Suppression of acquired docetaxel resistance in prostate cancer through depletion of notch- and hedgehog-dependent tumor-initiating cells

Abstract

Acquired resistance to Docetaxel precedes fatality in hormone-refractory prostate cancer (HRPC). However, strategies that target Docetaxel resistant cells remain elusive. Using in vitro and in vivo models, we identified a subpopulation of cells that survive Docetaxel exposure. This subpopulation lacks differentiation markers and HLA class I (HLAI) antigens, while overexpressing the Notch and Hedgehog signaling pathways. These cells were found in prostate cancer tissues and were related to tumor aggressiveness and poor patient prognosis. Notably, targeting Notch and Hedgehog signaling depleted this population through inhibition of the survival molecules AKT and Bcl-2, suggesting a therapeutic strategy for abrogating Docetaxel resistance in HRPC. Finally, these cells exhibited potent tumor-initiating capacity, establishing a link between chemotherapy resistance and tumor progression.

Figure 1. Phenotypical Characterization of Docetaxel-Resistant Cells

(A) Genes with at least 1.8-fold increase ↑ or decrease ↓ in transcript expression comparing parental and Docetaxel-resistant cells. (B) Gene ontology categories of overlapping genes. Categories with statistical significance (p ≤ 0.01) are represented. *GO categories related to cell proliferation, cell death, and response to drugs. **GO categories related to developmental processes. ***GO category related to antigen presentation. (C) Heatmap illustrates epithelial differentiation, prostate specific, HLAI, and developmental (Notch and Hedgehog) gene expression of parental and Docetaxel-resistant cells. (D) Immunoblotting and quantification of parental and Docetaxel-resistant cells for indicated proteins. SCaBER was used as a positive control for high molecular weight cytokeratins and p63. (E) Immunofluorescent staining of parental and Docetaxel-resistant cells for indicated proteins. See also Figure S1.

Figure 2. Docetaxel-Resistant Cells Are Present in Prostate Cancer Tissue Samples and Associate with Tumor Aggressiveness

(A and B) Hematoxylin and eosin (H&E) and immunofluorescent staining analyses of prostate cancer metastases for indicated proteins. White arrows point to CK⁻ cells. Corresponding box plots show protein expression in CK⁻ and CK⁺ cells, including (from top) upper outliers, maximum (excluding outliers), upper quartile, median, lower quartile, minimum (excluding outliers), and lower outliers. (C) CK18 and CK19 immunohistochemistry of clinical metastatic prostate cancer tissues nontreated and treated with Docetaxel. (D) Association between the percentage of CK⁻ cells with Gleason Score and pathological stage in primary prostate cancer tissues. (E) Kaplan-Meier analysis of biochemical recurrence free survival of primary prostate cancer patients (n = 31) with low CK⁻ content (≤1.3%) compared to high CK⁻ content (>1.3%). Representative samples with low and high percentage of CK⁻ cells. Black arrows point to CK⁻ cells. Data is represented as means ± SD. See also Figure S2 and Table S1.

Figure 3. Docetaxel Exposure Selects for Pre-Existing Resistant Prostate Cancer Cells

(A) Working hypotheses; transition versus enrichment-selection induced by Docetaxel. (B) Immunofluorescence and flow cytometry quantification of CK18 and CK19 expression in DU145 and 22Rv1. White arrows point to cells with a CK⁻ phenotype. (C) Flow cytometry analysis of DU145-pCK19-GFP and 22Rv1-pCK19-GFP treated with Docetaxel (72 hr). (D) Colony formation assay and quantification of sorted DU145-pCK19-GFP and 22Rv1-pCK19-GFP cells cultured with Docetaxel, 10 and 50 nM respectively, for 72 hr, or DMSO. (E) Time-lapse microscopy of DU145-pCK19-GFP treated with Docetaxel. Dotted area shows a CK19/GFP⁻ cell. (F) Immunoblots of GFP and Docetaxel-resistance markers in DU145-pCK19-GFP and 22Rv1-pCK19-GFP sorted cells, as well as in unsorted DU145-pCK19-GFP and 22Rv1-pCK19-GFP cells exposed to Docetaxel (72 hr) at the same concentrations as in (D). (G) Colony formation assays and quantification of DU145-pCK19-GFP and 22Rv1-pCK19-GFP sorted cells cultured with or without Mitoxantrone 125 and 500 nM, Cisplatin 5 and 2.5 μM, and Vinorelbine 500 and 750 nM, respectively (all 72 hr). Data is represented as means ± SD of triplicate experiments. *p < 0.0001. See also Figure S3 and Movies S1 and S2.

Figure 4. Docetaxel-Resistant Prostate Cancer Cells Are Dependent on Notch and Hedgehog Signaling

(A) Colony formation assay and quantification of DU145-pCK19-GFP- and 22Rv1-pCK19-GFP-sorted cells expressing shRNAs against GLI1, GLI2, and NOTCH2 alone, as well as double (GLI1 and GLI2) and triple knockdowns. (B) Flow cytometry analysis of DU145-pCK19-GFP and 22Rv1-pCK19-GFP treated with Cyclopamine (1 μM) and/or DBZ (1 μM, both 48 hr). (C) Immunoblots of indicated proteins in DU145-pCK19-GFP cells treated with the same conditions as (B). (D) Colony formation assay and quantification of DU145-pCK19-GFP-sorted cells exposed for 72 hr to Cyclopamine, GDC-0449, DBZ, and Compound E (all 1 μM), alone or in combination (Cyc+DBZ or GDC+CE). (E) Flow cytometry analysis after Docetaxel (48 hr) alone or in combination with Cyclopamine and/or DBZ (both 1 μM). DU145-pCK19-GFP and 22Rv1-pCK19-GFP cells were treated with 10 and 50 nM Docetaxel, respectively. (F) Colony formation assay and quantification of parental DU145 and 22Rv1 cells exposed for 72 hr to Docetaxel (10 and 50 nM, respectively) alone or in combination with Notch inhibitors (CE or DBZ, both 1 μM) and/or Hedgehog inhibitors (Cyclopamine or GDC-0449, both 1 μM). Data are represented as means ±SD of triplicate experiments. *p < 0.05. See also Figure S4.

Figure 5. Notch and Hedgehog Inhibition Abrogates the Acquisition of Docetaxel Resistance

(A) Changes in tumor volume of DU145 and 22RV1 xenografts treated with Dexamethasone alone, double combinations, triple combinations, and quadruple combination. Dose schedules were Dexamethasone (15 mg/kg/i.p. daily), Docetaxel (10 mg/kg/i.p. once a week for 3 weeks every 4 weeks), DBZ (10 mM/kg/ip daily for 15 days every 4 weeks), and Cyclopamine (50 μg/kg/sc daily). (B) Quantitative RT-PCR of Notch and Hedgehog target genes in DU145 and 22RV1 xenografts obtained from mice treated with the same drugs and concentrations as in (A). Bars represent fold change in mRNA levels relative to vehicle (control). (C) Microphotographs illustrate the expression of low molecular weight cytokeratins (CK18 and CK19) in DU145 and 22RV1 prostate cancer tumor xenografts in NOD/SCID mice treated for 4 weeks with the same drugs as in (A). Magnifications illustrate CK⁻ cells. Histogram represents the percentage of CK⁻ cells detected in DU145 and 22RV1 xenografts for each treatment arm. Four xenografts for each treatment group where analyzed. Data is represented as means ± SD. *p < 0.05. See also Figure S5.

Figure 6. Notch and Hedgehog Signaling Regulate Survival Molecules in Docetaxel-Resistant Cells

(A) Immunoblots of indicated proteins in CK19/GFP⁻ and CK19/GFP⁺ sorted cells from DU145-pCK19-GFP and 22Rv1-pCK19-GFP. (B) Immunoblots of indicated proteins in DU145-pCK19-GFP⁻ and 22Rv1-pCK19-GFP⁻ cells exposed for 72 hr to DBZ and/or Cyclopamine (both 1 μM). (C) Immunoblots of indicated proteins in DU145-pCK19-GFP⁻ and 22Rv1-pCK19-GFP⁻ cells exposed for 72 hr to LY294002 (50 μM) and/or ABT-737 (10 μM). (D) Colony formation assay and quantification of colonies derived from GFP⁻ and GFP⁺ cells exposed to the same drugs and concentrations as in (C). (E) Flow cytometry analysis after 48 hr administration of Cyclopamine (1 μM) and/or DBZ (1 μM) of DU145-pCK19-GFP and 22Rv1-pCK19-GFP cells stably transfected with empty vector (EV), MYR-AKT and BCL2. Immunoblots illustrate the overexpression levels of pAKT (Ser-473) and Bcl-2. (F) Colony formation assay and quantification of DU145-pCK19-GFP- and 22Rv1-pCK19-GFP-sorted cells stably transfected with empty vector (EV), MYR-AKT, or BCL2 and exposed for 72 hr to Cyclopamine (Cyc), GDC-0449 (GDC), DBZ, and Compound-E (CE, all 1 μM), alone or in combination. (G) Colony formation assays and quantifications of GFP⁺-sorted DU145-pCK19-GFP and 22Rv1-pCK19-GFP stably transfected with EV, MYR-AKT, or BCL2 and treated for 72 hr with Mitoxantrone 125 and 500 nM, Cisplatin 5 and 2.5 μM, and Vinorelbine 500 and 750 nM, respectively. Data is represented as means ± SD of three independent experiments. *p < 0.05. See also Figure S6.

Figure 7. Docetaxel-Resistant Prostate Cancer Cells Have High Tumor-Initiating Capacity

(A) Tumor incidence, T-IC frequency, and latencies 36 weeks after injection of limiting dilutions of parental and Docetaxel-resistant cells. (B) Tumor incidence, T-IC frequency, and latencies 38 weeks after injection of limiting dilutions of DU145 and 22RV1 HLAI-sorted cells. (C) Image of a mouse bearing tumors after injection of DU145 HLAI⁻ cells in the upper flanks and HLAI⁺ cells in the lower flanks. H&E and immunofluorescence of indicated proteins in representative tumor xenografts generated from DU145 and 22RV1 HLAI⁻ cells. White arrows point to CK⁻ cells with positive nuclear staining of transcription factors and lack of HLAI and AR. (D) Table summarizes prostate cancer patients' clinicopathological characteristics, tumor incidence, T-IC frequency, and latencies after 61 weeks of injection of limiting dilutions of HLAI-sorted cells from fresh human prostate cancer samples. (E) H&E and immunofluorescence analysis of indicated proteins in human tumors and primary and secondary xenografts generated from HLAI⁻ cells. Patient 9 is represented. White arrows point to CK⁻ cells with nuclear expression of transcription factors and lack of HLAI and AR. (F) Tumor incidence and latencies 24 weeks after injection of 100 HLAI⁻-sorted cells from prostate cancer xenografts treated with DMSO, Dexamethasone 15 mg/kg/i.p. daily, Cyclopamine 50 μg/kg/sc daily plus dexamethasone, DBZ 10 μM/kg/i.p. daily for 15 days every 4 weeks plus Dexamethasone, or with triple combination. Data is represented as means ± SD. *p < 0.05. See also Figure S7 and Table S2.