WWOX and p53 Dysregulation Synergize to Drive the Development of Osteosarcoma

Abstract

Osteosarcoma is a highly metastatic form of bone cancer in adolescents and young adults that is resistant to existing treatments. Development of an effective therapy has been hindered by very limited understanding of the mechanisms of osteosarcomagenesis. Here, we used genetically engineered mice to investigate the effects of deleting the tumor suppressor Wwox selectively in either osteoblast progenitors or mature osteoblasts. Mice with conditional deletion of Wwox in preosteoblasts (Wwox^Δosx1) displayed a severe inhibition of osteogenesis accompanied by p53 upregulation, effects that were not observed in mice lacking Wwox in mature osteoblasts. Deletion of p53 in Wwox^Δosx1 mice rescued the osteogenic defect. In addition, the Wwox;p53^Δosx1 double knockout mice developed poorly differentiated osteosarcomas that resemble human osteosarcoma in histology, location, metastatic behavior, and gene expression. Strikingly, the development of osteosarcomas in these mice was greatly accelerated compared with mice lacking p53 only. In contrast, combined WWOX and p53 inactivation in mature osteoblasts did not accelerate osteosarcomagenesis compared with p53 inactivation alone. These findings provide evidence that a WWOX-p53 network regulates normal bone formation and that disruption of this network in osteoprogenitors results in accelerated osteosarcoma. The Wwox;p53^Δosx1 double knockout establishes a new osteosarcoma model with significant advancement over existing models. Cancer Res; 76(20); 6107-17. ©2016 AACR.

Figure 1. Characterization of Wwox^ΔOB conditional knockout mice

A. Wwox expression in femurs of CTRL and Wwox^ΔOsx1 mice as assessed by qPCR on femur, calvaria and liver (left panel). Right panel shows immunoblot analysis of calvaria isolated from these mice. B. µCT three-dimensional images of control and Wwox^ΔOsx1 mice at 1-and 3-months of age. Representative image of femur metaphysis region at 1-month shows less trabecular bone in Wwox^Osx1 mice. Quantitation of all µCT is shown in Fig S1A. C. Alkaline Phosphatase activity detects mature osteoblasts (left panel) and Alizarin Red staining (right panel) on Wwox^ΔOsx1 and CTRL calvarial osteoblasts after 21 days in osteogenic media. D. Depletion of Wwox in bone, and no effect in liver levels in Wwox^ΔOc mice. E. Significant decrease in trabecular number and connectivity density and a decrease in trabecular spacing at 3-months of Wwox^ΔOc mice. Complete analysis is presented in Fig S1. F. Inhibition of osteoblast differentiation in the Wwox^ΔOc mouse model.

Figure 2. Defects in differentiation of Wwox deficient calvarial osteoblasts are partially rescued by p53 deletion

A. qPCR analysis on osteoblasts from Wwox^ΔOsx1 calvaria showing an upregulation of cell proliferation and apoptosis regulators Trp53, Cdkn1a (p21) Bax and Puma mRNA levels during differentiation (days 0, 14, 21). B. Relative quantification of p53 and its target genes in control and Wwox^ΔOB calvarial osteoblasts at 21 days post-differentiation (left and right panel) and in 1 month-old Wwox^ΔOsx1 femur bones (middle panel). C. qPCR analysis of Osx1 and Oc levels in control, Wwox^ΔOsx1 and Wwox/p53^ΔOsx1 (DKO^ΔOsx1) calvariae at 21 days of differentiation. D,EWwox^ΔOsx1 phenotype rescue by p53 deletion - Alizarin Red Staining of cells in C at day 21 and ALP activity in DKO cells are comparable to CTRL osteoblasts. F,G. Bone marrow derived p53^ΔOsx1 mesenchymal stem cells (BMSCs) sustains, while DKO depletion inhibits osteogenesis. BMSCs were isolated from bone marrow of p53^ΔOsx1 and DKO^ΔOsx1 (F) and p53^ΔOc and DKO^ΔOc (G) mice and induced to differentiate in osteogenic media. qPCR analysis of RNA extracted after 7 and 21 days of differentiation for osteoblast markers. H. BMSCs from DKO^ΔOsx1 mice show a reduced ability to form mineralized matrix than p53^ΔOsx1 cells as assessed by Alizarin Red staining at 21 days from differentiation. I. DKO^ΔOc BMSCs induced to differentiate show stronger osteogenic abilities than p53^ΔOc cells.

Figure 3. Osx1-Cre dependent deletion of Wwox results in the acceleration of osteosarcoma formation

A. Kaplan-Meier survival plots for the indicated genotypes. B. Histological sections of the tumors derived from DKO^ΔOsx1 mice. Areas of pale staining represent bone tissues. C. Kaplan-Meier survival plots for DKO^ΔOc and p53^ΔOc mice. D. H&E staining of DKO^ΔOc and p53^ΔOc tumors. E. qPCR on DKO^ΔOsx1 (n=7) and p53^ΔOsx1 (n= 5) tumors. Osteogenesis makers were downregulated in DKO^ΔOsx1 tumors when compared to p53^ΔOsx1 tumors. F. Sirius-Red staining shows low collagen content in DKO^ΔOsx1 tumors relative to p53^ΔOsx1 tumors. G–I. Primary cell lines from DKO^ΔOsx1 tumors show an inability to undergo osteogenic differentiation as assessed by qPCR (G) for markers of mature osteoblasts: ALP activity (H) and Alizarin Red staining (I) at day 0 and after 21 days of differentiation.

Figure 4. Transcriptomic analysis of p53^ΔOsx1 and DKO^ΔOsx1 tumors

A. Heat map generated from RNA-sequencing analysis shows the genes that are commonly altered in both DKO^ΔOsx1 tumors and human OS samples compared to control bones. B. Validation by qPCR of several downregulated (left panel) and upregulated genes (right panel) in DKO tumors and hOS. C. Upregulation of RUNX2 target genes in DKO^ΔOsx1 vs p53^ΔOsx1 tumors (left) and tumor cell lines (right) as assessed by qPCR.

Figure 5. DKO^ΔOsx1 tumors cells are more resistant to chemotherapy and display an impaired DDR

A. After 48h of treatment with increased concentration of doxorubicin and cisplatinum, DKO tumor cells present higher resistance than p53^ΔOsx1 cells to the drugs as assessed by XTT assay. B. Immunofluorescence on DKO and p53^ΔOsx1 tumor cell lines using anti γ-H2AX antibody and DAPI at different times after NCS treatment. C. Quantification of γ-H2AX staining from B. D. Comet assay. Control- (MC3T3 EV) and WWOX-depleted MC3T3 (MC3T3-shWWOX) cells were irradiated or not at 10Gy. Labeled DNA was visualized under a fluorescence microscope using 60× magnification. Representative images are shown. E. Quantification of the comet assay in D. Bars show the comet tail as measured using ImageJ 1.47g software.

Figure 6. Loss of WWOX associates with inactivation of p53 and impaired DDR in human OS

A. Representative pictures of WWOX and p53 immunostaining in p53 wt or mutated p53 hOS. Scale bars: 20µm. Arrows and arrowheads depict cytoplasmic and nuclear WWOX immunopositivity, respectively. B. Quantitative representation of WWOX expression related to p53 status in hOS cases. C. Statistically significant comparison of WWOX immunohistochemical scoring values between p53 wt (average score=4,7) and p53 mut samples (average score=1,75). *p<0.05 D. Analysis of DDR markers, γ-H2AX and p-Chk2, in relation to WWOX expression in human OSs showing a clear trend of impaired DDR in WWOX negative cases and vice versa.