Glucocorticoid receptor regulates accurate chromosome segregation and is associated with malignancy

Abstract

The glucocorticoid receptor (GR) is a member of the nuclear receptor superfamily, which controls programs regulating cell proliferation, differentiation, and apoptosis. We have identified an unexpected role for GR in mitosis. We discovered that specifically modified GR species accumulate at the mitotic spindle during mitosis in a distribution that overlaps with Aurora kinases. We found that Aurora A was required to mediate mitosis-driven GR phosphorylation, but not recruitment of GR to the spindle. GR was necessary for mitotic progression, with increased time to complete mitosis, frequency of mitotic aberrations, and death in mitosis observed following GR knockdown. Complementation studies revealed an essential role for the GR ligand-binding domain, but no clear requirement for ligand binding in regulating chromosome segregation. The GR N-terminal domain, and specifically phosphosites S203 and S211, were not required. Reduced GR expression results in a cell cycle phenotype, with isolated cells from mouse and human subjects showing changes in chromosome content over prolonged passage. Furthermore, GR haploinsufficient mice have an increased incidence of tumor formation, and, strikingly, these tumors are further depleted for GR, implying additional GR loss as a consequence of cell transformation. We identified reduced GR expression in a panel of human liver, lung, prostate, colon, and breast cancers. We therefore reveal an unexpected role for the GR in promoting accurate chromosome segregation during mitosis, which is causally linked to tumorigenesis, making GR an authentic tumor suppressor gene.

Keywords: DNA damage; aneuploidy; cancer; glucocorticoid receptor; mitosis.

Fig. 1.

GR enriches to the mitotic spindle. (A) GR interacting proteins were compared with a published spindle proteome that identifies interaction of GR with a number of components of the mitotic machinery. (B) Mitotic spindles were purified and immunoblotted for α-tubulin, GR, ^PS203GR, and ^PS211GR. (C) HeLa cells were permeabilized and then fixed and immunolabeled for α-tubulin and GR. Metaphase and anaphase cells are shown. (D) HeLa cells were transfected with either Halo or Halo-GR. Twenty-four hours later, cells were treated with vehicle (Veh) or nocodazole (Noc) for 16 h or with 100 nM dexamethasone (Dex) for 1 h and then lysed and immunoblotted for GR, Halo, or α-tubulin. (E) HeLa cells were transfected with Halo-GR, permeabilized, and then fixed and immunolabeled for α-tubulin and Halo. DNA was counterstained with Hoechst. (Scale bars: 10 μm.)

Fig. 2.

Reduction of GR expression delays anaphase transition and increases the frequency of chromosomal abnormalities. HeLa cells stably expressing H2B-GFP were transfected with 50 nM nontargeting (NT) or GR-specific (GR6) siRNA, and mitotic events were imaged in real time [values depict minutes from an arbitrary start point (t0)]. Representative fields are shown. (A) Images were scored for average time to completion of mitosis, frequency of abnormal mitotic events, and death in mitosis. (B) Ten fields were tracked per replicate. Graphs show mean ± SEM from four independent duplicate experiments. Movies S1–S6 are provided as supplementary information. NEB, nuclear envelope breakdown. (C–E) HeLa cells were treated with 50 nM NT or two different GR-specific (GR5 and GR6) siRNAs, together with siRNA-resistant GR plasmid (pcDNA3GR) or empty vector control (pcDNA3) for 48 h, and then fixed and labeled with an α-tubulin–specific antibody and Hoechst DNA dye counterstain. Cells were imaged and scored for mitotic phase (C) and for evidence of mitotic abnormalities, including spindle deformities (D) and interphase micronuclei (E, yellow arrowhead). Graphs depict mean ± SEM of at least three independent triplicate experiments. Data were analyzed by the Mann–Whitney test or ANOVA, where *P < 0.05. (Scale bars: 10 μm.)

Fig. 3.

GR regulation of chromosome segregation requires the GR ligand-binding domain. (A) Schematic showing GR deletant and mutant constructs. Stable HeLa Flp-In TRex-GR cells were transfected with NT or GR-specific siRNA (GR11) and then treated with 5 μM doxycycline (Dox). Forty-eight hours later, cells were lysed and immunoblotted for GR (B) or fixed and nuclei counterstained with Hoechst (C). (D) Aberrant mitoses were scored and quantified. The graph depicts mean ± SEM of three independent triplicate experiments. (Inset) Image shows expression of ΔLBDGR. Data were analyzed by ANOVA, where **P < 0.01. DBD, DNA-binding domain; FLGR, full-length GR; LBD, ligand-binding domain; NTD, N-terminal domain. (Scale bars: 10 μm.)

Fig. 4.

GR haploinsufficiency increases DNA damage and tumor formation in vivo. (A) Low-passage (p4) GR^+/+ and GR^−/− mouse embryonic fibroblasts (MEFs) were analyzed for DNA content by fluorescence-activated cell sorting (FACS) and also metaphase spread assays. High-passage (p20) GR^+/+ and GR^−/− MEFs were analyzed for DNA content by FACS. (B) Livers from 12-wk-old GR^+/+ and GR^+/− mice infused with BRdU were stained for BRdU or γ-H2A.X as an indicator of DNA damage. BRdU or H2A.X staining is shown in brown, and nuclei are counterstained with toluidine blue. A minimum of 1,000 cells for each section were scored from three to four different animals. (C) Liver and colon from 100-wk-old GR^+/+ and GR^+/− mice were analyzed for γ-H2A.X as an indicator of DNA damage. A minimum of 10,000 cells were scored for each section from two to four different animals. (D) Immunohistochemistry of serial liver sections from a GR^+/− mouse labeled for Ki-67 and GR. (Insets) Higher magnification images are shown. Liver tumors from multiple animals are quantified. (E) Immunohistochemistry of serial lung sections from a GR^+/− mouse labeled for Ki-67 and GR. Lung tumors from multiple animals are quantified. The graph shows mean ± SEM values. (Magnification: B, 10×; C, 40×; D and E, 2.5×.) Data were analyzed by the Mann–Whitney test, where *P < 0.05 and **P < 0.01.

Fig. 5.

GR is down-regulated in human cancers. B lymphocytes from a family with GR haploinsufficiency (A) were analyzed for chromosome complement by metaphase spread assay (B). Chromosome counts show counts for 150 cells from three independent passages (p5–p8). (C) Microarrays of tumor and matched normal tissues from 1,314 patients were analyzed for GR transcript. (D) GR transcript in liver, lung, and prostate is shown. GR transcript and protein are shown for colon (E and G) and breast (F and H) cancers. (Magnification: 20×.) Graphs depict median and interquartile range. n, number of patient samples within each group. Data were analyzed by ANOVA, followed by Tukey’s correction or the Mann–Whitney test, where *P < 0.05 and **P < 0.001.