Myofibroblast dedifferentiation proceeds via distinct transcriptomic and phenotypic transitions

Myofibroblasts are the major cellular source of collagen, and their accumulation — via differentiation from fibroblasts and resistance to apoptosis — is a hallmark of tissue fibrosis. Clearance of myofibroblasts by dedifferentiation and restoration of apoptosis sensitivity has the potential to reverse fibrosis. Prostaglandin E2 (PGE2) and mitogens such as FGF2 have each been shown to dedifferentiate myofibroblasts, but — to our knowledge — the resultant cellular phenotypes have neither been comprehensively characterized or compared. Here, we show that PGE2 elicited dedifferentiation of human lung myofibroblasts via cAMP/PKA, while FGF2 utilized MEK/ERK. The 2 mediators yielded transitional cells with distinct transcriptomes, with FGF2 promoting but PGE2 inhibiting proliferation and survival. The gene expression pattern in fibroblasts isolated from the lungs of mice undergoing resolution of experimental fibrosis resembled that of myofibroblasts treated with PGE2 in vitro. We conclude that myofibroblast dedifferentiation can proceed via distinct programs exemplified by treatment with PGE2 and FGF2, with dedifferentiation occurring in vivo most closely resembling the former.


Total and Small RNA Isolation -CCL210 Fibroblasts
For RNA extraction, each experimental condition was performed in triplicate and cells were harvested in Trizol (Qiagen, Germantown, MD). RNA extraction was performed using the Qiagen miRNeasy Mini kit per the manufacturer's instructions to extract all RNA moieties >18 nucleotides. Genomic DNA was digested on-column per the manufacturer's instructions using the RNase-Free DNase Set (New England Biolabs, Ipswitch, MA). RNA concentration was measured using NanoDrop and RNA integrity was measured using BioAnalyzer (Agilent, Santa Clara, CA) and submitted for library preparation and sequencing.

Total RNA Library Preparation -CCL210 Fibroblasts
RNA was assessed for quality using the TapeStation (Agilent, Santa Clara, CA).

Small RNA Library Preparation -CCL210 Fibroblasts
RNA was assessed for quality using the TapeStation (Agilent) using manufacturer's recommended protocols. Samples were prepared using the NEBNext Multiplex
For differential expression analysis, data were pre-filtered to remove genes with 0 counts in all samples. Differential gene expression analysis was performed using 5 DESeq2 (4), using a negative binomial generalized linear model with cutoffs of linear fold change > 1.5 or < -1.5, Benjamini-Hochberg FDR (Padj) < 0.05. Plots were generated using variations of DESeq2 plotting functions and other packages with R version 3.3.3. For 'totalRNA' results, genes were annotated with NCBI Entrez GeneIDs and text descriptions. For 'totalRNA' results, functional analysis, including KEGG pathway and GO-term enrichments (5), was performed using iPathway Guide (Advaita) (6, 7).

Differential Expression Analysis, Small RNA -CCL210 Fibroblasts
Data was analyzed using the CAP-miRSeq pipeline from the Mayo clinic (8), using human reference genome version hg19 (UCSC), and MirBase version 2 (9). Briefly, FastQC was used to ensure data quality. Reads were trimmed using Cutadapt. The miRDeep2 mapper and miRDeep2 module (10) are the core components for known and novel miRNA detection. Differential miRNA analysis was performed using edgeR (11). TargetScan 7.2 (12) (http://www.targetscan.org/) was used for predicted targets, and MiR-TarBase 8.0 (13) (http://mirtarbase.cuhk.edu.cn/) was used for experimentally validated targets. 6 Young (2 mo) and aged (18 mo) Col1α1-GFP mice were treated with bleomycin by intratracheal administration to induce fibrosis as previously described (14); mice receiving PBS instead of bleomycin were used as controls. Thirty days after bleomycin treatment, at the early stages of fibrosis resolution (14)

Differential Expression Analysis -Mouse Lung Fibroblasts
Analysis of RNAseq data from mouse Col1α1-GFP+ lung fibroblasts, Mayo Clinic's MAPR-Seq software was used to process the raw paired end reads from the RNA sequencing experiments. The raw gene count expression values from MAPR-Seq were then processed by the R package, edgeR, to evaluate differential expression.
Genes with an average raw gene count less than 25 in the samples were excluded from the differential expression analysis. Differentially expressed genes between the young and aged groups were identified using Smyth's moderated t test and Benjamini-Hochberg procedure for adjusted P value (FDR). Genes with a false 8 discovery rate (FDR) below 0.05, and absolute log2 fold change greater than 1 were defined as being differentially expressed.