PSCA expression is associated with favorable tumor features and reduced PSA recurrence in operated prostate cancer

While most prostate cancers have an indolent clinical course, the disease represents the third most common cause of cancer related death in men in Western societies [1]. Gleason grade and tumor extent on biopsies, preoperative prostate-specific antigen (PSA), and clinical stage are the currently established pretreatment prognostic parameters. Although these parameters are linked to cancer aggressiveness, the distinction between indolent and aggressive prostate cancer is difficult for the individual patient. Molecular marker may enable a better prediction of prostate cancer aggressiveness in the future.

Prostate stem cell antigen (PSCA) is a protein of unknown function anchored to the cell surface. It was discovered in an attempt to identify genes up regulated in human prostate cancer [2]. Though the name implies specificity for the prostate, PSCA is expressed in several tissues: Placenta, kidney, pancreas, and bladder [3‐5]. The function of PSCA has not been fully elucidated [6‐9]. Experiments suggest a possible role in cell adhesion, proliferation control and cell survival [2, 10]. Evidence is accumulating that – depending on the cell type involved – PSCA can have a tumor promoting or a tumor suppressive effect [11‐15]. For example, loss of PSCA was associated with poor outcome in cancer of the gallbladder and stomach [12, 16], but with improved prognosis in pancreatic adenocarcinoma, renal cell carcinoma and non-small lung cancer [17‐20]. The part of PSCA in prostate cancer remains unclear. Even if most available data suggest that prostate cancer may belong to the tumors with an oncogenic function of PSCA overexpression [21‐24], there are also studies that do not support such a conclusion [25] or suggest the opposite that prostate cancer aggressiveness and metastasis is driven by PSCA down regulation [26‐29].

To clarify the prognostic role of PSCA expression in prostate cancer, we analyzed PSCA expression by immunohistochemistry on a large preexisting tissue micro array (TMA).

A total of 9642 (70.6%) of TMA spots were interpretable. Non-informative cases (4023 spots; 29.4%) lacked tissue samples or unequivocal cancer tissue spots. PSCA staining was absent in glands, stromal tissue and inflammatory cells of the normal prostate. In cancers, positive PSCA staining was seen in 5581 of our 9642 (53.7%) interpretable tumors and was considered weak in 22.4%, moderate in 24.5% and strong in 6.8% of cancers. 4461 tumors (46.3%) showed no PSCA staining.

This study demonstrates that PSCA expression is significantly associated with favorable tumor phenotype and a reduced risk for PSA recurrence. A total of 54% of prostate cancers showed detectable PSCA expression in our IHC study. This is in the range of two other studies reporting in 88% of 126 [2] or 48% of 233 patients [4] IHC positivity. These IHC findings are not contradictory to further studies describing at least a low level PSCA expression in all prostate cancers utilizing polymerase chain reaction [5, 23]. It is well known that IHC negativity does not reflect the absence of expression but rather that a certain threshold for detection is not reached. The threshold of detection is greatly influenced by the IHC protocol [38]. Although Ross et al. [4] and Reiter et al. [2] used different antibodies and protocols; they obviously resulted in a comparable sensitivity as in our study. We consider our protocol suited for studying the prognostic impact of PSCA expression as the selected conditions enable a distinction of cancers with high and low levels of PSCA expression. Given that normal prostate epithelial cells usually did not stain for PSCA, we assume that PSCA up-regulation had occurred in a fraction of prostate cancers. The link between cancer specific up-regulation of a protein and better prognosis is uncommon and argues for a “protective” or tumor suppressive role. Our outcomes are in concordance with data from Larkin et al. [27], also reporting a link between elevated PSCA expression and favorable clinical course. A tumor suppressive role of high level PSCA expression is also supported from cell line models. For example, functional analysis of cell lines from gastric and gallbladder cancer demonstrated that forced overexpression of PSCA hampered cell proliferation [12, 16].

It is noteworthy that several investigators reported contrary results. IHC with various and partly custom made antibodies to conventional large sections of 40 [22] and 112 [5] prostate cancers revealed associations with high Gleason score, advanced stage and castration resistant disease. Also in TMA studies including 114 [23] and 246 [24] prostate cancers the authors reported associations between strong PSCA expression and high Gleason score. However, another TMA study on 64 prostate cancers could not confirm these findings [25]. We cannot explain the discrepancy between these studies and our data obtained on almost 10,000 successfully analyzed carcinoma.

The mechanism for a tumor suppressive function of PSCA is unknown. In our study, we compared the expression of PSCA with molecular attributes associated with genomic instability, chromosomal deletion and tumor cell proliferation [37]. We found previously that features with a role in cell cycle control (p16 [39] or APE1 [40]) were significantly associated with a high Ki67 labeling index. Molecular attributes linked to genomic instability (MSH6/PMS2/MLH1 [41], ELAV1 [42], or HOOK3 [43]) were found to be associated with chromosomal deletion. The lack of clear-cut association with these features argues against a relevant impact of PSCA on cell proliferation control and development of deletion in prostate cancer. This is in contrast to one experimental study, suggesting an accelerating effect of PSCA loss on cell proliferation in a gastric cancer cell line [16]. That PSCA expression was completely unrelated to TMPRSS2:ERG fusion further demonstrates that PSCA is not significantly affected by any of the hundreds of genes that are deregulated in ERG fusion positive prostate cancer [44‐47].

PSCA expression was significantly associated with favorable patient outcome in our cohort. A possible clinical relevance of this finding is supported by its statistical independence of classical prognostic markers, especially in a pre-operative disease state. However, in comparison with established features such as the Gleason score, the impact of PSCA expression on patient outcome was rather small. If traditional prognostic Gleason groups were used, a small prognostic impact was still found in Gleason 3 + 4 (p = 0.0206) or Gleason 4 + 3 (p = 0.0092). However, if these subgroups were further differentiated according to the fraction of Gleason 4 (quantitative Gleason grading [30]) these PSCA associated prognostic differences vanished. This further illustrates the high bar that molecular characteristics have to overcome if compared with optimized morphologic analysis.

PSCA expression is a statistically independent predictor of favorable prognosis in prostate cancer. Although its prognostic impact per se is not very strong, PSCA expression analysis could be considered for inclusion in multi-parametric prognostic tests to distinguish prostate cancers with need for radical therapy.