SLIDES & TRANSCRIPTS
Wednesday, September 15, 2000

Angiogenesis/Immunotherapy Breakout Group Summary: Therapeutic Directions
Everett E. Vokes, MD
Joan H. Schiller, MD

DR. SAXMAN: We are going to start off this morning with the angiogenesis immunotherapy breakout group. The Chairpersons for this group are Dr. Everett Vokes, who is a professor in the Department of Medicine and Radiation Oncology and Director of the Hematology Oncology Section at the University of Chicago, and Dr. Joan Schiller, who is Professor of Medicine at the University of Wisconsin Comprehensive Cancer Center in Madison.

DR. VOKES: Good morning. We had about a 4-hour meeting yesterday, focusing on what were 29 questions, 13 on angiogenesis and the remainder in immune therapies for small cell lung cancer. It was a very mixed group of discussants that I think very well represented both the basic science and the emerging clinical issues in this area.

I am not sure that at the end of the day we had answered specifically all these questions. What we have instead are some summary conclusions of where we think the field stands and our recommendations for how clinical research ought to be pursued from here.

TOP

So the first discussion centered on whether or not small cell lung cancer is a reasonable target for antiangiogenesis. It was felt that in general it is an excellent target because it is highly angiogenic as a disease, and since it is a fast-growing tumor it can be hypothesized to rely on angiogenetic mechanisms fairly extensively.

It was also felt that we have effective early therapy and that clinical trials would likely focus on minimal residual disease. For that setting too, angiogenesis is relevant biologically and clinically.

Looking then at some of the planned Phase I studies or even other Phase II and early clinical trials, an issue that came up was, "What are the appropriate intermediate end points?" As many of you know, some of the early Phase I trials have been very heavily loaded with intermediate end points so that, aside from a typical administration of the drug in measuring clinical toxicities and efficacy, there has been heavy emphasis on trying to understand how these drugs work in people and trying to measure other intermediate end points, realizing that those by themselves would also be experimental since they are not verified. What we are listing here is what is actually being done in some of the current trials. I want to emphasize that we don't recommend that this be done in all studies.

This is a menu of intermediate end points that we think need verification that should be part of one or two trials. As you will see, we also recommend that many other trials be done that would be less heavily focused on intermediate end points and much more on traditional clinical end points so that experience with these drugs could be broader and more rapidly obtained.

So this is the list, and we can go over this. This is largely put together with the help of Joan Schiller, whose institution is about to start a Phase I trial with endostatin and Jim Pluda. What you see is PET scanning or rapid MRIs (magnetic resonance imaging), with a focus of measuring change in blood flow, microvessel density as a well-established indicator of prognosis, as a second end point that could be measured in tumor specimens, and then systemic measurement of MMP activity, basic FGF and VEGF levels in urine or serum.

Paul Bunn brought up the potential that there may be array patterns that can measure angiogenic genes and we are not sure where the status of that is, but should that be available, this would be an elegant and rapid way of looking at multiple expression points. So we would look at array patterns before and after treatment. Here it was emphasized that we would have to look at multiple tissues because different tissues might actually have a different response in angiogenic propensity and would have to be evaluated separately.

Interference with wound healing and skin biopsies were recommended, and then we had a discussion about the matrigel assay and Beverly Teicher is familiar with this. This is a foreign substance that is now studied in mice and induces an angiogenic response that can then be quantified and measured after a week or two. An issue that was brought up was, "Could one put this kind of substance or foreign body into a human and then do the same experiment before and after a therapy?" and then, finally, nitride urinary excretion as a potential new test in direct measurement of angiogenesis.

TOP

So these are some of the intermediate end points we came up with that should be pursued and that may be of use, although again, we would caution that they not be done in all trials, but as proof of principle be incorporated into selected Phase I trials.

We want to then stress again that different tumors may have distinct microenvironments and therefore that as these drugs go into Phase II testing they might need separate evaluation within each tumor and maybe even normal tissue organ.

Endometriosis, finally, is just mentioned here as a possible model to study angiogenesis.

TOP

So the second large focus of our discussion was optimal clinical trial design, and I covered some of that. Clearly there will be extensive Phase I studies looking at a large number of intermediate end points. These are going to be extremely expensive to do. They will be very slow to perform because you will need to do a lot of additional end points and will have slow accrual, although with endostatin maybe not after everything that has been said in the press. Here were then some models where we thought a more simple traditional evaluation could also be pursued.

So we would consider a Phase I study with classic end points and maybe a large 1b component focusing on a higher dose range in patients with refractory small cell lung cancer to specifically evaluate activity in that setting in a preliminary fashion.

One would then go to extensive disease and non-extensive disease, measure complete response rate and time to progression, either doing this early on with, or right before, standard chemotherapy, or after a maximum response has been obtained. Looking at time to progression, one would also look at pattern of failure. Is it new lesions or old lesions growing back?

Then if enthusiastic results were seen here, it might be possible to propose a more rapid accelerated drug development schedule. Otherwise we would as a last resort go the marimastat route, that is, looking at limited stage disease and then have a randomized evaluation, plus/minus the drug on a very large scale. Clearly however that is our least immediate and favorite option.

TOP

Finally, we focused on what other kinds of agents are available. Many of these are either in early Phase I or sometimes further along. There is the anti-VEGF group including anti-KDR by Imclone, two drugs by Sugen, 5416, which is in clinical Phase II testing right now. There is also an oral drug coming along that has a broader range of targets and of course, greater convenience of administration if it is going to be given orally. Angiozyme and another VEGF antibody coming up for Decision Network review.

We looked at MMPIs. Obviously there are several, and while there has been some skepticism about this class of drugs in general, it was pointed out that there seems to have been a randomized trial in gastric cancer that has been completed that is maybe not entirely negative. So we will need to look at the data, but there may be a survival difference, although at a different time of evaluation than what the initial statistical section called for.

There are other agents, thalidomide, endostatin, shark cartilage, the anti-integrins and TNP470.

TOP

Then finally looking at a few additional issues that we discussed, one was whether the old drugs should be reinvestigated, such as suramin taxol, and perhaps other agents with antiangiogenesis as an end point. While that seems reasonable to do, I am not sure that any of us were convinced as to whether that would allow those drugs be more effective than what we already know them to be. So it might be of interest to pursue this at a scientific level, but we would not necessarily expect much added clinical benefit from this.

Studying this class of drugs for cancer prevention was discussed. There is angiogenic dysplasia, which could be a potential model. It was pointed out that early or preclinical, premalignant stages of disease might be susceptible to treatments with MMP inhibitors or other drugs coming out of this class and might be a new lead for prevention efforts in the future.

TOP

So that was it for angiogenesis, and Joan is going to present the immune therapy discussion.

TOP

DR. SCHILLER: As Everett pointed out, the first half was spent on angiogenesis, and the second half was spent on the discussion looking at or exploring immune therapies.

The first question on the list and so therefore the first question that we really asked is, "What antigens might be appropriate to explore for small cell lung cancer?"

We all agreed that the antigens that we know the most about were the GD2 and GD3 antigens and clearly thought that those continued to deserve further exploration.

A whole class of antineural antigens were also pointed out, including the HUD antigen, and everyone thought these were very exciting and interesting. The problem is that at this point there are no available antibodies or vaccines for them, although we thought that those clearly needed to be developed.

Bombesin, of course, is another possible target, and then there was a whole slew of antigens that have been identified through the small cell lung cancer antigen workshop, including SM1 and on the cluster 5a.

p53 of course is a possible antigen. Work looking at a p53 vaccine for non-small cell lung cancer is currently being done by Dave Carbone at Vanderbilt University. That work is underway, and we thought that needed to be completed before looking at it for small cell. Then a question was raised about the possibility of alloantigens and in particular, after induction with chemotherapy and radiation therapy, looking at the allogeneic mini-transplant.

So the feeling was that there are a number of possible antigens that are possibly worthy of exploration. Really the only one that we have a lot of information on ready to go into the clinic and into clinical trials are GD2 and GD3.

TOP

The discussion centered on what we should be doing to move the field forward, and clearly one way was to think about enhancing immune responses. Ways of doing that of course, include looking at different adjuvants - which is being done - as well as looking at multivalent vaccines combining two different antigens, such as the GM2/GD2 vaccine that Bristol is exploring.

Other thoughts were to look at the mechanism of immune mediation with vaccines. Many times we assume that they are B cell related and enhancing humoral responses. There is some data to suggest that some vaccines may also be acting on CD31 resorpted T cells.

In general, the vaccines that have been used have been primarily B cell vaccines looking at the humoral response. It was pointed out that maybe, in addition to that, we should be looking at T cell vaccines, ways of optimizing the T cell response or ways of enhancing NK and/or macrophage activation. This hasn't been done for small cell that we are aware of.

Another point that was brought up was that not many, but some, small cell lung cancer patients are cured of their disease. Perhaps we should be studying these patients more to try to determine what the differences are between this group of patients and the unfortunate group of patients who are not cured.

This group of patients also would be potentially a group of patients to look at prevention of second malignancies. Clearly they are at a very high risk of developing second malignancies, and this might be a target population for that approach.

The other discussion that followed was when to combine these vaccines with chemotherapy. Intuitively we assume that chemotherapy decreases the immune response. There is some data to suggest that giving the vaccine up front in combination with chemotherapy might be a better approach. Lastly some other agents were discussed, such as IL-12. I think the feeling with IL-12 was that because of the toxicities and scheduling difficulties this was probably lower on our list of things to explore.

TOP

Clearly the potential problem associated with this approach for small cell lung cancer was the difficulty of obtaining tissue, not only because of the fact that so many patients now are diagnosed with fine needle aspirations, but even with those patients in whom we are able to obtain the actual tissue biopsy very often that tissue is not in good shape. The tissue is often very necrotic and therefore difficult to evaluate for immune response type of studies.

Another question that was raised is that we sometimes assume that the ability to mount a good immune response is predictive of a good clinical response. It was pointed out that we don't necessarily know that. It could be that patients who have a good clinical response are the ones that are then able to mount a good immune response. Lastly, another problem with small cell lung cancer for this type of approach is that rapidly proliferating tumors are usually poorly immunogenic.

TOP

In terms of clinical trial design issues, one thing that was mentioned was that perhaps for some studies we should be mandating that patients going on study in fact have a tissue biopsy.

We realize that following treatment it may be impossible to get tissue. But clearly, if this was built into a clinical trial, it was felt that many patients would be willing to participate, particularly if there was funding associated with the patient care costs.

We all felt that the optimal setting, however, is probably going to be in the setting of minimal residual disease. Intermediate end points that we discussed were humoral responses looking at antibody levels. Another intermediate end point that was brought up was looking at circulating tumor cells.

Tony Elias at Dana Farber has an assay looking at this and has found that a very high proportion of patients who were thought to be in clinical complete response in fact still have circulating tumor cells. That also might be a good intermediate end point.

How does one identify the best schedule and dose? We all felt that the way to do that was by optimizing the immune response and then utilizing that schedule and dose. We all agreed that the major end points for this type of approach are going to be survival and/or time to progression. It was felt that a tumor response in patients with big bulky tumors was probably asking too much of the immune system, particularly given our current state of technology. The next question that was asked was how to decide when a particular vaccine is ready for a clinical trial. It was felt that it is necessary to optimize the vaccine as much as possible in preclinical and perhaps Phase I studies before launching a big Phase II or Phase III study.

I think that is how we left it.

TOP

Discussion

DR. SAXMAN: We have time for questions or discussion, and if you don't mind, allow me to go first. Perhaps this is a question less for the Chairperson and more for Beverly and Paul in the audience. One of the things we didn't really discuss yesterday but I was curious about: Is there any reason to believe that these two types of interventions may be used in combination and have synergistic activity? For instance, Beverly yesterday showed data that antiangiogenesis agents increase the influx of chemotherapy into the tumor. Is it possible that angiogenesis or antiangiogenesis therapies might also increase the delivery of antibodies or the influx of T cells or immune cells or antigen-presenting cells into that area? Has that been studied, and is that something reasonable to consider?

DR. TEICHER: Certainly in the laboratory Roquesh Jayne has shown that if blood vessels can be made lengthier there is better delivery of protein therapeutics. Reducing interstitial pressure in the tumor allows a greater delivery of macromolecules. There are other potential targets in the angiogenesis area that might also increase the immune response to vaccines. That is for example, inhibiting or neutralizing TGF-beta which is another potential target in that area, which may have some antiangiogenic effects but might also reduce suppression of the immune system in cancer patients and increase the effectiveness of vaccines. So yes, I think those are very compatible therapies.

DR. CHAPMAN: In terms of vaccines, I am not sure that this is our major problem. If you look at some of the T cell vaccines in melanoma, unfortunately you can get a lot of T cells into the tumor and it doesn't seem to do anything. An area where this approach may be useful would be in passive immunotherapy where we are infusing either regular monoclonal antibodies or radiolabeled monoclonal antibodies. This is an area where access to the tumor is a big problem for a lot of the reasons that Dr. Jayne has elucidated over the years, and so that might be an interesting combination in the future.

DR. BUNN: The NCI always wants to know how it can help. So I have three suggestions for angiogenesis inhibitors. One is to help companies develop commercial assays that would allow one to measure the serum levels of basic fibroblast growth factor, VEGF, angiopoietin and so on. This could be done with some kind of kit that could be done with antibodies and beads. It could be done rapidly and could be useful for all these trials. People laugh about arrays because they are so expensive and all that, but you ought to be able to figure out which genes you want to modulate in some in vitro assays. So it is not 20,000 genes you probably need to look at. If you knew 1000 genes the NCI could make the gene chips available at a reasonable rate, and you could find out whether in a person the same thing happened that happens in whatever your model system is. You could argue that compounds have different mechanisms and might be different genes, but I think there is a limited number of genes. If you have to buy 20,000 gene chips and it costs you $5,000 an experiment, it might be a little bit impractical. It seems to me that the NCI might help with that. The other place where the NCI might help is these matrix gel assays. If they work in animals they ought to work in people. There are lots of questions that were brought up about whether the FDA would be allowing this, and this is in an area where perhaps the NCI could have some discussions with the FDA about whether this type of assay could be developed in people.

I think all those assays would be really helpful for developing this whole class of compounds and might expedite trying to figure out what dose we ought to be using of each of these agents because it is otherwise hard to determine the dose.

DR. SAXMAN: There aren't any FDA representatives here but there are people who are in cancer diagnostics at the NCI. I don't know, Tracy, Sheila, would you like to say anything about that?

DR. BUNN: The idea is that you would inject matragel into patients.

DR. SAXMAN: I was talking about the first two things that you had mentioned.

DR. TAUBE: We are in the process right now of revamping the entire program to rapidly bring new tests to the clinic, and so you will be hearing more about that over the next few months in terms of the kinds of services that the NCI will provide for both the industrial community as well as the academic community.

DR. BUNN: I had a comment. For the angiogenesis inhibitors in the small cell patients, regarding the standard Phase I trials: there might be a bit of a wrinkle. One group of standard Phase I patients would be people who relapsed after some standard treatment, and certainly one could study those patients, but there is another group of patients. ECOG showed that you could give ineffective treatment to patients for up to 6 weeks without hurting them, as long as you give them active chemotherapy after 6 weeks. So you could actually envision a study in which for 6 weeks people are getting the angiogenesis inhibitor alone. You could concurrently do a variety of toxicity and pharmacology studies.

After 6 weeks, in the absence of toxicity, you could keep going, add the drugs in, see if there is any pharmacokinetic interactions of carboplatin and etoposide and at the same time see what the response rate is. You would know if anybody responded in the beginning, but you would also know what the response rate is afterward to see if it looked like the standard etoposide carboplatin combination. So I think you could do Phase I studies in untreated patients which would be rather unique.

DR. B. JOHNSON: I have a question for Everett. Of the intermediate markers that you proposed for these antiangiogenic agents, how many of them have been shown to be elevated or detectable in small cell lung cancer patients?

DR. SCHILLER: That is an easy question because I think the answer is probably none. Most of them haven't even been validated at all as having clinical meaning. We know that for VEGF, for example, or beta FGF it can go up or down but I don't think we are at this point aware of the significance of those end points, or how to interpret them.

DR. DENNIS: Could I make a scientific comment regarding this because part of my graduate work centered on the control of FGF family members by serum proteins? I just wanted to make a comment in terms of that particular end point. Many growth factors are often bound to other proteins that are quickly cleared from the circulation, probably serving to limit the effects of growth factors locally so that a newly forming capillary would be exposed to high local levels of basic FGF, perhaps released from the extracellular matrix or increased local levels, probably with TGF beta as well, so that you may get profound effects of angiogenesis without perhaps getting changes in serum levels of these proteins. I don't know about the binding proteins for VEGF in serum, but I know that you may be misled if you think that you are getting no changes in FGF levels. You still may be getting profound effects in angiogenesis without the serum levels changing very much.

DR. VOKES: So I think that that is a very important point, and something we want to emphasize that traditional clinical endpoints have a role here because a lot of the intermediate end points are not verified and are experimental in themselves. So we advocate that they be pursued, but we also think if therapy is going to be meaningful, then somewhere in a patient we will hopefully be able to see the difference.

DR. B. JOHNSON: A follow-up on that, getting back to something more fundamental: Can you comment about the presence of either the factors or the receptors and their proportion on small cell lung cancer cells that you think are important for this pathway, VEGF receptors or FGF receptors?

DR. VOKES: On small cell lung cancer cells I don't think so.

DR. BUNN: Can you comment on whether you want to see them in the endothelium or you want them in the tumor?

DR. B. JOHNSON: Either one.

DR. SCHILLER: Our conclusion that small cell was a good target for an angiogenesis approach was mostly based upon a clinical scenario of rapid metastases as contrasted to some biological data demonstrating that small cell was uniquely sensitive. Would you agree with that?

DR. BUNN: And high microvessel density counts in small cell. There are studies showing that has prognostic relevance. That is really the scientific rational.

DR. LANGER: With respect to detecting the circulating tumor cells, were there any specific markers that were identified, or would it be relying on something like a cytokeratin?

DR. ELIAS: We are actually working on that, looking at cytokeratin alone. That is the bulk of our work. We are also looking at cytokeratin plus anti-GD2 and anti-GD3 in particular. It is an immunofluorescence semi-automated system that we are looking at and has sensitivity, at least in breast cancer, where it has been tested in other laboratories at around one tumor cell in a million. So it is equivalent to other assays.

Obviously that is one area that is important, because not all small cells are going to express high enough quantities of cytokeratin. Second of all, we are more interested in when you show cells that are dual stained those are much more likely to represent tumor cells as opposed to false positives and so forth. So we quantitate the cells in bone marrow of limited stage patients in or near complete response, and 75 percent will be positive for keratin. Of those, one-half - or more than 50 per million - are well outside the range of equivocal, namely our normal controls. One-half are, however, low positive and could conceivably be within the range of false positives.

DR. MABRY: Going back to Bruce's point, does that mean that we should consider doing xenograft models with agents to demonstrate that there is either a four-fold growth delay or a tumor regression before we go forward?

DR. VOKES: The panel didn't address that. Beverly do you want to comment?

DR. TEICHER: I think we do that with these agents. I think most companies do have xenograft models. There are quite a few for small cell.

DR. GAZDAR: Would the fact that the vessels in the xenografts are of mouse origin, would that have any effect?

DR. TEICHER: That has an effect with some of the agents. Of course, that was one of the big issues with Vitaxin, in that murine vessels didn't respond. They have since found a way around that. There was the implantation or engrafting of tumor skin onto the nude mouse or SCID mouse and then implantation of tumor into that. That allowed a vasculature that was human or part human, part murine to develop. So there are model systems that can address most of the issues involved in angiogenesis. Of course, you still have the issue that the drug metabolism is part of the kinetics of mouse inoculum, but you can do human vessels, and you can do human tumors in preclinical models.

DR. SCHILLER: But I think it was brought up yesterday that there are problems with mouse models, including the fact that depending upon the site where the metastasis is located, the tumor bed may be different. Models are models, and certainly in the mouse you can implant the same tumor cells in different locations in the mouse and get very different responses to therapy, just the same as it would be in a human patient. So you can model that heterogeneity.

DR. LANGER: Getting to a very pragmatic issue, I would like to hear from clinicians, people on the panel who see a lot of patients. In terms of best investigating these agents, Paul had mentioned looking up front. While that would be an ideal approach, I think we all are worried about the pragmatics of that and the acceptance of that among other investigators. Do you think the model that has been used with anti-VEGF in non-small cell, chemotherapy with or without drug for at least the initial Phase II forays might not be better? Then given this panoply of agents which seems to be ever expanding on a weekly basis, how do we as clinicians figure out which one is worth investigating?

DR. BUNN: Everett's second point was to give the drug plus chemotherapy in untreated patients. That was the second group of patients, the standard group of patients. Before you do that, you still have to do the pharmacokinetics of the experimental drug plus the standard drug to make sure they don't interfere with one another's pharmacokinetics. The proposal was actually to do the Phase I of the drug alone and the Phase I of the drug plus chemo in the same patients. You give the study drug alone for 3 to 6 weeks and then you give the study drug with chemotherapy for as long as the chemotherapy works. Where are the ECOG people? You have done several trials. Were there accrual problems in your trials where people were getting experimental things up front?

DR. D. JOHNSON: We appreciate you pointing out the failures of ECOG.

(Laughter.)

DR. D. JOHNSON: We proved lots of drugs don't work. I don't see any reason why we cannot prove that these don't, too.

(Laughter.)

DR. D. JOHNSON: You are right. In all candor, I agree with you, Paul. I don't see any reason why these agents cannot be used, not in limited stage, but certainly in the patient population that has been used in the ECOG trials and the NCI has done the same thing. In extensive stage patients who had good performance status who did not have life-threatening organ involvement in whom we knew the standard therapy would not be curative, and therefore we felt it was ethical to treat those individuals with new agents, and we have clearly used drugs that had zero activity. Their survival after salvage treatment with "standard therapy" was identical to patients up front given the exact same agents in a randomized fashion. So it wasn't really a total patient selection phenomenon. I actually think that is a good idea and prefer it over some of the other options that have been used.

The other reason why that might be a good patient to try this on is that, although tissue is extremely difficult to come by, if there is ever going to be an opportunity to do so pre-and-post some treatment. This is going to be the group of patients in whom that is possible, and the microarray approach may be in what little I know about it may be the way to analyze those tissues.

DR. G. JOHNSON: I wanted to touch base on this microarray question that was mentioned in what Paul talked about and one of the things that the NCI can really do and I think all investigators here. The SPORE is trying to have four groups in this, and that is the work of the NCI for different organ systems to make diagnostic arrays of say 500 to 1000 genes. If the working groups can come together and identify those genes and then have the NCI contract with someone to make those last five, if you make them in large volume, they become very inexpensive. All you need is a resonance analyzer to be able to do it, and everybody has standardized arrays to be able to do this. If NCI would do this and make them available to investigators, it would be a way to standardize a lot of the analysis, and it would make it easy for people to exchange their tissues because they would be analyzing the same slides. I think this could be something that could really accelerate this kind of diagnostic approach.

DR. B. JOHNSON: I have a question for Joan. A lot of our epitope-directed therapy in small cell lung cancer was kind of backed into for melanoma, and that is, other than the P53 vaccine that is in trial, we really don't have too much specifically directed towards small cell epitopes on the surface. I wondered if you want to comment about the cluster workshops. Are there better epitopes that we could be directing the small cell therapy to? Because although in melanoma there is a lot of work, there are lots more cases of small cell lung carcinoma than there are cases of melanoma in the United States. Because it has so many humorally mediated paraneoplastic syndromes, this may be something to get at.

Do you want to comment about other potential targets in small cell lung carcinoma that we should be working toward?

DR. SCHILLER: I guess I am not sure that I am the best person to comment on that, and I would certainly --

DR. LAIRD-OFFRINGA: I have several comments related to that. The first one relates to the anti-Hu antibodies. There are antibodies. The proteins are very immunogenic. In fact, a human antibody has been cloned out of a patient, a paraneoplastic patient through phage display. This was the work of Dalmau and colleagues, but it is not clear in this disease whether the antibodies would actually help because you can immunize animals to get very high titer, and it is not clear whether this will help the disease. In neuroblastoma, which is the other disease that has Hu proteins in it, there was an interesting paper, also by Dalmau, where they immunized animals. They treated them with antibodies. They immunized them with protein, and they immunized them with DNA, and the only case in which they saw effect was when the animals were immunized with the DNA encoding the Hu protein.

There are other possible epitopes that could be looked at, and there is very interesting work from Alan Epstein at USC, who has this great idea for general epitopes in cancer. He has antibodies that are directed against necrosis, for example, double stranded DNA, and I think for small cell lung cancer this would be great because there is a lot of necrosis going on. So what he does is: he couples the antibody either to a radioactive component or some other molecule, toxic molecule, and the antibodies go to the necrotic center and from that point actually affect the living cells that are close. Over the course of several treatments, the area to which the antibodies go expands. I think that this is a very important area to think of what could be done in small cell. So this is very exciting work by Alan Epstein, and I am sure that there are numerous publications about that coming out.

DR. BUNN: Two comments about the HuD. The first is that the good thing is that the animals didn't get any neurologic toxicity, and they looked at the brains of the animals. The other thing is that there are two papers where people develop anti-HuD antibodies do better than people who do not. Yesterday we had the anecdotes of the GD2 vaccine patients; the ones that developed antibodies seemed to do better. So this is a little bit of suggestive evidence that it might actually be useful to develop such antibodies. I think those were the reasons that HuD was picked out as a potential target. But even though people have the DNA and so on, is there a company? You see, what you need is a company that is going to produce the stuff that we can give to people?

DR. LAIRD-OFFRINGA: I am not aware of this company right now, but an antibody has been cloned. It is not clear at this moment whether the antibody part of the immune response is the one that is protecting the patients. It could be a different part of the immune response.

DR. BUNN: With the DNA vaccine, was there any T cell response from those studies, or was it all humoral?

DR. LAIRD-OFFRINGA: I don't recall. I would have to check.

DR. D. JOHNSON: Do we know in those patients that those that have anti-Hu do generally have limited disease?

DR. MINNA: The answer is yes to your question. They did have a T cell response.

DR. D. JOHNSON: They are also usually women. They are not always, but they are more commonly women than men, a very high percentage, which is almost the reverse of what one sees with small cell in general. Is there any other factor there that comes into play other than the anti-Hu? I would be a little nervous about giving anti-Hu even if the mice were --

DR. MINNA: One interesting thing is you know whether or not Hu-D molecules would have mutations. I mean you get antibodies against p53. It has mutations and so there have been several groups including our own who have studied these and have not found mutations, but there is no question these are strongly immunogenic molecules. I must say that it is my understanding - in contrast to Eaton-Lambert, which is an antibody developed against a voltage gated calcium channel subunit - while there are antibodies, the anti-Hu antibodies, there are T cell infiltrates that the Memorial Posner group have shown are involved with neurologic disease. It is a T-cell mediated paraneoplastic, but I think any clinical trials with these agents would have to be approached very cautiously.

DR. BUNN: I think that is true for sure, but there are two animal models where they did not develop neurologic disease.

DR. LAIRD-OFFRINGA: One last comment with respect to the limited disease: that is only seen with patients with very high titer antibodies, and actually about 16 percent of the non-neurologically sick or non-paraneoplastic patients have antibodies but they do very poorly. There was a report that when they looked at the expression of the molecules, there was a group of patients that seemed to survive slightly longer if they had HuD expression or if they had the antibodies.

DR. D. JOHNSON: Again, women generally do better anyway, and then if you take limited disease, I am not sure the two are necessarily - they may be true and unrelated is my point, or there may be another factor there.

DR. SAXMAN: Are there any other questions or comments?

Thank you very much.

TOP