Rb regulates fate choice and lineage commitment in vivo

Abstract

Mutation of the retinoblastoma gene (RB1) tumour suppressor occurs in one-third of all human tumours and is particularly associated with retinoblastoma and osteosarcoma. Numerous functions have been ascribed to the product of the human RB1 gene, the retinoblastoma protein (pRb). The best known is pRb's ability to promote cell-cycle exit through inhibition of the E2F transcription factors and the transcriptional repression of genes encoding cell-cycle regulators. In addition, pRb has been shown in vitro to regulate several transcription factors that are master differentiation inducers. Depending on the differentiation factor and cellular context, pRb can either suppress or promote their transcriptional activity. For example, pRb binds to Runx2 and potentiates its ability to promote osteogenic differentiation in vitro. In contrast, pRb acts with E2F to suppress peroxisome proliferator-activated receptor gamma subunit (PPAR-gamma), the master activator of adipogenesis. Because osteoblasts and adipocytes can both arise from mesenchymal stem cells, these observations suggest that pRb might play a role in the choice between these two fates. However, so far, there is no evidence for this in vivo. Here we use mouse models to address this hypothesis in mesenchymal tissue development and tumorigenesis. Our data show that Rb status plays a key role in establishing fate choice between bone and brown adipose tissue in vivo.

Figure 1. Rb cooperates with p53 and modulates mesenchymal tumor fate in a dosage-dependent manner

a, Mesenchymal tumor distribution (percentage of animals analyzed up to 24 months of age) for Prx1-Cre;Rb and/or p53 compound mutant animals. b, H+E staining of representative sarcomas (20× magnification). c, PCR genotyping to detect Rb wildtype (wt) and recombined conditional mutant (loxp) alleles in control Rb^fl/+;p53^fl/+ tissues (lane 1) or cell lines derived from Prx1-Cre;Rb^fl/fl;p53^fl/fl (DKO) or Prx1-Cre;Rb^fl/+;p53^fl/fl osteosarcomas. Cell lines were cultured for ≥20 passages prior to genotyping to eliminate stromal cell contribution.

Figure 2. Rb regulates osteosarcoma-cell lineage plasticity in vitro and in vivo

The differentiation potential of 3 different Osx-Cre;Rb^fl/fl:p53^fl/fl (DKO) and Osx-Cre:p53^fl/fl (p53KO) OS cell lines was assessed 0, 7, 14, or 21 days after addition of differentiation media. a, Representative staining for: (left lane) alkaline phosphatase prior to differentiation induction; (middle lane) Alizarin Red to detect bone mineralization 14 days after culture in osteogenic-induction media and (right lane) Oil-Red-O to detect lipid droplets 14 days after culture in adipogenic-induction media. Expression of bone (Runx2, Alp, Coll1a and Bsp) and fat (Ap2, Pparγ, C/ebpα and Pgc1α) markers was assessed by qPCR of un-induced DKO (orange) and p53KO (black) OS cells. Bars represent the mean of three independent experiments (+/− SD). NS = not significantly expressed. b, Rb or control (Luc) shRNAs were expressed in the p53KO cell lines. Rb knockdown was confirmed by immunoprecipitation and qPCR showed that this caused downregulation of bone markers Bsp (and also Coll1a and Alp, data not shown), and upregulation of fat markers Pparγ (and also Ap2 and C/ebpα, data not shown) without culture in differentiating media. Bars represent the mean of three independent experiments (+/− SD). c, The osteogenic and adipogenic potential of shLuc- and shRb-p53KO cell lines was assessed 0, 7, 14 and 21 days after differentiation induction by Alizarin Red and Oil-Red-O staining. A representative timepoint (14 days) is shown. d, H+E staining of representative tumors derived from shLuc- and shRb-p53KO cell lines injected subcutaneously into immunocompromised mice. shRb-p53KO OS cells consistently (10/10 injections) yielded tumors that arose faster, and were more aggressive, than those arising from the parental p53KO OS controls (10 injections). Moreover, the shRb-p53KO OS derived tumors were frequently (6/10 injections) mixed lineage (top inset shows fat neoplasm; bottom inset bone/undifferentiated sarcoma), while the control shLuc-p53KO tumors were uniformly (10/10 injections) osteosarcomas. Additional analysis of these tumors (H+E, sirius red staining and Runx2 IHC) is shown in Supplementary Figure 4.

Figure 3. pRb modulates the activity and the expression of the master lineage regulators Runx2 and Pparγ

a, The stable DKO-Rb^Dox-ON OS cells were generated by drug selection of pools of DKO cells transfected with the doxocyline inducible construct pCW22-Rb. These were cultured for two days in the absence (Rb Off) or presence (Rb On) of doxocycline and then analyzed. Results are representative of three independent experiments. Promoter occupancy was assessed by chromatin immunoprecipitation. Sequence analysis identified two potential E2f binding sites (−278 and −160) within the Pparγ promoter. pRb induction caused a dramatic upregulation of both pRb and E2F4 binding to the proximal site. (No binding was observed at the distal element.) Similarly, pRb induction allowed pRb to bind to the known Runx2 response element of Col1α and also increased the binding of Runx2. These changes correlated with the downregulation of Pparγ mRNA and upregulation of Col1α mRNA as judged by qPCR. Bars represent the mean of three independent experiments (+/− SD). b, Western blotting detected Runx2 in pRb-immunoprecipitates from p53KO-OS cell lines (left, top panel). Western blotting of whole cell extracts confirmed that Runx2 was expressed in both DKO and p53KO OS cell lines (left, bottom panel). MSCs and osteoblasts were used as a positive control. Right panel: Runx2 transcriptional activity was shown to be higher in the p53KO- versus the DKO OS cell lines as judged by activation of the artificial Runx2-responsive reporter p6OSE2-Luc. Results are the average of six independent samples. Error bars indicate S.E.

Figure 4. Rb maintains the osteoblastic fate commitment in normal osteoblasts and regulates fate choice during normal development in vivo

a, Calvarial osteoblasts were prepared from e18.5 Rb^fl/fl or p53^fl/fl embryos and infected with Ad-GFP or Ad-Cre at P1. Five days later, the cells were induced with differentiation media and then assayed for osteogenesis and adipogenesis at 0, 14 and 25 days by staining with Alizarin Red and Oil-Red-O. A representative timepoint (25 days) is shown. b, qPCR was also used to assess osteogenic and adipogenic markers in the un-induced Rb^fl/fl (wt) versus Rb^fl/fl+Ad-Cre (Rb^−/−) osteoblasts. Bars represent the mean of three independent experiments (+/− SD). c, Alizarin Red (bone mineralization) and Alcian Blue (cartilage) staining of e15.5 skeletons (top panel), e18.5 calvaria (middle panel) and e18.5 limbs (botton panel) from Meox2-Cre;Rb^+/+ and Meox2-Cre;Rb^fl/fl littermate embryos. Arrows mark visible skeletal defects. qPCR was used to assess osteogenic (Runx2, Alp, and Bsp) and adipogenic (Ap2 and C/ebpα) markers in mRNA extracted from the calvarial bones of e18.5 Meox2-Cre;Rb^+/+ and Meox2-Cre;Rb^fl/fl embryos. Bars show the mean of three embryos arising in two independent crosses (+/− SD). d, Brown adipose tissue (BAT) was dissected from the backs of Meox2-Cre;Rb^fl/fl embryos (n=10) and their Meox2-Cre;Rb^+/+ littermate controls. All 10 showed a dramatic expansion of the brown fat compartment. A representative example is shown (upper two panels). Introduction of the LSL-LacZ reporter into this model, and LacZ staining confirmed equal, widespread expression of Cre in the control and Rb mutant BAT (third panel). H+E staining of BAT (bottom panel).