Novel HLA-Ilo/HLA-IIhi phenotype and immune evasion in triple-negative breast cancer.

Abstract

1095 Background: Immune checkpoint inhibitors (ICIs) are a therapeutic option for patients with triple-negative breast cancer (TNBC), however resistance and non-response remain an unmet clinical need. Human leukocyte antigens (HLAs) play an important role in induction and regulation of host immune responses. Downregulation of HLA class I expression is one mechanism by which cancer cells evade the immune control. HLA typing and understanding the role of HLAs can provide valuable insight into optimizing immunotherapy efficacy in TNBC. In this study, we evaluate the association of HLA I/II phenotypes with biomarkers of tumor immune interaction and overall survival (OS) in a real-world breast cancer cohort. Methods: A targeted RNA-seq assay was performed on 142 breast cancer samples. HLA I and HLA II expressions were estimated as average expression of HLA-A, HLA-B, HLA-C and HLA-DP, HLA-DQ, HLA-DR genes, respectively, by a percentile rank values [0-100]. Cohen D index was used to estimate ratio of HLA-I and II in each specimen. We particularly focused on the HLA-Ilow(lo)/HLA-IIhigh(hi) ratio given a proposed mechanism of immune escape by tumor cells (low expression of HLA-I) in response to a highly immune infiltrated state where there is increased expression of HLA-II by the immune cells. We studied the co-expression of HLA-Ilo/HLA-IIhi with other checkpoint targets using Spearman correlation(rs). The impact of HLA phenotypes on OS were analyzed using Kaplan-Meier analysis and Chi-squared assessed associations. Results: The cohort consisted of 106 non-TNBC (ER+/HER2- or HER2+) and 36 TNBC tumors. Most samples originated from metastatic visceral sites (96, 68%) followed by primary breast (33, 23%), and lymph nodes (13, 9%). Among TNBC, HLA-Ilo/HLA-IIhi phenotype displayed a significantly higher co-expression of immune checkpoint targets such as PD-1, PD-L1, TIM3 and LAG-3 than HLA-Ihi/HLA-IIlo phenotype (p<0.001). However, this difference was not observed between HLA-Ilo/HLA-IIhi and HLA-Ihi/HLA-IIlo phenotypes in non-TNBC tumors. In TNBC, there was no significant difference in OS between HLA-Ilo/HLA-IIhi when compared to HLA-Ihi/HLA-IIlo (43 vs. 37 months, p=0.21). Similarly, in non-TNBC, there were no significant differences in OS between HLA-Ilo/HLA-IIhi when compared to HLA-Ihi/HLA-IIlo (107 vs. 66 months, p=0.5). Conclusions: Concurrent loss of HLA I with increased HLA II expression was associated with co-expression of immune checkpoint molecules indicative of immune escape mechanism but not survival outcomes in TNBC. This same effect was not seen in non-TNBC with a more immunologically silent environment. Characterization of HLA phenotypes by immune profiling provides insight into the complex interplay of tumor immune microenvironment and offers an opportunity for developing novel approaches to overcome the immune evasion with cancer vaccines and adoptive cell therapies in TNBC.