Accumulation of blood-circulating PD-L1-expressing M-MDSCs and monocytes/macrophages in pretreatment ovarian cancer patients is associated with soluble PD-L1.

BACKGROUND: Previous studies have shown clinical relevance of programmed death-ligand 1 (PD-L1) and soluble PD-L1 (sPD-L1) in human cancers. However, still contradictory results exist. Our aim was evaluation of PD-L1-expressing monocytic myeloid-derived suppressor cells (M-MDSCs), monocytes/macrophages (MO/MA), tumour cells (TC) and immune/inflammatory cells (IC) as well as investigation of the sPD-L1 in ovarian cancer (OC) patients. METHODS: The group of 74 pretreatment women were enrollment to the study. The expression of PD-L1 on M-MDSCS and MO/MA was assessed by flow cytometry. The profile of sPD-L1 was examined with ELISA. The expression of PD-L1 in mononuclear cells (MCs) was analyzed using real time PCR. PD-L1 immunohistochemical analysis was prepared on TC and IC. An in silico validation of prognostic significance of PD-L1 mRNA expression was performed based microarray datasets. RESULTS: OC patients had significantly higher frequency of MO/MA versus M-MDSC in the blood, ascites and tumour (each p < 0.0001). In contrast, PD-L1 expression was higher on M-MDSCs versus MO/MA in the blood and ascites (each p < 0.0001), but not in the tumour (p > 0.05). Significantly higher accumulation of blood-circulating M-MDSC, MO/MA, PD-L1+M-MDSC, PD-L1+MO/MA and sPD-L1 was observed in patients versus control (p < 0.001, p < 0.05, p < 0.001, p < 0.001 and p < 0.0001, respectively). Accumulation of these factors was clinicopathologic-independent (p > 0.05). The expression of PD-L1 was significantly higher on IC versus TC (p < 0.0001) and was clinicopathologic-independent (p > 0.05) except higher level of PD-L1+TC in the endometrioid versus mucinous tumours. Interestingly, blood-circulating sPD-L1 positively correlated with PD-L1+M-MDSCs (p = 0.03) and PD-L1+MO/MA (p = 0.02) in the blood but not with these cells in the ascites and tumours nor with PD-L1+TC/IC (each p > 0.05). PD-L1 and sPD-L1 were not predictors of overall survival (OS; each p > 0.05). Further validation revealed no association between PD-L1 mRNA expression and OS in large independent OC patient cohort (n = 655, p > 0.05). CONCLUSIONS: Although PD-L1 may not be a prognostic factor for OC, our study demonstrated impaired immunity manifested by up-regulation of PD-L1/sPD-L1. Furthermore, there was a positive association between PD-L1+ myeloid cells and sPD-L1 in the blood, suggesting that sPD-L1 may be a noninvasive surrogate marker for PD-L1+myeloid cells immunomonitoring in OC. Overall, these data should be under consideration during future clinical studies/trials.