Knockdown of heme oxygenase-2 impairs corneal epithelial cell wound healing

Abstract

Heme oxygenase (HO) represents an intrinsic cytoprotective system based on its anti-oxidative and anti-inflammatory properties mediated via its products biliverdin/bilirubin and carbon monoxide (CO). We showed that deletion of HO-2 results in impaired corneal wound healing with associated chronic inflammatory complications. This study was undertaken to examine the role of HO activity and the contribution of HO-1 and HO-2 to corneal wound healing in an in vitro epithelial scratch injury model. A scratch wound model was established using human corneal epithelial (HCE) cells. These cells expressed both HO-1 and HO-2 proteins. Injury elicited a rapid and transient increase in HO-1 and HO activity; HO-2 expression was unchanged. Treatment with biliverdin or CORM-A1, a CO donor, accelerated wound closure by 10% at 24 h. Inhibition of HO activity impaired wound closure by more than 50%. However, addition of biliverdin or CORM-A1 reversed the effect of HO inhibition on wound healing. Moreover, knockdown of HO-2 expression, but not HO-1, significantly impaired wound healing. These results indicate that HO activity is required for corneal epithelial cell migration. Inhibition of HO activity impairs wound healing while amplification of its activity restores and accelerates healing. Importantly, HO-2, which is highly expressed in the corneal epithelium, appears to be critical for the wound healing process in the cornea. The mechanisms by which it contributes to cell migration in response to injury may reside in the cytoprotective properties of CO and biliverdin.

Figure 1

HO expression and activity in human corneal epithelial cells. (A) A representative Western blot of HO-1 and HO-2 proteins and (B) densitometry analysis (n=3–4) in primary cultures of human corneal epithelial cells. (C) A representative Western blot of HO-1 and HO-2 proteins and (D) densitometry analysis (n=6) in HCE cells (E) HO activity measured as CrMP-inhibitable CO production in HCE cells 4 h after scratch injury (n=4, *p<0.05 from control uninjured cells)

Figure 2

HO-1 and HO-2 expression in HCE cells. (A) Immunofluorescence assay for HO-1 (left panel) and HO-2 (right panel). Scratch injury was performed and expression of HO-1 and HO-2 determined 1, 2, 4, 8, 24 h after injury with 20X magnification of Zeiss Axioplan-2 fluorescent microscope. (B) Quantitative analysis of HO-1 and HO-2 fluorescence intensity 1, 2, 4, 8 and 24 h after injury (HO-1 n=6, *p<0.05 vs control and **p<0.05 vs 8h wound edge; HO-2 n=4).

Figure 3

Effect of biliverdin and CO on HCE wound healing. Scratch injury was performed in the absence and presence of biliverdin (BVD, 10μM) or CORM-A1 (100 μM) and percent healing was determined at 12 and 24 hr after injury. A) Representative images of wounded HCE cells. B) Quantitative analysis of healing (n=8–12, *p<0.05 vs control untreated cells).

Figure 4

Effect of biliverdin (A) and CORM-A1 (B) concentrations on HCE wound healing 6, 12 and 24 h after injury (n=8–12, *p<0.05, **p<0.01 and ***p<0.001).

Figure 5

Effect of HO inhibition on HCE wound healing. Scratch injury was performed in the absence and presence of CrMP (10 μM) with and without addition of biliverdin (BVD, 10μM) or CORM-A1 (100 μM). A) Representative images. Quantitative analysis of percent healing at 12h (B) and 24h (C) after injury. (D) Effect of CrMP, BVD and CORM-A1 at the concentration indicated above on cell proliferation (n=8–12; *p<0.05 vs CrMP, ^‡p<0.05 from control untreated cells).

Figure 6

Effect of HO inhibition on migration-driven wound healing. Scratch injury was performed in the absence and presence of hydroxyurea (HU, 0.5 mM) with and without addition of CrMP (10 μM), biliverdin (BVD, 10μM) or CORM-A1 (100 μM). Representative images of (A) wound at time 0; (B) wound in the presence of hydroxyurea at 24 h; and (C) wound in the presence of hydroxyurea and CrMP at 24 h. D) Quantitative analysis of percent healing (n=8–12, *p<0.05 vs corresponding time control, ^‡p<0.05 vs HU alone).

Figure 7

Effect of HO-1 and HO-2 siRNA on HCE wound healing. Cells were transfected with control and gene-specific siRNAs. Scratch injury was performed 48 h later and healing determined at 4, 6, 12 and 24 h after injury (n=8–12, *p<0.05 and **p<0.005 vs corresponding control siRNA-treated cells).

Figure 8

Specificity of gene knockdown by siRNA. HCE cells were transfected with control and gene-specific siRNAs. Cells were collected 24 after scratch injury and analyzed for HO-1 and HO-2 mRNA and protein levels. A) Real time PCR analysis of HO-1 and HO-2 mRNA (n=4–6, *p<0.05 vs control untreated cells or control siRNA-treated cells; ^‡p<0.05 vs HO-1). B) Representative Western blots of HO-1 and HO-2 proteins and densitometry analysis (in parenthesis; n=3).