AKR1C3 Converts Castrate and Post-Abiraterone DHEA-S into Testosterone to Stimulate Growth of Prostate Cancer Cells via 5-Androstene-3β,17β-Diol

Androgen receptor signaling inhibitors (ARSI) are used to treat castration-resistant prostate cancer (CRPC) to stop a resurgence of androgen receptor (AR) signaling. Despite early success, patients on ARSIs eventually relapse, develop drug resistance, and succumb to the disease. Resistance may occur through intratumoral steroidogenesis mediated by upregulation of aldo-keto reductase family 1C member 3 (AKR1C3). Patients treated with leuprolide (castrate) and those treated with leuprolide plus abiraterone (post-Abi) harbor a reservoir of DHEA-S which could fuel testosterone (T) biosynthesis via AKR1C3 to cause a resurgence of prostate cancer cell growth. We demonstrate that concentrations of DHEA-S found in castrate and post-Abi patients are (i) converted to T in an AKR1C3-dependent manner in prostate cancer cells, and (ii) in amounts sufficient to stimulate AKR1C3-dependent cell growth. We observed this in primary and metastatic prostate cancer cell lines, CWR22PC and DuCaP, respectively. Androgen measurements were made by stable isotope dilution LC-MS/MS. We demonstrate AKR1C3 dependence using stable short hairpin RNA knockdown and pharmacologic inhibitors. We also demonstrate that free DHEA is reduced to 5-androstene-3β,17β-diol (5-Adiol) by AKR1C3 and that this is a major metabolite, suggesting that in our cell lines 5-Adiol is a predominant precursor of T. We have identified a mechanism of ARSI resistance common to both primary and metastatic cell lines that is dependent on the conversion of DHEA to 5-Adiol on route to T catalyzed by AKR1C3. SIGNIFICANCE: We show that reservoirs of DHEA-S that remain after ARSI treatment are converted into T in primary and metastatic prostate cancer cells in amounts sufficient to stimulate cell growth. Pharmacologic and genetic approaches demonstrate that AKR1C3 is required for these effects. Furthermore, the route to T proceeds through 5-Adiol. We propose that this is a mechanism of ARSI drug resistance.