NF-kappaB-YY1-miR-29 regulatory circuitry in skeletal myogenesis and rhabdomyosarcoma

Abstract

Studies support the importance of microRNAs in physiological and pathological processes. Here we describe the regulation and function of miR-29 in myogenesis and rhabdomyosarcoma (RMS). Results demonstrate that in myoblasts, miR-29 is repressed by NF-kappaB acting through YY1 and the Polycomb group. During myogenesis, NF-kappaB and YY1 downregulation causes derepression of miR-29, which in turn accelerates differentiation by targeting its repressor YY1. However, in RMS cells and primary tumors that possess impaired differentiation, miR-29 is epigenetically silenced by an activated NF-kappaB-YY1 pathway. Reconstitution of miR-29 in RMS in mice inhibits tumor growth and stimulates differentiation, suggesting that miR-29 acts as a tumor suppressor through its promyogenic function. Together, these results identify a NF-kappaB-YY1-miR-29 regulatory circuit whose disruption may contribute to RMS.

Figure 1. YY1 represses miR-29b/c through binding to a conserved regulatory region

(A) An rVISTA schematic showing the degree of sequence conservation between human and mouse chromosome (Chr) 1 in a region upstream of the miR-29b/c cluster. Predicted YY1, MyoD, myogenin, SRF, and Mef2 sites are displayed. (B) EMSA performed from C2C12 MB or MT with probes corresponding to YY1 sites A-D. With MB extracts, a supershift EMSA was performed using YY1 antisera. Arrows denote YY1/DNA bound complexes. (C) ChIP assays for YY1 was performed with chromatin from C2C12 MB or MT. Precipitated DNA was amplified with oligonucleotides spanning regions A-D. Total inputs are indicated. (D) ChIPs as in (C) were repeated for Ezh2, H3K27, HDAC-1, SRF or Mef2. (E) MB were transfected with either an YY1 expression plasmid (pCMV-YY1) or YY1 siRNA oligos and then induced to differentiate for 24h, at which time miR-29b and miR-29c were measured by qRT-PCR and normalized to U6. Fold changes are shown with respect to control siRNA transfected cells where miR-29 levels were set to a value of 1. (F) MB were transfected with a miR-29b/c-enhancer-Luc reporter and maintained as MB or differentiated in MT for 48h, at which time luciferase activities were determined. (G) MB were transfected with a miR-29b/c-enhancer-Luc reporter, along with plasmids for YY1, SRF or Mef2. Cells were then switched differentiation conditions and luciferase activity was subsequently measured. (H) MB were transfected with the miR-29b/cenhancer- Luc reporter (YY1 wt) or with an enhancer reporter containing a deletion mutation in the YY1 “D” site (YY1 mut), and subsequently differentiated for 48h (MT) at which time luciferase activity was determined (left graph). Separate transfections were performed with YY1 wt and YY1 mut reporters along with an YY1 expression plasmid (pCMV-YY1) or YY1 siRNA. Cells were subsequently differentiated for 48h at which time luciferase was determined. Vector siRNA oligo transfected cells were used as a control (right graph). All luciferase data were normalized to β-galactosidase protein and represent the average of three independent experiments ± S.D.

Figure 2. NF-κB negatively regulates miR-29b/c

(A) C2C12 cells were treated with TNFα and miR-29 was measured by qRT-PCR normalized to U6. Fold changes are shown with respect to vector cells where miR-29 levels were set to a value of 1. (B) MB were transfected with vector or a p65 plasmid and miR-29b/c levels were measured 48h post-transfection. (C) MB were transfected with either vector or p65 siRNA oligos and miR-29 expression was then measured as in (B). (D) MiR-29b/c were measured in MB stably expressing vector or IκBα-SR. Fold changes are shown with respect to vector cells, which were set to a value of 1. (E) MyoD was stably expressed in p65^+/+ or p65^−/− mouse embryonic fibroblasts (Bakkar et al., 2008) and qRT-PCR was performed for miR-29b and miR-29c. (F) ChIPs with YY1 or control IgG were performed on chromatins derived from either vector control (V) or Iκ Bα-SR (SR) expressing MB. Primers specific to site D were used for the PCR amplification. Total inputs are indicated.

Figure 3. miR-29 is induced during muscle differentiation in vitro and in vivo

(A) C2C12 MB were induced to differentiate up to 4 days in differentiation medium (DM) and at indicated times miR-29a, miR-29b, and miR-29c were measured by qRT-PCR. (B) RT-PCR analysis of myogenic markers MyoD, myogenin, MyHC, skeletal actin (α-actin), troponin T performed at similar times to those in (A). (C) Westerns probing for YY1 and nuclear p65 in differentiating C2C12 cells. (D) Expression of miR-29b, and miR-29c in primary human myoblasts (GM) and myotubes (DM). (E) Measurement of miR-29 from either lower limb muscles at P2 and P8 or from TA muscles at P15, P23, and P90 in C57/BL6 mice. (F) Same measurement as in (E) from CTX injected TA muscles.

Figure 4. miR-29 accelerates muscle differentiation

(A) C2C12 cells were transfected with 0.2µg of MyHC-Luc or Troponin-Luc reporters along with pCMV-LacZ and 50µM of precursor control oligos (Pre-NC) or miR-29c (Pre-29c) oligos. Cells were differentiated for 48h and luciferase was determined and normalized to β-Galactosidase. Relative activity is shown with respect to control cells where normalized luciferase values were set to 1. The data represents the average of three independent experiments ± S.D. (B) C2C12 MB were transfected with Pre-NC or Pre-29c oligos. Cells were then maintained as MB or differentiated into MT. Lysates were prepared and probed for MyHC and troponin T. (C) MB were transfected with Pre-NC or Pre-29a, b, or c members and differentiated (DM) for 1 or 2 days (d), at which time cells were immunostained for MyHC. Scale bar = 100µm. Cell morphology was visualized by phase-contrast microscopy; scale bar = 200µm. (D) C2C12 cells transfected with a MyHC-Luc along with anti-miR control (Anti-NC) or anti-miR-29c (Anti-29c). Cells were then maintained as MB or differentiated into MT for 48hrs at which time luciferase activity was determined. (E) MB were administered Anti-NC or Anti-29c and then MyHC and troponin were probed in cells maintained as MB or differentiated into MT. (F) MB were transfected with Anti-NC or Anti-29c oligos and cells were subsequently differentiated for 3 days at which time cultures were stained for MyHC. Positively stained cells were counted from a minimum of 10 randomly chosen fields from 3 individual plates.

Figure 5. miR-29 suppresses YY1 through binding to its 3’UTR

(A) Predicted target site of miR-29c (green) in the 3’UTR of mouse YY1 (red) with the seed region underlined. (B) Predicted folding structure from mFOLD between miR-29c (green) and YY1 3’UTR (red). The minimal free energy (mfe) is indicated as well as the seed region shown by a line and arrow. (C) A wild type (WT) luciferase reporter was generated by fusing a ~ 500bp fragment of the YY1 3’UTR encompassing the miR-29 binding site downstream of the luciferase (Luc) reporter gene. The mutant plasmid was generated by deleting the miR-29 binding site. WT or Mutant reporter constructs were then transfected into MB with indicated precursor miRNA oligos. Luciferase was determined at 48h post-transfection and normalized to β-Galactosidase. Data represent the average of three independent experiments ± S.D. (D) MB were transfected with either precursor (Pre-NC) or Anti-miR control (Anti-NC) or precursor miR-29c (Pre-29) or anti-miR-29c (Anti-29) oligos. YY1 protein was then probed in extracts from cells differentiated for 48h. Blots were stripped and reprobed for α-tubulin. (E) P5 neonatal mice were injected with anti-miR control (Anti-NC) or anti-miR-29c (Anti-29) oligos into lower limb muscles. 48h post-injection, lysates were probed for YY1. (F) MB were transfected with siRNA control or siRNA to YY1. Cells where then differentiated for 24h, at which time they were photographed under phase contrast (scale bar = 200µm) or immunostained for MyHC (scale bar = 100µm). Numbers indicate averages of MyHC positive cells counted from a minimum of 10 randomly chosen fields.

Figure 6. miR-29 functions as a tumor suppressor in RMS

(A) Normal human skeletal muscle cells were cultured along with ARMS and ERMS cell lines and qRT-PCR was performed to measure miR-29b. (B) Total RNAs were obtained from ten RMS patient tumors (numbers represent patient identification) and expression of miR-29b was measured by qRT-PCR. Total RNAs from human skeletal muscle were used as the control. (C) RH30 cells were transduced with vector or miR-29b expressing lentiviruses, and stable cell lines were generated. The cell number was determined by Trypan blue staining over the course of ten days. (D) Total RNAs were extracted from vector or miR-29b expressing RH30 cells and semi-quantitative RT-PCR was performed probing for p21^CIP/WAF1 and cyclinD1. GAPDH was used as a control. (E) Same analysis as in (D) was done for differentiation markers, MyHC, α-actin and troponin T. (F) Phase contrast images of control and miR-29 expressing RH30 cells. Arrows indicate myotubelike structures. Cells were treated with differentiation medium for 2 days and immunostained for troponin. Scale bars = 200µm. (G) Vector control or miR-29 expressing RH30 cells were differentiated (DM) for 0, 1, 2, 3, or 4 days, and subsequently immunostained for troponin. The graph represents average number of troponin positive cells that were counted from a minimum of 10 randomly chosen fields from 3 culture plates.

Figure 7. miR-29 inhibits RMS tumor growth in vivo

(A) RH30 tumors were established in mice and then subjected injected with vector or miR-29b expressing lentiviruses. Tumor volume was recorded daily for 21 days. (B) Tumors were sectioned for stained for histology with H&E or with proliferation markers Ki67 and phospho-Histone H3 (P-H3). Apoptotic cells were stained by a standard TUNEL assay. (C) During the course of miRNA treatment, 3 tumors from vector control and miR-29b injected groups were resected. Total RNAs were prepared and semi-quantitative RT-PCR was performed probing for differentiation markers. (D) Ki-67, P-H3 and apoptotic cells were quantitated by counting positively stained cells from 10 randomly chosen fields from six sections per tumor. *p<0.05. (E) RH30 xenograft tumors were established as in (A) and then injected with precursor control miR (Pre-NC) or miR-29b (Pre-29b) oligos. Tumor volumes were recorded for 21 days.

Figure 8. NF-κBY–Y1–miR-29 circuitry is dysregulated in RMS

(A) Extracts were prepared from normal human muscle and RMS cell lines and immunoblots were performed probing for YY1, Ezh2, and p65. (B) Lysates were prepared five RMS patient tumors and adjacent normal muscle tissue and probed for YY1, Ezh2, and p65 proteins. (C) RH30 cells were transfected with siRNA control oligos or siRNA-YY1. Cells were differentiated and semi-quantitative RT-PCR was performed probing for differentiation markers, or GAPDH used as a control. (D) RH30 tumors were established in nude mice and then injected with siRNA oligos every 3 days for 1 week. RNA and protein lysates were prepared from tumors and subsequently probed for MyHC, α-actin and troponin by RT-PCR and YY1 by Western. (E) ChIPs with either an YY1 antibody or control IgG were performed on chromatins isolated from human skeletal muscle cells (control) or RH30 cells. Precipitated DNA fragments were amplified with oligonucleotides spanning regions A-D of the human miR-29b/c regulatory region. Total inputs are indicated. (F) RH30 cells were infected with adenoviruses expressing vector control or IκBα-SR. YY1 and miR-29b levels were measured at 48h post-infection by qRT-PCR. (G) RH30 cells were infected with control or IκBá-SR adenovirus and differentiation markers were probed by qRT-PCR. (H) RH30 cells were transduced with vector or IκBα-SR expressing retroviruses to generate stable cell lines. Cells were then treated with differentiation medium for 2 days and immunostained for troponin or quantified for myofibrillar expression. (I) The model depicts the role of the NF-κB-YY1-miR-29 regulatory circuit in both normal myogenic differentiation and RMS. In myogenesis this circuit involves constitutive activity of NF-κB in myoblasts regulating YY1, which subsequently epigentically suppresses miR-29 and maintains cells in an undifferentiated state. As differentiation ensues, downregulation of the NF-κB–YY1 pathway leads to upregulation of miR-29 that in turns further decreases YY1 levels to ensure proper differentiation into myotubes. In RMS, this circuit becomes dysregulated due to an increase in the NF-κB–YY1 pathway that constitutively represses miR-29. In the absence of miR-29 tumor suppressor activity, YY1 is left uncontrolled thereby impairing differentiation leading to Rhabdomyosarcomagenesis.