Clinical MedicineOncology Open Access | 10.1172/jci.insight.183158

Hannah E. Bergom,1,2 Ella Boytim,1,2 Sean McSweeney,3 Negar Sadeghipour,4 Andrew Elliott,4 Rachel Passow,1,2 Eamon Toye,1,2 Xiuxiu Li,5 Pornlada Likasitwatanakul,2,6 Daniel M. Geynisman,7 Scott M. Dehm,1,2 Susan Halabi,8 Nima Sharifi,5 Emmanuel S. Antonarakis,1,2 Charles J. Ryan,1,2 and Justin Hwang1,2

1Masonic Cancer Center, and

2Department of Medicine, University of Minnesota-Twin Cities, Minneapolis, Minnesota, USA.

3Glickman Urological and Kidney Institute, Cleveland Clinic, Cleveland, Ohio, USA.

4Department of Clinical and Translational Research, Caris Life Sciences, Phoenix, Arizona, USA.

5Desai Sethi Urology Institute, Sylvester Comprehensive Cancer Center, University of Miami, Florida, USA.

6Department of Medicine, Siriraj Hospital, Mahidol University, Bangkok, Thailand.

7Fox Chase Cancer Center, Philadelphia, Pennsylvania, USA.

8Department of Biostatistics and Bioinformatics, Duke Cancer Institute, Durham, North Carolina, USA.

Address correspondence to: Justin Hwang, 420 Delaware Street SE, MMC 480, Minneapolis, Minnesota 55455, USA. Email: jhwang@umn.edu.

Authorship note: HEB, EB, and SM are co–first authors. CJR and JH are co–senior authors.

Find articles by Bergom, H. in: JCI | PubMed | Google Scholar

1Masonic Cancer Center, and

2Department of Medicine, University of Minnesota-Twin Cities, Minneapolis, Minnesota, USA.

3Glickman Urological and Kidney Institute, Cleveland Clinic, Cleveland, Ohio, USA.

4Department of Clinical and Translational Research, Caris Life Sciences, Phoenix, Arizona, USA.

5Desai Sethi Urology Institute, Sylvester Comprehensive Cancer Center, University of Miami, Florida, USA.

6Department of Medicine, Siriraj Hospital, Mahidol University, Bangkok, Thailand.

7Fox Chase Cancer Center, Philadelphia, Pennsylvania, USA.

8Department of Biostatistics and Bioinformatics, Duke Cancer Institute, Durham, North Carolina, USA.

Address correspondence to: Justin Hwang, 420 Delaware Street SE, MMC 480, Minneapolis, Minnesota 55455, USA. Email: jhwang@umn.edu.

Authorship note: HEB, EB, and SM are co–first authors. CJR and JH are co–senior authors.

Find articles by Boytim, E. in: JCI | PubMed | Google Scholar

1Masonic Cancer Center, and

2Department of Medicine, University of Minnesota-Twin Cities, Minneapolis, Minnesota, USA.

3Glickman Urological and Kidney Institute, Cleveland Clinic, Cleveland, Ohio, USA.

4Department of Clinical and Translational Research, Caris Life Sciences, Phoenix, Arizona, USA.

5Desai Sethi Urology Institute, Sylvester Comprehensive Cancer Center, University of Miami, Florida, USA.

6Department of Medicine, Siriraj Hospital, Mahidol University, Bangkok, Thailand.

7Fox Chase Cancer Center, Philadelphia, Pennsylvania, USA.

8Department of Biostatistics and Bioinformatics, Duke Cancer Institute, Durham, North Carolina, USA.

Address correspondence to: Justin Hwang, 420 Delaware Street SE, MMC 480, Minneapolis, Minnesota 55455, USA. Email: jhwang@umn.edu.

Authorship note: HEB, EB, and SM are co–first authors. CJR and JH are co–senior authors.

Find articles by McSweeney, S. in: JCI | PubMed | Google Scholar

1Masonic Cancer Center, and

2Department of Medicine, University of Minnesota-Twin Cities, Minneapolis, Minnesota, USA.

3Glickman Urological and Kidney Institute, Cleveland Clinic, Cleveland, Ohio, USA.

4Department of Clinical and Translational Research, Caris Life Sciences, Phoenix, Arizona, USA.

5Desai Sethi Urology Institute, Sylvester Comprehensive Cancer Center, University of Miami, Florida, USA.

6Department of Medicine, Siriraj Hospital, Mahidol University, Bangkok, Thailand.

7Fox Chase Cancer Center, Philadelphia, Pennsylvania, USA.

8Department of Biostatistics and Bioinformatics, Duke Cancer Institute, Durham, North Carolina, USA.

Address correspondence to: Justin Hwang, 420 Delaware Street SE, MMC 480, Minneapolis, Minnesota 55455, USA. Email: jhwang@umn.edu.

Authorship note: HEB, EB, and SM are co–first authors. CJR and JH are co–senior authors.

Find articles by Sadeghipour, N. in: JCI | PubMed | Google Scholar

1Masonic Cancer Center, and

2Department of Medicine, University of Minnesota-Twin Cities, Minneapolis, Minnesota, USA.

3Glickman Urological and Kidney Institute, Cleveland Clinic, Cleveland, Ohio, USA.

4Department of Clinical and Translational Research, Caris Life Sciences, Phoenix, Arizona, USA.

5Desai Sethi Urology Institute, Sylvester Comprehensive Cancer Center, University of Miami, Florida, USA.

6Department of Medicine, Siriraj Hospital, Mahidol University, Bangkok, Thailand.

7Fox Chase Cancer Center, Philadelphia, Pennsylvania, USA.

8Department of Biostatistics and Bioinformatics, Duke Cancer Institute, Durham, North Carolina, USA.

Address correspondence to: Justin Hwang, 420 Delaware Street SE, MMC 480, Minneapolis, Minnesota 55455, USA. Email: jhwang@umn.edu.

Authorship note: HEB, EB, and SM are co–first authors. CJR and JH are co–senior authors.

Find articles by Elliott, A. in: JCI | PubMed | Google Scholar

1Masonic Cancer Center, and

2Department of Medicine, University of Minnesota-Twin Cities, Minneapolis, Minnesota, USA.

3Glickman Urological and Kidney Institute, Cleveland Clinic, Cleveland, Ohio, USA.

4Department of Clinical and Translational Research, Caris Life Sciences, Phoenix, Arizona, USA.

5Desai Sethi Urology Institute, Sylvester Comprehensive Cancer Center, University of Miami, Florida, USA.

6Department of Medicine, Siriraj Hospital, Mahidol University, Bangkok, Thailand.

7Fox Chase Cancer Center, Philadelphia, Pennsylvania, USA.

8Department of Biostatistics and Bioinformatics, Duke Cancer Institute, Durham, North Carolina, USA.

Address correspondence to: Justin Hwang, 420 Delaware Street SE, MMC 480, Minneapolis, Minnesota 55455, USA. Email: jhwang@umn.edu.

Authorship note: HEB, EB, and SM are co–first authors. CJR and JH are co–senior authors.

Find articles by Passow, R. in: JCI | PubMed | Google Scholar

1Masonic Cancer Center, and

2Department of Medicine, University of Minnesota-Twin Cities, Minneapolis, Minnesota, USA.

3Glickman Urological and Kidney Institute, Cleveland Clinic, Cleveland, Ohio, USA.

4Department of Clinical and Translational Research, Caris Life Sciences, Phoenix, Arizona, USA.

5Desai Sethi Urology Institute, Sylvester Comprehensive Cancer Center, University of Miami, Florida, USA.

6Department of Medicine, Siriraj Hospital, Mahidol University, Bangkok, Thailand.

7Fox Chase Cancer Center, Philadelphia, Pennsylvania, USA.

8Department of Biostatistics and Bioinformatics, Duke Cancer Institute, Durham, North Carolina, USA.

Address correspondence to: Justin Hwang, 420 Delaware Street SE, MMC 480, Minneapolis, Minnesota 55455, USA. Email: jhwang@umn.edu.

Authorship note: HEB, EB, and SM are co–first authors. CJR and JH are co–senior authors.

Find articles by Toye, E. in: JCI | PubMed | Google Scholar

1Masonic Cancer Center, and

2Department of Medicine, University of Minnesota-Twin Cities, Minneapolis, Minnesota, USA.

3Glickman Urological and Kidney Institute, Cleveland Clinic, Cleveland, Ohio, USA.

4Department of Clinical and Translational Research, Caris Life Sciences, Phoenix, Arizona, USA.

5Desai Sethi Urology Institute, Sylvester Comprehensive Cancer Center, University of Miami, Florida, USA.

6Department of Medicine, Siriraj Hospital, Mahidol University, Bangkok, Thailand.

7Fox Chase Cancer Center, Philadelphia, Pennsylvania, USA.

8Department of Biostatistics and Bioinformatics, Duke Cancer Institute, Durham, North Carolina, USA.

Address correspondence to: Justin Hwang, 420 Delaware Street SE, MMC 480, Minneapolis, Minnesota 55455, USA. Email: jhwang@umn.edu.

Authorship note: HEB, EB, and SM are co–first authors. CJR and JH are co–senior authors.

Find articles by Li, X. in: JCI | PubMed | Google Scholar

1Masonic Cancer Center, and

2Department of Medicine, University of Minnesota-Twin Cities, Minneapolis, Minnesota, USA.

3Glickman Urological and Kidney Institute, Cleveland Clinic, Cleveland, Ohio, USA.

4Department of Clinical and Translational Research, Caris Life Sciences, Phoenix, Arizona, USA.

5Desai Sethi Urology Institute, Sylvester Comprehensive Cancer Center, University of Miami, Florida, USA.

6Department of Medicine, Siriraj Hospital, Mahidol University, Bangkok, Thailand.

7Fox Chase Cancer Center, Philadelphia, Pennsylvania, USA.

8Department of Biostatistics and Bioinformatics, Duke Cancer Institute, Durham, North Carolina, USA.

Address correspondence to: Justin Hwang, 420 Delaware Street SE, MMC 480, Minneapolis, Minnesota 55455, USA. Email: jhwang@umn.edu.

Authorship note: HEB, EB, and SM are co–first authors. CJR and JH are co–senior authors.

Find articles by Likasitwatanakul, P. in: JCI | PubMed | Google Scholar

1Masonic Cancer Center, and

2Department of Medicine, University of Minnesota-Twin Cities, Minneapolis, Minnesota, USA.

3Glickman Urological and Kidney Institute, Cleveland Clinic, Cleveland, Ohio, USA.

4Department of Clinical and Translational Research, Caris Life Sciences, Phoenix, Arizona, USA.

5Desai Sethi Urology Institute, Sylvester Comprehensive Cancer Center, University of Miami, Florida, USA.

6Department of Medicine, Siriraj Hospital, Mahidol University, Bangkok, Thailand.

7Fox Chase Cancer Center, Philadelphia, Pennsylvania, USA.

8Department of Biostatistics and Bioinformatics, Duke Cancer Institute, Durham, North Carolina, USA.

Address correspondence to: Justin Hwang, 420 Delaware Street SE, MMC 480, Minneapolis, Minnesota 55455, USA. Email: jhwang@umn.edu.

Authorship note: HEB, EB, and SM are co–first authors. CJR and JH are co–senior authors.

Find articles by Geynisman, D. in: JCI | PubMed | Google Scholar

1Masonic Cancer Center, and

2Department of Medicine, University of Minnesota-Twin Cities, Minneapolis, Minnesota, USA.

3Glickman Urological and Kidney Institute, Cleveland Clinic, Cleveland, Ohio, USA.

4Department of Clinical and Translational Research, Caris Life Sciences, Phoenix, Arizona, USA.

5Desai Sethi Urology Institute, Sylvester Comprehensive Cancer Center, University of Miami, Florida, USA.

6Department of Medicine, Siriraj Hospital, Mahidol University, Bangkok, Thailand.

7Fox Chase Cancer Center, Philadelphia, Pennsylvania, USA.

8Department of Biostatistics and Bioinformatics, Duke Cancer Institute, Durham, North Carolina, USA.

Address correspondence to: Justin Hwang, 420 Delaware Street SE, MMC 480, Minneapolis, Minnesota 55455, USA. Email: jhwang@umn.edu.

Authorship note: HEB, EB, and SM are co–first authors. CJR and JH are co–senior authors.

Find articles by Dehm, S. in: JCI | PubMed | Google Scholar

1Masonic Cancer Center, and

2Department of Medicine, University of Minnesota-Twin Cities, Minneapolis, Minnesota, USA.

3Glickman Urological and Kidney Institute, Cleveland Clinic, Cleveland, Ohio, USA.

4Department of Clinical and Translational Research, Caris Life Sciences, Phoenix, Arizona, USA.

5Desai Sethi Urology Institute, Sylvester Comprehensive Cancer Center, University of Miami, Florida, USA.

6Department of Medicine, Siriraj Hospital, Mahidol University, Bangkok, Thailand.

7Fox Chase Cancer Center, Philadelphia, Pennsylvania, USA.

8Department of Biostatistics and Bioinformatics, Duke Cancer Institute, Durham, North Carolina, USA.

Address correspondence to: Justin Hwang, 420 Delaware Street SE, MMC 480, Minneapolis, Minnesota 55455, USA. Email: jhwang@umn.edu.

Authorship note: HEB, EB, and SM are co–first authors. CJR and JH are co–senior authors.

Find articles by Halabi, S. in: JCI | PubMed | Google Scholar |

1Masonic Cancer Center, and

2Department of Medicine, University of Minnesota-Twin Cities, Minneapolis, Minnesota, USA.

3Glickman Urological and Kidney Institute, Cleveland Clinic, Cleveland, Ohio, USA.

4Department of Clinical and Translational Research, Caris Life Sciences, Phoenix, Arizona, USA.

5Desai Sethi Urology Institute, Sylvester Comprehensive Cancer Center, University of Miami, Florida, USA.

6Department of Medicine, Siriraj Hospital, Mahidol University, Bangkok, Thailand.

7Fox Chase Cancer Center, Philadelphia, Pennsylvania, USA.

8Department of Biostatistics and Bioinformatics, Duke Cancer Institute, Durham, North Carolina, USA.

Address correspondence to: Justin Hwang, 420 Delaware Street SE, MMC 480, Minneapolis, Minnesota 55455, USA. Email: jhwang@umn.edu.

Authorship note: HEB, EB, and SM are co–first authors. CJR and JH are co–senior authors.

Find articles by Sharifi, N. in: JCI | PubMed | Google Scholar |

1Masonic Cancer Center, and

2Department of Medicine, University of Minnesota-Twin Cities, Minneapolis, Minnesota, USA.

3Glickman Urological and Kidney Institute, Cleveland Clinic, Cleveland, Ohio, USA.

4Department of Clinical and Translational Research, Caris Life Sciences, Phoenix, Arizona, USA.

5Desai Sethi Urology Institute, Sylvester Comprehensive Cancer Center, University of Miami, Florida, USA.

6Department of Medicine, Siriraj Hospital, Mahidol University, Bangkok, Thailand.

7Fox Chase Cancer Center, Philadelphia, Pennsylvania, USA.

8Department of Biostatistics and Bioinformatics, Duke Cancer Institute, Durham, North Carolina, USA.

Address correspondence to: Justin Hwang, 420 Delaware Street SE, MMC 480, Minneapolis, Minnesota 55455, USA. Email: jhwang@umn.edu.

Authorship note: HEB, EB, and SM are co–first authors. CJR and JH are co–senior authors.

Find articles by Antonarakis, E. in: JCI | PubMed | Google Scholar |

1Masonic Cancer Center, and

2Department of Medicine, University of Minnesota-Twin Cities, Minneapolis, Minnesota, USA.

3Glickman Urological and Kidney Institute, Cleveland Clinic, Cleveland, Ohio, USA.

4Department of Clinical and Translational Research, Caris Life Sciences, Phoenix, Arizona, USA.

5Desai Sethi Urology Institute, Sylvester Comprehensive Cancer Center, University of Miami, Florida, USA.

6Department of Medicine, Siriraj Hospital, Mahidol University, Bangkok, Thailand.

7Fox Chase Cancer Center, Philadelphia, Pennsylvania, USA.

8Department of Biostatistics and Bioinformatics, Duke Cancer Institute, Durham, North Carolina, USA.

Address correspondence to: Justin Hwang, 420 Delaware Street SE, MMC 480, Minneapolis, Minnesota 55455, USA. Email: jhwang@umn.edu.

Authorship note: HEB, EB, and SM are co–first authors. CJR and JH are co–senior authors.

Find articles by Ryan, C. in: JCI | PubMed | Google Scholar

1Masonic Cancer Center, and

2Department of Medicine, University of Minnesota-Twin Cities, Minneapolis, Minnesota, USA.

3Glickman Urological and Kidney Institute, Cleveland Clinic, Cleveland, Ohio, USA.

4Department of Clinical and Translational Research, Caris Life Sciences, Phoenix, Arizona, USA.

5Desai Sethi Urology Institute, Sylvester Comprehensive Cancer Center, University of Miami, Florida, USA.

6Department of Medicine, Siriraj Hospital, Mahidol University, Bangkok, Thailand.

7Fox Chase Cancer Center, Philadelphia, Pennsylvania, USA.

8Department of Biostatistics and Bioinformatics, Duke Cancer Institute, Durham, North Carolina, USA.

Address correspondence to: Justin Hwang, 420 Delaware Street SE, MMC 480, Minneapolis, Minnesota 55455, USA. Email: jhwang@umn.edu.

Authorship note: HEB, EB, and SM are co–first authors. CJR and JH are co–senior authors.

Find articles by Hwang, J. in: JCI | PubMed | Google Scholar |

Authorship note: HEB, EB, and SM are co–first authors. CJR and JH are co–senior authors.

Published August 29, 2024 - More info

Published in Volume 9, Issue 20 on October 22, 2024
JCI Insight. 2024;9(20):e183158. https://doi.org/10.1172/jci.insight.183158.
© 2024 Bergom et al. This work is licensed under the Creative Commons Attribution 4.0 International License. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. Published August 29, 2024 - Version history
Received: May 20, 2024; Accepted: August 23, 2024 View PDF Abstract

BACKGROUND. Prostate cancer (PC) is driven by aberrant signaling of the androgen receptor (AR) or its ligands, and androgen deprivation therapies (ADTs) are a cornerstone of treatment. ADT responsiveness may be associated with germline changes in genes that regulate androgen production, uptake, and conversion (APUC).

METHODS. We analyzed whole-exome sequencing (WES) and whole-transcriptome sequencing (WTS) data from prostate tissues (SU2C/PCF, TCGA, GETx). We also interrogated the Caris Precision Oncology Alliance (POA) DNA (592-gene/whole exome) and RNA (whole transcriptome) next-generation sequencing databases. Algorithm for Linking Activity Networks (ALAN) was used to quantify all pairwise gene-to-gene associations. Real-world overall survival was determined from insurance claims data using Kaplan-Meier estimates.

RESULTS. Six APUC genes (HSD3B1, HSD3B2, CYP3A43, CYP11A1, CYP11B1, CYP17A1) exhibited coalescent gene behavior in a cohort of metastatic tumors (n = 208). In the Caris POA dataset, the 6 APUC genes (APUC-6) exhibited robust clustering in primary prostate (n = 4,490) and metastatic (n = 2,593) biopsies. Surprisingly, tumors with elevated APUC-6 expression had statically lower expression of AR, AR-V7, and AR signaling scores, suggesting ligand-driven disease biology. APUC-6 genes instead associated with the expression of alternative steroid hormone receptors, ESR1/2 and PGR. We used RNA expression of AR or APUC-6 genes to define 2 subgroups of tumors with differential association with hallmark pathways and cell surface targets.

CONCLUSIONS. The APUC-6-high/AR-low tumors represented a subgroup of patients with good clinical outcomes, in contrast with the AR-high or neuroendocrine PCs. Altogether, measuring the aggregate expression of APUC-6 genes in current genomic tests identifies PCs that are ligand (rather than AR) driven and require distinct therapeutic strategies.

FUNDING. NCI/NIH 1R37CA288972-01, NCI Cancer Center Support P30 CA077598, DOD W81XWH-22-2-0025, R01 CA249279.

Introduction

In advanced prostate cancer (PC), persistent activity of the androgen receptor (AR) is a key driver of tumor progression, patient survival, and metastases (1, 2). Numerous aberrancies leading to the retained signaling of AR have been described, including non-hormonal drivers, promiscuity of other steroid hormone receptors (SHRs), amplification or mutation in the receptor itself, and, to a lesser extent, variation in the levels of ligands that lead to AR activation. In most cases, death from PC is marked by retained and persistent AR activity. Therefore, further exploration of novel inputs to its activity is needed.

Understanding the relationship between measurable androgens and patient outcomes is nascent. Previous studies have found positive associations between higher serum levels of androgens and patient outcomes in settings in which androgen-lowering drugs are subsequently deployed (2–4). Opposed to serum levels, intratumoral androgen production is a driver of tumor progression (5, 6), and our understanding of its relationship to tumor status and patient outcomes is incomplete. Androgens measurable in the serum typically reflect an adrenal source, whereas the technical feasibility of quantifying tumor-produced androgens is currently not possible. Alternatively, the availability of intratumoral sequencing techniques now enables the study of such production.

Androgens may enter tumor cells directly or be produced within the tumor through conversion of precursors to androgen molecules. This occurs through an enzymatic process involving a suite of more than 20 genes that govern androgen production, uptake, and conversion (APUC). Given the essential need for these steroid molecules or ligands in driving tumor growth, variation in such genes may significantly impact outcomes in PC, either singularly or in aggregate. One critical APUC gene is CYP17A1, which encodes cytochrome P450 17A1. This enzyme converts pregnenolone and progesterone to 17-OH pregnenolone and 17-OH progesterone and 17-OH-pregnenolone to DHEA primarily in the adrenal glands. Such steroid intermediates are critical precursors of testosterone and dihydrotestosterone (DHT). Testosterone and DHT are extremely potent activators of AR, which led to the pharmacologic development of abiraterone (2). Insights into androgens of adrenal (7) and/or intratumoral origin (6) drove the clinical development of the androgen synthesis inhibitor abiraterone acetate. This therapy is a standard-of-care treatment for patients with advanced PCs across many clinical states because it directly targets CYP17A1, preventing the generation of testosterone and DHT (2, 8–10). Beyond abiraterone, opevesostat, which inhibits CYP11A1, is the only other drug specifically designed to ablate the activity of an APUC gene that is available for patients with PC (11).

Germline genetic studies further support the examination of the clinical impact of APUC genes in PC. The APUC protein 3β-hydroxysteroid dehydrogenase-1 (encoded by HSD3B1) is primarily expressed in the peripheral, non-endocrine tissues of the body and catalyzes the conversion of DHEA into androstenedione and other downstream androgens. The protein product of HSD3B1 regulates the production of non-testicular testosterone or DHT (12), which are high-affinity ligands that activate AR. One variant of HSD3B1, in which sustained androgen synthesis is achieved through enzyme stabilization, is associated with resistance to AR-targeted therapies and subsequently unfavorable clinical outcomes (12–15). Beyond HSD3B1, variants of other APUC genes have been implicated in poor patient outcomes (16, 17). This includes the genes encoding the organic ion transport proteins, SLCO2B1 (18, 19), CYP19A1 (20), CYP3A4 (21, 22), and SRD5A2 (23), which results in the conversion of testosterone to DHT, creating the most potent AR agonist. Altogether, previous studies have demonstrated that germline genetic dysregulations in APUC genes can lead to poor patient outcomes, which suggests that intratumoral changes such as RNA levels measured by RNA-seq could also drive AR activation and therapy resistance.

The impact of intratumoral changes in either DNA or RNA of APUC genes is not well characterized. With 21 potential APUC genes, a challenge for defining their relationship to PC is that many intratumoral changes are detected in advanced prostate tumors and they are often not mutually exclusive. To address this and similar challenges, we previously developed the Algorithm for Linking Activity Networks (ALAN) (24), a bioinformatics tool that interprets gene networks within transcriptomic patient data by quantifying the relationship between all genes of interest. Given the complexity of APUC genes, and their many possible interactions, we sought to apply this unbiased approach to determine the intratumoral relationship between APUC genes within large PC patient datasets. In this study, we utilize a hypothesis-driven analysis of the associations of 21 genes known as enzymatic regulators of APUC (16) to determine their relationship PCs.

Results

Six APUC genes exhibit concordant behavior in metastatic PC. We first sought to utilize unbiased approaches to study the interplay between the 21 APUC genes. We used STRING (25) to conduct an iterative search in public databases in order to assess the strength of pairwise gene interactions (Figure 1A). While this confirmed that the family of APUC genes is largely associated with one another, we were unable to deconvolute which APUC gene interactions are relevant for the pathogenesis of PC. We next applied ALAN to quantify all pairwise gene-to-gene interactions based on a continuous measurement (i.e., RNA expression), which we previously utilized to model the changes in AR interactions in benign prostate tissue, primary PC, and metastatic PC (24). We performed ALAN on the transcriptomes from samples annotated as benign prostate tissue (Genotype-Tissue Expression [GTEx] project, n = 245) (26), primary PC (The Cancer Genome Atlas [TCGA], n = 493), and a cohort of metastatic PCs (Stand Up to Cancer/Prostate Cancer Foundation [SU2C/PCF], n = 208) (27). Unsupervised clustering of ALAN outputs indicated that 6 out of 21 APUC genes, HSD3B1, HSD3B2, CYP11A1, CYP11B1, CYP17A1, and CYP3A43, exhibited increasing association with respect to disease progression from benign to primary PC, but only exhibited robust association in metastatic PC, indicating context-specific coalescent behavior (Figure 1B). We examined the 6 genes in greater detail based on their ALAN profiles and again found they exhibited high concordance, but only in metastatic PC and not in benign prostate tissue (Figure 1C). Unsupervised clustering of prostate tumor biopsies (Caris, n = 4,490) and metastatic PC biopsies (Caris, n = 2,593) indicated that the expression of the 6 APUC genes was again clustered as evident by the origin at the same branch point, but interestingly not with AR, as it is the furthest branch point from APUC genes (Figure 1D). When we examined the expression of APUC genes in RNA-seq data from benign tissue (GTEx), at least 1 of the 6 APUC genes exhibited high expression in the adrenal gland, testis, and ovary (top 3 expressed tissues of 6 APUC genes, Figure 1E). Altogether, HSD3B1, HSD3B2, CYP11A1, CYP11B1, CYP17A1, and CYP3A43, hereby defined as APUC-6, demonstrated tissue- and cancer stage–specific interactions, with the most notable interaction in metastatic PC.

Figure 1

Interaction of APUC genes in prostate tissue. (A) STRING analysis was performed to indicate the degree of connection. Based on the output for molecular interactions, we then labeled uptake (blue), production (red), and conversion (yellow) genes. (B) The relationship between APUC genes was examined using ALAN outputs, values between –1 (blue) and 1 (red), based on WTS data from benign prostate tissue, PC, and metastatic PC tumors. Unsupervised hierarchical clustering was performed on ALAN outputs within each dataset. Six APUC genes are highlighted (red font). (C) The ALAN profiles for 6 APUC genes of interest are examined with greater detail in prostate tissue and metastatic PC. (D) Using WTS data from the Caris dataset, we conducted unsupervised hierarchical clustering of prostate and metastatic PC samples based on z score–scaled TPM data. (E) The median expression (TPM) of all APUC genes was examined in the GTEx database across all available tissue sites. Six APUC genes are highlighted (red font).

APUC-6 genes define a subset of metastatic PCs with reduced AR activity. As the APUC-6 genes did not exhibit the expected association with AR, we further examined the AR activity in tumors with elevated APUC-6 expression. In the SU2C/PCF study, metastatic PCs with elevated APUC-6 expression had reduced AR-V7 expression (Figure 2A) and AR activity (Figure 2B). These tumors were not associated with increased neuroendocrine PC (NEPC) scores and exhibited elevated basal- but not luminal-like profiles based on previously defined single-cell RNA-seq signatures (28) (Figure 2C). The lack of correlation with AR activity was observed even when we examined each of the APUC-6 genes in SU2C/PCF or tumors in the Caris cohorts (Figure 2D). In the Caris cohort, we found that NEPCs, as compared with the adenocarcinomas, overall had significantly lower expression of each of the APUC-6 genes (Figure 2E). When stratifying samples based on the expression of APUC-6 genes and the AR gene, we found that more than 75% of tumors that were APUC-6 high/AR low had low or no AR amplifications, whereas the APUC-6-low/AR-high group consisted of tumors that exclusively had high levels of AR amplifications (Figure 2F). In these samples, APUC-6 high or low status was not associated with significant differences in AR amplification status (Supplemental Figure 1A; supplemental material available online with this article; https://doi.org/10.1172/jci.insight.183158DS1). To establish which patients exhibited high AR expression or high APUC-6 expression, we determined the number of patients with high AR expression (>75th, >90th percentile; n = 52, n = 21) and high APUC-6 expression (>75th, >90th percentile; n = 52, n = 21) in the SUC2/PCF dataset. At the 75th percentile threshold, 11.8% (n = 11) of patients coexpressed these signatures, while only 5% (n = 2) coexpressed these signatures using a 90th percentile threshold (Figure 2G).

Figure 2

Key clinical correlates of APUC-6-high tumors. APUC-6 genes were used to stratify metastatic PC patients from the SU2C/PCF samples (27) in which we examined (A) the relative expression of AR-V7 (P value), (B) AR and NEPC signatures (adjusted P values), and (C) Luminal and Basal signatures (adjusted P values). P values for single tests or adjusted P values for multiple comparisons are shown: *Padj ≤ 0.05 and P > 0.01, **Padj < 0.01 and P ≥ 0.001, ***Padj ≤ 0.001. NS, Padj and P > 0.05. (D) The expression of each APUC-6 gene and AR activity was evaluated through Pearson’s correlations using the samples in Abida et al. (27) as well as the primary and metastatic samples from the Caris cohort. The correlation coefficients (R) are shown. (E) The expression of each APUC-6 gene is depicted in primary tumors that are adenocarcinomas or NEPCs. *q < 0.05, **q < 0.01, ***q < 0.001, ****q < 0.0001. (F) AR amplification status (no/low/high amplifications) was examined based on metastatic tumors as stratified by APUC-6 and AR expression. (G) Venn diagrams showing coexpression of AR-high and APUC-6-high PCs (SU2C/PCF) using 2 percentile thresholds — above the 75th and 90th percentiles of target gene(s) expression. (H) We aggregated the proliferation score for the 6 APUC genes based on our prior study (29). The scores are based on the z score of the specific gene as compared with all 17,255 genes in the overexpression screen, in which numbers reflect the standard deviation. We then presented the aggregate scores of genes based on 2 treatment conditions (No Treatment, ADT), as well as the differences in the proliferation scores for every gene (Differential Score).

We next evaluated the functionality of the APUC-6 genes in cells based on a prior overexpression screen in AR-dependent lymph node carcinoma of the prostate (LNCaP) cell lines that included 17,255 genes (29). In this experiment, the z scores reflect relative proliferative effects as based on standard deviation for each of the 17,255 ORFs tested. Based on z scores, each APUC-6 gene, with the exception of HSD3B2, promoted proliferation when LNCaP cells were treated with enzalutamide and cultured in androgen-stripped media (Supplemental Table 1). This supports the known tumor-promoting roles of HSD3B1 (12–15) and the rationale of inhibiting CYP17A1 with abiraterone (2) and CYP11A1 with MK-5684 (ODM-208) (11). As compared with the 17,249 other genes, the aggregate viability scores of APUC-6 genes were distinct (Figure 2H), but the pro-proliferation effects were only observed when the cells were cultured in conditions that mimicked androgen deprivation therapy (ADT) (no steroid hormones and treated with enzalutamide). Upon examining APUC-6 gene expression in tumors formed from C4-2 cells, we found that intratumoral testosterone levels generally exhibited patterns similar to aggregate APUC-6 gene expression (Supplemental Figure 1, B and C). These results support the functional relevance of APUC-6 genes in the setting of prostate tumors; however, increases in APUC-6 gene expression were found in tumors with elevated tumoral androgens, but surprisingly had reduced AR-V7 and AR activity. These results support the functional relevance of APUC-6 genes in the setting of ADT treatment; however, APUC-6 genes were surprisingly associated with reduced AR-V7 and AR activity.

APUC-6 genes exhibit robust associations with ESR1, ESR2, and PGR in prostate and metastatic biopsies. In order to identify the relative interaction of APUC-6 genes with respect to all detectable genes in metastatic PC, we next conducted a dimensional reduction of ALAN outputs (Figure 3A). This approach allowed us to visualize which genes behaved similarly to the APUC-6 genes. Here we found that APUC-6 genes were indeed in close proximity on the UMAP, but were distant from AR and AR cofactors, including HOXB13, FOXA1, GRHL2, PRMT1, and EP300. Surprisingly, when examining all alternative SHRs, including ESR1/2 (estrogen receptors, ERs), PGR (progesterone receptor, PR), NR3C1 (glucocorticoid receptor, GR), NR3C2 (mineralocorticoid receptor, MR), APUC-6 genes exhibited the most similarity to ESR1, ESR2, and PGR. Interestingly, ESR1 is a current therapeutic target in breast cancers. When revisiting the metastatic tumors stratified by APUC-6 levels, we found that tumors with high APUC-6 expression had increased expression of ESR1, ESR2, and PGR, but reduced expression of AR (Figure 3B). Using APUC-6 genes as a signature, we determined that the associations with ESR1, ESR2, and PGR were positive and significant (adjusted P value < 0.0001), while the associations with AR, GR, and MR were not significant (Figure 3C). Furthermore, we examined whether each APUC-6 gene was associated with ESR1/2 and PGR, which confirmed a positive correlation between each APUC-6 gene and ESR1/2 and PGR expression (Figure 3D).

Figure 3

APUC-6 genes are associated with the expression of alternative hormone receptors (ESR1, ESR2, PGR) instead of AR. (A) UMAP was used for dimensional reduction of the ALAN outputs from metastatic PC patients, in which the distance between genes (gray dots) indicates the similarity of ALAN gene behavior. Four groups of genes are specifically labeled as APUC (purple), APUC-6 (red), AR-Related (blue), and Alternative SHRs (green). (B) Based on stratifying patients by APUC-6 expression, we examined the relative expression of cancer-related hormone receptors. (C) Pearson’s correlation was used to examine the relative expression of hormone receptors with respect to APUC-6 genes (APUC-Score) in metastatic PC samples. (D) ESR1/2 and PGR expression levels were correlated with each APUC-6 gene. In C and D, the correlation coefficient (R) and adjusted P values (FDR adjusted for multiple comparisons) are shown: *Padj ≤ 0.05 and P > 0.01, **Padj < 0.01 and P ≥ 0.001, ***Padj ≤ 0.001. NS, Padj > 0.05.

We next aimed to confirm the relationship between APUC-6 genes with the expression of all hormone receptors in the larger Caris dataset. To do so, we cross-correlated the expression of each APUC-6 gene with the expression of all other SHRs. In the 4,490 prostate tumor biopsies, APUC-6 genes exhibited a robust correlation with one another, and all had strong positive correlations with ESR1, ESR2, and PGR (Spearman’s R 0.13–0.52), while the correlation with AR was relatively weak (Spearman’s R 0.06–0.20). In the 2,593 metastatic samples, we observed similar associations among the APUC-6 genes, but with greater correlation with ESR1, ESR2, and PGR (Spearman’s R 0.37–0.53), while the association with AR was further diminished (Spearman’s R –0.03 to 0.19) (Figure 4A). We next conducted the same analysis upon stratifying samples by metastatic biopsy sites, including lymph node (n = 833), bone (n = 533), liver (n = 360), bladder (n = 313), lungs (n = 114), brain (n = 24), and the adrenal gland (n = 22) (Figure 4B). APUC-6 genes exhibited robust correlation regardless of metastatic tissue sites, with the most robust association in the 22 adrenal gland samples. However, we also note that normal adrenal gland tissue also had the highest expression of APUC-6 genes (Figure 1E). The 3 SHRs ESR1, ESR2, and PGR exhibited the strongest correlation (Spearman’s R 0.49–0.71) with APUC-6 genes in the 533 bone metastasis samples. Similar to what we found in the SU2C/PCF cohort, APUC-6 genes exhibited the strongest correlation with ESR2 across all metastatic sites. The only tissue site in which APUC-6 genes had a notable AR correlation was in the adrenal gland, in which we also noted that APUC-6 genes are generally expressed in the benign prostate tissue (Figure 1E). When we stratified patients based on self-reported race (European American, African American, Asian Pacific Islander), the degrees of associations showed consistent trends across populations (Supplemental Figure 2). Altogether, APUC-6 genes exhibited consistent association with each other in prostate tumor biopsies and