Regulation of RB transcription in vivo by RB family members

Abstract

In cancer cells, the retinoblastoma tumor suppressor RB is directly inactivated by mutation in the RB gene or functionally inhibited by abnormal activation of cyclin-dependent kinase activity. While variations in RB levels may also provide an important means of controlling RB function in both normal and cancer cells, little is known about the mechanisms regulating RB transcription. Here we show that members of the RB and E2F families bind directly to the RB promoter. To investigate how the RB/E2F pathway may regulate Rb transcription, we generated reporter mice carrying an eGFP transgene inserted into a bacterial artificial chromosome containing most of the Rb gene. Expression of eGFP largely parallels that of Rb in transgenic embryos and adult mice. Using these reporter mice and mutant alleles for Rb, p107, and p130, we found that RB family members modulate Rb transcription in specific cell populations in vivo and in culture. Interestingly, while Rb is a target of the RB/E2F pathway in mouse and human cells, Rb expression does not strictly correlate with the cell cycle status of these cells. These experiments identify novel regulatory feedback mechanisms within the RB pathway in mammalian cells.

FIG. 1.

Direct binding of RB family members to the RB promoter. (a) Conservation of binding sites for the SP1, ATF, and E2F (bold) transcription factors in the proximal RB promoter. (b) Quantitative ChIP analysis of p107, p130, and E2F4 on the Rb promoter in NIH3T3 fibroblasts (n = 2). Enrichment is calculated over an unrelated DNA sequence (Actin) and normalized to the binding to a negative-control antibody (p16). (c) Representative ChIP analysis of GFP-RB, GFP-p107, and GFP-p130 fusion proteins on the RB promoter in Saos-2 cells. Untransfected (UT) cells and cells transfected with GFP serve as negative controls. ChIP was performed using GFP antibodies as well as antibodies for Dbf4 as negative controls. Input and immunoprecipitated DNA were amplified by PCR with primers specific for the human RB promoter and one non-E2F target gene (Albumin).

FIG. 2.

Regulation of RB transcription is not cell cycle dependent. (a) RT-qPCR of mouse Rb, p107, and Cdc6 mRNA in asynchronously cycling immortalized MEFs. G₁- and S-phase cells were sorted by DNA content through Hoechst staining and compared to samples rendered quiescent through 3 days of serum starvation (G₀) (n ≥ 2). (b) RT-qPCR of human RB and CDC6 in quiescent (0 h) T98G cells and cells synchronized in G₁/S phase (15 h). Hours are times poststimulation with 20% serum (n = 2). (c and d) Quantitative ChIP of E2F family binding (c) and ATF, Sp1, and RB family binding (d) to the RB (left panels) and CDC6 (right panels) promoters in quiescent (0 h) and G₁/S-phase (15 h) T98G cells. Antibodies used for ChIP are indicated (n = 3). (e) Wild-type (upper panel) and Rb null (lower panel) MEFs were cotransfected with luciferase constructs for the Rb and p107 promoters and with an empty expression vector (−) or a vector expressing E2F3 (E2F3) (n = 3). The Rb^E2FMut construct represents the Rb promoter carrying a mutation in the E2F site. The break in the x axis indicates that the luciferase activities were normalized to each wild-type construct cotransfected with the empty vector.

FIG. 3.

Generation of Rb-eGFP transgenic mice. (a) The RP24-370G12 mouse BAC is located 73.7 Mb from the origin of chromosome 14. This BAC encompasses more than 50 kb upstream of Rb and more than 100 kb downstream of Rb, but it does not contain the final coding exons. Two other cellular genes (Lpar5 and Itm2b) are contained within the BAC sequences, while a third (Rcbtb2) is outside of the BAC. (b) Schematic representation of the Rb BAC knock-in reporter. The eGFP cDNA is inserted at the ATG in exon 1 of Rb. The eGFP cassette contains two polyadenylation (polyA) sequences. Boxes indicate Rb exons 1 and 2, and B indicates BamHI. The location of the probe used in Southern blot experiments is indicated below exon 1. (c) Southern blot analysis of Rb 5′ regulatory sequences in the gDNA of Rb-eGFP BAC transgenic mice. BamH1 (B)-digested BAC DNA extracted from bacteria serves as a size control. gDNA from the two founder lines (G1 and G2) was analyzed. (d) Analysis by RT-qPCR of Rb and eGFP expression in organs collected from E15.5 Rb-eGFP transgenic embryos. mRNA expression for each gene was calculated relative to levels in the lung (n = 2). Lu, lung; In, intestine; Th, thymus; Br, brain; Sp, spleen; Li, liver.

FIG. 4.

eGFP expression in Rb-eGFP transgenic mice largely mimics Rb expression in early- and mid-gestation embryos. (a) Direct visualization of eGFP in an Rb-eGFP embryo at E9.5. NT, neural tube; LV, lateral ventricle; S, somites. (b) Immunofluorescence analysis of eGFP expression in E13.5 transgenic embryos. He, heart; Li, liver; SC, spinal cord; M, muscle; DRG, dorsal root ganglia. DAPI marks the DNA in blue. (c) Immunofluorescence analysis of eGFP in Rb-eGFP embryos (Rb eGFP⁺) in different organs at E13.5. Rb-eGFP⁻ embryos serve as negative controls. R, retina; L, lens; OE, olfactory epithelium; Ri, ribs; Dia, diaphragm muscle. Magnification, ×400.

FIG. 5.

eGFP expression in Rb-eGFP transgenic mice largely mimics Rb expression in late embryos. (a) Direct visualization of eGFP in the head region of an Rb-eGFP embryo at E17.5. Expression in the brain and in vibrissae is visible. (b to d) Immunofluorescence analyses of eGFP expression in sections of the vibrissae (b), the thymus (c), and the tongue (d) of a representative E17.5 transgenic embryo. Each time, a section from a control embryo (Rb eGFP⁻) is shown. Bar, 100 μm.

FIG. 6.

eGFP expression in Rb-eGFP transgenic mice largely mimics Rb expression in adult mice. (a) Analysis by RT-qPCR of Rb and eGFP expression in organs collected from Rb-eGFP transgenic mice (n ≥ 2). Data from line G1 are shown; similar data were obtained for line G2 (not shown). mRNA expression for each gene was calculated relative to that for the liver. Li, liver; Lu, lung; Spl, spleen; Mu, muscle; He, heart; Thy, thymus; Ki, kidney. (b) FACS analysis (left panel) of eGFP expression and RT-qPCR analysis of Rb mRNA expression (right panel) in B cells and T cells isolated from the spleens of wild-type Rb-eGFP transgenic mice. (c) Confocal immunofluorescence analysis of eGFP expression (green signal) and for the Müller glial cell marker α-glutamine synthetase (GS, red signal) in the retinas of Rb-eGFP transgenic (Rb-eGFP⁺) and control (Rb-eGFP⁻) mice. GCL, ganglion cell layer; INL, inner nuclear layer; ONL, outer nuclear layer. Bar, 100 μm. (d) Immunofluorescence analysis of eGFP expression in the cerebellum of an Rb-eGFP⁺ mouse. P, Purkinje cells. Bars, 100 μm.

FIG. 7.

RB positively regulates its own transcription in MEFs. (a) Schematic representation of the conditional Rb^lox allele. Blue arrows indicate locations of primers used to detect deletion of the Rb mRNA. Cre-mediated excision results in the genomic deletion of exon 3 and in the loss of Rb mRNA containing exon 3 sequences. (b) Analysis by RT-qPCR of Rb deletion (as for panel a) and eGFP and p107 mRNA expression in Rb^lox/lox; Rb-eGFP MEFs infected with Adeno-Cre (Cre) relative to those for cells infected with a control adenovirus (Ad) (n = 3). (c and d) Analysis by RT-qPCR of Rb, eGFP, and Cdc6 mRNA in Rb^lox/lox; Rb-eGFP MEFs infected with Adeno-Cre or a control adenovirus rendered quiescent through 3 days of serum starvation and then stimulated into synchronized cell cycle reentry through the addition of a 20% concentration of serum. Hours indicate time elapsed since the addition of serum (n = 3).

FIG. 8.

RB negatively regulates its own transcription in vivo. (a and b) Analysis by quantitative PCR of deletion of Rb in whole organs from adult Rb^lox/lox; Rosa26^CreERT2 mice (Rb^Δ/Δ) and Rb^lox/lox mice treated with tamoxifen, calculated relative to that for each control organ. Li, liver; Lu, lung; Spl, spleen. (a) qPCR on genomic DNA (n = 2) (b) RT-qPCR on mRNA (n = 3) (c) Analysis by RT-qPCR of eGFP in organs from Rb-eGFP; Rb^lox/lox; Rosa26^CreERT2 mice (Rb^Δ/Δ) and Rb-eGFP; Rb^lox/lox (Rb^lox/lox) mice treated with tamoxifen (n = 4). (d) RT-qPCR analysis of cyclin E (CycE), Cdc6, and p107 in organs from Rb^lox/lox and Rb^lox/lox; Rosa26^CreERT2 mice treated with tamoxifen (Rb^Δ/Δ) (e) Upper panels: representative (n = 3) FACS analysis for eGFP fluorescence in splenic Mac1⁺/Gr1⁺ myeloid cells, B220⁺ B cells, and CD3⁺ T cells from control (eGFP⁻, shaded area), Rb-eGFP; Rb^lox/lox, and Rb-eGFP; Rb^lox/lox; Rosa26^CreERT2 mice, all treated with tamoxifen. Lower panels: analysis by RT-qPCR of cyclin E, Cdc6, and p107 in sorted cells from the spleens of control and Rb mutant mice (n ≥ 2).

FIG. 9.

p107 and p130 control of Rb transcription in vivo. (a) Analysis by RT-qPCR of Rb mRNA expression in whole organs from wild-type and p107-deficient adult mice relative to the expression of Rb in the wild-type liver (n = 3). (b) Schematic representation of the conditional p130^lox allele. Blue arrows indicate locations of primers used to detect deletion of the p130 mRNA. Cre-mediated excision results in the genomic deletion of exon 2 and in the loss of p130 mRNA containing exon 2 sequences. (c) Analysis by RT-qPCR of p130 in the liver, lung, spleen, and thymus of p130^lox/lox; Rosa26^CreERT2 mice and control mice treated with tamoxifen. (d) Analysis by RT-qPCR of Rb in the same samples as for panel c. (e) Representative (n = 3) FACS analysis for eGFP fluorescence in Mac1⁺/Gr1⁺ splenic cells from control (eGFP⁻, shaded area), Rb-eGFP, and p107^−/−; Rb-eGFP mice. (f) Representative (n = 2) FACS analysis for eGFP fluorescence in Mac1⁺/Gr1⁺ splenic cells from control (eGFP⁻, shaded area), Rb-eGFP, and p130^lox/lox; Rosa26^CreERT2; Rb-eGFP mice. All mice were treated with tamoxifen.

FIG. 10.

The RB family regulates the Rb promoter in vivo. (a) Analysis by RT-qPCR of p130 and eGFP in conditional Rb^lox/lox; p130^lox/lox; p107^−/− (TKO) MEFs 4 days after infection with Adeno-Cre (Cre) or a control adenovirus (Ad) (n = 4). (b) Cell cycle profiles of conditional TKO MEFs infected with Adeno-Cre (Cre) or a control adenovirus (Ad) as for panel a. Percentages were calculated from the results of BrdU/PI FACS analysis (n = 2). (c) Analysis by RT-qPCR of Rb and p130 in the livers (Li), lungs (Lu), and spleens (Spl) of conditional Rosa26^CreERT2 TKO mice (TKO) and control mice (CTRL), all treated with tamoxifen (n ≥ 3). (d) Analysis by RT-qPCR of eGFP in samples similar to those for panel c from Rb-eGFP transgenic mice (n ≥ 4). (e) Analysis by RT-qPCR of CycE and Cdc6 in samples similar to those for panel c (n ≥ 2). (f) Representative FACS analysis of eGFP expression in control and TKO splenocytes. (g) Representative FACS analysis of eGFP expression in splenic Mac1⁺/Gr1⁺ myeloid cells, B220⁺ B cells, and CD3⁺ T cells sorted from control (eGFP⁻, shaded area), Rb-eGFP⁺, and TKO; Rb-eGFP⁺ mice. Green and blue numbers represent the percentages of BrdU⁺ cells in control (green) and TKO (blue) populations (n = 3). (h) Analysis by RT-qPCR of p130 in cells isolated as for panel g from control mice and TKO mice. (i) Analysis by RT-qPCR of CycE and Cdc6 from control and TKO mice.

FIG. 11.

The RB pathway regulates RB expression. (a to c) Analysis of MCF7 cells treated with vehicle (DMSO) or the CDK4 inhibitor PD 033291 (PD). (a) Quantification of BrdU incorporation (n = 3). (b) RT-qPCR for RB expression relative to that of GAPDH (n = 3). (c) Western blot analysis of RB and lamin B on protein extracts; bracket indicates hypo- and hyperphosphorylated species of RB.