Limits...
Modeling fibrosis using fibroblasts isolated from scarred rat vocal folds

View Article: PubMed Central - PubMed

ABSTRACT

Following injury, pathologically activated vocal fold fibroblasts (VFFs) can engage in disordered extracellular matrix (ECM) remodeling, leading to VF fibrosis and impaired voice function. Given the importance of scar VFFs to phenotypically appropriate in vitro modeling of VF fibrosis, we pursued detailed characterization of scar VFFs obtained from surgically injured rat VF mucosae, compared to those obtained from experimentally naïve, age-matched tissue. Scar VFFs initially exhibited a myofibroblast phenotype characterized by increased proliferation, increased Col1a1 transcription and collagen, type I synthesis, increased Acta2 transcription and α-smooth muscle actin synthesis, and enhanced contractile function. These features were most distinct at passage 1 (P1); we observed a coalescence of the scar and naïve VFF phenotypes at later passages. An empirical Bayes statistical analysis of the P1 cell transcriptome identified 421 genes that were differentially expressed by scar, compared to naïve, VFFs. These genes were primarily associated with the wound response, ECM regulation, and cell proliferation. Follow-up comparison of P1 scar VFFs and their in vivo tissue source showed substantial transcriptomic differences. Finally, P1 scar VFFs responded to treatment with hepatocyte growth factor and transforming growth factor-β3, two biologics with reported therapeutic value. Despite the practical limitations inherent to working with early passage cells, this experimental model is easily implemented in any suitably equipped laboratory and has the potential to improve the applicability of preclinical VF fibrosis research.

No MeSH data available.


Related in: MedlinePlus

Transcriptome-level comparison of in vitro and in vivo scar models(a) Venn diagrams showing limited overlap in the DE probeset/geneset identified in scar VFFs compared to naïve cells (at P1), and the DE probeset/geneset identified in scarred VF mucosa compared to naïve mucosa. Fourteen genes were identified as DE in both experimental systems. (b) Summary of DE probes/genes identified in direct in vivo versus in vitro comparisons of naïve VFFs and naïve VF mucosa, and scar VFF and scarred VF mucosa. (c) Heat maps showing mean-centered log2-expression data for the 6,346 DE genes identified in naïve VFF versus naïve VF mucosa, and the 6,700 DE genes identified in scar VFF versus scarred VF mucosa. DE genes are ranked by log2 fold change (in vitro normalized to in vivo) along the vertical axis. The annotations indicate a subset of DE genes associated with wound healing, fibrosis and ECM that were highly upregulated in the in vitro, compared to in vivo, condition (log2 fold change > 3) for both naïve and scar comparisons. All experiments were performed with n = 3–5 biological replicates per condition.
© Copyright Policy
Related In: Results  -  Collection


getmorefigures.php?uid=PMC4920689&req=5

Figure 5: Transcriptome-level comparison of in vitro and in vivo scar models(a) Venn diagrams showing limited overlap in the DE probeset/geneset identified in scar VFFs compared to naïve cells (at P1), and the DE probeset/geneset identified in scarred VF mucosa compared to naïve mucosa. Fourteen genes were identified as DE in both experimental systems. (b) Summary of DE probes/genes identified in direct in vivo versus in vitro comparisons of naïve VFFs and naïve VF mucosa, and scar VFF and scarred VF mucosa. (c) Heat maps showing mean-centered log2-expression data for the 6,346 DE genes identified in naïve VFF versus naïve VF mucosa, and the 6,700 DE genes identified in scar VFF versus scarred VF mucosa. DE genes are ranked by log2 fold change (in vitro normalized to in vivo) along the vertical axis. The annotations indicate a subset of DE genes associated with wound healing, fibrosis and ECM that were highly upregulated in the in vitro, compared to in vivo, condition (log2 fold change > 3) for both naïve and scar comparisons. All experiments were performed with n = 3–5 biological replicates per condition.

Mentions: Next, we compared the P1 naïve and scar VFF transcriptomes with previously reported in vivo data from naïve and scarred VF mucosae.28 These in vivo data were obtained using an identical rat strain and age at the time of VF injury, an identical surgical procedure and scar maturation period, and identical sample processing and microarray protocols. Comparisons of probes and genes that were DE in the scar, compared to naïve, conditions in both in vitro and in vivo models showed limited overlap: 18 probes, corresponding to 14 genes, exhibited DE in both models (Figure 5a). These transcriptome-level differences were further emphasized by an analysis of naïve VFF compared to naïve VF mucosa, as well scar VFF compared to scar VF mucosa (Figure 5b). Both analyses revealed a substantial number of DE probes (>10,000 in both naïve and scar comparisons) and genes (>6,000 in both naïve and scar comparisons) across experimental systems. Follow-up evaluation of relative expression levels revealed a number of wound healing, fibrosis and ECM-related genes (including the previously evaluated Col1a1 and Acta2 genes) that were highly upregulated (log2 fold change > 3) in the in vitro, compared to in vivo, condition (Figure 5c). Enrichment analysis of the 6,700 DE genes in the in vitro versus in vivo scar comparison highlighted an array of biological functions consistent with the system-wide repair program and involvement of epithelial and endothelial cells, myocytes, leukocytes, and neurons in vivo (Table S2).


Modeling fibrosis using fibroblasts isolated from scarred rat vocal folds
Transcriptome-level comparison of in vitro and in vivo scar models(a) Venn diagrams showing limited overlap in the DE probeset/geneset identified in scar VFFs compared to naïve cells (at P1), and the DE probeset/geneset identified in scarred VF mucosa compared to naïve mucosa. Fourteen genes were identified as DE in both experimental systems. (b) Summary of DE probes/genes identified in direct in vivo versus in vitro comparisons of naïve VFFs and naïve VF mucosa, and scar VFF and scarred VF mucosa. (c) Heat maps showing mean-centered log2-expression data for the 6,346 DE genes identified in naïve VFF versus naïve VF mucosa, and the 6,700 DE genes identified in scar VFF versus scarred VF mucosa. DE genes are ranked by log2 fold change (in vitro normalized to in vivo) along the vertical axis. The annotations indicate a subset of DE genes associated with wound healing, fibrosis and ECM that were highly upregulated in the in vitro, compared to in vivo, condition (log2 fold change > 3) for both naïve and scar comparisons. All experiments were performed with n = 3–5 biological replicates per condition.
© Copyright Policy
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC4920689&req=5

Figure 5: Transcriptome-level comparison of in vitro and in vivo scar models(a) Venn diagrams showing limited overlap in the DE probeset/geneset identified in scar VFFs compared to naïve cells (at P1), and the DE probeset/geneset identified in scarred VF mucosa compared to naïve mucosa. Fourteen genes were identified as DE in both experimental systems. (b) Summary of DE probes/genes identified in direct in vivo versus in vitro comparisons of naïve VFFs and naïve VF mucosa, and scar VFF and scarred VF mucosa. (c) Heat maps showing mean-centered log2-expression data for the 6,346 DE genes identified in naïve VFF versus naïve VF mucosa, and the 6,700 DE genes identified in scar VFF versus scarred VF mucosa. DE genes are ranked by log2 fold change (in vitro normalized to in vivo) along the vertical axis. The annotations indicate a subset of DE genes associated with wound healing, fibrosis and ECM that were highly upregulated in the in vitro, compared to in vivo, condition (log2 fold change > 3) for both naïve and scar comparisons. All experiments were performed with n = 3–5 biological replicates per condition.
Mentions: Next, we compared the P1 naïve and scar VFF transcriptomes with previously reported in vivo data from naïve and scarred VF mucosae.28 These in vivo data were obtained using an identical rat strain and age at the time of VF injury, an identical surgical procedure and scar maturation period, and identical sample processing and microarray protocols. Comparisons of probes and genes that were DE in the scar, compared to naïve, conditions in both in vitro and in vivo models showed limited overlap: 18 probes, corresponding to 14 genes, exhibited DE in both models (Figure 5a). These transcriptome-level differences were further emphasized by an analysis of naïve VFF compared to naïve VF mucosa, as well scar VFF compared to scar VF mucosa (Figure 5b). Both analyses revealed a substantial number of DE probes (>10,000 in both naïve and scar comparisons) and genes (>6,000 in both naïve and scar comparisons) across experimental systems. Follow-up evaluation of relative expression levels revealed a number of wound healing, fibrosis and ECM-related genes (including the previously evaluated Col1a1 and Acta2 genes) that were highly upregulated (log2 fold change > 3) in the in vitro, compared to in vivo, condition (Figure 5c). Enrichment analysis of the 6,700 DE genes in the in vitro versus in vivo scar comparison highlighted an array of biological functions consistent with the system-wide repair program and involvement of epithelial and endothelial cells, myocytes, leukocytes, and neurons in vivo (Table S2).

View Article: PubMed Central - PubMed

ABSTRACT

Following injury, pathologically activated vocal fold fibroblasts (VFFs) can engage in disordered extracellular matrix (ECM) remodeling, leading to VF fibrosis and impaired voice function. Given the importance of scar VFFs to phenotypically appropriate in vitro modeling of VF fibrosis, we pursued detailed characterization of scar VFFs obtained from surgically injured rat VF mucosae, compared to those obtained from experimentally naïve, age-matched tissue. Scar VFFs initially exhibited a myofibroblast phenotype characterized by increased proliferation, increased Col1a1 transcription and collagen, type I synthesis, increased Acta2 transcription and α-smooth muscle actin synthesis, and enhanced contractile function. These features were most distinct at passage 1 (P1); we observed a coalescence of the scar and naïve VFF phenotypes at later passages. An empirical Bayes statistical analysis of the P1 cell transcriptome identified 421 genes that were differentially expressed by scar, compared to naïve, VFFs. These genes were primarily associated with the wound response, ECM regulation, and cell proliferation. Follow-up comparison of P1 scar VFFs and their in vivo tissue source showed substantial transcriptomic differences. Finally, P1 scar VFFs responded to treatment with hepatocyte growth factor and transforming growth factor-β3, two biologics with reported therapeutic value. Despite the practical limitations inherent to working with early passage cells, this experimental model is easily implemented in any suitably equipped laboratory and has the potential to improve the applicability of preclinical VF fibrosis research.

No MeSH data available.


Related in: MedlinePlus