2.1 Virus titers in WT, CXCL9−/−, and CXCL10−/− mice

Mice infected with HSV-1 were assayed at times post infection (pi) for virus load in the eye. Whereas no differences were found with the cornea in the first 5 days pi, by day 7 pi CXCL10 −/− mice possessed significantly more HSV-1 in comparison to WT or CXCL9−/− mice (Fig. 1). In a similar fashion, virus shedding in the tear film was elevated at day 7 pi but not earlier time points surveyed in the CXCL10−/− mice compared to WT or CXCL9−/− mice. Further time points were not assessed since the majority of CXCL10−/− mice succumb to the infection between day 7 and day 9 pi. Since virus replication was unchanged in the CXCL10−/− mice until day 7 pi, the results imply the adaptive rather than innate immune response is most likely affected by the loss of CXCL10 expression.

2.2 CD4⁺ T cell infiltration into the cornea is reduced in CXCL9−/− mice

Since increases in virus shedding in the tear film and virus titer within the cornea were observed in CXCL10 deficient mice by day 7 pi, inflammation was assessed at this time point in all groups of mice measuring cytokine/chemokine expression and T lymphocyte infiltration. In measuring a battery of chemokines including CCL2, CCL3, CCL5, CXCL1, CXCL9, and CXCL10, with the exception of CXCL10, only CCL2 levels were modified (elevated) in the cornea of the CXCL10−/− mice in comparison to either the WT or CXCL9−/− mice (Table 1). In contrast, there was a noticeable reduction in CCL3, CCL5, CXCL1, and CXCL10 expression in the cornea of CXCL9−/− mice in comparison to WT and CXCL10−/− mice (Table 1). Likewise, IL-1β, IL-6, and TNF-α levels were not detectable or were significantly lower in the cornea of CXCL9−/− mice compared to the other two groups (Table 1). The expression of these cytokines/chemokines is in response to infection since uninfected mouse corneas contain significantly lower levels of the assayed analytes [11].

CXCL9 and CXCL10 recruit T cells to sites of inflammation thru CXCR3 [25], and T cells infiltrate the cornea during the latter stages of acute HSV-1 infection [1]. Therefore, T cell trafficking into the cornea was investigated following virus infection. There were no detectable CD3⁺ T cells in the cornea of uninfected mice and less than 10 cells in the peripheral cornea proximal to the ciliary body at day 3 pi (data not shown). By day 7 pi, CD3⁺CD4⁺ T cells were readily detectable in both the central and peripheral cornea in all groups of mice (Fig. 2). A reduction in the recruitment of CD4⁺ T cells to the peripheral and central cornea was found in CXCL9−/− mice compared to the WT controls (Fig. 2). Unlike the results in CXCL9−/− mice, CD4⁺ T cell infiltration into the cornea of CXCL10−/− mice was remarkably similar to WT mice (Fig. 2). CD8⁺ T cells are also known to infiltrate the cornea following virus infection depending on the strain of HSV-1 and mouse host employed for the study [6, 26]. Moreover, in the absence of CD4⁺ T cells, CD8⁺ T cells have been found to infiltrate the cornea and elicit transient immunopathology within the stroma following HSV-1 infection [27]. Therefore, it was possible that CD8⁺ T cells would compensate for the reduction in CD4⁺ T cells found in the cornea of CXCL9−/− mice following HSV-1 infection. However, a similar pattern of CD3⁺CD4⁻ T cell infiltration was observed in the CXCL9−/− mice compared to WT mice (Fig. 3). Taken together, the results reflect a deficiency of T cell trafficking into the cornea of CXCL9−/− mice following ocular HSV-1 infection.

Since changes in the trafficking of CD4⁺ T cells did not explain the increased sensitivity of the CXCL10−/− mice to HSV-1 infection, NK cell recruitment was investigated day 7 pi. NK cells are thought to contribute to stromal keratitis [28], control HSV-1 infection [29, 30], and respond to a CXCL10 chemical gradient through the receptor CXCR3 [31]. However, similar numbers of NK cells were found residing in the cornea comparing HSV-1-infected WT (225 ± 70 cells/cornea pair), CXCL9−/− (260 ± 90 cells/cornea pair), and CXCL10−/− (190 ± 90/cornea pair) mice.

In this study, the local infiltration of NK and T cells into the cornea following infection does not reflect host control of the virus. Specifically, CXCL10−/− mice possess significantly more HSV-1 in the cornea by day 7 pi compared to WT or CXCL9−/− mice and yet, displayed similar levels of NK and T cell recruitment compared to WT mice. The deficiency in the control of virus replication by CXCL10−/− mice may reside within the nervous system. CXCL10−/− mice possess significantly more HSV-1 in the innervating sensory ganglia for the cornea, the trigeminal ganglia (TG) within the first 72 hr of infection [Wuest et al., manuscript in preparation]. This inability to control HSV-1 replication may translate into an increase in virus found in the cornea during the latter stages of the acute infection. To this end, zosteriform spread of the virus from the TG back to the eye including the tear film, eye proper, and periocular tissue is noted only in virus that is capable of retrograde transport to the TG during acute infection [32].

A reduction in chemokine expression in the cornea of HSV-1-infected CXCL9−/−mice correlates with the drop in T cell recruitment. As activated T cells not only express the receptor for CXCR3 but also express CCR1, CCR2, and CCR5 for the ligands CCL3 and CCL5 among others, the drop in T cell recruitment into the cornea of the CXCL9−/−mice is consistent with the absence or significant reduction in CXCL9, CCL3, and CCL5 levels found in the tissue in response to HSV-1. The results also suggest chemokine expression in the cornea is not driven by levels of antigen alone since CXCL9−/− mice possessed similar titers of HSV-1 compared to WT mice and yet, showed dramatically reduced cytokine/chemokine levels. Such results are consistent with a recent finding in which highly resistant IFN-α1 transgenic mice display nearly identical levels of chemokines including CCL3, CCL5, CXCL9, and CXCL10 expressed in the cornea as their WT counterparts in response to ocular HSV-1 infection [33].

Both resident corneal cells and infiltrating leukocytes generate the chemokines detected in the present study including CXCL9 and CXCL10 [11, 34–37]. Out of the six chemokines investigated, only CCL2 levels were associated with sensitivity to infection since CXCL10−/− mice possessed significantly more virus and expressed elevated levels of CCL2 in the cornea in comparison to the less sensitive WT or CXCL9−/− mice. CCL2 has been linked to increased pathogenesis and mortality in mice infected with HSV-1 [38] perhaps thru a T_H1 to T_H2 switch [39]. However, we have not been able to detect IL-4 and rarely detect IL-10 at the protein level in the corneas of these mice. Moreover, the pattern of IFN-γ expression in the cornea of WT, CXCL9−/−, or CXCL10−/− mice does not support the change or reduction of T_H1 cytokine expression as an explanation for the increased sensitivity of CXCL10−/− mice to ocular HSV-1 infection.

The present results are inconsistent with a previous study in which the administration of anti-CXCL10 antibody to mice resulted in a reduction in edema and infiltrating mononuclear cells in the cornea in comparison to isotype antibody-treated mice following corneal infection with HSV-1 [12]. In the antibody depletion study, ICR mice which are more sensitive to HSV-1 compared to the C57BL/6 background used in the current study were employed. Likewise, the use of the anti-CXCL10 antibody did not neutralize 100% of the targeted chemokine but rather, reduced the level detectable in the infected tissue. The contribution of CXCL10 by resident cells in comparison to hematopoietic cells recruited to the inflamed tissue or residing in the organized lymphoid organs on maturation and trafficking of effector T and NK cells in response to ocular HSV-1 infection is currently unknown. Obviously more work is warranted evident by the inconsistent results in the current study using CXCL10−/− mice and the previous study incorporating neutralizing antibody to CXCL10 [12]. However, the discrepancy between these two studies is not unique since experimental models of multiple sclerosis show inconsistent results using anti-CXCL10 antibody versus CXCL10−/− mice [40,41].

Collectively, the current findings underscore the unique contribution CXCL9 and CXCL10 maintain in the cornea following ocular HSV-1 infection. The absence of CXCL10 renders the host sensitive to virus replication whereas CXCL9 deficiency has a significant impact on T cell recruitment to the cornea. The results emphasize the non-redundant nature of these CXCR3 ligands and the need to further study their role in the host in response to ocular infection.

3.1 Mice

Wild type (WT) male and female C57BL/6 (The Jackson Laboratory, Bar Harbor, ME), CXCL9−/− [44], and CXCL10−/− [22] mice (6–12 weeks of age) were housed in sterile cages under pathogen free conditions. CXCL9−/− mice were backcrossed onto a C57BL/6 background 8 generations and CXCL10−/− mice were backcrossed 9 generations. All mice were used in accordance with a detailed protocol approved by the Institutional Animal Care and Use Committee of The University of Oklahoma Health Sciences Center and the Dean A. McGee Eye Institute.

3.2 HSV-1 infection

The corneas of anesthetized mice were scarified with a 25 gauge needle prior to infection with 1,000 plaque forming units (PFU) of HSV-1 (McKrae strain) in a volume of 3 μl of RPMI 1640 medium. Mice were anesthetized and perfused with PBS (pH 7.4) at times post infection. The corneas were removed and processed for cytokine/chemokine content by suspension array or ELISA or prepared for whole mount. Alternatively, the corneas were removed and homogenized in RPMI-1640 containing 10% FBS. Following homogenization, the supernatant was clarified (10,000 x g, 1 min) and subsequently assessed for virus content by plaque assay [12]. Uninfected mice served as controls.

3.3 Virus and cells

The McKrae strain of HSV-1 originally obtained from Bryan Gebhardt (LSU Health Sciences Center, New Orleans, LA) was propagated in Vero cells and maintained at −80° C at a concentration of 2 x 10⁸ PFU/ml. Frozen stocks were diluted in RPMI 1640 medium immediately prior to use. Plaque assays using African green monkey kidney fibroblasts (Vero cells, ATCC CCL-81, American Type Tissue Culture Collection, Manassas, VA) were performed in RPMI-1640 medium supplemented with 10% FBS (Invitrogen, Calsbad, CA), gentamicin (Invitrogen), and antibiotic-antimycotic solution (Invitrogen) at 37° C, 5% CO₂, and 95% humidity as previously described [45].

3.4 ELISA

The detection of CCL2, CCL3, CCL5, CXCL1, IL-1β, IL-6, TNF-α, and IFN-γ were performed using a suspension array system (Bio-Rad, Hercules, CA) with a sensitivity of 1–2 pg/tissue for each targeted analyte. The corneas were removed from infected mice following perfusion at day 7 pi and homogenized in T-PER tissue protein extraction reagent (Pierce Chemical Co., Rockford, IL). Each sample was analyzed in duplicate along with known amounts of cytokine/chemokine provided by the kits to generate a standard curve used to determine the quantity of each cytokine/chemokine in a given sample. CXCL9 and CXCL10 were detected using commercially available kits (R&D Systems, Minneapolis, MN) according to the manufacturer’s instructions with a sensitivity of 25–35 pg/cornea. In this case, corneal samples were prepared as previously described [12]. The amount of cytokine/chemokine measured was normalized to the total wet weight of each cornea.

3.5 Whole mount preparation

The corneas from infected and uninfected perfused mice were removed from isolated eyes using a dissecting microscope (Fisher) and placed into 1.5 ml microcentrifuge tubes containing 4% paraformaldehyde (Sigma Chemical Co., St. Louis, MO) in PBS. After a 20 min incubation period at room temperature, the corneas were washed 5X with 1.0 ml PBS containing 1% Triton X-100 (Sigma) for 10 min/wash. Following the last wash, corneas were soaked in 100 μl of PBS-BGEN (1X PBS containing 3% BSA, 0.25% gelatin, 5 mM EDTA, and 0.025% nonidet 40) and 1 μl of anti-mouse CD16/32 (BD Pharmingen) for 60 min at room temperature. Following the incubation period, 5 μl of rat serum (Vector Laboratories, Burlingame, CA) was added and the samples were incubated overnight at 4° C. Next, the solution was removed, and the corneas were incubated with 2 μl of alexa fluor 647 conjugated anti-mouse CD3 (clone 17A2, BD Pharmingen) and either FITC conjugated anti-mouse CD4 or CD8 in 100 μl PBS-BGEN overnight at 4° C. The following day, the solution was removed, and the corneas were washed 5X in PBS containing 1% BSA and 0.1% Triton X-100 for 15 min/wash. The corneas were then soaked in Vectashield mounting medium containing DAPI (Vector Laboratories) overnight at 4° C and subsequently mounted on slides for subsequent analysis by confocal microscopy.

3.6 Confocal microscopy

Corneas (n=3/group/experiment) were imaged using an Olympus IX81-FV500 epifluorescence/confocal laser-scanning microscope with a UApo 40x water immersion lens. Samples were excited with 405, 488, 546, and 633 nm wavelength lasers (Olympus, Melville, NY). Scanning images were taken with a step size of 2 μm in the z-axis and image analysis was performed using FLUOVIEW software (Olympus). CD3⁺CD4⁺ and CD3⁺CD4⁻ T cells were enumerated from 6 corneas/group/experiment (3–4 experiments/group) of mice consisting of central and peripheral sections.

3.7 Preparation of cornea samples for detection of NK cells by flow cytometry

Corneal buttons from saline-perfused, HSV-1-infected WT, CXCL9−/−, and CXCL10−/− mice were removed 7 days following infection. The corneas were cut into small (1–2 mm segments) pieces and exposed to 1 mg/ml collagenase type I (Sigma Chemical Co., St. Louis, MO) for 60 minute triturating the sample every 20 minutes. Single cell suspensions were washed twice (300 x g, 5 min/wash) in PBS and then incubated on ice for 15 min with 4 μl anti-mouse Fc block (BD Pharmingen, San Diego, CA) in a total volume of 50 μl PBS-1% BSA. Following the incubation, the cells were centrifuged (300 x g, 5 min) and resuspended in 5% normal rat serum (Jackson Immuno Research Inc., West Grove, PA) for an additional 15 min on ice. Cells were then triple labeled with 6 μl containing 2 μl FITC-conjugated anti-mouse CD3, 2 μl phycoerythrin-Cy5-conjugated anti-CD45 (clone 30-F11), and 2 μl PE-conjugated anti-mouse NK1.1 and incubated in the dark on ice for 30 min. Following the incubation period, the cells were washed 3 times with PBS-1% BSA (300 x g, 5 min/wash) and resuspended in PBS containing 1% paraformaldehyde. After overnight fixation at 4° C in the dark, cells were pelleted (300 x g, 5 min) and resuspended in PBS-1% BSA. Immediately before analysis, CountBright absolute counting beads (Invitrogen, Eugene, OR) were added (28,000 beads/sample). Cell suspensions were gated on CD45^high labeled cells, and the percentage of NK cells (NK1.1⁺CD3⁻) were determined at this gate setting. A second gate was established to count the number of beads that passed through during the run (500 sec). The absolute number of NK cells was determined by calculating the number of input beads/28,000 X the percent NK cells in the CD45^high-gated sample.

3.8 Statistics

One-way ANOVA and Tukey’s post hoc t-test were used to determine significance (p<.05) of differences between WT, CXCL9−/−, and CXCL10−/− mice for each parameter under measure. All statistical analyses were carried out using the GBSTAT program (Dynamic Microsystems, Silver Springs, MD).