamma interferon 20 and G-CSF with or without irradiation prevented 21 apoptotic change. 22 Now I used to think of granulocytes as 23 being end stage cells that didn't do much in the way 24 of protein synthesis and I really didn't understand 25 these findings for quite some years, but I have been 195 1 educated. Granulocytes are, in fact, capable of new 2 protein synthesis and new messenger RNA synthesis, and 3 I have just summarized some recent studies here 4 showing that with these various stimuli, you can in 5 fact get new protein synthesis, as indicated here, and 6 this is accompanied by gene activation. And in fact 7 in I guess it is the August issue of Blood, there is 8 a new publication indicating that if you look at all 9 the different species of messenger RNA that are made 10 by granulocytes, there are over 700 different species 11 of messenger RNA that are actively generated by 12 granulocytes. So these aren't cells that are 13 incapable of protein synthesis. And that perhaps 14 explains how it is that G-CSF and some of these other 15 stimuli can modulate the apoptosis of neutrophils. So 16 I thought I would end here by showing that a variety 17 of different cells can modulate or delay apoptosis -- 18 interleukans 2415, GM-CSF, G-CSF of course, TNF alpha 19 after prolonged incubation, glucocorticoids, et cetera 20 -- while other functions, stresses, and agents may 21 accelerate granulocyte apoptosis. Well, after being 22 interested in granulocyte storage for 20 years, maybe 23 this is the holy grail. Maybe delaying apoptosis in 24 these cells will permit them to be stored for a longer 25 time. 196 1 But once again, I guess if I can say 2 anything important here this afternoon, it would be 3 that it doesn't matter whether the cells are apoptotic 4 or not if their pH is 5.5 and they are dead. So I 5 think with the modern collection techniques, we need 6 to be mindful of the concentration of granulocytes in 7 the medium and do something about that. And then in 8 the future, it may be possible to prolong or permit 9 longer storage using some of these agents. And in 10 response to Liana's question, I am going to sound like 11 a broken record. I think we need a trial of the 12 efficacy of granulocytes stimulated by G-CSF in donors 13 before we can really know whether they are working. 14 I would be -- I think we need to study granulocyte 15 storage, but I would be hesitant to store granulocytes 16 in such a trial. I think we have no current knowledge 17 that stored granulocytes, at least beyond 6 to 8 18 hours, really work, and I would be concerned at least 19 in a major trial that it would not be beneficial to 20 the potential efficacy of such a trial to include 21 stored granulocytes. I think that needs to be studied 22 separately. Thank you very much. 23 DR. STRONCEK: Thanks, Tom. It sounds 24 like once granulocytes are proven to be effective, we 25 will be busy for quite a while figuring out the best 197 1 way to store them. The next presentation will be by 2 Conrad Liles, who will talk about in vitro assays 3 predictive of product function. Dr. Liles is an 4 Assistant Professor in the Division of Allergy and 5 Infectious Disease of the Department of Medicine at 6 the University of Washington, Seattle. 7 DR. LILES: The title of my talk is as 8 introduced, and I thank the organizers for inviting me 9 to this workshop. It is in vitro assays predictive of 10 leukopheresis granulocyte product function. It is a 11 little bit of a difficult task because that is what 12 people have been talking about the entire day, but I 13 am going to try to talk about our studies in 14 evaluating leukocytes or granulocytes that are 15 mobilized with G-CSF and then also those granulocytes 16 during storage and storage plus or minus the 17 readdition of G-CSF ex vivo. 18 First of all, I wanted to talk about -- 19 you have seen this slide before, but this is why we 20 chose a regimen of 300 mcg of G-CSF and 8 mg of 21 dexamethasone to stimulate our donors in our 22 granulocyte collections. So we proceeded to use this 23 regimen because it seemed to give the maximal ANC and 24 it seemed to give it within 12 hours. The addition of 25 dexamethasone significantly increased the maximal ANC 198 1 induced by the dose of G-CSF alone. The greatest 2 mobilization was with 600 mcg of G-CSF and 8 mg of 3 dexamethasone. This drug regimen at least overall was 4 relatively well tolerated by the normal volunteers 5 that we used in this study. 6 The protocol involved treatment, as you 7 have seen earlier, with 600 mcg here and collection by 8 leukopheresis. And then the measurements were cell 9 numbers and morphology, immunophenotype analysis by 10 flow cytometry, chemiluminescence, bacteriocidal 11 function, and then blood and tissue kinetics in these 12 cells immediately after collection. The collections 13 were from 5 donors and you can see here that the mean 14 number of cells collected was 77 times 109 cells. 15 That is after a starting neutrophilia of 28,700 in the 16 donors. 17 First of all, we looked at 18 chemiluminescence. This just shows you the luminol- 19 enhanced chemiluminescence activity of the neutrophils 20 in these collected granulocyte fractions. We 21 evaluated baseline, that is, these were -- after 22 stimulation of these donors, we went ahead and looked 23 in their venous blood and looked to see what their 24 chemiluminescence activity was in response to PMA. We 25 also looked at the baseline activity right here prior 199 1 to collecting the actual neutrophils and stimulating 2 the donors, and then we also looked at the 3 leukopheresis product. What we found is that after 4 stimulating with G-CSF and dexamethasone, you actually 5 saw a priming effect of the stimulation procedure, so 6 there was greater chemiluminescence activity or 7 greater oxidative burst potential in response to PMA 8 after giving G-CSF and dexamethasone. But then if you 9 looked at the cells after leukopheresis, there is 10 actually a slight detriment. So the leukopheresis 11 procedure per se actually impairs the subsequent 12 oxidative burst, but still it was greater than just 13 cells obtained prior to the G-CSF stimulation. So our 14 conclusion was that the product here had good 15 respiratory burst activity and would have potential 16 activity in fighting or in having microbicidal 17 activity if retransfused. 18 We did look at bacteriocidal activity and 19 the staphylocidal assay ex vivo. We found that the 20 leukopheresis PMNs that were obtained after G-CSF and 21 dexamethasone stimulation were just as effective as 22 baseline PMNs in terms of killing the Staph aureus. 23 So the bacteriocidal activity appeared to be fine when 24 immediately collected. 25 When we looked at immunophenotype analysis 200 1 of these cells versus baseline cells, you can see that 2 in terms of L-Selectin, there is a slight decrease, 3 about half or 50 percent decrement, in the L-Selectin 4 expression on these cells as compared to baseline 5 PMN's. CD11B was about doubled. CD18 was also 6 doubled. We didn't really see a large induction of 7 CD14 at this dose of G-CSF. And then CD16, which is 8 FCgamma R3, was actually decreased on the 9 leukopheresis cells. CD32 or SCgamma R2 was actually 10 about the same. And then we did see an induction of 11 SCgamma R2 or CD64, about a doubling there. What we 12 concluded from this is that the leukopheresis cells 13 did have a slightly different immunophenotype, but it 14 was a favorable immunophenotype, and one -- given that 15 we had expression of the SCgamma receptors -- one in 16 which we though the cells would be effective in terms 17 of normal host defense function. 18 We then went on to look at the in vivo 19 kinetics of these cells when retransfused. David 20 already mentioned this earlier today. When we 21 retransfused the cells, we saw that they had a 22 prolonged half-life. What is not shown here is that 23 these cells not only circulated with the prolonged 24 half-life, but they could get to inflammatory tissue 25 sites. In other words, we did recover these cells 201 1 from the buccal mucosa when they were relabeled and 2 then retransfused, and they also migrated to skin 3 windows effectively. So that these cells not only 4 migrated, but they could also migrate to potential 5 inflammatory sites. 6 So our results overall show that we could 7 get a good yield of neutrophils from donors stimulated 8 with one dose of G-CSF and dexamethasone, and then the 9 leukopheresis performed 12 hours afterwards. The 10 respiratory burst activity -- I didn't show you all 11 the other respiratory burst activity to different 12 stimuli, but it was more or less normal or at least 13 there was significant activity to the point that the 14 cells would have microbicidal activity. Bacteriocidal 15 activity, at least against Staph aureus, was normal. 16 Immunophenotype showed increased CD11B and CD18 and 17 also an induction of CD64. And the kinetics showed an 18 increased blood half-life, but also the ability of the 19 cells to migrate to tissue sites. So it appeared 20 overall that these cells obtained from individuals 21 after a single dose of G-CSF and dexamethasone 22 appeared to be functional and would be viable 23 candidates in a neutrophil transfusion program. 24 So just to summarize again, G-CSF plus 25 dexamethasone allows much improved neutrophil 202 1 collection and the treatment is relatively well 2 accepted with a few adverse effects. Metabolic and 3 bacteriocidal functions are preserved and the cell 4 half-life is prolonged. And we have concluded that 5 transfusion of these cells to neutropenic patients may 6 be useful. 7 In the second part of this talk, I really 8 want to talk about storage, because that is really the 9 frontier at this point in terms of improving the 10 program. As you have heard, many blood banks would 11 not have the capability of collecting on weekends or 12 routinely on weekends. So the ability to obtain cells 13 and then store them for 24 to 48 hours would greatly 14 facilitate any sort of neutrophil transfusion program. 15 So we wanted to see whether or not with the knowledge 16 that we had concerning apoptosis and other storage 17 variables, whether or not we could come up with a 18 regimen to show effective storage for 24 to 48 hours 19 so that this could be adopted if we were to have 20 neutrophil transfusion programs instituted on a 21 nationwide basis. 22 The factors compromising the clinical 23 utility of granulocyte transfusion therapy have been 24 discussed, but one of the major factors, as was just 25 discussed in the talk preceding mine, is that 203 1 neutrophils rapidly undergo apoptosis during storage 2 in vitro. And as they undergo apoptosis, functional 3 activity declines. Granulocyte products obtained by 4 leukopheresis are currently transfused into the 5 recipient as rapidly as possible without storage. As 6 soon as that granulocyte product is obtained, it is 7 usually shipped and transfused as rapidly as possible 8 without any storage whatsoever. And this certainly 9 would hamper any sort of program to be instituted 10 nationwide. 11 If we could get effective storage of 12 granulocytes, then we could be able to obtain a 13 leukopheresis preparation and a granulocyte 14 preparation on a Friday and then use it through the 15 weekend and then get another donor on Monday, and it 16 would greatly facilitate the ability to maintain an 17 absolute neutrophil count in a recipient through the 18 weekend without undue stress in a blood bank program. 19 The factors that are known to decrease 20 neutrophil apoptosis during storage ex vivo include 21 anaerobic environment. We actually looked at this and 22 if you really do store isolated neutrophils, that is 23 neutrophils that you have obtained by venipuncture and 24 then culture them or maintain them in RPMI plus 10 25 percent fetal calf serum ex vivo, the anaerobic 204 1 conditions do prevent apoptosis and do significantly 2 prolong isolated PMN survival. However, when we 3 looked at the leukopheresis product, that is, the 4 product obtained after G-CSF and dexamethasone, 5 centrifugation leukopheresis, and then tried to look 6 at the effect of the anaerobic environment on that 7 leukopheresis product, we found no significant effect. 8 So you can't translate findings that you might see 9 with isolated PMNs with the actual survival of PMNs in 10 a granulocyte product. So anaerobic environment did 11 not appear to be a viable option to maintain cells 12 during storage. 13 We also found that reduced temperature 14 also reduced apotosis of neutrophils during storage or 15 during maintenance ex vivo, but if we tried to 16 translate this at 4 degrees into storage of the 17 leukopheresis product, we found that this was 18 impractical. At 4 degrees, there was a lot of 19 clumping of leukocytes that could never be 20 retransfused. However, when we looked at 10 degrees, 21 we found no significant clumping of leukopheresis 22 products at 10 degrees. So we subsequently tried to 23 look at whether or not 10 degrees was more effective 24 than room temperature, and I will get to those 25 studies. 205 1 It is also known that cytokines, 2 especially G-CSF and I suppose GM-CSF, are most 3 effective at decreasing neutrophil apoptosis. So we 4 wanted to look to see whether or not the readdition of 5 G-CSF ex vivo to the leukopheresis product would 6 further prolong neutrophil viability in addition to 7 just the viability that we could obtain by reduced 8 temperature. 9 Corticosteroids are also known to decrease 10 neutrophil apoptosis, but we didn't want to add 11 additional corticosteroids to the product that we 12 obtained. And, of course, we couldn't add other 13 agents like LPS which are also known to decrease 14 neutrophil apoptosis. 15 So we thought the best way to try to study 16 prolongation of neutrophil survival in the 17 leukopheresis product -- and this is the leukopheresis 18 product obtained after G-CSF and dexamethasone -- was 19 to look at reduced temperature and also the readdition 20 of G-CSF. This appeared to be most practical. 21 So the protocol for our study was to 22 stimulate donors with 600 mcg G-CSF subcutaneously and 23 also 8 mg of dexamethasone orally. Then 24 centrifugation leukopheresis was performed 12 hours 25 after stimulation. Then we looked at storage or 206 1 baseline and at 24 and 48 hours. We looked at room 2 temperature as one condition. Room temperature plus 3 the readdition of G-CSF with 100 nanograms per ml to 4 the storage bag. We looked at 10 degrees and then 10 5 degrees plus G-CSF added to the storage bag. 6 Then we looked at the following 7 parameters. We looked at white counts and 8 differential, respiratory oxidative burst activity, 9 immunophenotype, staphylocidal activity using a 10 conventional four plate assay of killing of Staph 11 aureus, and also fungicidal activity. In terms of 12 fungicidal activity, we didn't look at what has 13 usually been looked at and that is blastochlamydia 14 killing. We actually looked at hyphae damage, which 15 is more relevant for the clinical situation, which I 16 will get to. 17 Now in terms of storage of this product, 18 when you look at the ANC of the product, you can see 19 that it really doesn't change regardless of what 20 storage condition that we had. This is fresh product 21 here and you can see that there is no significant 22 difference if the product is stored at 10 degrees or 23 at room temperature, or if it is stored in the 24 presence of G-CSF, which is shown not on this slide 25 but another one right here. So under any condition at 207 1 reduced temperature, whether it is room temperature or 2 if it is at 10 degrees, you get preservation of the 3 ANC in the product, and the addition of G-CSF did not 4 appear to affect the subsequent storage. 5 Now next when we looked at the respiratory 6 burst activity, we used luminol-enhanced 7 chemiluminescence, which is a rapid and sensitive way 8 to look at the respiratory burst and allows for a 9 kinetic analysis. We used a variety of stimuli. We 10 used PMA as a soluble stimulus. We also looked at 11 opsonized zymosan as a particle stimulus, and its 12 activity is primarily related to its ability to bind 13 and then to be engulfed in the CD11B18. And then we 14 also looked at FMLP, and FMLP of course binds to a 15 cell surface receptor. I am just going to show you 16 the PMA and opsonized zymosan results, just because it 17 gets repetitive if we keep on going through it. 18 Here if we looked at the storage of this 19 product, and this is at 24 hours, with the stimulus 20 being PMA with and without G-CSF at the various 21 temperatures, you can see that the product is always 22 better in the baseline activity. It doesn't appear to 23 matter significantly, at least at this level, whether 24 or not the product is stored at room temperature or at 25 10 degrees. It also, although there appears to be a 208 1 relative greater benefit here at 10 degrees as opposed 2 to room temperature, this is probably not clinically 3 significant because the baseline product is certainly 4 no better than here and this probably reflects that 5 initial diminution in the product that I said that we 6 saw earlier. That is, after the product is first 7 collected, it appears to be relatively refractory to 8 a secondary stimulus, but it regains in response to 9 that stimulus. So if we were to look at this, we 10 would say in terms of overall oxidative capacity, it 11 is relatively well preserved either at room 12 temperature or at 10 degrees and the readdition of G 13 probably didn't make any difference. 14 Similar effects were also seen at 48 15 hours. We won't dwell there. And it is also seen 16 with opsonized zymosan as a stimulus. Here you see 17 the baseline product and then under any of these 18 storage conditions at 24 hours, you see enhanced 19 activity in response to opsonized zymosan. You still 20 see this enhancement at 48 hours. So overall, we can 21 say under these storage conditions here, oxidative 22 capacity appears to be preserved and the readdition of 23 G-CSF may not be necessary to retain that property of 24 the cells. 25 Now we also performed immunophenotype 209 1 analysis of the neutrophils during storage. We looked 2 at adhesion receptors at CD11B and CD18 and also L- 3 Selectin. We also looked at the three IgG CD 4 receptors, CD16, CD32, and CD64. We also looked at 5 CD14. I would just like to show you the CD16 and 6 CD32, and CD64 data for simplicity, because I don't 7 want to have to go through all the other ones here. 8 The major point here is that in terms of CD16, CD16 9 declines as cells undergo apoptosis. What you can see 10 here is that although there is a slight decline, CD16 11 expression is maintained throughout the storage period 12 or through 48 hours whether or not the cells are 13 maintained at 10 degrees or at room temperature and 14 whether or not G-CSF is present or absent. So what we 15 would conclude is that storage at reduced temperature 16 with or without G-CSF maintains CD16 expression, 17 retains CD32 expression, and also retains CD64 18 expression. So that reduced temperature of this 19 leukopheresed product during storage maintains 20 cellular viability and a favorable immunophenotype in 21 terms of FC receptor expression. A favorable 22 immunophenotype was also retained if you looked at 23 CD14 expression or in terms of adhesion receptor 24 expression. 25 Now I am not going to show the 210 1 bacteriocidal activity here, because the bacteriocidal 2 activity was always greater than 95 percent throughout 3 the study period. So bacteriocidal activity was 4 always intact regardless of the storage condition. 5 But what is more important when we are considering the 6 microbicidal activity of a storage product is actually 7 a fungicidal activity. Because if we were going to 8 envision a clinical trial, one would be most concerned 9 with serious fungal infections in our neutropenic 10 patients. That is really where the real problem is in 11 terms of the oncologic and infectious disease 12 standpoints. 13 I just wanted to emphasize the importance 14 of opportunistic fungal infections in neutropenic 15 patients. Prolonged neutropenia or abnormal 16 neutrophil function are the major risk factors for 17 opportunistic fungal infections. These opportunistic 18 fungal infections now represent the major cause of 19 infection-related mortality in bone marrow transplant 20 or marrow transplant patients. And of these 21 opportunistic fungal infections, invasive 22 Aspergillosis and Candidemia are the most common 23 opportunistic infections or mycotic infections in 24 these marrow transplant patients in patients with 25 prolonged neutropenia. 211 1 Our experience at the Fred Hutchinson 2 Cancer Research Center in Seattle from 1992 to 1996, 3 which includes the period -- and this is important for 4 trial considerations and while you really can't 5 probably rely on historical controls -- and that is 6 because our infectious disease prophylactic regimens 7 have changed significantly and thereby impacted 8 outcome and survival of patients as compared to years 9 past. From 1992 to 1996, this reflects the experience 10 during fluconazole and ceftazidine prophylaxis. At 11 the Fred Hutchinson Cancer Research Center during this 12 period, only 40 percent of patients who developed 13 fungemia during neutropenia had clearance of fungemia 14 within 10 days and survived for four weeks. More 15 importantly or just as important, less than 30 percent 16 of patients during this period who developed invasive 17 mold infections during neutropenia survived for 12 18 weeks. This just emphasizes the importance of these 19 infections in this patient population. We can treat 20 most bacterial infections fairly effectively nowadays. 21 The problem of CMVs still exist, but our methods to 22 control CMV problems are much better than they were 10 23 years ago. But fungal infections remain a major 24 problem and it is a problem that now defies current 25 antimicrobial prophylaxis and one that is really a 212 1 frontier, I think, of oncology and infectious 2 diseases. 3 Now in terms of the fungicidal activity of 4 the granulocyte product during storage, we wanted to 5 look at hyphae or pseudo-hyphae killing. The reason 6 being is that hyphae of true molds like Aspergillus or 7 rhizopus species or pseudohyphae of Candida species 8 are the predominant tissue forms of opportunistic 9 fungi during invasive infections. Usually people look 10 at Candida or blastocandida, which are easier targets. 11 The hard target are the hyphae and pseudohyphae, so we 12 really wanted to evaluate whether or not the cells 13 stored under these conditions could actually have 14 activity against hyphae and pseudohyphae. To do this, 15 we employed an XTT assay which measures leukocyte 16 mediated damage to hyphae or pseudohyphae in vitro. 17 What we found here -- we will first look 18 at Candida Albicans. It is that we looked at this 19 product that was stored at 10 degrees without the 20 addition of G-CSF, and we looked at neutrophils 21 obtained or the buffy coat of patients prior to 22 stimulation or of donors prior to stimulation, then at 23 baseline, and then at day 1 and 2 of storage of the 24 product. So this is the buffy coat of individuals 25 prior to getting G-CSF and dexamethasone. This 213 1 represents the activity of the leukopheresis or the 2 granulocyte product after leukopheresis on day zero 3 after individuals received G-CSF and dexamethasone for 4 stimulation. And then this is the storage of the 5 product at 10 degrees without the addition of G-CSF. 6 Then we looked at two effector to target ratios, the 7 effector cell being a leukocyte in the granulocyte 8 product, and the target being the pseudohyphal form of 9 Candida Albicans. And what you can see here is that 10 the activity against the pseudohyphae is maintained 11 throughout the storage period. We see good activity 12 that is maintained throughout the storage period. So 13 these cells even after 48 hours can still mediate 14 activity against Candida Albicans. 15 Similarly, activity is fairly well 16 maintained against Aspergillus fumigatus hyphae, a 17 very tough organism or a very tough form of the 18 organism to kill. So that we see especially at the 10 19 to 1 E to T ratio good maintenance of activity 20 throughout the storage period. It is certainly 21 greater at day 1 than day 2, but we still have a 22 significant level of activity at day 2. 23 So what can we conclude from these 24 studies? First of all we can conclude that the 25 granulocyte product obtained by centrifugation 214 1 leukopheresis from donors stimulated with G-CSF and 2 dexamethasone retain significant functional activity 3 when stored at reduced temperature for 24 to 48 hours. 4 From these studies, we could not conclude whether or 5 not 10 degrees was better than 22 degrees. In many of 6 the assays, 10 degrees looked to be slightly better, 7 but I don't know if that would be clinically 8 significant. So I think that just reduced temperature 9 -- in other words, storage at room temperature or 10 10 degrees would be superior to storage at 4 degrees or 11 at 37 degrees. Also, the addition of exogenous G-CSF 12 to this granulocyte product provides at best modest 13 benefit and probably would not be necessary to 14 maintain granulocyte products for 24 to 48 hours. 15 So what are the important questions I 16 think for future research in this area? Well, first 17 of all, we have done our ex vivo or in vitro analysis. 18 Now I think it is important to determine does the 19 stored granulocyte transfusion product function with 20 appropriate in vivo kinetics when transfused. We 21 should establish this to make sure that products could 22 be stored if we were going to do a clinical efficacy 23 trial. 24 The most important question, though -- and 25 this is to prevent the unwise use of granulocyte 215 1 transfusions in the future, and that is actually to 2 perform an appropriate randomized control multi-center 3 clinical trial. We have to know what the appropriate 4 clinical indications for granulocyte transfusion 5 therapy are. We can't really extrapolate from 6 historical controls because the practice has changed. 7 We have better antimicrobial regimens. So really some 8 sort of controlled multi-center clinical trial is 9 necessary to determine what are the appropriate 10 clinical indications and the specific clinical 11 indications for granulocyte transfusion therapy. We 12 have to know is it going to be effective for invasive 13 fungal infection or invasive Aspergillosis for mucal 14 mycosis. Will it be effective for fungemia or 15 Candidemia? Will it be effective for bacteremia in 16 the neutropenic host? I actually don't think that we 17 will see a benefit here just because bacteremia in the 18 neutropenic host is usually fairly well treated now 19 with the antibiotics that we now employ in clinical 20 practice. Then also another area, as David mentioned 21 earlier, was neonatal sepsis. We have to determine 22 the exact specific indications for neutrophil 23 transfusion therapy. If we don't, it will be used 24 maybe inappropriately in situations and actually could 25 probably cause much greater harm than good and also be 216 1 an unwise use of resources in the future. Only 2 through a randomized controlled clinical trial can we 3 determine the appropriate indications and avoid that 4 unwarranted use. 5 So I would like to acknowledge my other 6 collaborators during these studies, David Dale and Tom 7 Price. Milton Gaviria is a fellow that works with us 8 and he has been doing a lot of the antifungal assays. 9 Then Ellen Roger is a technician who has been working 10 with David Dale and myself for a long period and she 11 has been working with the granulocyte storage for the 12 last several years. Thank you. 13 DR. STRONCEK: I'd like to have the 14 speakers from this afternoon come up and we can answer 15 questions. Dr. Leitman? 16 DR. LEITMAN: Thank you. I have a 17 question for Dr. Liles. In a slide shown by you and 18 earlier by Dr. Dale on the kinetics of in vivo 19 recovery of autologous labeled G-CSF mobilized 20 granulocytes, the first column had to do with the 21 recovery. And the in vivo recovery, if I read that 22 right, was lower in the G-CSF treated products, a 65 23 percent in untreated versus 31 percent, although the 24 half-life was twice as long, 9.6 versus 20 hours. 25 Could you comment on that decreased recovery? 217 1 DR. LILES: Well, actually -- Tom, would 2 you want to comment? Or Tom or David, do you want to 3 comment on that? 4 DR. DALE: The recovery is lower. That 5 is, if you transfuse cells that have been collected by 6 leukopheresis, you don't get the same percentage of 7 those cells circulating initially. I think there is 8 an element of damage to the cells that occurs with 9 collection. And several people have commented that 10 the one-hour increment in the counts are not 11 necessarily the highest. And Tom and I found years 12 ago that if you collect cells by leukopheresis and 13 transfuse them, sometimes the counts go up and 14 actually cells that initially marginated will enter 15 the circulating pool. Those are the extrapolated 16 values you would get at the initial time of 17 transfusion. So I think that the values are lower 18 than normal, but I think the cells probably do recover 19 some function from being back in the warm, healthy 20 body. 21 DR. LEITMAN: So those are one-hour 22 recoveries. But from data presented by all of us here 23 today, it looks like the 4 or 8 hour post-transfusion 24 count is higher. 25 DR. DALE: That is right. These are 218 1 extrapolated based on radioisotopic. In normal 2 people, not in neutropenic people, you have to use an 3 extrapolated value based upon isotopic labeling. But 4 it is -- would be correct if you could measure the 5 recovery at four hours later approximately. I suspect 6 it would be higher. Does that make sense? 7 DR. LEITMAN: Yes. I have another 8 question. This is for Dr. Adkins. In your last trial 9 -- you went through a lot of different trials -- you 10 are giving allogeneic donors 15 mcg per kilo of G-CSF. 11 I want to point out that anytime you exceed 10, that 12 is two subcutaneous injections because most nursing 13 standards do not allow you to exceed 1.5 to 2 ml per 14 single subcutaneous dose, which doubles the discomfort 15 to the donor to get two subcutaneous shots rather than 16 one. The increment in your yield was 15 mcg per kilo 17 versus 5 mcg per kilo. It was not very great. I 18 think you had 10 times 1010, whereas Seattle and NIH 19 are getting around 8 times 1010. And in every study 20 I have seen, there is a dose-dependent increase in 21 adverse effects in the donor. So could you justify 22 why you are using 15 mcg rather than the lower dose? 23 DR. ADKINS: Well, actually in most of the 24 patients -- the vast majority of the patients in our 25 clinical trials, of which several you have seen, we 219 1 have used either 5 or 10 mcg per kilogram. We were 2 interested in defining whether or not there was a dose 3 response effect in terms of component yield. So we 4 have a very limited number of people that received 15 5 mcg per kilogram. In the autologous transplant trial, 6 which was the latter one that I mentioned, we were 7 giving 10 mcg per kilogram in that setting. So I am 8 not certain that we are going to necessarily pursue 9 doing 15. I think you can make an argument. If you 10 are going to use a single donor to donate granulocytes 11 for one patient over a course of a week or so and if 12 you are thinking of the strategy of storing overnight 13 a portion of the components, you can make an argument 14 for using a higher cell dose given that you get 15 greater yields with that approach. I mean, that is 16 just a strategy that one might take and that is 17 something that we are kind of looking at at Washington 18 University. I think that I agree with you about the 19 issue of toxicity. You know, we -- in the way we do 20 this, the donors clearly express a much greater 21 problem with toxicities during the phase of giving 22 growth factors to collect stem cells as opposed to the 23 time in which we give the growth factors to collect 24 the granulocytes. I don't that the problem with 25 toxicity is a very big problem when we give the growth 220 1 factors to mobilize the granulocytes. Their biggest 2 complaints occur the week prior when we are trying to 3 collect their stem cells, as I discussed earlier. 4 DR. SNYDER: Ed Snyder from Yale. Just a 5 couple of practical aspects. We had done some work 6 many years ago with stored granulocytes at room 7 temperature looking at the ability to put them through 8 an electromechanical pump because many of the oncology 9 units were doing that to decrease the flow rate and 10 yet make sure they went in in an appropriate time. So 11 studies, if they are going to be repeated with the G- 12 CSF, that might be a very practical point to look at 13 to see if the mechanical shear stresses don't have a 14 negative impact on granulocytes that go through the 15 pump. Because some of them can chew up the red cells. 16 But we didn't see any problems at that time without G- 17 CSF several years ago. 18 Another study that we did was to look at 19 the effect of the granulocytes on the platelets. 20 Because with all due respect to the neutrophils, we 21 think platelets are also beautiful cells that we need 22 to care for. And what we found was that stored for 24 23 hours at room temperature, there was a decrease -- if 24 I can remember this and I would have to go back and 25 look at the paper -- a decrease in GP1B on the 221 1 platelet, which we thought was due to release of 2 neutrophil enzymes during storage. So if the FDA is 3 going to consider neutrophils with platelets in them, 4 then someone needs to study the effect of storage, 5 certainly G-CSF stimulated granulocytes on platelet 6 function or whether the platelets should be removed 7 from the storage separately if they can co-exist 8 together in the bag. So I just mention those for the 9 record. 10 DR. DIAZ: A quick question for Dr. 11 Liles. If I interpreted your slide correctly, the 12 actual kinetics of the response at 48 hours or at some 13 of the late time points seem to be totally different 14 from the normal classical response of up and down in 15 15 minutes and then totalling up to 60 minutes at time 16 zero. Can you explain that or did I just read it 17 wrong? 18 DR. LILES: Which one? The stimulation 19 with PMA? You mean the chemiluminescence? 20 DR. DIAZ: Yes. 21 DR. LILES: I don't have a good 22 explanation of why that is in terms of why it is 23 flattened initially. Is that what you are saying? 24 DR. DIAZ: Yes. 25 DR. LILES: And then it rises out. No, I 222 1 don't have a good explanation of why. It is just that 2 that was a constant observation. 3 DR. DIAZ: Okay, thanks. 4 AUDIENCE MEMBER: This is a comment for 5 Tom Lane. I liked your presentation in reviewing all 6 of the details about storage. I think, though, when 7 we consider function, we can't go back to and rely on 8 these congenital defects in neutrophil function to 9 give us some idea on what we should expect or what we 10 should shoot for. An example of this is that you had 11 stated that we needed to have perhaps 10 percent -- at 12 least 10 percent or around 10 percent oxidase 13 activity. It depends on how you look at it. If you 14 look at the kind of classic patients with chronic 15 granulomatous disease, they have no activity. If you 16 look at variants, that data of 10 percent comes from 17 variants. And if you have perhaps 10 percent normal 18 cells, then you will have normal function. That is 19 very different than a patient that I have who on a 20 good day all of her cells have 20 percent activity and 21 she is always having problems with infection. And it 22 should remind us that one of the things that we have 23 to evaluate in terms of function is whether if we see 24 decreased function or increased function, whether that 25 is related to a subset of cells or whether it is all 223 1 cells with lower function or just several different 2 populations of the function. 3 DR. LANE: Yes. I think your point is very 4 well taken. That was just meant to give general 5 information. The other point that I neglected to make 6 is that certainly in the presence of multiple defects 7 in function, even minor defects may take on a lot more 8 importance. So I think that needs to be kept in mind 9 as well. 10 DR. STRONCEK: Along those lines, though, 11 I think the issue on storage is very confusing. I 12 agree with Tom Lane's summary of the literature that 13 probably if you are going to do one functional assay 14 in the laboratory on stored granulocytes, it should be 15 chemotaxis. But that said, the literature suggests 16 that storing granulocytes at less than room 17 temperature might preserve a lot of function but not 18 chemotaxis. I guess my question for you, Dr. Liles, 19 is have you looked at chemotaxis of your cells stored 20 at 10 degrees? 21 DR. LILES: No. You know chemotaxis 22 assays are the most probably variable of all the 23 assays and that is always a problem. With newer sort 24 of techniques, it is possible that you could. It 25 might be less laborious. But I think actually the 224 1 best test is actually to do the in vivo study that we 2 were talking about and to retransfuse and to see 3 whether or not you can get proper in vivo migration to 4 the potential inflammatory site, meaning the skin 5 window or to the buccal mucosa. That is really the 6 real test, even better than in vitro chemotaxis. So 7 I think that illustrates the point. Really we have to 8 look at these cells when retransfused in vivo to see 9 whether or not they can adhere to vasculature or to 10 the endothelium and then migrate through the 11 endothelium to an inflammatory site. That will be 12 most important prior to conducting a large clinical 13 trial to make sure that stored granulocytes can 14 migrate to an inflammatory focus. 15 DR. STRONCEK: Conrad, a minor point. 16 When you talked about bacteriocidal activity, was that 17 at a one to one ratio? Did you look at different 18 infectious ratios? 19 DR. LILES: Dan, we didn't look at 20 different ratios. It was at a one to one at that 21 point. So we didn't look under a stress situation, so 22 to speak. 23 AUDIENCE MEMBER: Conrad, I think you 24 probably mentioned this, but I probably missed it with 25 all the data. When you collected these granulocytes 225 1 and stored -- first of all, you collected them on the 2 CS3000 in what volume? And then when you stored them, 3 you just stored them as a product? You didn't isolate 4 the granulocytes? 5 DR. LILES: Exactly. So they were the 6 standard conditions that you saw earlier when Tom 7 Price discussed it. So they were stored under those 8 same conditions and under those same collection 9 parameters -- yes, COBE. 10 AUDIENCE MEMBER: COBE. All right. So 11 what kind of volume do your granulocytes -- what is 12 the final volume? 13 DR. LILES: The final volume is more like 14 300 cc or so. 15 DR. LANE: Again, to ad nauseam. I am 16 really concerned about the cell concentrations, 17 particularly when these cells are stored at room 18 temperature, and I think that is one of the first 19 things that people should look at. And I guess you 20 are going to present some work on that this afternoon. 21 AUDIENCE MEMBER: I would like to reflect 22 on a comment that was made before and actually second 23 that. That is we seem to be coming down to two basic 24 issues. One is a clinical trial to look at the 25 efficacy of granulocytes, and the other big area is 226 1 going to be to look at storage. And I think probably 2 we are not going to be able to look at storage -- we 3 are not going to be able to look at them both at one 4 time. It would probably be wiser to not store for a 5 clinical trial and then come back later when some 6 basic work is done on storage techniques and look at 7 that. 8 DR. LILES: Could I also make a comment. 9 I think also the issue of alloimmunization is very 10 important. Because when I am approached by an 11 oncologist regarding the possibility of granulocyte 12 transfusions from somebody in the pre-bone marrow 13 transplant setting, the question is always will giving 14 granulocytes from community donors obviate the chance 15 for a successful graft later. And we really don't 16 have good information to say whether or not that is 17 the case. And that is always going to be a hindrance, 18 I think, to the use or this practice unless those data 19 are available. So the issue of alloimmunization and 20 its clinical importance is still one that is out there 21 and is worth further investigation. 22 AUDIENCE MEMBER: With regards to 23 alloimmunization and the storage issue, has anybody on 24 the panel considered cryopreservation, which would 25 allow you to have autologous donations prior to 227 1 treatment? 2 DR. LILES: We tried cryopreservation by 3 different techniques and you get a gelatinous mess. 4 DR. LANE: I am sorry, I meant to comment 5 on that and I didn't. It is very interesting if you 6 review the literature on cryopreservation of 7 granulocytes in that you usually find one or at most 8 two publications by the investigative groups. Most of 9 the people who study cryopreservation disappear from 10 the face of the earth and are never heard of again. 11 And that always worried me. I think there may be 12 actually a few of us around. But cryopreservation is 13 a technology that is so far away from being 14 practicable for these kinds of cells that it is a 15 major area that would be wonderful if someone could 16 find out how to do it, but so far we are very far away 17 from that. 18 DR. DALE: I would just add that what 19 happens with the best cryopreservation methods now, if 20 you look by electron microscopy at the cells, they 21 blub. And it is probably that the freezing, even in 22 the best of circumstances, disrupts the membranes of 23 the granules and you get some damage to the 24 cytoplasmic components of the cells. We actually in 25 the last couple of years have taken this far enough to 228 1 measure chemiluminescence of thawed cells after 2 freezing, and there is a little activity there. But 3 it is really much diminished. And I don't know where 4 the breakthrough will come in getting the 5 cryopreservative into the cell to somehow preserve its 6 many very fragile parts, but I would not be optimistic 7 either. Really the hope is for two or three days of 8 storage. 9 AUDIENCE MEMBER: Well, I think one of the 10 issues is the clinical situation. If you look at 11 something like the red cells where you have the 12 ability to have wash steps and post-thaw treatments, 13 you might be in a much better situation than if you 14 needed something that would be directly transfusable. 15 Because there are new cryopreservation programs now 16 that are looking at water structuring molecules that 17 cross the membrane that are transfusable and nontoxic 18 that can, in fact, stabilize organelles and other 19 components. So while it hasn't been shown recently 20 with granulocytes, I think there has been some advance 21 in cryoprotectant technology that may allow them to be 22 applicable in this situation. It certainly appears 23 that we have a much hardier cell population now also 24 in terms of these stimulated programs and the ability 25 to reduce the amount of red cells and platelets in the 229 1 bags. 2 DR. LANE: I'd agree by saying that there 3 is a great opportunity for good work to be done in 4 this field. 5 DR. STRONCEK: If there are no more 6 questions, I guess that concludes this part of the 7 presentation. We will have abstracts, I guess. 8 CHAIRPERSON HARVATH: We have three 9 abstracts. Dr. Diaz is here from La Jolla, and Idun 10 Pharmaceuticals is going to speak on the preservation 11 of neutrophil viability through inhibition of cast 12 base dependent apoptosis. Then after he speaks, Dr. 13 Stroncek is going to present a couple of abstracts of 14 their work here that he has done in collaboration with 15 Dr. Susan Leitman and colleagues at the NIH. 16 (Whereupon, at 2:59 p.m. off the record 17 until 3:03 p.m.) 18 CHAIRPERSON HARVATH: We will let David 19 Stroncek present his abstracts first. They are going 20 to try and fix the bulb. Dave? 21 DR. STRONCEK: I'd like to thank everybody 22 who is still here at the end of a long two days of 23 meetings. I would like to talk about some studies. 24 They are preliminary studies that we have done looking 25 at first some of the safety aspects of giving G-CSF 230 1 and dexamethasone to people donating granulocytes. I 2 have had a lot of experience giving G-CSF to normal 3 donors, particularly with peripheral blood stem cells. 4 And the question comes up during every lecture similar 5 to the ones we got today about how safe is this really 6 to be giving donors a drug. 7 So with that in mind, I thought it was 8 worthwhile looking very carefully at granulocyte 9 donors to determine the effects of G-CSF on them. We 10 know from studies on giving G-CSF to stem cell donors, 11 and we heard some of this this morning, that there is 12 a marked effect on neutrophils giving G-CSF. After 13 the G-CSF is given and stem cells are collected, 14 platelet counts fall. The platelet counts fall not 15 only due to the dropoff from the apheresis collection, 16 but there seems to be some direct suppression of 17 platelet production by G-CSF. In addition, there is 18 a transient neutropenia and lymphopenia that occurs 19 after the collection of G-CSF mobilized stem cells. 20 We also know in stem cell donors there is a 21 marked increased in alkaline phosphatase and LDH, 22 slight increases in uric acid, and falls in potassium 23 and bilirubin. These falls are transient but they do 24 occur predictably in donors. As you have heard this 25 morning or today already, both stem cell donors and 231 1 people given G-CSF to donate granulocytes do 2 experience headaches, bone pain, myalgia, and fatigue. 3 The question I wanted to focus more on was 4 what happens with blood counts and blood chemistries 5 in donors given G-CSF. In specific, one question that 6 comes up is because we are using kind of a small group 7 of people that we have talked to that we have asked to 8 give G-CSF, we tend to use them over and over again to 9 donate granulocytes. One question that I have asked 10 and we have asked is how often can someone safely 11 donate granulocytes. I am not sure what the answer 12 is, but I do know I would like their blood counts and 13 chemistries to be normal before we give them G-CSF or 14 dexamethasone again and collect granulocytes. 15 So with that in mind, we designed this 16 study. We mobilized granulocytes three different 17 ways. Either with dexamethasone, G-CSF, or 18 dexamethasone plus G-CSF. We used the same protocol 19 that Dr. Leitman described earlier today. The 20 dexamethasone was an 8 mg dose 12 hours before the 21 collection. G-CSF was 5 mcg per kilogram 22 subcutaneously about 18 hours before the collection. 23 And then when we gave both, it was 8 mg of 24 dexamethasone and 5 mcg of G-CSF per kilogram. Only at 25 the NIH, but this is a double-blind study. So, yes, 232 1 we did have placebo tablets or dexamethasone, and we 2 did either give G-CSF or a placebo injection. 3 So we did look at symptoms and this was a 4 nice way to try and sort out the effects of 5 dexamethasone from G-CSF. We plan to enroll 24 6 donors. I have enrolled 10 so far and I am going to 7 show you the data on 6. Each donor would be 8 randomized to one arm and then we would study their 9 blood chemistries for several weeks and then they 10 would come back six weeks later to be enrolled a 11 second time and a third time. So all three donors got 12 each -- all of the donors got each of the three 13 mobilization regimens and had granulocytes collected. 14 What we measured was symptoms, blood 15 counts, and blood chemistries. I am going to focus 16 mostly on the blood count and chemistry data. We 17 analyzed the donors prior to mobilization, pre and 18 post-collection, and then one and two days after the 19 collection, and then once per week weekly for five 20 weeks. 21 This is a summary of the platelet counts 22 in the donors. First of all, as you would expect in 23 all three donor groups, the platelet counts fell after 24 the collection. These counts are premobilization. 25 And as expected, the counts fell about 20 to 30 233 1 percent. What we have seen in peripheral blood stem 2 cell donors is that the platelet counts remain low or 3 at about post-collection levels for almost a week. We 4 didn't see that in any of these groups. In the 5 dexamethasone donors, the counts started coming up by 6 day 4. And then about a week after collection, the 7 counts were back into the normal range. We did see a 8 slight overshoot in the counts after two weeks. And 9 by three and four weeks, the counts were back in the 10 normal range. The counts were almost the same in the 11 people who got G-CSF. Again, the counts started to 12 come up -- they were low one day after the collection 13 and started to come up two days and then a week later 14 they were back to normal and two weeks after that they 15 were above normal. A similar effect occurred with G- 16 CSF. So we didn't see any adverse effect of G-CSF on 17 platelet counts. It looks like the recovery of counts 18 is almost identical to donors given dexamethasone. 19 We also looked at neutrophil counts just 20 to see if there would be any post-collection 21 neutropenia. First of all, as you would expect, the 22 day 1 neutrophil count is about the same. They are 23 the same group of donors. And as you have seen 24 several times today, the counts are increased to about 25 20,000 or more with G-CSF and considerably more in 234 1 people that got G-CSF plus dexamethasone. The day 2 after the collection, the counts were still slightly 3 elevated in the G group and G plus dex, and even 4 slightly higher at day 2. We did not -- the next week 5 and the following week, we really didn't see any 6 difference in granulocyte counts from the pre-G-CSF 7 counts. So based on this data, their granulocyte 8 counts come back to normal baseline quickly, and at 9 least a week afterwards, there would be no reason why 10 a donor could not get another dose of G and donate 11 granulocytes again. 12 Concerning blood chemistries, it has been 13 well known that when you give G-CSF for three, four, 14 or five days, alkaline phosphatase double or triple. 15 A single dose of G-CSF though does not seem to have a 16 very marked effect on LDH levels. It does increase 17 them though. Baseline levels were at 148, 141, and 18 146, similar in the three groups. And as you would 19 expect, dexamethasone did not affect the LDH levels 20 after the second day. G-CSF though in both groups did 21 result in a slight increase in LDH levels. And then 22 the day afterward, actually the levels were below 23 baseline. I am not sure why that was. It may have 24 been due to some dilution following the apheresis. 25 In contrast to peripheral blood stem cell 235 1 donors, we did not see any change in alkaline 2 phosphatase. Apparently it takes more than one day of 3 G-CSF to cause changes in alkaline phosphatase. 4 These data were a little bit surprising in 5 that I did expect the potassium levels to fall when G- 6 CSF was given, and a single dose did result in quite 7 a significant drop in potassium from 4.2 to 3.7, which 8 was back to 4.0 again the day after apheresis. I was 9 surprised though to know that a single dose of 10 dexamethasone can also result in a fallen potassium 11 level. The level went from 4.3 to 3.9. The combined 12 effect of G-CSF and dexamethasone though does not seem 13 to be any different than either drug alone. This 14 change was very transient and by day 3 the potassium 15 levels seemed to be almost back to normal. 16 Another effect I didn't expect with 17 dexamethasone was a change in albumin levels. With 18 apheresis, donors tend to get some fluids, and I 19 thought it wouldn't be unexpected if they would get a 20 little bit of hemodilution and a fallen albumin. We 21 saw that, but we only saw it in donors that got 22 dexamethasone as opposed to donors that got G-CSF. So 23 at least by one measure dexamethasone may actually 24 cause more change in blood chemistry and G-CSF 25 doesn't. That wasn't the only parameter that changed 236 1 slightly with dexamethasone and not G-CSF. Again, 2 albumin went up very slightly. Although these numbers 3 might be statistically significant, I don't think they 4 are going to make any difference for donors. 5 Phosphorus did drop for all three donor groups, but 6 again transiently. 7 This kind of summarizes what happens with 8 the chemistries we measured. Sodium bicarb, 9 creatinine, calcium, magnesium, bilirubin, alkaline 10 phosphatase and SGPT, SGOT, and GGTs were unchanged. 11 As I pointed out, all donors in all three groups had 12 decreases in phosphorus and potassium. Donors in all 13 groups had a very slight change in cholesterol and 14 triglycerides. Triglycerides might be related to 15 fasting before apheresis. I am not sure on the 16 mechanism of change in cholesterol in the groups. 17 Dexamethasone, as I showed, had a slight decrease in 18 albumin. And then again we saw the data on the 19 bilirubin and the LDH. Chloride was slightly 20 increased in the dexamethasone group and uric acid had 21 a very slight increase in all three groups. The 22 bottom line is I don't think any of these changes are 23 really very significant, and I think that based on 24 these findings it is going to be safe from a blood 25 count and chemistry point of view to give G-CSF once 237 1 weekly. 2 So actually I was quite surprised. This 3 data suggests that in some ways dexamethasone might 4 actually be more potentially harmful to donors than G- 5 CSF. So in summary of this part of the data, the 6 mobilization of granulocytes with either G-CSF or 7 dexamethasone is associated with mild changes in blood 8 chemistries. These blood chemistry changes and cell 9 counts did return to baseline promptly and it may be 10 safe for donors to donate mobilized granulocytes at 11 one-week intervals. 12 I would like to continue on with the rest 13 of what I did with this study. On this same study, we 14 did collect granulocyte concentrates and we did want 15 to see how well these concentrates stored. The issues 16 I think a lot of people have already mentioned today, 17 though, that granulocyte standards limit storage to 24 18 hours. And as we have heard many times, G-CSF 19 inhibits apoptosis. It may be possible to store these 20 mobilized granulocytes more than 24 hours and still 21 maintain viability and function. We did have a 22 concern that high cell concentrations in the products 23 that we were collecting could counter the effects of 24 G-CSF and actually diminish viability. 25 Again, the study was exactly the same. 238 1 the same mobilization. We collected the products with 2 a CS3000 blood cell separator and 7 liters of blood 3 were processed. With the CS3000, our volumes tend to 4 be a little bit less than with the products collected 5 in Seattle with the COBE. I will show you later that 6 the products had about 225 ml of plasma. We stored 7 the products at 48 hours, and for this study we 8 measured cell counts, pH, and we also looked at 9 viability. My lab isn't geared up to measure -- not 10 yet anyway to measure a large number of granulocyte 11 functions, but as long as we had these products, we 12 thought we would get some data. 13 This slide summarizes the first 18 14 products we collected. Again, it is very similar to 15 what Dr. Leitman has shown earlier today. The volume 16 of these products are all about the same, 230 ml. The 17 white cell concentration though is extremely high in 18 the products. For the dex products, it was 144 times 19 109 cells per liter. It went up to 204 for the G-CSF 20 and 332 for the G plus dex products. I think Tom Lane 21 showed some data where his highest concentration was 22 80 times 109 cells per liter. So this is at least 23 double that concentration and this is four times that 24 concentration. For a lymphocyte culture, we would 25 usually go with 1 or 2 times 109 cells per liter. The 239 1 composition of the products, again, was similar to 2 reported before. It was 66, 76, and 84 percent 3 granulocytes. And these were the total numbers of 4 cells present, 2.4 times 1010 cells and 3.7 times 1010 5 and 6.5. So these are the dose of cells that you have 6 heard many times that people are transfusing 7 practically these days. 8 The first thing we looked at again was 9 actual counts on the products, either immediately 10 after collection, day 1 or day 2. And as you have 11 seen before, the counts stayed very steady from the 12 first day of storage, at least the dexamethasone 13 products did. We saw a very slight fall-off in 14 counts, less than 10 percent after the second day. 15 Again, the counts were well maintained in both the G 16 and the G plus dex products. So at least by a gross 17 measure by counts, there wasn't much change in the 18 number of cells present. 19 We looked at cell viability using a 7AAD 20 stain. The cells were almost 99 percent viable the 21 first day, maybe 98 percent viable the second day. We 22 did -- maybe the fall-off went to 90 to 95 percent the 23 second day, but really it really didn't change much. 24 What we were surprised about though was the pH of 25 these products. As Dr. Lane showed earlier, the pH 240 1 was 7.1 to start with. I think in his most 2 concentrated products, he had a similar change in pH 3 at 6.3 the first day and then we went to two days and 4 it was 6.1. That wasn't probably too detrimental. But 5 when you go over to the other products, the very 6 concentrated products collected from donors given dex 7 plus G, right immediately after collection the cells 8 were slightly acidic and a day later the pH was 5.5 9 and stayed that way the second day. 10 The G stimulated products weren't much 11 better. They were 6.2 pH after day one and 5.8. What 12 was remarkable about these is we saw one or two of 13 these six products have a little bit of hemolysis of 14 red cells by the first day and almost all of them had 15 some hemolysis of red cells present after two days. So 16 even though these cells might look viable, just based 17 on the pH and the presence of hemolysis, I wouldn't 18 transfuse cells with the hemolysis present. 19 So we think something needs to be done to 20 try and maintain the viability a little better of 21 these cells if we are going to store them. The first 22 thing we thought of is well maybe it has to do with 23 concentration. How far out can we dilute these 24 granulocyte concentrates in order to maintain the pH. 25 This is a summary of preliminary results with diluting 241 1 four concentrates out with autologous plasma. What we 2 did was we allocated the concentrates into teflon bags 3 and added autologous plasma and did a 1 to 2, 1 to 4, 4 1 to 8, and 1 to 16 dilution. I had one product that 5 was -- I lumped the data. One product was collected 6 after dex mobilization and one after G and two after 7 G plus dex. And what we found was that if we diluted 8 the product out 1 to 8 or 1 to 16, we began to have a 9 normal pH and those pH's were maintained in that 10 normal range even after two days. So it looks like 11 diluting a product out just two or four-fold won't be 12 adequate, but somewhere around 8 to 16 fold might 13 maintain the pH. Again, these products were 230 ml to 14 begin with and we can't collect 2 liters of autologous 15 plasma. We are going to try some various additives to 16 see if that will help maintain storage. It may be -- 17 we heard other ideas today. Maybe adding bicarbonate 18 might be another way to go without diluting out the 19 concentrate. 20 I do agree these are preliminary studies 21 and we do plan to start to measure chemotactic 22 activity on some of these stored cells to see if there 23 is fall-off in chemotaxis function. 24 So concluding this abstract, to optimize 25 granulocyte storage, we believe they should be diluted 242 1 6 to 18 fold, especially G and dexamethasone mobilized 2 concentrates, or at least some other additive should 3 be added to help maintain the pH. We will probably 4 need clinical grade diluents or additives to maintain 5 the pH. I would like to thank the people that helped 6 me with these studies. Dr. Leitman and members of her 7 apheresis unit, Yu Ying You, Janice Carr, Hatian 8 Chung, who started some of the assays in the 9 laboratory with me, and Dr. Tom Lightfoot, who will be 10 continuing some of these storage studies. Thank you. 11 CHAIRPERSON HARVATH: Are there any 12 questions for Dr. Stroncek before we start? Maybe 13 what we will do is we will hold the question until the 14 next abstract and we will do it up here. Sorry, 15 because that microphone apparently is not working 16 right now. Now we will try the overhead projector 17 once again. I think we have a new projector. 18 DR. DIAZ: So this is the last talk of the 19 day. It is Friday. So I promise there are only 14 20 slides. You only have to be awake for 7 of them. So 21 I think we can get through this very quickly. 22 Diaz Law Number 2 of presentations. When 23 someone from a company is presenting something, its 24 credibility is inversely proportional to how slick the 25 presentation is. So in order to aid my credibility, 243 1 you will see that I have got low-tech black and white, 2 and in the very first slide I have introduced a 3 spelling mistake. Any other spelling mistakes you 4 spot from now on will be due to my aberrant education 5 on a small island just off the coast of Europe. 6 At Idon, our expertise and our interest is 7 actually in apoptosis. So the neutrophil is really 8 just a by-product of what we have been looking at. 9 And one of the things that we are very interested in 10 is in the caspase dependent apoptosis and cell death, 11 caspases being the enzymes that are involved with the 12 end stage of the death of cells. 13 Very quickly, the neutrophils obviously 14 have short circulating half-lives in the body. They 15 also seem to lose function and dye when stored in ex 16 vivo, such as in the leukopheresis pack. We know that 17 death in circulating neutrophils is apoptotic. We can 18 only surmise that perhaps the death that we are seeing 19 in the leukopheresis packs is also an apoptotic cell 20 death. Therefore, what we have been trying to see is 21 if we can interfere with the normal progress of 22 neutrophil death by using inhibitors of apoptosis. 23 So the two inhibitors that I want to look 24 at very quickly is one which is a generic inhibitor of 25 caspase, I think it is mostly the IL1 beta converting 244 1 enzyme, which is Z VAD fmk that has been used a lot in 2 different apoptotic research. This is one of our own 3 molecules made by one of our chemists, a highly 4 experienced and wonderful chemist. He is my boss, by 5 the way. Which is the azile dipeptide fmk, indole 6 fmk. 7 So the assays we have been looking at we 8 set up in order to look at the neutrophils and where 9 we are really inhibiting some neutrophil apoptosis is 10 we looked at oxidative burst assays using zymosan. 11 The reason why we use zymosan is because probably it 12 is more physiologically relevant than something like 13 PMA. It has been shown that cells that are heavily 14 into apoptosis can still have some sort of burst with 15 PMA. And we have been looking at viability assays 16 through flow cytometries like hypodiploidy assay, 17 which is an assay that looks at how intact the 18 neutrophil nuclei are. And the other one is annexin 19 V labeling. 20 Basically the phosphotidylserine flips out 21 from the inside part of the cell to the outside part 22 of the bilipid layer when the cell goes into apoptosis 23 before it disintegrates. The other thing we have been 24 looking at is CD16. CD16, as Conrad Liles mentioned 25 before, is the FC gamma 3 receptor. It has been shown 245 1 in several studies that disappearance of the CD16 or 2 expression of CD16 in neutrophils seems to correlate 3 with onset of apoptosis. 4 So the series of slides I am going to show 5 you are all from isolated human neutrophils from 6 normal donors. So this is not from apheresis packs. 7 This is actually from isolated neutrophils. And what 8 we see is that if we look at the time course of the 9 onset of apoptosis, you can see here that annexin V 10 without treatment, we actually see very quick 11 expression of annexin V labeling with 12 phosphotidylserine on the outside after 24 hours. By 13 48 hours, just about every cell has gone. When we 14 treat the cell with 1965, this is our sort of 15 shorthand for that azile indole dipeptide, you can see 16 that right after 96 hours we have preservation. 17 We see similar preservation, although it 18 is not quite as spectacular, when you look at CD16 19 expression. So again, by 24 hours you have a huge 20 diminution of the CD16 being expressed in the cell 21 surface of isolated neutrophils. But when you have -- 22 when you incubate these cells in the presence of 1965, 23 you actually see that there is preservation right out 24 to 96 hours where still over 60 percent of the cells 25 are expressing. Just to show you that these numbers 246 1 aren't made up, we have some pretty colored slides 2 here. You can see here that this is the untreated 3 control at time zero. You can see here that there is 4 a nice population in the facts analysis of a 5 granulocyte population. It is expressing a nice 6 homogenous CD16 population, and it is not labeling 7 with annexin V. With time, 24, 48, 72, and 96, you 8 can see there is a very quick drop off of CD16 and 9 very quick expression of annexin V labeling on the 10 outside, and you can see that here as well in the 11 double staining experiments. Can you see it at the 12 back? With the 1965, you see that you do get 13 preservation. It is not just a figment of my graphs. 14 You can actually see that there is quite good 15 preservation here. Although of course you are 16 starting to see the break-up of the CD16 signal here. 17 Of course these are all surrogate 18 endpoints and we wanted to look also as well at 19 surface markers to see what do the nuclei of these 20 cells look like. So we did the hypodiploidy assay. 21 This is basically an assay where you, after certain 22 time points, you allow PI or propidium iodine to enter 23 into the cell and label the DNA of the cell and then 24 what you are doing is you do flow cytometry to look at 25 the cells or the nuclei is still attached as opposed 247 1 to a nuclei that is chopped up. You can see that out 2 to this time point, the 1965 protects the integrity of 3 the nuclei. The fmk has some slide protection, but it 4 is not that great, and you can see how this falls off 5 when it is untreated. 6 Perhaps more important is what is 7 happening to the actual functional part of the cell. 8 So we -- this is the oxidative burst, what happens in 9 the oxidative burst as I say in response to opsonized 10 zymosan. I think it is probably the most relevant 11 since in order for that to form a response, you have 12 got to have receptors being expressed -- SC receptors 13 being expressed on the outside to capture the cell and 14 to capture the opsonized zymosan. The cell has to 15 have the ability to restructure its cytoskeletan and 16 take in phagocytose and then it has to have a 17 response. So I think it is a very good way of looking 18 at the viability of a cell. You can see that the 19 ability to burst, isolated neutrophils fall very, very 20 rapidly after 24 hours and Z-VAD-fmk has very little 21 protective activity. The 1965 has protective activity 22 out to 48 hours, which is actually quite good. By the 23 way, I am just showing you -- we have a lot more data 24 than this, but obviously we just wanted to show you 25 representative samples. This is actually at 50 248 1 micromolar. 2 The IC50 of 1965 for 48 hours is about 10 3 micromolars. So at 10 micromolar, you are still 4 getting 50 percent protection. There are other 5 compounds that seem to be working even better. The 6 1965 is obviously one that we can show you the 7 structure of. 8 So not knowing very much about 9 leukopheresis or really about neutrophils, we decided 10 to do something really naive and just put some of 11 these compounds into an apheresis bag thinking we are 12 on to a winner here. If it works on the isolated 13 neutrophils the way that we are mistreating them, it 14 is bound to work in the bag. We set up a series of 15 assays and in particular set up a series of flow 16 cytometry assays looking at these different markers, 17 just so that we could make sure that what we are 18 looking at are neutrophils. So in each case we are 19 labeling with CD3 to identify lymphocytes and CD14 to 20 identify monocytes, although CD14 is expressed in 21 neutrophils, it is actually the percentage of 22 expression is a lot, lot lower than on monocytes. 23 CD66B is the expression of neutrophils. That is a 24 marker specific for neutrophils. And then obviously 25 this looking for platelets. 249 1 In terms of functional markers of cells, 2 we looked at CD16 and CD16B. Really, they are both 3 the same marker. The only difference is that CD16B is 4 a isoform of the CD16. It is actually specific for 5 neutrophils. So that is what mainly we are going to 6 be talking about, CD32. And then most importantly 7 also is CD62L, that is L-Selectin. What we have seen 8 is that when a cell is activated, it loses its L- 9 Selectin. 10 So we did a whole series of experiments. 11 Rather than go through all that data, we hit several 12 problems of inconsistency, which is sometimes we could 13 get some of our compounds to work and sometimes not. 14 We hit a whole series of problems. And I think a lot 15 of it were also as we went along what we found was 16 that we had other issues other than just stopping 17 apoptosis, which is a lot of these cells were actually 18 not dying of apoptosis I think in the first place. 19 But problems that obviously you are all aware of, but 20 we weren't aware until very recently, which is 21 problems of actual mechanical storage of the samples. 22 Conrad Liles and other people have obviously gone into 23 this. These are things that we started to sort out on 24 our own. 25 But really the problem was that we 250 1 couldn't get through the very first hurdle, which was 2 this. We were actually collaborating with our local 3 blood bank, who I don't think really have -- I am not 4 sure how much experience they have on granulocyte 5 apheresis products. But this is the first hurdle we 6 hit and why we can't give you a straight answer at the 7 moment as to whether these compounds work. This is at 8 time zero. We actually went through -- I think these 9 end up being 11 different leukopheresis packs tested 10 on these days with these different treatments. 11 Unfortunately, our local blood bank also didn't have 12 a G-CSF protocol that we could use. 13 So these are the different treatments and 14 these are the different dates and here is the problem. 15 If you look right off you can see there is a huge 16 variation from week to week on the cell count that 17 they are actually able to give us. And then even more 18 so the neutrophil count. So the total cell count and 19 the percentage of neutrophils collected in each time 20 point is totally different. 21 This was a short working day. I am not 22 quite sure how that happened but on the 11th of 23 February we actually got a sample which had no 24 neutrophils in it whatsoever. The other problem -- 25 even on the days where we actually had cells and even 251 1 better the days we had cells in the neutrophils, we 2 found that looking at CD62L there is a huge variation 3 in the percentage of cells that are actually labeling 4 with CD62L, which means the state of activation of 5 those cells is totally different. We haven't had a 6 chance to look into this, but I am sure that the state 7 of activation of the cells varies or influences 8 dramatically how the cell will survive and whether the 9 cell is going to go into apoptosis or not. 10 Lastly, the viability of the cells -- 11 well, at that particular time, there is not that much 12 difference. But I am not quite sure how important 13 that is or how much it is going to vary. Really it is 14 the fact that the number of cells we are collecting 15 and the site of activation of the cells varied so much 16 that we can't really go any further with that 17 particular collaboration. 18 So we were hoping to come here so that we 19 could get answers like how consistent and we have been 20 pleased with some of the responses here which is how 21 consistent can we expect the neutrophil apheresis 22 product to be. It is obvious that it can be a lot 23 more consistent. And then also what state can we 24 actually expect them to be by the time we get them. 25 I think the other thing that will be very 252 1 important to us is I think it is going to be very 2 difficult for us to do anything with apoptosis 3 inhibitors until we have a system that people agree on 4 of the best way to store the cells and then we can go 5 on from there. Because otherwise, we are going to be 6 trying to interrupt a death process that is more than 7 likely not going to be apoptosis but some mechanical 8 death due to pH, et cetera. 9 So what next? As I say, this is basically 10 summarizing what I just said. How representative is 11 our source? It looks like it is not very 12 representative of what other people are getting. 13 Certainly not -- we have seen some of the data in 14 terms of consistency historically that Susan Leitman 15 has shown us and that is amazingly different and 16 encouraging. 17 Very lastly, just a slide of who actually 18 really did the work. The work was actually done by 19 Theresa, Steve, and Shannon in terms of the science 20 and Karent Valentino and David Higgins, who collected 21 a lot of the data and who are our clinical development 22 people and got us in touch with everyone. I sort of 23 mainly drank coffee, discussed UNC basketball and 24 provided amusing anecdotes. Thank you. 25 CHAIRPERSON HARVATH: Okay. Dr. Ambruso? 253 1 DR. AMBRUSO: I guess this microphone is 2 on. Dave, what kind of bags did you use to collect 3 your samples in? 4 DR. STRONCEK: We collected them in 5 originally the bags that come with the kit and then we 6 transferred. We tried in larger bags, some live cell 7 bags, and teflon bags. Live cell is a Baxter bag that 8 is supposed to be more breathable, and it didn't seem 9 to make much difference with storage. We also tried 10 teflon bags because we wanted the smaller bags, some 11 with about a 30 or 40 ml capacity so we could aliquot 12 these products and try some different storage 13 conditions. What was available was from a local 14 company here that will manufacture teflon bags in 15 about any size we wanted. So that is why we picked 16 those. 17 DR. AMBRUSO: These are not appreciably or 18 don't allow enough gas exchange perhaps like 19 platelets? 20 DR. STRONCEK: Well, the teflon bags are 21 supposed to be extremely breathable. They are 22 supposed to be better than some of the other bags. 23 DR. AMBRUSO: You didn't see -- I mean one 24 of the ways to approach this might be if you allowed, 25 as with platelet storage, if you alloed CO2 to diffuse 254 1 out. Maybe that might help with the pH. 2 DR. STRONCEK: Yes. That would be one of 3 the things to try. 4 AUDIENCE MEMBER: Hi, Dave. I don't know 5 much about neutrophils, but if given a carbon source, 6 will they take one that would go into the Krebs cycle 7 rather than glycolysis? 8 DR. STRONCEK: I don't know. Do you mean 9 what if we incubated them in a CO2 incubator at room 10 temperature? 11 DR. TORLINI: Or acetate. People are 12 looking at platelets for acetate. So you don't make 13 an acid by-product. 14 DR. STRONCEK: That would probably be 15 worth a try rather than trying to add or dilute out 16 the products. 17 AUDIENCE MEMBER: Yes. That is what I was 18 thinking. 19 CHAIRPERSON HARVATH: Okay. I would like 20 to thank everyone who participated in this conference 21 and all of those of you who have stayed here to the 22 closing moments of the conference. I hope our 23 colleagues at NIH who were here to hear about the 24 areas that are in need of research support in this 25 area heard what all of you had to say. I know that 255 1 one of the reasons for having this meeting transcribed 2 was so that we could have a record of the current 3 thinking in this area. And also we look forward to, 4 I think, continuing the dialogue and looking at the 5 progress in this area, particularly those of you who 6 have already discussed your intentions to pursue a 7 multi-center type of study design to look at these 8 basic issues. I don't know if Dr. Snyder is still in 9 the audience. Is he here? No. Okay. I would like 10 to also thank him for his comments and input during 11 the meeting as well and some of the suggestions that 12 he had. It has been a long couple of days. We have 13 heard a lot and I want to again thank all of you on 14 behalf of the organizing committee and let you enjoy 15 what is left of your Friday afternoon. 16 Thank you. 17 (Whereupon, at 3:41 p.m., the workshop was 18 concluded.) 19 20 21 22 23 24 25