CHEMOTHERAPY/ALTERNATIVE DELIVER METHODS SECTION - LOCALLY DELIVERED THERAPIES by Jack Roth, MD

SLIDES & TRANSCRIPTS
Wednesday, June 19

CHEMOTHERAPY/ALTERNATIVE DELIVER METHODS SECTION - LOCALLY DELIVERED THERAPIES

Jack Roth, MD

Slide 1: Standard Local Control

DR. ROTH: This talk is meant to serve as an introduction to the next several talks in the session, really to discuss some of the issues revolving around biologic therapy and how that might be integrated into a treatment approach in early stage lung cancer. Really we know very little about this. So my talk is going to consist mostly of a lot of questions and hypotheses rather than actual data. However, I think it has become apparent that local regional control still is an issue. Studies in this area have been overshadowed by a real emphasis on systemic therapy and rightly so in lung cancer. We know metastatic disease is the major cause of death in most of these patients, but I think as we look at these earlier lesions and as we re-examine some of the data much of which was presented yesterday we see that there are areas of local control that we could improve considerably and in fact this may give us a real opportunity to make an impact on survival. For example, even under the best situations in clinical stage I lung cancer performing a lobectomy there is still a 7% isolated local failure rate. So, we have not only the local failure rate, but we have the morbidity of the procedure as well, a 3% operative mortality in general and these are just figures culled from a number of studies in the literature and in most cases we see in these patients who already have compromised pulmonary function a decrease in their overall performance status and pulmonary function and this is oftentimes reflected in an inability to tolerate postoperative adjuvant therapy. We saw some data on non-anatomic resections yesterday. Again, the local failure rate here is extremely high, and yet this may be a favored approach for many of these patients with early lesions because we want to preserve pulmonary function. We know these patients are at very high risk for second primary cancers, and with radiation therapy alone there is, also, a very high treatment failure rate, oftentimes manifest as poor local control. So, I think we have a long way to go here and to some degree this may be low-hanging fruit. That is we may be able to more easily make an impact on local regional control than we can on systemic effects.

Slide 2: Hypothesis

So, I would just propose a hypothesis that in fact improvements in local control can translate into survival benefit, and this mostly likely will occur in the context of an existing effective systemic treatment. One can argue to some degree about the effectiveness of systemic treatment for lung cancer, but I think there is no question that it improves survival in stage 4 disease and as we saw in local regional disease when combined with radiation therapy and with other local treatments can result in improved survival. So, there is an opportunity here. If one looks at the past couple of years the New England Journal there are several articles in breast, colorectal, also, in lung cancer showing that improvements in fractionation and delivery of radiation therapy have resulted in increases in local control and have translated in survival.

Slide 3: Study

Here is one of the most interesting, and this was in a paradigm of systemic disseminated lung cancer, small cell lung cancer. Drew Turrisi's study which showed an improvement in 5-year survival from 16% to 26% which was accompanied by an overall increase in local control accomplished by altering the fractionation of radiation therapy. The group showed a 25% to 58% increase in local control. So, I think, again, this does point out an opportunity.

Slide 4: Techniques

Now, when it comes to local regional delivery of agents, there are a number of routes that we can consider. There is intratumoral injection. There is direct instillation into the tracheobronchial tree via bronchoscopy or some other mechanical route and there is aerosolization. Now, these techniques may have differing objectives. With intratumoral injection obviously we are targeting a specific lesion or lesions. The aerosolization and bronchial instillation routes are more targeted toward the entire tracheal bronchial tree and the bronchial epithelium. So, this means that we can look at treatment of local tumors, but we, also, should consider treatment of the large and small airways.

Slide 5: Injection

Now, here is an example of delivering a gene directly to a tumor via intratumoral injection. This is data from Mitch Steiner's group although we have accumulated similar data in our laboratory. The gene here is the LAC-Z gene which produces beta-galactosidase. This gives a very intense blue stain. In this case this is a subcutaneous prostate cancer, but we have done this with lung and a variety of other tumors. There is just a single injection in the center of this 1 centimeter mouse tumor, and when the tumor is harvested and you look for gene expression 24 to 48 hours later you can see this intense blue color throughout the tumor. Now, there are areas of homogeneity here. This represents some areas of non-viable tumor. In general one can transduce these tumors quite effectively with the appropriate adenoviral vector. Now, we have looked at this in larger tumors as well in patients, and have found again that repetitive injections and central tumor injections can result in distribution of vectors throughout the tumor. So, this is an opportunity to develop vectors which can deliver therapeutic genes via intratumoral injection.

Slide 6: Direct Lung Delivery

Now, there is an advantage of direct lung versus oral delivery, and that is the size of the compartment. This is data from Jim Mulshine, but if one looks at what it takes to deliver a drug to the entire body and the volume that one has to deal with, even with intravascular injection this is considerably reduced if we consider delivery just to the pulmonary epithelium.

Slide 7: Techniques

. So, alternatives for delivery to the lung include direct injection via the bronchoscope or alternatively computed tomographic guidance, aerosolized delivery and a third and I think very interesting type of delivery system liquid ventilation. In most cases this is done with perfluorocarbons which have very low surface tension. They evaporate very readily. Oxygen can diffuse through these compounds and in fact they are used to ventilate patients, but also they can be used for gene delivery and delivery of other agents as well.

Slide 8: Public Health Tool

This is an example of a device that can deliver aerosolized medications to the lung. This again is from Jim Mulshine and this is really a portable unit that is quite efficient in delivering aerosolized liquid, 60 to 80% efficient. This can be an aqueous formulation.

Slide 9: Study

Studies of aerosolized retinoids actually show some efficacy in animal models. In this case the studies were done on mice which spontaneously develop adenocarcinomas, the AJ strain exposed to a variety of carcinogens and with aerosolized retinoids delivered for up to 16 weeks the pulmonary tumors were reduced in the treated animals, and this reduction was quite significant although higher doses were associated with some toxicity.

Slide 10: Research Objectives

What might the research objectives be in looking at these various types of delivery system? We might consider improving the results of definitive treatment of early stage lung cancer either by preventing recurrence or preventing second primary tumors. We could enhance local tumor control using this as some sort of an adjuvant, perhaps combine novel treatments with current local and systemic treatments, radiation and surgery and lower treatment morbidity perhaps by developing non-surgical approaches and finally prevent additional lesions from developing.

Slide 11: Classes of Agents

There is a variety of agents that could be delivered by these means including small molecules, proteins, peptides, DNA, cytokines and various other synthetic agents.

Slide 12: Animal Models

However, it is going to be necessary to get a considerable amount of preclinical data before taking these into patients, and this can be done in a variety of model systems including xenograft models, transgenic mice. There are strengths and limitations of each of these systems. The xenograft model is very easy to use. It can replicate disseminated human lung cancer. Transgenic mice are a little bit more difficult to work with in terms of therapeutic models. I think they are very good for looking at gene function, but you have relatively small numbers to deal with. The appearance of metastases is not quite as quantitative and of course it is very difficult to give repetitive treatments to extremely young mice.

Slide 13: Animal Model

This is one such model, a very interesting transgenic mouse that was developed by Tyler Jacks. This was reported in Nature several weeks ago. Basically what is done here is that a K-ras mutant allele is developed with a neomycin gene inserted in it. A transgenic mouse is created and what happens is that recombination events can occur, and the neogene can be deleted causing recombination of the K-ras gene into a functioning mutant K-ras gene, and in this particular model system two of these genes are inserted. Some mice get a single dose of the mutant gene. Some mice get a double dose and this results in the very reproducible appearance of adenocarcinomas in the lung which are multifocal very much like bronchoalveolar cancer.

Slide 14: Delivering Genes

John Curry has been working with this system, and he has developed a system for delivering genes directly to these tumors via instillation of the gene in the tracheal bronchial tree using calcium phosphate precipitates, and so the plasmid is combined with the calcium phosphates, delivered and I think you can appreciate the blue color here which indicates gene transduction in the lung. It appears to be quite efficient with this system.

Slide 15: Untreated Lungs

Here you can see that these are the untreated lungs here. This appears to be selectively taken up by these very small adenocarcinomas within the lung.

Slide 16: Clinical Trial Design

Once we select agents with this preclinical screening process, designing the clinical trials is going to be a real challenge. We need to focus on the natural history and biology of what we are treating. That still is not clear yet for these very early lesions. Differentiation of the lesions, progression, field effects will all have to be taken into account. What will the relevant endpoints be? In many cases given the long natural history of some of these lesions survival is not going to be a relevant endpoint and thus we are going to need biomarkers to assess intermediate endpoints.

Slide 17: Clinical Trial Design

This will probably need to be done through initial Phase I studies, and these can be done actually perhaps best in advanced disease patients to assess the biodistribution of these agents and the safety of these agents really is the primary endpoint rather than to try to take these agents immediately into patients with very early stage lung cancer. Stage I patients would require obviously large numbers of patients and long follow-up to validate this type of approach.

Slide 18: Combination

Now, a second approach rather than using these agents as single modalities would be in combination and one can envision the possibility of sensitization to chemotherapy or radiation therapy and evaluate synergistic, antagonistic and additive effects. One might think of induction adjuvant therapy for example, for microscopic residual disease, treatment of in situ carcinoma and prevention of second primary tumors.

Slide 19: Ad-P53

This type of prevention strategy will require a very careful type of analysis, a very rigorous type of statistical analysis to look at interaction among the various agents that are being combined. This is one such agent that we have worked with for several years, P53, a tumor suppressor gene which functions to regulate the cell cycle, which, also, will induce apoptosis in cells with extensive amounts of DNA damage. This is expressed in a serotype 5 adenoviral vector, and this is some of the initial data that was garnered with this construct. Here are H1299 human lung cancer cells and you can see at various multiplicities of infection the expression of the wild type P53 in this lung cancer cell line which is defective and wild type P53 expression has a profound growth suppressing effect whereas the control vectors here do not have any effect. However, if one incubates normal human bronchial epithelial cells with the adenovirus P53 one sees really no change in their growth even at very high levels of expression. So, there is a very large therapeutic window here, anywhere from 2 to 3 logs in most of the cell lines that we have looked at and I think it is this kind of therapeutic window with these biologic agents that we ought to be able to exploit in many of these early stage cancers.

Slide 20: Phase II Schema

We have done extensive Phase I testing with a single agent adenovirus P53, and I will just present a couple of slides on a combination trial where the adenovirus was injected intratumorally at the same time patients were receiving external beam radiation therapy. Again, this was done in non-small cell lung cancer.

Slide 21: Ad-P53

These patients had very poor performance status. They were not eligible to receive a combination of chemotherapy and radiation which would be standard treatment in the patients. So, they had three intratumoral injections. At 3 months we did a biopsy to evaluate the presence or absence of tumor and just in a very brief format these patients now have been followed for several years. About one-third of the patients remain alive with most of the patients at follow-up at 2 to 3 years. In terms of time to local progression we see that about 40% of these patients remain free of local progression.

Slide 22: Isobolograms

Now, in terms of looking at potential combinations with existing therapies it is going to take a really complex type of analysis to rigorously look at the contribution of each agent, and this is one type of analysis that has been very useful for us called isobologram analysis where one looks at the effectiveness of each agent individually and then develops curves based on an additive effect. In other words you add a little bit of one agent, take away a little bit of another. What would the expected cell kill be for that? One can develop envelopes of additivity based on these individual kill curves, and then if you combine the agents in vitro or this can be done in vivo as well one can look and see where the kill points fall. If they are outside of this envelope they may be in an area of super additivity or an area of antagonism, and in fact one can develop envelopes of additivity in three dimensions for combining three different agents.

Slide 23: P53

We examined possible interactions with docetaxel adenovirus P53 and radiation in vitro and found that in four different non-small cell lung cancer cell lines these agents acted in a synergistic way with the kill points all falling in this area of the envelope outside of the net of additivity.

Slide 24: Important Variables

So, what are the important variables in designing these trials? Well, the biologic relevance of the intervention to the disease process is extremely important. Efficiency of delivery to the target will be critical, and for many of these agents with intratumoral injection this still represents a problem. Biodistribution will need to be examined; exposure time and dose will all be important variables.

Slide 25: Patient Subsets

It is very likely that it is not going to be all patients that benefit from these interventions but certain subsets, and it will be necessary in these early stage patients to continue to identify important subsets to look at the efficacy of conventional treatment risk of these treatments and molecular profiling.

Slide 26: Future Directions

So, future directions in this area will certainly be in the development of more efficient delivery systems, both intralesional and organ based, looking at various small molecules, comparing this with delivery of genes and proteins, combining these agents in clinical trials with existing agents and identifying biomarkers for determining mechanisms of efficacy. Thank you very much.

Slide 27: Discussion

[Discussion not transcribed]