Preclinical validation of TGFβ inhibitors as a novel therapeutic strategy for post-traumatic heterotopic ossification

Abstract

Heterotopic ossification (HO) is characterized by the abnormal growth of ectopic bone in non-skeletal soft tissues through a fibrotic pathway and is a frequent complication in a wide variety of musculoskeletal injuries. We have previously demonstrated that TGFβ levels are elevated in the soft tissues following extremity injuries. Since TGFβ mediates the initial inflammatory and wound-healing response in the traumatized muscle bed, we hypothesized that targeted inhibition of the TGFβ pathway may be able to abrogate the unbalanced fibrotic phenotype and bone-forming response observed in post-traumatic HO. Primary mesenchymal progenitor cells (MPCs) harvested from debrided traumatized human muscle tissue were used in this study. After treatment with TGFβ inhibitors (SB431542, Galunisertib/LY2157299, Halofuginone, and SIS3) cell proliferation/survival, fibrotic formation, osteogenic induction, gene expression, and phosphorylation of SMAD2/3 were assessed. In vivo studies were performed with a Sprague-Dawley rat blast model treated with the TGFβ inhibitors. The treatment effects on the rat tissues were investigated by radiographs, histology, and gene expression analyses. Primary MPCs treated with TGFβ had a significant increase in the number of fibrotic nodules compared to the control, while TGFβ inhibitors that directly block the TGFβ extracellular receptor had the greatest effect on reducing the number of fibrotic nodules and significantly reducing the expression of fibrotic genes. In vivo studies demonstrated a trend towards a lower extent of HO formation by radiographic analysis up to 4 months after injury when animals were treated with the TGFβ inhibitors SB431542, Halofuginone and SIS3. Altogether, our results suggest that targeted inhibition of the TGFβ pathway may be a useful therapeutic strategy for post-traumatic HO patients.

Keywords: Fibrosis; Heterotopic ossification; SMAD2; SMAD3; TGFβ; TGFβ inhibitors; Trauma.

Fig. 1

TGFβ inhibitors selectively regulate SMAD2/3 phosphorylation in human primary MPCs. (A) Primary MPCs were treated with ALK5 inhibitors SB431542 or Galunisertib/LY2157299 [0.3 µM, 1 µM, and 3 µM]. SB431542 and Galunisertib/LY2157299 treatments were performed simultaneously with the TGFβ [10 ng/mL] treatment for 30 min. (B) Primary MPCs were treated with SMAD3 inhibitors Halofuginone [3 nM, 10 nM, and 30 nM] or SIS3 [5 µM, 10 µM, and 20 µM]. Halofuginone treatment was performed for 24 h before TGFβ [10 ng/mL] treatment for 30 min and SIS3 treatment was performed for 24 h before TGFβ [10 ng/mL] treatment for 15 min. GAPDH was used as the loading control. DMSO treatment was used as vehicle control. Molecular weight is shown in Kilodaltons (kDa). TGFB = TGFβ, SB = SB431542, LY = Galunisertib/LY2157299, Halo = Halofuginone.

Fig. 2

Targeted TGFβ inhibitors down-regulate the expression of fibrotic markers in human primary MPCs. Primary MPCs were treated with ALK5 inhibitors (SB431542 or Galunisertib/LY2157299) and SMAD3 inhibitors (Halofuginone or SIS3) and TGFβ [10 ng/mL] for 48 h. Relative expression levels were analyzed by qRT-PCR for the fibrotic markers (A) ACTA2, (B) COL1A1, (C) FN1, and (D) SERPINE1. GAPDH was used as a housekeeping control gene. DMSO treatment was used as vehicle control and used as a reference to calculate relative gene expression levels. Average ± SD from three to four independent donors. Adjusted p-value was calculated by 1-tail Student’s T-test in comparison with TGFβ treatment alone. TGFB = TGFβ, SB = SB431542, LY = Galunisertib/LY2157299 and Halo = Halofuginone.

Fig. 3

Targeted TGFβ inhibitors selectively regulate the expression of RUNX2 in human primary MPCs. Primary MPCs were treated with ALK5 inhibitors (SB431542 or Galunisertib/LY2157299) and SMAD3 inhibitors (Halofuginone or SIS3) and TGFβ [10 ng/mL] for (A) 2 days and (B) 1 week. Relative expression levels were analyzed by qRT-PCR. GAPDH was used as a housekeeping control gene. DMSO treatment was used as a vehicle control and used as a reference to calculate relative gene expression levels. Average ± SD from three to four independent donors. Adjusted p-value was calculated by 1-tail Student’s T-test in comparison with TGFβ treatment alone (**p ≤ 0.01). TGFB = TGFβ, SB = SB431542, LY = Galunisertib/LY2157299 and Halo = Halofuginone.

Fig. 4

Targeted TGFβ inhibitors selectively regulate the formation of fibrotic nodules in vitro. (A,B) Primary MPCs were treated with ALK5 inhibitors (SB431542 or Galunisertib/LY2157299) and SMAD3 inhibitors (Halofuginone or SIS3) and TGFβ [10 ng/mL] for 4 days. Results are representative of 2 independent donors. (C) Average ± SD number of nodules per well from three independent donors treated with ALK5 inhibitors (SB431542 or Galunisertib/LY2157299) and TGFβ [10 ng/mL]. (D) Average ± SD number of nodules per well from three independent donors treated with SMAD3 inhibitors (Halofuginone or SIS3) and TGFβ [10 ng/mL]. Non-treated control was used for comparison and DMSO treatment was used as a vehicle control. Adjusted p-value was calculated by 2-tail Student’s T-test in comparison with the TGFβ treatment alone (*p ≤ 0.05). TGFB = TGFβ, SB = SB431542, LY = Galunisertib/LY2157299 and Halo = Halofuginone.

Fig. 5

ALK5 TGFβ inhibitors downregulate the expression of fibrotic markers following the fibrotic nodule formation assay in human primary MPCs. Primary MPCs were treated with ALK5 inhibitors (SB431542 or Galunisertib/LY2157299) and SMAD3 inhibitors (Halofuginone or SIS3) and TGFβ [10 ng/mL] for 4 days following the fibrotic nodule formation assay conditions. Relative expression levels were analyzed by qRT-PCR for the fibrotic markers (A) ACTA2, (B) COL1A1, (C) COL3A1, (D) VIM, (E) FN1, (F) SERPINE1, (G) MMP9 and (H) CDH2. GAPDH was used as a housekeeping control gene. Non-treated control was used as a reference to calculate relative gene expression levels. DMSO treatment was used as a vehicle control. Average ± SE from three to seven independent donors. Adjusted p-value was calculated by 1-tail Student’s T-test in comparison with TGFβ treatment alone (*p ≤ 0.05). TGFB = TGFβ, SB = SB431542, LY = Galunisertib/LY2157299 and Halo = Halofuginone.

Fig. 6

ALK5 TGFβ inhibitors downregulate the expression of the osteogenic regulator RUNX2 following the fibrotic nodule formation assay in human primary MPCs. Primary MPCs were treated with ALK5 inhibitors (SB431542 or Galunisertib/LY2157299) and SMAD3 inhibitors (Halofuginone or SIS3) and TGFβ [10 ng/mL] for 4 days following fibrotic nodule formation assay conditions. Relative expression levels were analyzed by qRT-PCR for the osteogenic markers (A) RUNX2, (B) ALP, and (C) BGLAP. GAPDH was used as a housekeeping control gene. Non-treated control was used as a reference to calculate relative gene expression levels. DMSO treatment was used as vehicle control. Average ± SE from six to seven independent donors. Adjusted p-value was calculated by 1-tail Student’s T-test in comparison with TGFβ treatment alone (*p ≤ 0.05). TGFB = TGFβ, SB = SB431542, LY = Galunisertib/LY2157299 and Halo = Halofuginone.

Fig. 7

SB431542, Halofuginone, and SIS3 show a trend towards a lower extent of HO formation by radiographic analysis in an in vivo rat blast model of HO. Animals underwent routine X-ray assessments beginning after the first month following amputation for 2-week intervals for the remainder of their survival. (A) SB431542, (B) Galunisertib/LY2157299, (C) Halofuginone or (D) SIS3. Low dose, medium dose and high treatment dose, respectively, were administered as follows: SB431542 [5 µg/g, 10 µg/g and 30 µg/g bodyweight], Galunisertib/LY2157299 [16 µg/g, 30 µg/g and 36 µg/g bodyweight], Halofuginone [2.5 µg/g, 5 µg/g and 10 µg/g bodyweight] and SIS3 [2 µg/g, 10 µg/g and 50 µg/g bodyweight]. Grade scale of 0–2: grade = 0 represented no HO; grade = 1 represented mild HO (> 0 but less than 25% of the diaphyseal diameter), and grade = 2 represented moderate to severe HO (more than > 25% of the diaphyseal diameter; modified from Hoyt et al.). Grades for each radiograph were averaged to give a final score for each image. 1wk = 1 week; 2wks = 2 weeks; 1 M = 1 month; 1.5 M = 1.5 months; 2 M = 2 months; 2.5 M = 2.5 months; 3 M = 3 months; 3.5 M = 3.5 months; 4 M = 4 months. Statistical analysis was performed using a linear mixed model (*p ≤ 0.05).

Fig. 8

TGFβ inhibitors modulation of the expression of fibrotic markers in an in vivo rat blast model of HO. Following blast injury, animals were treated with ALK5 inhibitors (SB431542 or Galunisertib/LY2157299) and SMAD3 inhibitors (Halofuginone or SIS3) for 2 weeks (n = 02) and up to for 4 months (n = 03). Low dose, medium dose and high treatment dose, respectively, were administered as follows: SB431542 [5 µg/g, 10 µg/g and 30 µg/g bodyweight], Galunisertib/LY2157299 [16 µg/g, 30 µg/g and 36 µg/g bodyweight], Halofuginone [2.5 µg/g, 5 µg/g and 10 µg/g bodyweight] and SIS3 [2 µg/g, 10 µg/g and 50 µg/g bodyweight]. Relative expression levels were analyzed by qRT-PCR for the fibrotic markers (A) Acta2, (B) Col1a1, (C) Col3a1, (D) Fn1 and (E) Serpine1. β-actin (Actb) was used as a housekeeping control gene. Non-treated (vehicle) control was used as a reference to calculate relative gene expression levels. Average ± SD from independent animals. Adjusted p-value was calculated by 1-tail Student’s T-test in comparison with the control group from each timepoint.

Fig. 9

TGFβ inhibitors modulation of the expression of osteogenic markers in an in vivo rat blast model of HO. Following blast injury, animals were treated with ALK5 inhibitors (SB431542 or Galunisertib/LY2157299) and SMAD3 inhibitors (Halofuginone or SIS3) for 2 weeks (n = 02) and up to for 4 months (n = 03). Low dose, medium dose and high treatment dose, respectively, were administered as follows: SB431542 [5 µg/g, 10 µg/g and 30 µg/g bodyweight], Galunisertib/LY2157299 [16 µg/g, 30 µg/g and 36 µg/g bodyweight], Halofuginone [2.5 µg/g, 5 µg/g and 10 µg/g bodyweight] and SIS3 [2 µg/g, 10 µg/g and 50 µg/g bodyweight]. Relative expression levels were analyzed by qRT-PCR for the osteogenic markers (A) Alp and (B) Runx2. β-actin (Actb) was used as a housekeeping control gene. Non-treated (vehicle) control was used as a reference to calculate relative gene expression levels. Average ± SD from independent animals. Adjusted p-value was calculated by 1-tail Student’s T-test in comparison with the control group from each timepoint.

Fig. 10

Schematic representation of the rat blast and no-trauma control treatment groups.