Use of geospatial analysis with multiplex immunofluorescence to analyze tumor immune microenvironment patterns and prognosis in penile cancer.

Abstract

12 Background: Previous studies utilizing multiplex Immunofluorescence (mIF) have described cell density patterns of immune exhaustion implicating both innate and adaptive immune system components across varying stages of penile squamous cell carcinoma (PSCC). Here we studied geospatial clustering patterns of various immune effector cells within the tumor immune microenvironment (TIME). Methods: Tissue microarray (TMA) was constructed for 57 cases of invasive PSCC and immunostained for CD20, CD3, CD4, CD8, CD45RO, CD68, CD206, CD163, NKp46, FOXP3 using OPAL TM 7 kit (AKOYA Biosciences). Areas of tumor and stroma were identified using an image analysis system (InForm 2.2.4). Spatial analysis and co-clustering of cell phenotypes was performed using Ripley’s K. Survival curves were determined using Kaplan Meier method and tested using logrank and cox regression. Univariate and bivariate analysis describes co-clustering and proximity to tumor cells, respectively. Results: 57 PSCC patients (median age 60, [IQR (interquartile range) 31-92]); 60% HPV negative, and 31/57 with pN+ had mIF analysis. A high clustering pattern of CD68+, a general marker for macrophages, was associated with a significant OS benefit on univariate analysis (162 vs 36 mos, p=0.006) and an OS, RFS, and CSS benefit on bivariate analysis (84 vs 27 mos; NA vs 14 mos; NA vs 27 mos p=0.007, 0.014, 0.001). High clustering of CD3+CD4+, a marker for T-helper-lymphocytes, within the tumor and the tumor-stroma interface was associated with an OS benefit on bivariate analysis (84 v 20 mos, p=0.009). Low intra-tumoral clustering of CD68+CD163+, marker for M2 protumor macrophages, was associated with improved RFS on bivariate analysis (NA v 15 mos, p=<0.01). Low clustering of CD68+CD163+ in HPV+ patients was associated with a CSS benefit compared to high clustering in HPV- patients (p=0.024). High versus low clustering of CD68+CD163+ was inconsequential in HPV- patients (p=0.66).Overall, low intratumoral clustering pattern of FOXP3+, a marker for regulatory immune-suppression, was associated with an improved CSS on bivariate analysis (p=0.017). Low clustering of FOXP3+ in HPV+ patients was associated with an improved CSS versus low clustering of FOXP3+ in HPV- patients (p=0.041). Conclusions: Spatial analysis shows that proximity of CD68+ and CD3+CD4+, a general marker for macrophages and T-helper cells, to tumor cells has a positive impact survival in PSCC. Interestingly, proximity of protumor M2 macrophages to cancer cells confers a survival benefit specifically in HPV+ tumors and has no impact on survival in HPV- tumors. Decreased intratumoral FOXP3+ activity, a marker for regulatory immune-suppressive T cells, shows a preferential survival benefit in in HPV+ tumors. These findings point to diverging tumorigenic pathways related to HPV status that may be attributable to their poorer prognosis.