Ocular Surface Involvements in the Development of Sjögren's Syndrome–Associated Dry Eye in the IL14α Transgenic Mouse

Abstract

Purpose

To investigate the ocular surface changes during progress of the Sjögren's Syndrome (SS), using a previously described IL14α transgenic mice (IL14α TG) SS model.

Methods

The ocular surface of IL14α TG and C57BL/6 wild-type (WT) female mice were evaluated at the age of six, nine, 12, 15, and 18 months. Slit lamp microscopy observation, Oregon green dextran staining, Schirmer test, and periodic-acid–Schiff staining were assessed. Immunohistochemistry, immunofluorescence, and associated gene expression analysis by qPCR and ELISA were performed in cornea, conjunctiva, and lacrimal grand at different ages of the mice. Masson's trichome staining was conducted on lacrimal gland cryosections.

Results

Compared with C57BL/6 WT mice, IL14α TG mice showed corneal barrier function damage and losses in conjunctival goblet cell density starting at nine months, whereas decreases in tear secretion started at 18 months of age. Significant increases in CD4⁺ T cell infiltration in the conjunctiva of IL14α TG mice was first observed at 6 months. Higher expression levels of inflammatory cytokines IL-17A, IFN-γ, IL-1β, and TNF-α in the conjunctiva, whereas MUC5AC and MUC5B had lower expression levels at nine months in the IL14α TG mice. However, lacrimal gland function–associated gene expression levels mostly decreased in IL14α TG mice at 12 months of age.

Conclusions

Ocular surface tissue changes were involved in SS-like dry eye in a time-dependent manner in IL14α TG mice, and conjunctival T-cell infiltration may contribute to ocular surface pathological changes in an early stage of SS-related dry eye.

Sjögren's syndrome (SS) is a systemic autoimmune disease characterized by lymphocyte infiltration in the exocrine glands such as lacrimal gland and salivary gland, which results in severe dry eye (DE) and dry mouth in patients. It is one of the most common autoimmune diseases in adults, affecting 0.5% of the population.¹ In clinical practice, SS is greatly under recognized mostly because of its diverse symptomatic expressions, making the initial diagnosis difficult. Diagnosis is often delayed by an average of 6.5 years from symptom onset.² Early detection of SS is important because patients who start biological treatment within the first five years of disease onset are more likely to respond to treatment than those in whom therapy initiation is delayed.³

Previous studies have shown that up to 10% of dry eye patients have SS, and they could have dry eye symptoms for approximately one decade before SS is diagnosed.^4,5 However, ocular symptoms and signs in isolation have been poorly predictive of extraocular objective signs required for the diagnosis of SS patients. Therefore understanding the correlation between ocular symptoms and signs with those of SS progression may shed new light on how to diagnose SS at an earlier age. The underlying pathophysiological mechanism of DE development includes corneal barrier disruption, reduced conjunctival goblet cell (GC) density, aqueous tear deficiency, and squamous metaplasia of the ocular surface epithelium.⁶ SS-induced dysfunction of the lacrimal gland has been thoroughly studied,^7–10 although there is no thorough investigation that describes the association between changes in ocular surface integrity and SS progression.

Interleukin 14α (IL-14α, also known as alpha-taxilin) was identified as a high-molecular-weight B-cell growth factor.¹¹ It is produced by follicular dendritic cells, germinal center T cells, and some B-cell tumors.¹¹ IL-14α has also been shown to play an essential role in the pathophysiology of primary SS.¹² IL-14α also serves as a marker of immune activation in transplant rejection.¹³ Recently, a study showed that upregulation of IL14 enhances memory B-cell function by pushing normal low-affinity auto-reactivity into high-affinity memory responses.¹⁴ Constitutive expression of IL-14 in mice resulted in an enhanced memory B-cell response to systemic lupus erythematosus autoantigens and vaccinations.¹⁵ Elevation of serum IL-14 expression levels can also serve as a cytokine biomarker for the stratification of SS patient dry eye versus non-SS dry eye patient.²

A previous study using IL14α transgenic mice (IL14α TG) model has demonstrated that lymphotoxinα upregulation is involved during the initial stages of SS and IFN-α in a later stage of the disease. Experimental animals exhibit manifestations involving the lacrimal and salivary glands and others including systemic changes that develop in lung disease and lymphoma.¹⁶ In the current study, we describe more extensively the systemic ocular surface manifestations and the underlying SS progression mechanism. We found that corneal, conjunctiva, and lacrimal gland changes are all involved in a time-dependent manner in the development of SS-like dry eye in IL14α TG mice.

Material and Methods

Animal

C57BL/6 (B6) mice obtained from the Shanghai SLAC laboratory animal center (Shanghai, China) were used as WT control mice. IL14α TG mice, whose genetic background is C57BL/6 (B6), were obtained from Dr. Julian L. Ambrus, Jr.’s, laboratory in the State University of New York at Buffalo (Buffalo, NY, USA). All animals were housed and bred at the Experimental Animal Center of Xiamen University in accordance with institutional guidelines. Female mice were used in the experiments. All animal procedures were approved by the Experimental Animal Ethics Committee of Xiamen University (approval No. XMULAC20170014).

Measurement of Tear Production

Tear production (10 eyes of five mice in each group) was measured with phenol red cotton threads (Zone-Quick; Yokota, Tokyo, Japan) as previously reported.¹⁷ In brief, the thread was placed on the lower palpebral conjunctiva at approximately one-third of the lower eyelid distance from the lateral canthus for 15 seconds. The length of the wet red thread was measured in millimeters under the microscope.

Corneal Permeability

Corneal epithelial permeability was assessed through measurements of Oregon-green-dextran (OGD; D7173; Invitrogen, Eugene, OR, USA) penetrance in each group (10 eyes of 5 mice in each group) as reported previously.¹⁷ Briefly, 0.5 µL OGD (50 mg/mL) was instilled onto the ocular surface for one minute after anesthesia through pentobarbital injection. Corneas were then washed five times with 1 mL saline solution. Images were photographed with a microscope (AZ100; Nikon, Tokyo, Japan) under fluorescence excitation at 470 nm. Fluorescent intensity was analyzed by software (NIS Elements; version 4.1; Nikon, Melville, NY, USA).

Histology and Immunohistochemistry Staining

Eyelids, Meibomian glands, conjunctiva and eyeball tissues were harvested from IL14α TG mice and WT control mice and embedded in paraffin. Hematoxylin & eosin (H&E) staining and immunohistochemistry (IHC) staining for F4/80 (1:100, 565402; BD, New Jersey, USA), B220 (1:200, ab64100; Abcam, Cambridge, UK), CD8 (1:200, ab217344; Abcam), and CD4 (1:200, ab183685; Abcam) were performed as previously described.¹⁷ The images were observed under a light microscope (Eclipse 50i; Nikon, Tokyo, Japan) and captured for comparison and analysis. Three sections from each sample were applied for each staining.

Masson's trichome staining was used to detect fibrosis in the lacrimal gland. Sections cut from paraffin-embedded lacrimal glands were stained with Masson's trichome staining kit (D026; Nanjing Jiancheng, Nanjing, China) following the manufacturer's protocol. The collagen fibers were stained blue, and the background was stained red. Digital images of representative areas of the lacrimal gland were captured with a microscope (Eclipse 50i).

Immunofluorescent Staining

Immunofluorescent staining was performed in cryosections of the eyes, conjunctival tissues and lacrimal glands. Sections were fixed in acetone for 10 minutes at −20°C. After incubating with 0.2% Triton X-100 for 10 minutes, sections were blocked by 2% bovine serum albumin for 60 minutes. After that, sections were incubated at 4°C overnight with polyclonal anti-K10 (1:50, ab24638; Abcam), anti-Sprr1B (1:50, LS-C146187; Lifespan Bioscience, Seattle, WA, USA), anti-CD8 (1:200, ab217344; Abcam), anti-B220 (1:200, ab64100; Abcam), and anti-α-smooth muscle actin (αSMA) (1:50, ab5694; Abcam) antibodies. The next day, samples were incubated with Alexa Fluor488-conjugated donkey anti-goat (1:300; A11055; Invitrogen, Eugene, OR, USA) or anti-rabbit IgG (1:300; A21206; Invitrogen) for one hour in the dark at room temperature, followed by three washes in PBS. Sections were then counterstained with DAPI and mounted. Digital images of representative areas were captured with a laser confocal scanning microscope (Fluoview 1000; Olympus, Tokyo, Japan).

Measurement of Goblet Cell

Sections of paraffin-embedded tissues including eyelid, conjunctiva and eyeball were stained with periodic-acid–Schiff (PAS) reagent (395B-1KT; Sigma-Aldrich, St. Louis, USA) as instructed by the manufacturer. At least three sections from each sample were applied for staining. Digital images of representative areas of the conjunctiva were captured with a microscope (Eclipse 50i; Nikon, Tokyo, Japan).

ELISA Assay

To investigate the conjunctival and corneal tissue inflammatory response, ELISA assay was used to examine the expression of inflammatory cytokines. Total proteins in the conjunctiva and cornea were extracted with cold RIPA buffer (R0278; Sigma-Aldrich Corp.). The same concentration was used in all samples. Each group contained six samples, and each sample was measured in triplicate. TNF-α, IL-1β, IFN-γ, and IL-17A concentrations were measured by using recombinant mouse TNF-α (BMS607-3; eBioscience, San Diego, CA, USA), IL-1β (BMS6002; eBioscience), IL-17A (BMS6001; eBioscience), and IFN-γ (BMS606; eBioscience) ELISA kits in accordance with manufacturer's instructions. The minimum level of detection for cytokines measured by ELISA assay is 3.7 pg/mL for TNF-α, 1.2 pg/mL for IL-1β, 1.6 pg/mL for IL-17A, and 1.7 pg/mL for IFN-γ, respectively.

Total RNA Extraction, Reverse Transcription, and Quantitative PCR

Total RNA from the corneal epithelium, conjunctiva and lacrimal gland tissues was extracted using PicoPure RNA isolation kit (KIT0204; Arcturus Therapeutics, San Diego, CA, USA) according to the manufacturer's instructions. CDNA was synthesized with a reverse transcription kit (RR047A; Takara, Shiga, Japan). Real-time PCR was performed on a Step One Real-Time PCR System (Applied Biosystems, Foster City, CA, USA) through a TB Green Premix Ex Taq Kit (RR420A; Takara). The primer sequences are summarized in the Table. The amplification program consisted of pre-denaturation step at 95°C for 60 seconds, followed by 40 cycles of denaturation at 95°C for 10 seconds and annealing and extension at 60°C for 30 seconds. After that, a melt curve analysis was conducted to confirm amplification specificity. The results were analyzed using the comparative threshold cycle (△△CT) method, which was normalized to β-actin expression.

Table.

Primer Sequence Pairs Used for Quantitative Real-Time PCR

Gene	Sense Primer	Antisense Primer
β-actin	5′-CCTAAGGCCAACCGTGAAAAG-3′	5′-AGGCATACAGGGACAGCACAG-3′
IL-1β	5′-GGGCCTCAAAGGAAAGAATC-3′	5′-TACCAGTTGGGGAACTCTGC-3′
TNF-α	5′-GTGATCGGTCCCCAAAGGG-3′	5′-GCTACAGGCTTGTCACTCG-3′
IL-17A	5′-TACCTCAACCGTTCCACGTC-3′	5′-TTTCCCTCCGCATTGACACA-3′
IFN-γ	5′-GACAATCAGGCCATCAGCAAC-3′	5′-CTCATTGAATGCTTGGCGCT-3′
K10	5′-TCGAGGACCTTAAGGGGCAG-3′	5′-GTCAGCTCATCCAGTACCCTG-3′
Sprr1b	5′-GCGACCACACTACCTGTCCT-3′	5′-CTGGCAAGGCTGTTTCACTT-3′
Muc5ac	5′-ACACATGTTCTGGAGGGCA-3′	5′-ACACTTTCGCAGCTCAACCA-3′
Muc5B	5′-CTCTGTACTGCCCCCAGGATG-3′	5′-TGACTGTCTCCGGTGAGTTCTA-3′
Muc15	5′-TCCCAAATACATCAGACACCCC-3′	5′-TGGGTTGTAGTAACTGGTTCGTT-3′
α-SMA	5′-GTCCCAGACATCAGGGAGTAA-3′	5′-TCGGATACTTCAGCGTCAGGA-3′
VEGF	5′-CACTGGACCCTGGCTTTACT-3′	5′-GCAGTAGCTTCGCTGGTAGA-3′
EGF	5′-TTTTGACAAGTGGCAGGAGGTC-3′	5′-AGTGATAGGATCCAGGGGTGT-3′
HGF	5′-AGTCAGCACCATCAAGGCAA-3′	5′-ACCAGGACGATTTGGGATGG-3′
AQP4	5′-CACTGCCCAACGTTAGCTC-3′	5′-TGAGCCACCCCAGTTTATGG-3′
AQP5	5′-ATTGGCTTGTCGGTCACACT-3′	5′-ACGATCGGTCCTACCCAGAA-3′

Statistical Analysis

Statistical analysis was performed using GraphPad Prism software version 8.0 (GraphPad Software, La Jolla, CA, USA). Quantitative data are presented as means ± SD. Typically, two-way ANOVA was applied to evaluate significance between groups. A value of P < 0.05 was considered statistically significant.

Results

Spontaneous Dry Eye Development in IL14α TG Mice

Previous studies have shown that lymphocyte infiltration in the lacrimal gland of IL14α TG mice started from the age of six months, and gradually increased to the age of 18 months.^12,18 It was also found that IL14α TG mice displayed severe corneal damage after the age of 15 months. In general, characteristics of dry eye include corneal barrier disruption, losses in conjunctival GC density and tear volume reduction. In the current study, multiple examinations were performed to evaluate the extent of dry eye in IL14α TG mice. Slit lamp microscopy showed that corneal haziness in IL14α TG mice emerged at the age of 12 months and gradually increased up to the age of 18 months (Fig. 1A). OGD staining showed that corneal epithelial barrier disruption started at 9 months of age and gradually increased with time in IL14α TG mice (Fig. 1B). Fluorescent density evaluation confirmed this pattern change (Fig. 1C). Phenol red thread test showed that tear secretion in IL14α TG mice did not significantly change until age 15 months, although it had significantly declined at 18 months (Fig. 2A). PAS staining of the conjunctival tissues revealed a significant decrease of goblet cell number that started at nine months and progressively decreased up to 18 months (Fig. 2B), as image changes demonstrated in WT mice and IL14α TG mice (Fig. 2C). Taken together, clinical manifestations of dry eye in IL14α TG mice started at age nine months and became more apparent with time. Fluorescent staining and PAS staining pattern changes showed that initial lesions may occur in the cornea and conjunctiva around nine months, whereas the lacrimal gland function did not decline until 18 months based on tear secretion test.

Figure 1.

Dry eye appearance in IL14α TG mice. (A) WT and IL14 mice are shown from age six to 18 months. Representative slit-lamp microscopic images show that IL14 mice begin to develop corneal haziness at the age of 12 months that gradually increase up to the age of 18 months. (B) Representative OGD staining images show that corneal epithelial barrier disruption starts in the central cornea at the age of 12 months and increases with age in IL14α TG mice. (C) Mean intensity of OGD staining is significantly different between the two groups from nine months to 18 months. Data are shown as mean ± standard deviation (SD). For analysis, two-way ANOVA and Tukey's multiple comparisons test were used (n = 5). ****P < 0.0001.

Figure 2.

Tear secretion changes and conjunctival GC loss in the IL-14α TG mice. (A) Tear secretion of IL-14α TG (red bars) significantly decreased at age of 18 months compared with that of WT mice (black bars) (n = 10). (B) PAS staining shows significant decline in the number of goblet cells in IL-14α TG mice at different time points from nine months to 18 months. (C) Representative PAS staining images of each group at age of six, 12, and 18 months. Data are shown as mean ± SD. For analysis, two-way ANOVA and Tukey's multiple comparisons test were used (n = 5). ****P < 0.0001. Scale bar: 200µm.

Aberrant Differentiation of Corneal Epithelial Cells in the IL14α TG Mice

H&E staining showed that the corneal surface was roughened along with the progression of dry eye in IL14α TG mice (Fig. 3A). The corneal epithelium became thinner at the age of 18 months, and the central cornea displayed stromal edema with increased cell infiltration (Fig. 3A). K10, a biomarker of cornea aberrant differentiation of epidermis-specific cytokeratin in dry eye,^19,20 was expressed in the corneal superficial epithelial cells of IL14α TG mice at age nine months and spread to reach almost the full thickness of the corneal epithelium at 18 months (Fig. 3B). Sprr1B, a biomarker of squamous metaplasia, was not expressed in the WT mice corneal epithelium, although it was expressed in the corneal epithelium of the IL14α TG mice from the age of six months (Fig. 3C). Real-time PCR revealed that gene expressions of K10 (Fig. 3D) and Sprr1B (Fig. 3E) were significantly upregulated in the corneal epithelium of IL14α TG mice from nine months and six months, respectively, and gradually increased with time, although K10 and Sprr1B was not detected in the corneal epithelium of the WT mice throughout the observation period. These data confirmed that keratinization of corneal epithelium in IL14α TG mice was initiated at the early stage of nine months old. It was reported that inflammatory factors could destroy cornea epithelial barrier function and induce corneal epithelium aberrant differentiation. Furthermore, we found that gene expressions of inflammatory cytokines, including TNF-α (Fig. 3F), IL-1β (Fig. 3G), and IFN-γ (Fig. 3H), were significantly increased in the corneal epithelium of IL14α TG mice at 15 months old. Although IL-17A (Fig. 3I) mRNA transcripts in IL14α TG mice were not significantly changed, compared with that of WT mice. ELISA assay was applied to determine the protein levels of the factors, whereas the concentration of the factors in the cornea fell below the limit of detectability (data not shown). Taken together, the aberrant differentiation of corneal epithelium at the early stage was not related with the expression levels of the inflammatory factors in the corneal epithelium.

Figure 3.

Abnormal differentiation of corneal epithelial cells in the IL14α TG mice. (A) Representative images of H&E staining in the cornea from ages of six to 15 months. Although corneal epithelium thickness decreases, there is cell infiltration into the central corneal stroma at the age of 18 months. (B, C) Representative immunofluorescent staining images of K10 and Sprr1B. K10 (D) and Sprr1B (E) mRNA expression is significantly up-regulated in the corneal epithelium of IL14α TG mice from the ages of nine months and six months, respectively. (F, G, H) RT-PCR results show that mRNA expression levels of TNF-α, IL-1β, and IFN-γ are significantly up-regulated in the corneal epithelium of IL14α TG mice from the age of 15 months. (I) The IL-17A mRNA expression levels are not different between IL14α TG mice and WT mice. Data are shown as mean ± SD. For analysis, two-way ANOVA and Tukey's multiple comparisons test were used (n = 6). ****P < 0.0001. Scale bars: 100 µm.

T-Cell and B-Cell Infiltration in Conjunctiva of the IL14α TG Mice

To further investigate conjunctival pathological changes of IL14α TG mice, first we performed H&E staining. Interestingly, there was a cluster of cell infiltrates underneath the conjunctival epithelium of IL14α TG mice as early as six months of age, and they gradually increased with time and became more diffuse and compacted beneath the fornix conjunctiva (Fig. 4A). Moreover, PAS staining showed that there was an obvious decrease in goblet cell density in the IL14α TG mice (Figs. 2B, 2C).

Figure 4.

T-cell and B-cell infiltration in the IL14α TG mice conjunctiva. (A) Representative H&E staining images of conjunctiva from the age of six to 15 months. In the IL14α TG mice conjunctiva, cell infiltration appears underneath the epithelial layer at the age of six months (arrows) and clearly increases with age. (B) IHC staining shows that CD4⁺ T cells are present in the conjunctival stroma of the IL14α TG mice (arrows). (C) Immunofluorescent staining shows that CD8⁺ T cells are present in the conjunctival stroma (arrows). (D) B220 immunofluorescent staining in the conjunctiva of IL14α TG mice at the age of 15 months (arrows). Scale bars: 100 µm.

Previous studies have shown that increases in CD4⁺ T cell infiltration contribute to losses in goblet cell density, and corneal epithelial barrier disruption.^21,22 We then performed immunostaining to determine the phenotype of the infiltrated cells in the conjunctiva. CD4⁺ T cells and CD8⁺ T cells emerged beneath the conjunctival epithelium of IL14α TG mice at age six months and rapidly increased from nine months to 15 months, although there were sporadic CD4⁺ T cells and CD8⁺ T cells distributed in the conjunctiva of WT mice from 12 months to 15 months (Figs. 4B, 4C). We also detected B220⁺ B cells beneath conjunctival epithelium of IL14α TG mice at 15 months, although no B cells were present in the conjunctival tissue of WT mice (Fig. 4D). Taken together, there was a tendency toward increasing numbers of CD4⁺ T cells, CD8⁺ T cells and B220⁺ B cells in the conjunctival tissue with age in IL14α TG mice.

Conjunctival Inflammation and Malfunction in IL14α TG Mice

Ocular surface inflammation is the key pathophysiological alteration in dry eye.²³ Quantitative real-time PCR and ELISA were used to examine the expression levels of inflammatory cytokines. The results showed that there was a significant age-related increase of TNF-α, IL-1β, IFN-γ and IL-17A mRNA levels and proteins levels in IL14α TG mice between ages nine to 15 months, compared with that of the WT mice (Figs. 5A–H).

Figure 5.

Expression levels of proinflammatory cytokines, mucin, and differentiation related genes in the conjunctiva of the IL14 α TG mice. The gene expression levels of TNF-α (A) and IL-1β (B) increase in the IL14 α TG mice from nine months to 15 months. IFN-γ (C) is upregulated at the ages of nine and 12 months in the IL14 α TG mice. IL-17A (D) is upregulated at nine months and gradually increases with time. TNF-α (E), IL-1β (F), IFN-γ (G), and IL-17A (H) protein expression levels increase in the conjunctiva in the IL14α TG mice from nine months to 15 months. MUC5AC (I) and MUC5B expression levels start to decrease (J) at the age of 9 months in the IL14 α TG mice. The K10 (K) expression is significantly higher at the age of 15 months in the IL14α TG mice, whereas the conjunctional Sprr1B expression level is invariant (L) throughout the observation period. Data are shown as mean ± SD. For analysis, two-way ANOVA and Tukey's multiple comparisons test were used. For analysis, two-way ANOVA and Tukey's multiple comparisons test were used (n ≥ 4). *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

To determine whether the secretory function of goblet cells in the IL14α TG conjunctiva was impaired, we assessed the gene expression levels of MUC5AC and MUC5B. A significant decrease of MUC5AC and MUC5B expression were detected as early as nine months in IL14α TG mice compared to their age-matched WT controls, and the expression levels kept reducing from nine months to 15 months (Figs. 5I, 5J). To determine whether there was abnormal differentiation of conjunctival epithelial cells, we examined the expression levels of K10 and Sprr1B in the conjunctival tissue. We only detected a significant increase of K10 expression at an age of 15 months in IL14α TG mice, compared with WT mice (Fig. 5K). There was no dramatic change in the Sprr1b expression levels between the two groups at different time points (Fig. 5L). Based on the above data, we deduced that the changes in the conjunctival and corneal epithelial phenotype were caused by cell infiltration and rises in cytokine release in the conjunctiva from nine months.

Lymphocyte and Macrophage Infiltration in the Lacrimal Gland of the IL14α TG Mice

To clarify the effect of changes in lacrimal gland function on the ocular surface, we identified the temporal changes in the lacrimal gland in the IL14α TG mice. Previous study has shown that lymphocyte infiltration in the lacrimal glands of IL14α TG began at the age of 9 months, and significantly increased with age and became severe at age 18 months.¹⁸ To determine what types of immune cells infiltrated into the lacrimal glands of the IL14α TG, we performed CD4, CD8, F4/80, and B220 immunohistochemical staining. Eighteen-month-old WT lacrimal glands served as the control group. We found that CD4⁺ (Fig. 6A) and CD8⁺ (Fig. 6B) T cells and macrophages (Fig. 6C) were present at age nine months and increased with time, whereas B220⁺ B cells (Fig. 6D) appeared from 12 months. Interestingly, most of the macrophages were located along the edges of the infiltrated foci (Fig. 6C). These data indicated that CD4⁺ T cells, CD8⁺ T cells, B220⁺ B cells and macrophages were all involved in the process of inflammation in the lacrimal gland of IL14α TG mice from nine months.

Figure 6.

CD4+, B220+ and CD8+ dominant lymphocytic and macrophage infiltration in the lacrimal gland of the IL14α TG mice. Representative IHC images demonstrated CD4+ T cells (A), CD8+ T cells (B), and F4/80+ macrophages (C) were found in the lacrimal gland began at age of nine months; B220+ B cells (D) at age of 12 months. Cell infiltration gradually increased with time. Inserts stand for enlarged images of arrow areas. Scale bars: 200 µm.

Lacrimal Gland Dysfunction in the IL14α TG Mice

As shown in Figure 1A, there was a time-lag between lymphocyte infiltration and development of lacrimal gland dysfunction. Aqueous tear production was invariant until 18 months of age whereas lacrimal gland inflammatory cell infiltration started from nine months old in the IL14α TG mice. To determine this disconnect, we conducted a variety of lacrimal gland function tests to pin down the time at which functional losses began. The α-SMA is a specific myoepithelial cell marker, which can enhance the secretory function of exocrine glands.²⁴ RT-PCR and immunostaining showed reduced mRNA and protein levels of αSMA at 12 months of age in IL14α TG mice compared with WT group, and they both further decreased at 18 months (Figs. 7A, 7B). The aquaporins (AQPs), AQP4 and AQP5, belong to the family of water channel proteins, and they are expressed in the lacrimal gland. These conduits play an important role in lacrimal gland acinar cell aqueous tear secretion.²⁵ We found that AQP4 underwent significant declines in expression starting at 12 months of age (Fig. 7C), and this also occurred in AQP5 at 18 months (Fig. 7D) in IL14α TG mice. We observed decreases in VEGF (Fig. 7E) and epidermal growth factor (EGF) (Fig. 7F) expression levels in the lacrimal gland of IL14α TG mice at an age of 12 months compared to age matched WT mice, and more dramatic decreases at the age of 18 months. While hepatocyte growth factor (HGF) expression showed a transient increase in IL14α TG mice at age six months, which became more pronounced at age 18 months (Fig. 7G). Masson's trichome staining was performed to evaluate lacrimal gland fibrosis. We found positive staining in the lacrimal gland at ages of 15 and 18 months in IL14α TG mice (Fig. 7H). In summary, water and protein secretory function of lacrimal gland was impaired starting at an age of 12 months old in IL14α TG mice, and tissue fibrosis had become evident at age 18 months.

Figure 7.

Impairment of lacrimal gland function at a late stage in the IL14α TG mice. (A) α-SMA immunofluorescent staining (red) shows that it is less pronounced in the lacrimal gland of IL14α TG mice at the age of 12 months, and it decreases more at ages of 15 months and 18 months (star area). Scale bars: 50 µm. (B) The α-SMA mRNA expression level decreases at the age of 12 months. The AQP4 (C) and EGF (F) expression levels are significantly lower at the ages of 12 months and 18 months in the lacrimal gland of IL14α TG mice. The AQP5 (D)and VEGF (E) expression levels are significantly lower at the ages of 18 months in the lacrimal gland of IL14α TG mice. (G) HGF mRNA expression levels increase at the age of six months, whereas they decrease at the age of 18 months. (H) Representative Masson's trichome staining is positive at ages of 15 and 18 months in the lacrimal gland of IL14α TG mice (white arrows). Scale bar: 200 µm. Data are shown as mean ± SD. For analysis, two-way ANOVA and Tukey's multiple comparisons test were used (n ≥ 5). **P < 0.01; ****P < 0.001.

Lymphocyte Infiltration in the IL14α TG Mice Meibomian Gland

The previously documented SS patients exhibit greater Meibomian gland dropout of the upper eyelid than non-SS patients.²⁶ To explore the Meibomian gland histological changes in the SS, we assessed this change in IL14α TG mice. There were no prominent Meibomian gland changes in either the six- or 12-month age groups, whereas lymphocyte infiltration was present at 15 months of age (Supplementary Fig. S1). Our data show that T- and B-cell infiltration in the conjunctiva may be responsible for their infiltration into the Meibomian glands because the Meibomian gland is located close to the tarsal conjunctiva.

Discussion

The changes in the ocular surface resulting from SS progression have not been systematically studied. In the current study, we found that the IL14α TG mouse displayed a spontaneous dry eye, including corneal epithelial squamous metaplasia, barrier function damage, decreases in tear secretion, and losses in conjunctival goblet cell density. Moreover, IL14α TG mice progressively developed a chronic ocular surface inflammation.

A variety of animal models have been reported to study the pathological mechanism of SS. Nonobese diabetic (NOD) and IQI/Jic mice showed massive lacrimal gland lymphocytic infiltration, whereas the ocular surface integrity of the mice remained.¹⁰ NZB/W F1 and MRL/lpr mice are considered as secondary SS models.²⁷ Repeated intraperitoneal injection of short peptides from 60-kDa Ro antigen to BALB/c mice recapitulated hyposalivation and lymphocyte infiltration in the salivary glands²⁸ without lacrimal gland lymphocytic infiltration. The rabbit SS model was developed through injecting proliferating autologous mixed lymphocytes into lacrimal gland.²⁹ CD25 knock-out (KO) mice showed early onset of Sjögren's syndrome dry eye (SSDE) with no involvement of an immune response.⁸ Thrombospondin-1 KO mice, aire KO mice and C57BL6.NOD-Aec1Aec2 (Aec) mice were also used for investigating corneal and conjunctival epithelial injury associated with dry eye. However, thrombospondin-1 KO mice developed SSDE with increasing density of goblet cells without salivary gland involvement.³⁰ Aire KO mice exhibited severe lymphocytic infiltration of the lacrimal gland at five weeks of age, but it was difficult to identify the early onset of SSDE.³¹ Aec mice displayed losses in goblet cell density from four weeks to 20 weeks of age; however, the function of lacrimal gland remained normal.³² In our study, IL14α TG mice presented many features that mimic chronological changes in SS patients. This difference allowed us to clarify details about the early events and progression of SS that underlie ocular surface changes.

In the current study, we found that IL14α TG mice developed ocular surface inflammatory features of SS over a relative time frame similar to that in SS patients. Figure 8 summarizes the temporal pathological changes in the ocular surface of IL14α TG mice in accordance with the stages described in a previous report.¹² As noted, there was CD4⁺ T cell infiltration of conjunctiva and corneal epithelium squamous metaplasia at stage 1 (3–6 months). Squamous metaplasia is the hallmark of dry eye when the ocular surface epithelial cells undergo aberrant differentiation.³³ After that, corneal epithelial barrier function damage, CD8⁺ T cells infiltration of conjunctiva and losses in goblet cell density were detected at stage 2 (6–10 months). Although, lymphocytic infiltration of lacrimal gland was also identified at this stage, the function of lacrimal gland remained normal. At stage 3 (10–14 months), the expression levels of lacrimal gland function associated genes such as AQP4, AQP5, EGF, HGF, and VEGF started to decrease. EGF and HGF are expressed by the lacrimal gland and essential for promoting ocular surface tissue regeneration.³⁴ The tear VEGF level decrease was associated with inflammation and tissue injury of ocular tissues in the NFS/sublingual gland mutant SS mice model.³⁵ Lymphocytic infiltration of conjunctiva and corneal epithelial barrier function damage became even worse. At stage 4 (14–18 months), lacrimal gland fibrosis and decreases in tear secretion were observed (Fig. 6). Previous studies have demonstrated that the IL14α TG mice appeared interstitial lung disease and mild renal impairment appeared, at stage 4 (14–18 months).^12,36 Our results imply that the conjunctiva is more vulnerable at an early stage than the lacrimal gland to disruption of ocular surface function during SS progression.

Figure 8.

Time-dependent pathological changes in the IL14α TG mice ocular surface. The pathological changes of different ocular surface tissues in IL14α TG mice are summarized in a chronological format.

In the current study, lymphocyte infiltration is present in the conjunctiva at an age of six months. The function of conjunctiva was disturbed at age nine months whereas the function of lacrimal gland was not affected. Keratoconjunctivitis sicca was traditionally thought to be the result of tear deficiency caused by LG disease. However, many SS patients presented with clinically significant ocular surface disease, but without aqueous tear deficiency.³⁷ In our study, the increases in the number of T cells infiltrating the conjunctiva, paralleled with losses in conjunctival goblet cell density and increases in corneal epithelial barrier disruption. The expression levels of IFN-γ and IL-17A rose in the conjunctiva in nine-month-old IL14α TG mice. IFN-γ secreted by T helper 1 (Th1) cells can lead to corneal epithelial apoptosis³⁸ and decreases in goblet cell density.³² TNF-α, IL-1β, and IL-17A can disrupt corneal barrier function.^22,39,40 The increases in cytokine expression levels may lead to conjunctival dysfunction. Goblet cells in the conjunctiva produce gel-forming soluble mucin proteins such as MUC5AC and MUC5B,⁴¹ which are tear components playing pivotal role in the homeostasis of tear film.⁴¹ Based on these data, we propose that SSDE develops as a consequence of Th1 and Th17 cells infiltration into the conjunctiva. This response accounts for the development of conjunctival dysfunction and DE symptomology at an early stage of SS in IL14α TG mice. Whether this mechanism holds true in humans requires further clinical investigation.

Conjunctival inflammation in SS has been reported in many studies. Diebold et al.⁴² reported that inflammation in the conjunctiva is present in the 9-week-old MRL/MpJ-Fas^lpr mice. Pflugfelder et al.⁴³ found that antigen-presenting cells and mature dendritic cell infiltration in the conjunctiva is associated with the severity of DE in SS. Previous studies indicated that IFN-γ–secreting CD4⁺ T cell leads to corneal epithelial apoptosis in the desiccating stress model.³⁸ In our study, we claimed that CD4⁺ T cell infiltration may induce conjunctival goblet cell loss, corneal epithelial apoptosis, and abnormal differentiation before onset of aqueous tear deficiency. Our results are corroborated by findings from the Pflugfelder group. They also showed an increased number of CD4⁺ T cells infiltrating the conjunctival epithelium in C57BL/6.NOD-Aec1Aec2 mice, and this T-cell infiltration was accompanied by conjunctival goblet cell loss independent of lacrimal dysfunction.⁴⁴ Our study provided the evidence that conjunctival inflammation may play an important role in the SS-associated ocular surface epithelial change and ocular discomfort, even in the context of normal aqueous tear production or volume.

Our findings may help to diagnose SS at an early stage. Based on the classification criteria for SS established by the American-European Consensus Group, a diagnosis of SS in patients was made at a stage when the salivary gland and lacrimal gland were infiltrated by massive increases in lymphocytes and irreversible function loss.⁴⁵ Therefore it is important to classify SS or non-SS-related DE patients before the appearance of irreversible complications. In this study, we found that the conjunctiva was significantly infiltrated by CD4⁺ cells and had undergone losses in function before those appearing in the lacrimal gland. An evaluation of changes in the proinflammatory cytokine expression profile may also help to diagnose conjunctival dysfunction. Th-17 cells are classically identified by their production of IL-17A. IL-17A expression levels are reported to increase in tears and saliva from SS patients.^46,47

In a recent study, γδ T cells were 25.64% of the conjunctival resident immune cells in six- to eight-week-old mice by flow cytometry.⁴⁸ The role of γδ T cells in SS are largely unknown. IL-17A plays a crucial role in the development and progression of dry eye.⁴⁹ IL-17A-producing γδ T (γδ T17) cells ranged from 36.70% to 88.00% in various studies.^48,49 In our results, IL-17A was upregulated in the IL14α TG mice. These suggested that γδ T17 cells not only CD4⁺ but also CD8⁺ T cells could play essential physiological roles in the ocular surface. However, the underlying mechanism needs further study.

There are certain limitations of our current study. In this article, we found that conjunctival inflammation occurred at a relatively early stage in the IL14α TG mice. The mechanism of this pathological change is still not completely understood yet. How CD4⁺ and CD8⁺ T cells recruited into conjunctiva and the role of IL14α on the ocular surface remains largely unknown. IL14α inhibitor and neutralizing antibody would provide ways to identify the mechanism. On the other hand, the role of conjunctival tissue resident memory T cells in the pathological change of primary SS dry eye need further investigation.

In summary, a relatively more complete picture is presented on the ocular surface changes and progression that are associated with the development of early SS-related DE in the IL14α TG mice. This is the first demonstration that losses in conjunctival function resulting from SS damage are a frontline event that occurs during the early onset stage of SSDE. This preclinical evidence makes it possible for a clinician to provide an earlier and more-precise diagnosis. Accordingly, it may be possible to obtain more effective therapeutic management of SS and increase the chance for a more favorable outcome than could be obtained if therapeutic intervention is instead initiated at a later stage.