Controversial Role of the Immune Checkpoint OX40L Expression on Platelets in Breast Cancer Progression

Abstract

In conventional T cells, OX40 has been identified as a major costimulating receptor augmenting survival and clonal expansion of effector and memory T cell populations. In regulatory T cells, (Treg) OX40 signaling suppresses cellular activity and differentiation. However, clinical trials investigating OX40 agonists to enhance anti-tumor immunity, showed only limited success so far. Here we show that platelets from breast cancer patients express relevant levels of OX40L and platelet OX40L (pOX40L) inversely correlates with platelet-expressed immune checkpoint molecules GITRL (pGITRL) and TACI (pTACI). While high expression of pOX40L correlates with T and NK cell activation, elevated pOX40L levels identify patients with higher tumor grades, the occurrence of metastases, and shorter recurrence-free survival (RFS). Of note, OX40 mRNA levels in breast cancer correlate with enhanced expression of anti-apoptotic, immune-suppressive, and tumor-promoting mRNA gene signatures. Our data suggest that OX40L on platelets might play counteracting roles in cancer and anti-tumor immunity. Since pOX40L reflects disease relapse better than the routinely used predictive markers CA15-3, CEA, and LDH, it could serve as a novel biomarker for refractory disease in breast cancer.

Keywords: OX40L; biomarker; breast cancer; immunotherapy; platelets; prognosis.

Figure 1

pOX40L in breast cancer patients is associated with a distinct platelet immunophenotype and activation of cytotoxic lymphocytes. (A–C) Comparative analysis of the platelet immune checkpoint molecules pOX40L, pTACI and pGITRL (n = 65). The gating strategy used to analyze pOX40L expression on platelets is given in Supplementary Figure 3 . Co-expression of TACI (B) or GITRL (c) on pOX40L^low and pOX40L^high platelets, respectively. (D–J) Level of pOX40L in healthy donors (D) and breast cancer patients (G) and expression of CD69 (E, F) and CD25 (H, I) on NK cells and T cells of the respective donors (n = 10 each). (J–Q) Co-regulation of CD69/CD25 on NK/T cells and pOX40L in healthy donors and breast cancer patients (n = 10 each). r-t Comparative analysis of platelet activation in the presence or absence of TRAP-6 (10 µM) with regard to pOX40L status. Representative platelet aggregation studies (R, S) were performed in pOX40L high and low expressing platelets derived from breast cancer patients. Aggregation studies (T) were performed in platelets derived from 5 healthy donors (green) and 10 breast cancer patients (pOX40L low in blue, pOX40L high in red). (U) Expression of CD62P and OX40L on platelets from HD and breast cancer patients ex vivo (n = 10 each). (V) Expression of OX40L on platelets from healthy donors and the complete breast cancer cohort ex vivo (n = 25 and n = 65, respectively). (W) pOX40L level in healthy donors (HD), (left panel) and breast cancer patients (right panel) with regard to CD62P expression (n = 25 and n = 65, respectively). (X) Correlation of pTACI expression and platelet count (left panel) and platelet size (right panel). (B, C), (E, F), (H, I) Each dot represents a single patient. Boxes represent median and 25th to 75th percentiles, whiskers are minimum to maximum.

Figure 2

| pOX40L predicts disease relapse and metastasis in breast cancer. (A) pOX40L expression in different tumor stages (T), (n = 65) (B) Association of pOX40L expression and lymph node invasion (N), (n = 65). (C) pOX40L level in patients with different tumor grading (G), (n = 65). (D) Correlation between pOX40L expression on platelets of 65 BC patients and the Ki67 positive cells (%). (E) pOX40L expression in patients with and without metastatic disease (M), (n = 65). (F) Kaplan–Meier curves estimates of RFS (months) in patients with a pOX40L > mean (pOX40L^high = red) and pOX40L level < mean (pOX40L^low = blue), (n = 65). Median time to metastasis (pOX40L^high = red, pOX40L^low = blue). (G) pOX40L expression in patients with different number of metastasis (n = 65). (H) Mean expression of pOX40L in different metastatic organs (n = 20). (I) Heatmap depicting the relative mRNA level OX40 and a gene set associated with lung metastasis, anti-apoptotic and tumor-promoting signatures in breast tumors (TCGA-Firehose legacy), (n = 65). (J) Spearman correlation mRNA gene signatures and OX40 mRNA. (K–P) Scatter plot of OX40 mRNA and PD-L1 (K), CD80 (l), CD86 (M), VISTA (N), HVEM (O) and TGFß (P) mRNA level in breast cancer. Each dot represents a single patient. (A–E, G, K–P) Each dot represents a single patient. (A–E, G) Data are mean ± SEM.

Figure 3

Predictive value of pOX40L in breast cancer relapse. (A) Comparative analysis of pOX40L and the clinical markers HER2, ER, CA 15-3, CEA and LDH. (B) Association of pOX40L expression and different breast cancer subtypes. (C-E) Correlation of pOX40L and conventional tumor marker including CA15-3 (C), CEA (D) and LDH (E). (F–K) Predictive value of pOX40L (F), HER2 (G), ER (H), CA15-3 (I), CEA (J) and LDH (K) expression for disease relapse was analyzed using ROC (Area under the ROC curve) analysis.