Simultaneous miRNA and mRNA transcriptome profiling of human myoblasts reveals a novel set of myogenic differentiation-associated miRNAs and their target genes

Abstract

Background: miRNA profiling performed in myogenic cells and biopsies from skeletal muscles has previously identified miRNAs involved in myogenesis.

Results: Here, we have performed miRNA transcriptome profiling in human affinity-purified CD56+ myoblasts induced to differentiate in vitro. In total, we have identified 60 miRNAs differentially expressed during myogenic differentiation. Many were not known for being differentially expressed during myogenic differentiation. Of these, 14 (miR-23b, miR-28, miR-98, miR-103, miR-107, miR-193a, miR-210, miR-324-5p, miR-324-3p, miR-331, miR-374, miR-432, miR-502, and miR-660) were upregulated and 6 (miR-31, miR-451, miR-452, miR-565, miR-594 and miR-659) were downregulated. mRNA transcriptome profiling performed in parallel resulted in identification of 6,616 genes differentially expressed during myogenic differentiation.

Conclusions: This simultaneous miRNA/mRNA transcriptome profiling allowed us to predict with high accuracy target genes of myogenesis-related microRNAs and to deduce their functions.

Figure 1

Expression of myogenic differentiation markers and miRNA profiling. A. Expression of muscle differentiation markers MYOG (myogenin, myogenic factor 4, p-value=0.0016), TNNT1 (Troponin T type 1, skeletal slow, p-value=0.00064), MYL4 (myosine light chain 4, p-value=0.0016), COL15A1 (collagen, type XV, alpha 1, p-value=0.013), TNNT2 (troponin T type 2 (cardiac), p-value=4.8x10^-6), VIM (vimentin, p-value=0.00034), ID1 (inhibitor of DNA binding 1, p-value=1.9x10^-5) and CAV1 (caveolin 1, p-value=0.00015) in proliferating myoblasts (labeled as P) and differentiated myotubes (labeled as D) was measured using qRT-PCR, normalization was performed using ΔΔCt method using GAPDH as a control gene and proliferating sample #1 as a reference sample (expression level 1). The average of three independent experiments is shown. Numbers from 1 to 6 indicate the sample number (for full description refer to Table S1). Insets indicate the average of 6 samples of proliferating myoblasts (P) and differentiated myotubes (D). Asterisk corresponds to p-values<0.05. Error bars correspond to standard deviation (SD) in the case of individual samples and standard error of the mean (SEM) in the case of average expression levels (inset). B. Results of immunofluorescence microscopy analysis of cells stained with anti-Ki67 (red), and anti-Desmin (green) antibodies and DAPI nuclear staining (blue) showing normal cellular localization of these proteins in proliferating myoblasts and differentiated myotubes, 20x magnification. Bar=10 μm. C. The number of Ki67+ cells is significantly reduced during myogenic differentiation in vitro. Results of quantification of the DES+ and Ki67+ cells representing 300 individual cells. Statistically significant difference between cell cultures of proliferating myoblasts and differentiated myotubes is indicated by an asterisk (t-test p-value<0.05).

Figure 2

miRNA profiling in proliferating myoblasts (left) and differentiated myotubes (right). Out of 365 miRNAs tested, this table shows 60 miRNAs that were differentially expressed in myotubes as compared to myoblasts. Gray levels indicates the expression level of microRNA in each individual sample tested. Labels 1P to 6P correspond to samples of proliferating myoblasts described in Table S1, labels 1D to 6D correspond to the samples of differentiated myotubes described in Additional file 1: Table S1.

Figure 3

Bioinformatic predictions of microRNA target genes. A. Density plot of Pearson correlation coefficients between expression of miRNA and their target genes (continuous line) predicted by RNA22 algorithm. Number of genes is expressed as% of total predictions indicated within each plot. The distribution of Pearson correlations is clearly shifted towards negative values; the density plot of Pearson correlation coefficient after permutation of the list of microRNAs (dashed line). In this case the distribution is centered around zero. Several examples are shown, for other microRNAs see the Additional file 6. B. Diagram showing the proportion of bioinformatic predictions made by RNA22 algorithm (taken for 100%) that were supported by transcriptome profiling (black). Light grey shows unsupported predictions. The graphs corresponding to TargetScan predictions are presented in the Additional file 8. C. Comparison of the accuracy of target gene predictions by RNA22 and TargetScan algorithms.% of target genes predicted by both algorithms and supported by transcriptome data is shown.

Figure 4

Confirmation of bioinformatic predictions of miR targets. Human immortalized myoblasts (iMyo) cultivated in growth medium (P) or differentiation medium for 3 days (D) were transfected with either control LNA (anti-miR-C) or LNA against miR-1 and miR-206, miR-133a and miR-133b (A). Human rhabdomyosarcoma cells (RD) were transfected with plasmids coding for miR-128 and miR-30 precursors or scrambled sequence (scr) (B). Then the expression of corresponding microRNA and their randomly selected target genes was tested using qRT-PCR, normalization was performed using ΔΔCt method using GAPDH as a control gene and proliferating sample #1 as a reference sample (expression level 1). Transcriptome-supported target genes of miR-1/206 (C) and miR-133a/b (D) were downregulated during normal myogenic differentiation but not when it was accompanied by the transfection with corresponding anti-miRs. Transcriptome-supported target genes of miR-128 and miR-30 were downregulated when human rhabdomyosarcoma cells were transfected with lentiviral constructs overexpressing miR-128 (E) and miR-30 (F). The average of three independent experiments is shown. (*) indicates p-value <0.05; (G): Diagram showing the proportion of supported predictions that were qRT-PCR validated in this study. Black: validated targets, gray: non-validated targets.

Figure 5

Predicted functions of MR-miRs. Green: functions downregulated during myogenic differentiation, Red: functions upregulated during myogenic differentiation. Yellow: functions that are both up- and downregulated during myogenic differentiation. Framed are the functions that have not been previously ascribed to a given microRNA.