The Secreted Ly6/uPAR-Related Protein 1 (Slurp1) Modulates Corneal Angiogenic Inflammation Via NF-κB Signaling

Abstract

Purpose: Previously we demonstrated that the secreted Ly-6/uPAR related protein 1 (SLURP1), abundantly expressed in the corneal epithelium (CE) and secreted into the tear fluid, serves as an antiangiogenic molecule. Here we describe the Slurp1-null (Slurp1X^−/−) mouse corneal response to silver nitrate (AgNO₃) cautery.

Methods: Five days after AgNO₃ cautery, we compared the wild-type (WT) and Slurp1X^−/− mouse (1) corneal neovascularization (CNV) and immune cell influx by whole-mount immunofluorescent staining for CD31 and CD45, (2) macrophage and neutrophil infiltration by flow cytometry, and (3) gene expression by quantitative RT-PCR. Quantitative RT-PCR, immunofluorescent staining, and immunoblots were employed to evaluate the expression, phosphorylation status, and subcellular localization of NF-κB pathway components.

Results: Unlike the WT, the Slurp1X^−/− corneas displayed denser CNV in response to AgNO₃ cautery, with more infiltrating macrophages and neutrophils and greater upregulation of the transcripts encoding VEGFA, MMP2, IL-1b, and vimentin. At 2, 7, and 10 days after AgNO₃ cautery, Slurp1 expression was significantly downregulated in the WT corneas. Compared with the WT, naive Slurp1X^−/− CE displayed increased phosphorylation of IKK(a/b), elevated phosphorylation of IκB with decreased amounts of total IκB, and higher phosphorylation of NF-κB, suggesting that NF-κB signaling is constitutively active in naive Slurp1X^−/− corneas.

Conclusions: Enhanced angiogenic inflammation in AgNO₃ cauterized Slurp1X^−/− corneas and constitutively active status of NF-κB signaling in the absence of Slurp1 suggest that Slurp1 modulates corneal angiogenic inflammation via NF-κB signaling.

The cornea, a transparent and refractive tissue that enables sight by focusing the incident light on the retina, also serves a protective function safeguarding the rest of the eye from external dust, allergens, and pathogens.¹ Corneal avascularity, a prerequisite for its transparence, is supported by a host of antiangiogenic factors.^2–11 Any disruption in their function initiated by any angiogenic insult results in corneal neovascularization (CNV) involving invasion of new blood vessels into the cornea from the limbus. CNV is sight-threatening, affects more than a million patients a year worldwide, and is a common reason for corneal graft failure. Alkali burn injury, one of the more common causes of CNV, initiates oxidative stress, elevates reactive oxygen species (ROS), and activates NF-κB signaling culminating in CNV.¹² Diverse molecules and pathways confer angiogenic and immune privilege to the cornea, enabling it to tolerate mild insults without eliciting acute inflammatory response.^2–11 The secreted Ly-6/uPAR-related protein 1 (SLURP1) is one such immunomodulatory peptide that helps the cornea remain free of angiogenic inflammation while exposed to mild insults.^13–16 

SLURP1 is a member of the Ly6/urokinase-type plasminogen activator receptor (uPAR) family of proteins characterized by a three-finger structure with five disulfide bridges and is structurally similar to α-bungarotoxin.^15–26 It influences intracellular signal transduction, cell adhesion, and immune response and serves as a tumor suppressor.^{23,24,27–29} Previously, we demonstrated that SLURP1, abundantly expressed by the corneal epithelium (CE) and secreted into the tear fluid, serves as an immunomodulatory peptide that (1) regulates uPAR activity by acting as a soluble scavenger of urokinase-type plasminogen activator (uPA),³⁰ (2) inhibits leukocyte infiltration in healthy corneas,¹⁵ (3) inhibits tube formation by human umbilical vein endothelial cells (HUVECs),³¹ (4) suppresses neutrophil chemotaxis and transmigration through confluent monolayer of endothelial cells,³² (5) suppresses TNF-α–induced cytokine production,³³ and (6) stabilizes epithelial cell junctions.³³ SLURP1 is also expressed in the skin, where it regulates keratinocyte functions through cholinergic pathways via α7-nicotinic acetylcholine receptor.^{22,26,28,34–39} Mutations in SLURP1 cause Mal de Meleda, an autosomal recessive inflammatory palmoplantar keratoderma (PPK).^{22,26,28,34–39} Although the Slurp1X^−/− germline knockout mouse model with a point mutation (N35X) in Slurp1 developed severe PPK with epidermal barrier defect mimicking Mal de Meleda,⁴⁰ its corneal morphology and histology remained unaffected.⁴¹ We reported that the Slurp1X^−/− CE cell homeostasis is altered, with elevated expression of cyclins and downregulation of p15/CDKN2B consistent with SLURP1 being a prodifferentiation factor that stalls G1-S transition during cell cycle progression.⁴¹ Taken together, these reports suggest a key role for SLURP1 in the cornea and the skin, where it is abundantly expressed, and establish the availability of a mouse model to study its functions in greater detail. 

As the absence of spontaneous angiogenic inflammation in naive Slurp1X^−/− mouse corneas⁴¹ was counterintuitive in view of our previous finding that SLURP1 suppresses HUVEC tube formation,³¹ we evaluated the effect of proangiogenic insults on Slurp1X^−/− mouse corneas. Based on the outcomes of our studies comparing the wild-type (WT) and Slurp1X^−/− mouse corneal response to silver nitrate (AgNO₃) cautery and the status of NF-κB signaling pathway components in naive WT and Slurp1X^−/− corneas, we report here that Slurp1 suppresses corneal angiogenic inflammation by modulating NF-κB signaling. 

Materials and Methods

Animals and Reagents

Slurp1X ^−/− mouse with a point mutation in exon 2 of the Slurp1 gene, a kind gift from Dr. Stephen Young, UCLA, was re-derived in C57Bl/6J background at the University of Pittsburgh.⁴¹ The animal studies were done in compliance with the ARVO statement for the use of animals in ophthalmic and vision research and followed the guidelines set forth by the University of Pittsburgh Institutional Animal Care and Use Committee (IACUC Protocol #: 21059346). 

AgNO₃ Cautery

Wild-type and Slurp1X^−/− mice were anesthetized by intraperitoneal injection of a mix of ketamine (100 mg/kg body weight) and xylazine (10 mg/kg body weight). AgNO₃ applicators (StyptStix; Henry Schein, Melville, NY, USA) were gently placed on anaesthetized mouse corneas for 5 seconds, immediately followed by rinsing with a gentle stream of PBS following the published procedure.⁴² After 5 days, corneas were dissected and used for quantitative RT-PCR (qRT-PCR), immunofluorescent staining, or flow cytometry. 

Immunoblots

For immunoblots, two corneas dissected from each mouse were pooled together and homogenized in 50 µL radioimmunoprecipitation assay (RIPA) buffer (50 mM Tris-HCl [pH 8.0], 150 mM NaCl, 1% Triton X-100, 1 mM EDTA, 0.5% sodium deoxycholate, 0.1% SDS, with protease inhibitors and/or phosphatase inhibitors added just prior to use) using a handheld homogenizer. Equal quantities of total protein (20 µg), quantified by the bicinchoninic acid (BCA) method with bovine serum albumin (BSA) used for the standard curve, were separated by electrophoresis on 4% to 12% gradient polyacrylamide gels (Nupage Novex mini gels; Invitrogen, Carlsbad, CA, USA), electroblotted to polyvinylidene difluoride (PVDF) membranes, blocked with Intercept blocking buffer (Li-Cor, Lincoln, NE, USA), exposed to primary antibody overnight, probed with secondary antibody, and imaged on the Odyssey imaging system. Densitometric analysis was done using ImageJ (National Institutes of Health [NIH], Bethesda, MD, USA) to quantify the immunoblot signal intensity and normalized to actin band intensity of the same blot. Immunoblot data derived using lysates from three to six different WT and Slurp1X^−/− mouse corneas were used for densitometry. The list of antibodies, their source, and the dilutions used in this study are provided in Supplementary Table S1. 

Immunofluorescent Staining

For corneal whole-mount immunostaining, dissected corneas were fixed in 4% paraformaldehyde, blocked in 10% goat serum for 2 hours, incubated with PE-conjugated anti-CD31 antibody (that stains PECAM-1, a glycoprotein constitutively expressed by endothelial cells) and FITC-conjugated anti-CD45 antibody (that stains a receptor-linked protein tyrosine phosphatase expressed on nucleated hematopoietic cells) overnight at 4°C, washed thrice with phosphate-buffered saline containing 0.1% Tween-20 (PBST) for 30 minutes each time, and mounted in aquamount after making four radial incisions to facilitate flattening. Whole corneal images were captured using a dissection microscope equipped for fluorescence microscopy, and higher-magnification images were captured using an Olympus FluoView 1000 confocal system (Olympus, Center Valley, PA, USA). ImageJ (NIH) was used to measure the area covered by corneal neovascularization. Each experiment was performed with at least eight untreated control and cauterized WT and Slurp1X^−/− corneal wholemounts each for quantifying the mean percent area of the whole cornea covered by vasculature, and representative images are presented here. 

For immunofluorescent staining, 8-µm-thick cryosections from freshly dissected mouse eyeballs embedded in optimal cutting temperature (OCT) compound were fixed in 4% paraformaldehyde for 10 minutes, permeabilized in PBS with 0.1% Triton for 5 minutes, blocked for 1 hour in 10% donkey serum, incubated overnight with primary antibody, stained with secondary antibody and DAPI for 1 hour at room temperature (RT), washed 3 times for 30 minutes each time, and cover-slipped using aquamount. Images were captured using an Olympus FluoView 1000 confocal system (Olympus) or Keyence BZ-X810 microscope (Keyence, Itasca, IL, USA). Images were acquired under similar settings and processed using FluoView software or Keyence BZ-X800 analyzer software. Images of immunofluorescent staining of cryosections shown are representative of at least three corneas from three different animals. 

RNA Extraction, cDNA Synthesis, and Quantitative PCR

Two corneas were pooled together for RNA isolation. Equal amounts of total RNA (800 ng) extracted using the Biobasic RNA extraction kit (Biobasic, Marckam, ON, Canada) were converted to cDNA using mouse Moloney leukemia virus reverse transcriptase (Promega, Madison, WI, USA). The qRT-PCR assays were performed using SYBR green or TaqMan reagent in a StepOne Plus thermocycler (Applied Biosystems, Foster City, CA, USA). Reaction conditions included initial denaturation at 95°C for 10 minutes, followed by 40 cycles of denaturation at 95°C for 15 seconds and annealing and extension at 60°C for 60 seconds. This was followed by evaluation of melt curve when SYBR green reagents were used. Primers for SYBR green reactions were obtained from IDT (Indianapolis, IN, USA) and TaqMan probes from Applied Biosystems. Each experiment was performed with total RNA from a minimum of four mice per treatment. 

Flow Cytometry

Dissected mouse corneal stroma and epithelium were separated by treatment with dispase (1 hour at RT in supplemental hormonal epithelial medium [SHEM] containing 50 mg/mL dispase-II and 100 mM sorbitol). Single-cell suspensions of mouse corneal stromal cells were obtained by incubating dissected stroma in collagenase for 45 minutes and filtering through a 35-µm cell strainer cap. The filtered cells were stained for viability and fluorescent antibodies that detected CD11b (integrin αM, expressed by many leukocytes, including monocytes, neutrophils, natural killer cells, granulocytes, and macrophages), CD11c (integrin αX, a marker of dendritic cells), Ly6G (lymphocyte antigen 6 family member G, a marker for monocytes, granulocytes, and neutrophils) (all from BioLegend, San Diego, CA, USA), F4/80 (a marker of murine macrophages), Ly6C (lymphocyte antigen 6 family member C, a marker for monocytes, macrophages, and endothelial cells) (BD Pharmingen, Franklin Lakes, NJ, USA); washed with PBS with 2% fetal bovine serum; fixed in 2% paraformaldehyde; filtered through a mesh; and analyzed using Beckman Coulter Cytoflex LX (Beckman Coulter, Brea, CA, USA) as earlier.¹⁵ 

Statistical Analysis

The results presented here are representative of at least three independent experiments and shown as mean ± standard error of mean (SEM) unless otherwise stated. Statistical significance was tested by Student's unpaired t-test, with P ≤ 0.05 considered statistically significant. 

Results

Robust Angiogenic Response to Silver Nitrate Cautery in Slurp1X^−/− Corneas

To determine if WT and Slurp1X^−/− corneas differ in their angiogenic response to alkali burn injury, we employed AgNO₃ cautery. After 5 days of recovery from AgNO₃ cautery, we subjected the dissected corneas to whole-mount immunofluorescent staining with anti-CD31 antibody (Fig. 1). Compared with the WT, the Slurp1X^−/− corneas displayed a much denser angiogenic response to AgNO₃ cautery, although there was no significant difference in the area covered by the new vasculature expressed as a percentage of the total corneal surface in the WT and Slurp1X^−/− corneas (Fig. 1). 

Figure 1.

 

Robust angiogenic response to silver nitrate cautery in Slurp1X^−/− corneas. Left eyes of the WT and Slurp1X^−/− mouse corneas were subjected to silver nitrate cautery, and 5 days later, dissected corneas were processed for whole-mount immunofluorescent staining with anti-CD31 (PECAM-1) antibody. (A–D) WT and (E–H) Slurp1X^−/− corneas were subjected to whole-mount immunofluorescent staining with anti-CD31 (red) antibodies 5 days post-AgNO₃ cautery. Areas boxed in panels A, B, E, and F are shown magnified in panels C, D, G, and H, respectively. N = at least 8 corneas; representative images shown. Statistical significance determined using unpaired t-test is shown. *P < 0.05, **P < 0.005, ***P < 0.0005, and ****P < 0.0001. (I) The mean percent area of the overall cornea covered by vasculature quantified from at least eight untreated control and cauterized corneal wholemounts each is shown.

Increased Immune Infiltrate in Cauterized Slurp1X^−/− Corneas Compared With the WT

Whole-mount costaining of WT and Slurp1X^−/− corneas with anti-CD31 and anti-CD45 antibodies 5 days after AgNO₃ cautery revealed the presence of large numbers of CD45⁺ cells at the boundary of neovasculature in Slurp1X^−/− but not the WT corneas (Fig. 2A). Consistent with these results, flow cytometry of cells in the stroma of these corneas showed a greater influx of CD11b⁺ cells, macrophages, and neutrophils into the AgNO₃ cauterized Slurp1X^−/− (Fig. 2B). The number of neutrophils was significantly higher even in the control (naive) Slurp1X^−/− corneas compared with the WT corneas not subjected to cautery, suggesting that the absence of Slurp1 results in a proinflammatory environment in the cornea (Fig. 2B). The gating scheme that we used to identify the specific subsets of infiltrating immune cells using fluorescently tagged antibodies that recognize CD45, TCR-β, CD11b, Ly6G, and Ly6C5 is shown (Fig. 2C). Collectively, these results suggest that Slurp1X^−/− mice have heightened inflammatory responses and respond more aggressively to silver nitrate cautery, as illustrated by a larger influx of immune cells into the cornea compared to WT controls. 

Figure 2.

 

Slurp1X ^−/− corneas display greater influx of immune cells upon AgNO₃ cautery. (A) Whole-mount immunofluorescent staining. WT (i and ii) and Slurp1X^−/− (iii and iv) corneas were subjected to whole-mount immunostaining with anti-CD31 (red) and anti-CD45 (green) antibodies 5 days post-AgNO₃ cautery. High-resolution confocal images from the edge of vascular areas are shown. N = 3; representative images are shown. (B) Flow cytometry of immune infiltrate in control and AgNO₃-cauterized WT and Slurp1X^−/− corneal stroma. Five days after AgNO₃ cautery, mouse corneas were dissected and the stroma separated from the epithelium by dispase treatment. Single-cell suspensions of stromal cells obtained by collagenase treatment of dissected stroma were stained for different immune cell–specific markers and analyzed by flow cytometry. Flow cytometry revealed significantly elevated macrophage and neutrophil influx in AgNO₃-cauterized Slurp1X^−/− corneas compared with the WT. (C) The gating scheme used to identify myeloid cells in the mouse cornea. Single-cell suspensions from dissected corneal stroma were stained with fluorescently tagged antibodies that specifically recognize the cell-surface markers CD45, TCR-β, CD11b, Ly6G, and Ly6C5 and analyzed by flow cytometry following the gating strategy shown. CD45⁺ gate identifies all immune cells. CD11b⁺TCRb^– identifies myeloid-derived cells. F4/80⁺ gating marks macrophages, and Ly6G⁺Ly6C^– cells are identified as neutrophils. N = at least 8 corneas. Statistical significance determined using unpaired t-test is shown. *P < 0.05, **P < 0.005, and ***P < 0.0005.

Slurp1 Expression Is Decreased in Corneas Subjected to Silver Nitrate Cautery

Previously, we demonstrated that Slurp1 expression is decreased in response to a variety of proinflammatory insults, including herpes simplex virus (HSV) infection, bacterial lipopolysaccharides (LPS) injection,¹⁵ or exposure to Pam₃Csk₄, Poly(I:C), and Zymosan-A.¹⁶ In the current experiments, qRT-PCR revealed a significant reduction in Slurp1 transcripts in the WT corneas 2 days after silver nitrate cautery (Fig. 3A). Consistent with these results, immunoblots revealed that the Slurp1 protein is expressed in the untreated or contralateral WT corneas but undetectable in the cauterized corneas 2, 7, and 10 days after exposure to silver nitrate (Figs. 3B, 3C). These results suggest that the immune infiltrate in cauterized WT corneas detected above (Fig. 2) is accompanied by a significant downregulation in Slurp1 expression (Fig. 3). 

Figure 3.

 

Slurp1 is downregulated upon silver nitrate cautery. (A) qRT-PCR quantification of Slurp1 transcripts in the WT mouse corneas 5 days after silver nitrate cautery. N = 3 (qRT-PCR was performed with three independent total RNA samples, each isolated from two pooled corneas from different animals). (B) Detection of Slurp1 protein by immunoblot in WT corneas at 2, 7, and 10 days postcautery. N = 3 (immunoblot was repeated thrice; representative blot shown). (C) Densitometry of the immunoblots. CL, contralateral eye; DPT, days posttreatment; T, treated (cauterized); UT, untreated. Statistical significance determined using unpaired t-test is shown. *P < 0.05, **P < 0.005, and ***P < 0.0005.

Proangiogenic Inflammatory Environment in Cauterized Slurp1X^−/− Corneas

Next, we evaluated if the expression of select proangiogenic inflammatory molecules is altered in the Slurp1X^−/− corneas 5 days after cautery. We found no significant difference in the expression of the proinflammatory angiogenic molecules except IL-1β between the WT and Slurp1X^−/− corneas at homeostasis. Upon AgNO₃ cautery, the transcripts encoding VEGFA, MMP2, IL-1β, and vimentin were upregulated in both the WT and Slurp1X^−/− corneas. However, the extent of upregulation was markedly higher in the Slurp1X^−/− corneas compared with the WT (Fig. 4). Thus, Slurp1X^−/− corneas respond to AgNO₃ cautery by producing relatively more VEGFA, MMP2, IL-1β, and vimentin than the WT. 

Figure 4.

 

Differential expression of select transcripts in the WT and Slurp1X^−/− corneas subjected to silver nitrate cautery. Expression of transcripts encoding vimentin, IL-1B, MMP2, and VEGFA was quantified by qRT-PCR using total RNA isolated from dissected WT and Slurp1X^−/− untreated control or cauterized corneas. N = 3. Error bars represent standard error of mean. Statistical significance determined using unpaired t-test is shown where differences are significant. In panel B, the fold changes in expression of IL-1B appear exaggerated due to its negligible expression in the untreated control.

NF-κB Is Upregulated and Hyperphosphorylated in Unstimulated Slurp1X^−/− Corneas

Considering that (1) the above data are consistent with the existence of a proinflammatory environment within Slurp1X^−/− corneas, and (2) NF-κB signaling plays a key role in inflammation in many diverse scenarios, next we evaluated the status of NF-κB signaling in Slurp1X^−/− corneas. Immunofluorescent staining revealed relatively higher levels and nuclear localization of phospho-NF-κB in Slurp1X^−/− corneas compared with the lower expression and mostly cytoplasmic localization in the WT (Fig. 5A). There was a similar increase and altered localization in total NF-κB in Slurp1X^−/− corneas compared with the WT (Fig. 5A). Immunoblots revealed that the total NF-κB levels are moderately elevated in the Slurp1X^−/− corneas, consistent with the NF-κB expression as well as activity being elevated in the Slurp1X^−/− corneas (Fig. 5B). 

Figure 5.

 

NF-κB is upregulated and hyperphosphorylated in the Slurp1X^−/− corneas. (A) Cryosections from the WT (panels ii and v) and Slurp1X^−/− (iii and vi) corneas were subjected to immunofluorescent staining with anti–phospho-NF-κB (red, panels ii, iii, vii, and viii) and anti–total NF-κB (red, panels v, vi, ix, and x) antibodies. Nuclei are counterstained blue by DAPI in panels i to vi. No primary antibody controls (panels i and iv) show no nonspecific binding by the secondary antibody. Panels vii through x show digitally magnified images of the boxed-in regions from panels ii, iii, v, and vi in the red channel, omitting DAPI for clarity. Both phospho–NF-κB and total NF-κB levels are relatively lower in the WT, where they are retained in the cytoplasm (panels ii, v, vii, and ix; arrowheads) compared with the Slurp1X^−/− CE, where they also appeared in the nucleus (panels iii, vi, viii, and x; arrows). N = 3; representative images are shown. (B) Quantification of total NF-κB expression in the WT and Slurp1X^−/− corneas by immunoblot and densitometry. After imaging, the blots were reprobed with antiactin antibody as loading control. N = 3. Error bars represent standard error of the mean.

As NF-κB activity is regulated by IκB that binds and retains NF-κB in the cytoplasm in an inactive form,⁴³ next we tested the expression and phosphorylation status of IκB. Immunofluorescent staining revealed elevated levels of phospho-IκB in Slurp1X^−/− corneas, with the total IκB remaining comparable or slightly lower than that in the WT (Fig. 6A). This was further confirmed by immunoblots, which revealed that the total IκB content is about 15% lower in the Slurp1X^−/− compared with that in the WT corneas (Fig. 6B). Furthermore, immunofluorescent staining revealed elevated levels of phospho-IκB kinase (phospho-IKK) subunits a and b expression in Slurp1X^−/− corneas compared with the WT, with no discernible difference in the level of total IKK expression (Fig. 7). Taken together, increased phosphorylation of IKK(a/b) and IκB, decreased amounts of total IκB, and higher phosphorylation of NF-κB are consistent with the NF-κB signaling pathway being constitutively active in the Slurp1X^−/− CE cells. 

Figure 6.

 

IκB is phosphorylated and marked for degradation in the Slurp1X^−/− corneas. (A) Cryosections from the WT (panels ii, v, vii, and ix) and Slurp1X^−/− (panels iii, vi, viii, and x) corneas were subjected to immunofluorescent staining with anti–phospho-IκB and total IκB antibodies (red). No primary antibody controls (panels i and iv) show no nonspecific binding by the secondary antibody. Panels vii through x show images from panels ii, iii, v, and vi in the red channel, omitting DAPI for clarity. Phospho-IκB levels are relatively lower in the WT (panels ii and vii; arrowheads) compared with the Slurp1X^−/− CE (panels iii and viii; arrows). In contrast, intensity of staining for total IκB is comparable between the WT and Slurp1X^−/− CE (panels v and vi, or ix and x). Nuclei are counterstained blue by DAPI. N = 3; representative images are shown. (B) Quantification of total IκB expression in the WT and Slurp1X^−/− corneas by immunoblot and densitometry. The blots were reprobed with antiactin antibody as loading control. N = 3. Error bars represent standard error of the mean.

Figure 7.

 

IKK is hyperphosphorylated in the Slurp1X^−/− corneas. Cryosections from the WT (panels ii and v) and Slurp1X^−/− (panels iii and vi) corneas were subjected to immunofluorescent staining with anti–phospho-IKK or anti–total IKK antibodies (red). No primary antibody controls (panels i and iv) show no nonspecific binding by the secondary antibody. Nuclei are counterstained blue by DAPI in panels i through vi. Panels vii through x show images from panels ii, iii, v, and vi in the red channel, omitting DAPI for clarity. Phospho-IκK levels are relatively lower in the WT (panels ii and vii; arrowheads) compared with the Slurp1X^−/− CE (panels iii and viii; arrows). In contrast, intensity of staining for total IκK is comparable between the WT and Slurp1X^−/− CE (compare panels v and vi, or ix and x). N = 3; representative images are shown.

Discussion

Previously, we demonstrated that SLURP1, abundantly expressed by the CE and secreted into the tear film, serves as an antiangiogenic and anti-inflammatory factor.^{15,16,30,32,33,41} Although Slurp1 is highly expressed in the mouse cornea and our previous studies ascribed an important immunomodulatory role for Slurp1, Slurp1X^−/− corneas displayed only a subtle phenotype with no detectable change in corneal protein profile.⁴¹ Here, we report that (1) the Slurp1X^−/− corneas display robust angiogenic response to AgNO₃ cautery, (2) the Slurp1X^−/− corneas subjected to AgNO₃ cautery support elevated immune infiltrate, (3) Slurp1 expression is decreased in WT mouse corneas subjected to AgNO₃ cautery, (4) cauterized Slurp1X^−/− corneas harbor a more proangiogenic inflammatory environment, and (5) NF-κB signaling is constitutively active in Slurp1X^−/− corneas. Collectively, these results provide evidence that Slurp1 promotes healthy corneal avascularity by suppressing NF-κB signaling and that upon exposure to proinflammatory insults, its expression is downregulated, allowing angiogenic inflammation to proceed. 

An important observation in this study is that Slurp1 modulates NF-κB activity in healthy mouse corneas, which complements our previous finding that SLURP1 impedes with TNF-α–stimulated NF-κB translocation into the nucleus in human corneal limbal epithelial cells.³³ NF-κB is a heterodimeric transcription factor that plays a key role in inflammation in many parts of the body, including the cornea. In quiescent unstimulated cells, IκB binds NF-κB dimers and retains them in the cytoplasm. Upon stimulation by proinflammatory cytokines, the IκB kinase (IKK) phosphorylates IκB, releasing NF-κB. Free NF-κB dimers then translocate to the nucleus, where they activate proinflammatory gene transcription. Phosphorylated IκB is ubiquitinated by E3 ubiquitin ligase and cleared by proteasomal degradation (Fig. 8). The data presented in this report elucidate that these key steps in the NF-κB signaling pathway are constitutively active in naive Slurp1X^−/− CE, suggesting that Slurp1 facilitates retaining the NF-κB signaling pathway in an inactive state in healthy cells (Fig. 8). This active state of NF-κB keeps the signaling milieu in the Slurp1X^−/− cornea primed for hyperinflammatory response to proinflammatory insults. 

Figure 8.

 

Schematic representation of the findings reported here. The data presented in this report elucidate that upon silver nitrate cautery, the Slurp1X^−/− corneas display elevated angiogenic inflammation. Furthermore, the Slurp1X^−/− corneas display elevated IKK phosphorylation, IκB phosphorylation, and NF-κB activation and nuclear localization, consistent with the constitutively hyperactive state of NF-κB signaling in the absence of Slurp1. Taken together with our previous findings that Slurp1 serves an immunomodulatory role in the mouse cornea, these data suggest that Slurp1 modulates corneal angiogenic inflammation via NF-κB signaling. Whether Slurp1 affects each of these steps in the NF-κB signaling pathway independently or only affects an early step in the cascade of events remains to be determined. C, cytoplasm; N, nucleus; P, phosphorylated protein; Ub, ubiquitin.

Previously, we demonstrated that SLURP1 suppresses (1) TNF-α–induced HUVEC tube formation by inhibiting nuclear translocation of NF-κB,³¹ (2) neutrophil docking on HUVEC cells by decreasing E-selectin production,³² and (3) neutrophil transmigration through the HUVEC monolayer by stabilizing membrane localization of VE-cadherin and β-catenin,³² suggesting that SLURP1 promotes corneal angiogenic privilege. The immunophenotyping data presented in this report, that AgNO₃-cauterized Slurp1X^−/− corneas support significantly higher influx of CD11b⁺ cells, macrophages, and neutrophils, present clear in vivo evidence in support of a role for Slurp1 in protecting the transparent cornea from undesirable angiogenic inflammation. 

AgNO₃ cautery in the cornea causes oxidative stress that activates NF-κB, which in turn induces the expression of inflammatory cytokines that induce CNV and recruit immune infiltrate, causing further tissue damage. Consistent with the previous reports that corneal VEGF expression increased manyfold 7 days after alkali burn,^44–47 the data presented here reveal that the expression of a select few cytokines is elevated at 5 days after cautery and that this increase is significantly higher in the Slurp1X^−/− corneas. Although this finding provides evidence of an active inflammatory response to alkali burn, the complete timeline of changes in their expression following cautery needs further study. 

In summary, considering that (1) Slurp1 expression is decreased in mouse corneas subjected to AgNO₃ cautery, (2) Slurp1X^−/− corneas display elevated expression of proinflammatory cytokines coupled with robust CNV and immune influx in response to AgNO₃ cautery, and (3) NF-κB signaling is constitutively active in Slurp1X^−/− corneas, we conclude that Slurp1 promotes healthy corneal avascularity by suppressing NF-κB signaling and that upon exposure to proinflammatory insults, its expression is downregulated, allowing angiogenic inflammation to proceed (Fig. 8). Whether Slurp1 affects the canonical NF-κB pathway (that depends on IKKβ and IKKγ) or the noncanonical one (that depends solely on IKKα), and if Slurp1 affects each step in the NF-κB signaling pathway independently or only affects an early step in the cascade of events, remains to be determined. 

Acknowledgments

The authors thank Stephen G. Young and Loren G. Fong, UCLA, for sharing the Slurp1X^−/− mice. 

Supported by R01 EY 022898 (SKS), RO1 EY 031684 (SKS), unrestricted grants from Research to Prevent Blindness and the Eye and Ear Foundation of Pittsburgh to the Department of Ophthalmology, University of Pittsburgh School of Medicine, and the NEI Core Grant P30 EY08098. The funding agencies did not have any influence on design or conduct of the experiments described here or preparation of this manuscript. 

Disclosure: S. Swamynathan, (P); G. Campbell, None; P. Sohnen, None; S. Kaur, None; A.J. St. Leger, None; S.K. Swamynathan, (P)