Interleukin-1 as a phenotypic immunomodulator in keratinizing squamous metaplasia of the ocular surface in Sjögren's syndrome

Abstract

Chronic inflammation of the ocular surface in Sjögren's syndrome (SS) is associated with a vision-threatening, phenotypic change of the ocular surface, which converts from a nonkeratinized, stratified squamous epithelium to a nonsecretory, keratinized epithelium. This pathological process is known as squamous metaplasia. Based on a significant correlation between ocular surface interleukin (IL)-1beta expression and squamous metaplasia in patients with SS, we investigated the role of IL-1 in the pathogenesis of squamous metaplasia in an animal model that mimics the clinical characteristics of SS. Using autoimmune-regulator (aire)-deficient mice, we assessed lacrimal gland and ocular surface immunopathology by quantifying the infiltration of major histocompatibility complex class II(+) (I-A(d+)) dendritic cells and CD4(+) T cells. We examined squamous metaplasia using a biomarker of keratinization, small proline-rich protein 1B. We used lissamine green staining as a readout for ocular surface epitheliopathy and Alcian blue/periodic acid-Schiff histochemical analysis to characterize goblet cell muco-glycoconjugates. Within 8 weeks, the eyes of aire-deficient mice were pathologically keratinized with significant epithelial damage and altered mucin glycosylation. Although knockdown of IL-1 receptor 1 did not attenuate lymphocytic infiltration of the lacrimal gland or eye, it significantly reduced ocular surface keratinization, epitheliopathy, and muco-glycoconjugate acidification. These data demonstrate a phenotypic modulation role for IL-1 in the pathogenesis of squamous metaplasia and suggest that IL-1 receptor 1-targeted therapies may be beneficial for treating ocular surface disease associated with SS.

Figure 1

The squamous metaplasia biomarker, SPRR1B, as a predictor of IL-1β and the OSS in human patients with Sjögren’s syndrome. Superficial ocular epithelial cells from 15 patients with Sjögren’s syndrome were harvested by impression cytology for RNA extraction. Ocular surface epitheliopathy was quantified using the SICCA OSS. To examine the consistency of real-time PCR data, both TaqMan- (n = 6, blue dots) and SYBR Green dye-based chemistry (n = 9, black dots) were used to examine mRNA expression of SPRR1B and IL-1β. Data were normalized by GAPDH and ΔΔC_t values were analyzed by linear regression to test the null hypothesis that SPRR1B expression was independent of IL1-β expression (A) and OSS (B). Each graph shows raw ΔC_t values as well as a regression line. Squared correlation coefficients (R²) indicate the fraction of variance explained by the regression model. Correlations were considered statistically significant when computed P values were <0.01.

Figure 2

Autoimmune-mediated inflammation in aire-deficient mice mimics the lacrimal gland exocrinopathy and ocular surface keratinization of Sjögren’s syndrome. A: Gross anatomy of lacrimal glands from aire-deficient and aire-sufficient mice, stained with anti-CD4⁺ T cell antibody. B: H&E staining of the lacrimal gland. Inset is the higher power view of the boxed area. C: External eye pictures of the ocular surface. D: Immunofluorescence staining of SPRR1B in the central cornea. Dotted line indicates the corneal endothelium. Scale bar = 50 μm.

Figure 3

The IL-1 cytokine family in aire-mediated KCS is proinflammatory. A: Genomic DNA was isolated from tail clippings of heterozygous (+/−) and knockout (−/−) mice for each gene of interest (Aire and IL-1R1). Duplex PCR was used to detect the full-length target gene in wild-type mice or a shorter mutant form in knockout mice. As expected, the aire wild-type band was detected at 1000 bp, the aire knockout band at 600 bp, the IL-1R1 wild-type band at 350 bp, and the IL-1R1 knockout band at 172 bp. B: Transcriptional activity of proinflammatory IL-1β and anti-inflammatory IL-1RA among four genotypic groups. Data were obtained from three individual mice in each group and are shown as relative quantitation (using wild-type mice aire^+/−IL-1R1^+/− as onefold in gene expression). *P < 0.05, using analysis of variance. C: Confocal immunolocalization of IL-1β-expressing cells in the cornea. Whereas wild-type and IL-1R1 KO mice showed minimal expression of IL-1β in the corneal epithelium, both aire KO and aire/IL-1R1 DKO mice showed abundant expression of IL-1β across the corneal epithelium. Scale bar = 70 μm.

Figure 4

Dendritic cell activation and autoreactive CD4⁺ T-cell recruitment in aire-deficient mice with and without intact IL-1R1signaling. A: Immunohistochemical analysis against MHC class II surface antigen I-A^d (DAB chromogen shown as brown) was used to identify activated antigen-presenting cells in the central corneas of wild-type (aire^+/−IL-1R1^+/−), IL-1R1 KO (aire^+/−IL-1R1^−/−), aire KO (aire^−/−IL-1R1^+/−), and aire/IL-1R1 DKO (aire^−/−IL-1R1^+/−). Aire-sufficient wild-type and aire-sufficient IL-1R1 KO mice served as controls. Arrowheads indicate subepithelial infiltrating APCs, whereas arrows indicate stromal infiltrating APCs. B: Central corneal I-A^d+ dendritic cells (brown) were quantified by DAB staining intensity among the four groups. Data are shown as means ± SD in arbitrary units. C and D: Immunohistochemical analysis against CD4⁺ T cells (brown) demonstrated dense infiltration of the lacrimal gland parenchyma and the limbal area of the ocular surface in both aire KO and aire/IL-1R1 DKO mice, with scarce cells noted in aire-sufficient controls. Arrows indicate intraepithelial infiltrating CD4+ T cells, while arrowheads indicate intrastromal infiltrating CD4+ T cells. E: DAB staining intensity of CD4⁺ effector T cells was analyzed by densitometry. Data are shown as means ± SD in arbitrary staining units. Analysis of variance was used to test for significant differences between groups with P < 0.05 considered statistically significant. F: Immunofluorescence to detect Foxp3⁺ T regulatory cells in the limbal tissues. Sporadic expression of Foxp3⁺ cells (arrows) in the spleen served as a positive control. Scale bar = 50 μm.

Figure 5

Ocular surface epitheliopathy in aire-deficient mice with and without IL-1R1 signaling. A: Lissamine green staining of the ocular surface was used to reveal punctate epithelial cell damage (green). Compared with aire-deficient mice with IL-1R1 (aire KO), staining intensity was significantly reduced when IL-1R1 was simultaneously knocked-out (aire/IL-1R1 DKO). B: Lissamine green staining intensity was quantified by ImageJ with data reported as means ± SD. Analysis of variance was used to test for significant differences between groups with P < 0.05 considered statistically significant. Scale bar = 100 μm.

Figure 6

Ocular surface keratinization in aire-deficient mice with and without IL-1R1 signaling. A: SPRR1B staining of the ocular surface revealed keratinization in aire KO mice. SPRR1B expression was highest in the superficial corneal epithelial cells with a decreasing gradient toward the basal layer in aire KO mice. SPRR1B staining in DKO mice was largely reduced and undetectable in wild-type and IL-1R1 KO mice. Inset is the high power view of the boxed area. The grey mode of SPRR1B immuno-signals was used to provide a better contrast. B: SPRR1B fluorescence intensity was quantified by ImageJ with data reported as means ± SD. Analysis of variance was used to test for significant differences between groups with P < 0.05 considered statistically significant. Scale bar = 100 μm.

Figure 7

Quantification and characterization of conjunctival goblet cells in aire-deficient mice with and without intact IL-1R1 signaling. A: PAS staining was used to identify the GC-rich zone in the mouse tarsal conjunctiva. B: The total number of PAS-positive (pink) GCs in a defined area of the GC-rich zone was determined, and data are expressed as mean GC number ± SD. C: AB-PAS staining was used to identify the presence of GCs containing acidic mucin glycoconjugates (blue, arrows) or neutral glycoconjugates (pink). Aire KO mice exhibited dramatic acidification of GC mucins that was significantly reduced in aire/IL-1R1 DKO mice and nearly absent in wild-type and IL-1R1 KO controls. D: Percentage of AB⁺ GCs to overall GC number (sum of AB⁺PAS⁺ and AB⁻PAS⁺ cells) was determined for each group. Data are reported as mean percentage ± SD. Analysis of variance was used to test for differences between groups with P < 0.05 considered statistically significant. Scale bar = 100 μm.

Figure 8

Epithelial damage, corneal keratinization, and acidification of GC mucins are highly correlated in Sjögren’s syndrome-like autoimmune ocular surface disorder. A–C: Regression analysis was performed to examine the relationships between three phenotypic characteristics of the Sjögren’s syndrome ocular pathology: lissamine green staining (epithelial damage), corneal expression of SPRR1B, and the percentage of AB⁺ conjunctival GCs. Squared correlation coefficients (R²) were ∼0.50 for each analysis, suggesting that each phenotypic characteristic predicted approximately 50% of the variability in the other two outcomes. Raw data are plotted with the best-fit regression line to demonstrate statistically significant correlations (P < 0.05) between SPRR1B and lissamine green staining (A), SPRR1B and percentage of AB⁺ GCs (B), and lissamine green staining and percentage of AB⁺ GCs (C). D: A hypothetical diagram presents a central role for IL-1 in the pathogenesis of KCS and squamous metaplasia of Sjögren’s syndrome. In the progression of nonkeratinized KCS to keratinizing squamous metaplasia, IL-1 may initiate inflammatory epitheliopathy in the acute stages of disease and later act as an immunomodulator of epithelial phenotype in the setting of chronic inflammation.