Role of the C Domain of IGFs in Synergistic Promotion, with a Substance P–Derived Peptide, of Rabbit Corneal Epithelial Wound Healing

Abstract

purpose. Insulin-like growth factors (IGFs) and either substance P (SP) or an SP-derived peptide (FGLM-amide) synergistically facilitate corneal epithelial wound healing in vitro and in vivo. The mechanism of this synergism and the clinical potential of these agents were further investigated by determination of the relevant functional domain of IGFs.

methods. The effects of IGF-derived peptides on corneal epithelial cell migration were evaluated with the rabbit cornea in an organ culture system. Corneal epithelial wound closure in vivo was also evaluated in rabbits after epithelial debridement with n-heptanol.

results. In the presence of FGLM-amide, peptides corresponding to the C domain of IGF-1 or -2 significantly promoted corneal epithelial migration in vitro to an extent similar to that apparent with the full-length molecules. In contrast, peptides corresponding to the D domain of these growth factors had no such effect. Mutation of serine-34 in the C domain of IGF-1 to alanine abolished the synergistic effect with FGLM-amide on corneal epithelial migration. The C peptide of proinsulin did not affect corneal epithelial migration in the absence or presence of FGLM-amide. The administration of eye drops containing both the C-domain peptide of IGF-1 and FGLM-amide significantly promoted corneal epithelial wound closure in vivo.

conclusions. The C domain of IGF-1 or -2, for which no biological function has previously been identified, is essential for the synergistic effect of these growth factors with SP on corneal epithelial migration.

Wound healing is fundamental to the survival of an organism. The healing process is complex and comprises a series of events in which humoral and neural factors, extracellular matrix, and various cell types participate at the appropriate time. ¹ Epithelialization is one of the most critical steps of this process. We have studied the regulation of corneal epithelial wound healing in animals, both in vitro ^{2 3 4 5 6 7} and in vivo ^{8 9} as well as in humans clinically. ^{10 11} The cornea possesses a relatively simple structure and is avascular. It is one of the most sensitive tissues in the body, with the density of nerve endings in the cornea being 300 to 400 times that in the skin. These characteristics render the cornea amenable to the study of epithelial wound healing or of epithelial migration per se. We have developed an organ culture system to study the cornea, with which it is possible to evaluate epithelial migration quantitatively. ² With this system, we have previously shown that insulin-like growth factor (IGF)-1 acts synergistically with substance P (SP) to promote corneal epithelial migration in vitro. ¹² The combination of IGF-1 and SP also synergistically facilitates the closure of rabbit corneal epithelial wounds in vivo. ¹³ Both in vitro and in vivo studies demonstrated that SP exerts its synergistic action with IGF-1 through interaction with the neurokinin receptor NK-1. ¹⁴ The carboxyl terminal four amino acids of SP (Phe-Gly-Leu-Met-amide, or FGLM-amide) are sufficient for this effect. ¹⁵ Furthermore, we showed that the persistent corneal epithelial defects of patients with neurotrophic keratopathy are successfully treated by the administration of eye drops containing IGF-1 and either SP or FGLM-amide. ^{16 17 18}  

IGF-1 and -2 each comprise four domains: the A, B, C, and D domains (Fig. 1) . These growth factors exert various biological effects as a result of their interaction with IGF and insulin receptors. ¹⁹ IGFs and insulin share many structural similarities, especially in the A and B domains. ²⁰ However, the C domain of IGFs does not share sequence homology with the C peptide of proinsulin, which is not retained in mature insulin. In addition, IGFs possess the D domain at their carboxyl termini, whereas insulin has no counterpart sequence (Fig. 1) . IGF-1 binds with highest affinity to the IGF type 1 receptor, which is structurally homologous to the insulin receptor. IGF-2 also binds the IGF type 1 receptor but with lower affinity. In contrast, the IGF type 2 receptor binds IGF-2 with higher affinity and IGF-1 with lower affinity. Both IGF-1 and -2 also bind to the insulin receptor with low affinity. ²¹ No biological function has been attributed to the C domain of IGFs, although this domain modulates the interaction of the A and B domains with insulin and IGF receptors. ^{22 23 24}  

The synergistic effect of IGF-1 with SP on corneal epithelial migration is mimicked by IGF-2 but not by insulin, ²⁵ suggesting that the C or D domain of IGFs plays an important role in this action. To provide greater insight into the mechanism underlying the synergism of IGFs and SP, we have now investigated which domain of IGF-1 or -2 is responsible for the effects of these growth factors with FGLM-amide on rabbit corneal epithelial migration both in organ culture and in vivo. Our results demonstrate that the C domain of IGF-1 or -2, but not the D domain of these molecules, is responsible for the synergism with SP in the promotion of corneal epithelial migration. 

Methods

Materials

Human recombinant IGF-1 and -2 were obtained from BD Biosciences (Bedford, MA), and human recombinant insulin and C peptide of proinsulin were from Sigma-Aldrich (St. Louis, MO) and Biogenesis (Poole, UK), respectively. The C-domain (GYGSSSRRAPQT) and D-domain (PLKPAKSA) peptides of IGF-1, the C-domain (SRVSRRSR) and D-domain (TPAKSE) peptides of IGF-2, and FGLM-amide were synthesized by the Peptide Institute (Osaka, Japan). Human recombinant IGF-1 was also expressed in and purified from Escherichia coli as a fusion protein of Staphylococcus protein A encoded by the pEZZ 18 vector (Amersham, Piscataway, NJ), as described previously. ²⁶ Mature IGF-1 was released by hydroxylamine cleavage and purified by immunoglobulin-G–based affinity chromatography. A cDNA encoding the IGF-1(S34A) mutant was generated by a polymerase chain reaction–based strategy with Pfu polymerase (Promega, Madison, WI). The mutation was verified by nucleotide sequencing. IGF-1(S34A) was expressed and purified as described for the wild-type protein. 

Cell Culture

A cornea of a 65-year-old white man was obtained from Northwest Lions Eye Bank (Seattle, WA). The tissue was used in accordance with the tenets of the Declaration of Helsinki. Corneal fibroblasts were prepared from the tissue remaining after corneal transplantation surgery and were cultured as described previously. ²⁷ The cells were maintained in Dulbecco’s modified Eagle’s medium supplemented with 10% heat-inactivated fetal bovine serum until they had achieved approximately 90% confluence in 100-mm culture dishes. Cells were subjected to experiments in the third to fifth passage. 

Animals

Albino rabbits (body mass, 2–3 kg) were obtained from Kitayama Labs (Kyoto, Japan) and KBT Oriental (Saga, Japan). Their care and treatment conformed to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Rabbit Corneal Epithelial Migration Assay

The length of the path of epithelial migration over the stroma exposed on the sides of a block of cultured rabbit cornea was measured as described previously. ² In brief, rabbits were killed with an overdose of pentobarbital sodium injected intravenously, and both eyes were enucleated. The sclerocorneal rim was cut, and the cornea was excised and washed several times with sterile phosphate-buffered saline (PBS). Six blocks (each ∼2 × 4 mm) were cut from each cornea with a razor blade. The size of the corneal blocks was previously shown not to affect the rate of epithelial migration. ²⁸ Each corneal block was placed with the endothelial side up in a well of a 24-well tissue culture plate with unsupplemented Medium199 (control) or Medium199 containing the various test agents. Each treatment group consisted of three blocks. After incubation for 24 hours at 37°C under a humidified atmosphere of 5% CO₂ in air, the blocks were fixed overnight at 4°C with a mixture of glacial acetic acid and absolute ethanol (5:95, vol/vol). They were then dehydrated by exposure to a graded series of ethanol solutions, immersed in chloroform, and embedded in paraffin. Four thin sections (7 μm) were cut at 200-μm intervals from each block and, after the removal of paraffin, stained with hematoxylin-eosin. The sections were examined with a light microscope and photographed, and the length of the path of corneal epithelial migration down both sides of each section (toward the endothelial side) was measured on the micrographs with a computer-assisted digitizer. Preliminary experiments showed that the lengths of the paths of epithelial migration down each side of a corneal block are independent of each other. We therefore averaged the results obtained from the four sections for each side of each block separately. Data are expressed as the mean ± SE of the six determinations (one averaged value for each side of each of the three blocks in a treatment group). 

Immunoblot Analysis of Akt

Human corneal fibroblasts (2 × 10⁶) were cultured in 35-mm dishes first for 48 hours in culture medium containing 0.5% fetal bovine serum and then for 1.5 hours in serum-free medium. They were then incubated for 30 minutes at 37°C in serum-free medium containing test agents, washed with ice-cold PBS, and lysed in 100 μL of a solution containing 2% SDS, 150 mM NaCl, 2 mM NaF, 2 mM Na₃VO₄, 5 mM EDTA, and 20 mM Tris-HCl (pH 7.5). Protein concentration was determined with a protein assay kit (Dc; Bio-Rad, Hercules, CA). Cell lysates (38 μg of protein) were fractionated by SDS–polyacrylamide gel electrophoresis and subjected to immunoblot analysis with antibodies to Akt or to phospho-Akt (Cell Signaling, Beverly, MA). Immune complexes were detected with horseradish-peroxidase–conjugated secondary antibodies and enhanced chemiluminescence reagents (Amersham). 

Assay of Rabbit Corneal Epithelial Wound Healing

Rabbits were anesthetized with an intramuscular injection of ketamine and xylazine as well as with oxybuprocaine eye drops. The corneal epithelium of each eye was wounded with n-heptanol as described. ²⁹ The animals were divided into four groups of four and were treated with eye drops containing either 100 nM C-domain peptide of IGF-1, 1 mM FGLM-amide plus 100 nM IGF-1, 1 mM FGLM-amide plus 100 nM C-domain peptide, or PBS vehicle alone (control). Each eye received 1 drop of the respective solution immediately after n-heptanol treatment and again 2, 4, 6, 8, 10, 24, 26, 28, 30, 32, and 34 hours later. The epithelial defects were stained with 1 drop of 2% fluorescein and photographed immediately after wounding as well as 6, 12, 18, 24, 30, 36, and 48 hours later. The area of each epithelial defect was measured on the photographs with a computer-assisted digitizer. The healing rate of each eye was calculated by linear regression analysis of the data collected at 6, 12, 18, 24, and 30 hours. 

Statistical Analysis

Statistical analysis was performed with the unpaired Student’s t-test for comparison of two groups or with the Tukey-Kramer multiple comparison test for comparison of three or more groups. P < 0.05 was considered statistically significant. 

Results

Facilitation of Corneal Epithelial Migration by the C Domain of IGF-1

We first examined the effects of the C and D domains of IGF-1 in the presence of the SP-derived peptide FGLM-amide on corneal epithelial migration in an organ culture system (Fig. 2) . Incubation of corneal blocks with FGLM-amide (20 μM) alone had no significant effect on the length of the path of epithelial migration, as described previously. ¹⁵ Furthermore, neither full-length IGF-1 nor peptides corresponding to the C or D domains of IGF-1 (1 nM) affected corneal epithelial migration in the absence of FGLM-amide. In the presence of FGLM-amide, however, full-length IGF-1 and the C-domain peptide, but not the D-domain peptide, significantly increased the extent of epithelial migration. Essentially identical results were obtained with SP in place of FGLM-amide (data not shown). This effect of the C-domain peptide was dose dependent, reaching a plateau at a concentration of approximately 1 nM in the presence of 20 μM FGLM-amide (data not shown). 

To confirm the role of the C domain of IGF-1 in the synergistic promotion of corneal epithelial migration with SP, we prepared a mutant IGF-1 protein, IGF-1(S34A), in which serine-34 of the C domain was replaced by alanine. IGF-1(S34A) did not act synergistically with FGLM-amide to promote corneal epithelial migration (Fig. 3) . The C domain, but not the D domain, of IGF-1 thus appeared responsible for the synergism with SP in the promotion of corneal epithelial migration. 

Elimination of the IGF Type 1 Receptor as a Mediator of Signaling by the C Domain of IGF-1

We examined whether the action of the C domain of IGF-1 is mediated by the IGF type 1 receptor. We have shown that the interaction of IGF-1 with the type 1 receptor in human corneal fibroblasts results in the activation of phosphoinositide 3-kinase and the consequent phosphorylation of Akt (protein kinase B). ³⁰ FGLM-amide alone had no effect on Akt phosphorylation (Fig. 4) . Whereas IGF-1 at concentrations as low as 1 nM induced phosphorylation of Akt in a manner independent of FGLM-amide, the C-domain peptide of IGF-1 failed to induce Akt phosphorylation at concentrations up to 100 nM in the absence or presence of FGLM-amide (Fig. 4 , data not shown). These results thus indicate that the C domain of IGF-1 does not activate the IGF type 1 receptor and that the activation of this receptor is therefore not responsible for the synergistic effect of IGF-1 with SP on corneal epithelial migration. 

Facilitation of Corneal Epithelial Migration by the C Domain of IGF-2

We next examined whether the C domain of IGF-2, like that of IGF-1, exhibits synergistic activity with SP in the promotion of corneal epithelial migration. In the presence of FGLM-amide, the length of the path of epithelial migration was increased significantly by either full-length IGF-2 or the C domain of this growth factor, but not by the D domain (Fig. 5) . These results were thus consistent with those obtained with IGF-1. 

Lack of Effect of Insulin or the C Peptide of Proinsulin on Corneal Epithelial Migration

Proinsulin is processed posttranslationally by proteases to yield mature insulin and C peptide. To determine whether the C peptide of proinsulin exerts a synergistic effect with SP on corneal epithelial migration similar to that of the C domain of IGF-1 or -2, we added insulin or C peptide, with or without FGLM-amide, to the rabbit cornea in organ culture. Neither insulin nor C peptide affected corneal epithelial migration in the absence or presence of FGLM-amide (Fig. 6) . 

Promotion of Corneal Wound Healing In Vivo by the C Domain of IGF-1 and FGLM-Amide

Finally, we examined whether the C-domain peptide of IGF-1 was able to replace the full-length molecule in the synergistic facilitation with FGLM-amide of rabbit corneal epithelial wound healing in vivo. Twenty-four hours after epithelial wounding, the area of epithelial defects in rabbits treated with eye drops containing either IGF-1 plus FGLM-amide or the C-domain peptide of IGF-1 plus FGLM-amide was markedly smaller than that in animals treated with PBS (vehicle) or the C-domain peptide alone (Fig. 7A) . This effect of the combination of IGF-1 or the C-domain peptide with FGLM-amide was time dependent, being significant at 12, 24, and 36 hours after the onset of treatment (Fig. 7B) . There was no significant difference in the size of the initial defect among the various treatment groups. The mean healing rates were 0.992 ± 0.050, 0.945 ± 0.058, 1.205 ± 0.055, and 1.255 ± 0.053 mm² · h⁻¹ for animals treated with PBS, the C-domain peptide of IGF-1 alone, IGF-1 plus FGLM-amide, and the C-domain peptide plus FGLM-amide, respectively. These results indicate that the C-domain peptide of IGF-1 can act as a substitute for the full-length molecule in synergistic facilitation with FGLM-amide of corneal wound healing. 

Discussion

We have shown that peptides corresponding to the C domain of IGF-1 or -2 possess biological activity, in that, together with SP or FGLM-amide, they promote corneal epithelial migration in vitro and wound healing in vivo. The C-domain peptides of both IGF-1 and -2 were maximally effective at concentrations as low as 1 nM, at which the full-length molecules also promote corneal epithelial migration in the presence of SP or FGLM-amide. However, neither insulin nor the C peptide of proinsulin exerted such a synergistic effect with FGLM-amide on corneal epithelial cell migration. 

IGFs exhibit various biological activities including the stimulation of cell proliferation and the inhibition of apoptosis. ³¹ No function has been attributed to the C domain of these growth factors, however, although this domain modulates the interaction of the A and B domains of IGF-1 with insulin and IGF receptors. ^{22 23 24} We have now shown that, like the full-length molecules, synthetic peptides corresponding to the C domain of IGF-1 or -2 facilitate epithelial migration in the presence of FGLM-amide. Our data thus suggest that the synergistic action of IGFs with SP or FGLM-amide in epithelial migration is distinct from other biological functions of IGFs, in that it is mediated by the C domain. Our observation that the IGF-1(S34A) mutant, in which the second serine residue of the C domain of IGF-1 is replaced by alanine, failed to promote corneal epithelial migration in the presence of FGLM-amide further confirms the role of the C domain of IGF-1 in this synergistic action. The mutant protein retained the ability to stimulate the proliferation of 3T3 mouse fibroblasts through its interaction with the IGF type 1 receptor (Yanai R, Inui M, unpublished observation, 2003). Further studies are necessary to determine the sequence within the C domain of IGF-1 or -2 required for the synergistic action with SP in corneal epithelial migration. 

The distinct nature of the synergistic action of IGF-1 with SP in epithelial migration compared with other biological functions of IGF-1 was further demonstrated by our analysis of intracellular signaling mediated by the IGF type 1 receptor. The C domain of IGF-1 thus failed to activate Akt in human corneal fibroblasts in the absence or presence of FGLM-amide, whereas full-length IGF-1 induced Akt phosphorylation by binding to the IGF type 1 receptor, an effect that is mediated by the activation of phosphoinositide 3-kinase. ³² The synergistic action of IGF-1 with SP in corneal epithelial migration thus does not appear to be mediated by the IGF type 1 receptor. 

Persistent corneal epithelial defects in humans can be successfully treated by the administration of eye drops containing IGF-1 and SP. ^{18 33} FGLM-amide is able to substitute for SP in the promotion together with IGF-1 of corneal epithelial wound healing. ^{15 16 17} We have now shown that IGF-1 can be replaced by the C domain of IGF-1 or -2. Treatment with peptides rather than the corresponding full-length molecules is advantageous in the clinical setting, because it is usually associated with fewer unfavorable biological activities. Indeed, FGLM-amide lacks the ability to induce contraction of the pupil (miosis), which is one of the pharmacological actions of SP. ³⁴ Given that the C domain of IGF-1 does not reproduce actions of IGF-1 mediated by the IGF type 1 receptor, the administration of this peptide might be expected not to induce unfavorable effects such as hyperplasia or neovascularization in the cornea. Our results thus provide the basis for development of a new strategy in the treatment of corneal wounds. 

 Supported by grants from the Japan Society for the Promotion of Science (JSPS) and the Ministry of Education, Culture, Sports, Science, and Technology of Japan (MI, TN).

 Submitted for publication June 20, 2003; revised October 16, 2003; accepted November 24, 2003.

 Disclosure: N. Yamada, None; R. Yanai, None; M. Nakamura, Santen Pharmaceutical Co., Ltd. (E); M. Inui (P); T. Nishida (P)

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Comparison of the domain sequences of human insulin family proteins. Proinsulin consists of three domains, whereas IGF-1 and -2 each comprise four domains. The C domain (C peptide) of proinsulin is removed posttranslationally to yield the mature insulin molecule. Insulin and IGFs exhibit sequence homology in the A and B domains, which interact directly with insulin or IGF receptors. No specific biological function has previously been attributed to the C or D domains of IGFs. Amino acids that are identical in all three proteins are indicated in bold.

Figure 2.

 

Effects of C- or D-domain peptides of IGF-1 on corneal epithelial migration. (A) Rabbit corneal blocks were incubated for 24 hours with IGF-1 or a C- or D-domain peptide of IGF-1, each at a concentration of 1 nM, in the absence or presence of 20 μM FGLM-amide, after which sections of the tissue were stained with hematoxylin-eosin. Arrows: edge of the epithelial surface of the corneal block; arrowheads: edge of the migrating epithelial cells. (B) Corneal blocks were incubated as in (A), and the length of the path of epithelial migration was determined. Data are the mean ± SE of four to six determinations from a representative experiment. **P < 0.01, ***P < 0.001 versus corresponding corneal blocks incubated without FGLM-amide.

Figure 3.

 

Effect of point mutation of the C domain of IGF-1 on the promotion of corneal epithelial migration. (A) Rabbit corneal blocks were incubated for 24 hours with wild-type IGF-1 or IGF-1(S34A) at a concentration of 1 nM, in the absence or presence of 20 μM FGLM-amide, after which sections of the tissue were stained with hematoxylin-eosin. Symbols are as in Figure 2 . (B) Corneal blocks were incubated as in (A), and the length of the path of epithelial migration was determined. Data are the mean ± SE of four to six determinations from a representative experiment. **P < 0.01 versus corresponding corneal blocks incubated without FGLM-amide.

Figure 4.

 

Failure of the C domain of IGF-1 to activate the IGF type 1 receptor. Human corneal fibroblasts were incubated for 30 minutes with IGF-1 or the C-domain peptide of IGF-1, each at a concentration of 1 nM, in the absence or presence of 20 μM FGLM-amide. Cell lysates were then subjected to immunoblot analysis with antibodies to Akt and to phosphorylated Akt (P-Akt).

Figure 5.

 

Effects of C- or D-domain peptides of IGF-2 on corneal epithelial migration. (A) Rabbit corneal blocks were incubated for 24 hours with IGF-2 or a C- or D-domain peptide of IGF-2, each at a concentration of 1 nM, in the absence or presence of 20 μM FGLM-amide, after which sections of the tissue were stained with hematoxylin-eosin. Symbols are as in Figure 2 . (B) Corneal blocks were incubated as in (A), and the length of the path of epithelial migration was determined. Data are the mean ± SE of six determinations from a representative experiment. *P < 0.05 versus corresponding corneal blocks incubated without FGLM-amide.

Figure 6.

 

Effects of insulin or the C peptide of proinsulin on corneal epithelial migration. (A) Rabbit corneal blocks were incubated for 24 hours with IGF-1, insulin, or C peptide, each at a concentration of 1 nM, in the absence or presence of 20 μM FGLM-amide, after which sections of the tissue were stained with hematoxylin-eosin. Symbols are as in Figure 2 . (B) Corneal blocks were incubated as in (A), and the length of the path of epithelial migration was determined. Data are the mean ± SE of six determinations from a representative experiment. *P < 0.05 versus corresponding corneal blocks incubated without FGLM-amide.

Figure 7.

 

Effects of the C-domain peptide of IGF-1 together with FGLM-amide on rabbit corneal epithelial wound closure in vivo. Representative photographs of the cornea 24 hours after removal of the epithelium (A) and the time courses of the mean area of epithelial defects (B) are shown. After removal of the corneal epithelium by exposure to n-heptanol, eyes were treated with eye drops containing PBS vehicle (control), the C-domain peptide of IGF-1, IGF-1 plus FGLM-amide, or the C-domain peptide of IGF-1 plus FGLM-amide. The corneal wound is traced by a white outline in the control shown in (A). Data in (B) are expressed as a percentage of the value in PBS-treated eyes and are the mean ± SD of results in eight eyes. **P < 0.01, ***P < 0.001 compared with the corresponding control value.