miR-206 integrates multiple components of differentiation pathways to control the transition from growth to differentiation in rhabdomyosarcoma cells

Abstract

Background: Similar to replicating myoblasts, many rhabdomyosarcoma cells express the myogenic determination gene MyoD. In contrast to myoblasts, rhabdomyosarcoma cells do not make the transition from a regulative growth phase to terminal differentiation. Previously we demonstrated that the forced expression of MyoD with its E-protein dimerization partner was sufficient to induce differentiation and suppress multiple growth-promoting genes, suggesting that the dimer was targeting a switch that regulated the transition from growth to differentiation. Our data also suggested that a balance between various inhibitory transcription factors and MyoD activity kept rhabdomyosarcomas trapped in a proliferative state.

Methods: Potential myogenic co-factors were tested for their ability to drive differentiation in rhabdomyosarcoma cell culture models, and their relation to MyoD activity determined through molecular biological experiments.

Results: Modulation of the transcription factors RUNX1 and ZNF238 can induce differentiation in rhabdomyosarcoma cells and their activity is integrated, at least in part, through the activation of miR-206, which acts as a genetic switch to transition the cell from a proliferative growth phase to differentiation. The inhibitory transcription factor MSC also plays a role in controlling miR-206, appearing to function by occluding a binding site for MyoD in the miR-206 promoter.

Conclusions: These findings support a network model composed of coupled regulatory circuits with miR-206 functioning as a switch regulating the transition from one stable state (growth) to another (differentiation).

Figure 1

Expression of RUNX1 or ZNF238 leads to terminal differentiation of RMS cells. (A) qPCR for RUNX1 was performed in RD cells either infected with a control virus, or the forced MyoD ~ E dimer (MD ~ E) as well as control (0 h) human fibroblasts and fibroblasts differentiated into myotubes (96 h). (B) RT-PCR for the two isoforms of ZNF238 in RD cells and fibroblasts as in 1A. (C) Myosin heavy chain (MHC) immunostains in RD cells either not infected (no infection), infected with a control GFP-expressing lentivirus (GFP control) or RUNX1 or ZNF238 expressing lentivirus. All cells were infected at approximately equivalent MOIs, and cells were allowed to differentiate for 72 h in low-serum media before staining. GFP was detected directly, without the use of an antibody. (D) qPCR for muscle-specific creatine kinase (CKM) in RD cells infected with either ZNF238 or RUNX1 viruses compared to cells with control retroviruses. (E) After 24 h of differentiation, RD cells were pulsed for a further 24 h with EdU-containing differentiation media, before fixation and quantification of the percentage of EdU-positive cells. (F) qPCR for ID2 and ID3 in control and ZNF238-expressing RD cells. All qPCR data are normalized to TIMM17b expression, and the level in control cells is set to 1. All bar graphs represent the mean ± SEM of at least three independent experiments. *: P < 0.05; **: P < 0.01; ***: P < 0.001; ****: P < 1 × 10^-4.

Figure 2

MyoD ~ E, RUNX1, and ZNF238 increase miR-206. (A) microRNA northern blots to detect the mature form of the indicated microRNAs in RD cells infected with either empty (control) retrovirus, or retrovirus expressing MyoD ~ E (MD ~ E). (B) RT-PCR using primers located in the pre- and pri-miR-206 sequence to detect the primary miR-206 transcript. TIMM17b is the internal control. (C) microRNA northerns as in 2A, in RD cells expressing a transcription factor as indicated. (D) microRNA northerns in C2C12 cells at various stages of differentiation ranging from undifferentiated myoblasts (50% GM), through beginning differentiation (90% GM) to myotubes (DM). (E) Immunostains for MHC in RD cells transiently transfected with either a pre-miR-206 RNA construct, or a negative control construct. Nuclei were stained with DAPI. (F) qPCR for CKM in RD cells treated as in E. (G) RD cells treated as in E were pulsed with BrdU for 24 h and then stained and counted to determine the extent of co-localization of MHC + myotubes and BrdU + nuclei. (H) RD cells transiently transfected as in E were pulsed for 24 h with BrdU-containing differentiation media before fixation and quantification of the percentage of BrdU positive cells. The percent reduction of BrdU + nuclei almost exactly equals the percent of cells found to be MHC + (not shown). (I) Luciferase activity in RD cells using a miR-206 promoter driven reporter and transiently transfected factors as indicated. ‘206 RUNX mutant’ indicates that the reporter has had a putative RUNX1 binding site mutated to prevent RUNX1 binding. Control indicates transfection with an empty plasmid. All luciferase assays were normalized to the results from a co-transfected renilla plasmid. (J) RUNX1 ChIP assays, both with normal PCR and qPCR results, at the miR-206 promoter and a control locus before (Control) and after (RUNX1) infection of the cells with empty or RUNX1-expressing retrovirus. All PCRs in the imaged gel (upper) were performed for the same number of cycles. qPCR results (lower) represent the mean ± SEM of two independent ChIP experiments. Relative enrichment is calculated as the ratio of the % of input amplified with antibody to the % of input amplified with no antibody. All other graphs in this figure represent the mean ± SEM of at least three independent experiments. * : P < 0.05; ** : P < 0.01; *** : P < 0.001.

Figure 3

ZNF238 is a downstream target of MyoD and RUNX1. (A) qPCR for RUNX1 and ZNF238 in RD cells transduced with virus expressing the converse factor. (B) RUNX1 ChIP results at the intron of ZNF238 and a control locus before (Control) and after (RUNX1) infection of the cells with empty or RUNX1-expressing retrovirus. (C) qPCR for ID genes in control and RUNX1-expressing RD cells. PCRs in the imaged gel in 3B were performed for the same number of cycles. The graph in 3B represents qPCR data showing the mean ± SEM of two independent experiments. Relative enrichment is calculated as the ratio of the % of input amplified with antibody to the % of input amplified with no antibody. qPCRs in 3A and 3C are represented as the mean ± SEM of at least three independent experiments. *: P < 0.05; **: P < 0.01.

Figure 4

RUNX1, ZNF238, and miR-206 function through common mechanisms. (A) 3-way Venn diagram representing the overlap between significantly regulated (fold-change >2, FDR <0.05, overlap considers only changes in same direction) gene targets in RD cells differentiated either through RUNX1, ZNF238, or miR-206 expression relative to GFP-infected controls. The table indicates the breakdown of upregulated versus down-regulated genes for each portion of the Venn diagram. (B) Scatter plots showing pairwise comparisons of gene expression. All values are plotted as the log₂ of the fold-change relative to GFP-infected controls, as indicated along the x- and y-axes. Correlation is listed for each comparison in the matrix. (C) (upper panel) Bar graph demonstrating that the majority of the 95 genes listed as being ‘uniquely’ regulated by miR-206 in 4A are also regulated by RUNX1 and/or ZNF238, but at lower levels of expression change. FDR was kept constant (<0.05) in this analysis, and to be included as a ‘shared’ target, the change had to occur in the same direction (either up- or down-regulated) in RUNX1 and/or ZNF238 as in miR-206. (bottom panel) Analysis as in the top panel for genes in the ZNF238 unique and ZNF238:miR-206 intersection groups relative to RUNX1 changes. (D) RT-PCR for select gene targets from Additional file 9: Table S4. TIMM17b serves as the internal control.

Figure 5

MSC interferes with the ability of MyoD to positively regulate the microRNA miR-206 by blockading a necessary MyoD-binding site. (A) Site-specific ChIP for MSC in the miR-206 promoter and at a control locus. (B) MSC ChIP as in (A) in RD cells infected either with an empty retrovirus (Control), or RD cells differentiated through expression of RUNX1 (RUNX1). (C) Screenshot from the human UCSC Genome Browser of the region that corresponds to the miR-206 promoter. Mapped reads from ChIP-seq for MyoD in RD and HFF cells are indicated, with the number on the left-hand y-axis indicating the number of reads mapped at the peak of occupancy. The location of E-boxes are indicated at the bottom of the panel by the black rectangles. Vertical lines are drawn through the apparent highest points of occupancy for MyoD (red) and MSC (green) in RD cells. (D) Electrophoretic mobility shift assays using in vitro transcribed and translated proteins as indicated and probes that correspond to either the E-box located at the peak of MyoD binding as indicated by the red mark in 5C, or the E-box located at the peak of MSC binding, indicated by the green mark. The position of MyoD:E and MSC:E heterodimers are indicated by the arrows. The lane marked 2x E12 indicates protein mixtures that included double the amount of E12 compared to other lanes, and the triangles indicate decreasing amounts of either MSC or MyoD in the mixtures as other proteins were maintained at constant levels and total protein amounts were balanced with translation of empty CS2. (E) Luciferase assays in RD cells with constructs as indicated below the figure using either the miR-206 promoter luciferase reporter (206) or one in which the E-box that the peak of MSC occupancy is located over has been mutated (206 MSC-binding Ebox mutant). Control indicates transfection with an empty plasmid. All luciferase assays were normalized to the results from a co-transfected renilla plasmid. (F) qPCR for CKM from RD cells transduced with either an empty virus (control), or one expressing either the MD ~ E or MD ~ E2/5 forced dimer. Cells were differentiated for 24 h before collection of RNA for use in qPCR. (G) qPCR for RUNX1 from the RD cells assayed in E. (H) qPCR for pri-miR-206 from the RD cells assayed in E and F. For all ChIPs, relative enrichment is calculated as the ratio of the % of input amplified with antibody to the % of input amplified with no antibody. The control locus is located at hemoglobin beta, a silent gene in myogenic cells. Corrected relative enrichment (5B) is the ratio of enrichment at miR-206 to enrichment at the control locus. All graphs represent the mean ± SEM of at least three independent experiments. All qPCR of gene expression was corrected to TIMM17b and control cells set to 1. * : P <0.05; ** : P < 0.01; *** : P < 0.001; † : P = 0.058.

Figure 6

miR-206 integrates the output of oscillating circuits and acts as a genetic switch to transition from growth to differentiation. The experimental data support a network model composed of coupled oscillators with miR-206 functioning as a switch regulating the transition from one stable state to another. In myoblasts, MyoD, E-proteins, and ID proteins compose the first oscillating circuit: (1) MyoD:E heterodimers bind an E-box in the regulatory regions of the ID2 and ID3 genes and drive ID transcription; (2) the ID protein competitively forms dimers with the E-protein, limiting the production of active MyoD:E-protein heterodimers; (3) the decline in active MyoD:E-heterodimers results in decreased ID production; and (4) the decreased ID permits an increase in active MyoD:E-protein heterodimers and more ID production. The second oscillating circuit is composed of MyoD, E-proteins, MSC, and miR-206: (1) MyoD:E-protein and MSC:E-protein heterodimers compete for binding at the E-box in the miR-206 regulatory region, which oscillates between MyoD-activated and MSC-repressed states; (2) limiting amounts of E-protein prevent full activation by MyoD; and (3) low levels of miR-206 prevent full suppression by MSC. These circuits are coupled by their shared response to the concentration of active MyoD:E-protein heterodimers. The oscillating circuits bifurcate to a new determined state when the concentration and/or activity of the MyoD:E-proteins become sufficient to activate the expression of RUNX1 and ZNF238 in a feed-forward circuit that blocks the expression of the ID genes and permits the accumulation of active MyoD:E-protein complexes. The increase of MyoD:E-protein heterodimers together with RUNX produces higher miR-206 expression, and the increased miR-206 suppresses MSC and other inhibitors of differentiation.