TGF-β3 modulates the inflammatory environment and reduces scar formation following vocal fold mucosal injury in rats

Abstract

Transforming growth factor (TGF)-β1 and TGF-β3 have been reported to exert differential effects on wound healing, and possibly even account for tissue-specific differences in scar formation. Scarring is particularly detrimental in the vocal fold mucosa (VFM), where destruction of the native extracellular matrix causes irreparable biomechanical changes and voice impairment. Here, in a series of in vitro and in vivo experiments, we identified differences in TGF-β1 and TGF-β3 transcription and immunolocalization to various cell subpopulations in naïve and injured rat VFM, compared with oral mucosa (which undergoes rapid healing with minimal scar) and skin (which typically heals with scar). Treatment of cultured human vocal fold fibroblasts with TGF-β3 resulted in less potent induction of profibrotic gene transcription, extracellular matrix synthesis and fibroblast-myofibroblast differentiation, compared with treatment with TGF-β1 and TGF-β2. Finally, delivery of exogenous TGF-β3 to rat VFM during the acute injury phase modulated the early inflammatory environment and reduced eventual scar formation. These experiments show that the TGF-β isoforms have distinct roles in VFM maintenance and repair, and that TGF-β3 redirects wound healing to improve VFM scar outcomes in vivo.

Keywords: Fibrosis; Larynx; Tissue repair; Wound healing.

Fig. 1

Differential immunoactivity of TGF-β1 and TGF-β3 in naïve rat VFM. (A) Representative H&E-stained coronal section illustrating the lamina propria (LP), stratified squamous cells (SSC) along the superior aspect of the epithelium, and ciliated pseudocolumnar cell formation (CPF) along the inferior aspect of the epithelium. (B) Representative image showing positive immunostaining for TGF-β1 (red). (C,D) Enlarged images of the regions indicated by dashed boxes in B. Arrows indicate TGF-β1⁺ LP cells. Arrowheads indicate TGF-β1⁺ epithelial cells in the SSC and CPF regions. (E) Representative image showing positive immunostaining for TGF-β3 (red). (F,G) Enlarged images of the regions indicated by dashed boxes in E. Arrowheads indicate TGF-β3⁺ epithelial cells in the CPF region. Each image is representative of three independent animals and >3 replicate sections per animal. Dashed lines in C,D,F,G indicate the boundary between the LP (left) and epithelium (right). Scale bar: 300 μm (A,B,E); 30 μm (C,D,F,G).

Fig. 2

TGF-β1 and TGF-β3 transcription in injured VFM, OM and skin. (A-C) TGF-β1, TGF-β3 and macrophage-specific transcript Emr1 mRNA expression in VFM up to 7 days post-injury. (D) Comparison of basal TGF-β1 and TGF-β3 mRNA expression in naïve VFM, OM and skin. (E,F) Comparison of TGF-β1 and TGF-β3 mRNA expression in naïve and injured (72 hours post-injury) VFM, OM and skin. Results represent the means ± s.e.m. of three independent experiments with n=5 animals per time point, and are presented as fold change relative to naïve VFM. *P<0.01 versus naïve control within tissue type (A-C,E,F), versus naïve VFM (D) and across tissue types at 72 hours post-injury (E, F).

Fig. 3

TGF-β1 immunoactivity and CD68 colocalization in VFM post-injury. (A-D) Representative images showing low TGF-β1 (red) and CD68 (green) immunosignals in naïve VFM. Nuclei are counterstained blue. (E-X) Representative images showing an increase in TGF-β1⁺ lamina propria and epithelial cells, and extracellular TGF-β1 immunosignals, 12 hours to 7 days post-injury. A subpopulation of the TGF-β1⁺ lamina propria cells are CD68⁺ macrophages. Each image reflects the stratified squamous cell region of the VFM and is representative of three independent animals and >3 replicate sections per animal. White arrowheads indicate TGF-β1⁺CD68⁻ cells; white arrows indicate TGF-β1⁻CD68⁺ macrophages; yellow arrows indicate TGF-β1⁺CD68⁺ macrophages; white dashed lines indicate the boundary between the lamina propria (left) and epithelium (right). The absence or fragmentation of these lines depicts incomplete reepithelialization. Scale bar: 30 μm.

Fig. 4

Transient upregulation of TGF-β3 is associated with vocal fold epithelial cell recovery. (A-D) Representative images showing low TGF-β3 (red) and CD68 (green) immunosignals in naïve VFM. Nuclei are counterstained blue. (E-X) Representative images showing initial appearance of TGF-β3 immunosignals corresponding to early epithelial cell recruitment at 24 hours post-injury, peak TGF-β3 expression corresponding to the completion of reepithelialization at 72 hours post-injury, and reduction of epithelial cell TGF-β3 expression at 5 and 7 days post-injury. The majority of TGF-β3⁺ cells are CD68⁻. Each image reflects the stratified squamous cell region of the VFM and is representative of three independent animals and >3 replicate sections per animal. White arrowheads indicate TGF-β3⁺CD68⁻ epithelial cells; white arrows indicate TGF-β3⁻CD68⁺ macrophages; yellow arrows indicate TGF-β3⁺CD68⁺ macrophages; white dashed lines indicate the boundary between the lamina propria (left) and epithelium (right). The absence or fragmentation of these lines depicts incomplete reepithelialization. Scale bar: 30 μm.

Fig. 5

TGF-β3 is a less potent inducer of profibrotic molecule expression and myofibroblast differentiation in human VFF than TGF-β1 and TGF-β2. (A) Treatment with 5 and 10 ng/ml TGF-β3 induced less Col1a1 transcription than treatment with TGF-β1 or TGF-β2, in a dose-dependent manner. Western blotting was performed to evaluate collagen type 10 ng/ml condition. (B) Treatment with 5 and 10 ng/ml TGF-β3 induced greater Mmp1 transcription than treatment with TGF-β1 or TGF-β2, in a dose-dependent manner. Western blotting was performed to evaluate Mmp1 protein production in the 10 ng/ml condition. (C) Treatment with 5 and 10 ng/ml TGF-β3 induced less myofibroblast-specific marker Acta2 transcription than treatment with TGF-β1 or TGF-β2, in a dose-dependent manner. Western blotting and flow cytometry were performed to evaluate α-SMA protein expression and the α-SMA⁺ cell population in the 10 ng/ml condition. Results represent three independent experiments using primary human cells and are presented as fold change (means ± s.e.m.) relative to untreated controls (qRT-PCR) or representative images (western blots, flow cytometry plots). Separate control data for the 5 and 10 ng/ml conditions reflect separate qRT-PCR runs from the same experiment. Comparable data were obtained using an immortalized cell line. *P<0.01 for the TGF-β3 condition versus TGF-β1 or TGF-β2, and between doses for each isoform. All TGF-β treatment conditions exhibited significant differences compared with untreated controls for all genes of interest (for clarity of presentation, these comparisons are not denoted by an asterisk).

Fig. 6

TGF-β3 administration during the acute injury phase attenuates vocal fold scar formation in vivo. (A) Schematic showing experimental details. The red arrow indicates timing of vocal fold injury; green arrows indicate timing of TGF-β3 or placebo delivery; black arrows indicate experimental end points. (B) qRT-PCR data showing Emr1, Fn1 and Acta2 transcription in VFM at 72 hours post-injury (24 hours post-final TGF-β3 or placebo injection) in naïve, untreated, placebo-treated and TGF-β3-treated groups. (C) Representative Gomori’s trichrome-stained coronal sections at 60 days post-injury (left) and associated morphometric analysis of lamina propria cross-sectional area, muscle fiber invasion and collagen abundance (right). (D) Representative images showing immunostaining for collagen type III (left; red), Col1a1 and Col3a1 transcription (upper right), and morphometric analysis of collagen type III abundance (lower right), at 60 days post-injury. (E) Representative images showing immunostaining for lysyl oxidase (left; red), Lox transcription (center), and morphometric analysis of lysyl oxidase abundance (right), at 60 days post-injury. Results represent 3–5 independent animals per experimental group and time point and are presented as means ± s.e.m. **P<0.01 for all groups versus naïve, and for the TGF-β3-treated group versus placebo; *P<0.05 for all groups versus naïve, and for the TGF-β3-treated group versus placebo; n.s.=no significant difference.