Alternative splicing targeting the hTAF4-TAFH domain of TAF4 represses proliferation and accelerates chondrogenic differentiation of human mesenchymal stem cells

Abstract

Transcription factor IID (TFIID) activity can be regulated by cellular signals to specifically alter transcription of particular subsets of genes. Alternative splicing of TFIID subunits is often the result of external stimulation of upstream signaling pathways. We studied tissue distribution and cellular expression of different splice variants of TFIID subunit TAF4 mRNA and biochemical properties of its isoforms in human mesenchymal stem cells (hMSCs) to reveal the role of different isoforms of TAF4 in the regulation of proliferation and differentiation. Expression of TAF4 transcripts with exons VI or VII deleted, which results in a structurally modified hTAF4-TAFH domain, increases during early differentiation of hMSCs into osteoblasts, adipocytes and chondrocytes. Functional analysis data reveals that TAF4 isoforms with the deleted hTAF4-TAFH domain repress proliferation of hMSCs and preferentially promote chondrogenic differentiation at the expense of other developmental pathways. This study also provides initial data showing possible cross-talks between TAF4 and TP53 activity and switching between canonical and non-canonical WNT signaling in the processes of proliferation and differentiation of hMSCs. We propose that TAF4 isoforms generated by the alternative splicing participate in the conversion of the cellular transcriptional programs from the maintenance of stem cell state to differentiation, particularly differentiation along the chondrogenic pathway.

Figure 1. Expression analysis of TAF4 alternative splice variants in human tissues.

(A) RT-PCR analysis of human tissues using TAF4 transcript-specific primers. Primers for the full-length TAF4_v1 amplify all splice variants, whereas ASV-specific primers generate predominantly one PCR product. The numbers of PCR cycles and exposure times of the images for each set of primers vary and cannot be directly compared. The arrow indicates the canonical TAF4_v1 ASV. (B) Schematic representation of the human TAF4 gene structure and its alternative splice variants drawn in scale. The regions encoding the respective domains are indicated above the gene structure. Filled boxes represent the coding regions of the ASVs.

Figure 2. Expression of TAF4 splice variants and isoforms in human MSCs differentiated into adipocytes, osteoblasts or chondrocytes and treated with TAF4 RNAi.

(A) RT-PCR analysis of different individual hMSCs clones (hMSCs 1–5) using TAF4_v1a (full length) specific primers (see Table S2). Expression of GAPDH mRNA is shown at the bottom. (B) Expression of TAF4 ASVs encoding proteins with compromised hTAF4-TAFH domain is dominant in differentiated hMSCs. RT-PCR analysis using TAF4 ASV-specific primers was performed 7 days after induction of differentiation of hMSCs towards adipo-, osteo- and chondrogenic lineages. GAPDH mRNA expression was used for the normalization. (C) Expression of TAF4 ASVs following treatment of human MSCs with control and hTAF4-TAFH-domain targeting TAF4 siRNAs. Cells were treated with TAF4 or control siRNAs at the indicated time points and RT-PCR analysis performed using TAF4 ASV-specific primers. Analysis of GAPDH mRNA expression was used for normalization. (D) siRNA-mediated silencing targeting TAF4 exons V or VI induces changes in the expression of TAF4 protein isoforms as detected at 48 h post-treatment using Western blot analysis. The asterisk indicates the canonical form of TAF4 protein with the molecular weight of 135 kDa, two asterisks indicate TAF4_v2 isoform with a calculated molecular weight of about 73 kDa.

Figure 3. TAF4 siRNA treatment of human MSCs induces TP53-dependent cell-cycle arrest and switching from canonical to non-canonical WNT signaling.

(A) Effects of TAF4 siRNA treatments on hMSC proliferation at different time points as compared to negative control siRNA. hMSCs were transfected with 50 nM of TAF4 or control siRNAs and analyzed at the indicated time points by WST-1 proliferation assay. Experiments were done in triplicate and a comparison was made to control siRNA treatments (_** indicates significant differences between control and TAF4 siRNA groups with P<0.01). (B) Down-regulation of the canonical form of TAF4 affects the expression of cell cycle regulators. hMSCs were transfected with 50 nM TAF4-specific (ex5_ or ex6_TAF4 siRNAs) or control siRNAs. Relative expression of cell cycle regulators CDKN1A and CDK2 as compared to control siRNAs transfected cells was analyzed using real-time RT-PCR at 6 h post-treatment. Observed differences were statistically significant (Student's t-test) with *** P<0.001. (C) TAF4 siRNA-mediated RNAi affects expression of cell cycle regulator proteins. Western blot analysis of cell cycle regulators TP53 and CDKN1A/P21 24 h following transfection of hMSCs with control or both, ex5_ and ex6_TAF4-specific siRNAs. Expression of TAF4 and GAPDH was analyzed to ensure effective silencing and equal loading (left). Time-dependent expression of TP53 was analyzed by real-time PCR and compared to the mRNA levels in TAF4 and control siRNA treated hMSCs (right). (D) Down-regulation of the canonical form of TAF4 does not induce apoptosis. FACS analysis of the cell cycle progression of hMSCs treated with TAF4 siRNAs. (E) TAF4 siRNA treatment doesn't cause senescence of hMSCs. Quantitative SA-ß-gal assay. hMSC extracts were prepared from TAF4 and control siRNA transfected cells 48 h post-treatment. 1 µM 4-NQO was added to hMSCs for 1 h and used as positive senescence control. Fluorescence intensity of 4-MU hydrolysis was normalized to total protein. Error bars in experiments represent the standard deviations of three independent experiments (P<0.005) (F) RNAi of hTAF4-TAFH switches from canonical to non-canonical WNT signaling. 20 µg of control or TAF4 siRNA-treated whole cell lysates were analyzed by Western blot analysis for the expression of ß-catenin (CTNNB1) and GAPDH as loading control (left). Real-time PCR shows increased expression of non-canonical markers of WNT signaling in TAF4 siRNA treated hMSCs as compared to control siRNA treated cells. Differences are statistically significant with P<0.001 (right).

Figure 4. TAFH domain disruption by TAF4 siRNA attenuates adipogenesis and osteogenesis and accelerates chondrogenesis in human MSCs.

(A) TAF4 down-regulation slows down adipogenesis in hMSCs. Oil-Red-O staining of lipid droplets of TAF4 or control siRNA treated and adipogenesis stimulated hMSCs at day 6 post-induction. siRNAs were transfected 24 h prior to the stimulation of differentiation of hMSCs. Scale bar, 40 µm (top left). Western blot analysis reveals reduced expression of adipogenic markers, namely PPARG and ADIPOQ at day 5 after induction of differentiation of TAF4 siRNA treated hMSCs along adipocyte pathway (right panel). Real-time PCR analysis of the expression of adipocyte markers in hMSCs at different time points of post-transfection and differentiation. The expression is normalized to control siRNA treatments (middle panel). RT-PCR analysis of TAF4 ASVs expression using TAF4_v1b specific primers in TAF4 or control siRNA transfected and towards adipogenesis stimulated hMSCs at day 5 post-transfection. GAPDH mRNA was analyzed for gel loading normalization (left bottom). (B) RNAi of hTAF4-TAFH inhibits osteogenic differentiation of hMSCs. Alkaline phosphatase staining of TAF4 or control siRNA exposed and osteogenesis stimulated hMSCs at day 6 post-induction. siRNAs were transfected 24 h prior to the stimulation of hMSC differentiation. Scale bar, 40 µm (top left). Western blot analysis reveals reduced expression of osteogenic markers RUNX2 and OPN (right panel). Real-time PCR analysis of the expression of osteogenic markers in hMSCs at different time points post-transfection and differentiation. The expression is normalized to negative control siRNA treatment (middle panel). RT-PCR analysis of TAF4 ASVs expression using TAF4_v1b specific primers in TAF4 or control siRNA treated and stimulated to osteogenic differentiation hMSCs at day 5 post-transfection. GAPDH mRNA was analyzed for gel loading normalization (left bottom). (C) RNAi of hTAF4-TAFH accelerates chondrogenic differentiation of hMSCs. Immunofluorescence staining analysis of TAF4 or control siRNA treated hMSCs upon chondrogenic stimulation at day 6 post-treatment reveals induced expression of COL2A1 and SOX9. siRNAs were transfected 24 h prior to stimulation of differentiation of hMSCs. Scale bar, 40 µm (top left). Western blot analysis shows increased expression of chondrogenic marker genes in TAF4 siRNA treated hMSCs. COL2A1 expression was upregulated at day 2 post-differentiation; SOX9 and MMP13 level of expression were increased at 8 days after siRNA treatments (right panel). Real-time PCR analysis of the expression of chondrogenic markers in hMSCs at different time points post-transfection and differentiation. The expression is normalized to control siRNA treatment (middle panel). RT-PCR analysis of TAF4 ASV expression using TAF4_v1b specific primers in TAF4 or control siRNA treated hMSCs upon stimulation to chondrogenic differentiation at day 5 post-transfection. GAPDH mRNA was analyzed for gel loading normalization (left bottom). Effects of RNAi of hTAF4-TAFH on activation of TP53^Ser15 were observed in each differentiation study using immunoblot analysis. Real-time PCR differences were found to be statistically significant with P<0.005.