Molecular characterization of STEAP1 and -2 in advanced prostate cancer.

5072 Background: STEAP 1 and 2 (six-transmembrane epithelial antigen of prostate) are metalloreductase proteins involved in a variety of biologic processes. STEAP1/2 are tumor-associated cell surface antigens highly expressed in prostate cancer (PC), although their role in cancer is poorly understood. STEAP1/2 have emerged as successful targets for adoptive T-cell therapy trials for PC. We employed a multi-omics approach to investigate the molecular features associated with STEAP1 and STEAP2 expression in PC. Methods: NextGen Sequencing of DNA (592 genes or whole exome) and RNA (whole transcriptome) was performed for PC tumors (n = 7089) submitted to Caris Life Sciences (Phoenix, AZ). PC samples were stratified by STEAP1/2 mRNA levels into top (high) and bottom quartile (low). Immune cell infiltration in the tumor microenvironment (TME) was inferred by quanTIseq. Transcriptomic signatures of androgen receptor signaling (AR), neuroendocrine classification (NEPC), and interferon gamma signaling (IFN) were calculated. Mann-Whitney U and X2/Fisher-Exact tests were applied where appropriate, with P-values adjusted for multiple comparisons (q< .05). Results: Of 7,089 samples, 63.2% were from the prostate; 11.7% from lymph node metastases (LNM); 7.3% from bone; and 17.8% from visceral/soft tissue metastases (V/STM). STEAP -1 and -2 were significantly correlated to each other (R= 0.90, p<.001), with significantly higher expression of STEAP1 observed in primary prostate and LNMs, compared with reduced expression in V/STM ( STEAP1 TPM: 105.2 vs 140.6 vs 91.9 p<.001). Mutations in AR (3.8% v 1.9%), KDM6A (4.2% v 2.2%), SPOP (10.9% v 8.4%) and AR V7 (23.0% v 10.5%) were enriched in STEAP1 high PC (each q<.01). Mutations in KDM6A (3.8% v 2.5%) and AR V7 (17.6% v 14.3%) were enriched in STEAP2 high PC (each q<.05). STEAP1/2 expression negatively correlated with TMB count (R =-0.03, p<.05) and IFN score (R = -0.26, p<.001). Concordantly, fewer proinflammatory immune cell fractions (M1 Macrophages, NK cells, CD4+/CD8+ T cells, myeloid dendritic cells) were observed within the TME of STEAP1/2 high PC (p<.0001). However, STEAP1/2 expression correlated positively with the AR signature (R = 0.39, p<.001) and androgen response pathways, while correlating negatively with the NEPC signature (R = -0.15, p<.001). Conclusions: PC tumors expressing high STEAP1/2 display distinct genomic and transcriptomic profiles compared to STEAP1/2 -low PC, and STEAP 1/2 expression varies across sites of metastases. Immune biomarkers and immune cell infiltration data suggest that STEAP1/2 may be associated with a cold TME. The recent success of STEAP1-targeting T-cell redirecting therapies mechanisms by which adoptive T-cell strategies may overcome immunosuppressive factors within the TME. Ongoing development of T-cell immunotherapeutics targeting STEAP1 may account for the differential expression profiles in guiding patient selection and combination strategies.