Contribution of mouse adrenal glands to intratumor androgens and growth of castration-resistant prostate cancer xenografts.

Abstract

224 Background: Historical studies using radioimmunoassay did not detect adrenal androgens in serum from castrate mice, and it is widely held that rodent adrenal glands do not make adrenal androgens due to lack of CYP17A expression. This contrasts with the clinical setting in which circulating adrenal androgens are significant and inhibition of adrenal CYP17A markedly decreases tissue androgens. Methods: We evaluated CYP17A (by methylation, transcript and protein) and androgens (by mass spectrometry) in adrenal glands of CB17-NOD/SCID mice. We determined the impact of adrenalectomy (ADX) on suppressing tumor androgens and growth in two patient derived xenografts (PDX) models of castration resistant prostate cancer (CRPC). Results: CYP17A is unmethylated, and transcript and protein are present in adrenals from intact and castrate mice. In castrate mice adrenal levels of DHEA (0.75 pg/mg), androstenedione (AED; 44pg/mg), T (12.5pg/mg) and DHT (4.7 pg/mg) are detectable and nearly 2 orders of magnitude higher than in kidney, liver or muscle (p < 0.05 for all). In castrate mice bearing LuCaP35 and LuCaP96 tumors, ADX suppressed tumor steroids at 21-30 days after ADX vs castration alone (LuCaP35: AED 0.05 vs 0.02 pg/mg, p = 0.005; T 0.64 vs 0.03 pg/mg p < 0.001; DHT 2.3 vs 0.23 pg/mg, p = 0.0003; LuCaP96: AED 0.06 vs 0.02 pg/mg p = 0.08; T 0.81 vs 0.03 pg/mg, p < 0.0001; DHT 1.3 vs 0.04 pg/mg, p = 0.002), and delayed time to tumor re-growth (median survival: LuCaP35 33 vs 179 days, p = 0.006; LuCaP96 25 vs 301 days, p = 0.0012). Whereas tumor recurrence in LuCaP96 was uniformly delayed, a subset of tumors in LuCaP35 showed more rapid regrowth after ADX (66 days vs undefined, p = 0.008), with a trend toward increased levels of tumor steroids (progesterone 0.25 vs 0.02 pg/mg, p = 0.07; T 0.14 vs 0.02, p = 0.04; DHT 0.36 vs 0.21, p = 0.17). Recurrent tumors variably showed induction of full length and truncated AR variants and/or induction of steroidogenic enzymes as potential mechanisms of resistance. Conclusions: Adrenally-derived steroids are produced in mice and contribute to tumor androgen levels and growth in PDX models. Mice are an appropriate model for evaluation of steroidogenesis inhibitors in CRPC xenograft studies