High Glucose Suppresses Epidermal Growth Factor Receptor/Phosphatidylinositol 3-Kinase/Akt Signaling Pathway and Attenuates Corneal Epithelial Wound Healing

Culture media were purchased from Invitrogen (Grand Island, NY). Human recombinant HB-EGF was obtained from R&D Systems (Minneapolis, MN). N-acetylcysteine (NAC), l-buthionine sulfoximine (BSO), and 2, 7-dichlorofluorescein (DCF) were from Sigma-Aldrich (St. Louis, MO). A glutathione assay kit was purchased from Cayman Chemical (Ann Arbor, MI). Rabbit anti-human EGFR, mouse anti-extracellular signal–regulated kinase 2 (ERK2) and phospho (p)-ERK1/2 were from Santa Cruz Biotechnology (Santa Cruz, CA). Rabbit anti-Akt, phospho (p)-Akt (ser⁴⁷³), and phospho (p)-EGFR (Y⁸⁴⁵ and Y¹⁰⁶⁸) were obtained from Cell Signaling (Danvers, MA).

Porcine eyes were obtained from a local abattoir, transported to the laboratory on ice in a moist chamber, and processed for corneal culture the same day. An epithelial wound was made by demarcating an area on the central cornea with a 5-mm trephine and then removing the epithelium within the circle with a small scalpel, leaving an intact basement membrane (29). Corneal-scleral rims were excised and rinsed in sterilized PBS (pH 7.4). The excised corneas were placed epithelial side down into a sterile cup of silicon mold. The endothelial corneal cavity was then filled with minimum essential medium (MEM) containing 1% agarose and 1 mg/ml collagen maintained at 42°C. This mixture was allowed to gel. The cornea, along with its supporting gel, was inverted and transferred to a 35-mm dish. At that point, ∼2 ml of MEM was added drop-wise to the surface of the central cornea until the limbal conjunctiva was covered, leaving the epithelium exposed to the air. The corneas were then cultured in a humidified 5% CO₂ incubator at 37°C in MEM containing 5 mmol/l (normal glucose), 25 mmol/l d-glucose (high glucose), or 5 mmol/l d-glucose plus 20 mmol/l mannitol (high mannitol). The wound was allowed to heal for 48 h in MEM with or without HB-EGF (50 ng/ml) or NAC (25 μmol/l). In our preliminary study, we found that high NAC concentrations (≥50 μmol/l) caused an increase in opacity of cultured corneas, and as a result, 25 μmol/l was chosen for the study. After 48 h culture, the corneas were stained with Richardson's staining solution to mark the remaining wound area, photographed, and quantified using Photoshop software. At least four corneas were used for each treatment, and two independent experiments were performed (eight or more corneas for each treatment).

Porcine corneal epithelial cells, collected 48 h postwounding, were lysed with radioimmunoprecipitation assay (RIPA) buffer with a proteinase inhibitor cocktail, sonicated, and used for determination of protein concentration by the micro-BCA (bicinchoninic acid) method. Cell lysates containing 10 μg of proteins were deproteinized with 10% meta-phosphoric acid, and glutathione was measured in the resultant supernatant by the glutathione assay kit. The concentration of glutathione was obtained in reference to a standard curve.

To assess effects of high glucose on cell signaling ex vivo, porcine corneal epithelial cells were isolated at the end of a 48-h organ culture, lysed in RIPA buffer, and processed for Western blotting.

Primary human corneal epithelial cells (29) and an SV40-immortalized/transformed human corneal epithelial cell line (THCE) (30) were cultured in defined keratinocyte serum-free medium containing normal or high glucose for 48 h and starved overnight in keratinocyte basal medium with the same glucose conditions. Cells on 100-mm dishes were wounded by multiple linear scratches using a cut of 48-well shark tooth comb for DNA sequencing gel (Bio-Rad, Hercules, CA) going from one side of the dish to the other. The dish was then rotated, and scrapes were made similarly to the original scrapes at 45°, 90°, and 135°. At the indicated times, cells were lysed with RIPA buffer and used for Western blotting. Cell lysate of 10–30 μg proteins was used to determine phosphatidylinositol (PI) 3-kinase and ERK activation using antibodies against p-Akt and p-ERK1/2, respectively, with Akt and ERK2 levels for equal protein loading. EGFR tyrosine phosphorylation in wounded THCE cells was analyzed using antibodies against p-EGFR (Y⁸⁴⁵ and Y¹⁰⁶⁸).

Normal and diseased human autopsy corneas obtained from the NDRI (National Disease Research Interchange, Philadelphia, PA) were embedded in OCT (Ted Pella, Redding, CA) on arrival. The OCT blocks were immediately frozen in liquid nitrogen and stored at −80°C. The cryostat sections were used from corneas belonging to 10 diabetic patients (6 with type 1 diabetes, mean age 68.3 ± 4.8 years, and 4 with non–insulin-dependent type 2 diabetes, mean age 75.0 ± 7.5 years) and 4 patients without diabetes and eye disease (referred to as normal healthy; mean age 69.4 ± 4.6 years). The sections, usually three to four in each slide, were stained either with rabbit antibodies against p-Akt or with nonimmune rabbit IgG (negative control), and subsequently with goat anti-rabbit fluorescein-conjugated secondary antibody (Jackson ImmonoResearch Lab, West Grove, PA). The slides were then washed in PBS and mounted with Vectashield mounting medium containing 4′,6-diamidino-2-phenylindole (DAPI) nuclear staining and examined under a Carl Zeiss Axioplan 2 fluorescence microscope equipped with an ApoTome digital camera. Negative controls (rabbit IgG isotypes) exhibited no fluorescence (data not shown).

The level of intracellular ROS was determined on the basis of the oxidative conversion of cell-permeable 2′,7′-dichlorofluorescein diacetate (DCFH-DA) to fluorescent DCF on reaction with hydroxyl radical, hydrogen peroxide, or peroxynitrite. Primary human corneal epithelial cells at passage 3 were seeded onto 6-well plates. After cell attachment and growth for 1 day, the cells were treated with normal glucose, high mannitol, or high glucose in defined keratinocyte SFM for 24 h. Cells were then incubated with DCFH-DA (5 μmol/l) in Hank's balanced salt solution in 37°C for 45 min in the dark, detached from culture wells by 0.05% trypsin-EDTA, and washed twice with ice-cold PBS. Cells without DCFH-DA were used as the negative control; cells treated with BSO (250 μmol/l) to deplete intracellular glutathione were used as the positive control. Flow cytometric measurements (Becton Dickinson FACS DiVa cell sorter) were performed in triplicate for each treatment. The mean fluorescence intensity was quantified by CELLQuest software.

The intracellular accumulation of ROS was also detected in THCE cells plated on eight chamber slides cultured in normal glucose, high mannitol, or high glucose with or without NAC for 24 h by the addition of DCFH-DA to the culture. The distribution of fluorescent DCF on the cell monolayer was visualized under a fluorescence microscope, and the pictures were taken using a Spot digital camera.

Data were presented as the means ± SD. Statistical differences among three or more groups were identified using one-way ANOVA. Differences were considered statistically significant at P < 0.05.

We previously used cultured porcine corneas as an ex vivo model to study epithelial wound healing, and this air/liquid corneal organ culture model has been shown to preserve the structural integrity and the responsiveness of epithelial cells to wounding and growth factor stimulation (29,31,32). To assess whether high glucose influences corneal epithelial wound healing, we cultured porcine corneas in high glucose, using normal glucose or high mannitol as controls. The epithelial wounds were created by epithelial debridement at the center of the cornea and allowed to heal for 48 h. As shown in Fig. 1, a 5-mm epithelial wound was ∼75% covered in the corneas cultured in normal glucose. Interestingly, high glucose significantly impaired epithelial wound closure in cultured porcine corneas (only 29% of the wound was covered) in a manner similar to that observed in cultured human diabetic corneas (8). Importantly, high glucose–induced delay in epithelial wound healing was independent of osmotic changes because 20 mmol/l d-mannitol had minimal effects on the process. We concluded that high glucose impairs corneal epithelial wound closure in cultured corneas.

Glutathione is a major intracellular free radical scavenger, and its intracellular levels have been used as a parameter to measure cellular antioxidant ability (33). We measured the levels of intracellular glutathione in collected porcine corneal epithelial cells. Cells cultured in high glucose for 48 h had significantly reduced cellular glutathione concentration (Fig. 2) compared with those cultured in normal glucose or high mannitol.

We previously showed that wounding elicits EGFR-mediated ERK and PI 3-kinase/Akt activation in corneal epithelial cells (29,34). To elucidate the mechanism underlying high glucose–induced impairment in wound closure, we next investigated the effects of high glucose on EGFR signaling (Fig. 3). In the corneas cultured in both normal and high glucose, wounding stimulated the phosphorylation of ERK1/2. On the other hand, whereas the level of p-Akt in the wounded corneas cultured in normal glucose appeared to be higher than that of the unwounded controls, the level of p-Akt, clearly detectable, in the porcine corneal epithelial cells cultured in high glucose with or without wounding was much lower than in those cultured in normal glucose. The cellular protein levels of Akt and ERK2 were largely the same in all samples tested. Thus, high glucose appears to specifically target PI 3-kinase/Akt in corneal epithelial cells.

Having shown that the EGFR-mediated PI 3-kinase/Akt signaling pathway was altered in porcine corneal epithelial cells cultured in high glucose, we then investigated whether this pathway was also affected by hyperglycemia in the corneas of patients with diabetes. Figure 4 shows representative staining of p-Akt in each diabetes category, within which a similar staining pattern was observed: normal (four subjects), type 1 diabetic (six subjects), type 2 diabetic (three subjects), and type 2 diabetic (diet controlled, one subject). In the basal layer of the corneal epithelium of a normal subject (74 years old), p-Akt staining was strong and continuous, consistent with the metabolically active and proliferative characteristics of the basal layer. A similar staining pattern of p-Akt was noticed in the cornea of a patient with non–insulin-dependent, diet-controlled diabetes (64 years old). However, in a patient with insulin-dependent type 1 diabetes (63 years old with 54 years of diabetes), a few cells at the basal layer were p-Akt–positive, and the staining pattern was patchy and discontinuous. A similar pattern of p-Akt staining was also observed in the cornea of a patient with non–insulin-dependent type 2 diabetes (78 years old with 8 years of diabetes).

To understand the effects of elevated glucose on cell response to wounding, we investigated EGFR signaling in THCE cells cultured in high glucose. In cells cultured in both normal and high glucose, wounding resulted in a rapid increase in the phosphorylation of ERK1/2 and Akt (Fig. 5). However, the intensity of such activation in cells cultured in high glucose was much reduced compared with cells cultured in normal glucose. Furthermore, the activation of ERK and Akt in high glucose was more transient, as indicated by phosphorylation levels that declined to a level similar to those of the unwounded cells at 30 min postwounding versus 2 h postwounding in cells cultured in normal glucose. The low levels of ERK and PI 3-kinase/Akt activation in cells cultured in high glucose appeared to be related to the wound-induced EGFR activation because site-specific phosphorylation of EGFR at Y¹⁰⁶⁸ and Y⁸⁴⁵ declined more quickly in cells cultured in high glucose than in normal glucose. These results suggest that chronic exposure to high glucose could blunt EGFR signaling and its downstream Akt and ERK activation in corneal epithelial cells.

To examine whether high-glucose treatment induces ROS production, we measured intracellular ROS levels in primary human corneal epithelial cells using the redox-sensitive fluorescent dye DCFH-DA and flow cytometry. Figure 6,A shows that high glucose increased the number of cells with high intensity of fluorescence compared with cells cultured in normal glucose or high mannitol. High glucose caused a 40.1% increase in fluorescence intensity over that in normal glucose, indicating oxidative stress on high-glucose treatment (Fig. 6,B). High-glucose–induced generation of ROS in cultured human corneal epithelial cells was confirmed by fluorescent microscopy. Although only a few cells were fluorescent in normal-glucose and high-mannitol cultures, high glucose greatly increased in the number of fluorescent cells after 24 h culture. Furthermore, the production of ROS induced by high glucose was sensitive to antioxidant present in the culture medium because NAC, a nonspecific ROS scavenger, attenuated ROS generation in THCE cells cultured in high glucose for 24 h (Fig. 6 C). HB-EGF, on the other hand, exhibited no effects on fluorescence-positive cells in high-glucose–cultured cells with or without NAC (data not shown).

Having demonstrated that the antioxidant NAC suppresses high-glucose–induced ROS generation in human corneal epithelial cells, we next investigated whether high-glucose–delayed epithelial wound closure could be reversed by antioxidants. We previously showed that HB-EGF accelerates epithelial wound healing in normal corneas (29). Hence, we first tested the effects of HB-EGF and NAC on epithelial wound closure in cultured porcine corneas. Unlike HB-EGF, NAC alone had no effect on wound healing of corneas cultured in either normal or high glucose (data not shown). Although both HB-EGF and NAC noticeably accelerated epithelial wound closure, the rate of closure in high-glucose–cultured corneas in the presence of HB-EGF or NAC was not significantly higher than that in the absence of HB-EGF or NAC (Fig. 7). Remarkably, the presence of both HB-EGF (50 ng/ml) and the antioxidant NAC (25 μmol/l) resulted in significantly accelerated wound healing with a rate similar to that observed in corneas cultured in normal glucose (P < 0.01) (Fig. 7 B). This accelerated wound healing was sensitive to the presence of BSO, which inhibits the generation of intracellular glutathione. Hence, we conclude that EGFR activation alone may not be sufficient for restoring corneal epithelial wound healing impaired by high glucose, and antioxidants may serve as adjunctive therapy to EGFR ligands for treating diabetic keratopathy.

The effects of high glucose on ex vivo corneal epithelial wound healing were dramatic and highly reproducible. To elucidate the underlying mechanisms for antioxidants to affect epithelial wound closure, we assessed the signaling pathways in healing epithelial cells of porcine corneas cultured in normal and high glucose with or without antioxidants. Figure 8 shows that whereas high glucose appeared to increase ERK activation, phosphorylation of Akt was reduced in high-glucose–cultured cells. Addition of HB-EGF or NAC in high-glucose culture media restored partially high-glucose–impaired Akt phosphorylation. Consistent with greater acceleration of high-glucose–impaired wound closure, a combination of HB-EGF and NAC in high-glucose–cultured porcine corneas also enhanced wound-induced Akt phosphorylation to a larger extent than either HB-EGF or NAC acting individually. NAC appeared to have some effects on ERK phosphorylation as well because NAC treatment increased the levels of p-ERK in migrating cells in cultured pig corneas. Taken together, our results suggest that high glucose may affect EGFR signaling, particularly the PI 3-kinase/Akt pathway, through elevated ROS, a process that can be reversed by the antioxidant NAC (35,36).

In this study, we investigated the effects of high glucose on corneal epithelial wound healing. We showed that high glucose delayed epithelial wound closure in cultured porcine corneas and impaired both basal and wound-induced PI 3-kinase/Akt signaling pathway. Hyperglycemia-induced impairment of the PI 3-kinase/Akt pathway was also apparent in the corneal epithelium of diabetic patients, but not in the age-matched nondiabetic control subjects. When human corneal epithelial cells were exposed to high glucose for 48 h, their response to wounding was blunted in terms of the duration and intensity of EGFR phosphorylation and of Akt and ERK activation. The downregulated cell signaling was related to oxidative stress because antioxidant NAC prevented high-glucose–targeted impairment of PI 3-kinase/Akt signaling in cultured corneas. Furthermore, the addition of HB-EGF was unable to overcome the inhibitory effects of high glucose on corneal epithelial wound healing, whereas the presence of both the antioxidant and EGFR ligand greatly accelerated wound healing in cultured porcine corneas. Taken together, we conclude that oxidative stress induced by high glucose weakens EGFR signaling and suppresses the PI 3-kinase/Akt pathway, resulting in the impairment of epithelial wound healing in diabetes corneas.

In the literature, both in vitro and animal models of type 1 and type 2 diabetes have been used to study diabetic corneal complications and wound healing, and a wealth of information has been obtained from these studies (7,9,25,37,,,,–42). For example, hyperglycemia was found to decrease cell proliferation but have no effects on apoptosis and necrosis (43) in the corneal epithelium of streptozotocin-induced diabetic rats, leading to decreased cell density and increased intercellular spaces (44), characteristics of diabetic keratopathy. However, the limitation in tissue size of diseased animals such as rats has hampered the biochemical analysis of hyperglycemia-induced alterations in the cornea. As for the in vitro studies, discrepant results have been reported regarding the effects of high glucose on human corneal epithelial cell proliferation; whereas McDermott et al. (40) reported that high glucose (17.5 mmol/l) increases cell proliferation, Fujita et al. (39) showed a decrease in cell proliferation in the presence of high glucose. Organ culture represents a more relevant experimental approach compared with cultured cells (45,–47). Using human corneas, Kabosova et al. (8) showed that organ-cultured human corneas from diabetic retinopathy patients exhibited much-delayed epithelial wound healing, in complete accordance with clinical data in diabetic patients. Moreover, the corneas of patients in organ culture preserved the same marked abnormalities and altered expression of proteinase and growth factors as in vivo corneas. We previously used cultured porcine corneas as an ex vivo model to study epithelial wound healing and showed that healing an epithelial wound requires EGFR and its downstream signaling pathways, notably ERK and PI 3-kinase/Akt (29). Interestingly, high glucose induced a highly reproducible and marked delay in epithelial wound healing in cultured porcine corneas, similar to that observed in patients and in cultured human corneas (8), validating the porcine corneal organ culture system as a valuable ex vivo model to study the effects of high glucose on corneal wound healing. It should be mentioned that Zagon et al. (48) reported that there was no abnormality in corneal reepithelialization after epithelial debridement within the first 2 weeks of streptozotocin-induced diabetes in rats. The reason for the discrepancy in the results of in vivo (rats) (9) and our ex vivo studies (cultured porcine corneas) is not clear and may be related to the lack of neuronal nourishment in cultured corneas. Indeed, the combination of substance P and IGF-I has been shown to promote delayed corneal epithelial wound healing in diabetic rats (49). The neurological protection against corneal diabetic complications is of great interest and should be further explored in ex vivo studies and in animal models. Whereas the described keratopathy/epitheliopathy can be attributed in part to defects in epithelium–basement membrane adhesion and to diabetic corneal neuropathy (17,18), we propose that many epithelial defects and abnormalities may also be related to alterations in EGFR signaling in the epithelial cells themselves.

Unlike cancerous epithelial cells, there are very few reports regarding EGFR signaling in diabetic tissues/cells. Although a dramatic decrease in hepatic EGFR transcripts was observed during the progression of diabetes (50), type 2 diabetic mice were shown to have upregulated EGFR and elevated EGFR phosphorylation, which were suspected to play a role in kidney enlargement in diabetes (51). In diabetic glomeruli and in mesangial cells treated with high glucose, elevated Akt phosphorylation, through the EGFR/PI 3-kinase pathway, was shown to mediate high-glucose–induced collagen I upregulation (52). On the other hand, high glucose specifically downregulates Akt activities in endothelial cells (53,54) and endothelial progenitor cells (35,36), and disruption of Akt pathways has been linked to endothelial dysfunction and diabetic complications, both macro- and microvascular (55). Thus, the effects of hyperglycemia on the PI 3-kinase/Akt pathway may be tissue-/cell type–specific.

Our study was the first to show that high glucose is specifically targeting PI 3-kinase in corneal epithelial cells. Interestingly, 48-h exposure to high glucose did not significantly affect another major EGFR-mediated signaling pathway, the ERK pathway, in either normal unwounded controls or in healing corneal epithelial cells. Our previous studies revealed that the basal and wound-induced ERK and Akt phosphorylation/activation are EGFR dependent, and those pathways are required for epithelial wound healing in corneal epithelial cells (34). Hence, high-glucose–mediated impairment of the PI 3-kinase/Akt pathway may contribute to delayed wound healing in porcine corneas cultured in high glucose. Most strikingly, we observed a severely altered staining pattern of p-Akt in the corneas of both insulin-dependent and insulin-independent diabetic patients compared with that of age-matched normal subjects. Our data suggest a direct correlation between hyperglycemia and p-Akt expression/distribution in the human corneas. Because PI 3-kinase and Akt have been shown to control diverse cellular activities, including cell survival, growth, proliferation, metabolism, and migration (56), low Akt activity induced by hyperglycemia may be a contributing factor to abnormalities such as basal cell degeneration (6), decreased cell proliferation (9), and, more importantly, the delayed epithelial wound healing (8,9,57) observed in diabetic corneas.

Oxidative stress caused by hyperglycemia is important for the development of diabetic abnormalities (23,24). In addition to being an avascular and transparent tissue, the cornea is unique in that it is constantly exposed to and absorbs most of the harmful UV light. Therefore, a strong, intrinsic defense mechanism against oxidative stress must exist in corneal epithelial cells. Two enzymes, aldehyde dehydrogenase 3 (58,59) and transketolase (60,61), abundantly expressed in corneal epithelial cells have been suggested to function as “corneal crystallins.” They play a protective role against UV-induced oxidative stress through the generation of antioxidant NADPH, the scavenging of ROS, and/or the possession of chaperone-like activity (62). Nevertheless, severe abnormalities, including epithelial cell damage in the corneas of diabetic patients (7,8), suggest the dysfunction of intracellular antioxidant apparatus induced by hyperglycemia. To that end, we used DCFH-DA as a probe for intracellular oxidative stress and showed that high glucose increased levels of DCFH in cultured primary human corneal epithelial cells and in the number of fluorescent cells in THCE cell culture. The notion that high glucose increases in oxidative stress was further affirmed by the detection of a decrease in cellular glutathione levels, an indicator of cellular antioxidant capacity, in porcine corneal epithelial cells cultured in high glucose and by the observation that l-buthionine sulfoximine, an inhibitor of glutathione synthesis, reversed the effects of NAC on restoring high-glucose–delayed epithelial wound healing. The latter observation also suggests that depletion of the intracellular glutathione contents may be an underlying mechanism for hyperglycemia-induced oxidative stress in diabetic cornea. Taken together, our data suggest that oxidative stress may also be a cause of diabetic keratopathy and contribute to the delay of epithelial wound healing in diabetic corneas.

The observation that the antioxidant NAC partially reversed high-glucose–induced suppression of Akt phosphorylation suggests that the generated ROS is a contributing factor for high-glucose–induced disruption of the PI 3-kinase/Akt signaling pathway. As such, it is possible that the Akt pathway may be particularly sensitive to oxidative stress, as shown in culture bovine retinal (63) and human umbilical vein (53) endothelial cells. In the literature, oxidative stress has been shown to target p85 of PI 3-kinase (63), tumor suppressor PTEN (phosphatase and tensin homolog deleted on chromosome 10) (53), or inositol 5-phosphatase SHIP2 (SH2-containing inositol 5′-phosphatase 2) (64), all of which influence Akt activation and signaling. Which molecule(s) in the PI 3-kinase/Akt pathway might be the target of high-glucose–induced oxidative stress in the diabetic cornea is under investigation in our laboratories.

Because delays in epithelial wound healing may result in sight-threatening complications and increase the possibility of microbial infections in diabetic patients with an already weakened immune system (2,57), efforts have been made to identify reagents that accelerate epithelial wound healing in diabetic corneas (57). Growth factors such as EGF (65,66), insulin (9), and antioxidants such as trolox (25) have been tested in animal models and were shown to have limited success in managing diabetic wound healing. Using corneal organ culture, we showed that HB-EGF or NAC alone does not effectively restore the high-glucose–delayed epithelial wound healing. However, treatment of high-glucose–cultured corneas with a combination of antioxidant and HB-EGF restored the rate of wound closure to a level close to that of corneas cultured in normal glucose. Hence, HB-EGF elevating EGFR signaling and antioxidants such as NAC reducing oxidative stress may have synergistic effects on epithelial cells and therefore may prevent or reverse epithelial defects, including compromised barrier function and delayed wound healing, in diabetic corneas. The unique feature of treating diabetic keratopathy is with the use of topical application of therapeutics. Hence, further studies aimed at understanding how aberrant EGFR signaling, including PI 3-kinase/Akt pathways and enhanced production of ROS by hyperglycemia, results in diabetic corneal defects and complications should provide the molecular basis for developing better combinational therapies to treat diabetic corneal complications, including delayed epithelial wound healing.

This study is supported by grants from the National Institutes of Health (R01EY10869 and EY17960 to F.-S.X.Y. and EY13431 to A.V.L.), research grants from Midwest Eye Banks (to K.-P.X.), and an unrestricted grant from Research to Prevent Blindness (Department of Ophthalmology, Wayne State University).

No potential conflicts of interest relevant to this article were reported.

We thank Jenny Huang for editorial assistance.