Fig 1.
CXCL12/CXCR4 axis-mediated myofibroblast phenoconversion is not coupled to Smad3 phosphorylation.
N1 fibroblasts (A) and primary prostate fibroblasts (B) were treated in defined serum-free Ham’s media with CXCL12 (100pM), or 0.01% BSA vehicle, and TGFβ (4 ng/mL) or 20mM citric acid vehicle. In both types of cells EGFR, Akt and Erk1/2 were phosphorylated upon CXCL12 treatment. TGFβ treatment activated Smad-mediated signaling and transient Erk1/2 phosphorylation, but not EGFR or Akt phosphorylation. Total antibodies for each kinase, as well as GAPDH and actin, were used as loading control. Protein molecular weight in kilodaltons is indicted by arrows. Signal intensity quantification (C) for N1 fibroblasts shows that EGFR activation occurs at 5 minutes post-treatment, followed by a robust activation of Erk at 10 minutes as well as mild Akt activation. The same pattern was observed in primary fibroblasts (D), however EGFR activation surpassed that of Erk.
Fig 2.
CXCL12/CXCR4-mediated signaling and myofibroblast phenoconversion requires EGFR transactivation.
(A) N1 fibroblasts were pre-treated with chemical inhibitors of CXCR4 (25 μM AMD3100), EGFR (500 nM AG1478), and TGFβRI (500 nM SB431542) or with 50 ng/ml EGF or 4ng/ml TGFβ for three hours. No activation of downstream signaling kinases (Akt, smad3, Erk) was observed in inhibitor-only treatment of fibroblasts. (B) N1 fibroblasts were pre-treated with chemical inhibitors of CXCR4 (25 μM AMD3100), EGFR (500 nM AG1478), and TGFβRI (500 nM SB431542) for two hours prior to CXCL12 (100 pM) treatment. Treatment with AMD3100 and AG1478 completely ablated the phosphorylation and activation of downstream targets. Inhibition of TGFβRI activation, with SB431542, had no effect of CXCL12 signaling.
Fig 3.
Activation of the CXCL12/CXCR4 and TGFβ/TGFβR axes independently promote ACTA2 and COL1A1 gene expression.
(A) N1 fibroblasts were treated with vehicle (DMSO) or EGFR inhibitor (AG1478, 500 nM) prior to CXCL12 treatment. Phosphorylation of EGFR, Akt, Smad and ERK1/2 were assessed via western blot. (B) qRT-PCR analysis of myofibroblast marker expression after CXCL12 (100pM) treatment in the presence or absence of AG1478 (500 nM). Expression levels of α-smooth muscle actin (ACTA2) and collagen 1α1 (COL1α1) were analyzed over the course of 24 hours of treatment. Treatment with AG1478 reduced or ablated the CXCL12/CXCR4 -mediated stimulation of both the ACTA2 and COL1α1 genes. (C) Western blot analysis of fibroblast markers, α-smooth muscle actin and collagen 1, at 24 and 48 hours after CXCL12 treatment. Secretion and incorporation of collagen 1 was inhibited in the presence of AG1478, α-smooth muscle actin production was inhibited as well. (D) Signal intensity quantification for 7 and collagen 1 western blots. * = p-value < 0.05. Error bars, SE.
Fig 4.
The CXCL12/CXCR4 and TGFβ/TGFβRI axes function independently to promote myofibroblast phenoconversion.
N1 fibroblasts were transfected with 20nM Scramble, anti-CXCR4 (A), and anti-TGFβRI (B) siRNAs for 24 hours and treated with CXCL12 (100pM) for 1 hour. CXCR4 partial knockdown (A) reduced the CXCL12-mediated activation of EGFR, Akt and Erk1/2. However, TGFβRI knockdown (B) had no effect in the CXCL12-mediated activation of EGFR, Akt and ERK1/2. siRNA-mediated knockdown were validated using antibodies against target receptor. Total antibodies for each kinase and actin were used as loading control (Western blot quantification is provided in S1 Fig). N1 fibroblasts were transfected with 20nM Scramble, anti-CXCR4, and anti-TGFβRI siRNAs for 24 hours and treated with CXCL12 (100 pM) (C) and TGFβ (4 ng/mL) (D) for 24 hours. qRT-PCR analysis of α-smooth muscle actin (ACTA2), collagen 1α1 (COL1α1) and TGFβ (TGFβ 1) was performed. (E) Western blot analysis to confirm effective knockdown of CXCR4 and TGFβRI during gene expression analysis. * = p-Value <0.05, ** = p-Value < 0.001. Error bars, SE.
Fig 5.
Inhibition TGFβ and TGFβRII-mediated signaling does not affect CXCL12-mediated myofibroblast phenoconversion.
(A) Immunofluorescence analysis of N1 fibroblasts left untreated (-CXCL12) or treated with 100pM CXCL12 (CXCL12) for 48 hours in the absence or presence of CXCR4 (250 μM AMD3100), EGFR (250 uM AG1478), ALK-5 (TGFβRII) (20μM A-83-01) small molecular inhibitors or an antibody against TGFβRII (200 ng/ml TGFβ MAb). Figure depicts photomicrographs images of cells stained for α-smooth muscle actin (green) or collagen 1 (red) proteins or DAPI (blue nuclear stain); orange color indicates α-smooth muscle actin and collagen 1 colocalization in the merged image. (B) Western blot analysis of fibroblast markers, α-smooth muscle actin and collagen 1 at 24, 48 and 72 hours after CXCL12 (100pM) of scramble, anti-CXCR4, and anti-TGFβRI siRNAs transfected fibroblasts. CXCL12-driven expression of myofibroblasts markers does not require the presence or activation of TGFβRI. (C) Quantification of western blot images for myofibroblast markers.
Fig 6.
Proposed mechanism of action used by CXCL12 to drive myofibroblast phenoconversion.
The canonical pathway driving myofibroblast differentiation is driven by TGF-β-mediated Smad and MEK/Erk signaling14. However, upon binding to CXCR4, CXCL12 promotes the transactivation of EGFR, potentially through MMP/ADAM as previously described by Kasina et.al in prostate cancer cells35 and Src activation, as shown in the supplementary data (S2 Fig). Active EGFR then activates Akt and MEK/ERK signaling pathways to promote myofibroblast marker gene expression and phenoconversion. CXCL12-mediated phenoconversion acts independently of TGFβR/Smad-mediated myofibroblast phenoconversion.