Figure 1.
Characterization of LuCaP35 and LuCaP96 prostate cancer xenografts and responses to systemic androgen suppression.
(A) Representative FFPE samples of each xenograft were stained with hematoxylin and eosin (H&E) and for expression of the androgen receptor (AR) and PSA as indicated. The scale bar (depicted on the PSA figures for ease of visualization) are 50µm. Kaplan-Meier plots of progression free survival (defined as tumor size <750 mm3) in mice bearing LuCaP35 (B) or LuCaP96 (C) xenografts. Intact male SCID mice were implanted subcutaneously with 30 mm3 pieces of the indicated xenografts. When tumors reached ∼300 mm3, mice were randomly enrolled into cohorts that were either left intact (No Cx) or castrated (Cx). P-values for curve comparisons were generated using the Mantel-Haenszel logrank test. (D) Mean and standard deviation of tissue testosterone (T, black bar) and DHT (gray bar) levels measured by mass spectrometry in tumors of the indicated xenograft (passaged in intact mice). (E) Relative expression of transcripts for the indicated steroidogenic genes was calculated using the delta dCt method (fold change = 2∧ddCt). Genes differentially expressed in LuCaP35 vs. LuCaP96 within one order of magnitude are indicated within the gray lines. Significant differences (by Welch’s t test; p<0.05) are indicated by black circles; white circles indicate genes that were not significantly different between LuCaP35 and LuCaP96 (all values given in Supplementary Data 2).Upward triangles indicate highly differentially expressed genes specifically leading to increased T (AKR1C3, 40 fold) or increased DHT levels (SRD5A1, 5.0 fold; 17BHSD10 4.8 fold; RLHSD, 99 fold). Downward triangles indicate highly differentially expressed genes specifically mediating DHT catabolism (AKR1C2, 7 fold; UGT2B15, 3000 fold).
Figure 2.
Tumor growth and androgen levels in prostate cancer xenografts treated with castration and dutasteride.
Mean tumor volumes in mice bearing LuCaP35 (A) and LuCaP96 (B) xenografts. Intact male SCID mice were subcutaneously implanted with 30 mm3 pieces of the indicated xenograft. When tumors reached ∼300mm3, mice were castrated and randomly enrolled into cohorts treated with either vehicle (Cx) or dutasteride (Cx+Dut) for 8 weeks (denoted by black line above x-axis). Mean tumor volumes are depicted for each treatment group at the indicated days post enrollment (Cx, black squares; Cx+Dut, white circles).
Figure 3.
Response to dutasteride by tumor size at enrollment in LuCaP35 xenografts.
Kaplan-Meier plots of progression free survival (defined as tumor size <750 mm3) in all LuCaP35 tumors treated with castration alone (Cx) vs. castration + dutasteride (Cx + Dut) (A), and in tumors enrolled into treatment when tumors were <250 mm3 (B). P-values for curve comparisons were generated using the Mantel-Haenszel logrank test. Mean tumor volume growth curves at the indicated days post enrollment in LuCaP35 tumors enrolled into treatment when tumors were <250 mm3 (C), between 250–400 mm3 (D), and >400 mm3 (E). Dutasteride treatment was continued for 8 weeks (denoted by black line above x-axis) in the castration + dutasteride group.
Figure 4.
Androgen levels and AR expression in prostate cancer xenografts treated with castration and dutasteride.
Tissue testosterone (T, black bars) and DHT (gray bars) levels were measured by mass spectrometry in LuCaP35 (A) and LuCap96 (B) tumors resected from intact mice (No Cx), and from mice treated with castration alone (Cx) or castration + dutasteride (Cx+Dut) at early time points (d3-21, while still on therapy, indicated by double-headed arrows), or at castration-resistant re-growth (defined as >750 mm3). P values computed from Welch’s two sample t test (p<0.05 were considered significant). Single stars indicate a statistically significant difference in DHT levels between Cx vs. Cx+Dut treated samples at d3-21 of treatment. Double stars indicate a significant difference in DHT levels between Cx vs. Cx+Dut treated samples even after castration-recurrent re-growth. No other comparisons between Cx vs. Cx + Dut treated groups were significant. Expression of full length (FL) AR and the AR variant 7 (ARV7) truncated splice variant was measured in LuCaP35 (C) and LuCaP96 (D) at the time of tumor re-growth to 750 mm3. Transcript expression was measured by qRT-PCR and normalized to expression of the housekeeping gene RPL13A within each sample to yield the delta cycle threshold (dCt). The relative difference in expression between the indicated treatment groups was calculated using the delta dCt method (fold change = 2∧ddCt). P values computed from Welch’s two sample t test (p<0.05 were considered significant).
Figure 5.
Androgen levels and expression of AR isoforms in castration resistant prostate tumor metastases.
(A) Androgen levels were measured by mass spectrometry in 1–3 soft tissue metastases obtained from each of 8 patients via rapid autopsy. The graph depicts the absolute androgenicity index in each tumor, calculated using a 5∶1 ratio for the relative potency of DHT to T (e.g (5×DHT)+(1×T)). The portion of the total androgenicity contributed by T or DHT is represented by the stacked black and gray bars, respectively. The gray line represents a hypothetical cut point in the androgenicity index between tumors with relatively higher vs. relatively lower tissue androgenicity. Data for T and DHT in LuCaP35 and LuCaP96 tumors from intact mice (No Cx) or recurring after combined hormonal therapy (castration + dutasteride, Cx+D) is presented for comparison. (B) IHC staining scores for AR and PSA expression in 4–5 separate tumor metastases from five of the patients presented in panel A (as indicated by arrows). The % of each patient’s tumors demonstrating no, faint, weak or strong staining for the indicated antibody is presented. AR stains were separately performed using either N or C terminal antibodies to identify N+ but C– tumors consistent with the presence of C terminal truncated AR variants.