[No text is associated with this slide].

TOP

In terms of antigen specific immunotherapy for melanoma, there are some new and promising ideas, and I am going to go through them relatively rapidly.

I think there are new ideas in immune assay development, and I think it should be easy to convince you that immune assays or surrogate markers are critical in the development of antigen-specific immunotherapy.

Anyone who has any experience in immunology would accept that you need to generate memory T cell responses. I would also think we need to generate T helper responses.

We desperately need new molecular adjuvants in immunotherapy, and there are examples that we will describe, such as CTLA-4 antibody, and I will very briefly mention toll-like receptor binding agents.

Also, as Jim Mule will discuss in the next talk, we desperately need to exploit the idea of homeostatic lymphoid proliferation, using either adoptive T cell transfer or dendritic cell immunization.

TOP

At any rate, in the next slide which we reverse, you will see that there are a number of tumor antigens that have been defined over the last decade.

I would like to divide them into five categories. If you look at the second category, the vast majority of antigens that have been employed in antigen-specific therapy for melanoma fall into the so-called differentiation antigen category.

Among them are antigens we have heard mentioned by other speakers -- Mart1, GP100, tyrosinase, TRP-2. For other tumors there are a variety of similar types of antigens, like PSA or CEA.

I will not talk about the so-called true mutated tumor specific antigens because frankly, as many will agree, they are found either on individual tumors or on very small subsets.

Knowing an antigen doesn't mean that you can immunize it, nor does it guarantee that you can even perform an assay.

TOP

The choice of a surrogate immunologic assay in our antigen-specific vaccine trials is very important to guide further development.

The ELISPOT assay is used by a number of groups, which essentially monitors spots on a plate coated with an antibody where cells interact on the plate, adhere to the plate, and then a second antibody is added to develop a spot, which indicates a single antigen-presenting cell, effector cell interaction.

That methodology is rather limited in its reliability, its flexibility and reproducibility, but it is commonly used.

In the future, I think, we will see flow based assays that have much greater potential for use. They can actually enumerate -- that is, you can actually count cells very sensitively with flow cytometry tetramer assays.

I think I will show you that you can generate functional data. You can easily standardized flow cytometry. You can control them, and any good university medical center has a good flow cytometer.

I will briefly show you a slide or two on a new flow-based assay that, for the first time, will yield data on functional high avidity T cells that may be important for tumor regression.

TOP

This just shows a typical tetramer based assay in two different patients, pre-vaccine, showing that the CD8 tetramer positive cells before vaccination are nil, nil defined as one in 20,000 or less, and significant reactivity in this patient, and in this patient, is seen after a peptide vaccine.

TOP

However, this is just enumeration. It doesn't tell you whether these are functional cells. It doesn't tell you whether these are clinically useful cells.

It turns out there is a new development that tells you how you might begin to develop a functional assay.

CD107A is a lysosome membrane protein upregulated in activated lytic CD8 cells. There has been shown to be a correlation of lysis and avidity, and avidity with lysis of peptide expressing tumor cells, in work done by Peter Lee at Stanford.

The highly avid tumor-specific T cells turn out to be CD107A positive, and if you can sort tetramer positive CD107A positive cells, those are the cells that are functional. Those are the cells that lyse tumor lines.

If you look at the CD107A negative, but tetramer positive cells, those cells do not lyse. They are not functional cells.

So, the question arises, could you devise a combined peptide tetramer, CD107A assay, to use to detect functional highly avid tumor specific CD8 cells after vaccination.

The answer is, you probably can.

TOP

These are data that have already been submitted, and I will just show you one slide suggesting that, if you look at clones of cells from patients who have been vaccinated against GP100 -- and this is a Mac slide, so it did not come out well -- but if you look at these four clones that are CD107A and tetramer positive, they all lyse tumor cell lines that are antigen specific, and don't lyse the antigen negative control.

If you look at any number of clones -- and here are two clones that are CD107A negative and tetramer positive, so they would be counted as cells that score positive in any sort of tetramer assay, they don't lyse any of the targets, they don't lyse peptide pulse targets either. So, they are non-functional cells.

For the first time, we might actually have a functional flow based assay, which I think will be very useful in monitoring vaccine trials.

TOP

What we desperately need, though, as I have stated in the past, is a correlation between the assay and any clinical benefit.

The tetramers count cells. You don't get functional data. The new assay that I mentioned may provide functional flow data.

What we need to do is perform phase II trials of our antigen specific vaccines with sufficient numbers of patients to be able to make a correlation between time to relapse, survival, or clinical response, and the immunologic end point.

We desperately need a useful surrogate marker, and we really don't have one at this time. Tetramer arrays, which is a new idea, or tetramer spot assays, where you actually spot tetramers onto a plate, and you can actually count very few antigen specific T cells that bind to the plate, may give high sensitivity detection of antigen specific T cells, and you could literally use 1,000 or 10,000 cells in your assay.

Those are new developments, and those are coming up the pike. So, I think there is hope for the future that we will be able to devise functional assays that may correlate with clinical benefit in our vaccine trials.

TOP

Now, molecular adjuvants are a key area. Interleukin-12 has already been shown by investigators in this room, including Tom Guyevsky(?), to boost immune responses. I have shown that both IL-12 and GM-CSF can do the same.

In peptide vaccine trials, when they are used as local adjuvants, they can augment responses by either presumably acting directly on T cells, which is IL-12, or acting on APC, which is GM-CSF.

There are newer adjuvants coming up the pike, such as those which bind to toll-like receptors, that include the CPG oligodeoxynucleotides, which bind to toll-like receptor 9 on plasmacytoid DC, or Imiquimod, which is the prototype of the whole family of reagents that bind to TLR-4 on myeloid DC, and those are just entering testing.

More advanced toll-like receptor ligands like resiquimod and newer oligodeoxynucleotides are in the biopharmaceutical pipeline, and I think hold some hope for the future.

TOP

This just shows an example of how IL-12 can boost antigen specific immune responses. These are public data showing that, in a peptide trial, in patients with resected stage III and IV melanoma, the addition of local IL-12 at the vaccine site boosts T cell immunity directed specifically against GP100, compared to its absence.

TOP

In a trial with GM-CSF, when the same peptides were used to immunize patients, there is no question that you can begin to see epitope spreading, where you actually begin to generate CD8 T cell reactivity against antigens that were not in the vaccine mix.

In this case, it enumerates Mart1 specific cells, either to the so-called Mart wild type or the Mart substituted heterolytic epitope, suggesting that immunizing with GP100 can cause epitope spreading to different antigens, in this case, Mart1.

TOP

So, what are the clinical data with these peptide vaccines?

Unfortunately, very rare responses in patients with stage IV melanoma, five percent or less.

However, when used in patients with resected melanoma, in one case, in one particular trial with IL-12, we had 22 patients with resected stage IV and we had 26 patients with resected stage III disease. The median survival in that trial is greater than 48 months with 28 of the 48 patients still alive at 48 months, very favorable compared to any other matched group in any other similar kind of trial.

Similar kinds of data can be seen in a stage 2A, 2B trial, where there have been only three deaths in 48 over 50 months.

Those data, as I said, are equivalent to, or better, than any published data for this population of patients. I am not ready to have a news conference but, again, these are encouraging data which suggest that antigen specific vaccines should be taken further in the setting of resected stage III and IV melanoma.

TOP

Now, any immunologist in the audience will tell you about the importance of long-term memory T cells.

In patients in my laboratory, who have been vaccinated with GP100 and tyrosinase epitope peptides with incomplete Freund's(?) adjuvant, we had the luxury of being able to analyze their T cells 18 to 24, and even 36 months after they were vaccinated. These were patients who would have a long-term survival. They had resected melanoma.

We detected effector and effector memory T cells up to 36 months after finishing their vaccination. These were cells that were gamma interferon secreting, activated cytolytic T cells that, in some cases, could retain potency for up to three years after the vaccination.

The cell numbers decayed at different rates, but one patient that we had actually had a 2.5 percent tetramer specific response and maintained his response over two years, without any diminution.

TOP

Again, if you just look at one slide, it shows two different patients who have T cell reactivity, ELISPOT assays, pre-vaccine at different, nice responses to GP100 after vaccination at six months, and 18 months and two years later, still have very nice memory effector T cells that secrete gamma interferon, suggesting that these, indeed, were memory effector T cells generated years after vaccination.

TOP

If you look at their ability to lyse peptide expressing targets, if you look at this patient pre, another patient pre, no evidence of lytic reactivity.

After six months of vaccination or 18 months later or 24 months later, on a cell for cell basis, the effector cells in the circulating blood were just as effective at lysing peptide expressing targets.

TOP

What we need are long-term memory T cells. We need to see studies in our antigen specific vaccine trials on the maintenance of T cell immunity.

Patients with resected melanoma are obviously appropriate to explore these issues. We need to understand how anti-melanoma T cell reactivity decays over time, and what signalling mechanisms control it. There has been almost no work done on this area.

As Kim Lowery(?) will tell you, he feels that maintenance of T cell numbers over some critical level may be important for protection against clinical relapse. If that is the case, understanding how you can maintain a response is very important.

TOP

Now, what about T help? Well, in a very small pilot trial done at my institution, we have actually vaccinated patients with peptides that are specific to A1, A3, A11 and B44. We have given them in an aqueous adjuvant called AS02B.

For the first time, we actually immunized patients who matched at DR4 and DR11, class II epitope peptides. We actually pheresed them before and after six monthly vaccinations.
We actually were able to show for the first time that patients could respond to a class I epitope peptide.

We found that the DR4 restricted, Mart1 5173 peptide generated T cell responses, as demonstrated by proliferative response, CD4 lysis, in fact, and ELISPOT assays.

TOP

Just one or two slides suggesting that in three different patients who were matched at DR4 but had different DR4 alleles, DR4B01, 04 and DR7, which is actually a serologic DR4 allele, it shows that fresh PBMC can proliferate after the patients are vaccinated with this class II helper peptide.

These are controls over here, our negative controls.

TOP

If you look at ELISPOT reactivity, again, three different patients with different DR4 alleles, DRB0404, 0401 and DR7. No question that you can generate from fresh PBMC or stimulated PBMC ELISPOT reactivity against this class II epitope.

It turns out that they make gamma interferon and the cells also make IL-5, so their T helper one and T helper two cells.

TOP

What do we need? I think we need induction of T help. There is no question that, at least in this small pilot trial, that you can see significant levels of CD4 T helper cell reactivity after vaccination with at least this one peptide.

You can also see ELISPOT and proliferative recognition. It turns out that you can generate T helper cells that are lytic and the lysis is actually class II restricted.

I think follow up trials where we employ class I peptides or any sort of class I restricted construct with class II help should be conducted.
We actually planned a trial with class I peptides with or without class II peptides, with a Th1 polarizing adjuvant.

The key question that is unanswered is, do broad T helper responses augment or prolong class I specific immune responses and improve clinical outcome.

TOP

I will close in talking about what I think is a promising reagent, which is CTLA-4 antibody.

It is a human antibody that has been prepared that binds to CD154, a molecule expressed on T cells that acts as a down modulatory influence on T cell responses.

In murine models, Jim Allison has elegantly shown that the antibody administered with a GM-CSF transduced cell vaccine boosted immune responses and prolonged survival in a mouse model.

A phase I trial of escalating doses of that antibody was performed at our institution in patients with resected melanoma, who also received a multi-peptide vaccine with incomplete Freund's adjuvant.

Again, the peptides were from three melanoma antigens, GP100, Mart1 and tyrosinase.

TOP

To our surprise, or I should say perhaps it shouldn't be such a surprise, with this human antibody, the half-life was quite long.

When we got up to what we thought was going to be our maximal or optimal dose at three milligrams per kilo, given intravenously every month, we had very high levels and, again, this is micrograms per ml in the serum, so we are between 50 and 100 micrograms per ml, and the half life is very long. This is four weeks. So, we were at very impressive levels out to three and four weeks.

TOP

In terms of the trials, we have actually conducted a trial using this as an adjuvant for the first time. We saw immune responses in about half of our patients.

The PK, as I indicated, was very prolonged. Fifteen of our 19 stage III-IV patients are alive and free of disease at a one year median follow up, including all of those in the highest cohort.

There is no question that in this trial there are immune responses. There is no question that in trials where CTLA-4 antibody alone was used, as a stand-alone reagent given once, there was evidence of responses in patients with melanoma.
CTLA-4 antibody is an active drug against melanoma and merits further study. We will probably be doing increasing single dosing and every two month dosing schedules as an adjuvant in cancer vaccine trials.

Very briefly, I will tell you about a trial that Steve Rosenberg performed at his institution, which Pat Hwu talked about last week publicly,

TOP

where, instead of immunizing patients with peptides IFA with antibodies in the adjuvant setting, it was done in the metastatic setting, where the CTLA-4 antibody was given every three weeks.

There was no clear impact on immune responses, but clear evidence of objective responses associated with evidence of autoimmunity, which appears now, for the first time, to be a surrogate marker for the effectiveness of this human antibody.

Again, I should point out that response rates with the same peptides IFA alone are less than five percent. So, currently the NCI, as Steve tells me, is doing a trial with a loading dose of the antibody followed by a lower dose given every three weeks.

TOP

Assessment and recommendations, since my time is up. I think it is very unlikely that an antigen specific vaccine approach alone will result in a high rate of regression in stage IV melanoma.

We need to generate higher levels of CTL in our PBMC and in the lymph nodes.

We need to induce long-term effector memory T cell responses.

We need to immunize with multiple epitopes from different antigen classes.
I strongly think that our trials should treat patients with stage III and IV resected disease. You have a rapid clinical end point in the space of months, yet the patients will stay alive and free of disease long enough to be able to collect important information.

We desperately need to correlate clinical benefit with surrogate immune markers, and you can take your pick.

We need to test new adjuvant preparations. We need to generate T help. Again, drug approval, looking up the road, I think, would require a phase III trial in resected disease.

I think to me the optimal end point in an antigen specific immunotherapy trial is overall survival in patients at high risk for relapse. Thank you for your attention.

TOP

[No text is associated with this slide].

TOP