Reversible modulation of myofibroblast differentiation in adipose-derived mesenchymal stem cells

Abstract

Unregulated activity of myofibroblasts, highly contractile cells that deposit abundant extracellular matrix (ECM), leads to fibrosis. To study the modulation of myofibroblast activity, we used human adipose-derived mesenchymal stem cells (ADSCs), which have much potential in regenerative medicine. We found that ADSCs treated with TGF-β developed a myofibroblastic phenotype with increases in α-smooth muscle actin (α-SMA), a myofibroblast marker, and ECM proteins type I collagen and fibronectin. In contrast, treatment with bFGF had the opposite effect. bFGF-differentiated ADSCs showed marked down-regulation of α-SMA expression, collagen I, and fibronectin, and loss of focal adhesions and stress fibers. Functionally, bFGF-differentiated ADSCs were significantly more migratory, which correlated with up-regulation of tenascin-C, an anti-adhesive ECM protein, and vimentin, a pro-migratory cytoskeletal protein. On the other hand, TGF-β-differentiated ADSCs were significantly more contractile than bFGF-differentiated cells. Interestingly, cells completely reversed their morphologies, marker expression, signaling pathways, and contractility versus migratory profiles when switched from culture with one growth factor to the other, demonstrating that the myofibroblast differentiation process is not terminal. Cell differentiation was associated with activation of Smad2 downstream of TGF-β and of ERK/MAP kinase downstream of bFGF. Reversibility of the TGF-β-induced myofibroblastic phenotype depends, in part, on bFGF-induced ERK/MAP kinase signaling. These findings show that ADSC differentiation into myofibroblasts and re-differentiation into fibroblast-like cells can be manipulated with growth factors, which may have implications in the development of novel therapeutic strategies to reduce the risk of fibrosis.

Figure 1. Differences in cell morphology with TGF-β and bFGF.

ADSCs in SSFM were treated with 10/mL bFGF plus 5 µg/ml heparin (bFGF), 1 ng/mL TGF-β1 (TGFβ), or no additions (Unt). Phase contrast images were taken at 2 days (2d) and 4 days (4d) of treatment. Images are representative of more than three independent experiments. Scale bar = 50 µm.

Figure 2. Cytoskeletal responses to TGF-β and bFGF.

Cell lysates prepared from ADSCs treated with TGFβ, bFGF or untreated (Unt) were subjected to SDS-PAGE and immunoblotted. (A) 4-day lysates were probed with anti-α-SMA monoclonal antibody or anti-GAPDH antibody as a loading control. Blot is representative of three independent experiments. (B) Lysates prepared at 1 hour of treatment were immunoblotted with anti-phospho-Smad2 (pSmad2) and with total Smad antibody. Blot is representative of three independent experiments. Dash represents 37 kD (A) or 50 kD (B). (C, D) After 4 days of treatment, cells were replated onto collagen-coated glass coverslips for two hours before fixation, permeabilization, and staining with anti-vinculin antibody (C, top), rhodamine-phalloidin (C, bottom), or anti-α-SMA monoclonal antibody (D). Insets in (D) show rhodamine-phalloidin staining of the same fields. All images are representative of three independent experiments. Scale bars = 50 µm.

Figure 3. Effects of growth factors on ECM expression.

Total RNA was isolated from untreated (Unt), bFGF-treated, or TGF-β-treated ADSCs after 4 days of treatment and qRT-PCR was performed for type I collagen (A) and fibronectin (C) using ubiquitin C as the normalization control. Graphs represent the average of multiple samples from three independent experiments; error bars indicate SEM. *p<0.05, **p<0.01, and ***p<0.001. (B) Anti-type I procollagen antibody was used to detect collagen in conditioned media separated in an 8% polyacrylamide-SDS gel. Sample loadings were normalized to total RNA amounts isolated from the cells. Blot is representative of three independent experiments. (D) Equivalent amounts of protein in urea lysates were separated in a 5% polyacrylamide-SDS gel, and fibronectin was detected with an anti-fibronectin monoclonal antibody. Anti-GAPDH was used to confirm equal loading (not shown). Blot is representative of two independent experiments. Dash represents 150 kD (B) or 250 kD (D).

Figure 4. bFGF induces a migratory cell phenotype.

(A) Migration of ADSCs either untreated (Unt) or treated with bFGF or TGF-β for four days was measured using a Transwell migration assay. Cells were plated on a type I collagen-coated Transwell filter in SSFM. SSFM containing 2% serum was placed in the bottom chamber and cells were allowed to migrate for 21.5 hours. Cells were fixed and stained with DAPI and counted by fluorescence microscopy. Quantification represents the average number of cells per field +/− SEM from four independent experiments. ***p<0.001. (B, C) Total RNA was isolated from untreated (Unt), bFGF-treated, or TGF-β-treated ADSCs after 4 days of treatment. RNA was used for qRT-PCR with primers for tenascin-C (B) or vimentin (C). Ubiquitin C was used as the normalization control. Graphs represent the average of multiple samples from three independent experiments +/− SEM. *p<0.05, **p<0.01.

Figure 5. Growth factor-induced myofibroblast differentiation is reversible.

(A) Phase images are shown of ADSCs treated with TGF-β (TGF-β, 4d) or with bFGF (bFGF, 4d) for 4 days (left). SSFM containing bFGF was added to TGF-β-differentiated cells and TGF-β was added to bFGF-differentiated cells. Phase images were captured after 4 days (right). Images are representative of more than three independent experiments. Scale bar = 50 µm. (B) Lysates were prepared from cells grown under the six indicated conditions and were used for immunoblots with anti-α-SMA monoclonal antibody. GAPDH was used as the loading control. Blot is representative of three independent experiments. Legend indicates cell treatments in lanes 1–6. Unt., 4d = untreated for 4 days.

Figure 6. Inhibition of TGF-β-receptor signaling is not sufficient to reverse ECM protein expression.

(A) Lysates were harvested for cells grown as indicated in the legend for lanes 1–4, and immunoblotted with anti-phospho Smad2 or anti-total Smad antibody. GAPDH was used as the loading control. Blot is representative of three independent experiments. (B, C) bFGF, TGF-β receptor inhibitor (TGFβ-RI), or DMSO was added to TGF-β-differentiated ADSCs on day 4. Total RNA was isolated at 6, 24, and 48 hours of treatment and qRT-PCR was performed for type I collagen (B) and tenascin-C (C) expression relative to expression in TGF-β-differentiated ADSCs on day 4, using ubiquitin C as the normalization control. Graphs represent the average of multiple samples from three independent experiments +/− SEM. *p<0.05, **p<0.01, and ***p<0.001. The p-values are for comparisons to expression in TGF-β-differentiated ADSCs, unless otherwise indicated.

Figure 7. bFGF and ERK/MAP kinase induce tenascin-C expression.

(A) Lysates were harvested after growth factor treatment as indicated in the legend for lanes 1–4. Immunoblots were probed with phospho-specific anti-ERK1/2 monoclonal antibody or anti-total ERK antibody. GAPDH was used as the loading control. Blot is representative of three independent experiments. Dash represents 37 kD. (B) bFGF, along with 20 µM PD98059 or DMSO, was added to TGF-β-differentiated ADSCs at day 4. Total RNA was isolated after 48 hours of treatment and qRT-PCR was performed for tenascin-C using ubiquitin C as the normalization control. Graphs represent the average of multiple samples from three independent experiments +/− SEM. *p<0.05, ***p<0.001.

Figure 8. Reversibility of contractile and migratory phenotypes.

(A) The % contraction of fibrin-fibronectin matrices by differentiated and re-differentiated cells was averaged for 6–14 total reactions (+/− SEM) from three independent experiments. **p<0.01, ***p<0.001 (B) A Transwell migration assay was performed with re-differentiated cells. Unt, bFGF, and TGFβ data are from Figure 4A. Quantification represents the average number of cells per field +/− SEM from three to four independent experiments. **p<0.01, ***p<0.001. (C) Comparison of myofibroblast and fibroblast features. TGF-β-differentiated ADSCs are myofibroblastic, contractile, and produce increased type I collagen, fibronectin, α-SMA, and phospho-Smad2. Fibroblast-like cells lack myofibroblastic features, are more migratory, and have increased levels of tenascin-C, vimentin, and phospho-ERK1/2. These fibroblastic and myofibroblastic phenotypes are not terminal and are reversed by switching the respective growth factor treatments.