The Role of Adhesion Molecules as Biomarkers for the Aggressive Prostate Cancer Phenotype

Background Currently available methods for diagnosis and staging of prostate cancer lack the sensitivity to distinguish between patients with indolent prostate cancer and those requiring radical treatment. Alterations in key adherens (AJ) and tight junction (TJ) components have been hailed as potential biomarkers for prostate cancer progression but the majority of research has been carried out on individual molecules. Objective To elucidate a panel of biomarkers that may help distinguish dormant prostate cancer from aggressive metastatic disease. Methods We analysed the expression of 7 well known AJ and TJ components in cell lines derived from normal prostate epithelial tissue (PNT2), non-invasive (CAHPV-10) and invasive prostate cancer (LNCaP, DU145, PC-3) using gene expression, western blotting and immunofluorescence techniques. Results Claudin 7, α –catenin and β-catenin protein expression were not significantly different between CAHPV-10 cells and PNT2 cells. However, in PC-3 cells, protein levels for claudin 7, α –catenin were significantly down regulated (−1.5 fold, p = <.001) or undetectable respectively. Immunofluoresence showed β-catenin localisation in PC-3 cells to be cytoplasmic as opposed to membraneous. Conclusion These results suggest aberrant Claudin 7, α – and β-catenin expression and/or localisation patterns may be putative markers for distinguishing localised prostate cancer from aggressive metastatic disease when used collectively.


Introduction
Prostate cancer (CaP) is the most common cancer in men in the UK. Every year nearly 35,000 cases of CaP are diagnosed and it accounts for approximately 12% of all male deaths from cancer in the UK [1]. Early diagnosis of organ-confined CaP is essential since radical prostatectomy or radiotherapy offers the only chance of complete cure and treatment of advanced disease is palliative and in-effective (in the long term). In addition, early organconfined CaP is not always life threatening and can follow and indolent course which does not require treatment. It is generally accepted that currently available methods used for diagnosis and staging of CaP (prostate specific antigen levels, gleason score and clinical and pathological grade) lack the sensitivity to distinguish between patients with indolent, organ-confined CaP, those requiring radical treatment and those at risk of relapse after radical treatment. Clearly there is a need to identify molecular markers of CaP progression, invasion and metastasis to predict diagnosis and guide therapy.
For a cancer to metastasise, cells must first break away from the primary tumour. In epithelial tissue, cells are connected to one another by membrane structures called tight junctions (TJ), adherens junctions (AJ) and desmosomes [2]. Together they maintain the architecture of the epithelium. AJ proteins comprise the cadherin and catenin families. E-cadherin is primarily present in epithelia and represents the prototypic member of the cadherin family. The cytoplasmic domain of E-cadherin binds to cytosolic proteins called catenins (a, b and p120) [3]. b-catenin binds directly to E-cadherin while a-catenin binds indirectly via its interaction with b-catenin [4]. Together with desmosomes they are primarily responsible for adhesion between adjacent cells [5] and forming stable cell-cell contacts. The TJ separate the apical and basolateral regions of the plasma membrane and regulate the passage of ions, water and macromolecules across the epithelium [6]. TJs are made up of membrane-bound proteins (claudins, occludin, tricellulin) and their adapter and scaffolding proteins (junctional adhesion molecule, ZO-1, ZO-2, ZO-3 cingulin, MUPP1) [7].
Until recently, TJs have only been perceived as cellular seals [8]. Now, however, losses of TJ proteins are being recognised for their association with a variety of cancers. Deregulation of TJ proteins is associated with the loss of epithelial cell polarity and dedifferentiation which is a known event in early stage carcinogenesis [7]. In poorly differentiated breast [9,10], thyroid [11], endometrial [12] and gastrointestinal [13] cancers the expression of tight junction proteins have been shown to be reduced, while in breast cancer patients' loss of functional TJ proteins has also been associated with a poorer prognosis [14,15]. Since TJ proteins are defined as the point where the membranes of two cells join together they perform a vital function by holding cells together and are a crucial barrier that cancer cells must overcome in order to spread [16].
While evidence continues to grow regarding the expression of TJ and AJ components in cancer the majority of studies are aimed at investigating individual molecules. Cancers are heterogeneous by nature and thus, it is impossible to diagnose cancer or predict disease progression using a single biomarker.
We utilised five commercially available prostate cell lines derived from normal prostate epithelium, non-invasive prostate cancer and metastatic cancer to analyse the expression of 7 well known AJ and TJ components, identified as aberrantly expressed in prostate cancer tissue samples [17,18,19,20,21], in the first step of potentially elucidating a panel of biomarkers that may distinguish indolent cancer from aggressive metastatic disease when used collectively.

Cell culture
Prostate epithelial cells (PNT2) and prostate cancer cells derived from cancer metastatic to the lymph nodes (LNCaP) and bone (PC-3) were obtained from the European Collection of Cell Cultures. Prostate cancer cells derived from non-invasive prostate cancer (CAHPV-10) and cancer metastatic to the brain (DU145) were purchased from the American Type Culture Collection. PNT2, LNCaP, DU145 and PC-3 were routinely cultured in RPMI 1640, 2 mM L-Glutamine Glutamine and 10% (v/v) foetal calf serum (Invitrogen, Paisley UK). CAHPV-10 cells were cultured in Keratinocyte serum-free medium with 5 ng/ml human recombinant epidermal growth factor and 50 mg/ml bovine pituitary extract (Invitrogen, Paisley UK).

Gene expression analysis
Total cellular RNA was extracted using the RNeasy extraction kit (Qiagen, Crawley, UK) and residual DNA was removed by treating with DNA-free TM (Ambion, Cambridgeshire, UK) according to the manufacturers' instructions. cDNA was synthesised from 1 mg RNA using random primers and the High Capacity cDNA synthesis reverse transcription kit (Applied Biosystems, Warrington, UK). Real-time PCR reactions were performed in the iCycler iQ Thermal Cycler (Bio-Rad Laboratories, Hemel Hempstead, UK) using the SYBR Green detection methodology. Primer sets (Table 1.) were designed to be exon-exon spanning in order to minimise the possibility that any contaminating genomic DNA would be amplified. The primer sets used were also tested to ensure they all demonstrated equally amplification efficiencies. All reactions were performed in triplicate with 2 ml cDNA, 12.5 ml iQ SYBR Green Supermix (Bio-Rad Laboratories, Hemel Hempstead, UK) and 0.2 mM forward and reverse primers, in a final reaction volume of 25 ml. b-actin and HPRT were used as reference genes. PCR amplification conditions were 95uC for 3 min, followed by 40 cycles of 94uC for 30 s, 60uC for 30 s and 72uC for 30 s. Fold expression was normalised against the normal prostate epithelial PNT2 cells.

Immunofluorescence
Following mRNA and protein analysis, Claudin 7 a-catenin and b-catenin appeared to have similar expression levels to normal prostate cells which became altered in cells derived from an aggressive metastatic tumour. Therefore, immunofluoresence was only performed on these molecules to ascertain whether alterations in protein expression were followed by aberrant localisation. Cells were seeded at 1610 5 cells/ml onto sterile microscope slides

Statistical analysis
Analysis carried out using the parametric ANOVA test. Dunnett's post hoc analysis was performed to compare cancer cells with the normal epithelial PNT2 control cells and Tukey post hoc analysis was performed for multiple group comparisons. A p-value,0.05 was considered significant.
NOTE: Only differences in gene expression and protein levels that satisfy both the p-value and a biological fold-change of ,1.5. would be deemed a significant change.

Gene Expression Analysis
Quantitative real-time PCR (Fig. 1) showed occludin mRNA levels were significantly different between all groups (p,0.001). A significant reduction in occludin mRNA levels was observed in CAHPV-10 (219.08 fold, p = 0.001), DU145 (23.2 fold, p = 0.003) and PC-3 (24.3 fold, p = 0.003) prostate cancer cells compared to normal prostate epithelial cells (PNT2). There was no significant difference in occludin mRNA expression between the cancer cells regardless of invasive capacity. However, the mRNA levels were significantly up regulated (2.2 fold, p = 0.001) in the LNCaP cancer cells compared to prostate epithelial cells (PNT2).
With regards to ZO-1, there was no significant difference in transcriptional levels between normal prostate epithelial cells and any of the prostate cancer cells. Claudin 7 gene expression levels were similar between PNT2 and CAHPV-10 (expression value 1.1). In contrast, Claudin 7 transcription was significantly up regulated in LNCaP, (4.63 fold, p = ,0.001), but down regulated in DU145, (25.3 fold, p = 0.004) and PC-3, (224.8 fold, p = 0.01) when compared to PNT2. There was also a significant down regulation in expression levels when Claudin 7 mRNA levels were compared between CAHPV-10 and (DU145, (p = 0.02) and PC-3 (p = ,0.001).
b-catenin gene expression was significantly down-regulated in CAHPV-10 cells, (23 fold, p = 0.015) compared to PNT2 while a significant up-regulation was observed in LNCaP,

Protein Expression Analysis
Western blot and densitometry were performed (Fig. 2) to ascertain whether the mRNA expression levels were comparable at the protein level (Table 2). Protein levels for occludin were down regulated in the prostate cancer cells CAHPV-10 (22.8 fold), DU145 (21.6 fold) and PC-3 (29.3 fold) but up-regulated in LNCaP (2.4 fold) compared to PNT2. However, due to the high variation between replicates (data not shown), there was no significant difference between any of the prostate cells.
Compared to PNT2, Claudin 1 levels were reduced in all prostate cancer cells. CAHPV-10 was down regulated 22.1 fold, DU145 22.3 fold and PC-3 21.8 fold. LNCaP was the only cell line to show a significant down-regulation compared to PNT2 (29.3 fold, p = 0.04).
Claudin 7 protein levels showed no difference between CAHPV-10 (1.2 fold) and PNT2 but levels were significantly down regulated in LNCaP (21.1 fold), DU145 (21.1 fold) and PC-3 (21.5 fold). All had a p = value of ,0.01. However, LNCaP and DU145 did not meet the biological fold-change and therefore, these changes were not taken into account. In addition, when Claudin 7 protein levels for CAHPV-10 were compared to PC-3, a significant down-regulation was also documented (p = ,0.01).
With regards to acatenin protein levels, there was no significant difference between CAHPV-10 and PNT2, while for LNCaP (4.4 fold, p = 0.03), DU145 (22 fold p = 0.04) a-catenin levels were significantly down regulated by various degrees when compared to PNT2. For, PC-3 a substantial down regulation in acatenin protein levels was observed, such that a protein band was undetectable and a densitometry value could not be obtained. In addition, CAHPV-10 protein levels were also significantly different to LNCaP (p = 0.03) and DU145 (p = 0.04).
For E-cadherin protein levels a significant down regulation was evident in CAHPV-10 (21.7 fold, p = ,0.01) and DU145 (21.5 fold, p = ,0.01). Again, a substantial down regulation in PC-3, such that a protein band was undetectable and a densitometry value could not be obtained. In LNCaP cells E-cadherin was up regulated at the protein level (2.7 fold p = 0.01).

Immunofluorescence
In order to determine whether the AJ and TJ proteins were present at their correct sub-cellular location (cell membrane), immunofluorescence was utilised. Immunofluorescence (Fig. 3) for Claudin 7 showed strong staining at the cell membrane for PNT2, CAHPV-10 and LNCaP. For both DU145 and PC-3, fluorescent staining was reduced. Staining was markedly reduced in both the DU145 and PC-3 cells with staining evident both at the cell membrane and in the cytoplasm. Staining for acatenin showed strong specific staining at the cell membrane for PNT2, CAHPV-10 and LNCaP cells, while for DU145 and PC-3 signal intensity was markedly reduced and non specific. With regards to b-catenin fluorescence staining, PNT2, CAHPV-10 and DU145 exhibited specific staining at the  cell membrane. b-catenin intensity in LNCaP and PC-3 cells was strong at the cell membrane and at sites of cell-cell contact, but in PC-3 staining was also evident in the cytoplasm.

Discussion
The first step in the metastatic cascade is the loss of cell-cell adhesion in order for cells to break away from the primary tumour. Thus, AJs have been a major focus of cancer studies and now TJ components are also being recognised for their involve-ment. Alterations in key AJ and TJ components such as a-catenin, E-cadherin, b-catenin and Claudin 1 have been hailed as potential biomarkers for prostate cancer progression [18,22,23,24] but the majority of research has been carried out on individual molecules. Cancer is heterogeneous by nature, thus there is a need for a panel of biomarkers, as opposed to single molecules, to help determine cancer progression and in the case of prostate cancer, distinguish dormant cancer from aggressive metastatic disease. As a result, the aim of this study was to evaluate the expression of several well known TJ and AJ components, collectively, as an important first step in identifying a panel of putative biomarkers that may help distinguish dormant prostate cancer from aggressive metastatic disease. Seven putative biomarkers were chosen due to their aberrant expression, which have been documented in the scientific literature and investigated in previous patient centred studies [17,18,20,22,23,24,25,26,27]. The 7 AJ and TJ proteins chosen were E-cadherin, aand bcatenin, ZO-1, occludin, claudin 1 and claudin 7 and their expression was investigated using an in vitro model of prostate cancer progression. Out of the 7, we highlighted 3 (Claudin 7, aand bcatenin) whose mRNA, protein levels and/ or localization are similar in cells derived from both normal and organ-confined prostate cancer but their levels and/or localization all alter in cells derived from aggressive metastatic tumours. Claudin 7 mRNA and protein levels in non-invasive prostate cancer cells (CAHPV-10) were similar to those found in prostate epithelial cells. In addition both mRNA and protein levels were  significantly down regulated in the most aggressive metastatic cell line PC-3 when compared to CAHPV-10. These results are supported by Sheehan et al, who showed a decrease in Claudin 7 expression correlated with high grade prostatic tumours [19]. Therefore, Claudin 7 protein levels may be a reflection of the aggressiveness of these prostate cancer cells and may only be altered in the most aggressive form of the disease. In addition, further analysis by immunofluorescence revealed that Claudin 7 localisation in the PC-3 cells was cytoplasmic as well as membrane bound. Aberrant localisation of Claudin 7 to the cytoplasm has been shown to occur in breast cancer cells [28] and oesophageal squamous cell carcinoma [2]. This re-distribution of Claudins have been attributed to several mechanism including protein down regulation resulting in cytoplasmic internalisation [2]. Based on these results it may be hypothesised that Claudin 7 could be a possible candidate for distinguishing indolent cancers from the most aggressive metastatic forms, which requires validation in patient samples.
In combination with Claudin 7, a -catenin levels in non-invasive prostate cancer cells were similar to levels in prostate epithelial cells, but like Claudin 7, was significantly decreased in the most aggressive metastatic cells. A significant relationship between reduced a-catenin expression and high gleason score has been reported in prostate tumours [24]. Similarly, a-catenin has been reported to be present in prostate tumours with a low gleason grade [17]. It has been suggested, therefore, that loss of a-catenin leads to the loss of e-cadherin function leading to a poorer prognosis [18], whereas, repletion of a-catenin in a-catenin null prostate cancer cells has been reported to increase adherens junction formation and reduced transcriptional activity of bcatenin, cyclin D1 levels and cell proliferation [26].
For b-catenin mRNA protein levels and immunofluorescent staining intensity of CAHPV-10 cells were similar to normal prostate epithelial cells and interestingly, immunofluorescence revealed b-catenin in PC-3 cells to be both membrane and cytoplasmic. b-catenin localises to the cytoplasm as a result of adherens junction complex breakdown and then translocates to the nucleus where it exerts it transcriptional effect [29]. This aberrant localisation to the cytoplasm has been shown to contribute to thyroid carciongenesis [30] and linked to a poor prognosis in breast cancer patients [31]. Our results suggest that it may be pertinent to look at cytoplasmic localisation of b-catenin, as opposed to quantitatively assessing it's mRNA or protein levels, as a putative biomarker of aggressive disease.
It should be noted that the mRNA and protein expression levels in the LNCaP cell line rarely followed the same trend as DU145 and PC-3. This difference in expression levels could be explained by the metastatic potential and differentiation status of the cell line. LNCaP has been documented as having limited metastatic potential and remains well differentiated when introduced into mouse models [32]. Thus it may be postulated that following metastasis to the lymph nodes, this cell line has reverted back to a more epithelial form as a result of colonisation [33] and growth. In addition, there were also some discrepancies between mRNA and protein levels for Claudin 7 and occludin in LNCaP, and b-catenin in CAHPV 210 and DU145 and a-catenin in CAHPV-10 and LNCaP. As mRNA is translated into protein it can be assumed that there should be a correlation between mRNA and protein levels. However, studies that have investigated a correlation between mRNA and protein expression levels have reported minimal or limited correlations [34]. The amount of protein detected may likely be affected by translation and post-transcriptional modifications providing a secondary degree of expression control that could account for these differences.

Conclusion
Due to the heterogeneous nature of cancer, analysis of single molecules provides limited information and it is impossible to diagnose cancer or predict disease progression using a single biomarker. The 7 biomarkers analysed here, have been reported to be aberrantly expressed in prostate cancer tissue samples but mainly on an individual basis. However, our in vitro investigation has shown, that out of these 7 only 3 (Claudin 7, a-catenin and bcatenin) may collectively be used to distinguish localised prostate cancer from cells representing aggressive metastatic disease. We believe that this in vitro identification of alterations in several key genes in combination highlights the importance of utilising several molecular markers and is an important first step in identifying a panel of putative biomarkers that may help distinguish dormant prostate cancer from aggressive metastatic disease. A large patient IHC study is now needed to validate these results.