NEW YORK – A new study has found that metastatic urothelial carcinoma patients with a mutation in ARID1A and baseline tumor expression of immune cytokine CXCL13 had better overall survival on immune checkpoint therapy.
The expected five-year survival rate in urothelial cancer is around 5 percent. Recently, the US Food and Drug Administration approved immune checkpoint inhibitors for metastatic urothelial cancer showing that patients on these therapies have durable responses. However, there is a need for more refined biomarkers that can help identify patients most likely to experience an enduring response.
Researchers led by Padmanee Sharma from the MD Anderson Cancer Center used patient samples from two ongoing clinical trials at the institute to discover whether certain characteristics in the immune microenvironment and the tumor mutational profile were associated with patients' ability to respond to immune checkpoint therapy. The results of their study were published in Science Translational Medicine on Wednesday.
Patient samples were obtained from a Phase II study assessing the safety and efficacy of nivolumab (Bristol Myers Squibb's Opdivo) in metastatic or surgically unresectable urothelial carcinoma patients who progressed or recurred on platinum-based chemotherapy and from a Phase I/II open-label study of nivolumab monotherapy or nivolumab combined with ipilimumab (Bristol Myers Squibb's Yervoy) in advanced or metastatic solid tumor patients.
Further, the researchers used patient profiles from the CheckMate275 and IMvigor210 trials as independent confirmatory cohorts. CheckMate275 studied nivolumab monotherapy in platinum-resistant urothelial cancer, and IMvigor210 studied atezolizumab (Roche's Tecentriq) in platinum-treated urothelial carcinoma.
Whole-exome sequencing was performed on the bladder tumor samples of 24 patients from the discovery cohorts to identify tumor-specific mutations. From these patients in the ongoing clinical trials, the team found that ARID1A was the only gene mutation that was enriched in responders compared to non-responders. Among the 11 patients who responded to treatment with a checkpoint inhibitor, four patients had ARID1A mutations, whereas none of the non-responders had an ARID1A mutation.
ARID1A plays a role in transcription. In preclinical animal studies of bladder cancer, researchers found that when they disrupted normal function of ARID1A, the animals had an increased response to anti-PD-1 therapy.
When cross-referencing this finding against the data in CheckMate275 and IMvigor210, the researchers found that patients with the ARID1A mutation tended to have higher tumor mutational burden (TMB). Patients with higher TMB and an ARID1A mutation had better progression-free survival and overall survival compared to patients with higher TMB but without the ARID1A mutation.
On the same day the study was published, the FDA approved pembrolizumab (Merck's Keytruda) for patients with refractory solid tumors and high TMB as defined by 10 mutations/Mb. As such, urothelial cancer patients with metastatic or unresectable tumors with no other options can now receive pembrolizumab if they have high TMB status.
In addition to TMB, the study authors explored how other biomarkers may interact with ARID1A mutations in impacting immunotherapy response in patients. In the IMvigor210 cohort, for example, patients with the ARID1A mutations had lower levels of transforming growth factor β1. TGFβ1 has been previously associated with reduced responses to anti-PD-L1 therapy in patients with urothelial carcinoma.
ARID1A mutations, however, did not correlate with differences in PD-L1 expression in tumors or in immune cells of patients from the IMvigor and CheckMate275 studies.
ARID1A mutations were associated with better overall survival in both the IMvigor and CheckMate275 studies. Among the 139 patients in CheckMate275 with ARID1A mutations, median overall survival was 11.4 months compared to six months in patients without ARID1A mutations. In the 275-patients in the IMvigor210 cohort, those with ARID1A mutations had a median overall survival of 15.4 months compared to 8.2 months in patients without these mutations.
Sharma and colleagues then wanted to explore if there was also an immunological biomarker present that could predict response to immune checkpoint therapy. To do that, they performed immunohistochemistry testing in the baseline bladder tumor samples of 31 patients from the discovery cohorts. They found that responders to therapy had greater immune infiltration in their tumors from CD4 and CD8 T lymphocytes and CD20 B cells.
To identify genes in the baseline tumor samples that correlated with immune infiltration, the team then used a custom 739-gene NanoString panel to analyze gene expression. CXCL13 was the one gene they found that was involved in both T cell and B cell infiltration.
The researchers further observed how CXCL13 affected therapy response in animal models of bladder cancer and found that mice with CXCL13 expression responded to anti-PD-1 therapy while mice without CXCL13 expression did not. They went back to the CheckMate275 and IMvigor210 cohorts and found that indeed, responders to immunotherapy tended to have higher expression of CXCL13 compared to non-responders.
In the CheckMate275 cohort, bladder cancer patients with high CXCL13 expression had a median overall survival of 13.5 months, bladder cancer patients with medium CXCL13 expression of 6.6 months, and bladder cancer patients with low CXCL13 expression of 5.7 months. A similar trend was observed in the IMvigor210 cohort. Patients with high, medium, and low CXCL13 expression had a median overall survival of 17.1 months, 6.7 months, and 8 months, respectively.
"Based on previous biomarker studies, we have learned that a single genomic or immunological biomarker could not reproducibly predict responses to immune checkpoint therapy because of dynamic interactions between tumor cells and immune cells," Sharma and colleagues noted in their paper. "Therefore, we tested whether [the] combination of ARID1A mutation and CXCL13 expression have improved predictive capacity compared to either single biomarker."
In CheckMate275, patients with ARID1A mutations and high CXCL13 had a median progression-free survival of 3.7 months, compared to 1.7 months in those with ARID1A mutations and low CXCL13. Progression-free survival was 1.9 months in patients with no ARID1A mutation and high CXCL13, and 1.9 months in patients with no ARID1A mutation and low CXCL13.
The median overall survival was 19.1 months in patients with ARID1A mutations and high CXCL13, versus 5.7 months in patients with ARID1A mutations and low CXCL13. For those with no ARID1A mutation and high CXCL13, median overall survival was 9.5 months, and 5.3 months in patients with no ARID1A mutation and low CXCL13.
Similar trends were seen in IMvigor210, where patients with ARID1A mutations and high CXCL13 had a median overall survival of 17.8 months, compared to 10.5 months in patients with ARID1A mutations and low CXCL13. Among those with no ARID1A mutation and high CXCL13, the median overall survival was 10.2 months, and 7.1 months in patients with no ARID1A mutation and low CXCL13.
"The association between CXCL13 expression and response was stronger in patients harboring ARID1A mutation compared to patients without ARID1A mutation … Therefore, composite biomarkers identified through interrogation of the tumor-immune ecosystem are needed to guide treatment decisions," the researchers wrote. "Combination of the two biomarkers in baseline tumor tissues suggested improved prediction of responses to immune checkpoint therapy compared to either single biomarker, which will need to be tested prospectively."