CD137+ tumor infiltrating lymphocytes predicts ovarian cancer survival

Abstract

Objective: Ovarian cancer (OC) is the leading cause of death from gynecologic malignancy in the United States, and biomarkers of patient outcomes are limited. Data using immunohistochemical (IHC) analysis are mixed regarding whether and which tumor infiltrating lymphocytes (TILs) impact survival, and IHC does not adequately quantify rare cell populations, including CD137+ (4-1BB) tumor-reactive TILs. Our study investigates if a higher percentage of CD3+ CD137+ TILs is associated with improved overall survival (OS) in OC.

Methods: Flow cytometry was performed on viably banked OC digests. Chart review and statistical analysis were performed. Forty-seven patients were included, 40 of whom were diagnosed with high-grade serous ovarian carcinoma (HGSOC), papillary serous carcinoma, or undifferentiated histology.

Results: A high percentage of CD3+ CD137+ TILs correlated with improved OS (n = 40, r = 0.48, P = 0.0016). Subjects were divided into CD3+ CD137+ TIL high and low groups by the median. Subjects with high CD3+CD137+ TIL frequencies (>9.6%) had longer OS (Wilcoxon rank-sum test; P = 0.0032) and improved OS (logrank test; P = 0.007). Differences in CD3+ or CD3+ CD8+ TILs did not impact survival. CD3+ CD137+ TILs were predictive of OS regardless of germline mutation or debulking status. Analysis of subgroups including late stage HGSOC and late stage HGSOC with primary optimal cytoreduction indicated CD3+ CD137+ TILs correlated with improved OS after adjusting for age and PARP inhibitor use (P = 0.034 and P = 0.016, respectively).

Conclusions: Prevalence of CD3+ CD137+ TILs in digested OC specimens is associated with improved OS, while general TIL markers are not. CD137 has the potential to be a novel biomarker for survival in OC.

Keywords: CD137; Ovarian cancer; TILs; Tumor infiltrating lymphocytes.

Figure 1:. Associations between TIL populations and overall survival in subjects with high-grade ovarian carcinoma.

Analysis included subjects with HGSOC, papillary serous, or undifferentiated ovarian cancer histology (n = 40). The distinction between high and low populations was determined by the median value for each TIL population (CD3+CD137+ (Median 9.6, Range 0.2 – 40,0), CD3+ (Median 37.7, Range 0.6 – 89%), and CD3+CD8+ (Median 33.2, Range 16.9 – 63.9). A) Pearson correlation of TIL populations and overall survival in months. B) Wilcoxon analysis comparing overall survival based TIL populations. C) Kaplan-Meier curve of high versus low TIL populations. D) Kaplan-Meier Plotter online tool generated Kaplan-Meier curves based on high or low gene expression of TNFRSF9 (CD137), CD3E, and CD8A (n = 975).

Figure 2:. Impact of germline mutation and debulking status on CD3+CD137+, CD3+, and CD3+CD8+ TILs and overall survival in high-grade ovarian carcinoma subjects.

Subjects with either HGSOC, papillary serous, or undifferentiated histology were included for analysis (n = 40). A-B) Subjects with known germline genetic testing were evaluated (n = 30). Germline genetic mutations identified in our population were BRCA1, BRCA2, and RAD51C. Subjects with germline positive and negative were evaluated independently. A) Pearson correlation between increasing proportion of TIL populations and overall survival. B) Kaplan-Meier curves revealing the probability of survival based on mutational status and high versus low TIL subgroup. C-D) Subjects with known optimal (n = 35) or suboptimal (n = 10) debulking were evaluated independently. C) Pearson correlation of increasing proportion of TIL subtype and overall survival, stratified by debulking status. D) Kaplan-Meier curve of high and low TIL subgroup proportion stratified by debulking status.

Figure 3:. Associations between TIL populations and overall survival in subjects with high-grade ovarian carcinoma diagnosed at late stage who underwent primary, optimal debulking surgery.

Analysis included subjects with HGSOC, papillary serous, or undifferentiated ovarian cancer histology diagnosed at stage IIIC+, who were optimally debulked and did not receive neoadjuvant chemotherapy (n = 22). The distinction between high and low populations was determined by the median value for each TIL population (CD3+CD137+ (Median 9.6, Range 1.2 - 31.1), CD3+ (Median 37.9, Range 12.2 - 89.1), and CD3+CD8+ (Median 34.5, Range 16.9 - 63.9). A) Pearson correlation of TIL populations and overall survival in months. B) Wilcoxon analysis comparing overall survival based on TIL populations. C) Kaplan-Meier curve of high versus low TIL populations. D) Kaplan-Meier Plotter online tool generated Kaplan-Meier curves based on high or low gene expression of TNFRSF9 (CD137), CD3E, and CD8A (n = 495).