Analysis of CA9 and PSMA expression in clear cell renal cell carcinoma (ccRCC) and the effect of prior therapies.

547 Background: In ccRCC, carbonic anhydrase 9 (CA9) and prostate-specific membrane antigen (PSMA) have emerged as potential therapeutic targets. CA9 is overexpressed in ccRCC tumor cells, while PSMA, classically linked to prostate cancer, is expressed in ccRCC tumor-associated vasculature. An ongoing phase 1/2 trial (NCT06245915) is evaluating AB-2100, a programmable circuit T cell therapy using a PSMA-gated CA9 chimeric antigen receptor (CAR), for the treatment of ccRCC. We analyzed CA9 and PSMA expression in primary and metastatic ccRCC, assessing whether expression is retained in metastases and after prior treatment with immunotherapy (IO) and/or tyrosine kinase inhibitors (TKI). Methods: Single-plex immunohistochemistry (IHC) for CA9 and PSMA was performed on serial ccRCC samples. CA9 was scored by the percentage of tumor cells with membranous staining (0–100%), and PSMA by vascular staining positivity (0–3+). A total of 103 ccRCC samples (17 primary, 86 metastatic) from 78 patients (pts) were analyzed. Paired Wilcoxon test compared matched primary-metastatic pairs (n=16; most recent if multiple available). Independent metastatic samples (one per patient, most recent if multiple available, n=77) were analyzed using the Wilcoxon rank-sum test to compare CA9 and PSMA expression in treatment-naive versus treatment-exposed samples. Correlation between the two markers was also evaluated (Spearman’s ρ). Results: Across the 78 pts, 61 received a TKI-based therapy, 46 an IO-containing regimen, and 13 a HIF-2α inhibitor. Of the 103 samples, 54.4% (n=56) were treatment-naive and 45.6% (n=47) were collected after prior exposure to systemic therapy. CA9 expression was uniformly high (median=100%, IQR: 99-100%, range: 10-100%) across 103 evaluable samples. Among matched pairs (n=16), there was no significant loss of CA9 expression in metastatic samples compared to their corresponding primary tumors (paired Wilcoxon test, p = 0.78). PSMA was detected in the tumor vasculature in 102 evaluable cases (median = 2, IQR: 2-3), with comparable expression between 16 matched pairs (paired Wilcoxon test, p = 0.48). No correlation between CA9 and PSMA was found (Spearman's ρ = -0.17). Expression of both markers was not significantly affected by prior TKI- or IO-containing regimens (Table 1). Conclusions: CA9 and PSMA expression is consistently retained in metastatic ccRCC compared to primary tumors, even after systemic treatment. CA9 remains expressed in tumor cells, while PSMA is observed in the tumor vasculature, supporting the rationale for their therapeutic targeting in advanced ccRCC. Treatment-exposed vs naive samples. CA9 PSMA Treatment Group n Median p* n Median p* Naive (ref) 35 100.0 — 35 2.0 — TKI only 11 100.0 0.879 11 2.0 0.231 TKI-containing regimen 35 100.0 0.119 34 2.0 0.434 IO-containing regimen 26 100.0 0.152 26 3.0 0.293 *P-values from Wilcoxon rank-sum test vs naive (reference group).