Breakout Session B Summary: Immunotherapy, Biologics, and Gene Therapy: Therapeutic Directions by Jack Roth, MD

SLIDES & TRANSCRIPTS
Thursday, June 15, 2000

Breakout Session B Summary: Immunotherapy, Biologics, and Gene Therapy: Therapeutic Directions
Jack A. Roth, MD

Slide 1:

DR. ROTH: Thank you, Scott. I would like to express my gratitude to David Parkinson for filling in for David Johnson. He really did an outstanding job. So I do owe you one, David.

Our session was to cover three very broad topics, immunotherapy, biologics, and gene therapy. The participants realized that much of the discussion in this session would probably tread on the areas discussed by the other workshop groups, and for that we apologize, but we went ahead and did it anyway.

Unlike the previous session that was presented by Paul, we really don't have a long list of therapeutic agents that are in advanced clinical trials here. In fact, it's a rather short list. Our charge was to take a look at research that is going on in these various areas, and to try to develop specific candidates, either genes or biologics or immunomodulator agents that would be suitable for bringing into clinical trials; to look at the top candidates there, and also to address some of the issues of clinical trial design, and endpoints as were brought up in the previous discussion of angiogenesis.

As it turned out, we weren't able to achieve a great deal of consensus, but you will see that by the conclusion slide we were able to address some broad issues which we thought were of very high priority for further research development.

Slide 2:

These were three major areas. We had some research presentations in all of these areas, which I will go through shortly.

Slide 3:

There were four research presentations. I think it's worthwhile to just briefly comment on these, because each of them was on subjects and agents that were very likely good candidates for bringing into clinical trials, or are in early clinical studies at the present time.

Dr. Kian Ang from M.D. Anderson discussed the relationships of epidermal growth factor receptor and COX-2 expression to radioresistance, and presented some very interesting data that a humanized antibody to the EGF receptor can actually enhance radiosensitization in cancer cell lines. And he had very similar data with COX-2 inhibitors.

The humanized antibody for the EGF receptor is currently in clinical trials. It has completed Phase II studies, combined with radiation, and also with platinum-based chemotherapy in head and neck cancer patients, and is currently entering Phase III studies in head and neck cancer patients. At the present time there are not ongoing studies in non-small cell lung cancer, but there are plans for studies with both of these agents in non-small cell lung cancer. The panel felt that these were very important targets for clinical trials, and for further mechanistic studies.

Dr. Joan Schiller, from the University of Wisconsin, presented data on her experience and with collaborators at Vanderbilt with an adenovirus p53 construct in treating bronchoalveolar lung cancer. Now this is a construct that is made in an E1 deleted adenovirus, which is replication deficient, and expresses the wild type p53 gene driven at a very high level, with a cytomegalovirus promoter. There have been Phase I studies that have been completed with this in non-small cell lung cancer, as well as head and neck cancer and several other tumor types which we won't discuss.

John Nemunaitis discussed some of his experiences in Phase I studies in combination with cisplatin in non-small cell lung cancer. This was reported to have some anti-tumor activity in bronchoalveolar lung cancer when the vector was delivered by a bronchoalveolar lavage, although there were some issues pointed out with respect to potential toxicity. There were a couple of patients who developed a pneumonitis-like syndrome, although it wasn't completely clear that this was directly related to the vector. Nevertheless, this was an area that certainly merited further study. In addition, it was felt that this is an agent where pathways and interaction with other genes, as well as interaction with other therapeutic modalities such as radiation and chemotherapy, merited further study.

There are a couple of clinical trial proposals for adenovirus p53 in lung cancer that are currently being considered. Dr. Choy from Vanderbilt has proposed a trial to look at direct intra-tumor injection of adenovirus p53 in patients with clinical early stage lung cancer who are not candidates for surgical resection, the idea being to use the ad-p53 as a radiation sensitizer. I believe this study is going to be considered by the RTOG.

There is a second trial that is being considered which will be a randomized study that would evaluate intra-tumoral injection of the ad-p53 plus or minus concomitant chemoradiation therapy. This was based on some observations in a Phase II study that there may be enhanced local tumor control when radiation and ad-p53 are combined together in patients with unresectable non-small cell lung cancer. So that study is under consideration as well.

Slide 4:

Mario Sznol presented some very interesting data on a Salmonella typhimurium bacteria therapy of tumors. This is a very interesting organism, because it seems to target to the tumors and selectively replicate in tumors. Although it was pointed out that the replication appears to be extracellular for the most part, and not intracellular. But this was thought to have very interesting anti-tumor properties in preclinical models, and to also be a very interesting vector for delivery of agents selectively to tumors. This will also be going into early clinical studies.

Finally, David Schrump discussed demethylation and histone deacetylation in the modulation of lung cancer gene expression, and used decitabine and depsipeptide as agents to modulate the methylation state and acetylation state. It is interesting. David pointed out that there were possibly two applications here. One is in the enhancement of antigen expression by lung cancer cells. He has been investigating a novel antigen, NY-ESO-1, which actually is a testis antigen, but is expressed in a fairly significant percentage of lung cancers at a rather high level. This expression can be upregulated by these agents. So we have here a potential for immunomodulation and immunotherapy using these agents. Also, there was observation that apoptosis was induced in non-small cell lung cancer cells as well. So there is a possibility for some therapeutic efficacy. I think this brings up the interaction of these various modalities of gene modality, immunotherapeutic modalities that was brought out in the committee discussion.

Slide 5:

So let's focus on the gene therapy area for a minute. Although we cannot come up with a list of ongoing clinical trials and agents that are ready for clinical trials, what we could come up with is a list of genes. These fall into some major areas such as oncogenes, receptor tyrosine kinases. Now this of course is really a subset of oncogenes, but it was felt by our group that functionally this is rather distinct, and there are agents now in the clinic that directly address the inhibition of tyrosine kinases in cancer cells. So this would be an area for focus.

Tumor suppressor genes merited a fair amount of discussion. Again, p53 is probably the most advanced gene in clinical trials. But it was also pointed out there are a number of other good candidate genes such as p16, p10, and others that would probably be very useful to look at, at least in preclinical models, and potentially have some application in lung cancer.

Cell cycle regulatory genes are another group that certainly merits consideration. Pro-apoptotic or inhibition of anti-apoptotic genes could be looked at. This presents some challenge in the area of gene therapy because for viral vectors, for example, one has to develop these vectors in packaging cell lines. Oftentimes the genes are quite toxic to the packaging cell line. So novel vectors will need to be developed for producing viral vectors at least that express apoptotic genes at high levels. But nevertheless, such genes as BAX and BAC for example, genes that would down-regulate bcl-2 would also be very good potential target genes.

Slide 6:

So the therapeutic approaches that were discussed, and that were thought to be worthy of further evaluation, include gene replacement therapy, where a defective gene is directly replaced in the tumor, or there is some manipulation to up-regulate the expression of a gene that has been down-regulated in the tumor.

Another approach that is not direct gene therapy involves selectively replicating oncolytic viruses. And for the most part, these viruses have been designed to take advantage of the mutations that exist in certain genes within the cancer that normally can prevent viral replication. For example, the onyx 015 virus is deleted in E1B, and when E1B is present, generally functions to down-regulate p53. So in those cells that have a defective p53 pathway, oftentimes one can see selective replication of this vector within the cells.

But there are other strategies for doing this as well, based on selective promoters, and based on inactivating other genes within the adenoviral genome, or using other types of virus. For example, reoviruses were pointed out, which can selectively replicate in those cells which have an abnormal ras pathway.

Finally, another technique of modulation of gene expression which can be focused on looking at demethylation of genes, which might enhance the expression of the appropriate genes within the cancer. So these could all be thought of as related to gene manipulation.

Slide 7:

What are the key research questions here? I think there is enough information available on a variety of genes that could be delivered to cancer cells. But at the present time the major limitation that the group observed was novel delivery systems. We really have to have a way of delivering these genes efficiently to tumors. There are a number of techniques that are evolving now that may be able to accomplish this -- alternating the tropism of viral vectors; developing nonviral vectors which have a minimal immunogenicity and inflammatory response; and liposomes were pointed out as delivery systems which could function very well in this capacity. There are some novel liposomes that have been described to date which may function here; targeted vectors to minimize the exposure of normal tissues; and altering tropism, altering binding of the viral vectors may be very helpful.

But the ultimate goal is systemically delivering these vectors. There are some clinical trials out there with adenovirus looking at systemic delivery, and to date the toxicity has been very acceptable with those. But I think there is also a recognition that this going to be limited by immunogenicity and inflammatory response, so there will need to be vector modifications.

Slide 8:

Other key research questions in this area involve selectively replicating organisms. We pointed out the Salmonella bacteria which appears to have selective replicating properties. The development of synthetic vectors, totally synthetic vectors such as DNA protein complexes which can be engineered with targeted specificity, and also be engineered to reduce any immune or inflammatory response. Also, gene augmented vectors that are designed in fact to augment the expression of genes within a cancer cell might be therapeutically beneficial.

Slide 9:

What other key research questions are involved here? The group spent a lot of time discussing ways to improve and understand the interaction of various genes and these delivery systems with radiation and chemotherapy. I know there is a fair amount of information accumulating on this. It was felt that much more is needed.

It is clear that there can be additive or super-additive interactions, among various tumor suppressor genes for example, and radiation and chemotherapy. Some of these gene interactions can sensitize to radiation very strongly, but the mechanisms behind these interactions are very poorly understood. It's clear that we are going to need to devote additional resources if we are going to get a clear understanding of what direction we need to go to in this area.

Another area that was thought to be key is the area of tumor-specific promoters. One that was pointed out, for example is the human telomerase promoter, which seems to be selectively expressed in very rapidly dividing cells, and selectively expressed in human cancers. This might be very useful for gene delivery. But there are other types of promoters as well that can be examined. One qualification here is that this has to be a strong promoter, and has to have very good selectivity. So this is another area for active research.

Slide 10:

Now with respect to immunotherapy there are also a variety of key research questions. Now here it was really not clear among members of the working group as to what the best approach should be in terms of cancer immunotherapy in lung cancer. It was recognized that at the present time we really don't have a good idea of what we are doing in terms of tumor antigens. This has progressed greatly in the area of melanoma for example, and some other types of tumors, but we are really behind in lung cancer and we need to do further definition. Also, the use of novel whole tumor immunogenic approaches seems to be promising, the idea of using intra-tumor dendritic cells to enhance the tumor microenvironment to be more immunogenic was considered and thought to be a promising area.

Slide 11:

And importantly, we need to develop agents to reverse tumor associated immunosuppression. It is clear that lung cancer patients are highly immunosuppressed, this can be observed even in the early stages of disease. We really need agents to directly address this, to enhance the immunotherapeutic protocols.

Finally, we need to explore the interface between angiogenesis and immunologic manipulations. It is really not clear how antiangiogenic agents may interact in this situation, and whether there is going to be a beneficial or a detrimental type of interaction. And in fact, this goes not only for immunologic manipulations but also chemotherapeutic and radiation manipulations. There really needs to be a great deal of work done to look at potential mechanisms of interaction and for synergies as well as antagonistic interactions.

Slide 12:

Now what about other biologic approaches and other novel targets? Well, here is a list of perhaps the top six targets, not in any order of preference -- just the ones that were thought to be potentially of interest for further investigation. Of course several of these are already in advanced clinical trials, but a number of them have not been looked at directly in lung cancer.

So this includes again the humanized EGF receptor antibody or agents which can antagonize EGF receptor, the VEGF receptor, PDGF receptor. We heard yesterday about fibroblast growth factor, and the possibility of enhancing response to chemotherapeutic agents by manipulating expression of this. C-kit was thought to be another way to target, as was HER-2/neu. So these are all targets that could be potentially looked at in clinical trials.

Slide 13:

Now what about key research questions here? Well, again, a comparison of antibody and small molecule-based approaches would be very important. It is not really clear which approach is going to be superior in terms of anti-tumor activity, or if these approaches might even be complementary, but there is not enough information known about the relevant benefits and risks in terms of using antibodies as kinase inhibitors, or small molecules. Another point that was raised was do we really know the mechanism of these anti-kinase receptor antibodies? And is this primarily an immunologic mechanism, or are we really looking at kinase inhibition here?

Another important area is defining pathway effects for selective target inhibition. It was felt that just focusing on individual pathways would really be less informative at this point, as opposed to developing a modular approach. There are so many agents out there, and so many potential targets, we really need to look at this as a whole, and look at these mechanisms within a modular context, as opposed to an isolated single pathway context.

Finally, defining chemotherapy-radiotherapy interactions, along with signal transduction inhibition is going to be extremely important. Not a lot of information is out there as to the additive effects that might occur with these various agents, and the appropriate studies to look at mechanistic interactions really need to be done.

Slide 14:

Another point that was brought up and emphasized by the group was the critical importance of trial design and intermediate endpoints in all of these studies. We are again faced with the same problem -- what's the appropriate endpoint here? At the end of Phase III, I think everybody agrees that survival is going to be the key endpoint. But in terms of moving from Phase I to Phase II and then to Phase III, what are the appropriate biologic endpoints that we need to look at? These really don't exist for most of the genes, and most of the agents that we currently are studying. We need to find biologic correlates of therapeutic action and clinical effects, which should permit more efficient and accurate decision-making with respect to therapeutic development, which includes dosing and schedule selection, which is very empiric with all of these agents.

Slide 15:

We need to define the contribution of patient and tumor heterogeneity in therapeutic responses. Many of these agents are being tested in a broad spectrum of cancer types and cancer stages. And it is clear that responses are going to vary. So we need to focus more on the influence of heterogeneity in the patient selection, and its effect on the outcome of these trials.

Slide 16:

We need to validate pathologic response and functional imaging as intermediate endpoints. It was felt that this would be a major advance in decreasing the cycle time of therapeutic development -- the type of trial design that Frances Shepherd mentioned, the sort of Phase II to Phase III, which saves a large number of patients and allows you to abort if there appears to be safety concerns was also mentioned. But it was pointed out that although lung cancer is not the most accessible of tumors, there are still good techniques for doing serial biopsies in these patients. This has been done in other studies that were reported and could provide valuable information in terms of intermediate endpoints.

Slide 17:

What are the therapeutic research priorities? These emerged as the top three considerations, again, not in any particular order. Number one, develop a technology platform for vector delivery, to enhance gene delivery and efficiency. Number two, tyrosine kinases as therapeutic targets, looking at their interactions with conventional modalities. This appeared to be a very promising area to which a great deal of emphasis should be given. And finally, a clearer definition of complex molecular pathway interactions, and determining their relevance to these various therapeutic effects.

Thank you.

Slide 18:

DR. SAXMAN: We can open this for discussion. Can I just start with a question. Can you just describe a little bit more what you mean by technology platform?

DR. ROTH: That's kind of a catchall term I guess for what are vectors basically, what are engineered vectors. We pointed out that the viral vectors that are being used presently have limitations in terms of immunogenicity for systemic delivery. It was felt that although local regional control is an important endpoint and could contribute to survival, ultimately we want to bring these vectors into systemic use. How could viral vectors, for example, be engineered to be used as systemic delivery agents? There are trials out there showing that even despite very robust immune responses, you can give repeated doses of adenoviruses, and in fact systemically, and you can do this without major -- at least within a relatively small number of patients -- without major safety concerns. But it was felt that ultimately this would probably be a limitation in terms of long-term treatment. So how can we make the viral vectors less immunogenic? Is that going to be possible? And if that is not possible, or if we need to administer viral vectors, for example, with other agents to reduce their immunogenicity or enhance their targeting, can we develop technology in that respect?

The second area involves going into nonviral vectors, and developing either lipid-based vectors, or engineered vectors using protein DNA complexes, which could successfully be delivered systemically, and which ultimately could be targeted. I think that is the concept behind a technology platform. A lot of research in this area is moving very rapidly. But as yet, nothing has emerged as a dominant vector for use systemically.

David, do you want to comment on that as well?

DR. PARKINSON: No, I think you described it well. The concept is one of cassettes. For example, you could use different genes in frameworks that are quite different. One could be for systemic delivery, another kind of framework would be for generating immunogenicity, and others might be tailored for organ-specific replication. That is what we mean by technology platforms. And then you can shuttle genes in and out, depending on the situation. I think a lot of that will come out of industry, but what can't be done in industry and where the academic community and the government community can do this is to compare and contrast these, and to try and sort out the kinds of issues that Jack pointed out with respect to sorting out advantages and disadvantages to antibodies versus small molecules for some of the tyrosine kinase inhibitors.

These are key issues. How are they similar? How are they different? What are advantages? What are disadvantages? That kind of comparative analysis would I think, well suit therapeutics development.

DR. ROTH: I think just to add to that, there are a number of steps that could be done to engineer a given vector to make it more specific. One, you could add a targeting moiety. In an adenovirus this can be done by substituting a sequence in the fiber protein. In a liposomal vector this could be done by chemically combining a peptide with the lipid that is used in the vector. So targeting moieties are certainly one consideration. A second level would be the DNA expression cassette itself, altering the promoter to make it more tumor-specific. A third level would be to deliver these agents with other types of modulating agents that might reduce the immune response or that might enhance transduction efficiency.

These are just three. There are probably another half dozen or a dozen steps that could be done in terms of vector modification. The feeling was that this is really an area that requires intensive support, but would be the one area that could really enhance gene delivery to the level where we could I think begin to consider real systemic delivery.

DR. SAXMAN: You also said that the group felt that the interactions between these agents, particularly the gene therapeutics and radiation needed to be better studied. Did they have any thoughts about what types of studies need to be done? Do we currently have the models and tools to do that? And if not, what types of tools need to be developed? Perhaps some of the other radiation biologists or radiation oncologists in the audience would have thoughts about that. If that is a clear deficiency, particularly as we talk about locally advanced non-small cell lung cancer, what needs to be done to explore that?

DR. MC BRIDE: I suppose that there are a lot of things in there. One is how do you identify the best approach in these systems? I think that obviously the NCI initiatives in this might be able to help in terms of trying to develop well characterized tumor models, well characterized systems for looking at the effects of p53, as opposed to p16, as opposed to any other kind of approach. I think particularly with respect to human tumors, there is the question of how much variation there is. I think obviously you need to regard radiation resistant tumors and radiation sensitive or less radiation resistant tumors perhaps as being different. How much does tumor heterogeneity impact upon the type of response that is seen?

In terms of the effects of the vector, again this is something which is perhaps overlooked, and different vectors really have different kinds of effects perhaps, or contributing factors, in terms of the radiation response as well. So a replicating adenovector which has an E1A intact is different from the E1A deleted one in terms of the response. There are reports on wild type A and B radiosensitizers. In terms of A and B vectors, I don't know whether there is anything with regard effects on radiosensitization that has been worked out. I think there are a lot of features with respect to the vector, the genes, and the tumors. We need to develop systems that are really very well characterized for looking at these interactions.

DR. SCHMIDT-ULLRICH: Just with respect to the targets that you identified, and I think it will come up in our report as well, it's very important that one appreciates the complexity of the interactions in these molecules. If you just perturb EGFR, you change the interactions with other EGRF molecules and they have compensatory activities. That is one level. So you obviously need to look at the target itself, and the effect of any agent on the target. But you also need to look at secondary interactions, depending on the receptor system. In addition, you have to look at downstream effects, because the downstream effects may be modulated in different ways, depending on how you change the receptor functionality.

So it is very important that you really consider, in your correlative studies, the complexity of these systems. It is very difficult, but I think it can be done experimentally as long as you really keep in mind what the function of these targets is and how they affect ultimate endpoints within the overall cellular behavior.

DR. ROTH: Yes, that's an important point, and I think that was emphasized by many of the individuals in our workshop. I think the question is do we really have the technology yet to begin to look at these complex interactions? It was pointed out for example that array technology, while very helpful, sometimes brings up more questions than answers and that the support, in terms of informatics and being able to look at these complex, multifaceted interactions, still is very much in the beginning stages. We needed a lot more emphasis on research in that area to be able to do it. The focus should be not on a single pathway, but on complex interactions.

If I can just make a comment on the animal model issue also, because I think that's an important one, I think it was recognized that we still don't have the best model systems available, to be able to look at these agents. The standard currently appears to be a xenograft model. I think most of the investigators in the group use that and we know their limitations.

Transgenic models have generally not worked out very well for acute studies, so there probably needs to be some emphasis on developing animal models where spontaneous cancers develop rapidly, but in some type of a targeted way. Perhaps models that look at inducing gene alterations -- for example, maybe with a cre/lac system or something along those lines -- would be extremely helpful in investigating these types of interactions. But we really are lacking spontaneous tumor models, which might provide some useful information here.

DR. OKUNIEFF: Just to make another comment about the gene therapy interactions. In looking for a gene therapy to combine with radiation, you actually don't want one or need one that is particularly toxic. And the reason is that radiation can reproducibly kill cells. The problems are mostly in terms of repairing DNA damage, hypoxia. There are certain very well characterized modulators of effectiveness.

So that in sort of designing a gene therapy to add to radiation, you actually don't want something that is too toxic. You want something, for example, which will keep cells from going into S phase. We mentioned anaerobic bacteria that might grow in hypoxic regions, bring a gene there, and maybe kill those cells, which would be synergistic with radiation. Anything that inhibits DNA repair, particularly DNA PK would be helpful, and it doesn't have to be a very strong inhibitor. It just has to be one that helps a little bit, because you know you are going to get your 20 strand breaks for every 2 Gray, and double strand breaks, and you've just got to keep them from getting repaired correctly.

DR. KOMAKI: I am a clinician. I have been involved in Phase I and Phase II studies for gene therapy with Dr. Jack Roth and Steve Swisher. We have just finished a Phase II study of the adenovirus with p53 and radiation therapy for those patients who have poor performance status, and more peripheral lesion around the chest wall or endobronchial lesion.

We started to escalate the dose of the adenovirus. Then eventually we achieved the MTD of the adenovirus, and we injected this wild type of p53 in adenovirus, plus radiation therapy, which was the standard dose. But for the human beings we achieved 11 out of 14 patients who had good partial response or complete response; 6 out of 11 patients who had pathological CR. We have done biopsies three months after completion of the radiation therapy and the p53 and adenovirus injections. There was an incredible response. We did not see much toxicity. This report was presented by Dr. Swisher at the last ASCO meeting. I think this type of treatment can be done, and we are going to do a Phase III study very soon clinically, combining with chemotherapy.

DR. SAXMAN: In that vein, Paul discussed earlier some of the things that came out of the first session regarding the continued importance of Phase II trials, and what we need to learn out of the Phase I and Phase II studies. For these classes of agents, would that list be the same? Or are there specific things that you would add to or remove from that list in terms of the decision-making, going from preclinical to early clinical to late clinical trials? Did the group have any thoughts or discussions about that?

DR. ROTH: We did discuss those issues and really couldn't resolve it completely. So I don't think I can really present a consensus. But the issues that were brought up were certainly safety could be assessed in the Phase I context, and to date, at least for gene-based therapies, the safety record has been pretty good. Now that doesn't mean it's going to continue that way, and that doesn't mean there aren't going to be problems down the line with it. But it was felt at least that the Phase I approach would be good for safety evaluation.

I think that the issue of whether you need the Phase II study would still be controversial. There were some that felt that one could go directly to Phase III with these agents. I think that here the mechanistic decision points might be some assay to look at gene expression, and other assays to look at some evidence that the mechanistic pathway that you are trying to alter is indeed altered.

It was felt that could be done with serial biopsies. That might be a very reasonable technique, even though the number of cells obtained is small, with PCR technology that is currently available and with current laser capture technology one can analyze these samples fairly rigorously for this type of thing. But it was thought that it's very important that these mechanistic studies be built in even at the Phase I level.

My own bias on this is that I think we need to be much more creative in terms of going to larger scale studies. There are some agents out there that we're going to have to test, and the real bottom line is going to be the Phase III result. The combination of sort of a Phase II/Phase III study, the randomized Phase II that can be extended to the Phase III study, these sorts of innovative approaches are going to need to be extended for these agents. But it was also thought by the group that, at the end of the day when we are looking at the final results, the criteria for these agents should be no different than any other agents, and that is survival.

DR. SHEPHERD: Doing these studies are really very difficult. We participated in a gene therapy study in which there was direct injection of a p53 adenoviral vector gene therapy product. For each of our patients we had to do a biopsy beforehand. If we had only cytology, we had to go back and do a biopsy. Then in the radiology suite, they had to have the injections done of the therapy, and then two days later they had to go back and have another biopsy to know if we had gene incorporation. We did that three times in a row, once every three weeks. That's an extremely onerous type of study to do, and I don't think you can do that in large randomized trials, because you just aren't going to get: (a) the patients that will be able to go through it; and (b) the radiology departments that will be willing to support it, et cetera. So I think you are only going to be able to do those in a select setting.

DR. ROTH: I would agree with you. They are very difficult studies to do. I think they are essential for the initial evaluation though. A tremendous amount of information comes up. If you are not seeing gene expression post-treatment, or you are looking for apoptosis and it's not there, this raises significant questions about whether you ought to pursue with a given gene or a given agent. Those studies can be done, but they require a very dedicated group, usually in a large center, to carry them off. I would agree with you that, when it comes to the Phase III, I don't think those kinds of studies can be done, at least in all the patients.

DR. BUNN: I'll address this question to Dr. Komaki. Whenever I have a problem in the skin, and I see a lot of patients who have cancer in their skin, I send them, if it's local, to the dermatologist who injects vinblastine. We know the dose, and it works over 90 percent of the time. Vinblastine is actually a good radiosensitizer as well. So we want to go through the difficult process of injecting local tumors with the needle. And you want to add it to radiation, and we know there is an interaction between vinblastine and radiation. Why wouldn’t we be willing to use vinblastine and radiation, rather than a virus which makes a tumor shrink 8 percent of the time, and you don't know what the interaction with radiation would be?

DR. KOMAKI: Direct injection of the chemotherapy and the radiation therapy to accessible tumor, that's the question?

DR. BUNN: Yes, the direct injection of the chemo plus radiation, as opposed to just the injection of a virus.

DR. KOMAKI: Well, we talked about that with endoscopies. We see a lot of bronchial lesions. That is one possibility. We can add some of the chemotherapeutic agents directly into the tumor. But one problem we have, in some of the patients we do not know the exact extent of disease. Sometimes our patients do have lymph node metastases, and we would like to combine more systemic chemotherapeutic agents, rather than direct injection of the chemotherapy.

We have all these discussions, the exact staging. The patient, do they have microscopic metastases to somewhere else? Or is this truly just a chest wall lesion, or endobronchial lesion without any occult metastases to the lymph node? That's a difficult problem we face. So we prefer more systemic chemotherapeutic agents at the present time.

DR. SCHMIDT-ULLRICH: I think with p53 there is a distinct biological rationale to combine that with radiation. That is probably one of the rationales why you try this with radiotherapy.

DR. ROTH: That's correct. I don't know if there is any direct comparison between intra-tumoral injection of chemotherapeutic agents versus systemic injection, whether that has really been done.

DR. BUNN: As you might expect, response rates are higher with local injections than they are with systemic injections.

DR. ROTH: At least in the animal models that have been looked at to date, when you administer chemotherapy systemically and look for radiosensitization, you can augment that amount of radiosensitization when you inject with p53. Whether you can do that with an intra-tumoral injection or not, or chemotherapy, I don't know at this point. But again, there are mechanistic differences in the agents. I think it would be reasonable to eventually look at that sort of trial and do some kind of comparison study and see. But there is a mechanistic rationale for using p53, as opposed to other chemotherapeutic agents. But perhaps the combination might even yield a better result.

DR. MC BRIDE: I think one of the questions surely is the extent to which you damage normal tissue. I think that this came up during discussion as well. Presumably the aim of a gene therapy approach is to get cell activity in the long term anyway, whether using tissue promoters or using genes which are going to preferentially affect cancer cells. I think that the opportunities are really in the long-term, particular with gene therapy are enormous from that perspective, and certainly with radiation therapy, the issue is the late normal tissue effects. That's what a lot of the radiation therapy is based on in terms of at least the dose that we give. If you can augment the response within the tumor without affecting the normal tissue response, then that presumably is the end.

DR. GUMERLOCK: Jack, was there any discussion regarding the role of p53 and triggering DNA repair following double stranded breaks from radiation?

DR. ROTH: There was certainly discussion as to a balance of possible beneficial versus detrimental effects for any of these genes that we are looking at.

DR. GUMERLOCK: But specifically p53?

DR. ROTH: I don't think that specifically came up.

DR. GUMERLOCK: Measuring response of GAD45 in this type of a situation? I think when it is combined with radiation it's a different story than when it's just a high level that might induce apoptosis.

DR. ROTH: Yes, I think that's a very valid point. In studies that have been looked at so far, there are very high levels of a gene expressed, which may be very different in terms of the DNA repair capabilities and the ultimate apoptosis versus relatively low levels that are more equivalent to wild type.

DR. GUMERLOCK: But perhaps we better include some of the p53 radiation response genes in the analysis, the downstream genes.

DR. SAXMAN: Have all the other problems that have been going on with gene therapy, all the stuff that has been in the news lately, impacted upon the studies in lung cancer at all or do you think those are moving forward at the same pace? I'm just curious whether this type of work in malignancy or particularly in lung cancer has been affected.

DR. ROTH: We have not noted any change in the number of patients that have been enrolled in the clinical studies to date for lung cancer. My understanding is for the other studies going on in at least our institution, there hasn't been any marked change in the enrollment.

DR. SCHILLER: We have a couple of gene therapy studies going on too, and no, accrual has not been impacted from the patient standpoint. It's been impacted from the regulatory standpoint a lot.

DR. ROTH: Right.

DR. SAXMAN: Okay, thank you.