The association between genetic variants in HSD3B1 and clinical management of PCa

The Abstract Androgen is an important factor in the occurrence and progression of prostate cancer. The principal clinical strategy is


INTRODUCTION
It is widely accepted that androgen and androgen receptor (AR) play an important role in the occurrence and progression of prostate cancer (PCa) [1] . Androgen deprivation therapy (ADT) via medical or surgical castration, which has been applied since 1941 [2] , is the principal therapeutic strategy to advanced prostate cancer. The general clinical efficiency of ADT depends on its blockage of gonadal androgen [3][4][5] . However, the development from castrate-sensitive prostate cancer (CSPC) to castrate-resistant prostate cancer (CRPC) is eventually inevitable [4] . One of the key mechanisms is the intertumoral synthesis of androgen, including testosterone and dihydrotestosterone (DHT), which originate from adrenal precursor steroids, such as dehydroepiandrosterone (DHEA) [6,7] . The androgen synthesis is considered to induce the reactivation of AR. When the disease finally progresses to the state of CRPC, it becomes highly lethal. To solve this clinical problem, next-generation hormone therapies were introduced nearly a decade ago [8] . Since then, the PCa prognosis is noted to be largely improved by several new drugs such as docetaxel (a microtubule inhibitor), abiraterone (a selective inhibitor of cytochrome P450 17A1 (CYP17A1), which is a key enzyme in androgen synthesis), and enzalutamide (a targeted androgen receptor inhibitor) [9][10][11] .
In terms of the intertumoral androgen synthesis, several key enzymes and genes in the synthetic process are identified as potential targets for diagnosis or treatment. One of them is 3β-hydroxysteroid dehydrogenase isoenzyme-1 (3βHSD1, encoded by HSD3B1), which catalyzes the rate-limiting step in the metabolic conversion from DHEA to testosterone and DHT in the adrenal gland [12] . A specific germline missenseencoding variant of HSD3B1(1245A>C, rs1047303) leads to a divergence of enzyme level and downstream androgen synthesis. HSD3B1(1245A) is known as an adrenal-restrictive allele as it codes an enzyme that is degraded more rapidly, while HSD3B1(1245C), an adrenal-permissive allele, codes a stable enzyme resistant to proteasomal degradation that promotes robust conversion from DHEA to DHT [13,14] . A series of studies indicated that genetic variants in HSD3B1 are associated with the progression of PCa and resistance of ADT. In this review, the homozygous variant genotype is referred to as HSD3B1 (CC), while the heterozygous genotype and homozygous wild type are referred to as HSD3B1 (AC) and HSD3B1 (AA), respectively.
In this review, we summarize the published studies regarding the association between HSD3B1 and the clinical management of PCa, including both CSPC and CRPC. We also discuss the contemporary significance of HSD3B1 and its potential value in the therapeutic decision in PCa.

EVIDENCE SYNTHESIS
We searched the published research articles using the keywords "prostate cancer" and "HSD3B1", in PubMed and Embase database. In total, 45 results were found. After review of the abstracts and full articles, 16 original research articles were finally included in this review.

HSD3B1(1245C) promotes resistance to ADT and development to CRPC
Recently, plenty of evidence shows that PCa patients carrying HSD3B1(1245C) variant were more likely to become resistant to ADT and progress to CRPC. This variant was also reported to be associated with worse survival outcome for patients with PCa treated with ADT, especially for those with low-volume diseases.
As shown in Table 1, since Ross et al. [15] first reported in 2008 that SNPs rs1870050 in CYP19A1 [hazard ratio (HR) = 0.60; P = 0.0007], rs1856888 in HSD3B1 (HR = 0.58; P = 0.0047), and rs7737181 in HSD17B4 (HR = 0.70; P = 0.0096) were related to shorter time to progression, the role of HSD3B1 in PCa has been paid close attention by researchers because 3βHSD1 encoded by HSD3B1 is necessary for the synthesis of non-testicular testosterone or DHT. Wu et al. [16] in 2015 first summarized that HSD3B1(1245C) PCa is more likely to progress to CRPC, according to a retrospective study involving 85 patients with AA genotype and 18 with AC genotype who were diagnosed as advanced PCa and underwent surgical castration. However, no significant disparity of overall survival time was shown related to HSD3B1.
Hearn et al. [17] reported that HSD3B1(1245C) is significantly associated with PCa resistance to ADT. As a multi-cohort study, they enrolled 443 patients treated with ADT after prostatectomy from three cohorts: the post-prostatectomy cohort from the Cleveland Clinic registry, the post-prostatectomy validation cohort from the Mayo Clinic SPORE registry, and the metastatic validation cohort from the Mayo Clinic metastatic prostate cancer registry. The frequency of variant was 26%-36%. In the primary cohort, compared with the AA genotype group, the CC genotype groups (HR = 2.4; P = 0.029) and the AC group (HR = 1.7; P = 0.041) were associated with worse progression-free survival (PFS). However, in the other two cohorts, CC showed the same significant effect on development to CRPC, while the association between the AC genotype and the progression of CRPC was not significant (HR = 1.1; P = 0.38). In addition, the variant allele was predictive to a worse overall survival.
Subsequently, a few studies came to similar conclusions. Agarwal et al. [18] retrospectively analyzed 102 patients with metastatic CSPC accepting ADT. The frequency of variant was 31%. Compared with the PFS in the AA genotype group, in the CC genotype group, PFS was shorter (11 months vs. 21 months; HR = 2.16; P = 0.046), while that in the AC genotype groups was similar (19 months vs. 21 months; HR = 1.04; P = 0.86). Shiota et al. [19] also performed an analysis on a primary ADT cohort with 104 Japanese patients of metastatic CSPC where the frequency of variant was 5%. The results shown that patients with the CC genotype and the AC genotype were more likely to be resistant to ADT (HR = 2.34; P = 0.03) but had no significant difference for mortality.
Hearn et al. [20] furthered their studies on the basis of their original report. Instead of post-prostatectomy, focusing on patients undergoing ADT post-radiotherapy, the study proved that HSD3B1(1245C) was also associated with rapid development of metastases. Furthermore, analyses were performed on the E3805 Chemohormonal Therapy Versus Androgen Ablation Randomized Trial for Extensive Disease in Prostate Cancer cohort which included patients with metastatic PCa undergoing ADT with or without docetaxel. They concluded that HSD3B1(1245C) was related to earlier development of CRPC (HR = 1.89; P = 0.02) and shorter overall survival (HR = 1.74; P = 0.045) in patients with low-volume disease but not in patients with high-volume disease [21] .
The current clinical studies on the role of HSD3B1 in CSPC patients treated with ADT are discussed above. With regard to the effect of HSD3B1 (CC) for accelerating the resistance to ADT and progression to CRPC, the results of current studies are in agreement. A meta-analysis by Han et al. [22] confirmed this conclusion and also concluded that HSD3B1 had no association with mortality. Except the outcome in the primary cohort of the Hearn's study in 2016 [17] , HSD3B1(1245C) was noted to have no impact on the mortality of CSPC. At a more fundamental level, Chen et al. [23] reported that HSD3B1 was marginally significantly associated with impaired survival outcome. Based on the study by Hearn et al. [21] , the effect of HSD3B1 was only found in patients with low-volume PCa. It was difficult to derive a precise conclusion from the analysis for the overall cohort. In addition, it is worth noting that the association between the AC genotype and Ross et al. [15] , 2008* ADT 529 62 For TTP: HR = 0.58; 95%CI: 0.41-0.81; P = 0.0047 The polymorphism in HSD3B1 was associated with time to progression during ADT for PCa Wu et al. [16] , survival outcome remains unclear. The effect of the AC genotype on promoting resistance to ADT was statistically significant in only one study [17] . In contrast, in the other two cohorts of Hearn's study in 2016 and the cohort in Agarwal's study, no difference in overall survival was observed between the AC and AA genotype groups.

Impact of HSD3B1(1245C) on the other medical management is waiting to define
Since the function of germline variants in HSD3B1 was confirmed to promote the development of resistance to ADT, several works focused on how it influences the outcome when coming to the state of CRPC. It has been reported that HSD3B1(1245C) might have a negative effect on the overall survival of CRPC. However, there are contradictions in the existing results about its association with the response to medical management, such as abiraterone and enzalutamide, in the state of CRPC.
Stangl-Kremser et al. [24] and Chen et al. [23] both summarized the association between HSD3B1(1245C) and survival outcome of patients with CRPC via genetic sequencing of prostate tissue. The former concluded that there was no relation between HSD3B1(1245C) and survival outcome of CRPC [24] . Similarly, the latter reported that HSD3B1(1245C) was related to a trend of worse prognosis for CRPC because of increased tumor expression of cell proliferation and cell cycle genes [23] , but without significant difference.
It is not clear whether HSD3B1(1245C) induces an impaired survival outcome in the state of CRPC. It would also be valuable to explore if germline HSD3B1 is predictive for the reaction to the related medical management for CRPC. Hearn et al. [21] reported no relation between HSD3B1 and response to docetaxel. In addition, as mentioned above, the key mechanism of CRPC development is intertumoral androgen synthesis and reactivation of AR. Some medical therapies aiming at blocking this process, known as AR pathway inhibitors (ARPIs), are widely accepted to be applied in the treatment of CRPC, such as AR antagonists including enzalutamide and CYP17A1 inhibitors including 17α-hydroxylase/17,20-lyase. In addition, CYP17A1 inhibitors include nonsteroidal and steroidal types. Steroidal CYP17A1 inhibitors, such as abiraterone, are converted by 3βHSD1, the downstream metabolites of which act as AR agonists, inducing an opposite effect [25,26] .
Biologically, HSD3B1(1245C) is supposed to invalidate both abiraterone and enzalutamide. Alyamani et al. [27] [ Table 2] conducted a study on the pharmacokinetics and metabolites of the steroidal CYP17A1 inhibitor abiraterone. They concluded that HSD3B1 might negate the efficiency of abiraterone based on the results that the downstream metabolite of abiraterone, 3-keto-5α-abiraterone, which is an AR agonist, significantly increased with more copies of HSD3B1(1245C), while the level of another metabolite, D4A, which is an AR antagonist, did not increase. Theoretically, in addition to abiraterone, enzalutamide may also be negated by HSD3B1(1245C). The permissive adrenal effect of HSD3B1(1245C) causes an increase of testosterone. It has been reported that increased the AR natural ligand might decrease the activity of enzalutamide [28] .
Several clinical studies were performed to identify the role of HSD3B1. Almassi et al. [29] focused on the relation between HSD3B1 and the curative effect of the nonsteroidal CYP17A1 inhibitor ketoconazole among patients with metastatic CRPC. They concluded that HSD3B1(1245C) marginally significantly prolonged time of progression. Compared with the AA genotype group, the HR of the AC genotype group for disease progression was 0.6 (P = 0.06) and the HR of the CC genotype group was 0.5 (P = 0.08).
Hahn et al. [30] focused on the predictive effect of HSD3B1(1245C) for patients with CRPC treated with abiraterone as first-line therapy, showing that there were no significant associations.

No. of Carriers Results Conclusions
Almassi et al. [29] , Both Khalaf et al. [31] and Lu et al. [32] investigated the association between HSD3B1 and abiraterone/enzalutamide. In the work of Khalaf et al. [31] , 547 patients from two cohorts were involved. For patients with metastatic CRPC treated with abiraterone/enzalutamide, the CC genotype had worse time to progression (HR = 1.31; P = 0.032) and a nonsignificant trend to shorter time to PSA progression (HR = 1.28; P = 0.064). In addition, the CC genotype was less likely to achieve a PSA response according to the PSA response rates (48% for CC, 62% for AC, and 65% for AA, P = 0.019). However, no association was shown between HSD3B1(1245C) and overall survival. In contrast, Lu et al. [32] proposed that HSD3B1 (CC) was related to worse survival outcome (HR = 1.78; P = 0.04) and had no association with response to management in patients with mCRPC treated with abiraterone/enzalutamide according to the analysis of a cohort of 266 patients.
Shiota et al. [19] reported a completely opposite conclusion based on a cohort of 99 Japanese patients with CRPC using abiraterone. Patients carrying variant genotypes (CC or AC) had significantly lower progression risk (HR = 0.32; P = 0.006) and lower all-cause mortality risk (HR = 0.40; P = 0.04) compared with others carrying the AA genotype. Furthermore, Shiota et al. [33] also studied the combinational use of HSD3B1 and 5α-reductase (encoding by SRD5A2), a key enzyme for conversion of testosterone into DHT. They showed that HSD3B1 (AA) with variant genotype of SRD5A2 led to the worst response to abiraterone in the state of CRPC.
In summary, germline variants of HSD3B1 influence the therapeutic effects of patients with CRPC, especially for abiraterone and enzalutamide. However, there exists a contradiction with regard to the exact function. Further studies are necessary to figure out the role of HSD3B1. Additionally, further studies are worth performing to demonstrate the effect of HSD3B1 variants on clinical management of PCa. In theory, HSD3B1(1245C), as an adrenal-permissive allele, is considered to induce a switch to a reduced tumor dependence on gonadal androgens and augmented dependence on extragonadal androgens. According to the study by Khalaf et al. [31] , the treatment effect and survival outcome were more favorable in the cohort where ARPIs were used as first-line therapy than other cohorts where previous firstline therapies were allowed. These two facts suggest that using ARPIs in the state of CSPC might benefit patients carrying HSD3B1 variants. In recent years, abiraterone with ADT was also shown to improve the prognosis of CSPC [34,35] . As a result, a predictive role of HSD3B1 variants for abiraterone used in the state of CSPC is a potential topic of further research.
Besides, some points provide potential directions for future study in order to identify the association between HSD3B1 and the clinical management of PCa. First, HSD3B1 variants present a significant interethnic disparity. For example, the frequency of HSD3B1(1245C) is 34% for European, 20% for American, 16% for South Asian, 9% for African, and 8% for East Asian patients [36] . Inter-ethnic disparity may be one of the explanations to the opposite outcomes about the effect of HSD3B1(1245C) on abiraterone. Second, loss of heterozygosity (LOH) is presented in the HSD3B1 gene. Hearn et al. [17] observed that 3 of 11 (27%) patients with HSD3B1 (AC) had developed CRPC with LOH of the wild-type allele, whereas none had lost the variant allele. This observation might explain the better treatment outcome when ARPIs were used as first-line therapy. Third, Hearn et al. [21] concluded that the influence of HSD3B1 depends on the tumor burden (the extent of metastatic disease), which might be worth in-depth study. At last, it remains to be studied whether HSD3B1 can be treated as a therapeutic target for PCa, about which no work has been reported up to now. However, the therapeutic potential of targeting HSD3B1 was reported in triple-negative breast cancer based on an in vitro study [37] .
CONCLUSION HSD3B1(1245C) has been found to be associated with ADT resistance and poorer prognosis. Its association with next-generation hormone therapy is still unclear (i.e., abiraterone). Very few commercial genetic profiling products have included this variant in routine testing. However, its clinical impact as well as its high cost-effectiveness (single variant testing is usually inexpensive) should not be underestimated. Based on the review of the current evidence, clinicians should consider a more aggressive treatment or follow-up strategy for patients with HSD3B1(1245C) variants who are undergoing ADT.