AACR 2015: Report from Day 3

Day 3 was a big news day at the AACR annual meeting. Today we learned the results of two high-profile clinical trials—in melanoma and triple-negative breast cancer (TNBC). We also learned how basic research is radically changing the way we think about factors that can influence the immune system and its ability to fight cancer.

By far, the most anticipated development was the unveiling of results from a phase II trial of a combination of ipilimumab (ipi) and nivolumab (nivo) versus ipilimumab alone, for patients with advanced melanoma. These drugs, made by Bristol-Myers Squibb, are immune checkpoint inhibitors, which “take the brakes off” the immune system and allow it to attack cancer. Ipi is currently standard of care for advanced melanoma, but previous research has suggested that the combination might bring added benefits. The current trial included patients with and without BRAF V600 mutations. The objective response rate (ORR) was 61% for the combination vs. 11% for ipi alone in patients without a BRAF mutation. For patients with a BRAF mutation, the ORR was 52% vs. 10%. For both groups, the rate of complete responses was 22% for the combo, but 0% for ipi alone. Importantly, both PD-L1-positive and PD-L1-negative patients (defined on the basis of having 5% of tumor cells with detectable PD-L1 expression in a pre-treatment biopsy) had similar objective response rates (58% vs. 55%). This implies that PD-L1 expression is not a useful predictive biomarker for determining who might benefit from the drug—at least in this type of cancer. According to F. Stephen Hodi, MD, director of Immuno-Oncology at Dana-Farber Cancer Institute and a member of our clinical trials network, the results of the trial, published online today in the New England Journal of Medicine, indicate that “the nivo + ipi combination represents a new treatment option for treatment-naïve patients with advanced melanoma.”

The other big clinical trial news of the day was the report of results from a phase Ia trial of the PD-L1-blocking drug MPDL3280A, made by Genentech/Roche, in patients with triple-negative breast cancer (TNBC). TNBC is one of the most challenging forms of breast cancer to treat. Unlike women with HER2+ or hormone-positive breast cancers, those with TNBC have few treatment options. Neither targeted therapies like Herceptin nor hormone antagonists like tamoxifen work for them. And immunotherapy vaccines designed to prevent recurrence after surgery are currently available only for patients with HER2-positive tumors. The results from this new trial, presented by Leisha A. Emens, MD, PhD, of the Johns Hopkins Sidney Kimmel Comprehensive Cancer Center, who is also the reviewer of CRI’s breast cancer immunotherapy information page, showed that MPDL3280A was well tolerated and had efficacy in 4 out of the 21 patients (19%) whose data were ready for evaluation. These patients all had high pre-treatment expression of PD-L1 on their tumors. An additional 33 patients with other levels of PD-L1 expression were not yet ready for evaluation. Among the four responders, two had a complete response and two had a partial response. Three patients appeared to have progressive disease, but were later determined to have “pseudoprogression”—meaning that their tumors appeared to get worse before they got better. There was good reason to suspect that TNBC would respond to PD-L1 checkpoint therapy. This is a type of breast cancer with a higher rate of mutation compared to other types. More mutations means a higher likelihood of generating a neoantigen that the immune system will recognize as foreign. TNBC also tends to have more tumor-infiltrating lymphocytes, indicating the presence of a baseline immune response to the tumor. Finally, TNBC tumors tend to have more PD-L1 expression than other subtypes; they therefore have more of the target for the drug. A phase III trial of MPDL3280A is currently being planned for patients with TNBC.

Besides the exciting clinical trial data, there was lots of other interesting science happening today at the meeting. CRI Scientific Advisory Council member Robert H. Vonderheide, MD, DPhil, from the University of Pennsylvania, chaired a session on the tumor microenvironment that had the audience buzzing. Of particular note was a talk by Elizabeth Repasky, PhD, of Roswell Park Cancer Institute, discussing the role of stress on immune system function and tumor progression. Various kinds of stress, including both emotional and physical stress, are relayed through body by the sympathetic nervous system. Nerve fibers produce the neurotransmitter norepinephrine, which binds to beta-adrenergic receptors on cells in various tissues, like the heart, to mediate the “fight-or-flight” response. Repasky chose as an example of stress the body’s response to cold temperature. She found that tumor-bearing mice housed at colder temperatures had more aggressive tumor growth than mice housed at balmier temps. This effect on tumor growth was mediated by the immune system, as shown by the fact that mice housed at the warmer (“thermoneutral”) temp have double the number of T cells in their tumors than cold mice. Moreover, cold-stressed mice had larger spleens filled with cells of a type that is known to suppress the immune response. Most interestingly, this entire effect was dependent upon the sympathetic nervous system response to stress. Repasky found that giving the cold-stressed mice a beta-blocker (the drug propranolol) completely eliminated the detrimental effect of cold stress on the mouse immune system and its ability to control cancer. Repasky’s results are interesting in the context of growing recognition of the role of norepinephrine in tumor progression, as shown by other groups. The Cancer Research Institute is currently funding translational research looking at beta-blocker therapy in breast cancer.

Dr. Vonderheide concluded the session with a talk of his own, on ways of making pancreatic cancer a suitable target for immunotherapy. Pancreatic cancer has the lowest 5-year survival rates of common cancers—6%. So far, this cancer type has failed to respond to immunotherapy with checkpoint inhibitors. One possible reason for this failure is that the immune system in pancreatic cancer patients has not mounted an initial attack on the cancer. “Releasing the brakes” on the immune response therefore has no effect because there is nothing to release. Why the immune system is unable to mount even an initial attack against pancreatic cancer is an unsolved mystery. One possibility is that the dense tissue, or stroma, surrounding the tumor prevents immune cells from entering. But as Vonderheide showed, pancreatic tumors have many kinds of immune cells inside them, including macrophages. What they lack, however, are T cells—the immune system’s main attack cells. Vonderheide and his team wanted to know if it was possible to coax T cells into a pancreatic tumor by stimulating them with a kind of vaccine. The vaccine in this case was a combination of chemotherapy and an agonist (or “stimulating”) antibody specific for an immune molecule called CD40, found on antigen-presenting cells. He showed in a mouse model that this immunotherapy, when combined with checkpoint blockade, was effective at curing pancreatic cancer in 65% of the mice. Interestingly, this effect was entirely dependent upon T cells. “If you first deplete the T cells, you can’t cure a single mouse,” he said. Vonderheide stressed that there are various combinations of treatments that might serve as an effective immunotherapy in pancreatic cancer—including radiation—and noted that there are now quite a few clinical trials currently open or opening soon that are testing these vaccines in combination with checkpoint inhibitors. These trials will go far toward the urgent goal of making pancreatic cancer an effective target of immunotherapy.

Check back tomorrow for updates from day 4 of AACR.