Co-expression of PD-L1 and DDR2 in cervical cancer

Abstract

<h3>Objectives:</h3> Many cancers are dependent on programmed cell death ligand (PD-L1) to evade immune destruction. A clinical benefit of PD-1/PD-L1 inhibitors in cervical cancer has been established based on a combined positive score (CPS) of >1. However, cumulative evidence that the level of PD-L1 expression alone is not predictive of response to anti-PD1 axis therapy in cervical cancer suggests that identification of more complex molecular signatures are needed to select patients appropriate for immunotherapy. There is some indication that kinase pathway activation impacts anti-PD1 response, and recently discoidin domain receptor 2 (DDR2) was identified from an unbiased screen as a leading target for enhancement of response to anti-PD1 axis therapy. We hypothesized that DDR2 and PD-L1 co-expression occurs in some cervix cancers, and sought to determine the co-expression pattern of PD-L1 with DDR2 in patients with cervical cancer. <h3>Methods:</h3> A tissue microarray of pathologic specimens from treatment naïve cervical cancer patients was used for analysis of PD-L1 and DDR2 expression utilizing immunohistochemical (IHC) staining, with interpretable scores for 64 patients. A CPS was used to report PD-L1 expression. Positive PD-L1 staining was defined as CPS score ≥1. DDR2 expression was quantified by identifying percent positive staining pixels within the IHC samples and then categorized as low (0-20% positive), moderate (21-40% positive), or high (41-100% positive) expression. Correlation between PD-L1 expression and survival was calculated by the Kaplan-Meier method and log rank test. <h3>Results:</h3> Of the 64 patients, histology was as follows: 56 squamous, 5 adenocarcinoma, 1 adenoid cystic, 1 adenosquamous, 1 MMMT. Positive PD-L1 staining was observed in 58/64 samples. There was no independent association with PD-L1 positivity and recurrence or survival. Sixty-one of the samples were available for concurrent staining of DDR2. 17/61 samples demonstrated moderate DDR2 expression, and 44/61 samples had high DDR2 expression. Of the PD-L1 positive samples, all expressed at least moderate staining of DDR2, and 70% expressed high levels of DDR2. There was no significant correlation between CPS for PD-L1 and level of DDR2 expression. <h3>Conclusions:</h3> In this cohort of cervical cancer patients, the rate of PD-L1 and DDR2 expression was high. These data suggests that a subset of these patients may benefit from dual inhibition with anti-PD1 therapy and DDR2 inhibition, as this combination treatment has been shown to enhance anti-PD1 response in other tumor types. Further work to test this hypothesis in cervical cancer is needed.