Topical and oral peroxisome proliferator-activated receptor-α agonist ameliorates diabetic corneal neuropathy

Abstract

Diabetic corneal neuropathy (DCN) is a common diabetic ocular complication with limited treatment options. In this study, we investigated the effects of topical and oral fenofibrate, a peroxisome proliferator-activated receptor-α agonist, on the amelioration of DCN using diabetic mice (n = 120). Ocular surface assessments, corneal nerve and cell imaging analysis, tear proteomics and its associated biological pathways, immuno-histochemistry and western blot on PPARα expression, were studied before and 12 weeks after treatment. At 12 weeks, PPARα expression markedly restored after topical and oral fenofibrate. Topical fenofibrate significantly improved corneal nerve fibre density (CNFD) and tortuosity coefficient. Likewise, oral fenofibrate significantly improved CNFD. Both topical and oral forms significantly improved corneal sensitivity. Additionally, topical and oral fenofibrate significantly alleviated diabetic keratopathy, with fenofibrate eye drops demonstrating earlier therapeutic effects. Both topical and oral fenofibrate significantly increased corneal β-III tubulin expression. Topical fenofibrate reduced neuroinflammation by significantly increasing the levels of nerve growth factor and substance P. It also significantly increased β-III-tubulin and reduced CDC42 mRNA expression in trigeminal ganglions. Proteomic analysis showed that neurotrophin signalling and anti-inflammation reactions were significantly up-regulated after fenofibrate treatment, whether applied topically or orally. This study concluded that both topical and oral fenofibrate ameliorate DCN, while topical fenofibrate significantly reduces neuroinflammation.

Figure 1

Representative corneal section images of mice of different experimental groups showing PPARα immunostaining. Intense PPARα immunosignals were observed in the corneal epithelial layer of non-diabetic mice (NC group) compared to diabetic mice (PC, TF and OF groups). PPARα expression increased after topical and oral fenofibrate treatment in diabetic corneas. Scale bar: 50 μm.

Figure 2

Effects of fenofibrate treatment on CCT for 4 groups at different time points. (A) Mice in the PC group tended to develop neurotrophic keratitis, which manifested as impaired corneal healing and focal corneal ulceration (red arrow). (B) Diabetic mice presented an increased CCT compared to non-diabetic mice at 4 weeks. Asterisk represents p < 0.05, double asterisk represents p < 0.01.

Figure 3

Effects of fenofibrate treatment on corneal sensitivity and corneal nerve metrics for 4 groups at different time points. (A) Representative IVCM images of corneal nerves of 4 groups at different time points. The bar charts showing the (B) corneal sensitivity of 4 groups at different time points. The bar charts showing the (C) CNFL, (D) CNFD, (E) CNBD and (F) tortuosity coefficient of 4 groups at different time points. Asterisk and hash represent p < 0.05, double asterisk represents p < 0.01, triple asterisk represents p < 0.001. Asterisk, double asterisk and triple asterisk represent statistical significance for intergroup comparisons. Hash indicates statistical significance for comparisons across different time points.

Figure 4

Corneal epithelial morphology and CECD for 4 groups at different time points. (A) The morphology of corneal epithelial cells shows restoration of regular cellular polygonal shape, decreased cell size, and enhanced CECD in TF and OF group compared to PC. (B) The bar chart showing that fenofibrate treatment increased CECD in TF and OF groups compared to PC group at 8 weeks and 12 weeks. Asterisk and hash represent p < 0.05, double asterisk represents p < 0.01, triple asterisk represents p < 0.001. Asterisk, double asterisk and triple asterisk represent statistical significance for intergroup comparisons. # indicates statistical significance for comparisons across different time points.

Figure 5

Slit lamp evalution of diabetic keratopathy for 4 groups at different time points. (A) Representative slit-lamp images and (B) bar chart showing that topical and oral fenofibrate treatment significantly improved diabetic keratopathy in diabetic mice. Fenofibrate eye drops were more efficacious than oral treatment in reducing corneal epitheliopathy at 4 weeks and 8 weeks. Hash represents p < 0.05, double asterisk represents p < 0.01, triple asterisk and triple hash represent p < 0.001. Hash and triple hash indicate statistical significance for comparisons across different time points. Double asterisk and triple asterisk represent statistical significance for intergroup comparisons.

Figure 6

β-tubulin staining and quantification of the whole mount corneal sections of 4 groups at different time points. (A) Whole mount corneal sections and (B) bar chart showing that intensity of immunosignals and density of positively stained nerves were significantly lower in diabetic mice (PC, TF and OF groups) compared to non-diabetic mice (NC group), however they increased after both topical and oral fenofibrate treatment. Asterisk represents p < 0.05, double asterisk represents p < 0.01. Asterisk and double asterisk represent statistical significance for intergroup comparisons.

Figure 7

Tear neuromediators profile for 4 groups at different time points. In contrast to oral treatment, topical fenofibrate treatment increased (A) tear NGF levels at 6 weeks and 12 weeks as well as (C) substance P levels at 12 weeks compared to baseline. (B) Fenofibrate treatment did not significantly alter the tear CGRP levels. Hash represents p < 0.05, double hash represents p < 0.01, triple asterisk represents p < 0.001. Hash and double hash indicate statistical significance for comparisons across different time points. Triple asterisk represents statistical significance for intergroup comparisons.

Figure 8

mRNA expression in trigeminal ganglion for 4 groups at different time points. (A) Topical and oral fenofibrate treatment significantly decreased CDC42 mRNA expression at 12 weeks compared to baseline. (B) Topical fenofibrate treatment significantly increased β-III-tubulin (Tuj1) mRNA expression at 12 weeks compared to PC group. No significant difference was observed in (C) NGF and (D) GAP43 mRNA expression between groups at different time points. Asterisk and hash represent p < 0.05, double hash represents p < 0.01, triple hash represents p < 0.001. Asterisk represents statistical significance for intergroup comparisons. Hash, double hash and triple hash indicate statistical significance for comparisons across different time points.

Figure 9

PPARα protein expression and GSEA plots. (A) PPARα protein expression analysis for 4 groups at different time points show that the expression was significantly lower in the corneas of diabetic mice (PC, TF, and OF groups) compared to non-diabetic mice (NC group) at baseline, but it increased with both topical and oral fenofibrate treatment. Asterisk represents p < 0.05, double asterisk represents p < 0.01. Scatter plots for GSEA results [ES and −log10 (p value)] showing the enrichment pathways (p < 0.05) at (B) 12 weeks versus baseline in TF group, and (C) 12 weeks versus baseline in OF group. Red dots represent the significantly up-regulated pathways, and blue dots represent significantly down-regulated pathway.