Dr. Michael B. Atkins was recently appointed deputy director of the Georgetown Lombardi Comprehensive Cancer Center. He was formerly the director of the cutaneous oncology and biologic therapy programs and the cancer clinical trials office in Beth Israel Deaconess Medical Center's division of hematology/oncology and was the leader of the kidney cancer program at Dana-Farber/Harvard Cancer Center. His research interests include immunotherapy of malignancy—specifically, using cytokines, understanding and overcoming cancer-associated immune suppression, novel treatments for melanoma and renal cell carcinoma, and antiangiogenic and targeted therapies. He is currently investigating the potential of CTLA-4 and PD1 pathway antibodies, and mechanisms of enhancing sensitivity of melanoma to selective RAF pathway inhibitors.
—Interviewed by Ian Ingram
CancerNetwork: There seem to be two principal promising avenues of therapy now for patients with metastatic melanoma: targeted, oncogene-directed therapy, and immunotherapy. The mechanisms of action involved in these two approaches are very different. Could you briefly summarize what you see as the greatest strengths and weaknesses of the two approaches?
Dr. Atkins: The benefit of immunotherapies is that they can produce durable responses in a small subset of patients off the treatment and appear to work in patients with both BRAF-mutant and BRAF-wild-type melanomas. The benefit of the molecularly targeted therapies is they produce a high level of tumor shrinkage and survival benefits in a large proportion of patients with BRAF mutant melanomas. Those responses are likely short-lived, on a median of 6 to 8 months progression-free survival, but for patients with symptomatic disease who need a response, the high degree of response is a major benefit.
CancerNetwork: I believe a phase I/II trial of the combination—ipilimumab and vemurafenib—is now underway. Last year at ASCO there was excited speculation about the beneficial synergies that might result from using these two approaches together. However, a poster presentation this year, by James Harding and colleagues at Memorial Sloan-Kettering, reported an increased rate of grade 3 skin rash in patients who received vemurafenib after ipilimumab; the increase was most marked in patients who started vemurafenib within one month of completing treatment with ipilimumab. So clearly there is also a potential for negative interactions between targeted agents and immune-enhancing antibodies. What is your take on how successful this combination treatment is likely to be?
Dr. Atkins: Well, I think many of us are still hopeful that immune therapy and molecularly targeted therapy with a BRAF inhibitor will produce the benefits of both classes, meaning maintaining a high response rate and increasing the number of patients who have durable benefit. But the questions that remain are which agents should be given first if you're giving them in sequence and which agents to use for both BRAF inhibitor and immune therapy. The abstract that was presented from the group at Memorial Sloan-Kettering suggests a potential problem if you give an immune therapy first—which is if you start the vemurafenib afterwards you can have increased skin toxicity from the vemurafenib. That might not be an issue if you wait longer and allow for the antibody to clear from the circulation or if you use dabrafenib, which is a BRAF inhibitor that doesn't cause as much skin toxicity. There are equal if not more problems with starting with vemurafenib and trying to switch to ipilimumab, as an abstract that was presented in the poster session by Allison Ackerman suggested that very few patients are able to get immune therapy after progression of disease following a molecularly targeted therapy, and those that did receive it were unlikely to benefit. So the hope is that potentially by giving them simultaneously we might reduce some of those issues—but once again it may require using immunotherapies that are more selective, such as the PD1 blockers, and molecularly targeted agents that may have less overlapping toxicities with the immune therapies, such as potentially dabrafenib.
CancerNetwork: One of the hot areas of research in melanoma right now involves the discovery of other inhibitors on the surface of T cells, in addition to the CTLA-4 protein that ipilimumab targets. One of these, programmed death 1 (PD1), was discussed in an oral abstract presented at ASCO (#8507). How might targeting PD1 be different from targeting CTLA-4 and what might be the advantage (if any)?
Dr. Atkins: CTLA-4 is a very early checkpoint inhibitor, and when you target it, the interaction is likely blocking an interaction between antigen-presenting cells and immune cells that takes place in a lymph node, and therefore the effects are not selective to the tumor and you can release immunity not just against tumor antigens but against host antigens. The blocking of PD1 or the PD1 ligand (PDL1) appears to take place more peripherally because PDL1 and PD1 only seem to interact in peripheral tissues, and so providing an antibody that blocks that interaction might limit the effects to the tumor site. So that means that there's potential for more selectivity and less toxicity.
CancerNetwork: Do you think a combination of a PD1 antibody and a CTLA-4 antibody will be a promising approach, or might there be an unacceptable increase in autoimmune-related adverse effects?
Dr. Atkins: Well I think there's a lot of rationale for combining them from a therapeutic standpoint because when you activate the immune system with a PD1 antibody or a PDL1 antibody, those cells might still express CTLA-4 on their surface, and it's possible that the CTLA-4 expression—particularly if those cells are circulating—may limit their effectiveness. And blocking CTLA-4 at the same time may therefore enhance their efficacy. In addition, when immune cells that have been activated by shutting off the CTLA-4 inhibitory interaction reach the tumor, if the tumor is expressing PDL1, then those cells which likely also have PD1 on their surface may not be able to kill the tumor. So there's a strong rationale for putting these together. There is a concern, given the autoimmunity side effects related to ipilimumab, that you might enhance those. That's why many of us would recommend that if a combination be tested that it should be built off PD1 rather than using a full dose of CTLA-4 antibodies.
CancerNetwork: Do you think immunotherapy will soon become another area in which a "personalized medicine" approach will prevail, with some tumors responding better to one of these antibodies and some to another?
Dr. Atkins: There is some data that was presented that I would view as very preliminary suggesting that patients in their tumors PDL1 expression were more likely to benefit from a PDL1 inhibitor than those that did not have that expression and so that doesn't mean that PDL1 is actually a target of the antibody in the typical sense; it means that PDL1 is a biomarker of a immune reaction that's taking place where the immune system is recognizing something on the tumor. But because it's an inducible marker, its expression at even a low level, 5% or so, is sufficient to suggest that an immune reaction is taking place. It could be very difficult or possible to miss the expression with the type of pathology review that's typically done. So a lot of work has to be done to both figure out what specimens should be used, how to get a more reliable antibody, and to try to make certain that enough samples are looked at to accurately report whether a tumor is expressing PDL1 or not, so that we can address the question of whether this is a true predictive biomarker of response. For that reason I don't think that it's appropriate right now to restrict PD1 antibody treatments to patients whose tumors have the expression of the PDL1 ligand.
CancerNetwork: What progress has been made in melanoma with regard to the problem of patients developing resistance to targeted agents such as vemurafenib and dabrafenib? What have we learned about the mechanisms of acquired resistance—and have we learned enough to design ways to prevent or minimize it?
Dr. Atkins: There was a very nice oral presentation from a group led by Jeff Sosman in the melanoma oral session at ASCO that looked at analysis of tumor specimens pre- and at-time-of resistance in a number of patients receiving vemurafenib. And their conclusion was that the majority of resistance mechanisms involved reactivation of the MAP kinase pathway and specifically phospho-ERK, and they reported a number of mechanisms by which that can happen—a number of these have been published—but what they also mentioned were the frequency of mutations in NRAS and in MEK that were seen in their population. Some of those mutations were bypassed pathways that may be sensitive to MEK inhibitors—but not all of them—and therefore one of the proposals for prolonging the benefit to selective RAF inhibitors could be to combine them with a MEK inhibitor, and there was some data presented in the MEK inhibitor session that suggested that a combination of a RAF inhibitor and MEK inhibitor was tolerable and may produce a longer progression-free survival—although once again, resistance does develop. In addition, I think with more sensitive assays we may be able to pick up these mutations in the pretreatment specimens because they likely did not develop as a result of treatment but were there to begin with and were selected for. And that may allow one to choose whether a patient should receive a RAF inhibitor or a combination of a RAF inhibitor and a MEK inhibitor or some other treatment approach.
CancerNetwork: And have there been lots of adverse events associated with MEK inhibitors as well?
Dr. Atkins: The MEK inhibitors do have toxicities because they're less selective when they're given at high doses, but they don't seem to add toxicities to the selective BRAF inhibitors. Particularly when the MEK inhibitor from GlaxoSmithKline was added to dabrafenib, there seemed to be less skin toxicity and less typical toxicity associated with the MAP kinase pathway bypass, but there may have been more pyrexia.
CancerNetwork: Are there any targeted agents in the pipeline for melanoma patients who do not have the V600E mutation?
Dr. Atkins: I think the MEK inhibitors also show value in NRAS-mutant melanomas although they're not selective inhibitors of the mutation so they're limited by their toxicity because MEK is pretty important even in normal cells. There's been data that's presented for a while about c-kit inhibitors particularly in mucosal melanoma. But we don't really know what's driving the rest of the tumors that are classified as BRAF wild-type, and hopefully some research into that may uncover some driver mutations that potentially could be targeted.
CancerNetwork: Most of the buzz at ASCO is about therapies that involve drugs that can be mass-produced by pharmaceutical companies. What about other approached to melanoma therapy that do not involve a drug, such as adoptive cell transfer? Are any of these promising?
Dr. Atkins: I think the adoptive T-cell therapy approach has always been promising but very impractical and limited to sites like the National Cancer Institute which has the financing to be able to do these research studies or to a few institutions that have taken the effort to gather the resources to have a program. The trouble with adoptive T-cell therapy is that a randomized trial has never been done to prove the value of the adoptive transfer of the T cells above and beyond the selection of the patients that take place at multiple steps along the way in their process. Until that happens, I think it's going to be hard for programs to enthusiastically endorse this or for insurance companies to pay for it. There is a company that has developed some interest in trying to move this effort forward to perform that phase III trial and so it's possible that such a phase III trial with a company having some intellectual property rights to the process of preparing the cells. If and when the trial is done, and assuming the results show benefit for the use of the cell. Regarding other approaches, I think most of those are within the pharmaceutical industry, although some places are still looking at their own developed vaccines, alone or in combination with dendritic cells or fused to dendritic cells—and there's a new opportunity for vaccine approaches combined with some of the more novel immunotherapies, so we may see some of those move forward again.
CancerNetwork: Dr. Atkins, thank you very much for your time.
Dr. Atkins: You're very welcome, Ian.