Abstract 2002: Assessment of the co-localization between PD-L1 positive macrophages and PD-1 positive cells is associated with benefit from PD-1 axis immunotherapy in NSCLC

Background: The immune checkpoint target PD-L1 has been found to be highly expressed in macrophages and immune cell PD-L1 expression is associated with response to therapy in multiple tumor types. Here, we examine the co-localization between PD-L1 and its receptor PD-1 and especially in macrophages, using interaction analysis to determine whether this co-localization carries the predictive value of PD-1 axis therapy in NSCLC. Methods: We utilized two quantitative immunofluorescence (QIF) platforms (AQUA analysis with the PM2000 platform and InForm analysis with R studio with the Perkin Elmer Polaris Automated Quantitative Pathology Imaging System platform) and assessed the co-localization between PD-L1 and PD-1 both in the stroma and specifically in macrophages by phenotyping each cell. Co-localization was confirmed with high resolution images using confocal microscopy. The PD-1/L1 co-localization was measured using the 5-plex QIF panel including PD-1, PD-L1, CD68, cytokeratin, and DAPI. This panel was performed on three retrospective Yale NSCLC cohorts (1 PD-1 axis immunotherapy treated cohort with 62 cases and 2 non-immunotherapy treated cohort with 457 cases) represented in tissue microarrays. Results: PD-1/L1 co-localization measured by AQUA analysis was significantly associated with response to PD-1 axis therapy using overall survival as the endpoint(P=0.0176) but it was not prognostic in standard of care treated patients. PD-1/L1 co-localization was found to also co-localize with CD68. The co-localized signal of PD-1/L1 was 69.5% in CD68 expressing macrophages and 30.5% in other, CD68 negative cells. Phenotyping by InForm analysis showed that patients with clinical benefit had a significantly shorter distance of nearest neighbor between PD-1 positive cells and PD-L1/CD68 double positive cells as compared to non-clinical benefit group (P=0.03). However, no difference was found between two response groups for the distance between PD-1 positive and PD-L1/cytokeratin double positive cells. Conclusion: The association between PD-1/L1 co-localization and clinical outcome to PD-1 axis therapy infers the biological significance of the immune inhibition release by checkpoint inhibitors. Similar to previous work on melanoma (Johnson et al, CCR 2018), we find that PD-1/L1 co-localization is associated with clinical benefit in NSCLC. In addition, work shows this co-localization predominantly occurs in CD68 positive macrophages and the clinical benefit is associated with CD68 cell expressed PD-L1 instead of cytokeratin cell expressed PD-L1, confirming our previous work showing PD-L1 expression in CD68 positive cells is associated with immune checkpoint inhibitor benefit (Liu et al, CCR 2019). Citation Format: Yuting Liu, Jon Zugazagoitia, Kurt A. Schalper, Roy S. Herbst, David L. Rimm. Assessment of the co-localization between PD-L1 positive macrophages and PD-1 positive cells is associated with benefit from PD-1 axis immunotherapy in NSCLC [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 2002.