Immunohistochemical Expression of IDO and PD-L1 in Distinct Compartments of Breast Cancer Tissue: Correlation with Clinicopathological Features and Outcomes

Background: Indoleamine 2,3-dioxygenase (IDO) is an immune checkpoint that has been shown to play a key immunomodulatory role in various solid tumors, including breast cancer (BC). Although increased IDO expression has been previously observed in some BC subtypes, mainly triple-negative BC (TNBC), the clinical relevance of this protein across the entire range of BC and its exact correlations with other immune checkpoints remain to be elucidated. We herein aimed to further investigate the differential expression patterns of IDO and programmed death-ligand 1 (PD-L1) in variable BC subtypes and in distinct compartments of breast cancer tissue, and to explore their potential associations with standard patient- and tumor-related clinicopathological parameters as well as prognosis. Methods: This was a retrospective multi-center cohort study of 150 female patients with BC. The clinicopathological parameters analyzed were retrieved from the medical records of patients while sections from archival formalin-fixed, paraffin-embedded (FFPE) tissue blocks were also obtained for the performance of immunohistochemistry. The expression of IDO and PD-L1 was evaluated separately on tumor cells (IDO/CA, PD-L1/CA), lymphocytes (IDO/L, PD-L1/L) and stromal cells (IDO/S, PD-L1/S) and the results were correlated with the remaining clinical and pathological features of patients, as well as with local recurrence, metastasis and survival. Results: The mean age of patients was 59.5 years (SD = 13.4 years). Positive expression of IDO/CA, IDO/L and IDO/S was found in 6%, 93.3% and 90.7% of tissue samples, respectively, while 4%, 11.2% and 6.7% of tumors were positive for PD-L1/CA, PD-L1/L and PD-L1/S, respectively. A significantly higher rate of positive IDO/CA expression was observed in triple-negative BC (TNBC) patients (p = 0.037). Positive expression of IDO-CA was also significantly associated with positivity for PD-L1/L and PD-L1/S (p = 0.001 and p = 0.015, respectively). Multivariable logistic regression analysis showed independent correlations between IDO/CA and IDO/L and the presence of invasive ductal carcinoma (IDC) (OR = 1.10; p = 0.026) and N1 status (OR = 10.93; p = 0.039), respectively, IDO/S and both N1 (OR = 14.64; p = 0.018) and positive HER2 status (OR = 6.11; p = 0.019), PD-L1/L and high Ki67 (OR = 7.96; p = 0.001) as well as negative ER (OR = 0.08; 0.003) and PR status (OR = 0.09; p = 0.002), PD-L1/S and both NST (no special type) histology (OR = 4.68; p = 0.032) and negative ER status (OR = 0.21; p = 0.044). No statistically significant associations were observed between the expression patterns of the examined biomarkers and recurrence, metastasis or survival. Conclusions: In our study, IDO expression on tumor cells was predominantly observed in TNBC and was found to correlate with PD-L1 expression in the lymphocytic and stromal compartments. Furthermore, expression of PD-L1 among lymphocytes was found to independently correlate with unfavorable clinicopathological parameters, including high proliferation rate and negative hormone receptor status.