Upregulation of distinct collagen transcripts in post-surgery scar tissue: a study of conjunctival fibrosis

Abstract

Excessive accumulation of collagen is often used to assess the development of fibrosis. This study aims to identify collagen genes that define fibrosis in the conjunctiva following glaucoma filtration surgery (GFS). Using the mouse model of GFS, we have identified collagen transcripts that were upregulated in the fibrotic phase of wound healing via RNA-seq. The collagen transcripts that were increased the most were encoded by Col8a1, Col11a1 and Col8a2 Further analysis of the Col8a1, Col11a1 and Col8a2 transcripts revealed their increase by 67-, 54- and 18-fold, respectively, in the fibrotic phase, compared with 12-fold for Col1a1, the most commonly evaluated collagen gene for fibrosis. However, only type I collagen was significantly upregulated at the protein level in the fibrotic phase. Type VIII and type I collagens colocalized in fibrous structures and in ACTA2-positive pericytes, and appeared to compensate for each other in expression levels. Type XI collagen showed low colocalization with both type VIII and type I collagens but can be found in association with macrophages. Furthermore, we show that both mouse and human conjunctival fibroblasts expressed elevated levels of the most highly expressed collagen genes in response to TGFβ2 treatment. Importantly, conjunctival tissues from individuals whose GF surgeries have failed due to scarring showed 3.60- and 2.78-fold increases in type VIII and I collagen transcripts, respectively, compared with those from individuals with no prior surgeries. These data demonstrate that distinct collagen transcripts are expressed at high levels in the conjunctiva after surgery and their unique expression profiles may imply differential influences on the fibrotic outcome.

Keywords: Collagen; Conjunctiva; Fibrosis.

Fig. 1.

Upregulated collagen transcripts in the day 2 and day 7 post-experimental GFS transcriptomes. (A) Collagen genes are ranked in descending order from the highest to lowest transcript induction on day 7. Three paired groups at each time point were analysed (n=3). Each paired group consisted of 10 operated tissues from the left eyes pooled together for comparison with 10 unoperated tissues from the contralateral right eyes similarly pooled together. FC, mean fold-change of three groups comparing operated against paired unoperated conjunctival tissues. P values shown have been adjusted using the Benjamini-Hochberg False Discovery Rate (FDR) method. Fold, fold-change between day 7 and day 2 (day 7 FC/day 2 FC); -, P>0.05 (not significant); ECM, extracellular matrix; FACIT, fibril-associated collagens with interrupted helices. (B) The 10 collagen transcripts that were expressed at the highest levels in the day 2- and day 7- operated conjunctival transcriptomes, calculated relative to the unoperated counterparts (n=3). Day 2 expression is represented by white diamonds, whereas day 7 expression is represented by coloured diamonds. Day 2 expression levels that were not significantly different from the respective unoperated baselines are not presented in the plot.

Fig. 2.

Verification of collagen transcripts that were upregulated the most in the late phase of conjunctival wound healing by qPCR. (A) Real-time PCR analyses of the 10 most highly expressed collagen genes in mouse conjunctival tissues harvested 2, 7, 14 and 21 days after experimental surgery. Data are calculated as fold-change relative to paired unoperated controls. The mean fold-change of five groups, each consisting of operated tissues pooled from three mice, for each time point is indicated (n=5). *P<0.05 for fold-change compared with all other time points; ^δP<0.05 for fold-change on day 7 compared with both days 14 and 21; ^ψP<0.05 for fold-change on day 7 compared with day 21 only. (B) Immunoblot analyses of COL8A1, COL11A1 and COL1A1 in mouse conjunctival tissues harvested on day 2 and day 7 post-experimental surgery. Three paired groups for each time point are shown (n=3). Op, operated tissues pooled from five independent eyes per group; C, paired untreated controls pooled similarly. Fold-change in expression in operated relative to control tissues, both normalized to GAPDH, and the associated P values corrected by Bonferroni adjustment, where significant, are shown below the respective immunoblot.

Fig. 3.

Immunolocalization of COL8A1, COL11A1 and COL1A1 in the day 7-operated mouse conjunctival tissue. (A) Co-immunolabelling of COL8A1 (red), COL11A1 (green) and COL1A1 (magenta). Upper panels, insets (i) show magnified images of the boxed areas in the first and third images, with yellow and white arrowheads indicating high COL8A1/low COL1A1 and low COL8A1/high COL1A1 co-immunolabellings, respectively; inset (ii) left, second image, shows a magnified image of the boxed area co-immunolabelled for COL8A1 and COL11A1; inset (ii) right, second image, shows a magnified image of the boxed area co-immunolabelled for COL11A1 and COL1A1; inset (ii), fourth image, shows a magnified overlay image of the boxed area co-immunolabelled for all three collagens. None of the insets includes DAPI staining. Lower panels, a section from the same eye incubated with only the secondary antibodies used in the upper panel showed minimum non-specific staining in the bleb area in comparison with non-specific staining, which was observable in the choroid (arrow). S, sclera. (B) Co-immunolocalization of COL8A1 (green) and COL1A1 (magenta) with ACTA2 (red). Yellow and white arrowheads in insets indicate COL8A1 and COL1A1 co-immunolabelling patterns, as indicated in A. Co-immunolabelling of ACTA2 with COL8A1 (left inset, second image) or COL1A1 (right inset, second image) appears yellow or blue, respectively. Co-immunolabelling of COL8A1 and COL1A1 appears white (right inset, third image). None of the insets includes DAPI staining. (C) Co-immunolocalization of COL11A1 (green) with F4/80+ cells (red). Inset, fourth image, shows a magnified overlay image of the boxed area co-immunolabelled for COL11A1 and F4/80. All pictures were captured by confocal microscopy. Scale bars: 100 μm.

Fig. 4.

Induction of top collagen genes by TGFβ2 in primary mouse conjunctival fibroblasts. (A) Real-time PCR analyses of collagen genes in fibroblasts stimulated for 72 h with TGFβ2. Data shown are representative of three independent experiments, and show mean fold-change±s.d. relative to untreated controls. *P<0.05 for fold-change in treated versus control. (B) Immunoblot analyses of COL8A1, COL11A1 and COL1A1 in mouse fibroblasts treated as indicated for 72 h. Three independent sets of experiments are shown. Fold-change in expression in TGFβ2-treated relative to control cells, both normalized to GAPDH, and P value (corrected by Bonferroni adjustment) where significant, are shown in the densitometric analyses below the immunoblot.

Fig. 5.

Induction of top collagen genes by TGFβ2 in primary human conjunctival fibroblasts and in individuals with surgical failure. (A) Real-time PCR analyses of COL8A1, COL11A1, COL8A2 and COL1A1 in primary human conjunctival fibroblasts stimulated with TGFβ2 for 72 h. Data shown are representative of three independent experiments, and calculated as mean fold-change±s.d. relative to untreated controls. *P<0.05 for fold-change in treated versus control. (B) Immunoblot analyses of COL8A1, COL11A1 and COL1A1 in human fibroblasts treated as indicated for 72 h. Three independent sets of experiments are shown. Fold-change in expression in TGFβ2-treated relative to control cells, both normalized to GAPDH, and P values (corrected by Bonferroni adjustment) where significant, are shown in the densitometric analyses below the immunoblot. (C) Real-time PCR analyses of COL8A1 and COL1A1 in human conjunctival tissues from individuals requiring repeat surgeries (n=15) relative to those with no prior GFS (n=20). Data for COL8A1 in tissues of individuals with no prior surgeries was only available from 15 patients as five samples did not produce detectable values. Data shown are calculated as fold-change in expression in tissues of repeat surgery patients relative to that from patients with no prior surgery. Significant fold-changes between the two groups of patients and the associated P values (*) are indicated. Each symbol represents one patient.