UDP-glucuronosyltransferases and biochemical recurrence in prostate cancer progression

It is estimated that there will be approximately 161,360 new cases of prostate cancer (PC) and approximately 26,730 deaths from PC in the US population in 2017 [1]. While overall the average annual percent change in PC incidence and death from 2009 to 2013 has decreased by 8.6%, the death rates in black men was 2.3 times those in white men. Critical to the treatment and prognosis of prostate cancer is the identification of biomarkers by tumor clinicopathologic characteristics to direct the development of therapies and treatment options. Biomarkers that are involved in adrenal and instraprostatic androgen biosynthesis and metabolism have emerged as important targets [2]. Common androgen deprivation therapy (ADT) agents used in individuals with metastatic prostate cancer point to the relevance of androgens/androgen receptor signaling in tumor progression in the prostate cancer cell microenvironment [3]. Various expression levels of steroideogenic enzymes, involved in the biosynthesis of testosterone precursors, have been differentially associated with prostate tumor progression to castration-resistant prostate cancer (CRPC) [2, 4‐6]. Moreover, recent evidence suggests that steroideogenic enzymes responsible for the catabolism of intraprostatic testosterone metabolites may be linked to biochemical recurrence (BCR) and tumor progression to CRPC, however, those associations are less understood [2, 7‐12].

Uridine 5′-diphosphate-glucuronosyltransferases (UGT) are part of a superfamily of enzymes which are part of the phase II drug detoxification pathways in the hepatic tissue in mammals. The UGTs catalyze the addition of the hydrophilic moiety, glucuronide to acceptor molecules in a process called glucuronidation [13, 14]. In humans there are two major classes, UGT1 and UGT2, which each contain multiple genes on chromosome 2 and 4, respectively. The UGT2B subfamily is comprised of UGT2B4, UGT2B7, UGT2B11, UGT2B15, UGT2B17, and UGT2B28 along with six pseudogenes that are interspersed among the genes in the region [15]. Enzymes UGT2B15 and UGT2B17 exhibit substrate specificity for androgens such as testosterone, dihydrotestosterone (DHT), DHT metabolites, androsterone and 5α-androstane-3α,17β-diol [16, 17]. Furthermore, those enzymes conjugate androgens present in the lumen and basal epithelial tissue of the prostate [18, 19]. The UGT2B28 enzymes have been shown to be expressed in human testis, prostate, and prostate cancer cell line, LNCaP, where they conjugate 5α-androstane-3α,17β-diol, 5β-androstane-3α,17β-diol and androsterone, and testosterone [20] (Fig. 1). The conjugation of androgens and their metabolites by the UGT2B enzymes in prostate tissue is an irreversible reaction, effectively regulating the levels of DHT available for androgen receptor signaling [19, 21]. Thus the expression levels of UGT2B17, B15, and B28 may be predictors of intraprostatic levels of androgens and prostate cancer phenotype.

A couple of studies have suggested that expression levels of the UGT2B17, B15, and 28 enzymes may be prognostic for clinicopathologic characteristics of prostate cancer. In one study the expression levels of UGT2B17 and B15 in prostate tumor samples from ADTh treated patients were compared to expression levels in patients that were untreated or diagnosed as benign prostatic hyperplasia (BPH) [22]. The results showed that UGT2B15 expression in patients that were treated up to 12 months with ADTh was significantly higher when compared to expression levels in patients that received no treatment for a comparable time or were diagnosed with BPH. Results from the same study showed that UGT2B17 expression was significantly elevated for only 5 months in ADTh treated patients. In another study examining the association of UGT2B28 expression, Belledant et al. [21] demonstrated that higher cytoplasmic expression of the protein was significantly associated with higher Gleason scores and positive nodes. In contrast, lower nuclear expression was associated with decreased PSA and positive margins. These studies suggest that UGT2B expression was altered by intraprostatic levels of androgens and tumor aggression.

To assess whether expression levels of UGT2B enzymes would predict worse prostate cancer outcomes, we analyzed the association between expression of UGT2B15, B17 and B28 in tumor microarrays at the time of radical prostatectomy and risk of BCR. We hypothesized that higher expression of the UGT2B enzymes is associated with increased prostate cancer progression risk.

The demographic characteristics of the study population and their clinical and pathologic phenotypes are summarized in Table 1. The median age of the cohort was 63 years (Q1-Q3: 58-67) and less than half (48%) of the patients were black. The median follow-up time was 118 months (Q1-Q3: 85-144 months). Expression of the UGT2B isoforms was assessed by immunohistochemical detection (Fig. 2). The median percentage of cells staining positive for UGT2B15, UGT2B17, and UGT2B28 was 39%, 74%, and 77% respectively. The median of H-score for UGT2B15 was 40, 76 for UGT2B17, and 86 for UGT2B28.

There was no association between UGT2B15 or UGT2B28 percent positive or H-score on univariable or multivariable analysis (all HR = 0.98-1.08, p > 0.5; Table 2) when biomarkers were modeled per 10-unit increase. Similarly, there was no association between UGT2B17 H-score on univariable or multivariable analysis (HR 1.05, p > 0.2; 1.03, p > 0.4). However, there was a trend between higher UGT2B17 percent positive and increased risk of BCR (HR 1.11, 95% CI 0.98-1.25, p = 0.09) on univariable analysis, although this did not reach statistical significance. This trend was weakened on multivariable analysis (HR 1.10, 95% CI 0.98-1.25, p = 0.11).

Notably in secondary analysis, patients with UGT2B17 total positive above the median had a significant increased risk of BCR on univariable analysis (HR = 1.57, 95% CI 1.02-2.43, p = 0.041; Table 3). However, this association was attenuated in the adjusted model (HR = 1.50, 95% CI 0.94-2.40, p = 0.088). There was no association between UGT2B17 H-score groups and BCR on univariable (p = 0.20) or multivariable (p = 0.32) analyses. Upon further inspection, pathological Gleason score and seminal vesicle invasion were the greatest mediators between UGT2B17 and BCR. Specifically, after adjusting for these two variables, the HR is 1.28 (95% CI 0.81-2.00; p = 0.29).

Finally, we found no association between H-scores of UGT2B15, UGT2B17, or UGT2B28 and pathological Gleason score (p = 0.34, 0.92, 0.80, respectively; Table 4). There were no interactions between the biomarkers and race (p = 0.65 for UGT2B15, p = 0.10 for UGT2B17, p = 0.27 for UGT2B28).

There is increasing evidence that standard approaches to ADT that involve targeting enzymes involved in androgen biosynthesis and orchiectomy are not sufficient to prevent the development of CRPC [2]. The goal of these therapies is to diminish the availability of DHT activated androgen receptor (AR) to promote cancer cell proliferation and growth through AR specific gene activation [6]. The UGT2B enzymes, UGT2B15, B17, and B28, which are regulated in part by the AR, are responsible for the removal of DHT metabolites in an irreversible glucuronidation reaction [19, 23]. In vitro experimental evidence has shown that the amounts of DHT are lower in culture media in control prostate cancer cell lines not treated with UGT2B17/B15 si-treated cell. Thus the enzymes may play a critical role in hormone levels in prostate cancer influencing tumor aggression and promotion. We sought to evaluate the expression of three UGT2B enzymes for their association with clinical and pathological characteristics of a multiethnic population from the Durham Veterans Affairs Medical Center, and evaluate whether their expression levels would predict prostate cancer outcomes after radical prostatectomy. Our results show that UGT2B17 overexpression was associated with a significant increased risk of BCR and that association was attenuated in adjusted models. That association was not modified by race. However, there were no significant associations between UGT2B17 expression and pathological Gleason score. The expression of UGT2B15 and B17 enzymes was not associated with any clinical and/or pathological characteristics.

To date no study has compared the expression of UGT2B17 and UGT2B17 to clinicopathologic characteristics of prostate tissue. Results from Grosse et al., [22] showed that expression of the UGT2B enzymes in samples from ADT treated patients were significantly higher than those of untreated patients. However UGT2B17 and B15 expression in the ADT untreated samples was not evaluated for association with clinical and pathological characteristics of prostate cancer such as BCR, PSA, pathological Gleason score, positive surgical margins, extracapsular extension and seminal vesicle invasion. The tissue used in this study was from patients who had no hormonal ablation or radiation therapy prior to prostatectomy surgery, therefore any comparison to the previous study may not be accurate. The results from our study showed no significant associations with BCR and Gleason score for UGT2B28 expression. In contrast, previous report showed that UGT2B28 overexpression was associated with lower PSA, high Gleason score, and BCR [21]. In that study, antibody staining was targeted in the basal cells of the prostate tumor and were cytoplasmic or nuclear. Further patient tissue samples with strong nuclear staining were associated with low PSA and positive margins while strong cytoplasmic staining was associated with positive nodal status and higher Gleason score. Interestingly in that same study, prostate secretory cells showed low to moderate staining. The results in our study emanated from cytoplasmic stained patient tissue samples that were not enriched for specifically for prostate basal cells. The population make-up in the studies may be different since the above study included patients from hospitals in Quebec, Canada and our study included Black and non-black patients in equal proportions. It is not clear if individuals of African descent have been characterized for UGT2B28 expression or polymorphisms. These differences in staining approaches and sample population may account for the differences in results.

Genetic studies have demonstrated a relationship between UGT2B genes and clinical and pathologic characteristics of prostate cancer. For example, a study has shown that deletions of UGT2B17 and B28 genes were significantly associated with decreased survival in Caucasians and Asians while deletion of only UGT2B17 was associated with Gleason score in Asians [24]. While significantly lower steroid glucuronides were found in that study, no other association was shown with PSA, Gleason score and tumor node metastasis. The expression levels of the UGT2B genes were not observed. Previous studies have shown that low expression levels of UGT2B17/15 in the LNCaP prostate cancer cell line resulted in lower intracellular glucuronide levels [19, 23]. In the current study, the results suggest that UGT2B17 overexpression was associated with BCR, which may be an indication overexpression may be a marker for tumor progression to CRPC independent of intracellular glucuronide levels. Another UGT gene has recently been implicated as a probable prognostic marker. Single nucleotide polymorphisms of the UGT1A locus (UGT1A9, UGT1A10) have been significantly associated with decreased survival following BCR [25]. In that same study, some SNPs from the same genes were protective and conferred decreased risk of BCR. Interestingly, the UGT1A risk alleles did not significantly change androgen glucuronide levels. UGT1A class expression was shown to be decreased in prostate cancer stem cells and in metastatic CRPC cells while in vitro experiments with LNCaP prostate cancer cell lines suggests this decreased expression is associated with cell survival [26]. This may be an indication that the expression of UGT2B genes in the prostate are critical to the hormonal microenvironment.

Our study has some limitations. The small sample size may have limited the power to detect significant associations between UGT2B17 expression and BCR on multivariable analysis. In addition we had a few number of events for CRPC and metastases, thus we could not evaluate the associations between UGT2B expression levels and those long-term prostate cancer outcomes. UGT2B17 expression has been found to be significantly expressed in prostate cancer metastases and CRPC tumors, while UGT2B15 is negatively regulated in those tissues [26]. Elevated circulating testosterone has been associated with UGT2B28 nuclear staining which may decrease expression of UGT2B15 and UGT2B17 since studies show that their expression is elevated in patients and prostate cancer cell cultures treated with antiandrogens [21, 22]. This suggests that other factors may be acting in tissues with deleted UGT2B17 and have higher biochemical recurrence.

In conclusion, the results from this study are suggestive that UGT2B17 overexpression may be associated with an increased risk of BCR. This study contributes to an emerging field of investigation that seeks to uncover the novel predictors of intraprostatic levels of androgens that persist in CRPC and the subsequent development of prostate tumor metastases [27]. Future analyses with increased sample sizes and data on UGT2B gene variants are needed, as those variants continue to show impact on the risk for prostate cancer [28]. The UGT2B enzymes, UGT2B15, B17, and B28 are responsible for much of the catabolism of testosterone and DHT metabolites in the prostate and must be further investigated to assess their importance as putative biomarkers and therapeutic targets.