CCL20, γδ T cells, and IL-22 in corneal epithelial healing

Abstract

After corneal epithelial abrasion, leukocytes and platelets rapidly enter the corneal stroma, and CCR6⁺ IL-17⁺ γδ T cells migrate into the epithelium. γδ T-cell-deficient (TCRδ^−/−) mice have significantly reduced inflammation and epithelial wound healing. Epithelial CCL20 mRNA increased 19-fold at 3 h, and protein increased ∼16-fold at 6 h after injury. Systemic or topical treatment of wild-type C57BL/6 mice with anti-CCL20 reduced γδ T-cell accumulation in the cornea by >50% with a concomitant decrease in epithelial healing and stromal inflammation. In addition to CCR6 and IL-17, corneal γδ T cells stained positively for RORγt, IL-23R, and IL-22. Anti-IL-22 reduced peak cell division of the healing epithelium by 52%. Treatment of TCRδ^−/− mice with rIL-22 significantly promoted wound closure, with peak epithelial cell division increased >3-fold. In addition, rIL-22 restored neutrophil and platelet influx in the TCRδ^−/− mice to wild-type levels and increased CXCL1 production by wounded corneal explants >2-fold. These results indicate that an important aspect of the healing response to corneal epithelial abrasion includes CCL20-dependent influx of CCR6⁺ IL-17⁺ IL-22⁺ γδ T cells and that IL-22 contributes to the inflammatory response and promotes epithelial healing.—Li, Z., Burns, A. R., Byeseda Miller, S., Smith, C. W. CCL20, γδ T cells, and IL-22 in corneal epithelial healing.

The surface stratified epithelium of the cornea functions as a physical and chemical barrier against infection and provides physiological and biochemical mechanisms necessary for sustained visual clarity. It also contains leukocytes of the innate immune system (1–9) that reside mostly in the peripheral regions of the cornea near the junction with conjunctiva (4, 5, 7, 10). Animal models of corneal injury reveal an acute inflammatory response of the limbal blood vessels and substantial accumulation of leukocytes within the avascular dense connective tissue stroma and the stratified epithelium. Neutrophils migrate to corneal wounds mostly within the anterior stroma (3, 7, 11); lymphocytes, macrophages, and dendritic cells migrate within both stroma and stratified epithelium (7, 10, 12–14); and platelets accumulate within and outside of limbal vessels (15, 16). CCR6⁺ IL-17⁺ Vγ3⁻ γδ T cells accumulate in the murine cornea after abrasion injury to the epithelium and promote an acute neutrophil dominant inflammatory response required for early healing of injured corneal nerves (16) and epithelium (7, 17).

γδ T cells are reported to assist in healing epithelial surfaces of the skin, gastrointestinal tract, and lung, and the specific mechanisms of their contributions to healing may differ depending on the tissue involved (2, 18). γδ T cells resident in the epidermis of mice represent a distinct subset with an invariant TCR (Vγ3Vδ1) that apparently contributes to wound healing though release of keratinocyte growth factors, hyaluronan, and cytokines, including TNF-α, IFNγ, and IL-22 (2, 19). In the lung, tissue integrity appears to be maintained and restored by γδ T cells that express IL-17 (20). γδ T cells that populate the postnatal thymus, spleen, lymph nodes, and some mucosal surfaces are predominately of 2 major functional groups. One subset expresses IFNγ when activated, is characterized by surface markers such as CD27 and NK1.1, and expresses the transcription factor T-bet (21–23). A second subset expresses IL-17A, IL-17F, and IL-22 when activated; is characterized by surface receptors CCR6 and IL-23R; and is CD27 negative (22–25). These cells express the transcription factor RORγt (18, 24). Each of the subsets is programmed for rapid response to tissue stress (injury or infection), thereby influencing the very early inflammatory and host-protective events polarized by IFNγ or IL-17 (25, 26). In the current study, we demonstrate that CCL20 is a dominant chemokine attracting the γδ T cells to the cornea and that these cells express IL-22, a cytokine that induces inflammation and promotes wound closure and epithelial recovery.

MATERIALS AND METHODS

Animals

TCRδ^−/− mice on the C57BL/6 background and C57BL/6 mice were purchased from The Jackson Laboratory (Bar Harbor, ME, USA). Mice were controlled for sex (female) and age (12–16 wk). To examine the contribution of IL-22, CCL20, and CXCL1 to inflammation and corneal epithelial reepithelialization, some mice were injected intraperitoneally with anti-mouse IL-22, CCL20, or CXCL1 antibody (R&D Systems, Minneapolis, MN, USA) at 0.2 mg in 200 μl PBS 2 h before wounding. For IL-22 replacement in TCRδ^−/− mice, a single dose of 1000 ng recombinant IL-22 (R&D System) in 300 μl PBS was intraperitoneally injected. To observe the local effect of IL-22, CCL20, or CXCL1 on corneal inflammation and reepithelialization after wounding, some mice received anti-mouse IL-22, CCL20, or CXCL1, or nonbinding antibody (200 μg/ml; R&D Systems) dissolved in lubricating eyedrops (AdvancedMedical Optics, Santa Ana, CA, USA) every 4 h for 24 h. The control animals received an equal concentration of nonblocking IgG antibody in eyedrops.

In vivo and in vitro corneal epithelial wounding model

The central corneal wound was performed as described previously (7, 15–17). In brief, mice were anesthetized, and the central corneal epithelium was demarcated with a 2-mm trephine and then removed using a Golf Club Spud for refractive surgery (Accutome, Malvern, PA, USA) under a dissecting microscope. All animals were bred and housed in our facility according to the guidelines described in the Association for Research in Vision and Ophthalmology Statement for the Use of Animals in Vision and Ophthalmic Research and Baylor College of Medicine Animal Care and Use Committee policy. At various times following injury, mice were also killed, and corneal tissue including ∼1 mm outside the apparent limbal area was excised and processed for immunohistology, ELISA, or mRNA isolation.

For the in vitro wound model, 2-mm corneal epithelial wounds were also created. Debriding the epithelium within a central circular wound was as described above. Six corneas with the complete limbus (1 mm) were excised and placed in a single well of a 24-well culture plate with 1 ml of culture medium. The medium was made up as follows: 48.5 ml Advanced DMEM:F12 (Life Technologies, Carlsbad, CA, USA), 1 ml FBS (Atlanta Biologicals, Atlanta, GA, USA), 0.5 ml 100× Pen/Strep with or without mouse recombinat IL-22 (R&D Systems) for different culture times (6, 12, and 18 h) in conditions of 5% CO₂ at 37°C. Corneas and culture supernatant at each time point were collected and frozen at −70°C for the ELISA assay.

Immunofluorescence and deconvolution imaging

The corneal whole-mount technique was as described previously (7, 15–17). In brief, wounded corneas with the complete limbus were dissected, fixed (2% formaldehyde), permeabilized (0.01% Triton X-100), and incubated with the following labeled antibodies: anti-TCRδ PE or FITC (hamster IgG, GL3 clone; BD Pharmingen; Franklin Lakes, NJ, USA) and anti-TCRδ APC (hamster IgG, GL3 clone; eBiosciences, San Diego, CA, USA); anti-CD31 (IgG2a, MEC 13.3 clone; BD Pharmingen); anti-CD41 (rat IgG1, MWReg30 clone; BD Pharmingen); anti-mouse NK1.1 (mouse IgG2a, PK136 clone; eBiosciences) and anti-NKp46 (goat IgG-FITC; R&D Systems); anti-mouse CD27-PE (hamster IgG1, LG3A10; eBiosciences); anti-mouse Ly6G (rat IgG2a, 1A8 clone; BD Pharmingen); anti-mouse IL-22 PE (rat IgG2a, clone 1H8PWSR; eBiosciences); anti-mouse IL-22R1 (rat IgG2a, Clone 496514 clone; R&D Systems); anti-mouse CCR6 FITC (rat IgG2a, 140706 clone; R&D Systems); anti-mouse IL-23R (goat IgG-PE; R&D Systems); and anti-mouse RORγt (rat IgG1, B2D clone; eBiosciences). Controls using IgG isotype- and species-matched antibodies were in all cases negative. Radial cuts were made in the cornea so that it could be flattened by a coverslip, and the cornea was mounted in Airvol (Celenase, Dallas, TX, USA), containing 1 μM 4′,6-diamidino-2-phenylindole (DAPI; Sigma Chemical, St. Louis, MO, USA) to assess nuclear morphology. Image analysis and quantification of wounded corneas were performed using Delta Vision (Applied Precision, Issaquah, WA, USA) as described previously (7, 15–17). The limbus was evident in all whole mounts (limbal vessels stained with anti-CD31) as a reference for counting leukocytes and epithelial cells.

Quantitative real-time PCR

Total RNA was isolated from corneal epithelium with the RNeasy Mini kit (Qiagen, Valencia, CA, USA) according to the manufacturer's protocol. Quantity and quality of the extracted RNA were verified using a Nanodrop-1000 spectrophotometer (Nanodrop Technology, Wilmington, DE, USA). Purified RNA was stored at −80°C until analysis. First-strand cDNA synthesis was performed with the TaqMan reverse transcription kit (Applied Biosystems, Foster City, CA, USA) using 2 μg total RNA, per the manufacturer's recommendations. The resulting cDNA was stored at −20°C until further analysis. For the amplification of target genes, 5 μl cDNA was added to a corresponding 20× TaqMan MGB probe primer set for each message, multiplexed with primers for GAPDH and 2× TaqMan Universal PCR master mix (Applied Biosystems). PCR was performed in a 7500 real-time PCR system (Applied Biosystems) using the manufacturer's suggested thermal settings. Relative mRNA expression was calculated using the Δ comparative threshold (C_t) method. GAPDH was used as internal control. Each experiment was repeated 3 times.

Assessment of epithelial wound healing

Wound closure was measured as described previously (7, 15–17) by photographing the ocular surface at a known scale for calculating the surface area of the wound at each time point. Assessment of cell division and leukocytes in the cornea was performed in corneal whole mounts. Peak epithelial cell division as a measure of epithelial healing was determined as described previously (15), neutrophil accumulation in the cornea was assessed by counting total Ly6G⁺ cells in 4 microscopic fields within the original wound, and platelets (CD41⁺ cells) were counted as described previously (15) in 8 microscopic fields in the limbus. The pattern showing regions for counting is schematically illustrated in Fig. 2B. Histological assessment of corneal epithelial thickness was analyzed in corneal cross-sections at 96 h after epithelial abrasion. We have previously shown that thickness of the stratified epithelium returns to unwounded levels in wild-type mice within 96 h (15, 17). Cross-sections were prepared from enucleated eyes fixed overnight at 4°C in 0.1 M sodium cacodylate buffer (pH 7.2) containing 2.5% glutaraldehyde. The corneas with the limbal tissue were then excised and postfixed in 1% osmium tetroxide for 1 h at room temperature, dehydrated through an ethanol series, and embedded in resin. Thick (0.5 μm) transverse sections passing through the center of the cornea were cut on an ultramicrotome (RMC 7000; Venana Medical Systems, Tucson, AZ, USA) equipped with a diamond knife. Sections were stained with toluidine blue O and viewed on an inverted microscope (DeltaVision Spectris; Applied Precision) using a ×40 objective, and transverse measurements of the central epithelial thickness were made using the calibrated linear measurement tool contained in the supplied imaging software (SoftWorx; Applied Precision).

Figure 2.

Effects of anti-CCL20 on γδ T cells, epithelial division, and inflammation. A, B) Wild-type mice treated intraperitoneally with control or anti-CCL20 antibody before corneal abrasion were analyzed at 18 h after injury for γδ T-cell migration into corneal epithelium (A) or stroma (B). GL3⁺ cells per field are plotted for five 150- × 150-μm fields of view in the corneal limbus (L), paralimbus (PL), parawound (WM), original wound edge (W), and center of cornea (C), using the sampling pattern schematic (B, right panel) for each cornea (n=6). C) Additional corneas were prepared at 12 h after corneal epithelial abrasion from wild-type mice treated intraperitoneally with control or anti-CCL20 antibody or from untreated TCRδ^−/− mice and analyzed for neutrophils (stained with anti-Ly6G) or for platelets (stained with anti-CD41; n=4). D) Epithelial cell division at 18 h after epithelial abrasion (13) was analyzed in wild-type mice treated with control or anti-CCL20 antibody and in TCRδ^−/− mice (n=4). E) Effects of anti-CCL20 and anti-CXCL1 antibodies were analyzed by topical application of antibodies every 4 h for 18 h after epithelial abrasion. Neutrophils were counted as Ly6G⁺ cells in 4 fields at the wound margin; γδ T cells were counted as GL3⁺ cells in 8 fields in the limbus and paralimbus (n=4). *P < 0.01.

ELISA assay

CXCL1 and CCL20 from in vivo or in vitro wounded corneal samples were analyzed by ELISA (R&D Systems) in extracts of corneas collected at different times after wounding. Six corneas (in vivo) at each time point were pooled in 1 ml of RPMI 1640 medium. Six corneas with culture supernatant (in vitro) were collected. All samples were subjected to 3 freeze-thaw cycles and sonication for 30 s at 250 Hz. All homogenates were frozen at −70°C until they were used in the assay.

Statistical analysis

Data analysis was performed using ANOVA and pairwise multiple comparisons using Tukey's test. A value of P < 0.05 was considered significant. Data are expressed as means ± se.

RESULTS

In this model of corneal abrasion, full-thickness epithelium and subbasal nerves are removed without disrupting the stroma of the cornea. Reepithelialization of the wound occurs within 24–30 h, and full-thickness stratification of the wound is present at 96 h after injury (17). Regeneration of the subbasal nerve plexus is much slower, reaching <20% of uninjured density at 96 h after injury (16). The inflammatory response is evident in the limbus within 2 h, with vessel dilatation and neutrophil accumulation that reached a peak at 12–18 h after injury. Neutrophils arrive in the corneal stroma at the wound edge within 6 h and increase throughout the stroma to a peak between 12 and 18 h (13).

In the current study, at 3 h after central epithelial abrasion, the remaining epithelium from the wound edge to the limbus significantly increased expression of a variety of chemokines (Fig. 1A). In earlier studies, we found that the γδ T cells accumulating in the corneal epithelium after abrasion injury are of the CCR6, IL-17A phenotype (16) and that they increase in number, apparently migrating within the epithelium to reach the wound edge within 12–18 h after epithelial abrasion (7, 17). Since CCL20 was expressed at 3 h, we performed ELISA analysis on extracts of a pool of 6 corneas collected 6 h after epithelial abrasion and a pool of 6 uninjured corneas cultured ex vivo for 6 h. Uninjured corneas yielded 1.8 pg/ml of CCL20, abraded corneas collected after 6 h in vivo yielded 29.9 pg/ml, and explants yielded 33.3 pg/ml. Since the injured corneas produced an equivalent level of CCL20 ex vivo as in vivo, it appears that the emigrated leukocytes are not necessary. In addition, quantitative PCR analysis of corneal epithelium from TCRδ^−/− mice revealed a 15-fold increase in CCL20 mRNA at 3 h after central epithelial abrasion compared to unwounded epithelium (P<0.01; n=3). Corneal epithelium is known to express several cytokines (8, 9, 13, 27), but typically, analysis has been at later times with different types of injury or with infection. The epithelium at this time after injury contained leukocytes including CD11c⁺ dendritic cells that are abundant in the periphery of uninjured epithelium (14, 28), and γδ T cells are often in close association with these cells (Fig. 1B).

Figure 1.

A) Quantitative PCR (qPCR) analysis of corneal epithelium at 3 h after central abrasion. Mean values for 3 separate experiments are plotted for chemokines that exhibited significantly (P<0.01; n=3 pools, 6 corneas/pool) increased expression in corneal epithelium following central epithelial abrasion. Values plotted are ratios of values from wounded to unwounded corneas. B) Typical image of cells in the paralimbal epithelium 3 h after abrasion, showing CD11c⁺ dendritic cells (red), γδ T cells (green; arrows), and nuclei of basal epithelial cells (blue).

CCL20 has a dominant influence on γδ T-cell migration into the corneal epithelium after central epithelial abrasion

As a known ligand for CCR6 (29), we evaluated the requirement for CCL20 in the migration of the CCR6⁺ γδ T cells into the corneal stroma and epithelium after central epithelial abrasion. Systemic administration (intraperitoneal injection) of anti-CCL20 30 min before epithelial abrasion reduced by >80% the accumulation of γδ T cells in the epithelium and stroma (Fig. 2A, B). This treatment also reduced the accumulation of neutrophils by >70% in the wound margin and platelets in the limbus by >70% (Fig. 2C). We have previously shown that γδ T-cell depletion significantly reduces epithelial healing, as measured, e.g., by epithelial cell division (7). Anti-CCL20 reduced epithelial cell division to near the levels of mice deficient in γδ T cells (Fig. 2D).

To determine whether anti-CCL20 applied topically would influence the inflammatory response to corneal abrasion, corneas were abraded, and anti-CCL20 was applied every 4 h over an 18-h period to one eye, and control nonbinding antibody was applied topically to the contralateral eye. In this protocol, anti-CCL20 significantly reduced both neutrophil and γδ T-cell accumulation in comparison to their levels in the control eye (Fig. 2E). With the use of the same protocol, anti-CXCL1 significantly reduced neutrophil accumulation at the wound edge but failed to influence the accumulation of γδ T cells (Fig. 2E).

IL-22 is an important cytokine for corneal epithelial wound healing

γδ T cells of the CCR6⁺ IL17A⁺ phenotype, similar to Th17 cells (30), may also express IL-22 (24), a cytokine known to promote healing of some epithelial tissues (31) and to activate keratinocyte proliferation in IL-23-dependent disease states such as psoriasis (32). We evaluated the possible role of IL-22 in corneal epithelial healing. γδ T cells in the limbal and conjunctival epithelium were found to be positive for IL-22, RORγt, and IL-23R (Fig. 3A) and negative for CD27 and NK1.1. NK cells (NKp46⁺ NK1.1⁺ GL3⁻ CD3⁻) found at the junction of the limbus and conjunctiva were not labeled with anti-IL-22 (Fig. 3B). IL22R1 was expressed in resting epithelium and increased (Fig. 3C) after injury, a time when epithelial migration and division are increasing (7), and the corneal epithelial cells were positive for IL-22R1 (Fig. 3D).

Figure 3.

Phenotypic characteristics of γδ T cells in murine corneas after central epithelial abrasion. A) GL3⁺ (anti-TCRδ⁺) cells (γδ T cells) were costained with anti-IL-22, anti-RORγt, or anti-IL-23R; tissue was counterstained with DAPI. Images are deconvolved Z stacks at 0.5-μm steps. B) NKp46⁺ cells were costained with anti-IL-22. Arrows (bottom panel) indicate position of NKp46⁺ cells because they fail to stain with anti-IL-22. Tissue was counterstained with DAPI. C) Gene expression of IL-22R1 in peripheral corneal epithelium at different times after central epithelial abrasion injury. Results are expressed relative to GAPDH ×10⁻³. D) Paralimbal corneal epithelium showing staining with GL3 (anti-TCRδ) and anti-IL-22R1, counterstained with DAPI. Image is a deconvolved Z stack at 0.5-μm steps.

To assess the possible contribution of IL-22 to corneal epithelial healing, anti-IL-22 was administered systemically to wild-type mice 30 min before epithelial abrasion, and wound closure was evaluated at 24 h after injury. Epithelial migration was significantly retarded (Fig. 4A, B). In contrast, systemic administration of rIL-22 in TCRδ^−/− mice restored the typically delayed epithelial migration into the abraded area to a rate near that of wild-type mice (ref. 7; Fig. 4C). Epithelial division was significantly reduced by anti-IL-22 in wild-type mice to levels seen in TCRδ^−/− mice (Fig. 4D), and epithelial division was restored in TCRδ^−/− mice by systemic administration of rIL-22 to levels seen in wild-type mice (Fig. 4D). Topical administration of rIL-22 to TCRδ^−/− mice using the protocol described above significantly increased (325%; n=6; P<0.001) total epithelial cell division across the cornea, most evident in the regions of corneal epithelium not directly damaged at the time of abrasion (Fig. 4E). Topical application of anti-IL-22 reduced epithelial cell division at 18 h after wounding by 52.2% (n=4; P<0.01). Longer-term effects of altering IL-22 were evident at 96 h, a time when normal murine corneas restore full thickness stratified epithelium after the standardized abrasion used in these studies, but TCRδ^−/− mice exhibit significantly reduced healing (Fig. 5A, B). Anti-IL-22 significantly reduced epithelial stratification in wild-type mice and rIL-22 restored stratification to near normal levels in TCRδ^−/− mice.

Figure 4.

Influence of IL-22 on epithelial wound healing. A) Representative images of wild-type murine corneas at 24 h after central epithelial abrasion. Mice were given control nonbinding antibody or anti-IL-22 intraperitoneally 30 min before corneal abrasion. B) After 24 h, corneas were stained by topical application of fluorescein to reveal open wounds, and the percentage of each wound remaining open was assessed (n=6). C) TCRδ^−/− mice received central corneal epithelial abrasion 30 min after intraperitoneal injection of vehicle or rIL-22. Percentage of each wound remaining open was assessed at 12, 24, and 30 h by application of fluorescein (n=6). D) At 18 h following central epithelial abrasion in either wild-type or TCRδ^−/− mice, epithelial cell division was analyzed. Results are expressed as percentage change for TCRδ^−/− mice treated intraperitoneally with rIL-22 vs. vehicle, or wild-type mice treated with anti-IL-22 vs. nonbinding antibody (n=6). E) TCRδ^−/− mice were treated topically after central epithelial abrasion by administration of rIL-22 in lubricating eye solution every 4 h for 18 h (open circles) to one eye, and vehicle only to the contralateral eye. Dividing epithelial cells were determined in 9 microscopic fields across the cornea from limbus to limbus (n=4). Values are means ± se. *P < 0.01.

Figure 5.

Influence of IL-22 on epithelial recovery after central abrasion. A) Representative images of the central cornea showing recovery of stratified epithelium in wild-type (WT) mice treated intraperitoneally with nonbinding antibody or anti-IL-22, or TCRδ^−/− mice treated intraperitoneally with vehicle or rIL-22. B) Measured epithelial thickness from sectioned corneas (n=5). Values are means ± se. *P < 0.05 vs. WT + nonbinding; ^#P<0.05 vs. TCRδ^−/− + vehicle.

Since IL-22 has been shown to induce proinflammatory gene expression in keratinocytes (33), we evaluated corneal inflammation after treatment with rIL-22. In addition to epithelial division, systemic administration of rIL-22 to TCRδ^−/− mice increased platelet localization in the limbus by 280% (P<0.01; n=6) to a level not significantly different from the wild-type control (7, 15). Systemic or topical administration of rIL-22 to TCRδ^−/−mice significantly increased neutrophil numbers at the wound margin (Fig. 6B) to a level not significantly different from wild-type controls. Corneas cultured ex vivo immediately after epithelial abrasion were analyzed by ELISA for production of CXCL1, a chemokine important for migration of neutrophils to the wound edge (Fig. 2E). Without injury, corneal extracts yielded 15 pg/ml of CXCL1; with injury alone and culture ex vivo for 6 h, they yielded 528 pg/ml; and with injury and culture ex vivo with 200 ng/ml of IL-22, they yielded 1201 ng/ml CXCL1 (Fig. 6C). Further evidence for a proinflammatory effect of IL-22 was obtained by systemic or topical administration of anti-IL-22 to wild-type mice. These treatments significantly reduced neutrophils at the wound margin (Fig. 6D).

Figure 6.

Influence of IL-22 on corneal inflammation after epithelial abrasion. A) TCRδ^−/− mice were treated intraperitoneally with vehicle or rIL-22 at the time of central epithelial abrasion. At 12 h after abrasion, corneas were collected and analyzed for platelets (anti-CD41-PE, red) in the region of the limbal vessels (anti-CD31-FITC, green) and neutrophils arriving at the wound margin. B) Quantification of data from A. Solid bar, TCRδ^−/− mice treated topically with vehicle; open bar, = TCRδ^−/− mice treated intraperitoneally with rIL-22; shaded bar, TCRδ^−/− mice treated with rIL-22 topically every 4 h in lubricating eyedrops (n=3). *P < 0.05. C) Six normal or abraded corneas were immediately collected and cultured for 6 h in complete medium without (No injury and Injury bars) or with rIL-22 (40 or 200 ng/ml). Culture media were analyzed for CXCL1 by ELISA. D) Wild-type (WT) mice were treated intraperitoneally with nonbinding antibody or anti-IL-22 30 min before central epithelial abrasion. At 12 h after abrasion, corneas were collected and analyzed for neutrophils arriving at the wound margin. Solid bar, WT mice treated topically with nonblocking antibody; open bar, WT mice treated intraperitoneally with anti-IL-22; shaded bar, WT mice treated with anti-IL-22 topically every 4 h in lubricating eyedrops (n=6). *P < 0.01.

DISCUSSION

In the current murine model, both epithelium and the subbasal nerve plexus are removed in the abraded area (16). Healing thus involves both epithelium and sensory nerves that have a trophic effect for the epithelium (34–36). In the current studies, we focused on the possible roles of leukocytes in epithelial healing, but it is clear that γδ T cells, neutrophils, and platelets are important for both epithelial (7, 15, 17) and corneal nerve recovery (16).

Multiple factors control leukocyte immigration into the cornea in response to injury. ICAM-1, for example, is necessary for γδ T-cell migration into the healing epithelium (17, 37), apparently serving as an adhesive ligand for CD11a/CD18 (LFA-1)-dependent migration. In the current study, the evidence for CCL20 as an important determinant of γδ T-cell migration is consistent with the finding that these cells express CCR6, a receptor for this chemokine (29, 38). However, anti-CCL20 not only reduced the number of γδ T cells responding to the injury but reduced neutrophil and platelet accumulation (Fig. 2). Since the only leukocytes found to be positive for CCR6 in the cornea were γδ T cells, the effect of anti-CCL20 on neutrophils and platelets is likely secondary to inhibition of γδ T cells. Depletion of γδ T cells (TCRδ^−/− or treatment with anti-TCRδ) significantly reduces neutrophil and platelet accumulation (7). Although CCR6 expression has been observed in human neutrophils exposed to combinations of cytokines, such as TNF-α and IFNγ in vitro (39), CCR6⁺ neutrophils were not observed in the abraded murine corneas. CCL20 serves numerous diverse constitutive, protective (e.g., direct antimicrobial activity; ref. 40), and pathogenic functions in a wide array of tissues (38), and in corneal abrasion it appears to be critical for the γδ T-cell-dependent inflammatory response.

While γδ T cells promote an early inflammatory response to corneal abrasion, they may also directly contribute to epithelial healing through expression of IL-22, a cytokine that can induce motility and division of epithelial cells (41). Support for this interpretation comes from the finding that anti-IL-22 significantly reduced wound closure, epithelial cell division, and epithelial stratification in the current model, and rIL-22 restored these parameters of healing in TCRδ^−/− mice to near wild-type levels. However, the corneal γδ T cells also stain positively for IL-17A, they express IL-23R and RORγt, and they are negative for CD27 and NK1.1, characteristics consistent with a known subset of γδ T cells linked to the attraction of neutrophils at sites of inflammation (16, 18, 22–24, 24–26). Current data indicate that both IL-17A (16) and IL-22 (Fig. 6) are involved in the attraction of neutrophils. IL-17A can act at several sites including limbal vessels (42, 43) and bone marrow (44) to influence neutrophil kinetics or emigration, and IL-22 may directly stimulate the IL-22R1⁺ epithelial cells to express chemokines (e.g., CXCL1) chemotactic for neutrophils (33) in the cornea (see Fig. 2). Neutrophils may also promote epithelial healing (16, 45). We obtained evidence in a recent study that neutrophils may rapidly deliver VEGF-A (16), a known trophic factor promoting neurite generation (46–51), through prestored VEGF-A in their granules (52–56). VEGF-A protein levels in the healing corneas were substantially influenced by neutrophil influx (16), and topical application of anti-VEGF-A antibody inhibited early neurite generation and epithelial cell recovery. Neutrophils are now known to be rich sources of prestored and expressible proteins (45, 57, 58) that may directly promote wound healing in the cornea. Neutrophils in large numbers can delay wound healing (15, 59) but in controlled levels promote healing (13, 15, 60, 61). Although γδ T cells may promote healing directly, a more significant contribution may be their ability to attract neutrophils and platelets.

In addition to neutrophils, platelets contribute to corneal wound healing (15). Platelets accumulate around intact blood vessels in the limbus (16), and their depletion significantly reduces corneal nerve regeneration (16), epithelial wound closure, and neutrophil influx (15). The γδ T cells in the cornea induce not only neutrophil but also platelet accumulation (7). Wild-type mice treated before corneal epithelial abrasion with antibody GL3 (that transiently blocks TCRδ) exhibit >65% reduction in both neutrophils and platelets within the first 18 h after injury (7). In corneal tissue, the recruitment of platelets and neutrophils appears to be interdependent (15), since depletion of neutrophils also significantly reduces platelet localization (15), a finding consistent with observations in another microvascular bed (62). Platelet P-selectin appears to be necessary, since the deficient colocalization of neutrophils and platelets in the limbus of P-selectin-deficient mice is partially restored by the passive intravenous transfer of isolated wild-type platelets (15). In addition to a role in neutrophil recruitment, platelets may directly support corneal healing through an array of growth factors that can be rapidly released on activation (63–67), including VEGF-A, which can induce early sensory nerve regeneration (16, 46–51).

In summary, evidence supports CCL20 as a major contributor to the migration of CCR6⁺ IL-17⁺ IL-22⁺ IL-23R⁺ RORγt⁺ γδ T cells into the corneal epithelium following central corneal epithelial abrasion, and IL-22 as a major contributor to epithelial wound healing. Considering previously published data and the results in this study, we argue that both CCL20 and IL-22 initiate mechanisms that induce neutrophil and platelet accumulation in the wounded cornea, cells necessary for efficient wound healing. We propose that CCL20 recruits γδ T cells, which, in turn, recruit neutrophils and platelets through IL-17-dependent pathways, and that IL-22 recruits neutrophils and platelets by activating IL-22R1-bearing epithelial cell expression of proinflammatory chemokines (e.g., CXCL1).