Mechanisms of sensitivity to platinum-based chemotherapy in PD-1 blockade-refractory head and neck cancer.

Abstract

e18003 Background: Our clinical observation from recurrent/metastatic head and neck patients who treated with platinum based combination after progression on pembrolizumab showed a striking response rate of 80% (8 out of 10 patients), this impressive ORR is dramatically higher than the historical ORR of 27% in the first-line setting. Our mouse model of ‘immune-cold’ head and neck cancer is developed to investigate the mechanism of sensitivity to platinum- based chemotherapy in PD-1 blockade-refractory head and neck cancer. Methods: For ‘immune-cold’ syngeneic model of HNSCC, C57BL/6 mice were injected with 5 × 104 MOC2 tumor cells in the right flank. After 11 days, mice with palpable tumors were randomized and treated concurrently with 4 groups (n = 5 per group): control, anti–PD1 antibody (200 mcg three times/week × 3 cycles), cisplatin (3mg/kg) +paclitaxel (5mg/kg) once a week x 3 cycles, or anti-PD1 antibody (200 mcg three times/week × 1 cycle) followed by cisplatin (3mg/kg) + paclitaxel (5mg/kg) once a week x 2 cycle. Cisplatin + paclitaxel arm received total of 4 cycles of treatment, while PD-1 followed by cisplatin and paclitaxel arm received total of 3 cycles of cisplatin and paclitaxel, this schedule was set up to maintain every 3 week regimen with same duration of treatment time to mimic patient’s treatment schedule in clinic setting. Tumor volume was measured 3 times per week. Results: Our initial pilot mouse model study showed similar tumor response seen in patient population. Starting on D24, average tumor size of PD1 inhibitor followed by cisplatin+paclitaxel group is significantly decreased compared to PD1 inhibitor or cisplatin+paclitaxel group and the difference is statistically significant. Most of mice in control and PD1 inhibitor group died at day 24 due to tumor progression. Our IHC data showed that the expression levels of immune markers such as CD8+, CD4+ and PDL1 in tumors treated with PD1 inhibitor followed by cisplatin and paclitaxel were increased, compared to almost no or scant expression (CD8+, CD4+) and lower expression (PDL1) in tumors treated with cisplatin and paclitaxel alone. With our expansion cohort, we will be further exploring key changes in immune signatures using both peripheral blood samples and tumor blocks for flow cytometry (myeloid/lymphoid subtypes, alterations in cytokine expression and circulating tumor DNA levels) and RNA sequencing analysis. Conclusions: We successfully established an ‘immune-cold’ mouse model of head and neck and reproduced similar tumor response seen in clinic. Preliminary data showing increased expression of immune makers such as CD8+, CD4+ and PDL warrants further investigation on tumor immune microenvironment remodeling as a potential mechanism of sensitivity to platinum-based chemotherapy in PD-1 blockade-refractory head and neck cancer.