Prevention of diabetic keratopathy

Patient morbidity related to diabetic induced ocular complications has increased year on year commensurate with the worldwide increase in the incidence of diabetes. These complications include retinopathy, neovascular glaucoma, optic neuropathy, keratopathy, and dry eye. Diabetic retinopathy, because of its clinical importance as a leading cause of blindness, has attracted the major thrust of both clinical and basic research. Clinical ophthalmological management of this condition now routinely includes photocoagulation and vitreoretinal surgery. Various systemic and local medications are now also being extensively examined both through basic research and clinical trials to determine their clinical efficacy in managing the complications of diabetic retinopathy.

Diabetic keratopathy has featured as the “poor relation” with regard to both clinical and research interest. The condition is not thought to represent a serious clinical or pathological entity and hence has been overlooked by both physician and scientist alike. Yet with only cursory investigation it is obvious that many patients have visual loss secondary to diabetic keratopathy. Diabetic keratopathy comprises several symptomatic corneal conditions inducing superficial punctate keratopathy and persistent corneal epithelial erosion.¹ The latter can be encountered especially after vitreoretinal surgery, where oedematous and cloudy corneal epithelium, often manually removed to restore clarity, results postoperatively in a poorly healing corneal epithelial surface. De novo epithelial erosion in diabetic patients can often be resistant to routine clinical management of corneal erosions including topical medication and bandage contact lenses. These poorly healing epithelial surfaces have compromised defences against general microbial attack, predisposing these patients to bacterial and fungal infective keratopathies.

Keratopathy in the presence of diabetes should be considered as a potential sight threatening condition and thence must be given appropriate clinical attention and increased research interest. For this reason, it is important to attempt to analyse the mechanism of diabetic keratopathy and from this, hopefully establish improved techniques to prevent and treat the condition. Before symptomatic diabetic corneal complications, subclinical abnormalities can develop in diabetic corneas; these include a decrease in epithelial barrier function,^2–4 abnormalities in shape of epithelial and endothelial cells,^5–9 basement membrane thickening,^10,11 and decreased corneal sensation.^12–15 These subclinical abnormalities, however, have often a close temporal relation to the development of symptomatic corneal conditions in diabetes. Several molecular mechanisms are thought to exist, which are related to and may underpin the development of diabetic keratopathy.

It may now be possible to use this simple organ, the cornea, with all its inherent advantages of accessibility, clarity, ease of observation, and lack of cellular complexity to investigate diabetic pathology secondary to increases in polyol pathway and deposition of AGEs

Firstly, an increase in the polyol metabolism in the corneal epithelial cells is reported as a mechanism of diabetic keratopathy.^16,17 There is a strong similarity in the spatial distribution of aldose reductase, an enzyme entry into polyol pathway and the target organs affected by typical diabetic pathology including kidney and blood vessels.^16,17 Akagi et al reported the accumulation of polyol and the expression of aldose reductase in the corneal epithelium and endothelium.¹⁶ These data are consistent with the clinical findings that the corneal epithelium and endothelium are targets of diabetic complications. The association between diabetes and the polyol pathway inducing corneal changes are further demonstrated using a galactose fed animal model where significant increases in polyol accumulation were noted within the corneal epithelium and endothelium.^17–19 Furthermore, inhibition of aldose reductase activity using aldose reductase inhibitor (ARI) ameliorates corneal changes in both diabetic and galactose fed animals models. In these models, ARI was effective in inhibiting the loss of corneal sensation,¹² delaying corneal epithelial wound healing,²⁰ enlargement of epithelial and endothelial cell size,^6,7,21 breakdown of corneal epithelial barrier function,⁴ and accumulation of polyol.¹⁹

Although there are a lot of anti-diabetic drugs effective in diabetic animal experiments, few of them have proved efficacy in human studies. ARI treatment, however, has been shown (although only in uncontrolled case studies) to ameliorate corneal changes in diabetic patients.^8,20,22 In a controlled study using topical ARI treatment Hosotani et al have demonstrated an ameliorative effect upon the enlargement of the corneal epithelial cells in diabetic patients.⁹ The study in this issue of the BJO by Nakahara et al (p 266) is now the second controlled study dealing with the effect of ARI treatment on diabetic keratopathy. In this issue, the authors have shown that topical ARI treatment was effective in the restoration of corneal epithelial barrier function, but not in the prevention of superficial punctate keratopathy. These results appear to indicate that there may be different mechanisms implicated in the breakdown of the corneal epithelial barrier function and the development of superficial punctate keratopathy.

Decrease in the corneal sensation²³ and loss of nerve derived trophic factor have been postulated as causative factors in the development of diabetic keratopathy. Nakamura et al have revealed that insulin-like growth factor 1 (IGF-1) and substance P, a neuropeptide present in sensory nerves, accelerate corneal epithelial wound healing.²⁴ In addition, the authors showed that topical application of substance P and IGF-1 accelerated the corneal epithelial wound healing process in diabetic animals. These studies help to strengthen the potential pathogenic link between decreased corneal sensation and diabetic keratopathy.

Other putative causes of diabetic keratopathy, in addition to enzymatic and neural dysregulations, include structural abnormalities in the corneal epithelium basement membrane.^10,25–27 Kenyon et al were the first to highlight the abnormal interaction of the corneal epithelium and basement membrane.²⁷ They showed that corneal epithelial basement membrane in addition to corneal epithelium was removed with manual epithelial removal during vitreoretinal surgery. For this reason, they speculated that bare corneal stroma, without basement membrane, after corneal epithelial abrasion was the reason for a delay in corneal epithelial wound healing.²⁷ Histologically, thickening and multilamination of the basement membrane²⁵ and a decrease in the penetration of anchoring fibrils (type VII collagen)¹⁰ were noted in diabetic corneas. These structural changes of the basement membrane in diabetic cornea may account for the loose attachment of corneal epithelial cells.

Advanced glycation end products (AGEs) have been implicated in the development of diabetic keratopathy and maybe at least partly explain some of the structural changes noted.^26,28 AGEs are known to deposit in the basement membrane of the corneal epithelial cells of diabetic patients.²⁶ When this happens the molecular structure of basement membrane components changes and they lose adhesive property. In this way, the corneal epithelial cells lose a clue for the attachment on the basement membrane. In addition, aminoguanidine, an antioxidant, was effective in inhibiting AGE formation and thus ameliorated the attachment of corneal epithelial cells to the basement membrane.²⁶ However, the in vivo effect of aminoguanidine on diabetic keratopathy remains unknown.

This review has alluded to several common molecular mechanisms previously implicated in the pathogenesis of systemic diabetic complications, and now also implicated in the pathogenesis of diabetic keratopathy. Potentially, diabetic keratopathy provides a pathogenic mechanistic model to shed light upon complications within other more complex organs. The value of using such a simple model as the cornea to shed light on complications within structurally much more complex organs, has previously already been elegantly demonstrated by investigators such as Gimbrone et al.²⁹ It may now be possible to use this simple organ, the cornea, with all its inherent advantages of accessibility, clarity, ease of observation, and lack of cellular complexity to investigate diabetic pathology secondary to increases in polyol pathway and deposition of AGEs.

I think that the potential value of diabetic keratopathy as a simplistic model of diabetic complications cannot be overstated. For this reason, I postulate that the model should be considered for adoption throughout diabetic research laboratories and within institutions performing double blinded clinical studies determining the effect of novel treatments upon systemic diabetic complications.