The following primary antibodies (BD Pharmingen, San Diego, California) were used for immunohistochemical staining: FITC-conjugated rat antimouse CD11b (monocyte/macrophage marker, catalog No. 557396; isotype FITC-conjugated rat antimouse IgG2bk, catalog No. 553988), purified hamster antimouse CD3e (T-cell marker, catalog No. 553057; isotype purified hamster IgG1, catalog No. 553969), biotin-conjugated rat antimouse GR-1 (neutrophil marker, catalog No. 553124, isotype biotin-conjugated rat IgG2b, catalog No. 553987), biotin-conjugated rat antimouse Iab (C57BL/6 major histocompatibility complex [MHC] Class II marker, catalog No. 553546; isotype biotin-conjugated mouse IgG2bk, catalog No. 559531). The secondary antibodies (Jackson Laboratories, Bar Harbor, Maine) included Cy3-conjugated goat anti-Armenian hamster (code No. 127165-160) and Cy3-conjugated Strepavidin antibodies (code No. 016-160-084). For whole-mount immunofluorescence corneal staining, freshly excised corneas were washed in PBS and acetone fixed for 15 minutes. Nonspecific staining was blocked with anti-FcR CD16/CD32 antibody (BD Pharmingen, catalog No. 553142), and Strepavidin and Biotin blocking solutions (Vector Laboratories, Burlingame, California). Next, the specimens were immunostained with primary or isotype antibodies for 2 hours, washed with PBS, incubated with secondary antibodies, and mounted using Vector Shield mounting medium (Vector Laboratories). Whole-mount corneal images were taken using confocal microscope (Leica TCS 4D; Lasertechnik, Heidelberg, Germany). Cells were counted in 8 to 10 areas each in the periphery (0.5-μm area from the limbus) and the center (central 2-μm area) of the cornea in a masked fashion using an epifluorescence microscope (model E800; Nikon, Melville, New York) at ×40 magnification. The mean number of cells were obtained by averaging the cell number in the 8 to 10 areas studied. Cell number was compared using 1-way analysis of variance (ANOVA), followed by pairwise comparisons adjusted for multiple comparisons by the least significant difference method. P values less than .05 were deemed statistically significant.