Melanoma 2003

SLIDES & TRANSCRIPTS
Monday, May 5, 2003

Gene and Cellular Therapy for Melanoma

James J. Mulé, Ph.D.

Slide 1:

So, Vern asked me to go over a broad area which includes both gene therapy and cell based immunotherapies for melanoma.

What I decided to do, actually, is give you a history of the development of both arms of these therapeutic approaches in melanoma and then share with you at least my belief of why animal models, particularly new animal models, may help us tremendously, in giving us some new clues as to how these therapies have been relatively ineffective, and how these models can help us to design new therapeutic approaches, again, based on genetic manipulations and adoptive transfers.

Slide 2:

So, it starts in the 1970s, of course, when Norm Wolmark transferred pig lymphocytes to humans, and the field rapidly developed after that, including the use of PHA activated killer cells.

In the 1980s, we saw a large amount of effort using lymphocyte activated killer cells either alone, combined with IL-2.
In the 1990s, the transfer of tumor infiltrating lymphocytes, T cell clones, vaccine primed lymph nodes, cells that were expanded ex vivo and given systemically.

Again, in the 1990s we saw sort of the advent of gene transfer approaches. The key early studies, of course, were genetically modifying tumor infiltrating lymphocytes, giving those to melanoma patients using a reporter gene such as NeoR and then later, perhaps, a therapeutic gene such as TNF-alpha.

Also in the 1990s, we saw immunization with a battery of autologous tumors that have been genetically modified with genes to express 4NHLA, a variety of cytokines including IL2, TNF.

More recently, including last week, we continued to hear about the approach of genetically modifying tumor cells to secret GM-CSF.

Also, in the 1990s, we have seen an explosion, really, of immunization with antigen pulsed dendritic cells, and work by a number of groups have shown that DC can now be genetically manipulated to either enhance their ability to secrete cytokines or enhance their ability to present antigens. I will touch a little bit more on that aspect a little bit later in the talk.

Then we move to the 2000s. We heard a little bit about allogeneic PBSCT or peripheral blood stem cell transplants, mini-transplants in renal cell cancer, but also in melanoma, not a lot of activity in that respect due to the significant toxicities associated with that approach, coupled with the fact that, at least in melanoma, it did not appear to have evidence of therapeutic benefit.

Also, in this millennium we have really some exciting data coming out of Steve Rosenberg's group with Mark Dudley's paper in Science which is, using tumor infiltrating lymphocytes in the setting of melanoma patients, that have been rendered lymphopenic.

I will talk a little bit more about that aspect again later in the talk, but it suggested that induction of lymphopenia may have some benefit in adoptive T cell transfers.

Whether it is due to making space or whether it is due to eliminating some tumor induced immunosuppressive mechanisms is something that I think many of us will focus on in the very near future.

Also, in this millennium, we are seeing the beginnings of the use of T cells that have been genetically modified to change their recognition of tumors by inserting genes that encode specific T cell receptors.

The concept there, of course, is to arm the patient with a large number of killer cells that have now been redirected in their specificity to tumor.

Lastly, I will share with you some new data on dendritic cell-based vaccines and gene modified tumor cells in the setting of induction of lymphopenia, and some surprising data, I think, are now coming out of those types of studies.

Slide 3:

So, I will go through the preclinical data very quickly, because these have been published not only by us, but by a number of other groups.

Again, what I want to send home today is the importance of developing relevant animal models.

Slide 4:

What we have now available, which we did not have before, is this green mouse, which is a GFP transgenic mouse, where essentially the promoter is a ubiquitin promoter, in which every cell in this animal is green.

In fact, not unlike a scorpion, when one turns the light out in the animal room and shines a UV light, these animals glow in the dark.

So, this animal, now, will allow us to transfer cells and monitor in real time in adoptive transfer, or in vaccine strategies, how long these cells persist, where they disseminate to and so forth.

I think this animal is going to help us quite a bit.

Slide 5:

What you can see here is, the lights are on in this animal room. We just shined a light. The eyes will light up green.

Essentially, if one shaves the animal and the lights are out, the animal essentially glows.

Slide 6:

This is important because, again, every cell in this animal now is brightly stained for GFP.

This animal essentially can show a fully competent immune response. It can reject allografts. It can reject challenges of tumor when immunized.

Even though the expression of GFP is present in this animal, it does not appear to adversely affect the immune response generated in these animals.

So, what you see here are wild type dendritic cells that were generated in GM-CSF and IL-4 from bone marrow, from wild type mouse, compared to the MFP transgenic mouse. Essentially every dendritic cell in this culture is now stained green.

Slide 7:

This is a picture of a six day culture of dendritic cells, where you begin to see the cells piling on each other, which is common in these types of cultures, again, very brightly stained dendritic cells. In fact, every cell in this dish is stained green, unlike the wild type animal.

Slide 8:

Now, the expression is not limited to dendritic cells. As you see by this flow based analyses, where we compare wild type splenocytes with GFP splenocytes, we look for T cells, we look for CD4 cells, and we look for CD8 cells.

Essentially, very brightly stained cells are observed in the lymphoid, at least the T compartment.

Slide 9:

This is also seen for B cells and NK cells. Again, if one wants to now design and monitor different types of cells, one has this available. So, here you see CD19 for B cells brightly stained in the GFP animals, and NK1.1 cells, again, are brightly stained.

Slide 10:

One could transfer very small numbers of cells into wild type animals and then begin to detect them by flow-based assays.

Here is one example where we transferred as few as 500,000 cells. We then analyzed the spleen and one can detect CD4 cells and transferred CD8 cells.

These are fresh cells. We have done this with antigen-activated cells as well. So, again, I think these new animal models should give us some advantage in helping us decide how to design further immunotherapeutic approaches for melanoma.

Slide 11:

So, I want to now focus on one aspect of the field, and that is dendritic cells, and give you a sense of how these animal models have helped us to design some new clinical approaches, which I will share with you.

If one has dendritic cells readily available in humans through leukophereses, columns and culture-based methods, one can load up these cells with a variety of tumor forms, including lysate.

If one has the luxury of having to find antigens, one can use protein, purified tumor antigens.

What I am going to share with you is how one can overcome a major limitation in the biology of a dendritic cell by inserting a particular gene.

Slide 12:

So, once one has these cells available -- and this is a typical morphology of dendritic cells and you can see the processes here, and it is a very homogeneous population of cells.

One can introduce into these antigen presenting cells genes of interest.

Slide 13:

By doing that, we were able to design phase I clinical trials that were published. So, I won't share with you that data.

I will tell you that the results of the phase I clinical trial, although we convinced ourselves that dendritic cells were non-toxic, we also were struggling to show any anti-tumor effects as shown by clinical responses.

This was a phase I trial of dendritic cells from patients mobilized peripheral blood that were pulsed with the patient's own tumor, immunized interdermally.
Although we could show immunologic responses, again, there was no evidence of clinical response.

Slide 14:

We are now conducting, based on the animal studies, a phase 1b trial to assess, again, antigen loaded dendritic cells using the patient's own autologous tumor, combined with or without IL-2.

It is a three-armed study comparing high dose IL-2 to low dose IL-2 versus vaccine alone, and this is a trial that Bruce Redman has just initiated at Michigan in stage IV melanoma patients.

Slide 15:

So, the major limitation that we observed, not only in patients but, clearly, in animals, which it is much easier to do is that, when one injects dendritic cells interdermally, the vast majority of these cells never find a lymph node.

Most of them are lodged at the injection site and it is thought that that perhaps may give us a clue as to why this approach has not been terribly effective for us.

What we have done, now, we want to instruct dendritic cells to traffic to and accumulate in multiple lymph nodes by transfer of the L-selectin gene.

Ideally, if one injects dendritic cells intravenously, they will never find a lymph node, due to the fact that ex vivo expanded dendritic cells do not express L selectin.

Slide 16:

What we did was, by confocal microscopy, if you show dendritic cells here, which are colored in red, and you fill them up with tumor that is labeled in green, the confocal very nicely shows uptake by the dendritic cells of killed tumor material.

Slide 17:

What this also allows you to do is inject the cells into mice. This is done interdermally, as you see on the left-hand panel.

The double staining, again, is showing dendritic cells that are loaded with tumor material.

Twenty-four hours or 48 hours later,

Slide 18:

the problem is the sentinel lymph node draining the site, one can see some accumulation of dendritic cells carrying the tumor material with them, but the reality is that this represents much less than .5 percent of the injected dendritic cells.

Slide 19:

So, what is known about trafficking from blood to lymph is that cells must migrate in lymph nodes or traverse what is known as the high endothelial venules, to enter lymph nodes.

That is the problem. The molecule that is responsible for that is L-selectin, which is not expressed, at least, by ex vivo generated dendritic cells.

So, what we did, with Kevin McDounough in internal medicine, collaborating with me and Lloyd Stillman in pathology, was to take dendritic cells which have CCR7, a critical receptor for secondary lymphoid tissue chemokine, was to now introduce into those cells, by retroviral gene transfer, the L-selectin molecule.

Slide 20:

This shows the facts analysis. What we did was, we also generated a mutated form of L-selectin, because this wild type molecule is readily cleaved from the surface of the cell.

So, we wanted the L-selectin to stay on the surface a dendritic cell and, as you can see, we are doing pretty well with the level of gene transfer, hitting about 80 percent of these bone marrow derived dendritic cells.

Slide 21:

I will skip this slide.

Slide 22:

One of the assays we have done is the standard Stamper-Woodruff assay, which is a frozen section of a lymph node.
If you look carefully, what you can see is, we overlay the dendritic cells that have been genetically modified.

If you look carefully, you can see them attaching here to high endothelial venules on the lymph node, showing sort of this pearl-like necklace appearance throughout the lymph node structure.

Slide 23:

Here is a higher power of that response. You can block it with a ligand, and show that it is specific for L-selectin.

Slide 24:

The key, of course, is what happens when you introduce these cells systemically. Here is our first results that I can show you, is that it is pretty remarkable.

In fact, the systemic administration of these cells, now, can target every lymph node in the mouse. Here is one example of 24 hours with the inguinal lymph node, in which the DC assay, control without L-selectin, shows no evidence of migration to that node. Here, we can see numerous dendritic cells now showing up in the inguinal lymph node.

Slide 25:

This is also true for para-aortic lymph nodes. This is 24 hours. In fact, one can see this interesting dendritic cell, which looks like the surgery branch cancer, or CRAB.

You can begin to see, actually, activation of dendritic cells. Without going into details, these dendritic cells were also pulsed with a nova peptide in a transgenic mouse to elicit a response in the animals.

So, what you can see are numerous dendritic cells with the processes, and the evidence in 24 hours of some beginning activation of T cells next to those dendritic cells.

Slide 26:

Remarkably, in inguinal lymph node at 72 hours, one sees what one would presume by the biology, of dendritic cells entering the subcapsular sinus from the periphery, and you see them lining up here in the lymph node.

Again, at 72 hours, the activation of the T cells is dramatic, and you can see the blasts here, as well as here, and another one here.

Again, I think this is one approach which may allow us to inject, for the first time -- at least, to continue this in the animal models -- giving dendritic cells systemically to target lymph nodes in a more efficient way.

Slide 27:

Another way is to create lymph nodes at injection sites. Because of time, I think what I will do is focus more on the efforts of using vaccines and T cell transfers in the setting of lymphopenia.

Slide 28:

[No text is associated with this slide].

Slide 29:

[No text is associated with this slide].

Slide 30:

[No text is associated with this slide].

Slide 31:

[No text is associated with this slide].

Slide 32:

[No text is associated with this slide].

Slide 33:

[No text is associated with this slide].

Slide 34:

[No text is associated with this slide].

Slide 35:

[No text is associated with this slide].

Slide 36:

So, as I mentioned, there was a very exciting piece of information that came out in Science from Steve Rosenberg's group that suggested that a lymphopenic environment may give us clues as to how to deliver immunotherapies, at least with T cell transfers, more effectively.

What I am going to talk about is the role of homeostatic controlled lymph node proliferation, and why this may give us an advantage in pushing ahead with vaccine strategies, using dendritic cells or other types of vaccine approaches such as peptides, and also combining that with T cell transfers.

Slide 37:

So, if one scans the literature over the last couple of years, what one sees is a link between dendritic cells and the induction of homeostatic T cell proliferation.

What this mean is that Polly Massinger, in a paper in Science years ago, showed that, unlike B cells, which is an antigen presenting cell, dendritic cells can prime neonatal T cells to antigen.

That is, as the immune system is generating itself in neonates, if one provides an antigen on a dendritic cell, that you can educate the developing T cell repertoire to recognize that antigen. No other antigen presenting cell can do that.

It was also shown by Mike Bevin and Herman Eisen's group at MIT that, in a lymphopenic host, naive T cells can masquerade as memory T cells by phenotype, hypersensitivity to antigen stimulation.

In other words, one could trigger these cells to antigen and show release of gamma interferon as early as two hours after stimulation in vitro.

They also show dramatic increased gamma interferon production as a result of homeostatic driven T cell proliferation.

At a key point in time -- and the window is narrow -- the lymphocytes that are mature T cells undergo, for some unknown reason, this wave of proliferation.

They immediately become masqueraded memory cells. If one provides antigen at that critical point in time, it suggests that can essentially sensitize these T cells in a very efficient way.

Herman Eisen also published in PNAS a follow up paper that, in a T cell dendritic cell co-culture system, naive T cells can proliferate in the absence of foreign antigen.

IL-15 was responsible to some extent, in that T cell dendritic cell co-culture system, and it was inhibited by he addition of CD4, CD25 regulatory cells. Hence, bringing in new cell types suggesting that this cell may play some role in the down regulation of immune responses in fully immunocompetent animals.

Slide 38:

Our first clue was this experiment. What we did, we gave a mouse a syngeneic bone marrow transplant and, very early on, six days after the transplant -- and keep in mind these animals receive total body irradiation, they receive a whole marrow transplant -- and contaminating that marrow transplant are about 500,000 mature T cells. That is all.

So, by flow based assays, one cannot readily define or see reliably T cells in many of the lymphoid organs.

However, if you begin to immunize these animals as early as six days after the transplant with unpulsed dendritic cells -- these are unmanipulated dendritic cells -- or tumor loaded dendritic cells, what we saw essentially was T cell recovery being expedited.

If you look at day 19, for instance, in lymph nodes, a dramatic increase in the level of T cell recovery that were produced by dendritic cells.

I don't show you the controls, but B cells, monocytes, had absolutely no effect on expediting T cell recovery. Only dendritic cells were shown to do this.

If you look in the spleen, of course, there are numerous cells, more so than in this lymph node, but the pattern remained the same.

One could see an expedited recovery of lymphocytes. These are naive cells. They are CD4 and CD8 cells. There does not seem to be a preferential recovery of one subset versus the other.

Slide 39:

Remarkably, what we saw was that if animals were bearing or, in this case, had received a bone marrow transplant, and then were immunized with tumor loaded dendritic cells as early as seven days post transplant, when essentially they were 90 percent depleted in all their lymphoid organs of CD4 and CD8 cells, and then were re-challenged at 28 days when recovery was coming back in these mice, one could see a hint that a challenge dose of tumor could be inhibited by giving one injection of tumor loaded dendritic cells at this time point.

Slide 40:

Now, of course, if you gave dendritic cell vaccines in a more aggressive manner -- that is, every seven days starting on day seven and ending at day 21, and challenged the animals, essentially, all animals were protected completely even though, when this challenge was administered, full recovery of the animal's immune system had not come back to normal.

In fact, at this point in time, there was about 50 percent recovery of lymphocytes and, even then, one could see dramatic protection of these animals to a challenge of a lethal, in this case, mammary tumor.

What one also saw is, even though the mice remain somewhat lymphopenic at this time point, if we took the spleens out of these animals, we could show specific gamma interferon release, when the cells were taken out and triggered with the appropriate tumor, but not an H2 matched A20 lymphoma, and some controls are shown here.

So, not only were we able to educate the animals early on, not only were we able to expedite T cell recovery, but we were also able to educate the animal to reject a challenge dose of tumor.

Slide 41:

Remarkably, one can now take the marrow of these animals and transfer those, because, unlike the spleen and unlike lymph node -- and we are not exactly sure why -- we see large numbers of antigen specific T cells in the marrow of these animals, again, early on as T cell recovery is coming.

Here is an example of the bone marrow of these animals, showing, in this case, this is in a B16 melanoma tumor model, where these T cells are specific gamma secreters, for B16. The controls are shown here. There is essentially no production with some of the other control groups, including normal bone marrow T cells, tumor bearing bone marrow T cells and so forth.

Again, what we are doing is to take the marrow of these animals, transferring those, and immunizing further to see whether vaccines can now improve or give increased potency to the vaccine, using this bone marrow transfer approach.

Slide 42:

Not surprisingly, animals with established tumors that then undergo bone marrow transplant, followed by immunization with dendritic cells early on, can show dramatic regressions of the pulmonary metastases.

One always see, when you immunosuppress mice, to some extent, the amount of tumor appears to be greater in the bone marrow transplant alone mouse, compared to the immunocompetent no BMT mouse.

Nonetheless, the vaccine appears to be as efficacious in the setting of lymphopenia as it is in the setting of a fully immunocompetent mouse. We have data to suggest that it may actually be more effective.

Slide 43:

Given that, we have planned -- and it is before an IRB and it is with the FDA with Jim Geiger, a pediatric surgeon, working with me, we are going to pursue a phase II trial of tumor loaded dendritic cell vaccine early after autologous peripheral blood stem cell transplants, to prolong progression-free survival in pediatric patients with sarcoma neuroblastoma.

The beauty of this trial, of course, is that these patients have minimal residual disease post-transplant.

Then, with Jeff Weber, Jeff has just, with us, put together a protocol which we hope to get off the ground, soon, which will be a phase I trial of escalating doses of fludarabine, followed by internodal delivery of MH class I/II peptide-pulsed matured dendritic cells in patients with chemotherapy naive metastatic melanoma.

Again, what I wanted to do today is give you a sense of how the animal models are giving us clues as to how we should perhaps design the next generation trials.

I think those models have provided us with some very exciting new approaches that I think we will hear more about in the future. Thanks.