Patients with ovarian cancer can respond to immunotherapy, but rationally designed synergistic combinations will be necessary to enhance upfront efficacy and to sustain durability, said Daniel J. Powell, Jr, PhD, Scientific Director of Immunotherapy, Division of Gynecologic Oncology, University of Pennsylvania, Philadelphia, at the 2019 ASCO-SITC Clinical Immuno-Oncology Symposium.
The only approved form of immunotherapy for ovarian cancer is bevacizumab, which targets the vascular endothelial growth factor receptor signaling pathway. The paucity of approved immunotherapies is based on a common misconception that ovarian cancer is not immune responsive, said Dr Powell.
The defining role of immune cells that control tumor progression was actually elucidated in ovarian cancer, he pointed out, in which dense accumulation of T cells within the tumor was shown to predict survival. An early experiment showed a reduction in CA125 levels and a reduction in tumor size with a strategy of vaccination using a granulocyte-macrophage colony-stimulating factor–modified whole tumor vaccine in combination with ipilimumab in a patient with recurrent ovarian cancer.
This observation led to clinical trials of a variety of checkpoint inhibitors in the context of recurrent ovarian cancer. The KEYNOTE-100 trial, which tested pembrolizumab, reported data from a training set of 100 patients and showed an overall response rate (ORR) of just 9%, with a higher ORR in patients with PD-L1 expression (14% with combined positive score [CPS] ≥1 and 25% with CPS ≥10). The ORR in KEYNOTE-100 was consistent with that observed in other trials of single checkpoint inhibition in ovarian cancer, Dr Powell said.
“It’s important to point out that unlike other cancers, like melanoma and lung cancer, ovarian cancer is a disease of gene amplification and homologous recombination deficiency,” he said. “It does not have a high tumor mutational burden. So I think we need to think about ovarian cancer as a different type of disease—one that could be immune responsive but doesn’t fit within the paradigm of neoantigen recognition.”
Targeting CTLA-4 and PD-1
CD137-positive T cells are tumor-reactive T cells that are enriched within the tumor microenvironment and, once isolated, produce proinflammatory cytokines in response to autologous tumor, whereas CD137-negative cells do not. CD137-positive tumor-reactive cells express high levels of PD-1 and CTLA-4 and other negative immunoregulatory molecules at the global level, Dr Powell explained.
Recent research shows that within the tumor-reactive tumor-infiltrating lymphocyte subset are 4 phenotypic metaclusters. The distribution of these subtypes skews toward exhausted T cells, with far fewer effector T cells. Effector T cells express very high levels of CTLA-4, whereas exhausted T cells have high levels of PD-1 and also express high levels of CD28, a costimulatory molecule required for cell proliferation. “Importantly, we know that CTLA-4 is required during the priming phase, while PD-1 is induced during the effector phase,” said Dr Powell. “So we rationalize that ovarian cancer could be treated better with a combination of targeting these 2 molecules.”
In patients with persistent or recurrent ovarian cancer, a phase 2 randomized trial of nivolumab with or without ipilimumab, each followed by maintenance nivolumab, demonstrated a 3 times higher rate of response within 6 months in the nivolumab-ipilimumab arm compared with nivolumab single-agent arm (31.4% vs 12.2%, respectively). The complete response rate, however, was nearly identical in the 2 arms—5.9% versus 6.1%, respectively. The median progression-free survival (3.9 vs 2.0 months, respectively) was also significantly superior with the combination (P = .041), with a nonsignificant trend toward improved overall survival (OS; 28.1 vs 21.8 months; P = .43), although the study was underpowered to detect a clinically important difference in OS, said Dr Powell.
An increased likelihood of response was observed in the subset with clear cell cancer, similar to observations in renal cell carcinoma.
The responses were independent of age, performance status, and number of previous cytotoxic regimens, which would suggest that these T cells are recognizing cancer or are recruiting new T cells, or perhaps the T cells are resistant to the effects of chemotherapy, he said.
“Lastly, it raises questions of whether this therapy should be brought up front immediately after surgical debulking and chemotherapy.” The potential to synergize immunotherapy with a vaccine that bolsters the endogenous immune response is warranted by the results, he added.
Although ovarian cancer has naturally occurring tumor-reactive T cells, not all patients with ovarian cancer have them, he said. Tumor intrinsic factors can limit response, such as loss of class I or loss of antigen-processing machinery. Synthetic biology provides a possible solution by engineering tumor-reactive T cells that can recognize cancer, according to Dr Powell.
Folate Receptor Alpha CAR
One target that has been analyzed in the context of cellular therapy is folate receptor alpha (FRα). Its surface glycosylphosphatidylinositol-anchored protein transports folate, internalizes it, and allows for DNA synthesis and other transcriptional activity, which renders a growth advantage for cancer cells by promoting proliferation, migration, and loss of adhesion. It is an ideal target in ovarian cancer, Dr Powell said, because 90% of cases express high levels of target antigen.
Toxicity has not been reported with any FRα-targeted therapy, whether it be with an anti-FRα antibody, an antibody-drug conjugate, or bispecific antibodies. ORRs with the latter 2 approaches have been approximately 25%.
The first chimeric antigen receptor (CAR) T-cell product targeting FRα did not produce any responses and caused toxicity related to the high-dose interleukin-2 that was coadministered. Failure was blamed on poor transgene expression and lack of persistence of T cells. To overcome the problem of poor T-cell persistence, CD137 has been incorporated into CARs as a costimulatory domain. “These cells that are outfitted with costimulatory domains have a capacity to persist in vivo,” he said. Based on potent antitumor effects observed in a xenograph model, a clinical trial to target FRα has been opened to patients with persistent or recurrent stage IV high-grade serous epithelial ovarian, fallopian tube, or primary peritoneal carcinoma.