Loss of Somatostatin Receptor Subtype 2 in Prostate Cancer Is Linked to an Aggressive Cancer Phenotype, High Tumor Cell Proliferation and Predicts Early Metastatic and Biochemical Relapse

Somatostatin receptor subtype 2 (SSTR2) is the most frequently expressed SSTR subtype in normal human tissues. SSTR2 expression is differentially regulated in various tumor types and therapeutic somatostatin analogs binding to SSTR2 are in clinical use. In prostate cancers highly contradictory results in terms of SSTR2 expression and its consequences have been published over the past years. The aim of this study was to clarify prevalence and clinical significance of SSTR2 expression in prostate cancer. Therefore, quantitative immunohistochemistry (IHC) using a tissue microarray containing samples from 3,261 prostate cancer patients with extensive clinical and molecular cancer characteristics and oncological follow-up data was performed. IHC data was compared to publicly available Gene Expression Omnibus datasets of human prostate cancer gene expression arrays. While membranous SSTR2 staining was always seen in normal prostate epithelium, SSTR2 staining was absent in more than half (56.1%) of 2,195 interpretable prostate cancer samples. About 13% of all analyzed prostate cancers showed moderate to strong cytoplasmic and membranous SSTR2 staining. Staining intensities were inversely correlated with high Gleason grade, advanced pT category, high tumor cell proliferation (p<0.0001 each), high pre-operative PSA levels, (p = 0.0011) and positive surgical margins (p = 0.006). In silico analysis confirmed lower SSTR2 gene expression in prostate cancers vs. normal adjacent tissue (p = 0.0424), prostate cancer metastases vs. primary cancers (p = 0.0011) and recurrent vs. non-recurrent prostate cancers (p = 0.0438). PSA-free survival gradually declined with SSTR2 staining intensity (p<0.0001). SSTR2-negative cancers were more likely to develop metastases over time (p<0.05). In conclusion, most prostate cancers are indeed SSTR2-negative and loss of SSTR2 strongly predicts an unfavorable tumor phenotype and poor prognosis. Therefore, SSTR2 expression seems an important factor in the pathogenesis of prostate cancer and re-introduction of the receptor in SSTR2-negative prostate cancers may feature a promising target for novel gene therapy approaches.


Introduction
In Western males, prostate cancer is the most frequently diagnosed malignant tumor with the second highest rate of cancerattributed death [1]. Prostate cancers feature considerably variable courses of disease from slow local growth to aggressive and invasive metastatic proliferation. Therefore, a thorough molecular characterization of different prostate cancer subtypes is essential in order to discriminate between aggressive and non-aggressive cancer phenotypes.
Somatostatin (SST) is a cyclic neuroendocrine hormone which was originally isolated from sheep hypothalamus as in vitro inhibitor of growth hormone action (GH, [2]). SST is produced by neuroendocrine cells throughout the whole body and is therefore widely expressed. This comprises the central and peripheral nervous system and a variety of cells within the digestive, genitourinary and reproductive tracts [3].
SST exists in two biologically active isoforms (SST-14 and SST-28) which bind to 5 distinct membrane surface receptors (SSTR1-5) with variable binding affinities [4]. Upon activation by SST-14/-28 all SSTRs inhibit the generation of cyclo-AMP thereby diminishing mitogen-activated (MAP) kinase mediated cell proliferation in a broad variety of cell types [3]. The most widely expressed SSTR subtype in normal tissues [3] is SSTR2.
Various synthetic SSTR agonists are available, of which the anti-proliferative octapeptide octreotide is the most popular representative that is now routinely used for detection and therapy of neuroendocrine tumors [5]. In addition, by agonist profiling octreotide was identified as most potent binding partner of SSTR subtype 2 (SSTR2, [4]).
Using labeled octreotide binding assays, in situ hybridization, immunohistochemistry and RNA-binding assays SSTR2 expression has been shown to be also expressed in a variety of malignant and non-malignant tumors such as pituitary adenoma, meningeoma, neuroblastoma, Non-Hodgkin's lymphoma, neuroendocrine carcinoid tumors, breast tumors, renal, pancreatic and small cell lung cancers [3]. In these cancers SSTR2 activation leads to an inhibition of tumor cell proliferation, mostly mediated via growth arrests [3].
In prostate cancer there are multiple contradictory reports regarding the role and expression of SSTR2 (for a summary see [6]): Hansson et al., for instance, suggest an up-regulation of SSTR2 expression in prostate cancers [7], Morichetti and coworkers found increased SSTR2 immunostaining in 80% of incidental prostate cancers [8]. Other groups found diminished or absent SSTR2 expression in prostate cancers [9,10].
In the light of these conflicting reports, we analyzed SSTR2 expression by immunohistochemistry on a large (3,261 tumors) prostate cancer tissue microarray (TMA) and by publicly available Gene Expression Omnibus (GEO) datasets of human prostate cancer gene expression arrays in order to comprehensively clarify prevalence and clinical significance of SSTR2 expression in prostate cancers.

Tissue microarray construction
The prostate cancer prognosis tissue microarray (TMA) consisted of cancer samples from 3,261 patients distributed over 7 paraffin blocks. The sampling and constructions have been previously described in detail [11]. In brief, specimens from radical prostatectomies performed between 1992 and 2005 at the Department of Urology, University Medical Center Hamburg-Eppendorf were paraffin-embedded and afterwards matched with clinico-pathological data. TNM classification (American Joint Committee on Cancer, 2002, 6 th edition) was used for tumor staging in order to define primary tumor size and local invasiveness (pT), regional lymph node affection (pN) and distant spreading/metastatic disease (pM). Grading of cancers was performed using the Modified Gleason Score [12] and by evaluation of cancer-free surgical margins [13]. In all patients undergoing radical prostatectomy, prostate specific antigen (PSA, [14]) concentrations were measured at the time of diagnosis and for post-operative surveillance quarterly in the first year followed by biannual measurements in the second and annual measurements after the third year following surgery. Recurrence was defined as a postoperative PSA of 0.2 ng/ml. Time of recurrence was defined by the first PSA value above or equal to 0.2 ng/ml. Patients without evidence of tumor recurrence were censored at last follow-up. No patient of the cohort received neo-adjuvant or adjuvant therapy. For TMA construction, representative tissue cylinders with a diameter of 600 mm were punched from tumor areas of a paraffin-embedded donor tissue block and transferred to the corresponding coordinates on the recipient paraffin block in a half-automated process using precision instruments. Four-micrometer thick sections of each microarray block were transferred to adhesive slides for immunohistochemistry analyses. Use of tissues and clinical data was approved by the ethics committee of the Hamburg Chamber of Physicians and in accordance with local law (Hamburgisches Krankenhausgesetz, HmbKHG) and the Declaration of Helsinki. In agreement with HmbKG, 112, 1-3 and 112a, 1-5 specific informed consent was neither required nor obtained for the present study. All patient records/information were anonymized and de-identified prior to analyses.

Immunohistochemistry (IHC)
Freshly cut TMA sections were stained in one experiment on a single day. TMA sections were de-paraffinized followed by heatinduced antigen retrieval in an autoclave in acetate buffer pH 6.0 for 5 min. Primary polyclonal rabbit anti-SSTR2 antibody (HPA007264, Atlas Antibodies, Stockholm, Sweden) was used in a final dilution of 1:150. SSTR2 expression was visualized utilizing the Envision System (DAKO, Glostrup, Denmark).
Membranous and cytoplasmatic staining was evaluated separately for each spot. The staining intensity (negative = 0, weak = 1+ , intermediate = 2+, strong = 3+) and the fraction of positive tumor cells (in %) were recorded for each tissue spot. A final score was built from these two parameters as previously described [15,16]. In brief, negative scores had staining intensity of 0, weak scores had staining intensity of 1+ in #70% of tumor cells or staining intensity of 2+ in #30% of tumor cells; moderate scores had staining intensity of 1+ in .70% of tumor cells, staining intensity of 2+ in .30% and #70% of tumor cells or staining intensity of 3+ in # 30% of tumor cells and strong scores had staining intensity of 2+ in .70% of tumor cells or staining intensity of 3+ in .30% of tumor cells. Ki67 IHC data generated on the same TMA were available from a previous study [17].

In silico cDNA microarray analysis
A Gene Expression Omnibus (GEO) search (www.ncbi.nlm.nih. gov/gds) was conducted for human gene array datasets with information on SSTR2 expression using the string ''prostate cancer somatostatin receptor 2 homo sapiens''. Additional requirements were: expression data within the same dataset for comparison of (1) healthy prostate and prostate cancer, (2) primary prostate cancers and metastatic prostate cancers, or (3) nonrecurrent and recurrent prostate cancers. Furthermore, suitable datasets needed to include a minimum of n = 22 per group in order to accomplish a Type I error probability of 0.05 to reach statistical power of 0.9 and a difference in groups of one standard deviation in a non-paired sample setting.
Out of a total of 158 datasets identified with expression data on SSTR2 in human prostate cancers only 2 datasets fulfilled all of the criteria above, namely GDS2545 [18,19] and GDS4109 [20]. Deviations from total are due to missing data in the subcategories and rounding. doi:10.1371/journal.pone.0100469.t002 GDS2545 contains expression data from 65 primary prostate cancers, 63 normal tissues adjacent to prostate cancer and 25 prostate cancer metastases hybridized to the Affymetrix Human Genome U95 Version 2 Array platform (GPL8300). GDS4109 contains expression data from 39 recurrent and 49 non-recurrent primary prostate cancers hybridized to the Affymetrix Human Genome U133A Array platform (GPL96).
Normalized gene expression values for SSTR2 (GDS2545+ GDS4109) as well as SSTR3, 4 and 5 (GDS2545 only) were extracted and analyzed using the GEO Dataset Browser data analysis online tools. No information was available for SSTR1 from GDS2545.

Statistical analysis
Statistical analyses were accomplished using JMP 5.0.1 software (SAS Institute Inc., Cary, NC, USA) performing Pearson's chisquare test for contingency tables. Analysis of Variances (ANOVA) was used to test the association of Ki67 labeling index (LI) and SSTR2 staining, Kolmogorov-Smirnov test for determination of normal distribution and subsequent unpaired t-test or Mann-Whitney-U test calculations were performed for in silico cDNA expression analysis using PRISM 6 (Graphpad Software, Inc, La Jolla, CA, USA). Survival curves were calculated by Kaplan-Meier analysis and compared with log rank test. Multivariate analysis utilized the Cox regression model to identify indepen-dence of clinical parameters and SSTR2 immunohistochemistry to predict PSA-recurrence, cancer-specific survival and time to onset of metastatic disease after radical prostatectomy.

Baseline characteristics of prostate cancer patients
A TMA was constructed from cancer tissues after radical prostatectomies from 3,261 patients treated at the Department of Urology, University Medical Center Hamburg-Eppendorf between 1992 and 2005. Follow-up data for biochemical cancer relapse were available for 2,385 patients (73.1%) with a mean observation period of 34.7 months (range: 1 to 143 months, see Table 1 for baseline characteristics). TMA cores of 2,363 patients contained prostate cancer cells (72.5%). These cores were included in this study. The group of interpretable cases for statistical analysis was formed by 2,195 spots (67.3%) with immunohistochemical information regarding SSTR2 expression.
Loss of SSTR2 expression in prostate cancer is strongly linked with biologically aggressive cancers SSTR2 expression as determined by IHC was absent in 1,231 (56.1%) of interpretable prostate cancer samples. Weak SSTR2 staining was detected in 680 of these tumors (31.0%). Moderate and strong staining occurred in 187 (8.5%) and 97 prostate cancers (4.4%), respectively. Staining pattern was mostly both membranous and cytoplasmic throughout positive TMA spots. For comparison, normal prostate tissue always showed a positive receptor staining ( Figure 1A-D).
Contingency analysis of tumor phenotype and clinical features revealed significant, inverse associations of SSTR2 staining intensity with de-differentiation of tumors (as indicated by high Gleason grade, p,0.0001), advanced tumor stage of prostate cancers (as indicated by advanced pT category, p,0.0001, and cancer-positive surgical margins, p = 0.006) as well as high preoperative PSA levels (p = 0.0011). Statistical details on the correlation of clinico-pathological factors with SSTR2 IHC intensities and frequencies are given in Table 2.
The same strong inverse correlation with SSTR2 staining intensity was detected with tumor cell proliferation as determined by Ki67 LI. Ki67 LI increased with decreasing SSTR2 IHC intensity in prostate cancer cells from 2.860.4 (strong SSTR2 IHC) over 3.860.4 (intermediate SSTR2 IHC) and 5.060.2 (weak SSTR2 IHC) to 5.960.2 in SSTR2-negative prostate cancers (arithmetical mean 6 standard error of the mean, p,0.0001, Figure 2).
To test if the changes in SSTR2 expression detected by IHC could be confirmed on the transcriptional level, we performed in silico expression analysis utilizing two publicly available GEO datasets of human prostate cancer gene expression arrays (GDS2545 and GDS4109).
Using the GDS2545 dataset, we compared expression of SSTR subtypes 2, 3, 4 and 5 in normal prostate tissue with physically adjacent primary prostate cancer in an unpaired fashion.  Figure 3A).

Loss of SSTR2 in prostate cancers predicts metastatic and biochemical cancer relapse
In Kaplan-Meier analysis, all tested clinical and pathological features were strongly linked to PSA recurrence, prostate cancer specific survival and time to onset of metastatic disease. This applied for Gleason score, pT stage, pN Stage, pre-operative PSA  Table 3 and Figure 4A-C).
Except for prostate cancer-specific survival (p = 0.5942, not shown) SSTR2 staining was inversely correlated with biochemical recurrence and metastasis-free survival: Respectively, PSA free survival (p = 0.0009, Figure 5A) and metastasis-free survival (p = 0.0452, Figure 5B) gradually declined from strong SSTR2 staining over medium and weak to negative.
In multivariate analyses, all tested parameters (Gleason score, pT stage, pN Stage, surgical margin status and pre-operative PSA levels, all p,0.001) but not SSTR2 staining intensity (p = 0.6938) were identified as independent risk factors for biochemical relapse.

Discussion
Immunohistochemically detectable cytoplasmic and membranous SSTR2 protein was seen in 44% of our 2,195 interpretable prostate cancer samples. Previous studies had analyzed smaller patient cohorts (14-45 cases) and found highly variable results including lower [9,10] and higher [7,8] numbers of ''SSTR2positive'' prostate cancers as compared to our data. By RT-PCR, Halmos et al. only found three SSTR2-positive prostate cancers Table 3. Association of clinico-pathological features of prostate cancer samples included in this study with PSA recurrence, cancerspecific survival and time to onset of metastatic disease.  80% of prostate cancers from radical prostatectomies [8].

Log-Rank test (Kaplan-Meier analysis)
The comparison of normal and cancerous prostate epithelium in the present study by IHC demonstrated that SSTR2 is not overexpressed, but instead generally downregulated in prostate cancers. Our IHC data were confirmed by two independent cDNA microrarray datasets evaluated for SSTR2 gene expression. Furthermore, our findings are also supported by early data from studies using in vitro receptor autoradiography to compare SSTR2 expression between normal and neoplastic prostate epithelium [21]. Our data show, that SSTR2 downregulation is strongly linked to unfavorable tumor phenotype, early PSA relapse and onset of metastatic disease. This observation is also consistent with a previous study showing IHC positivity in 100% of 12 cases with Gleason grade 1 or 2 but in only 20% of 20 Gleason 4 or 5 cancers [10].
The strong association between high Ki67 LI and downregulated SSTR2 protein found in our study implies that SSTR2 downregulation exerts unfavorable biologic effects upon prostate epithelial cells through diminished cell proliferation control. Several studies have indeed suggested a role of SSTR2 in the regulation of tumor cell proliferation in various tumors, since tumors with reduced SSTR2 protein levels revealed increased cell proliferation [22][23][24][25][26][27][28][29]. For example, Qui et al. found significantly higher Ki67 LI in colorectal cancer cells with absent SSTR2 [25]. Over-expression of SSTR2 in MCF-7 breast cancer cells, which naturally express low levels of SSTR2, led to increased apoptosis and cell cycle arrest [29]. In C6 glioma cells, proliferation was inhibited by activation of SSTR2 as measured by [ 3 H]thymidine incorporation assays [28,30]. Additionally, infection of pancreatic and non-small cell lung cancer cells with SSTR2 expressing adenoviral vectors significantly decreased tumor growth and proliferation rate [27].
The mechanism, by which SSTR2 is able to confer its antiproliferative effects, has recently been investigated. Zou et al. found a strong inhibitory effect of SSTR2 on the cell cycle in the aforementioned pancreatic cancer cells [27]. In their animal model for pancreatic adenocarcinoma, SSTR2 overexpression led to strong up-regulation of cyclin-dependent kinase inhibitor p16, which then inhibited tumor cell cycle progression from G1 to S phase. Recent data suggest that methylation may be a relevant mechanism for controlling SSTR2 expression in cancer. Torrisani et al. reported the regulation of human SSTR2 expression in various (cancer) cell lines by epigenetic modifications [31]. Their data showed that DNA methylation and histone acetylation can regulate the activity of the SSTR2 promoter and demonstrated that promoter activity directly correlates with SSTR2 expression in breast cancer, pancreatic cancer, hepatic cancer, melanoma and retinoblastoma cell lines in an inverse fashion. Moreover, treatment of these cells with demethylating agents and acetylase inhibitors rescued SSTR2 mRNA expression [31].
As a surface membrane receptor, SSTR2 is suitable as a therapeutic target. Since somatostatin's short half-life limits its therapeutic use, synthetic analogs have been developed since the 1980s [32]. Octreotide is the best-characterized analog, and it binds to SSTR2 with higher affinity and much higher therapeutic potency than somatostatin [32]. Several other therapeutic ligands of SSTR2 are currently available or under clinical testing, including long-acting formulations (lancreotide, vapreotide, seglitide and pasireotide/SOM230) [33], and chimeric molecules coupled with cytotoxic agents (e.g. AN-238, a doxorubicin/ somatostatin conjugate, [34]).
Octreotide and its derivatives have been used routinely for the detection and therapy of neuroendocrine tumors for years [35]. In C6 glioma, activation of SSTR2 by various somatostatin analogs led to a strong inhibition of in vivo cancer cell proliferation, intratumoral neoangiogenesis and Ki-67 expression [30].
In concordance with our data, human pancreatic adenocarcinomas lose SSTR2 expression [44]. Re-introduction of SSTR2 in pancreas cancer by gene transfer robustly inhibited tumor cell proliferation and tumorigenicity [24,45]. In the SSTR2-negative pancreatic and non-small lung cell cancer models of Zou and colleagues, re-expression of SSTR2 in the cancer cells led to significantly impaired tumor growth which was aggravated in a dose-dependent manner by application of octreotide and its derivate vapreotide (RC-160, [27]).
Lastly, we show that loss of SSTR2 was linked to metastatic progression of prostate cancers. Patients with SSTR2-negative prostate cancers had impaired metastasis-free survival, and metastases of prostate cancers expressed even lower levels of SSTR2 mRNA than primary prostate cancers. Similar findings have been described for colorectal cancers [46] and seminonas Figure 5. Clinical impact of SSTR2 staining on event-free survival. PSA recurrence-free survival gradually declines from strong staining of cancer spots over moderate and weak to SSTR2-negative prostate cancers (A, p = 0.0009, Kaplan-Meier analysis with Log-Rank test). Prostatectomized patients with SSTR2-negative prostate cancers also have impaired metastasis-free survival (B, p = 0.0452, Kaplan-Meier analysis with Log-Rank test). doi:10.1371/journal.pone.0100469.g005 [47]. In addition, gene transfer of SSTR2 inhibited metastatic progression in two different pancreatic carcinoma models [48,49].
How SSTR2 might regulate tumor spreading is less clear, but published data suggests a mechanism promoting cell-cell adhesions. In pancreatic cancer the Src homology region 2 domaincontaining tyrosine phosphatase 1 (SHP-1) is able to dephosphorylate the epithelial cell adhesion molecule E-cadherin, thereby stabilizing inter-epithelial cell junctions in a SSTR2dependent fashion [48,50]. SHP-1 has also been shown to be downregulated in biologically aggressive prostate cancers [10]. This mechanism is further supported by work of Lahlou et al. showing that SSTR2 facilitated restoration of functional gap junctions in pancreatic adenocarcinoma cells through up-regulation of connexins Cx26 and Cx43 [51].
Most interestingly, in prostate cancer cells SSTR2 is also able to inhibit another feature of cancer metastasis, cell migration [52]. Mechanistically, this seems to be mediated by ligand-dependent activation of the Y-27632-sensitive Rho-GTPase pathway [52]. Data from normal, healthy, primary keratinocytes corroborate this mechanism and suggest underlying SSTR2-(amongst others) mediated inhibition of Rac1 activity [53].

Conclusions
We have been able to clarify that loss of SSTR2 is strongly linked to an aggressive tumor phenotype and predicts poor prognosis of prostate cancers. Indeed, most prostate cancers are SSTR2 negative and SSTR2 expression seems to be an important factor in the pathogenesis of prostate cancer. Although, the suitability of SSTR2 as a target for gene therapy needs to be evaluated, loss of SSTR2 is strongly linked to invasiveness, early PSA relapse, and metastatic spread in prostate cancer. Our data suggest that mechanistically these effects are mediated by increased cancer cell proliferation in cells with down-regulated SSTR2.