Read the Full Video Transcript

Andrea Miyahira: Hi everyone. I'm Andrea Miyahira at The Prostate Cancer Foundation. Today I'm joined by Dr. Rana McKay, an associate professor at UCSD. Dr. McKay will discuss her group's recent paper, "Association Between Adrenal-Permissive HSD3B1 Allele, and Prostate Cancer Outcomes in the Veterans Affairs Health System," that was published in JAMA Network Open. Dr. McKay, thank you for joining us today.

Rana McKay: Of course, it's my pleasure. It's wonderful to be able to be here with you today. By way of background, the 3-beta-HSD1 protein, which is encoded by the HSD3B1 gene, actually catalyzes a rate-limiting step necessary for non-testicular testosterone and dihydrotestosterone production. HSD3B1 has two common germline missense alleles, each resulting in very distinct functional activity of the 3-beta-HSD1 protein. The HSD3B1 adrenal-permissive allele encodes for a more rapidly degraded enzyme that restricts the conversion of DHEA to DHT. However, the HSD3B1 adrenal-permissive allele actually encodes a more stable enzyme that's resistant to ubiquitination and degradation, resulting in enzyme accumulation and then more robust conversion of DHEA to more potent androgens like DHT. And you can appreciate the significance of this with regards to sensitivity to hormonal therapy and also the role that it may play in prostate cancer pathogenesis, given that you now have more potent androgens in circulation with the adrenal-permissive allele.

The frequency of the adrenal-permissive allele varies widely by ancestry, and while the adrenal-permissive allele frequency is highest among white men, existing studies have really limited the number of African-American men that were represented. And there are some key cohorts here, and you can see the sample size across these key cohorts, ranging on the order of 100 patients or so, the largest series here, with the mCRPC series of just about 500 patients. But the percentage of individuals that were African-American that were enrolled was actually quite low across these series.

And there's growing evidence that demonstrates a link between HSD3B1 inheritance and prostate cancer outcomes. I think one of the most robust analyses prior to the work that we ventured on, was from the CHAARTED analysis that included 475 individuals looking at outcomes based off of HSD3B1. This work was conducted only in white individuals, and for those individuals that had low-volume disease, there seemed to be significant discrimination in outcomes based on the presence of the adrenal-permissive enzyme. And this analysis that was conducted was a little bit unique in that they actually lumped the heterozygote and homozygote adrenal-permissive patients into one group, as opposed to focusing on the homozygote adrenal-permissive as an independent entity. They lumped the AC and CC groups together. There was no prognostic significance in the high-volume groups, but largely in the low-volume groups.

So, really the objective of our study was to evaluate the difference in outcomes in men with prostate cancer based on HSD3B1 genotype. And we leveraged the MVP and the VA database to be able to do that. This is the largest study to date of men with prostate cancer having undergone systematic HSD3B1 genotyping, including over 5,000 men that all were profiled through the MVP. And what's great about the VA is actually the diversity of patients that are included in the database and in our analysis. We had a significant number of black individuals that were actually represented and we were able to discreetly look at the prevalence rates across different populations. And you can see that there definitely is an enrichment for the HSD3B1, adrenal-permissive AC and CC genotypes among white individuals. We look at that 77.2% for the AC and white individuals, 91.8% for CC. And the white individuals juxtapose 17.2 and 4.7. So dramatically different inheritance patterns based on race and ethnicity. Additional factors we captured around Gleason score, PSA diagnosis, there was really no differences in Gleason score based on HSD3B1 status. PSA, again, no difference among the groups that we observed. Interestingly, we did find slight variations in smoking history, whether these are statistically significant, but they really were not largely clinically significant. The differences were not large.

And additional baseline characteristics. The cohort largely included individuals who had been diagnosed with localized disease at diagnosis as opposed to metastatic disease. Less than 10% of the population was diagnosed with metastatic disease at diagnosis included in our cohort, and there was no difference in prevalence rates among the genotype groups and really no differences in prior treatments among the genotype groups.

And so our key finding is that the cumulative incidence of prostate cancer-specific mortality for patients with the CC adrenal permissive HSD3B1 genotype actually had the worst outcomes. So the cumulative incidence of prostate cancer-specific mortality at five years after prostate cancer diagnosis was higher among men who were homozygous for the adrenal permissive allele, compared to the heterozygotes and the adrenal restrictive homozygotes. And we were able to actually do a differential analysis discreetly looking at these three independent groups from 4% to 1.2% in the adrenal restrictive group.

Additionally, we looked at the cumulative incidence of prostate cancer-specific mortality among patients with localized disease at diagnosis and metastatic disease at diagnosis, as breaking up the cohort. And for patients who were non-metastatic at diagnosis, we saw that prostate cancer-specific mortality was higher in the CC genotypes compared to AA and AC at 1.8% compared to 0.6% for the homozygous AA. We did not really observe statistically significant differences among the genotype groups for patients who had metastatic disease at diagnosis. But keep in mind that this group really was less than 10% of the population that we looked at. So the sample size was really limited. We did observe numerical differences, 29.9% compared to 21% and 19%, but they were not statistically significant, probably related to the fact that these groups were on the order of a couple of hundred patients with only 33 patients having CC, as opposed to hundreds and thousands in the other analyses that were conducted.

The cumulative incidence of the development of metastases in patients with localized prostate cancer was similar across the group. So we looked at, for those people who had localized disease, did the rate of metastasis development differ? And that did not seem to be the case in our analysis.

And then we did an analysis looking at the cumulative incidence of prostate cancer-specific mortality among patients who developed metastases, and in patients who developed metastases at any time, the incidence of prostate cancer-specific mortality at five years was higher in the CC group, 36% compared to 8.5%.

So really in this analysis, we demonstrated that patients homozygous for the adrenal permissive CC genotype had the worst prostate cancer-specific mortality. There was a higher prevalence of the AC and CC genotypes among white men, and really our study adds to the growing evidence of the prognostic importance of HSD3B1 status, and its implications for people with prostate cancer.

Preclinical work from the Sharifi Lab has really been important in characterizing the function of the HSD3B1 protein in the cell and its implications in prostate cancer pathogenesis, and HSD3B1, for it to function, actually requires phosphorylation to take place for its cellular activity. And that phosphorylation can actually be inhibited by BMX directly. And preclinical studies have actually demonstrated that blockade of BMX can inhibit CRPC development in preclinical models. And I think this basically identifies what we believe to be a new hormonal therapeutic vulnerability for patients who have HSD3B1 mutated altered protein.

It's led to our design of a study called MAVERICK. Right now, this study is not enrolling patients. It was actually closed, given that the company who was supporting this study, unfortunately, went bankrupt. But this trial was looking at investigating abiraterone with abivertinib, which was a BMX inhibitor, in patients with MCRPC progressing on abi or naive to abi. So there's a tremendous interest in testing BMX inhibition in patients who have the adrenal permissive HSD3B1 genotype, not just in the mCRPC setting, but actually also in the hormone-sensitive setting as we think it has implications for sensitivity to ADT and ARSIs. So thank you so much.

Andrea Miyahira: Thank you so much, Dr. McKay, for sharing this with us. So, Black patients with prostate cancer have been demonstrated to have better outcomes with certain treatments such as abiraterone. Do you think there is any relationship between these observations and your findings that Black patients are less likely to have the permissive HSD3B1 CC allele?

Rana McKay: I definitely think that it has implications because the rates of the adrenal permissive allele in African-American men are dramatically lower in these individuals. And there's a series of data that demonstrate that African-American men tend to respond better to abiraterone with greater PSA responses, longer time on therapy. And I do think that it could be driven by underlying genetic germline alterations that could be driving some of these outcomes. I think there are a lot of reasons for patients to have variable responses to therapy, but I think it's definitely one of the drivers.

Andrea Miyahira: Thanks. So what treatment types might benefit patients with the permissive CC allele, for instance, docetaxel or non-antiandrogens, or are there any AR-targeted therapies that have a mechanism of action that this allele cannot impact?

Rana McKay: I think this brings up a good point. These patients tend to not have great responses to ADT, a shorter time to developing castration resistance, and I think thinking about alternative strategies that are non-AR targeted makes a lot of sense, such as taxane chemotherapy. We saw the data from CHAARTED, but we don't have patients who received an ARSI in the context of CHAARTED to be able to answer that question. But we've got the ARASENS study, we've got the PEACE-1 study, and that could be a really cool thing to investigate in the context of those studies. I think what's going to be really exciting is integrating strategies of BMX inhibition or other ways to target these vulnerabilities in the tumor by prolonging their time on an ARSI and ADT by integrating BMX inhibition.

Andrea Miyahira: Okay, thanks. What are the next steps that are needed to make a clinical impact with these findings? Do you envision this being used for prognostication, treatment selection, etc.?

Rana McKay: I think with regards to next steps, this allele, the adrenal permissive allele, is highly prognostic. And when we think about its prevalence in the population, so heterozygotes, at least 50% of the population, much higher prevalence in Caucasian individuals. But if you take the all-comer population of all people with prostate cancer, at least 50% are going to be AC. And then at least 10% are going to be CC. And that's even more prevalent than BRCA. And the effect that we're seeing is almost to the degree and even more profound than what's seen with BRCA with CC. So I do think that this warrants understanding about the prognostic significance. I do think we as a field in prostate cancer need to be moving towards more biomarker-enriched strategies to better optimize therapy for patients. So I would hope that we would be able to launch an HSD3B1 specific study with a therapeutic strategy. I feel very strongly about that, and I do think that we can do it. And I think this data really showcased the need for that. This is a greater than 5,000 patient study highlighting the prognostic significance. And I do want to highlight just how amazing the MVP project has been. This is just such a rich tool to be able to investigate and leverage this data and no other dataset could we have done this project.

Andrea Miyahira: Okay. Well, thanks again for coming on and sharing this with us today.

Rana McKay: Thank you for having me.