Insulin Growth Factor Promotes Human Corneal Fibroblast Network Formation In Vitro

Abstract

Purpose.: Corneal fibroblast cell (CFC) reticulation is involved in the structural development of corneal stroma and in wound healing. In an earlier paper, it was reported that the expression of VEGFR-1 by CFCs is related to their reticulogenic properties in vitro and decreases with the age of the donors. The present study was focused on the nonreticulogenic corneal fibroblast population and explored whether these cells can be induced to form cell networks in vitro.

Methods.: The network formation was analyzed using an array of signaling pathway inhibitors: wortmannin for PI3 kinase, U0126 for MEK-1/2 kinase, Rottlerin for PKC, farnesyl transferase inhibitor (FTI-277) for Ras, and picropodophyllin (PPP) for IGFR-1. Among the several growth factors studied, IGF seemed to be crucial to cell network formation. The presence of IGF signaling was demonstrated using gene array analysis and was confirmed by RT-PCR and immunocytochemistry and by cell network formation on reduced synthetic basement membrane arrays. The pleiotropic effect of IGF-1 on the cells was analyzed by protein cytokine array.

Results.: The genesis of reticulation was found to occur via MEK-1/2 and IGFR pathways, since inhibitors of these signaling pathways reduced or prevented cell network formation. The addition of exogenous IGF-1 generated a cell network structure in corneal fibroblasts obtained from healthy donors, indicating the involvement of IGF-1.

Conclusions.: IGF signaling and the MEK-1/2 pathway are involved in the cell network formation of corneal fibroblast cells from aged donors.

The corneal fibroblast cells (CFCs) within the corneal stroma are interconnected, indicating the presence of a functional communication network. ¹ Electron microscopic images reveal that CFCs extend processes to create a network with the extracellular matrix (ECM) and also form intercellular junctions. ² CFC migration in response to cytokine stimulation is vital to corneal repair. This property is reduced considerably in aging CFCs. Therefore, the accumulation of senescent keratocytes with age may adversely affect corneal wound healing. ³ The migratory activity of CFCs in vitro is influenced by the composition of the surrounding extracellular matrix, either through integrin-mediated cell–matrix interactions or through matrix–matrix interactions. ⁴ CFCs play a primary role in stromal healing response to corneal trauma or surgical wounds. In response to corneal injury, cytokines and growth factors intervene by influencing epithelial–stromal interactions. The healing and reparative processes may result in tissue remodeling and fibrosis. ⁵ The wound-healing mechanisms of the cornea are complex; several aspects remain to be elucidated. Several growth factors such as EGF, FGF, IGF, KGF, HGF, PDGF, and TGF undoubtedly play key roles in the process. ²  

In an earlier work, we demonstrated that the formation of cell reticulum by corneal fibroblasts is directly related to the expression of VEGFR1 and that this expression decreases with age. ⁶ In the present work, we extended this finding and further investigated the involvement of the IGF-mediated pathway in reticulation and corneal wound healing. 

Material and Methods

Isolation and Culture of Primary Stromal Cells

Human corneas were obtained from a tissue bank (the Banque Française des Yeux) after informed consent was received from the donors, in agreement with the revised rules of the Declaration of Helsinki. Pieces of cornea of 2 to 3 mm² were gently scraped to remove and discard epithelial cells, after which corneal pieces were seeded on 24-well plates containing Dulbecco's modified Eagle's medium:F12 (DMEM/F12; PAA, Mureaux, France); supplemented with 10% fetal calf serum, penicillin (100 U/mL), and streptomycin (100 μg/mL; PAA); and incubated at 37°C in a humidified atmosphere containing 5% CO₂. Nonadherent cells were removed and discarded after 3 days, and the culture dishes were washed with 1× phosphate-buffered saline (PBS; PAA). After an additional week in culture, some cells began to spread out of the corneal explants. These cells were primary corneal stromal cells and were detached with trypsin treatment (PAA) for 15 minutes. The cells were centrifuged, and the pellet was washed and resuspended and then seeded in the same complete medium for incubation for 2 weeks in six-well plates (4 × 10 ⁴ cells/cm²) in a humidified atmosphere containing 5% CO₂ at 37°C. Samples were taken and used for experiments from the first passage through the sixth. 

Analysis by Gene Array

The gene expression patterns of corneal primary culture cells (n = 2) and immortalized cells (by SV40 T antigen; n = 1) were analyzed and compared by a two-color, topic-defined microarray (PIQOR Microarray; Miltenyi Biotec. GmbH, Bergish Galdbach, Germany), as previously described. ^7,8  

Immunochemistry

Cells (10⁵) were cultured for 24 hours in glass chamber slides (Labtek; Nunc, Naperville, IL) in complete DMEM/F12 (PAA). They were washed with serum-free medium and fixed for 15 minutes with 3% paraformaldehyde in PBS (PAA). After three washes with PBS, the cells were incubated for 20 minutes with 1% (wt/vol) bovine serum albumin (BSA; Sigma-Aldrich, Saint Quentin Fallavier, France) to avoid nonspecific binding. The cells were then incubated for 60 minutes at 4°C with a specific primary antibody diluted 1:100. The antibodies were mouse anti-IGF-1, mouse anti-IGF-2, goat anti-IGFR-1, and goat anti-IGFR-2 (R&D Systems, Lille, France). After three washes with PBS, the cells were treated with anti-goat secondary biotinylated antibody for 1 hour at 4°C (1:100 dilution). The cells were washed twice in PBS-BSA and exposed to streptavidin fluorescein (Sigma-Aldrich) for 45 minutes at 4°C. The antibody fluorescence was observed on slides by microscope (Nikon, Tokyo, Japan). 

PCR Procedure

Total RNA was extracted (Nucleospin RNA-II kit; Macherey-Nagel, Hoerdt, France), and the cells (5 × 10⁶) were lysed at 0°C. RNA and DNA were collected on silicate membrane, and DNase I was added directly to the membrane for 15 minutes at room temperature. After several washes, total RNA was eluted with RNase-free water. After quality control in 2% agarose gel and quantification by spectrometry, we used 1 μm of total RNA for reverse transcription to cDNA. The RNAs were denatured for 5 minutes at 65°C. Appropriate buffer (5 M Tris-HCl, pH 8.8; 2 M (NH₄)₂SO₄, and 1 M MgCl₂) was added with 1.25 mM dNTP (Sigma-Aldrich) and 200 units of MM-MLV reverse-transcriptase (Invitrogen-Gibco, Paisley, UK) for 30 minutes at 42°C in a final volume of 20 μL. We used synthetic primers IGF-1, 5′-AAA TCA GCA GTC TTC CAA C-3′ sense and 5′-CTT CTG GGT CTT GGG CAT GT-3′; IGF-2, 5′-AGT CGA TGC TGG GCT TCT CA-3′ sense, and 5′-GTG GGC GGG GTC TTG GGT GGG TAG antisense, (Eurobio, Paris, France); IGFR-1 (314 bp), 5′-CAC CAT GTC CTC CTC GCA TCT-3′ sense, and 5′-ATC CAC GAT GCT GTC TGA GG-3′ antisense; IGFR-2 (484 bp), 5′-CCT GGA GAC GTA CTG TGC TAC C-3′ sense, and 5′-GCT CAC TTC CGA TTG CTG G-3′ antisense; and β2-microglobulin (control) primer 5′ CCA GCA GAG AAT GGA AAG TC 3′ sense, 5′ GAT GCT GCT TAC ATG TCT CG 3′ antisense. The reaction mixture consisted of 15 ng of cDNA, 2.5 units Taq polymerase (Invitrogen-Gibco), 200 mM dNTP, 0.2 mM of the respective oligonucleotide primers, 10 mM Tris-HCl (pH 8.3), 50 mM KCl, and 1.5 mM MgCl₂ in a final volume of 450 μL. The amplification was performed with a thermal cycler (GeneAmp 9600; Perkin Elmer, Wellesley, MA): 40 cycles with denaturation at 94°C for 1 minute, followed by Tm for 45 seconds at 52°C for IGFR-1, for IGFR-2, 64°C for IGF-1, 67°C for IGF-2, and 60°C for β2-microglobulin and an extension at 72°C for 1 minute. PCR products were collected in a 2% agarose electrophoresis gel in 0.45 BET. Samples with H₂O and without Taq polymerase were tested as controls. The analysis was performed with UV light. The PCR cycle number was normalized to the amount of human β2-microglobulin. 

Cell Motility Assays

Cell motility and behavior were assessed in a 96-well microplate precoated with a natural murine fibrosarcoma extracellular matrix (Matrigel; BD Biosciences, San-Jose, CA). It provides a physiologically relevant environment for studies of cell morphology, migration, and invasion as cultures grown on the top of it present a two-dimensional arrangement within the formation of cellular networks. Standard Matrigel contains multiple active growth factors, ⁹ including IGF-1 (15.6 ng/mL), bFGF (0.01 pg/mL), EGF (05–1.3 ng/mL), PDGF (12 pg/mL), NGF (<0.2 ng/mL), and TGF-β (2.3 ng/mL). For this study, two types of Matrigel, one rich and the other reduced (poor) in the growth factors IGF-1 (5 ng/mL), bFGF (0.01 pg/mL), EGF (<0.5 ng/mL), PDGF (<5 pg/mL), NGF (<0.2 ng/mL), and TGF-β (1.3 ng/mL) were used (BD Biosciences). A 96-well microplate was coated with 35 μL of the matrix at 0°C. After 1 hour at 37°C, 10⁴ cells were seeded on each well with 100 μL of medium. Two hundred microliters of medium DMEM:F12 with a final concentration of 200 ng/mL IGF (PromoCell GmbH, Heidelberg, Germany) was added to primary culture cells from 62 donors and were seeded on the matrix as reported elsewhere. ⁶ Every 3 hours, the cultures were observed, and the number of interconnections and tubules were counted (Archimede; Microvision, Paris, France). 

Determination of Signaling Pathway Involved in the Formation of Network Structures by CFCs

To elucidate the mechanism of the formation of the corneal fibroblast network on the matrix, we tested the effect of five signaling inhibitors: picropodophyllin (PPP), an IGFR-1 inhibitor; Rottlerin, a PKC inhibitor; wortmannin, a PI3 kinase inhibitor; U0126, an MEK-1/2 kinase inhibitor; and FTI-277, a farnesyl transferase inhibitor. All inhibitors were purchased from Calbiochem (La Jolla, CA) and were used at a final concentration of 10 μM, added to the CFC suspension just before plating on the matrix. The cells were examined after an 8-hour incubation at 37°C in a humidified atmosphere of 5% CO₂, as described above, and the inhibition of network structure formation was analyzed. 

Quantification of IGF-1

The supernatants (serum-free DMEM:F12, DMEM:F12; Invitrogen-Gibco) were collected from cultures of human CFCs at 10⁶ cells/mL after the cells had incubated for 24 hours in a humidified atmosphere of 5% CO₂ and 95% air at 37°C. The concentration of IGF-1 in the supernatant was determined by ELISA (R&D Systems). 

Cytokine Array

To analyze the pleiotropic role of IGF, we examined the supernatant of corneal fibroblasts (n = 2) using a protein cytokine array (Human Cytokine Antibody Array; Ray Biotech, Norcross, GA). This technique is based on the principle of “sandwich immunoassay.” It was composed essentially of screening, in duplicate, 174 different membrane-coupled anti-cytokines along with appropriate controls. 

CFCs (10⁶ cells per mL) were incubated in the presence (or absence) of IGF-1 (200 ng · mL⁻¹) in DMEM without fetal calf serum at 37°C in a humidified atmosphere of 5% CO₂ for 24 hours. Supernatants containing the cytokines were retrieved, and the cytokines were allowed to couple with their specific antibodies, which had been fixed on membranes. The membranes were saturated for 2 hours at room temperature with BSA. Array membranes with the supernatants (along with the controls) were incubated with the corresponding antibodies overnight at 4°C. After several successive washes, the membranes were incubated in the presence of a mixture of anti-cytokines biotinylated at 4°C overnight. Streptavidin, coupled with HRP, was added to the membranes for 2 hours at room temperature. The presence of antibody-coupled proteins was revealed by applying ECL (enhanced chemoluminescence) to the membranes, according to the recommendations of the manufacturer. Membranes were then exposed to photosensitive film (X-Omat AR; Eastman Kodak, Rochester, NY). 

The intensity of chemoluminescence captured on the photosensitive film was measured and recorded. After the background noise was subtracted, the results were expressed as the ratio of the chemoluminescence intensity of experimental versus control spots. The positive control was set at 1. A ratio less than 0.5 indicated a reduction of the cytokine, and a ratio of greater than 1.5 indicated an increase in the presence of IGF-1 cytokine expression. 

Statistical Analysis

Results are expressed as the mean ± SEM (or SD) and compared using a two-tailed nonparametric Mann-Whitney test (InStat software; Sigma-Aldrich). P < 0.05 was considered significant. 

Results

Corneal Fibroblasts Form Networks on Synthetic Matrix

Fibroblast cell–cell interaction led to the formation of a cell network, as described previously. ⁶ Network formation by CFCs is presented in Figure 1. From 62 donors (cultures) in all, we selected five primary cells and also included one immortalized reticulogenic corneal cell (Figs. 1a, 1b) and seven nonreticulogenic cells (Figs. 1c, 1d) for seeding on the matrix. The results obtained after cell culture showed two distinct types of CFCs. These were classified as types 1 and 2, according to whether they formed or did not form a network on the matrix, respectively. 

Figure 1.

 

Network formation by CFCs on synthetic basement membrane. (a) Type 1 and (c) type 2 CFCs seeded on the matrix and photographed before incubation, representing time 0. The same plates were photographed after 8 hours of incubation. (b) Type 1 CFCs showed cell elongation and the beginning of network formation, whereas (d) type 2 fibroblasts showed no network formation.

Mechanisms Involved in Corneal Fibroblast Network Formation

The effectors of signaling pathway on network formation was tested with different inhibitors on type 1 CFCs grown on synthetic matrix rich in growth factor, so as to determine the molecular mechanisms involved in reticulogenesis. 

The inhibition of tubule formation, compared with the control (Fig. 2a), was 60% with PPP (P < 0.0001; Fig. 2b) and 38% with U0126 (P < 0.0001; Fig. 2c), whereas with Rottlerin (P < 0.05; Fig. 2d), wortmannin (P < 0.05; Fig. 2e), and FTI-277 (P < 0.05; Fig. 2f), the inhibition was not significant. The histogram presents the number of cell intersections (Fig. 2g) and tubules (Fig. 2h) formed on the synthetic matrix at 8 hours. 

Figure 2.

 

The effect of signaling pathway inhibitors on network formation on synthetic matrix. Cell intersections and tubulelike formations were analyzed after treatment with several signaling pathway inhibitors, such as (b) PPP, an inhibitor of IGFR-1 (10 μM, P < 0.0001); (c) U0126, an MEK-1/2 kinase inhibitor (10 μM, P < 0.0001); (d) wortmannin, a PI3 kinase inhibitor (10 μM, P < 0.05); (e) Rottlerin, a PKC inhibitor (10 μM, P < 0.05); and (f) FTI-277, a farnesyl transferase inhibitor (10 μM, P < 0.05). They were compared with the nontreated control cells (a). The number of cell intersections (g) and tubules (h) formed were counted after 8 hours.

In light of the role played by IGF in the reticulation of the type 1 cells and the ability of its specific inhibitors to block reticulation, we were prompted to extend the investigation to the type 2 cells, to determine their eventual response to added IGF. 

Figure 3 shows the results of fibroblast network formation on the matrix. Figure 3a represents the control using the type 2 cells without exogenous addition of IGF. Figure 3b shows that addition of IGF induced the formation of CFC networks on reduced growth factor matrix. 

Figure 3.

 

Induction of a CFC network by IGF-1. The addition of exogenous IGF-1 (100 ng/mL) significantly increased the number of cell interconnections in only type 1 cells after 18 hours (b) and on type 2 fibroblasts, this effect was inhibited by PPP (c). The number of cell intersections (d) and tubules (e) formed on the matrix were counted at 18 hours. The control using the type 2 cells is represented by (a).

The effect on type 2 CFCs of simultaneous addition of IGF and its inhibitor PPP (a specific inhibitor of IGFR-1) before plating is shown in Figure 3c. The histogram (Fig. 3d) shows a 60% diminution in the number of formed tubules and intersections after 18 hours of incubation. The results indicate that IGF induces the formation of tubules and their intersections. Its effect can be blocked by simultaneous addition of the specific inhibitor of IGFR-1 (PPP). 

Expression of IGFs and Their Receptors, IGFR-1 and -2, by Corneal Fibroblasts

Using RT-PCR analysis, we investigated the presence of mRNAs in fibroblasts type 1 and 2 for IGF-1 (410 bp), IGF-2 (470 bp), and their receptors IGFR-1 (314 bp) and IGFR-2 (144 bp), with β2-microglubulin as a control (114 bp). Figure 4A shows that both type 1 (Fig. 4A, lanes c, f) and type 2 (Fig. 4A, lanes b, d, e) fibroblasts expressed IGF-2 and IGFR-1 and -2. Our results showed that IGF-1 was not expressed by type 1 or 2 cells, whereas both expressed IGF-2 and their receptors IGFR-1 and -2. 

Figure 4.

 

(A) IGF and IGF receptors analysis by RT-PCR of CFCs. The presence of IGF-1, IGF-2, IGFR-1, and IGFR-2 mRNAs was analyzed with specific primer in type 1 and 2 CFCs. In five tested cells, only IGF-2, IGFR-1, and IGFR-2 were expressed. No differences were observed between CFCs type 1 (lanes c, f) and type 2 (lanes b, d, e). The RT-PCR control with H₂O is presented in lane a. β-2Microglobulin was used as the control for cell extraction of the RNA. (B) Immunodetection of proteins for IGFs and their receptors. Immunoreactivity of IGFs and their receptors using specific antibodies in type 1 and 2 CFCs was the same. Compared with negative control, with an anti-mouse antibody and (Ba) and an anti-goat antibody (Bb), CFCs treated with IGF-2 (Be), IGFR-1 (Bd), and IGFR-2 (Bf) antibodies showed strong immunoreactivity in cell cytoplasm. Weak immunoreactivity against IGF-1 was detected (Bc), probably owing to absorbed exogenous IGF. Scale bar, 7 μm.

The RT-PCR results were reconfirmed by immunocytochemical analysis using specific antibodies against IGF and its receptors (Fig. 4B). Micrographs Ba (anti-mouse) and Bb (anti-goat) were negative controls. Micrograph Bc shows that IGF-1 was not expressed, whereas cells expressed IGF-2 (Fig. 4Be). Micrographs Bd and Bf show intense fluorescence, indicating that IGFR-1 and -2 were highly expressed, which confirms the observations using RT-PCR analysis. 

Protein Cytokine Array Analysis of In Vitro Cultured Human Corneal Fibroblasts

The supernatants of the CFCs were processed by using the cytokine array technique. The spot intensity, revealed after chemoluminescence and exposure to photosensitive films, was measured. The results, presented in a histogram (Fig. 5) showed an increase in the presence of IGF-1 and a diminished expression of cytokines. These results indicate that the expression of IL13, TRAILR4, endoglin, and ErbB3 decreased by approximately 50% compared with the controls. We also found an increase in the expression of 18 cytokines. 

Figure 5.

 

Cytokine expression by human CFCs regulated by exogenous IGF-1. The data show an increase (ratio > 1.5) in expression of 18 cytokines and a decrease in expression of 4 proteins (ratio < 0.5).

In parallel, we found that CFCs secreted IGFBP-1, -2, -3, -4, -5, and -6 and soluble IGFR (Table 1). Expression of these proteins was also determined in a gene array analysis (results not shown). 

Table 1.

 

Regulated Cytokine Proteins by Exogenous IGF and Their Ratios in the Array

Name	Protein Array Ratios*
IGFsR	0.52
IGFBP1	1.23
IGFBP2	1.1
IGFBP3	0.87
IGFBP4	1
IGFBP5	0.94
IGFBP6	1

*  Tested/control.

Discussion

The mechanisms of wound healing of the cornea in all its complexity remain to be worked out in complete detail. In a rabbit CFC culture, fibroblasts passaged at low density become α-SMA⁺ myofibroblasts and mimicked the situation in the cornea after wounding. ¹⁰ Some earlier independent studies have shown that the phenotype of CFCs can be maintained in vitro, ¹¹ and among these, a small fraction are indeed PAX6⁺ corneal fibroblast progenitor cells. ¹² In our experiments, CFCs were tested and precharacterized using the α-SMA analysis with specific antibodies (Sigma-Aldrich) in separate cultures at high density (85%–90% confluence) The results obtained (not shown) indicated that both type 1 and 2 CFCs were negative, and therefore they were not characterized as myofibroblasts. 

We distinguished two cell types: type 1 capable of reticulation and type 2 unable to reticulate. Type 1 predominates during most of life and type 2 appears with increasing age when physiological body functions are affected and vision may begin to diminish. ⁶ Corneal fibroblast reticulation remains a salient feature in the structural development of corneal stroma and wound healing. Several growth factors, notably VEGF and IGF and also the receptors, are expressed in the corneal cells and intervene in their proliferation. ¹²  

If the two IGFs, IGF-1 binds to two distinct cell surface receptors: IGFR-1, and the insulin receptor (INSR). The binding of IGF-1/IGF-R1 occurs at a higher affinity than IGF-1/INSR. IGFR-1 is a transmembrane receptor tyrosine kinase and is activated by both IGF-1 and -2. ^{14 –17} IGF-2 binds to the IGF-2 receptor and acts as a signaling antagonist. IGF-1 and -2 form complexes with insulin-like growth factor binding protein (IGFBP). 

In the cornea, IGF-1 is released from corneal epithelial cells and upregulates the expression of N-cadherin in corneal fibroblasts. ¹⁸ It has also been shown that the IGFs induce the recruitment of fibroblasts to the site of wounds in vivo ¹⁹ and their proliferation. ²⁰  

IGF-1R is believed to play a role in cell differentiation and in inhibiting apoptosis. After binding IGF-2 at the cell surface, IGF-2R accumulates to form clathrin-coated vesicles and is internalized, where it is ultimately eliminated. ²¹  

In the present work, we attempted to further our current knowledge of the role of growth factors in stromal networking and organization of the cornea. The use of inhibitors brought to light the importance of IGF in reticulation of type 1 cells. We were tempted therefore to find out whether the nonreticulogenic type 2 cells could be induced to form networks by exogenous addition of IGF. The type 2 cells unexpectedly responded to exogenous addition of IGF by forming tubules on the synthetic matrix, suggesting a role of IGF in CFC interaction. Our observations are consolidated by experiments in which we used an inhibitor of the IGF signaling pathway, PPP, and also an inhibitor of MEK-1/2 kinase, U0126. 

The IGF induction of tubules and intersection formation can be considerably diminished by simultaneous addition of antagonists. We also demonstrated the presence of IGF-2 and the receptor of IGF-1 and -2 in type 2 cells. These results speak in favor of an inherent resident capacity of type 2 cells to reticulate, provided the appropriate signals are available. IGF family and their receptors constitute such a signal. 

Cytokine analysis of supernatants of cultured cells, stimulated by exogenous IGF, may indicate that IGF participates in the regulation of cellular proliferation and differentiation. The exogenous IGF may inhibit the commitment to apoptosis by diminishing 50% of the expression of TRAILR4. Cytokine analysis also showed an increase in adhesion molecule expression (Tie-1, Tie-2, L-selectin, Siglec-5, and IL-13β) and seemed to orient them via the TGFβ axis (CD105, Sgp130, BMP5, BMP7, TGFB2, and EGFR) toward differentiation of fibroblasts into myofibroblasts. 

Our results, however, revealed that IGF was not secreted by the corneal fibroblasts whereas they harbored functional IGF receptors. The cells therefore presumably regulate their homeostasis of IGF through the secretion of different IGF-binding proteins and sIGFR. Gene array experiments revealed that these proteins and IGFBPs mRNAs were expressed in CFCs. 

The absence of IGF-1 signal in the RT-PCR and immunocytochemical analyses indicates that the CFCs expressed only IGF-2, not IGF-1. Physiologically, maintenance of the network structure of the corneal stroma can be regulated by IGF-1 released from epithelial cell layers. ¹⁸ Altogether, the results suggest a crucial role of IGF -1 signaling in epithelial-stromal interactions. 

It should be remembered that IGFBP2 and IGFR-1 are expressed in corneal fibroblasts of rat, ¹⁶ and we have shown that IGFBPs are also expressed in human cornea. 

Apoptosis in CFCs is the first detectable response of the corneal stroma to injury. IGF-1 and -2 induce Akt-activation in human corneal fibroblasts in a PI3-kinase-dependent manner and have a role in the inhibition of apoptosis and in the promotion of cell survival. ²² However, the molecular mechanisms of cell reticulations induced by IGF-1 remain obscure. Recently, we demonstrated the implication of IGF-1 in vasculomimicry in leukemic bone marrow stromal cells. ²³ It indicates yet another novel, important role for IGF-1 in other diseases. 

The present work emphasizes the role of IGF in inducing reticulation in type 2 cells, which appear in older age when repair mechanisms are not at their best. The importance of IGF in corneal stromal cells is emphasized in the present work and is described for the first time. Our finding is reinforced by a recent study demonstrating the importance of IGFI and IGFII. ²⁴ It is clear that IGF signaling and MEK-1/2 pathway is involved in the cell network formation by human corneal fibroblasts cells from aged donors, suggesting that IGF-related components, such as those that mimic IGF's action, could be novel pharmacologic derivatives. Clinical application is a definite possibility, and studies are already under way with this in view. 

Footnotes

 Supported by Rétina France and the Dalloz Foundation.

Footnotes

 Disclosure: A. Berthaut, None; P. Mirshahi, None; N. Benabbou, None; E. Ducros, None; A. Agra, None; A. Therwath, None; J.M. Legeais, None; M. Mirshahi, None

The authors thank the Banque Française des Yeux for providing human corneas.