SNW1 is a prognostic biomarker in prostate cancer

Prostate cancer was estimated to be the most frequent diagnosed cancer and the third most common cause of cancer related death in men with western lifestyle in 2012 [1]. About 1 in 5 patients diagnosed with prostate cancer probably dies due to the cancer in developed countries. Established pretreatment prognostic parameters include Gleason grade and tumor extent on biopsies, preoperative prostate-specific antigen (PSA) levels, and clinical stage. Unfortunately, a reliable distinction between indolent and aggressive prostate cancer is still not possible in individual patients. It is hoped that upcoming molecular markers may enable a better prediction of prostate cancer aggressiveness.

SNW domain-containing protein 1 (SNW1) alias SKI-interacting protein 1 (SKIP1) is a multifunctional protein implicated in the regulation of several genes and pathways connected to cell growth and homeostasis. It acts either by direct protein interactions, by mRNA splicing regulation, or by transcriptional control. SNW1 has been originally named after its direct interaction with the SKI oncogene in the regulation TGF-ß signaling [2, 3], but it also associates with the retinoblastoma tumor suppressor protein to impair its function [4], and stabilizes beta-catenin to enhance WNT-signaling [5]. In its function as a selective splicing factor for p21, SNW1 is known to counteract p53-mediated apoptosis [6]. Finally, SNW1 is an important co-regulator for nuclear receptors [7] including vitamin D receptor [8], progesterone and androgen receptor [9]. Given these numerous interactions with pathways relevant in cancer, it is not surprising that SNW1 can be up regulated in many cancer types. For example, overexpression of SNW1 was reported in 43% of breast cancers [10], 50% of liver cancers [11], 30% of bladder cancers [12] and 40% of malignant pleural mesothelioma [13]. High SNW1 expression was associated with high tumor grade and proliferation in breast cancer [10] and with poor prognosis in the other above-mentioned tumor types [11‐13]. Given its functional association with the androgen receptor, it appears possible that SNW1 expression may also be relevant in prostate cancer biology. Studies investigating the potential prognostic impact of SNW1 in clinical prostate cancer samples are currently missing.

To test the potential of SNW1 as a clinical marker in prostate cancer, the validated antibody HPA002457 from the Human Tissue Atlas Project [14] was employed for an immunohistochemical analysis of a highly annotated tissue micro array (TMA) with more than 12,000 prostate cancers.

A total of 10,310 (83%) tumor samples were interpretable in our TMA analysis. Non-informative cases (2117; 17%) lacked tissue samples or unequivocal cancer cells in the TMA spot (Additional file 1: Table S1). Normal prostate glands showed variable levels of cytoplasmic and nuclear SNW1 staining, ranging between negative and moderate intensity (Additional file 1: Figure S1). In prostate cancers, positive SNW1 staining was found in 83.1% of cases, and was considered weak in 31.5%, moderate in 37.7% and strong in 14% of cases. Representative images are shown in Fig. 1.

The results of our study demonstrate that SNW1 overexpression is linked to aggressive tumor features and poor prognosis in prostate cancers. Our immunohistochemical analysis revealed at least weak SNW1 staining in 83% of our 10,310 interpretable prostate cancers. Normal prostate glands showed variable levels of SNW1 expression, ranging from negative to moderate staining in our study. That a strong SNW1 immunostaining - at a level not observed in normal prostate epithelium - was found in 14% of prostate cancers suggests that SNW1 becomes up regulated during tumor development in a fraction of prostate cancers. Comparable data derived from immunohistochemical analyses of SNW1 in prostate cancer are currently lacking in the literature. Data from RNA sequencing of prostate cancers do not show major SNW1 mRNA expression changes [21], suggesting that stabilization may account for the increased levels identified by immunohistochemistry. However, data supporting SNW1 up regulation in malignant tumors comes from breast cancer [10], hepatocellular carcinoma [11], bladder cancer [12] and malignant pleural mesothelioma [13].

Highly significant statistical associations of SNW1 overexpression with unfavorable tumor phenotype, accelerated cell proliferation and poor clinical outcome in our cohort of more than 10,000 patients identifies SNW overexpression as a sign for increased cancer aggressiveness. These observations fit well to the role of SNW1 as co-activator of known oncogenic pathways [2, 3], such as transforming growth factor β (TGF-β) [2] and WNT signaling [5]. Functional studies have further demonstrated, that SNW1 counteracts p53-mediated apoptosis [6] and inhibits the transcriptional repressor activity of the retinoblastoma tumor suppressor gene [4]. Accordingly, associations of SNW1 up regulation with unfavorable tumor features have also been reported from several other malignancies. For example, SNW1 overexpression is linked to high tumor grade and accelerated cell proliferation in breast cancer [10], and to poor prognosis in urinary bladder cancer, hepatocellular carcinoma, and malignant pleural mesothelioma [11‐13].

The availability of a molecular database attached to this prostate cancer TMA enabled us to investigate the role of SNW1 in molecularly defined cancer subgroups, the most relevant of which are ERG positive and ERG negative cancers. TMPRSS2:ERG fusions occur in about 50% of prostate cancers [22]. They predominantly occur in younger patients [18] and lead to a constitutive overexpression of the transcription factor ERG [22]. ERG overexpression by itself lacks any prognostic impact, at least in patients not receiving systemic therapy. However, ERG modulates the expression of more than 1600 genes in prostate epithelial cells ERG [23‐25]. The biological effects of various proteins may be mitigated or intensified in such a modified cellular microenvironment. Our data identify SNW1 expression as an ERG-dependent feature being clearly more prominent in ERG positive than in ERG negative cancers. It appears possible, that SNW1 up regulation in ERG-fusion positive tumors is mediated by ERG-dependent activation of TGF-ß signaling. Earlier studies have shown that TGF-ß signaling is massively up regulated in ERG-positive cancers [24], and that TGF-ß1 increases SNW1 expression in mouse models [26]. The strong association between SNW1 overexpression and increased AR expression observed in our study fits well to earlier reports on a functional relationship between these proteins. For example, Abwanka et al. [9] demonstrated that binding of testosterone to the AR and nuclear translocation of activated AR is massively increased when SNW1 forms complexes with the AR.

The prognostic effect of SNW1 expression was stronger in ERG negative than in ERG positive cancers but also retained in the latter group. A modified cellular microenvironment may serve as an explanation for the particularly strong prognostic role of SNW1 expression in ERG negative cancers. In earlier studies we had described several molecular features that were either prognostic in ERG positive (for example, SOX9, [27] and SENP1 [28] or in ERG negative cancers (for example, GGH [29] and NBS1 [30]) but not in both groups. An alternative explanation for different prognostic effects between ERG positive and ERG negative cancers is the experimental set-up. It cannot be excluded that our immunohistochemistry protocol was better suited to distinguish expression differences in cancers with generally lower expression levels (ERG negative cancers) than in those with higher expression (ERG positive cancers). For example, the group of SNW1 negative cancers – the best prognostic group – contained 1182 cancers in ERG negative, but only 164 cancers in ERG positive cancers. Irrespective of its cause, the different prognostic impact of SNW1 in ERG positive and ERG negative cancers demonstrates that the applicability (and perhaps thresholds) of prognostic markers may depend on individual tumor features. This clearly represents a challenge for the development of prognostic cancer tests that shall be applicable to every patient.

The data of this study suggest that SNW1 expression may represent a clinically useful marker. It is of note that the search for clinically useful prognostic markers is not predominantly about finding factors that are independent of established parameters. Most of all, parameters are needed that are more reproducible and, thus, reliable in individual patients. The established prognostic factors (Gleason grade, preoperative PSA-level, clinical stage, pathological stage and nodal status) are statistically strong but suffer from significant shortcomings in clinical practice. pT stage and nodal status are not available during the preoperative therapeutic decision-making process. The quality of pN data greatly depends on the extent of surgery and the pathological work-up of the removed materials. The Gleason Grade – i.e., the most powerful preoperatively available prognostic marker - suffers from very substantial inter-observer variability reaching beyond 40% in individual biopsies [31]. SNW1 expression measurement thus clearly has potential to become an element in a future multi-parametric prognostic test for prostate cancer. To this end some of the limitations in the present study (retrospective analysis of a single spot per patient from prostatectomy specimen) have to be addressed.

SNW1 overexpression occurred in a relevant fraction of prostate cancers. The moderate prognostic impact of SNW1 makes it a candidate for future marker panels in prostate cancer.