H. Kim Lyerly, MD: RNA Modified Androgenetic Cell Therapy

With the introduction of anti-HER2/Neu antibodies as an effective therapy for breast cancer, we're beginning to realize some of the potential of immunotherapy. What I'll talk about today is the use of T Cell therapies as distinct from antibody based therapies to recognize or at least understand the induction of T Cell recognition in patients with cancer.

[ Slide 01 ]   A Killer T Cell's Victory

This is a slide that I've copied from the 2000 Howard Hughes Medical Institute Annual Report showing by electron microscopy evidence that recognition of tumor cells by T Cells, and then pore formation and then lysis of the tumor cells. What we will ask today is does it really occur in a manner that's going to be meaningful?

[ Slide 02 ]   The Molecular Mechanisms of T Cell Recognition

Some of the fundamental advances that have occurred that give us some hope that this will occur is the conceptual development and actual demonstration of the molecular mechanisms of T Cell recognition. This slide shows a cartoon of a virally infected cell in which we actually know that the processing of viral antigens on the surface of the cell leads to expression of a viral associated peptide in the context of the MHC molecule with some co-stimulatory molecules. The critical component of this slide is the fact that the peptide that's recognized on the surface of the cell is processed within the cell and presented on the cell surface. So it turns out that the most potently recognized target on virally infected cells are intranuclear and intracytoplasmic proteins. The concept that had precluded this was the fact that perhaps a virally infected cell has got a big protein hanging on its cell surface so it could be recognized and be killed. Well, it turns out that's in fact not the case. It's mutations or over expressions of protein within the nucleus or within the cytoplasm that gets processed and presented on the cell surface. We can see how the body can recognize cancer cells, by recognition of over expression or mutation of antigens which are processed and presented on the cell surface.

[ Slide 03 ]   Activation of T Cells

The second major theme that has been developed over the last five to ten years is the fact that activation of T Cells requires more stringent signals than mere recognition. I showed you a cartoon of recognition of an antigen on a tumor cell, but here's another cartoon that obviously is more complex. This tumor cell has these peptide antigens on its cell surface. This is well documented and dozens of antigens have been described on tumor cells. If the antigen's there, why isn't it recognized and why don't we go around killing all our tumor cells? Well this second concept is that activation of the immune system requires processing, presentation by antigen presenting cells, induction of T-helpers cells, with growth factors, etc., etc. The mere presence of these antigens will not be sufficient to induce an anti-tumor response or any type of anti-antigen response.

[ Slide 04 ]   Active Immune Therapy Inhibits Lung Metastasis

One strategy to overcome that was to use gene therapy strategies to modify tumor cells to provide them with those second signals. What if we genetically modify the tumor cell to produce those secondary growth factors like IL-2, or GM-CSF, etc. or co-stimulatory molecules like B7? This is an example of an experiment done with a metastatic breast cancer model. If we immunize animals that have had a breast cancer implanted, then resected, all the animals go on to die of metastatic breast cancer. If you immunize them with the control or with the tumor cells, or with gene modified tumor cells with the irrelevant genes, there's no effect. If we immunize with genetically modified tumor cells, to secrete cytokines like IL-2, we get a marked reduction in the number of metastases.

[ Slide 05 ]   Active Immunotherapy with Cytokine Gene Modified Tumor - Ex Vivo Strategy

This conceptual development, genetically modifying tumor cells, to trigger immunity to induce the appropriate immune response was attractive and strategies like this in cancer patients in which patients with cancer would have tissue harvesting, genetic modification of their tumor to express a variety of molecules, and immunization. Studies like this were carried out in melanoma and in renal cell carcinoma and they do look very interesting and promising. All of us who've gone to the surgical pathology laboratory after our cases and reviewed the pathology in our patients recognize that the one small practical problem in this strategy is the ability to harvest large amounts of purified tumor cells. Pancreatic cancer is a particularly difficult one, and the strategies in which pancreatic cancer was targeted have been altered in which investigators have tried to now use allogeneic cell lines rather than harvesting the patient's own tumor cells. Although this strategy of gene modification looks very promising, technically it's very difficult.

[ Slide 06 ]   Presentation of Antigen to CTL

Well what else can we do? Let's go back to the viral models. Here's a virally infected cell. Let's remind ourselves of gene expression, which occurs, by DNA being transcribed to RNA and the RNA is then processed. And part of the processed RNA goes on to make proteins, which are then degraded and presented along with MHC molecules on the cell surface which are then recognized. But along this pathway, we can see that these events can be recapitulated perhaps in very potent antigen presenting cells.

[ Slide 07 ]   Dendritic Cells

A parallel development occurred in which the actual cell in the body appears to trigger immunity was isolated and could actually be shown to grow to large numbers in tissue culture using progenitors and growth factors in establishing in-vitro cultures.

These are dendritic cells, named for their processes that are present here, not because they have anything to do with the central nervous system. Dendritic cells have these long processes, they are sort of dendrite-like and they appear to be the cell that's pivotal in induction and modulation of immunity. They have all the co-stimulatory molecules, all the accessory molecules that we knew were important in triggering immunity.

[ Slide 08 ]   Breast W42852 CEA Response

You can imagine one of the things we wanted to demonstrate is that you can use dendritic cells to trigger immunity. This is an experiment in which we modify dendritic cells to get them to trigger immunity, and in this experiment we show cytolytic activity against two colon cancer cell lines, one CEA+ and one CEA-.

What's remarkable about these data is that for the first time we're able to mount primary immune responses by repeated in-vitro stimulation using autologous tissues. It's sort of a vaccination in a test tube experiment, which five to six years ago would have been considered inconceivable by the immunology field.

[ Slide 09 ]   RNA Transfection, Processing and MGC Presentation of peptide Antigen by Human DC

One of the strategies we developed, again, is to genetically modify dendritic cells and there are a variety of strategies. One way that was pioneered by Smita Nair and Eli Gilboa, a collaborator at Duke, was to actually take messenger RNA encoding for the appropriate antigen, transfecting that messenger RNA process, and then have the peptide on the surface of the dendritic cell. And remarkably enough, this actually works.

[ Slide 10 ]   Messenger RNA

Now why use messenger RNA if it's not stable? We've tried a lot of things to modify the message, to cap it, to alter the poly-A sites, etc. to help in its stability. But it's the transfer is not that efficient. In fact, for some of the defined antigen strategies targeting CEA, HER2/Neu, etc., we actually use modified adenovirus, which is a very potent strategy to generate modified dendritic cells.

The reason RNA is so provocative, is the potential to use RNA to extract the total antigen content of a tumor cell. This is a metastatic colon cancer in which we used a mechanical microdissection to isolate the tumor cells.

[ Slide 11 ]   Target: Autologous Tumor Cells

We actually can take the RNA from that one pathology slide and we can load them onto dendritic cells and stimulate T Cell response very similar to the CEA response that you'd seen earlier. In this case if we stimulate with CEA specific dendritic cells, we get recognition and killing of a small fraction of the patient's own autologous tumor cells. If we take dendritic cells loaded with the tumor RNA from that patient's own tumor, we can generate a very potent and conceivably multi-antigenic immune response. By cold target competition we can show that many of the antigens present here turn out to be not CEA.

[ Slide 12 ]   Correlation of Immune Response and Clinical Response

We're on the verge now of beginning to thoughtfully explore the potential use of vaccine strategies in cancer. The potential use of any single antigen like HER2/Neu, CEA, p53 and so forth is very exciting, but now we have strategies that we think can conceivably be utilized in clinical trials. The big problem in my opinion is the fact that we are currently at a stage where we are implementing some type of immunotherapeutic strategy in patients with cancer and we achieve, at this point, detectable responses. The problem is "What is the real role and magnitude of this detectable immune response?" To achieve clinically effective immune responses, do we need to go to a superthreshold level, do we need to achieve a sustained level rather than having a high level that really is abrogated by the other normal mechanisms of the body? The major focus of the laboratory right now is in creating real data that describes these events in cancer patients rather than this cartoon which is what we hypothesize may be happening.

[ Slide 13 ]   Therapeutic Tumor Specific T Cell Responses

Finally, I think this is really an exciting time for Surgical Oncology as we have the potential to methodically explore the activities of induction of anti-tumor immunity and then correlate those with real clinical events. None of the animal models that I've seen from any vaccine studies ever demonstrated large tumors which shrunk to nothing. So the classical cytoreductive development strategies in Oncology, take a gigantic tumor and shrink it to nothing, are not likely to occur if we can't get them to work in mice. I think a strategy is to use our cancer patients is to clinically debulk them to undetectable clinical tumor and we know that there's predictable and oftentimes rapid relapses. The potential for large amounts of activity is to begin to develop the correlation between tumor burden and induction of anti-tumor immunity. We can extend these types of strategies to include not only antitumor immunity but also small molecules or molecular targeted anticancer therapies.

I do want to acknowledge my many collaborators, my laboratory group, Eli Gilboa's laboratory and the other investigators that help us to develop our vaccines and our ability to analyze them. Thanks very much.