Abstract
Purpose.:
The goal of this study was to determine the molecular mechanism by which transforming growth factor-α (TGF-α) is a more potent activator of epidermal growth factor receptor (EGFR)-mediated corneal wound healing than epidermal growth factor (EGF).
Methods.:
Telomerase immortalized human corneal epithelial (hTCEpi) cells and primary human corneal epithelial cells were tested for their ability to migrate in response to EGF and TGF-α. In parallel, the endocytic trafficking of the EGFR in response to these same ligands was examined using indirect immunofluorescence, immunoblots, and radioligand binding.
Results.:
TGF-α, compared with EGF, is a more potent activator of corneal epithelial cell migration. Although both TGF-α and EGF were able to induce EGFR internalization and phosphorylation, only those receptors that were stimulated with EGF progressed to lysosomal degradation. EGFRs stimulated with TGF-α recycled back to the plasma membrane, where they could be reactivated with ligand.
Conclusions.:
This study reveals that EGFR-mediated cell migration is limited by ligand-stimulated downregulation of the EGFR. This limitation can be overcome by treating cells with TGF-α because TGF-α stimulates EGFR endocytosis, but not degradation. After internalization of the TGF-α/EGFR complex, EGFR recycles back to the plasma membrane, where it can be restimulated. This sequence of events provides the receptor multiple opportunities for stimulation. Thus, stimulation with TGF-α prolongs EGFR signaling compared with EGF.
The cornea is the clear, dome-shaped structure that covers the anterior of the eye. It has two main functions. First, the transparent layer of cells control and focus the entry of light into the eye. Second, it prevents external agents, such as particles, viruses, and bacteria, from entering the eye. A leading cause of blindness worldwide is the result of damage to and infection of the cornea.
1
The cornea's first line of defense is its most exterior layer, which is made up of epithelial cells. On wounding of the corneal epithelium, the epithelium immediately begins to reestablish its structural integrity. The three major cellular events in the re-formation of the corneal epithelium are the migration of cells from the surrounding basal epithelium to the wounded area, the proliferation of these cells, and the differentiation of the cells into the stratified layers. Failure of these events to occur can result in painful corneal ulcerations and distorted vision.
1,2
A key regulator for maintaining a healthy cornea and promoting regrowth of the wounded cornea is the epidermal growth factor receptor (EGFR).
3 The EGFR is the prototypical tyrosine kinase receptor localized to basal and differentiated epithelia in the cornea.
4 It has been shown using an ex vivo organ system that activation of the EGFR can promote all three processes—proliferation, migration, and differentiation—involved in corneal epithelial wound healing.
4–6 Further, patients taking EGFR inhibitors (e.g., Iressa [AstraZeneca Pharmaceuticals, Wilimington, DE] or Tarceva [Genentech, South San Francisco, CA]) therapeutically for the treatment of non-small cell lung carcinomas report an increased incidence of corneal ulcerations.
7,8 Thus, the stimulation of the EGFR is necessary and sufficient for corneal wound healing.
Evidence from a rodent corneal wounding model is consistent with a ligand other than EGF as the primary mediator of wound healing.
4 There are seven endogenous EGFR ligands.
9 Four of these (EGF, TGF-α, HB-EGF, and amphiregulin) have been demonstrated to promote corneal wound healing.
4,10 Further, TGF-α, but not EGF, is transcriptionally upregulated after corneal wounding.
4,11
A role for TGF-α in wound healing makes sense physiologically. TGF-α is more efficacious than EGF at promoting corneal wound healing,
12,13 despite evidence that they stimulate the same receptor population and that they do so by way of the same molecular mechanism.
14 The differences in efficacies are not a function of ligand binding; EGF and TGF-α bind the EGFR with comparable affinities (dissociation constant [
K d] = 5.7 nM and 4.6 nM, respectively).
15 Thus, the mechanism of ligand activation of the receptor cannot account for the reported differences in signaling by these ligands.
In this study, we use primary and immortalized corneal epithelial cells to test the hypothesis that the differences in efficacies among EGFR ligands are a consequence of variations in cellular mechanisms for receptor inactivation. Ligand-mediated receptor desensitization is widely regarded as the primary method of EGFR inactivation.
16 It has been well established that the EGF/EGFR complex is internalized by clathrin-coated pits on ligand binding. Once inside the cell, the EGF/EGFR moves through a series of defined endocytic compartments (early endosome to late endosome to lysosome) and ultimately results in degradation of the ligand and receptor.
17 We postulated that in the human corneal epithelium, the endocytic trafficking of TGF-α/EGFR complex bypasses the degradation process. If this hypothesis is correct, after stimulation with TGF-α, the receptor would retain its activity longer or be able to undergo multiple rounds of stimulation. Identifying the molecular basis for the differences in ligand activity is the first step in developing strategies for accelerating and enhancing corneal wound healing.
We found that, as has been reported in ex vivo models, TGF-α is a more potent activator of corneal epithelial cell migration than EGF. In addition, stimulation with EGF results in the internalization and degradation of the EGF/EGFR complex. In contrast, treatment with TGF-α promotes internalization of the EGFR but ultimately leads to recycling back to the plasma membrane. After treatment with TGF-α, the receptor can be restimulated with additional ligand. Therefore, the activity of the TGF-α–stimulated EGFR is enhanced because of the absence of receptor degradation and the ability of the receptor to be restimulated with ligand.
As a secondary approach for analyzing the endocytic trafficking of the EGFR, we monitored the movement of the ligand/receptor complex using radioligands. By incubating cells with either 125I-EGF or 125I-TGF-α, we were able assess the amount of EGF/EGFR and TGF-α/EGFR inside the cell. If the rates of ligand internalization, degradation, and recycling were similar, that would be consistent with identical trafficking itineraries. Alternatively, differences in the amount of time the radioligand remained in the cell would provide additional evidence of an altered route of trafficking.
After incubation with either
125I-EGF or
125I-TGF-α, a comparable percentage of cell-associated radioligand was intracellular (
Fig. 4). When cells were incubated with
125I-EGF, 50% of the internalized radioligand remained associated with the cell after 100 minutes. In contrast, when the cells were treated with
125I-TGF-α, within 50 minutes only 50% of the internalized radioligand remained intracellular. The more rapid removal of radioligand from the cell and the slowed rate of receptor degradation are consistent with the notion that TGF-α promotes EGFR recycling and that EGF causes EGFR degradation.
To more clearly examine the EGFR endocytic trafficking itinerary in individual cells, we used indirect immunofluorescence to monitor the intracellular localization of the EGFR (
Fig. 5). Primary corneal epithelial cells were treated with either EGF or TGF-α, and the subcellular distribution of the EGFR was monitored by indirect immunofluorescence. Shown are confocal micrographs collected from the center (along the
z-axis) of each section from primary corneal epithelial cells (
Fig. 5). Similar results were observed when the hTCEpi cells were used (
Supplementary Fig. S1).
In cells that have not been stimulated (with either ligand), the distribution of the EGFR (shown in green) is along the plasma membrane of the cell. Once either EGF or TGF-α has been added for 15 minutes, there is a dramatic redistribution of the receptor from the cell surface to intracellular vesicles. In cells that have been treated with EGF, the staining of the EGFR remains punctate and becomes increasingly perinuclear over time (
Fig. 5, upper panels). With time, the intensity of EGFR staining decreases and is consistent with lysosomal degradation of the EGFR.
In cells that have been treated with TGF-α, after the initial localization of the receptor to endosomes, there is a time-dependent appearance of the EGFR back to the plasma membrane (
Fig. 5, lower panels). The amount of plasma membrane staining of the EGFR increases over time. Twenty-four hours after treatment with TGF-α, the distribution of EGFRs is indistinguishable from that of untreated cells.
It is important to note that though, in principle, this experiment was similar to the immunoblot-based analysis of EGFR degradation, the assay provided different readouts that precluded their direct comparison. The more sensitive immunofluorescence assay used in
Figure 5 indicates the distribution of the immunoreactive protein (the EGFR) present in the cell; the assay does not distinguish whether the immunoreactive protein is the native full-length receptor or a partially degraded fragment. In contrast, the immunoblot assay used in
Figure 3 indicates whether there is a 180-kDa immunoreactive protein. As a result, the loss of the immunoreactive band in
Figure 3 does not kinetically overlap with the diminished immunofluorescent signaling seen in
Figure 5. Both sets of data indicate that stimulation with EGF promotes more rapid degradation than TGF-α.
Finally, we wanted to know whether the EGFR that reappears at the cell surface after TGF-α–mediated internalization is capable of restimulation. If so, this would provide strong data that there is a physiological consequence to EGFR recycling. To test this idea, we monitored the localization of the EGFR after pulsing the cells with TGF-α.
Cells were incubated in media alone or in media supplemented with TGF-α (10 ng/mL) for 2 hours to induce internalization. External ligand was removed from the media by a series of washes, and the cells were incubated in media without ligand. After 4 more hours (6 hours total), cells were reincubated with TGF-α (10 ng/mL) for 15 minutes (
Fig. 6A). Cells were collected at various times during this treatment, and the localization of the EGFR was monitored by indirect immunofluorescence (
Fig. 6B). As predicted by our previous data, the EGFR internalized and recycled back to the plasma membrane after TGF-α treatment. When those cells were retreated with TGF-α, the EGFR became internalized in endosomes (
Fig. 6; 6.25 hours). In cells that were pulsed with EGF, the EGFR remained intracellular after washout, and restimulation with EGF had no effect on EGFR distribution (data not shown).
Using this system of introducing pulses of TGF-α to the cells, we were able to monitor EGFR trafficking. These data are consistent with the ability of TGF-α to stimulate EGFRs multiple times, whereas cells treated with EGF were only able to be activated once because the receptor was retained in the cell and then degraded.
Corneal wound healing and tissue homeostasis are critical to the maintenance of a healthy eye. It has been established using in vivo models that TGF-α is a potent ligand for promoting EGFR-mediated corneal epithelial growth after wounding of the tissue.
4,10,13 Here we show, using tissue culture models, that the critical first step in corneal wound healing—cell migration—is more potently activated by TGF-α than EGF. In these studies, we identify the endocytic trafficking itinerary as the critical regulatory process that confers this difference. As indicated by the radioligand trafficking data, receptor degradation studies, and indirect immunofluorescence, stimulation with EGF promotes EGFR degradation. In contrast, when cells are stimulated with TGF-α, the ligand/receptor complex internalizes but recycles to the plasma membrane. At the plasma membrane, the EGFR can bind ligand and be activated. Thus, the difference in ligand efficacy can be accounted for by the duration and frequency of receptor stimulation.
Although the exact mechanism by which these ligands confer different routes of EGFR endocytic trafficking is unknown, one strong candidate rests in the binding properties of EGF compared with those of TGF-α. It has been shown in other systems that dissociation of TGF-α from the EGFR is more sensitive to pH than EGF dissociation. The pH at which 50% of the maximally bound TGF-α dissociates from the receptor has been reported to be 6.83; for EGF, 50% of the ligand is dissociated at pH 5.89.
25 Given that early endosomes have a pH of approximately 6.8,
26 it is likely that TGF-α dissociates from the EGFR. On ligand dissociation, the receptor's kinase becomes inactive and the receptor no longer is able to sustain tyrosine phosphorylation and communication with downstream effectors. These dephosphorylated receptors recycle back to the plasma membrane
27 (
Fig. 7).
Although the model that TGF-α promotes multiple rounds of EGFR stimulation is attractive, one cannot rule out other models. For instance, it has been shown in other cell lines that the endocytic pathway also provides spatial regulation of EGFR signaling.
28,29 If the model of spatial regulation is correct, enhanced signaling by the TGF-α/EGFR complex would be the result of the liganded EGFR in the early endosome rather than of the EGF/EGFR complex. If true, this model also supports a role for the ligand-mediated itinerary of endocytic trafficking as a key determinant in ligand efficacy.
These data support the notion that manipulation of the endocytic pathway may be an effective mechanism for enhancing EGFR-mediated corneal wound healing. The use of ligands that cannot internalize or undergo degradation may provide a therapeutic strategy for reepithelialization of a wounded corneal epithelium. Alternatively, if the activities of proteins (rab5, tsg101, rab7) that regulate movement of the EGFR through the endocytic pathway are blocked, wound healing may occur more rapidly. However, as mentioned, the endocytic location of the EGFR may provide the necessary specificity for coupling to and activating effectors. Thus, simply disrupting the trafficking of the receptor may not be sufficient to promote the necessary signals for wound healing. Deciphering whether EGFR signaling is enhanced by spatial placement of the receptor or prolonging receptor activity is challenging. However, making these distinctions will prove to be important for designing therapeutic strategies that promote corneal wound healing.