11-Ketotestosterone is the predominant active androgen in prostate cancer patients after castration

Abstract

BACKGROUNDContinued androgen receptor (AR) signaling constitutes a key target for treatment in metastatic castration-resistant prostate cancer (CRPC). Studies have identified 11-ketotestosterone (11KT) as a potent AR agonist, but it is unknown if 11KT is present at physiologically relevant concentrations in patients with CRPC to drive AR activation. The goal of this study was to investigate the circulating steroid metabolome including all active androgens in patients with CRPC.METHODSPatients with metastatic CRPC (n = 29) starting a new line of systemic therapy were included. Sequential plasma samples were obtained for measurement of circulating steroid concentrations by multisteroid profiling employing liquid chromatography-tandem mass spectrometry. Metastatic tumor biopsy samples were obtained at baseline and subjected to RNA sequencing.RESULTS11KT was the most abundant circulating active androgen in 97% of patients with CRPC (median 0.39 nmol/L, range: 0.03-2.39 nmol/L), constituting 60% (IQR 43%-79%) of the total active androgen (TA) pool. Treatment with glucocorticoids reduced 11KT by 84% (49%-89%) and testosterone by 68% (38%-79%). Circulating TA concentrations at baseline were associated with a distinct intratumor gene expression signature comprising AR-regulated genes.CONCLUSIONThe potent AR agonist 11KT is the predominant circulating active androgen in patients with CRPC and, therefore, one of the potential drivers of AR activation in CRPC. Assessment of androgen status should be extended to include 11KT, as current clinical approaches likely underestimate androgen abundance in patients with CRPC.TRIAL REGISTRATIONNetherlands Trial Register: NL5625 (NTR5732).FUNDINGDaniel den Hoed Foundation and Wellcome Trust (Investigator Award WT209492/Z/17/Z).

Keywords: Endocrinology; Oncology; Prostate cancer.

Figure 1. Androgen biosynthesis.

Schematic overview of the conversion of adrenal precursor steroids to the potent androgens T, DHT, and 11KT. The molecular structures of the active androgens are shown, with the 11-keto group highlighted in gray. DHEA, dehydroepiandrosterone.

Figure 2. Patient and sample selection.

Selection and exclusion of CIRCUS study samples for multisteroid profiling, glucocorticoid quantification, survival analysis, and tumor biopsy analysis. LC-MS/MS, liquid chromatography-tandem mass spectrometry.

Figure 3. 11-KT is the most abundant circulating active androgen in CRPC patients at baseline.

(A) Active androgen concentrations of all patients with CRPC before the start of the first treatment after enrollment (n = 29). Box plot depicts the upper and lower quartiles, with the median shown as a solid line; whiskers indicate the range. Dots indicate individual data points. Statistical analysis was performed by 1-way ANOVA (P < 0.0001) with Tukey’s multiple-comparison test. *P < 0.05, ***P < 0.001. (B) The relative abundance of each androgen is shown as a percentage of the total androgen pool. Box plot depicts the upper and lower quartiles, with the median shown as a solid line; whiskers indicate the range. Dots indicate individual data points. (C) Active androgen concentrations are shown for all baseline samples (n = 34). Values below the analytical limit of quantification are shown if relevant calibrator and spiked quality control samples were accurate and reproducible with signal-to-noise ratio greater than 10:1. Samples with undetectable concentrations were set to 0.5 times the lowest accurate calibration sample for statistical purposes. Conventional clinical cutoff values for castrate testosterone levels (0.69 and 1.74 nmol/L, or 20 and 50 ng/dL, testosterone) are indicated on the y axis for reference. OT, on treatment; PD, progressive disease.

Figure 4. Effects of exogenous glucocorticoid treatment on circulating steroid concentrations.

(A) Differences between steroid concentrations at baseline (white boxes) and OT (red boxes) were assessed in patients who were exogenous glucocorticoid untreated at baseline and who started treatment with glucocorticoids (n = 13) by Wilcoxon’s signed-rank test. The individual data points are shown for (B) T and (C) 11KT at baseline and OT in patients who started therapy with glucocorticoids (red lines, n = 13) or without glucocorticoids (black lines, n = 6). (D) Differences between concentrations at baseline (gray boxes) and OT (blue boxes) were assessed in patients who were glucocorticoid treated at baseline and discontinued glucocorticoid treatment (n = 5). The individual data points are shown for (E) T and (F) 11KT at baseline and OT in patients who continued treatment with glucocorticoids (gray lines, n = 10) or were withdrawn from glucocorticoids (blue lines, n = 5). Box plot depicts the upper and lower quartiles, with the median shown as a solid line; whiskers indicate the range. Effects of treatment were assessed by Wilcoxon’s rank-sum test, while group differences were assessed by Mann-Whitney U test. Lines connect individual patients and group medians (squares) are shown beside the individual data points. *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 5. Effects of TA concentration on PFS and intratumor gene expression.

PFS curves are shown for patients stratified into 2 groups with concentrations above or below median (A) TA (defined as the sum of T, DHT, and 11KT), (B) 11KT, or (C) T. Log-rank test for survival was used to determine difference between the low- and high-TA groups. (D) Heatmap of differentially expressed genes (n = 24) across TA concentration in the tumor samples. Differential gene expression was determined using TA concentration as a continuous variable. Heatmap displays mean-centered and normalized (variance-stabilizing transformation) read counts. Unsupervised hierarchical clustering (Euclidean distance and Ward.D2 method) was performed on genes and samples. Upper tracks display biopsy site and TA concentrations.