Integrated analysis of mismatch repair, PD-L1, and immune microenvironment status in pregnancy-associated breast cancers

Introduction: Breast cancer (BC) occurring during gestation or lactation is rare, yet highly challenging under both biological and clinical standpoints. This condition, referred to as pregnancy-associated (PA) BC (PABC), shows enrichment in mismatch repair (MMR) deficiency mutational signature, higher expression of immune-checkpoint genes, and less tumor-infiltrating lymphocytes (TILs) compared to non-PA BCs. Despite these insights, no comprehensive data on MMR protein status, immune checkpoints, and immune microenvironment are currently available for these tumors. The aim of this study was to characterize the MMR status and immunologic milieu of PABC. Methods: Among a multi-Institutional database comprising 142 PABCs, we conducted a comparative analysis of a cohort of PABC (n=29) and a control group of age-matched non-PA BCs (n=74). Distinct areas of each tumor and the corresponding normal breast tissue were incorporated into a tissue microarray (4-6 cores per case, mean 4.5). For all cases, both stromal and intratumoral TILs were quantified according to the International TILs Working Group recommendations. Representative slides were subjected to immunohistochemical (IHC) analysis of the MMR proteins (i.e. MLH1, MSH2, MSH6, and PMS2), programmed death-ligand 1 (PD-L1), CD4, and CD8. Cases were classified as MMR-proficient (pMMR), MMR-deficient (dMMR), and, when the protein was expressed only in a part of the tumor, MMR-heterogeneous (hMMR). PD-L1 expression was evaluated separately in the tumor cells, stromal TILs, and intratumoral TILs. Finally, the relative proportion of CD4+ and CD8+ cells was assessed in both stromal and intratumoral TILs. Results: The study group included 4 (14%) Luminal A, 10 (35%) Luminal B, 4 (14%) non-luminal HER2+, and 11 (37%) triple-negative (TN) PABCs. Taken together, both the dMMR and hMMR status were more common in PABCs than in non-PA BC (n=3/29, 19% vs. n=6/74, 8% and n=7/29, 24% vs. 14/74, 19%, respectively). Specifically, dMMR was seen in Luminal A (n=2, 50%) and in TN (n=1, 9%) PABCs. Conversely, in non-PA BCs, all Luminal A (n=15, 100%) were pMMR, while 5/45(11%) Luminal B and 1/13 (7%) TNBC were dMMR. Despite no differences were observed in intratumoral TILs, PABCs showed significantly higher levels of stromal TILs (p=0.01). Compared to the control group, PD-L1 expression in PABCs was significantly higher in the tumor cells and in stromal TILs (p<0.01) but not in intratumoral TILs. Hence, 9 (31%) PABCs were PD-L1+, with tumor proportion scores (TPS) ranging from 2 to 10 (mean 5), while only 4 (5%) non-PA BCs were PD-L1+ (TPS 2-20, mean 8)(p=0.01). The expression of PD-L1 in stromal TILs was higher in PABCs (n=9, 31%) than in the controls (n=19, 26%)(p=0.005). In Luminal tumors, both the CD4+ and CD8+ populations were more represented than in non-PA BCs (p=0.01), while in TNBC only the relative proportion of CD4+ cells was significantly higher (p=0.02). Discussion: This study is the first to investigate the immune response alongside the MMR status by IHC in PABCs. Our findings broaden the understanding of the immunobiology underpinning PABC, suggesting that in these tumors i) MMR protein alterations occur at higher frequency than in non-PA BCs; ii) the tumor cells and tumor microenvironment may be capable to suppress the adaptive arm of immune system through the expression of PD-L1; and iii) lymphocytes located at the periphery rather than those inside of the tumor are likely to be implicated in the immune modulation. Conclusion: The MMR system and immune microenvironment may play a consistent role in the natural history of PABCs. An intimate knowledge of the multifaceted interplay between tumor and tumor immune microenvironment is likely to unveil clinically relevant mechanisms that may have a positive net health impact for women with BC during gestation or lactation.