Background
Water and solute movement across the epithelia lining the male reproductive tract are essential prerequisites for seminal fluid formation and homeostasis, and are of paramount significance for the modulation of the luminal environment in which sperm cells mature and reside.
The mechanism by which water crosses through epithelial borders had remained a matter of debate until the discovery and elucidation of the function of the aquaporin water channels (AQP) by the later Nobel laureate Peter Agre in the early 1990s. AQPs are a family of transmembrane pore-forming proteins that selectively allow water and other small, uncharged molecules such as urea, glycerol and pyrimidines to pass along hydrostatic and osmotic gradients. They play a fundamental role in numerous physiological processes, most notably in fluid absorption and secretion. To date, 13 different mammalian AQPs have been identified at the molecular level and localised in specific tissues [
1]. Analysis of several human diseases has confirmed that AQPs are functionally involved in various pathological conditions and thus may provide promising drug targets [
2]. Moreover, there is strong presumptive evidence that AQPs play a role in carcinogenesis, specifically in tumour angiogenesis and cell migration [
3].
To date, the presence and significance of AQPs in the human prostate remain largely uninvestigated, with reports of the individual expression of AQP 1, 3, 5 and 9 documented in previous studies [
4‐
8]. These findings suggest that fluid reabsorption and secretion in the prostate could be modulated by AQPs. Yet, despite investigations examining expression for individual AQPs, the human prostate has not been systematically studied in relation to all 13 members of the AQP family.
The principal aim of this study was to systematically characterize the expression pattern of all 13 AQP channels in cultured normal and malignant prostate epithelial cells, as well as freshly-isolated benign and malignant human prostate tissues. This approach allowed us to systematically study expression both at the mRNA and protein level of the whole AQP family in prostate tissue, and to correlate the pattern of expression with clinicopathological parameters. The potential biological and clinical significance of our findings are discussed.
Discussion
The expression and function of AQPs has been investigated in the majority of human tissues [
9]. Previous findings have suggested that fluid reabsorption and secretion in numerous organs are modulated by AQPs. However, human prostate tissue has not yet been systematically analysed in relation to AQP channel expression. The objective of the present study was to systematically investigate the expression and localisation of AQPs not only in human prostate (normal and malignant) cell lines in vitro, but also to investigate AQP expression in surgical specimens of BPH and PC of various malignancy grades.
Our study is the first to characterize human prostate tissue in relation to all 13 members of the AQP family. Transcripts for AQP 1, 3, 4, 7, 8, 10 and 11 were consistently detected in all four cell lines, while there was differential expression of AQP 5, 6 and 9. AQP 0, 2 and 12 were not detected in our study. Our mRNA transcript findings were corroborated at the protein level by immunofluorescence microscopy, where we detected AQP protein expression for those AQP family members where commercially available antibodies were able to be used. Our mRNA and protein findings in vitro are collectively in agreement with previous studies, which showed expression of AQP 1, 3, 5 in the human prostate at both the transcript and protein levels [
7,
15]. Having characterized AQP expression in normal and malignant human prostate cell lines, we then investigated AQP expression in human BPH and PC specimens of various tumour grades via qPCR and immunohistochemistry.
The expression of AQPs by human prostate tissue and recent reports on their possible role in carcinogenesis in epithelial tissues raise interesting questions about the potential functional (biological) significance of AQPs in the development and progression of PC. We contend that it is particularly important to investigate whether the pattern (and localisation) of AQP expression could be of prognostic and/or of therapeutic value. Our results provide strong presumptive evidence that there is a correlation between the loss of AQP 3 expression and increased PC tumour grade. To our knowledge, this is the first report of progressive loss of AQP 3 expression in high-grade PC. Previous investigations into the significance of AQPs in non-urological tumours have almost invariably shown overexpression of AQP 3 and it has been hypothesized that AQP 3 may be a promising drug target in the treatment of various epithelial tumours [
16‐
19]. The contrasting expression pattern of AQP 3 with intense labelling of AQP 3 in the case of BPH and well-differentiated PC, but non-homogeneous expression or loss of AQP 3 in the case of high-risk tumours, is noteworthy and may reflect tumour heterogeneity. Although this does not completely corroborate with cell line data, it is well known that in vitro cell lines do not necessarily reflect what happens in vivo. PC is well known for its tumour heterogeneity, which is reflected by diverse morphological manifestations and various molecular alterations associated with different tumour phenotypes [
20].
With regard to clinicopathological parameters, downregulation of AQP3 was associated with higher preoperative PSA values, higher risk according to the D’Amico classification and a higher Gleason-/ISUP-grade. Interestingly, in support of our prostate cancer findings here, we previously reported a similar observation in urothelial carcinoma, with studies suggesting that the loss of AQP 3 may play a role in bladder cancer progression. We showed a significant correlation between AQP 3 protein expression and tumour stage and grade, with AQP 3 expression being reduced or lost in urothelial carcinomas of higher grade and stage [
10,
21]. Furthermore, loss of AQP 3 expression was associated with worse progression-free and cancer-specific survival in patients with muscle-invasive bladder cancer [
22].
The pro-tumorigenic effect of AQP 3 loss has also been reported for non-urological tumour entities. For instance, knockdown of AQP 3 expression resulted in increased migration and proliferation in gastric adenocarcinoma cell lines [
16]. By contrast, overexpression of AQP 3 has been demonstrated for most other tumour entities, such as for squamous cell carcinomas [
17]. Although establishing a functional role for AQP 3 expression in BPH and PC was beyond the scope of this study, our findings suggest that progressive loss of AQP 3 expression may be associated with worse clinical PC outcomes.
When studying AQP 4 expression, we found expression in half of carcinomas but not in any BPH specimens in immunohistochemistry. Moreover, a significant positive correlation was observed between mRNA expression in qPCR and D’Amico risk classification, Gleason- and ISUP-grading. Differential expression of AQP 4 in benign and cancerous tissue and its potential biological and clinical roles need to be elucidated in further studies.
AQP 5 and 7 were detected both in BPH and in PC specimens of all grades. However, there was no correlation of AQP 5 and 7 mRNA expression with any clinicopathological parameter in our cohort. These findings are similar to the observations made in a recent study by Park et al. [
15]. Pust et al. previously reported highly variable AQP 5 expression in PC with both negative and intense expression of AQP 5 being linked to unfavourable outcomes [
23]. In contrast to our findings, Li et al. demonstrated that AQP 5 expression was upregulated and associated with advanced stage, circulating tumour cells and inferior survival rates in PC [
24]. These discrepancies and a wealth of somewhat contradictory results are interesting, and indicate that further studies are required to scrutinize the significance of AQP 5 expression in PC.
AQP 9 has previously been shown to be ubiquitously expressed along the male reproductive tract [
25]. In concordance with this, we found expression of AQP 9 in normal prostatic cells in vitro and in BPH in vivo, whereas loss of expression was consistently demonstrated in cancer cell lines and in PC specimens. The demonstration of an inverse correlation of AQP 9 mRNA levels in qPCR and PSA supports this. Our study is the first to report the consistent loss of AQP 9 expression in PC, the biological significance of which still remains to be established. In hepatocellular cancer, for instance, decreased expression of AQP 9 resulted in increased resistance to apoptosis [
26]. In rat prostates, AQP 9 expression has been shown to be regulated by androgens [
8]. Similarly, Jiang et al. and Tian et al. suggested that expression of AQP 1, 3–8, 10–12 in PC is modified by androgens, at least in rats [
27,
28]. However, there are little published data on its role in the human prostate. Hence, the potential role of AQP 9 in the carcinogenesis of PC needs to be addressed in future studies.
Despite the clear differential expression of AQP demonstrated in our study (particularly AQP 3), we do accept that our observational investigation was conducted on a relatively limited number of BPH and PC specimens. As such, despite their significance, our findings would certainly be strengthened by a bigger cohort of clinical specimens. One definite weakness of our study, however, is the lack of healthy prostate tissue serving as a control. Hence, we can only speculate on the relationship between AQP expression and function in normal human prostate tissue. Nevertheless, we did observe clear differential expression of AQP and a correlation with increasing malignancy in PC. While such an approach goes beyond the scope of the present study, future studies must certainly seek to prove a functional role for AQP 3 in PC progression. This would involve mechanistic studies and silencing with the RNAi-mediated knockdown of AQP 3 to assess biological endpoints relevant to tumour progression (proliferation, migration/invasion, and/or resistance to apoptotic stimuli) in the PC cell lines studied here. In addition, pre-clinical PDX models from patients with castration resistant PC might also be helpful. This would provide mechanistic insight into how changes in AQP expression may regulate tumour biology. We are currently addressing such questions, and are carrying out further studies that include sufficient patient numbers and long-term survival data to elucidate the potential clinical significance of AQP expression and its role in PC.