Loss of neurokinin-1 receptor alters ocular surface homeostasis and promotes an early development of herpes stromal keratitis

Abstract

Substance P neuropeptide and its receptor neurokinin-1 (NK1R) are reported to present on the ocular surface. In this study, mice lacking functional NK1R exhibited an excessive desquamation of apical corneal epithelial cells in association with an increased epithelial cell proliferation, increased epithelial cell density, but decreased epithelial cell size. The lack of NK1R also resulted in decreased density of corneal nerves, corneal epithelial dendritic cells, and a reduced volume of basal tears. Interestingly, massive accumulation of CD11c+CD11b+ conventional dendritic cells (cDCs) was noted in the bulbar conjunctiva and near the limbal area of corneas from NK1R^−/− mice. After ocular HSV-1 infection, the number of cDCs and neutrophils infiltrating the infected corneas was significantly higher in NK1R^−/− than C57BL/6J mice. This was associated with an increased viral load in infected corneas of NK1R^−/− mice. As a result, the number of IFN-γ secreting virus specific CD4 T cells in the DLNs of NK1R^−/− mice was much higher than infected C57BL/6J mice. An increased number of CD4 T cells and mature neutrophils (CD11b+Ly6g^high) in the inflamed corneas of NK1R^−/− mice was associated with an early development of severe HSK. Collectively, our results show that the altered corneal biology of uninfected NK1R^−/− mice along with an enhanced immunological response after ocular HSV-1 infection cause an early development of HSK in NK1R^−/− mice.

Introduction

Neurokinin-1 receptor (NK1R) is the highest affinity receptor for substance P (SP), an eleven amino acid long neuropeptide (1, 2). NK1R is reported to express in corneal epithelial cells, and SP-NK1R interaction promotes the chemokine expression in primary cultures of human corneal epithelial cells (3, 4). Blocking of NK1R signaling, while using NK1R antagonists, is shown to ameliorate many pro-inflammatory conditions, including airway and ocular inflammation in animal models (5–8). Therefore, NK1R serves as a promising target to control chronic inflammation. Currently, NK1R antagonists are approved to prevent nausea and vomiting associated with cancer chemotherapy in clinics (9, 10). In addition to promoting inflammation, NK1R signaling is also reported to accelerate corneal epithelial wound healing in animal models (11, 12). However, the role of functional NK1R in regulating ocular surface homeostasis under steady-state condition is not known.

Recurrent corneal HSV-1 infection can cause the development of herpes stromal keratitis (HSK), a chronic ocular inflammatory condition. If not controlled with steroids and anti-viral, HSK can cause the loss of vision. Mouse models have long been used to study the pathogenesis of HSK. Recently in a mouse model, we demonstrated that blocking NK1R signaling during the clinical period of the disease with spantide I (NK1R antagonist), ameliorated the severity of HSK (7). This suggested the pro-inflammatory nature of NK1R in an ongoing inflammatory condition However, it is not clear whether the lack of functional NK1R on the ocular surface under steady-state condition increases or decreases the susceptibility of eyes to develop HSK after corneal HSV-1 infection.

To address the above question, we used mice lacking functional Nk1r gene (NK1R^−/−) (13). Contrary to our expectation, an early development of severe HSK was noted in infected NK1R^−/− mice when compared with infected C57BL/6J (B6) mice. While trying to understand the mechanism, we determined that in comparison to uninfected B6 mice, uninfected NK1R^−/− mice exhibited excessive cell sloughing at the apical surface of the corneal epithelium in association with an increased epithelial cell proliferation, increased epithelial cell density, but decreased epithelial cell size. Additionally, a significant decrease in the number of resident corneal epithelial dendritic cells, but an increased accumulation of cDCs near limbal area was detected in naïve corneas of NK1R^−/− mice than wild type B6 mice. Upon ocular HSV-1 infection, increased infiltration of cDCs and neutrophils was detected in the infected corneas from NK1R^−/− mice. In addition, NK1R^−/− mice corneas exhibited an increased viral titer at early time-points (days 2 and 4) post-infection. This was associated with an increased priming of virus specific IFN-γ secreting CD4 T cells in the DLNs of NK1R^−/− mice. An increased number of CD4 T cells and mature neutrophils (CD11b+Ly6g^high) in inflamed corneas of NK1R^−/− mice was associated with an early development of severe HSK. Together, our results indicate the contribution of NK1R signaling in maintaining the homeostasis of the ocular surface under steady-state condition, and that the lack of functional NK1R increases the susceptibility of eyes to develop severe HSK upon ocular HSV-1 infection.

Materials and Methods

Mice and Ethics statement

Eight to twelve week old male and female C57BL/6J mice were procured from the Jackson laboratory (Bar Harbor, ME). Breeding pairs of NK1R^−/− mice were obtained from Dr. Norma P Gerard, and the mice were bred in an animal facility at Wayne State University School of Medicine (WSUSOM). Functional ablation of the NK1R gene in NK1R^−/− mice was confirmed using tail biopsy and performing the PCR followed by electrophoresis (Supplementary Figure 1). Eight to twelve week old male and female B6 and NK1R^−/− mice were used to carry out the experiments. All of the animals were housed in an Association for Assessment and Accreditation of Laboratory Animal Care (AAALAC) approved specific pathogen free animal facility at Wayne State University School of Medicine (WSUSOM). The Institutional Animal Care and Use Committee (IACUC) of Wayne State University approved all of the experimental protocols and procedures. In addition, all of the experimental procedures were in complete agreement with the Association for Research in Vision and Ophthalmology resolution on the use of animals in research.

Corneal HSV-1 infection and viral titration

A virulent HSV-1 RE Tumpey strain used in the current study was propagated on a monolayer of Vero cells (American Type Culture Collection, Manassas, VA; CCL81) as described previously (14). To carry out an ocular HSV-1 infection, mice were first anesthetized by intra-peritoneal injection of Ketamine (33mg/Kg bodyweight) + Xylazine (20mg/Kg bodyweight) in 0.2 mL PBS. The corneas were scarified with trephine (Fine Science Tools, Foster city, CA) while twisting three to four times over the corneal surface. 1×10⁴ plaque forming unit (p.f.u) of HSV-1 virus was then topically applied with a pipette to the eye in 3μL of 1x PBS followed by a gentle massage of the eyelids. The HSV-1 load in infected corneas of C57BL/6J and NK1R^−/− groups of mice was determined on day 2, 4 and 5 post-infection as described previously (14). The same infected eye was not used to measure the viral load at different days post-infection. After collecting the eye-swabs, mice were euthanized on the indicated time-points post-infection and the corneas were processed for flow cytometry analysis.

Clinical Scoring of HSK

The eyes were examined on different day post-infection, while using a hand-held slit lamp biomicroscope (Kowa, Nagoya, Japan), to determine the extent of corneal opacity and angiogenesis. A standard scale for corneal opacity, ranging from 0–5, was used as described earlier (14). The neovascularization (NV) of the cornea was determined by measuring the centripetal growth of newly formed blood vessels in each quadrant of the cornea as described earlier (14).

Lymph nodes and corneal cell preparation for flow cytometry

HSV-1 infected B6 and NK1R^−/− mice were euthanized on day 7 post-infection. A single cell suspension of the draining lymph nodes (DLNs) from individual mouse was prepared as described earlier (14). The infected eyes from both groups of mice were enucleated and collected in ice-cold RPMI medium with antibiotics. Under a dissecting microscope, a radial incision was made at the limbal region, and the cornea was separated from the underlying lens, iris, ciliary body, and scleral tissue using curved fine forceps (Miltex surgical instruments, PA). The individual cornea was suspended in 250 μL of RPMI, and 20 μL of liberase TL (2.5 mg/mL) was added followed by incubation at 37°C for 45 minutes on a tissue disruptor. At the end of the incubation period, the samples were triturated using 3 mL syringe plungers and passed through a 70 μm cell strainer. This was followed by pelleting down the cells at 315× g for 8 minutes in a refrigerated centrifuge.

Cell surface staining

Cell surface staining was carried out as described earlier (14). Briefly, the single cell suspension obtained from the individual cornea and DLN was washed with FACS buffer, followed by blocking of Fc receptors and incubation with fluorochrome conjugated antibodies. The following antibodies were used for the cell surface staining: Percp-Cy5.5 conjugated- anti-CD4 (RM4-5), FITC conjugated anti-Gr1 (RB6-8C5), Percp-Cy5.5 conjugated anti-CD11b (M1/70), PE-Cy7 conjugated anti-Ly6g (1A8), PE conjugated IA^b (AF6-120.1), and Alexa 647 conjugated anti-CD11c (N418). All of the antibodies were purchased from BD biosciences and BioLegend, San Diego, CA. At the end of the cell surface staining, samples were acquired using a LSR II flow cytometer, and the data was analyzed using the FlowJo software (Ashland, OR, USA. V8.8.7).

Corneal FITC painting assay

HSV-1 infected C57BL/6 and NK1R^−/− mice were anesthetized with isoflurane on day 2 post-infection. A 1% FITC solution was prepared by dissolving FITC isomer 1 (Sigma F7250) in 1:1 (vol/vol) acetone: dibutyl phthalate solution. The corneas of infected mice were painted under anesthesia with 3 μl of FITC solution. After 16h, the DLNs were collected and a single cell suspension was prepared while using a collagenase type II (Gibco 17101-015) enzyme solution. Cells were stained for IA^b, CD11b, and CD11c molecules and the samples were acquired on a LSRFortessa flow cytometer. The data was analyzed using the FlowJo software.

Intracellular cytokine and nuclear Ki-67 staining

Single cell suspension of the lymph nodes was stimulated with 3 MOI (multiplicity of infection) of UV inactivated HSV-1 as described earlier (14). At the end of incubation period, cell surface staining was carried out, followed by permeabilization of the cells with Cytofix/Cytoperm (BD biosciences). PE-conjugated anti-IFN-γ (XMG1.2) antibody was used to stain IFN-γ expressing CD4 T cells. The cells were washed in perm wash (BD Biosciences) buffer with the final washing given in FACS buffer. For nuclear Ki-67 staining, cells were stained using a PE mouse anti-human Ki-67 kit (BD Pharmingen) as per manufacturer instruction. Samples were then acquired on the LSR II flow cytometer and the data was analyzed using the FlowJo software.

Mouse cytokine protein array analysis

Cytokine and chemokine levels in HSV-1 infected corneal tissues from B6 and NK1R^−/− mice were measured using a Mouse Cytokine Array Panel A kit (R&D systems, Minneapolis, MN). Corneas dissected from naïve or infected eyes (day 4 and day 9 post-infection) of B6 and NK1R^−/− mice were transferred to 1x PBS with protease inhibitor cocktail (PIC) and kept at −80°C. Sonication of the corneas was performed at 50% amplitude with a cycle of 15 second pulse, followed by 1 minute of resting on ice. A total of six cycles were given to each cornea. The tissue lysates (sonicated samples) obtained were centrifuged at 4°C at 15,000 rpm for 10 minutes. The supernatant was collected and the amount of protein in each sample was estimated using a BCA protein assay kit (Thermo Scientific). A total of four corneas from each group was pooled to obtain a minimum of 200 μg of protein for cytokine and chemokine array analysis as per the manufacturer instructions. Positive signals on the membranes were quantified using Image Studio Lite Version 4.0.

Corneal whole mount staining

Mice were euthanized and their eyes were enucleated with a pair of forceps. The eyeballs were immediately fixed in 4% paraformaldehyde for 30 min at room temperature. The corneal tissue was separated with a bard-parker #21 blade under a dissection microscope. Tissues were flattened with 6–8 partial cuts from the limbal to central cornea and stored in 30% sucrose solution overnight at 4°C. The next day, the corneas were permeabilized using 0.1% EDTA-0.01% Hyaluronidase type IV-S solution at 37°C for 90 min, and immediately followed by blocking with a 3% bovine serum albumin-1% Triton X100 solution for 90 min at room temperature. The corneal tissues were incubated with a primary antibody at an appropriate dilution in a humidified chamber (Supplementary table 1) for 16 hours in a cold room (4°C). The next day, the tissues were washed four times with 0.3% Triton X100-PBS solution (10 min each wash). If the primary antibody was unconjugated, a secondary antibody was added at an appropriate dilution as stated in Supplementary table 1 followed by overnight incubation in a cold room. Prior to acquisition, the corneal tissue was mounted with vectashield medium containing DAPI (H1200, Vector Laboratories, Burlingame, CA, USA). Images were acquired using a Leica TCS SP8 confocal microscope. An Image J version 1.49b software was used to quantitate the CD11c+ cells in the corneal wholemount from both groups of mice. The Image J “multi-point” tool was used for manual counting of the corneal nerve leash and limbal nerve trunk branch points.

Cochet and Bonnet Esthesiometry

Corneal sensitivity was measured using Cochet and Bonnet Esthesiometer attached with 12/100 mm nylon thread. This test was performed at 23–25°C room temperature by touching the central cornea of B6 or NK1R^−/− mice. Numerical values were derived from a conversion table provided by the manufacturer (Luneau SAS).

Phenol red thread tear test

To measure the tear levels in the eyes of unmanipulated B6 and NK1R^−/− mice, pH indicator phenol red treated cotton threads were obtained from Zone-Quick (ZQ-1010, Showa Yakuhin Kako, Japan). The mice were restrained by holding the lose skin at base of the neck without touching the facial skin or whiskers. The eyes were tested for tear level by placing the tip of the cotton thread on the bulbar conjunctiva of the lateral canthus for 30 seconds. The picture of the threads was taken and tear absorption was measured using a 10 cm scale.

Bromodeoxyuridine (BrdU) incorporation assay

For BrdU labeling of the corneal epithelium, 8 to 12-week old B6 and NK1R^−/− mice were given a single intra-peritoneal injection of BrdU (10 mg BrdU/mL saline; 0.2 mL/mouse). The mice were euthanized on the next day after 18 hours, and their eyes were collected to snap freeze in OCT media. Longitudinal sections were cut at an 8μm thickness, and mounted on positively charged slides (Thermo Shandon Limited, UK). The sections were air dried overnight at room temperature and fixed in 4% buffered paraformaldehyde for 15 min followed by 3 washings with 1x PBS. For antigen retrieval, sections were immersed in 2N HCl in a covered Coplin jar for 30 minutes at room temperature. The acid was neutralized with 0.1M Borate buffer (pH-9.0) for 5 minutes with two buffer changes. The slides were blocked with 3% BSA in 1X PBS and incubated with a primary antibody (Abcam, MA, USA, BU1/75, ab6326) at a 1:100 dilutions in blocking buffer for 18 hrs at 4°C. The next day, the slides were washed three times and incubated with Alexa-488 conjugated anti-Rat antibody (Molecular Probes) at a 1:200 dilutions overnight at 4°C. The slides were washed and mounted with vectashield medium containing DAPI prior to acquisition.

Statistical Analysis

Statistical analysis was carried out using the Graph Pad Prism software (San Diego, CA). All p values were calculated using an unpaired, two-tailed Student’s t test. P < 0.05 was considered statistically significant. The results are displayed as the mean.

Results

Ocular HSV-1 infection causes an early development of corneal opacity and angiogenesis in NK1R^−/− mice

Previously, we determined that blocking NK1R signaling in HSV-1 infected inflamed corneas, during clinical period of disease, ameliorated the severity of HSK in a B6 mouse model (7). In this study, we analyzed whether the lack of functional NK1R on the ocular surface, under steady-state condition, also reduces the severity of HSK upon ocular HSV-1 infection. The corneas from both groups of mice were infected with HSV-1 as described in the methods section. Unexpectedly, our results showed significantly increased corneal opacity and angiogenesis on day 9 and 11 post-infection in HSV-1 infected corneas from NK1R^−/− mice in comparison to the infected group of B6 mice (Fig. 1A, C and D). The incidence of corneal opacity and angiogenesis was also much higher in NK1R^−/− mice when analyzed on day 9 and 11 post-infection (Fig. 1B). By day 14 and 16 post-infection, the extent of corneal opacity and angiogenesis was similar in eyes from both groups of mice. Collectively, our results show an early development of severe HSK in NK1R^−/− mice after ocular HSV-1 infection.

Figure 1. Lack of NK1R promotes an early development of HSK.

(A) Representative slit lamp pictures of B6, NK1R^−/− mice eyes on day 11 post-ocular HSV-1 infection. (B) Frequency of eyes with severe opacity in both groups of mice. (C and D) Scatter plots showing corneal opacity and angiogenesis grading on day 9, 11, 14 and 16-post ocular infection. Data shown is the representative of three independent experiments (n = 5–10 mice per group). p values were calculated using two-tail student’s t-test (****p<0.0001, **p<0.01, *p<0.05 significant and ns p>0.05 non-significant).

Excessive sloughing of apical corneal epithelial (ACE) cells in unmanipulated NK1R^−/− mice

To determine the underlying mechanism for our unexpected results, we microscopically evaluated the corneal surface of the eyes from uninfected B6 and NK1R^−/− mice. No phenotypic differences in the appearance of the eyes were noted when comparing B6 and NK1R^−/− mice (Fig. 2A). However, hematoxylin and eosin staining of paraffin-embedded corneal sections from NK1R^−/− mice showed a much thinner apical layer of the corneal epithelium with occasional delaminating superficial epithelial cells in comparison to the corneas from B6 mice (Fig. 2B). The laser scanning confocal microscopy of whole mount corneas from B6 and NK1R^−/− mice demonstrated significantly increased number of desquamating ACE cells (denoted by the * sign) in knockout mice (Fig. 2D). Excessive cell sloughing at the corneal apical surface could possibly stimulate compensatory cell proliferation in the corneal epithelium. Therefore, we next evaluated the corneal epithelial cell proliferation between B6 and NK1R^−/− mice. Short-term Bromodeoxyuridine (BrdU) incorporation assay showed a significant increase in labeling of basal corneal epithelial cells in NK1R^−/− compared to B6 mice (Fig. 2E). In addition, a significant increase in the number of hematoxylin stained nuclei was determined in the corneal epithelium of NK1R^−/− than B6 mice (Fig. 2C). Increased epithelial cell density, but decreased epithelial cell size was also evident in the whole mount corneas from NK1R^−/− mice (Fig. 2F). Together, our results demonstrated an excessive exfoliation and increased mitotic index of corneal epithelial cells in NK1R^−/− mice.

Figure 2. Lack of NK1R causes an excessive sloughing of apical corneal epithelial cells.

(A) Slit lamp bio-microscope images of naïve WT (top) and NK1R^−/− (lower) showing corneal surface. (B) Representative H & E stained corneal tissue sections (8μ thick) from both groups at 20X magnification, boxed area shows the higher magnified view of that region. Apical corneal epithelial cell sloughing from the ocular surface of NK1R^−/− mice indicated by arrow heads. Scale bar-60 μm. (C) Graphical representation for the number of epithelial cells quantified from H&E stained corneal sections of B6 and NK1R^−/− mice. (D) Whole mount naïve corneas stained for ZO-1 (green) in B6 and NK1R^−/− mice. Asterisks and scatter plot denote empty space resulting from the desquamation of ACE cells in outermost layer of the corneal epithelium. Scale bar-50 μm. (E) Representative corneal sections stained with anti BrdU Ab after 18 hours of BrdU pulsing in WT and NK1R^−/− mice (BrdU positive cells indicated by arrows). Scale bar-50 μm. (F) Whole mount naïve corneas stained with E-cadherin (green) and counterstained with DAPI to demonstrate an increased epithelial cell density but decreased epithelial cell size at wing cell zone in NK1R^−/− mouse cornea. Scale bar-25 μm. Images are representative of two independent experiments (n = 3–6 corneas per group). p values were calculated using two-tail student’s t-test (****p<0.0001, **p<0.01 significant).

Loss of NK1R reduces the density of corneal epithelial DCs, but dramatically increases the number of cDCs in the bulbar conjunctiva and near the limbal area of uninfected cornea

The normal corneal epithelium is populated with resident CD11c+ DCs, which are in close association with surrounding epithelial cells. The resident corneal DCs have been shown to promote the re-epithelialization of corneal wounds (15). An alteration in the homeostasis of the corneal epithelium might affect the homeostasis of corneal DCs. Therefore, we next compared the density of corneal DCs by carrying out the immunofluorescence staining of CD11c+ cells in the corneal whole mounts from unmanipulated B6 and NK1R^−/− mice. As is shown in Fig. 3A, the resident corneal epithelial DCs present in the peripheral and central areas of the corneas from NK1R^−/− mice were almost half in number when compared to DCs in the corneas from B6 mice. However, a more than two-fold increase in the number of CD11c+ cells stained with anti-CD11c (clone N418) antibody was evident in the bulbar conjunctiva and limbal region of the corneas from NK1R^−/− mice when compared to the limbal area of the corneas from B6 mice (Fig. 3B and supplementary Figure 2). The majority of CD11c+ cells near the limbal area was also co-stained with CD11b molecule (Fig. 3C). Moreover, z-scans of the corneal wholemount near the limbal area showed that CD11b+CD11c+ cells were localized in both the corneal epithelium and anterior stroma, whereas CD11b+CD11c− macrophages were deep seated in the stroma of the corneas from both groups of mice (Supplementary movie S1). Together, our results showed an altered homeostasis of corneal resident epithelial DCs along with an increased accumulation of conventional dendritic cell type near the limbal area of the corneas from unmanipulated NK1R^−/− mice.

Figure 3. Lack of NK1R changes the homeostasis of corneal epithelial and limbal dendritic cells.

(A) Whole mount (montage images with 10X) corneas of naïve B6 and NK1R^−/− mice stained with FITC conjugated anti-CD11c antibody (HL3 Clone). Scatter plot shows the total number of CD11c positive cells/each cornea from both the groups. (B) Low magnified limbal region of WT and NK1R^−/− whole mount corneas stained with FITC conjugated anti-CD11c antibody (N418 clone, 5X). Dotted line denotes the limbal region. Scale bar-200 μm. Scatter plot shows the total number of CD11c positive cells at limbus/cornea from both the groups. (C) Co-localization of CD11c (N418) (Green) and CD11b (Red) in the limbal region of NK1R^−/− mice corneal whole mounts. Nucleus was stained with DAPI (Blue) (20X). Scale bar-50 μm. Data shown is the representative of three independent experiments (n = 3–4 corneas per group). Cj-Conjunctiva, C Cornea-Central cornea. p values were calculated using two-tail student’s t-test. (**p<0.01 and *p<0.05 significant).

Reduced epithelial nerve density, stromal nerve trunk branching and basal level of tears in the corneas from NK1R^−/− mice

The corneal epithelium is a highly innervated tissue and the corneal nerves are reported in close association with corneal epithelial cells and dendritic cells (16, 17). We next ascertained whether the altered homeostasis of corneal epithelial and dendritic cells influences the density of corneal nerves. The corneal whole mount staining for class III β-tubulin, using Tuj-1 antibody, showed a significant decrease in the number of corneal subbasal nerve leashes in the peripheral region of the corneas from NK1R^−/− mice (Fig. 4A). Additionally, the number of corneal stromal nerve trunk branch points near the limbal area was also significantly more reduced in NK1R^−/− than B6 mice (Fig. 4A). The corneal nerve density was similar in the central corneal region when comparing both groups of mice (data not shown). Moreover, no significant differences (p>0.05) in corneal sensation were determined in the central corneal region of the eyes from both groups of mice, while using the Cochet-Bonnet Esthesiometer (Fig. 4B). Lastly, the phenol red thread tear test was carried out to evaluate the volume of unstimulated tears in the eyes from NK1R^−/− and B6 mice. Our results showed a significant decrease (p<0.0001) in the volume of unstimulated tears measured in the eyes of NK1R^−/− than B6 mice (Fig. 4C).

Figure 4. Lack of NK1R reduces the corneal nerve density and basal levels of unstimulated tears.

(A) Confocal images of Tuj1 stained whole mount corneas from uninfected B6 and NK1R^−/− mice. Images are montages images acquired with 20X objective lens. Scatter plots denote number of nerve leashes (bunches) and limbal stromal nerve trunk branch points from both groups of mice (n = 4 corneas per group). (B) Measurement of corneal sensitivity with Cochet and Bonnet Esthesiometer. (C) Image of phenol red threads used for measuring the tear quantity in B6 (top) and NK1R^−/− (bottom) mice. The extent of color change from the thread edge denotes the abundance of tears. Scatter plot represents the volume (in mm) of tear absorption by phenol red coated cotton thread from both the groups. Data shown is the representative of two independent experiments (n= 10 corneas per group). p values were calculated using two-tail student’s t-test. (****p<0.0001, **p<0.01 and ns p>0.05).

Increased viral load in the corneas of NK1R^−/− mice

The intact apical surface of the corneal epithelium provides resistance to the establishment of HSV-1 infection in the mouse model. Our results clearly showed that in NK1R^−/− mice, the apical layer of the corneal epithelium is much thinner and has more desquamating ACE cells. Therefore, viral load of HSV-1 on the corneal surface was compared at different time-points post-infection between B6 and NK1R^−/− mice as described in the methods. Our results showed a moderate, but statistically significant increase in viral load in the corneas of NK1R^−/− mice when compared with B6 mice on day 2 and 4 post-infection (Figure 5). No significant difference in viral load was determined between both groups of mice on day 5 post-infection.

Figure 5. Increased viral load in corneas of NK1R^−/− mice.

The viral load in individual corneas from both groups of mice was measured as plaque forming unit (p.f.u.) on day 2, 4, and 5 post-ocular infections. Each circle represents an individual eye from HSV-1 infected mice. Data shown is the sum of two independent experiments (n = 8–10 corneas per group). p values were calculated using two-tail student’s t-test. (*p<0.05 significant and p>0.05 non-significant).

Pre-clinical and clinical phase of HSK document a differential level of cytokine/chemokines in infected corneas of NK1R^−/− and B6 mice

Development of HSK in a mouse model is categorized into a pre-clinical and a clinical phase of disease as reported earlier (18). During the pre-clinical phase (day 1 through 6 post-infection), active viral replication is reported with minimal corneal opacity and angiogenesis whereas, in the clinical stage of the disease (day 7 though 18 post-infection) severe corneal opacity and angiogenesis develop in response to chronic immunoinflammatory reactions in the inflamed corneas. We ascertained whether the presence or absence of functional NK1R affect the level of cytokines/chemokines in HSV-1 infected corneas during the pre-clinical and clinical phase of HSK. Infected corneas obtained from both groups of mice on day 4 (pre-clinical) and day 9 (clinical phase) post-infection were sonicated as described in the method section. Uninfected corneas from both groups of mice were also processed to determine the basal level of cytokines/chemokines. The amount of cytokines in the corneal lysates from both groups of mice was measured using a Mouse Cytokine Array assay kit as per the manufacturer’s instructions. Our results demonstrated that infected corneal lysates from NK1R^−/− mice exhibited a marked increase in the amounts of neutrophil attracting chemokines CXCL1 (KC), CXCL2 (MIP-2) and CCL2 (MCP-1) in comparison to infected B6 mice, when measured on day 4 post-infection (Fig. 6). We also noted an increased amount of IL-1 receptor antagonist (IL-1Ra) in infected corneas of NK1R^−/− than B6 mice on day 4 post-infection. On the other hand, during the clinical phase of HSK (day 9 post-infection), reduced levels of a number of cytokines/chemokines including the tissue inhibitor of metalloproteinase (TIMP-1), CCL2, and interleukin-1 receptor antagonist (IL-1ra) was measured in the corneal lysates of NK1R^−/− than B6 mice (Fig. 6). No difference in the amounts of TIMP-1, CCL2, and IL-1ra protein was seen in uninfected corneal lysates from both groups of mice.

Figure 6. Differential amount of cytokine and chemokines in infected corneas of NK1R^−/− and B6 mice during pre-clinical and clinical phase of HSK.

The cytokine and chemokine array blots of the corneal lysates prepared from uninfected, and HSV-1 infected corneas of B6 and NK1R^−/− mice on day 4, and day 9 post-ocular infection. Bar diagram denote quantitation of the molecules involved in regulating the pathogenesis of HSK. A mouse cytokine array layout is provided to read the protein array blot images. Four corneal samples were pooled to carry out the assay. n = 8 corneas per group.

Increased number of cDCs and neutrophils in infected corneas of NK1R^−/− mice at an early time point after corneal HSV-1 infection

In light of our observation demonstrating the massive accumulation of cDCs near the corneal limbal area of unmanipulated NK1R^−/− mice (Fig. 3), we investigated the influx of cDCs in the peripheral, paracentral, and central region of HSV-1 infected corneas of B6 and NK1R^−/− mice at an early time-point post-infection. As is shown in the corneal whole mount staining of CD11c+ cells, NK1R^−/− mice exhibited an increased influx of CD11c+ cells in all three regions of infected corneas as stated above on day 2 post-infection (Fig. 7A). Moreover, the flow cytometry analysis of HSV-1 infected corneas on day 2 post-infection showed that almost all of CD11c+ cells co-expressed CD11b molecule, suggesting their cDCs phenotype (Fig. 7B). Our results also showed an increased influx of neutrophils in infected corneas of NK1R^−/− than B6 mice on day 2 and 4 post-infection (Fig. 7C). Collectively, an increased immune cell influx was detected in infected corneas of NK1R^−/− mice at early time-points post-infection.

Figure 7. Increased influx of cDCs and mature neutrophils in HSV-1 infected corneas of NK1R^−/− than B6 mice.

(A) Corneal whole mounts stained with FITC conjugated anti-CD11c (clone N418) antibody from B6 and NK1R^−/− mice on day 2 post-ocular HSV-1 infection. More magnified limbal, paracentral and central corneal images for both groups are included (n = 3–4 corneas per group). (B) Representative FACS plots denote that majority of infiltrating CD11c+ cells co-expressed CD11b molecule (n = 5 corneas per group). Graphs show frequency and absolute number of CD11b+CD11c+ cells in infected corneas on day 2 post-infection. (C) Representative FACS plots showing influx of leukocytes (CD45+) and neutrophils (CD11b+Ly6G+) in HSV-1 infected corneas on day 2 and 4 post-infection from both groups of mice. Graphs demonstrate the frequency and absolute number of leukocytes and neutrophils in infected corneas (n = 6 corneas per group). Data shown is the representative of two independent experiments. p values were calculated using two-tail student’s t-test. (*p<0.05, **p<0.01, ***p<0.001 significant and p>0.05 non-significant).

Increased numbers of CD11c+ dendritic cells observed in the DLNs of infected NK1R^−/− mice did not migrate from the infected corneas

The infiltrating DCs can pick up the viral antigens and migrate to the draining lymph nodes (DLNs) to prime virus specific T cells (19). Therefore, we next ascertained the number of CD11c+ cells in DLNs of HSV-1 infected B6 and NK1R^−/− mice at early time-points post-infection. Our results showed a significant increase in the absolute number of CD11c+ cells in the DLNs of NK1R^−/− than B6 mice on day 3 post ocular infection (Fig. 8A). No significant difference in the basal number of CD11c+ DCs was determined in the DLNs from both groups of uninfected mice. To address whether increased numbers of CD11c+ dendritic cells measured in the DLNs of infected NK1R^−/− mice are the result of more CD11c+ cells migrating from the infected cornea to the DLN, we carried out FITC-painting of the infected corneas from both groups of mice. FITC painting of virus infected corneas was carried out on day 2 post-infection, and 16h later the number of IA-^{b high}CD11C+CD11b+FITC+ cells were quantitated in the DLNs isolated from both groups of mice (Fig. 8B). Unexpectedly, our results showed a lesser number of FITC+CD11b+CD11c+ cells in the DLNs of NK1R^−/− than C57BL/6 mice, when measured on day 3 post-infection. Taken together, our results showed that the increased number of CD11c+ DCs noted in the DLNs of infected NK1R^−/− mice is not the outcome of an increased migration of cDCs from the infected corneas to DLNs.

Figure 8. Increased number of CD11c+ dendritic cells in the DLNs of HSV-1 infected NK1R^−/− mice did not migrate from the infected corneas.

(A) FACS plots demonstrating the frequencies of CD11c+Gr1-ve DCs in naïve DLNs, and the DLNs obtained on day 3 post-infection, from both groups of mice. Bar diagrams denote the absolute number of DCs (n = 4–5 mice per group). (B) FACS plots showing the frequency of FITC+ DCs in the DLNs of infected mice, from both groups, on day 3 post-infection. The infected corneas of B6 and NK1R^−/− mice were FITC-painted on day 2 post-infection and 16h later, the DLNs were removed. Bar diagrams show frequency and absolute number of FITC+ DCs in the DLNs of B6 and NK1R^−/− mice on day 3 post-infection (n = 6 mice per group). Data shown is the representative of two independent experiments. (*p<0.05, **p<0.01, ****p<0.0001 significant and p>0.05 non-significant).

Increased expansion of CD4 T cells and IFN-γ secreting virus specific CD4 T cells in NK1R^−/− mice

Our results showed an increased viral load in infected corneas of NK1R^−/− mice. Therefore, we next ascertained whether this results in an increased priming of CD4 T cells in the DLNs of NK1R^−/− mice. The proliferation of CD4 T cells in the DLNs of infected B6 and NK1R^−/− mice was compared by measuring the expression of Ki-67 nuclear antigen on day 3 and 5 post-infection using a flow cytometer. A significant increase in the proliferation of CD4 T cells in the DLNs of NK1R^−/− mice was noted on day 3, but not on day 5 post-infection (Fig. 9A). To address HSV-1 antigen-specificity of CD4 T cells in the DLNs of infected B6 and NK1R^−/− mice, a single cell suspension of the DLNs from both groups of infected mice was prepared on day 7 post-infection. As described in the methods section, the single cell suspension from both groups of mice was stimulated with UV-inactivated HSV-1 followed by intracellular cytokine staining to determine the number of IFN-γ secreting CD4 T cells. As is shown in Fig. 9B, about a two-fold increase in the number of IFN-γ secreting virus-specific CD4 T cell was detected in the DLNs of infected NK1R^−/− mice, when compared to infected B6 mice.

Figure 9. Increased expansion and Th1 differentiation of CD4 T cells in DLNs of HSV-1 infected NK1R^−/− mice.

(A) Representative FACS plots are showing the frequencies of Ki67 expressing CD4 T cells in the DLN of B6 and NK1R^−/− mice on day 3 post-ocular HSV-1 infection. Scatter plots show frequency and absolute number of CD4+ Ki67+ cells in DLNs of uninfected and HSV-1 infected mice from both groups on day 3 and 5 post-infection (n = 3–6 mice group). (B) Representative FACS plots are denoting IFN-γ producing CD4 T cells after ex-vivo re-stimulation of lymph node cells from both groups of mice with UV inactivated HSV-1 on Day 7 post-infection. Scatter plots demonstrate frequency and absolute number of IFN-γ producing CD4 T cells from both groups of mice (n = 9–10 mice per group). p values were calculated using two-tail student’s t-test. (****<0.0001, *p<0.05 significant and p>0.05 non-significant)

Increased influx of CD4 T cells and neutrophils into the inflamed corneas of infected NK1R^−/− mice

CD4 T cells and neutrophils play a pivotal role in the development of HSK (20, 21). Therefore, we next ascertained the influx of CD4 T cells in inflamed corneas of HSV-1 infected NK1R^−/− and B6 mice. Single cell suspensions of individual corneas obtained from B6 and NK1R^−/− mice on day 11 and 17 post-infection were stained for CD4 T cells, and the samples were acquired on a BD flow cytometer. Our results showed an increased frequency and absolute number of CD4 T cells in infected corneas of NK1R^−/− mice on day 11 and 17 post-infection (Fig. 8). Similarly, an increased frequency and absolute number of CD11b^highLy6G^high mature neutrophils were seen in inflamed corneas of NK1R^−/− mice on day 11 post-infection (Fig. 10).

Figure 10. Increased influx of CD4 T cells and mature neutrophils (CD11b^highGr1^high) in infected corneas of NK1R^−/− mice.

(A) Representative FACS plots of CD4 T cells in infected corneas from both groups of mice on day 11 and 17 post-infection. Bar diagrams demonstrate the frequencies and absolute number of CD4 T cells in individual corneal tissue from both groups on day 11 and day 17 post-infection. Data shown is the representative of two independent experiments (n = 5–7 corneas per group). (B) Representative FACS plots of CD11b^highGr1^high granulocytes in inflamed corneas on day 17 post-infection from both groups of mice. Bar diagrams denote frequency and absolute number of CD11b^high cells in infected corneas from both the groups of mice on day 17 post-infection. (n = 7 corneas per group). p values were calculated using two-tail student’s t-test. (***p<0.001, **p<0.01, and *p<0.05 significant).

Discussion

Our results clearly demonstrate an altered homeostasis of naive uninfected corneas in NK1R^−/− mice. This involves lesser tear volume, reduced nerve density, reduced numbers of resident corneal epithelial DCs, increased accumulation of cDCs near the corneal limbal area, and an excessive exfoliation of apical corneal epithelial cells. As a result of this, the corneas of NK1R^−/− mice are more susceptible to HSV-1 infection, as is evident from an increased viral load measured in the infected corneas at early time-points post-infection. An increased viral load was associated with the generation of an increased number of viral antigen specific CD4 T cells (as measured with intracellular cytokine assay) in NK1R^−/− mice. Collectively, an altered corneal biology of uninfected NK1R^−/− mice along with an enhanced immunological response after ocular HSV-1 infection cause an early development of HSK in NK1R^−/− mice as depicted in the schematic of Figure 11.

Figure 11. Schematic of events in uninfected and HSV-1 infected corneas of NK1R^−/− mice to cause an early development of HSK.

Cells in the apical layer of the corneal epithelium of uninfected NK1R^−/− mice slough off prematurely due to an abnormal localization or expression of junctional complex proteins. This leads to an increased basal epithelial mitotic index, reduced corneal nerve density, and reduced volume of tears. Lack of NK1R survival signal is the most likely cause of the reduced number of resident corneal CD11c+ cells noted in NK1R^−/− mice. Increased CD11b+CD11c+ cells found near the limbal area of uninfected NK1R^−/− mice is possibly the outcome of low-grade inflammation developed due to lesser volume of tears. Altered ocular surface in NK1R^−/− mice ease the establishment of corneal HSV-1 infection, and an increased influx of neutrophils in infected cornea further facilitate HSV-1 replication. Increased amount of viral antigens drained to the DLNs cause an increased expansion of HSV-1 specific effector CD4 T cells, which migrate to the infected cornea to cause an extensive tissue damage. Together, an altered ocular surface along with an enhanced immunological response after ocular HSV-1 infection cause an early development of severe HSK in NK1R^−/− mice.

Our results showed an excessive exfoliation of apical corneal epithelial (ACE) cells in uninfected NK1R^−/− mice. The exfoliation of ACE cells is regulated by tight junction proteins and adherens junction proteins. The ACE cells at the apical surface form tight junctions (TJs) and serve as a barrier to protect the ocular surface from microbial infections (22, 23). In addition, the adherens junction (AJ) proteins in the corneal epithelium are involved in stabilization of cell-cell adhesion (24). Thus, degradation, improper localization or reduced expression of membrane bound tight junction or adherens junction proteins in the corneal epithelium may increase the rate of exfoliation of ACE cells. NK1R signaling has been shown to enhance the expression of E-cadherin adhesion and ZO-1 tight junction protein in cultured human corneal epithelial cells (25, 26). This suggest that the lack of functional NK1R could cause a reduced expression or an improper localization of E-cadherin AJ and ZO-1 tight junction proteins in the corneal epithelium resulting in an excessive exfoliation of ACE cells, as determined in uninfected NK1R^−/− mice.

To maintain corneal epithelium homeostasis, excessive loss of ACE cells should be compensated by hyper-proliferation of basal epithelial cells, as determined in the corneas from NK1R^−/− mice (Fig. 2). Additionally, rapid exfoliation of ACE cells may also impact the innervation of corneal nerves, as the latter has been reported to form complex structures with ACE cells (16). Due to the excessive loss of ACE cells, the ends of nerve fibers innervating the corneal epithelium may not make an intimate association with superficial corneal epithelial cells, and thereby retract from the corneal epithelium. Furthermore, the retraction of nerves fibers from the selective places in the corneal epithelium may also reduce the branching of stromal nerve trunks as depicted in our results.

Afferent sensory nerves in the cornea are reported to play an important role in regulating tear secretion from the lacrimal gland through a neural reflex arc, which originates from the ocular surface (27). The reduced innervation of the corneas, as determined in NK1R^−/− mice, might affect the neural reflex arc, and thereby reduce the secretion from the lacrimal gland, resulting in a lesser volume of basal tears as measured in the eyes of NK1R^−/− mice. In addition, excessive exfoliation of ACE cells may also compromise tear film retention on the corneal surface, and promote corneal desiccation. This may also increase the severity of HSK upon ocular HSV-1 infection. In fact, a strong correlation between corneal desiccation, decreased density of corneal nerves, and pathogenesis of HSK are reported in human and mouse models of HSK (28–30).

It is widely accepted that the corneal epithelium is populated with CD11c+ dendritic cells with long dendritic processes, which cover a significant area of the non-inflamed cornea (31). These DCs and their dendritic processes are in close association with corneal epithelial cells (15). However, the nature of the interaction between epithelial cells and resident DCs under steady state condition in corneal epithelium is not clear. In normal skin, E-cadherin adhesion protein mediate the interaction between epidermal keratinocytes and Langerhans cells (LCs), and this interaction is required for the retention of LCs in the skin (32). An aberrant expression of E-cadherin in human papillomavirus infected cervical tissue was associated with a significant reduction in Langerhans cells (33). Considering the role of SP-NK1R interaction in regulating the expression of E-cadherin in corneal epithelial cells (26), the lack of functional NK1R on the ocular surface may cause an aberrant expression of E-cadherin in the corneal epithelium, resulting in a decrease in the number of corneal epithelial DCs as noted in NK1R^−/− mice. Moreover, signaling through NK1R has also been shown to promote the survival of bone-marrow derived DCs (34). Thus, the lack of NK1R on the ocular surface may result in the apoptosis of corneal epithelial resident DCs, and thereby decrease their number as determined in corneas of NK1R^−/− mice.

The massive accumulation of cDCs noted in the bulbar conjunctiva and near the corneal limbal area of NK1R^−/− mice could be due to persistent low-grade inflammation in the conjunctival tissue of these mice. In support, we observed an increased number of Gr1+ and NKp46+ cells near the corneal limbal area of uninfected NK1R^−/− than B6 mice (data not shown). Moreover, an increased vasodilation of feeder vessels was detected near the limbal area of NK1R^−/− mice when compared with B6 mouse normal corneas (Supplementary Figure 3). Vasodilation of blood vessels along with an enhanced expression of selectins on vascular endothelial cells are reported to cause an increased infiltration of leukocytes from the peripheral blood into the skin tissue (35). At present, the possible cause of an increased vasodilation, as depicted in the feeder blood vessels near the corneal limbal area of NK1R^−/− mice, is not clear. However, we suspect that it may be due to conjunctival inflammation in NK1R^−/− mice. We are currently investigating this possibility.

The severity of HSK in a mouse model is largely determined by viral load and the number of effector CD4 T cells and neutrophils in infected corneas. Increased viral load noted in NK1R^−/− mice is possibly the outcome of two events. First, the easiness in establishing the primary HSV-1 infection, as the apical surface of NK1R^−/− mice is disrupted due to excessive exfoliation of ACE cells. An increased viral load was associated with an increased influx of neutrophils as shown in our results on day 2 and day 4 post-infection (Figure 7). A recent study showed that infiltrating neutrophils in HSV-1 infected corneas can facilitate HSV-1 replication and dissemination (36). Thus, increased neutrophils in NK1R^−/− mice may play a role in increasing the viral load at early time-points post-infection. The possible cause of an increased influx of neutrophils is the higher amounts of neutrophil attracting chemokines (CCL-2, CXCL1 and CXCL2) measured in infected corneas of NK1R^−/− mice.

An increased viral load may cause an increased generation of effector CD4 T cells as noted in the DLNs of NK1R^−/− mice. A recent study showed that CD4 T cell priming in DLNs of HSV-1 infected mice could be due to the drainage of soluble viral antigens from infected corneas or the migration of cornea-infiltrating DCs loaded with viral antigens to the DLNs (19). Our results showed an increased number of cDCs in infected corneas and the DLNs of NK1R^−/− than B6 mice. However, the results obtained from the corneal FITC-painting assay, to determine DC trafficking to DLNs, showed a lesser number of corneal cDCs migrating to the DLNs in NK1R^−/− mice. DC migration is dependent upon their maturation (37), and NK1R signaling is reported to promote the maturation of DCs (38). The lack of functional NK1R on DCs in NK1R^−/− mice may affect their maturation, and thereby reduce their migration from infected corneas to DLNs. Thus, the increased priming of CD4 T cells detected in DLNs of NK1R^−/− mice is possibly the outcome of the drainage of soluble viral antigens from infected corneas.

Together, our findings characterized a novel role of NK1R in maintaining the homeostasis of the ocular surface under a steady state condition. Moreover, the lack of NK1R increases the susceptibility of the eyes to ocular HSV-1 infection resulting in an increased incidence and early development of HSK. Several NK1R antagonists are in clinical trials for non-ocular conditions, including alcoholism and psychiatric conditions. It would be of interest to determine whether long-term treatment of NK1R antagonists, given for non-ocular conditions, increases the incidence of HSK or other microbe-induced ocular pathologies.