Deletion of the p16INK4a tumor suppressor and expression of the androgen receptor induce sarcomatoid carcinomas with signet ring cells in the mouse prostate

Abstract

The tumor suppressor p16Ink4a, encoded by the INK4a gene, is an inhibitor of cyclin D-dependent kinases 4 and 6, CDK4 and CDK6. This inhibition prevents the phosphorylation of the retinoblastoma protein (pRb), resulting in cellular senescence through inhibition of E2F-mediated transcription of S phase genes required for cell proliferation. The p16Ink4a plays an important role in tumor suppression, whereby its deletion, mutation, or epigenetic silencing is a frequently observed genetic alteration in prostate cancer. To assess its roles and related molecular mechanisms in prostate cancer initiation and progression, we generated a mouse model with conditional deletion of p16Ink4a in prostatic luminal epithelium. The mice underwent oncogenic transformation and developed prostatic intraepithelial neoplasia (PIN) from eight months of age, but failed to develop prostatic tumors. Given the prevalence of aberrant androgen signaling pathways in prostate cancer initiation and progression, we then generated R26hARL/wt:p16L/L: PB-Cre4 compound mice, in which conditional expression of the human AR transgene and deletion of p16Ink4a co-occur in prostatic luminal epithelial cells. While R26hARL/wt:PB-Cre4 mice showed no visible pathological changes, R26hARL/wt:p16L/L: PB-Cre4 compound mice displayed an early onset of high-grade PIN (HGPIN), prostatic carcinoma, and metastatic lesions. Strikingly, we observed tumors resembling human sarcomatoid carcinoma with intermixed focal regions of signet ring cell carcinoma (SRCC) in the prostates of the compound mice. Further characterization of these tumors showed they were of luminal epithelial cell origin, and featured characteristics of epithelial to mesenchymal transition (EMT) with enhanced proliferative and invasive capabilities. Our results not only implicate a biological role for AR expression and p16Ink4a deletion in the pathogenesis of prostatic SRCC, but also provide a new and unique genetically engineered mouse (GEM) model for investigating the molecular mechanisms for SRCC development.

Fig 1. Generation of p16^Ink4a conditional deletion and AR conditional transgenic mice.

(A) A scheme was shown of the conditional p16^Ink4a deletion target construct. (B) Genomic PCR was used to confirm the p16^Ink4a targeting allele (blue solid arrow) and deleted allele (blue empty arrow) of different mouse tissues. (C) Representative H&E of the prostate tissue was shown for dorsal/lateral prostate (DLP), C1, and ventral prostate (VP), C2, for p16^L/L:PB-Cre4 mice. (D) A scheme was shown of the conditional AR transgene target construct and the p16^Ink4a deletion target construct. For the AR transgene construct, a PGK-neomycin cassette was flanked with loxP sites (LSL), red triangles, inserted between the CAG promoter and a FLAG-tagged AR coding sequence. For the p16^nkK4a deletion, two LoxP sites, red triangles, were inserted to flank exon 1 alpha. Targeting constructs were shown, upper and Cre-driven recombined alleles for both constructs are shown below. (E) Genomic PCR was used to confirm the AR (black solid arrow) or p16^Ink4a (blue solid arrow) targeting allele and AR recombined (black empty arrow) or p16^Ink4a deleted (blue empty arrow) allele from different tissues of R26hAR^L/wt:PB-Cre4 and R26hAR^L/wt:p16^L/L:PB-Cre4 mice. (F-H) Immunofluorescence (IF) assays. Prostate tissues of R26hAR^L/wt:PB-Cre4, p16^L/L:PB-Cre4, and R26hAR^L/wt:p16^L/L:PB-Cre4 mice were used for IF analyses and tissues sections were probed for AR (red, F1, G1, H1) or p16^INK4a (green, F2, G2, H2) and DAPI (blue, F3, G3, H3), respectively.

Fig 2. Histological and immunohistochemistry analyses of the prostate from R26hAR^L/wt:PB-Cre4, p16^L/L:PB-Cre4 and R26hAR^L/wt:p16^L/L:PB-Cre4.

(A-B) Representative H&E images of the prostate tissue was shown for AP, DLP, and VP for R26hAR^L/wt:PB-Cre4 (A1-A3, magnification A1’-A3’) and R26hAR^L/wt:p16^L/L:PB-Cre4 mice (B1-B3, magnification B1’-B3’). Scale bar used was 100 μm or 25 μm. (C). Representative H&E and IHC images of adjacent prostate tissue sections of AP from the p16^L/L:PB-Cre4 mice were shown for staining with different antibodies as labeled in right bottom corner. (D). Representative H&E and IHC images of adjacent prostate tissue sections from the p16^L/L: R26hAR^L/wt:PB-Cre4 mice were shown for staining with different antibodies (please see S1B and S1B’ Fig for positive staining of synaptophysin). Scale bar used was at 25 μm for all panels (C1-D9). (E) Pathological abnormalities of R26hAR^L/wt:PB-Cre4, p16^L/L:PB-Cre4, and R26hAR^L/wt:p16^L/L: PB-Cre4 mice reported for different time points.

Fig 3. Synergistic effect of conditional AR expression and p16^Ink4a deletion accelerates formation of adenocarcinoma.

(A) Pathological abnormalities of the prostate were listed for both p16^L/L:PB-Cre4 and R26hAR^L/wt:p16^L/L:PB-Cre4 mice. (B-C) Representative image of prostatic tumors from R26hAR^L/wt:p16^L/L:PB-Cre4 mice. Scale bar used was at 200 μm. B1-C4. Pathological dissemination of typical morphological characteristics of HGPIN (B1, C1) adenocarcinoma (B2, C2 and C3) and regions of necrosis (B3) and sarcomatoid carcinoma (B4, C4). D1-D8. Representative H&E and IHC images of adjacent prostate tissue sections of adenocarcinoma lesions from the R26hAR^L/wt:p16^L/L:PB-Cre4 mice were shown for staining with different antibodies (right bottom corner). Scale bar used was at 25 μm for all panels (B1-D8).

Fig 4. Simultaneous conditional expression of AR transgene and loss of p16^Ink4a expression induce sarcomatoid carcinoma development.

(A-B) Representative images of two tumor samples display the region of sarcomatoid carcinoma. Scale bar = 200 μm. Representative images of different lesions are shown, including sarcomatoid carcinoma (A1, B1), area of necrosis (A2), invasion into muscle (B2), and images of mitotic figures (A3, B3, see the arrows). (C) Representative H&E and IHC images from the R26hAR^L/wt:p16^L/L:PB-Cre4 mice were shown for staining with different antibodies as labeled in the right bottom corner. (D-E) Co-IF from the R26hAR^L/wt:p16^L/L:PB-Cre4 mice of secretory epithelial cell marker CK8 (green), mesenchymal cell marker vimentin (red), and DAPI (blue), upper panels; Co-IF of secretory epithelial cell marker E-cadherin (green) and mesenchymal cell marker vimentin (red) and DAPI (blue), lower panels. F. R26hAR^L/wt:p16^L/L: PB-Cre4 mouse tissues were stained and quantified for Ki67 for the above tissues were measured in total 1,000 epithelial cells. Representative data shows mean and standard deviation. ** p < 0.005 Scale bar = 25 μm and magnifications were at 12.5 μm (A1-E4).

Fig 5. Histological and Immunochemical identification of signet ring cell carcinoma.

(A-B) Representative image of the prostate signet ring cell carcinomas, SRCCs, in R26hAR^L/wt:p16^L/L:PB-Cre4 mice. Scale bar = 200 μm. A1-B2. Pathological characterization of typical morphological characteristics of SRCC was highlighted in two different tumor samples with magnified views (A1’- B2’). (C-F) Representative view is shown for PAS/PAS (C-D, magnified view C’-D’) staining, Alcian Blue staining (E, magnified view E’), and vimentin immunoreactivity (F, magnified view F’). Scale bar used was at 50 μm and magnified views scale bar used was at 12.5 μm. (G). Representative view of IHC images of adjacent prostate tissue sections of signet ring cell carcinoma lesion from the R26hAR^L/wt:p16^L/L:PB-Cre4 mice are shown for staining with different antibodies (right bottom corner). Scale bar used was at 12.5 μm. (H-J) Representative images of H&E of metastatic lesions in the lung (H), liver (I), and Lymph node (J) with magnified views (H’-J’). (K) Pathological abnormalities of the different aged mice were listed for R26hAR^L/wt:p16^L/L: PB-Cre4 mice. Scale = 50 μm, or 12.5 μm (C-F, H-J).

Fig 6. Co-expression of AR transgene and loss of p16^Ink4a expression induces an EMT signature.

(A) Heatmap showing a subset of differentially expressed genes (DEGs) that corresponded to an overlap of hallmark EMT genes comparing p16^L/L: PB-Cre4 and R26hAR^L/wt:p16^L/L: PB-Cre4 mice. (B) qRT-PCR confirmation of the gene alterations in prostate tissues of p16^L/L: PB-Cre4 and R26hAR^L/wt:p16^L/L: PB-Cre4. The data is presented as the mean ± (SD) (n = 3) ** p < 0.005 and * p < 0.01 by student’s T test. (C-D) Representative view of IHC images of adjacent prostate tissue sections of sarcomatoid carcinoma lesion from the R26hAR^L/wt:p16^L/L: PB-Cre4 mice (C1-C2) and p16^L/L: PB-Cre4 mice (D1-D2) and their magnified views (C1’-D2’). Scale bar = 50 μm or 12.5 μm.