Challenges Persist With CAR T-Cell Therapy in Pancreatic Cancer

Gregory L. Beatty, MD, PhD

The success of chimeric antigen receptor (CAR) T-cell therapy observed in hematologic malignancies has not yet translated into the solid tumor setting; however, efforts continue to try to bring this new modality into the treatment paradigm for solid tumors, including pancreatic cancer.

Barriers to efficacy of CAR T cells in this setting were discussed by Gregory L. Beatty, MD, PhD, at the 2019 American Society of Clinical Oncology-Society for Immunotherapy of Cancer Clinical Immuno-Oncology Symposium.¹

Multiple CAR targets have been developed, with CD19 being the most attractive one for therapy. CD19 is largely expressed by normal tissue, which is of importance with the recognition of on-target but off-tumor toxicities. Multiple clinical trials of CAR T-cell therapy have been conducted with a range of targets across several solid malignancies, with minimal dose-limiting toxicities observed.

“[CAR T-cell therapy] has been largely tolerable,” said Beatty, director of translational research at the Pancreatic Cancer Research Center at the University of Pennsylvania (Penn) in Philadelphia. “Nonetheless, there has been limited intratumoral activity that has been appreciated thus far.” Unlike with hematologic malignancies, getting CAR T cells into peripheral tissues and working in the microenvironment of solid malignancies is challenging, according to Beatty.

The overall survival rate of patients with advanced pancreatic cancer in response to cytotoxic chemotherapy is <20% at 2 years and <5% at 5 years, and attempts at introducing immunotherapy have had limited success. The response rate is 0% to anti—PD-1/L1 and anti&ndash;CTLA-4 therapy and <10% in second and later lines with immunotherapy combinations.²

&ldquo;The tumor microenvironment shows a desmoplastic reaction [and] is poorly vascularized, and there is a large infiltration of myeloid cells and often a scarcity of T cells,&rdquo; Beatty said, explaining some of the challenges.

Mesothelin-Specific CAR

At Penn, Beatty&rsquo;s laboratory developed a CAR specific for mesothelin, a protein that is overexpressed by pancreatic ductal adenocarcinoma cells. In vitro, CAR T cells engineered to express mesothelin &ldquo;will react and kill tumor cells that are mesothelin positive but not those that are negative,&rdquo; he said. Even in autologous T cells, the CAR T cells were effective in recognizing mesothelin-positive tumor cells.

One strategy to engineer CAR T cells uses an RNA platform and electroporation to transiently express the CAR in T cells. An alternative strategy involves using a lentiviral platform and transduction to permanently express the CAR. &ldquo;Although permanent transduction leads to superior activity,&rdquo; Beatty said, both exhibited antitumor activity in a mesothelioma human tumor model; however, the transient expression of CAR elicited through the RNA platform might appeal to safety.

Early Phase Activity and Safety

The safety, feasibility, and activity of RNA mesothelin-specific CAR T cells (CARTmeso) was explored in a phase I study in 6 patients with chemotherapy-refractory metastatic pancreatic ductal adenocarcinoma.³All patients had received 2 or more prior lines of therapy. The RNA CAR T cells were infused 3 times weekly for a total of 9 doses. &ldquo;We observed no dose-limiting toxicities, and we found no evidence of cytokine release syndrome,&rdquo; Beatty said. The most common adverse events included fatigue and abdominal pain, possibly a result of the pancreatic cancer itself. The best overall response was stable disease in 2 patients who experienced a progression-free survival time of 3.8 months and 5.4 months.

18Fludeoxyglucose (18F-FDG)—PET/CT imaging was used to monitor the metabolic active volume of tumors. The total metabolic activity remained stable (for 1 month) in 3 patients and decreased by 69.2% after 1 month in 1 patient with biopsy-proven mesothelin expression who had significant liver tumor burden at baseline. All liver lesions had a complete reduction in 18F-FDG uptake at 1 month compared with baseline, although the primary lesion was not affected.

The study demonstrated that CAR T cells can have biologic activity in patients with pancreatic cancer and that significantly different immune escape mechanisms may operate within the distinct anatomic location where the disease is present, Beatty said.

A lentiviral CAR T-cell trial enrolled 15 patients with either mesothelioma (n = 5), ovarian cancer (n = 5), or pancreatic cancer (n = 5). Two doses (1-3 &times; 107/m2 and 1-3 &times; 108/m2) were infused (with or without prior cyclophosphamide) with a lymphodepleting regimen. The mean number of prior therapies was 5 (range, 1-11). At both doses, lymphodepletion improved CAR T-cell expansion in the peripheral blood. Lymphodepletion did not have an impact on CAR T-cell persistence; CAR T cells became undetectable in the peripheral blood by 2 months in nearly all patients. The best overall response was stable disease in 11 of the patients.

Potential Resistance Mechanisms

Potential resistance mechanisms that could affect CAR T-cell therapy in solid malignancies are cellular fitness and trafficking, tumor microenvironment CAR target expression, and cellular fate. Patients with pancreatic cancer screened for the phase I study showed little T-cell proliferation in vitro. Further analysis showed a decreased proliferative expansion of CD8 T cells in vitro, along with a decrease in the number of naïve T cells and an increase in the number of effector cells in the peripheral blood in patients with pancreatic cancer, as well as altered cytokine secretion capacity.

&ldquo;We also found that CAR T cells were capable of trafficking to tumors in some patients. Levels of CAR T cells, though, that were detected within the tumors were actually low,&rdquo; Beatty said. &ldquo;CARTmeso cells were detectable in some but not all tumor biopsies, so trafficking was a challenge.&rdquo;

A large degree of heterogeneity was found in the tumor microenvironment, with some regions of the tumor lacking T cells and other regions having a strong influx of T cells. PD-L1 is not expressed in pancreatic cancer, and most T cells are not actively proliferating. The same biology exists within metastatic tumors, he said.

Multiple patterns of T-cell infiltration were identified: ones in which T cells are not present, those in which T cells are present but fail to engage tumor cells, and those that engage but fail to kill tumor cells.