Prostate-specific membrane antigen expression in tumor-associated vasculature of breast cancers

Alla Gabriella Wernicke,Sonal Varma,Eleni A Greenwood,Paul J Christos,K S Clifford Chao,He Liu,Neil H Bander,Sandra J Shin

DOI: https://doi.org/10.1111/apm.12195

Apmis

Abstract:Prostate-specific membrane antigen (PSMA) has been found to be expressed in the tumor-associated neovasculature of multiple solid tumor types including breast cancers. However, thus far, the number of cases studied from some tumor types has been limited. In this study, we set out to assess PSMA expression in the tumor-associated vasculature associated with invasive breast carcinomas in a sizable cohort of patients. One hundred and six patients with AJCC stage 0-IV breast cancer were identified. Ninety-two of these patients had primary breast cancer [invasive breast carcinoma with or without co-existing ductal carcinoma in situ (DCIS) (74) or DCIS alone (18)]. In addition, 14 patients with breast cancer metastases to the brain were identified. Immunohistochemical staining for PSMA and CD31 was performed on parallel representative tumor sections in each case. Tumor-associated vascular endothelial cell PSMA immunoreactivity was semi-quantitatively assessed based on two parameters: overall percent of endothelial positivity and staining intensity. PSMA expression for tumor-associated vascular endothelial cells was scored 0 if there was no detectable PSMA expression, 1 if PSMA staining was detectable in 5-50%, and 2 if PSMA expression was positive in >50% of microvessels. CD 31 staining was concurrently reviewed to confirm the presence of vasculature in each case. Tumor-associated vasculature was PSMA-positive in 68/92 (74%) of primary breast cancers and in 14/14 (100%) of breast cancers metastatic to brain. PSMA was not detected in normal breast tissue or carcinoma cells. All but 2 cases (98%) showed absence of PSMA expression in normal breast tissue-associated vasculature. The 10-year overall survival was 88.7% (95% CI = 80.0%, 93.8%) in patients without brain metastases. When overall survival (OS) was stratified based on PSMA score group, patients with PSMA scores of 0, 1, and 2 had 10-year OS of 95.8%, 96.0%, and 79.7%, respectively (p = 0.12). When PSMA scores of 0 and 1 were compared with 2, there was a statistically significant difference in OS (96.0% vs 79.7%, respectively, p = 0.05). Patients with a PSMA score of 2 had a significantly higher median tumor size compared with patients in the lower PSMA score groups (p = 0.04). Patients with higher nuclear grade were more likely to have a PSMA score of 2 compared with patients with lower nuclear grade (p < 0.0001). Patients with a PSMA score of 2 had a significantly higher median Ki-67 proliferation index compared with patients in the lower PSMA score groups (p < 0.0001). Patients with estrogen receptor (ER)-negative tumors were more likely to have a PSMA score of 2 compared with patients with ER-positive tumors (p < 0.0001). Patients with progesterone receptor (PR)-negative tumors were more likely to have a PSMA score of 2 compared with patients with PR-positive tumors (p = 0.03). No significant association was observed between PSMA score group status and lymph node involvement (p = 0.95). Too little variability was present in Human epidermal growth factor receptor-2 (Her2/neu) amplified tumors to correlate with PSMA score group status. To date, this is the first detailed assessment of PSMA expression in the tumor-associated vasculature of primary and metastatic breast carcinomas. Further studies are needed to evaluate whether PSMA has diagnostic and/or potential therapeutic value.