Background
An estimated 238,590 new cases of prostate cancer (PCa) are expected in 2013, accounting for 28 % of total malignancies in men [
1]. After surgery, some patients experience biochemically recurrent disease, resulting in unfavorable prognosis and an estimated cancer-specific mortality of 11–22 % [
2]. Treatment options are limited for patients that progress to metastatic disease, as androgen ablation inevitably leads to lethal castration-resistant prostate cancer. The identification of prognostic factors that predict metastatic recurrence after prostatectomy would allow the stratification of at-risk patients and assist in more effective individualized treatment. The reliability of the currently accepted biomarker used to detect PCa recurrence, prostate-specific antigen (PSA), has recently been questioned [
3], and highlights the urgent need for more effective PCa-specific biomarkers to guide patient treatment.
Matrix metalloproteinases (MMPs) are proteinases that play a role in tissue remodeling and disease progression through regulation of the extracellular microenvironment. Extracellular matrix metalloproteinase inducer CD147 (EMMPRIN, basigin) is a membrane-associated glycoprotein that stimulates the synthesis of several MMPs, including interstitial collagenase (MMP-1), gelatinase A (MMP-2), stromelysin-1 (MMP-3), and gelatinase B (MMP-9) [
4‐
6]. Elevated CD147 expression has been previously associated with cancer progression and invasion in malignancies such as breast and colorectal cancer [
7,
8].
Previous studies have observed similar trends in PCa, with increased CD147 expression in PCa samples compared to matched benign samples and an association with poor prognosis after prostatectomy [
9‐
13]. However, two recent studies have reported contrary results, with a decrease in CD147 observed with PCa progression [
13,
14], suggesting that CD147 may play a protective role in PCa. The prognostic role of CD147 in PCa is still debated, as results thus far are mixed [
11‐
14]. Furthermore, the sensitivity of previous studies has been limited due to semi-quantitative methods of analyzing immunohistochemical staining. The purpose of this study was to investigate the prognostic role of CD147 in PCa and the expression of CD147 in PCa progression using quantitative multispectral imaging.
Discussion
CD147 is a multifunctional glycoprotein that acts to induce the expression of MMPs, among other roles [
4‐
6,
25]. CD147 also forms heterodimers with proton-coupled monocarboxylate transporters (MCT) MCT1, MCT3, and MCT4 [
13,
25], helping to maintain lactate and pH homeostasis in epithelial cells, while N-glycosylation of CD147 leads to self-aggregation and MMP induction [
26]. Overexpression of CD147 of MMPs and CD147 has been noted in many malignancies, including breast and colorectal cancer [
7,
8], as the mechanistic role of highly glycosylated CD147 is suspected to assist in tumor invasion and metastatic spread [
26,
27]. However, the prognostic role of CD147 is still controversial, particularly in PCa.
Our results show that CD147 is decreased in malignant prostate samples compared to tumor-adjacent normal tissue, and further decreases in expression are associated with advanced pathologic stage and Gleason score, indicating that CD147 may be important in the progression to advanced stages of PCa. Using a xenograft-derived cell line model of prostate cancer progression, we confirmed that CD147 expression was decreased in aggressive PCa. Investigation of CD147-encoding BSG mRNA expression in publicly available microarray data shows that CD147 protein changes may be regulated transcriptionally. In this study, we also found that CD147 predicts biochemical recurrence after prostatectomy.
Our results are in disagreement with earlier studies on the expression and prognostic role of CD147 in PCa [
9‐
13]. However, one recent study of 11,152 patients found a decrease in CD147 expression between PCa and BPT and with increasing stage and Gleason score [
14], and these results are in agreement with one earlier study on CD147 expression in PCa [
13]. However, these studies concluded that CD147 does not play a significant prognostic role in determining post-surgical PSA recurrence [
14] or that PCa patients with higher CD147 expression performed significantly worse [
13]. This is the first study, to our knowledge, to demonstrate that low expression of CD147 is indicative of poor prognosis after prostatectomy independent of clinico-pathological features.
One limitation of previous studies is the semi-quantitative approach of evaluating immunohistochemical staining. While these methods are generally effective in quantitating “on/off” proteins, this approach is less effective when analyzing proteins with a heterogeneous staining pattern, particularly when proteins are localized to the membrane or cytoplasm, as small differences are not readily detected using manual methods of quantification. In this study, we show that CD147 is primarily localized to the cellular membrane. Quantitation of E-cadherin in the membrane portion of the epithelium resulted in a significant decrease in expression in all PCa samples compared to benign prostatic tissue, while HGPIN and BPH samples showed no significant differences in expression. Furthermore, membrane-associated E-cadherin expression predicted biochemical recurrence after prostatectomy in our patient cohort. This data is in agreement with previous studies on E-cadherin in PCa [
16,
28], and thus validates the epithelial cell membrane segmentation for investigation of CD147.
Antibodies to N-terminal synthetic peptides or recombinant fragments of CD147 have shown predominant localization to the basal and lateral plasma membrane, and studies using these antibodies have shown similar decreases in CD147 expression along PCa progression [
13,
14]. As has been argued previously [
14], the use of different antibodies may account for some of the contrasting results found in current literature on CD147. Some previous studies have shown cytoplasmic and nuclear staining of CD147 in PCa specimens [
11], rather than expected membrane-associated localization. Like previous studies with similar localization and expression results [
13,
14], CD147 protein expression in our cohort of patient samples was largely present in the plasma membrane. Additionally, our immunoblot results showed bands at molecular weights consistent with both glycosylated and nonglycosylated forms of CD147. When considering the known roles and expected localization of CD147 [
5,
27], our results suggest high antibody specificity for known forms of CD147.
Our results differ from previous studies on CD147 and other malignancies [
7,
8]. This may be attributable to the diverse roles and unusually high amount of post-translational modifications to CD147. The mechanistic role of CD147 of inducing MMP expression and promoting extracellular matrix degradation and reconstruction has implicated CD147 in tumor invasion. Caveolin-1 is a putative tumor suppressor in other malignancies [
26,
29,
30] that binds CD147 and suppresses N-glycosylation and CD147-induced fibroblast MMP activation. Caveolin-1 serves a unique role in PCa, as secretion and overexpression of caveolin-1 is associated with advanced disease [
31]. Though the interaction of caveolin-1 and CD147 was not investigated in this study, the overexpression of caveolin-1 in PCa progression may serve to bind CD147 and preventing N-glycosylation, self-aggregation, and MMP induction, thus accounting for the decrease in CD147 expression in advanced PCa that we observed.
Furthermore, the N-terminal domain of CD147 is essential for MMP induction [
27], as N-glycosylation of CD147 is associated with self-aggregation and MMP expression [
26]. In HT1080 fibrosarcoma and A431 epidermoid carcinoma cells, purification of a 22-kDa CD147 fragment from culture medium demonstrated cleavage and shedding of CD147 by membrane-type MMPs in the linking region between the two Ig-like domains [
27]. This soluble 22-kDa fragment was adequate for augmentation of MMP-2 expression in human fibroblasts, while the cleavage of membrane CD147 is expected to downregulate the membrane-specific MMP-activating function of CD147 [
27]. N-terminal cleavage and shedding of CD147 has not yet been studied in PCa, but may provide a mechanistic explanation for the decrease in CD147 with PCa progression that was observed in this and previous studies [
13,
14], as antibodies to the N-terminal Ig1 domain may not recognize cleaved forms of CD147. In this study, we also found that
BSG mRNA expression was not significantly associated with recurrence, indicating that post-translational modifications of CD147 may be important in the metastatic spread of PCa. Further studies are needed on post-translational modified forms of CD147 in PCa to reconcile the opposing results found throughout the literature.
Limitations of this study include both sample size and the use of tumor-adjacent normal as a baseline for comparison of expression. Compared to previous studies [
14] our sample size is relatively low, but the use of a quantitative multispectral platform increases the sensitivity of our assay and helps to make our sample size sufficient to draw conclusions. The use of tumor-adjacent normal tissues is common practice due to the difficult of obtaining truly normal tissues. However, future studies should incorporate autopsy normal prostate tissues to avoid the well-documented field effects of tumors.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made.
The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
TMB and WAR conceived the study and wrote the article. WH oversaw immunohistochemical staining and assisted in analysis. JAE developed methodology and performed Western blot analysis and public microarray data analysis. TMB carried out experimentation and performed statistical analysis. All authors have read and approved the final manuscript for publication.