Limbal stem cell deficiency causes conjunctivalization characterized by the covering of the corneal surface with conjunctival epithelium. However, the driving force for the encroachment of

these conjunctival cells is unclear. Conjunctival stem cells are bipotent stem cells that can proliferate and differentiate into conjunctival epithelial cells and goblet cells to maintain

regeneration of the conjunctival epithelium. Here, we show a robust proliferative response of conjunctival stem cells and upregulation of Wnt2b and Wnt3a gene expression in the conjunctivae

of mice with induced limbal stem cell deficiency. Topical application of the Wnt/β-catenin signaling activator CHIR resulted in increased proliferation of ΔNp63α-positive stem cells in the

preventing the differentiation of these cells. These expanded stem cells formed a stratified epithelium containing goblet cells under airlift culture conditions. Our data reveal that

Wnt/β-catenin signaling contributes to the pathological process of limbal stem cell deficiency by promoting the self-renewal of conjunctival stem cells and suggest that these cells are a

driving force in corneal conjunctivalization.

The cornea is a highly organized transparent tissue that covers the front portion of the eyeball. Limbal stem cells (LSCs) play a crucial role in maintaining corneal epithelial integrity,

which is essential for corneal transparency1,2,3. Found in the basal epithelial layer of the limbus, a transition zone between the cornea and conjunctiva, LSCs act as a barrier to prevent

the migration of the conjunctival epithelium onto the cornea4. Injury to the limbal epithelium, immune disorders affecting the ocular surface (such as Stevens–Johnson syndrome), and

long-term contact lens wearing can result in partial or total limbal stem cell deficiency (LSCD)5, which leads to a process referred to as conjunctivalization. This pathological process

entails encroachment of conjunctival epithelium containing goblet cells onto the corneal surface, resulting in corneal vascularization, corneal opacification, and loss of vision6,7,8.

Although LSCD is characterized by conjunctivalization, the molecular and cellular mechanisms underlying this process remain largely unclear.

The conjunctival epithelium is a self-renewing tissue consisting of nonkeratinized stratified columnar cells and interspersed goblet cells. Previous studies have reported the existence of

mitotically active stem cells in the bulbar and forniceal conjunctivae9,10,11. These cells have the capacity to self-renew, proliferate, and differentiate into epithelial cells and goblet

cells, as has been shown in clonal culture assays12,13. This indicates that conjunctival stem cells (CjSCs) can constitute an ultimate source for conjunctival epithelial renewal. However, it

is still unknown whether CjSCs contribute to corneal conjunctivalization.

Wnt signaling plays important roles in stem cell self-renewal and differentiation in adults as well as in embryonic development14,15. Wnt genes encode evolutionarily conserved secreted

glycoproteins that activate autocrine and paracrine cell signaling pathways, including the β-catenin/TCF pathway, the planar cell polarity pathway, and the Wnt/Ca2+ pathway16,17. The

canonical β-catenin/TCF pathway is activated by the binding of Wnt ligands to FZD receptors and the coreceptors LRP 5/6, resulting in inactivation of GSK-3β, which stabilizes and

translocates β-catenin into the nucleus. Nuclear β-catenin acts as a transcriptional coactivator for TCF/lymphoid enhancer factor (LEF) DNA-binding transcription factors.

Several Wnt genes are expressed in the cornea and limbus and are upregulated after injury18,19. Nuclear localization of β-catenin is detectable in proliferating cells of the limbal

epithelium. Moreover, activation of Wnt/β-catenin signaling by inhibition of GSK-3β drives the proliferation of LSCs19. Similar to the scenario during activation of Wnt/β-catenin signaling

in the limbus, some conjunctival epithelial cells were reported to express β-catenin and TCF4 in the nucleus, along with CjSC markers, during in vitro expansion20,21,22. In addition, a

previous study reported that deletion of Dkk2, an antagonist of Wnt/β-catenin signaling, resulted in corneal epithelial hyperplasia and invasion of conjunctival tissue into the corneal

periphery in the developing ocular surface23. These results suggest the possibility that Wnt/β-catenin signaling acts as a key modulator of CjSCs, which promote conjunctivalization. Here, we

show that LSCD causes upregulation of canonical Wnt ligand expression and the proliferative activity of CjSCs in the conjunctival epithelium. We also demonstrate that activation of

Wnt/β-catenin signaling regulates CjSC self-renewal and differentiation and enhances conjunctivalization of the corneal surface.

Human conjunctival tissues were obtained from the eye bank of Eversight (Chicago, IL, USA). The conjunctiva was retrieved from seven donors with a mean age of 65 years. This study was

approved by the Institutional Review Board of Daejeon St. Mary’s Hospital (DC17TNSI0055) and was conducted in accordance with the tenets of the Declaration of Helsinki. Human conjunctival

tissues were incubated in Hank’s balanced salt solution (Welgene, Gyeongsangbuk-do, Republic of Korea) containing 1% antibiotic-antimycotic (Corning, Inc., Corning, NY, USA) for 2 h at 37 

°C. The conjunctival epithelium sheets were divided into several fragments of approximately 0.5–1 mm. Four or five fragments of the conjunctival epithelium were placed in 3-cm culture

dishes. Primary conjunctival cells were maintained in CjSC medium (1:3 mixture of Dulbecco’s modified Eagle’s medium and Ham’s F12 medium [Welgene] supplemented with 5% fetal bovine serum

[FBS; Corning, Amsterdam, The Netherlands], 1% penicillin–streptomycin [Welgene], 10 ng/mL human epidermal growth factor [Sigma, St. Louis, MO, USA], 5 µg/mL insulin [Sigma], 30 ng/mL

cholera toxin [Sigma], and 500 ng/mL hydrocortisone [Sigma]). The cultures were maintained at 37 °C in an atmosphere containing 5% CO2, and the culture medium was changed every other day.

For CjSC culture, conjunctival stem cells were cultured on a feeder layer. Cells were detached using trypsin-EDTA (Welgene) solution. Cells were seeded on the prepared NIH3T3 feeder layer at

a density of 2.5 × 103 cells/cm2. Cocultures were treated with 10 µg/mL Y-27632 (Tocris Bioscience, Bristol, UK) and maintained in CjSC medium. CjSCs were treated with DMSO or 3 µM

CHIR99021 (Tocris Bioscience) to activate Wnt/β-catenin signaling. For differentiation, CjSCs were maintained in Roswell Park Memorial Institute (RPMI) 1640 medium supplemented with 10% FBS

(Corning, Inc.) and 1% penicillin–streptomycin.

For airlift culture, CjSC colonies treated with CHIR99021 or DMSO for 5 days were isolated mechanically and dissociated into single cells by trypsin-EDTA digestion. Cells were seeded at a

density of 4 × 104 cells/cm2 in Transwell inserts coated with 0.1% gelatin. After the CjSCs were confluent, they were cultured for 14 days under airlift conditions by lowering the

compartment containing the CjSC medium to the bottom of the insert.

Eight-week-old female C57BL/6 J mice were used in this study. All of the animals were treated according to the standards of the Association for Research in Vision and Ophthalmology Statement

for the Use of Animals in Ophthalmic and Vision Research. For induction of LSCD, mice were anesthetized with a 1:10 dilution of a 1:1 mixture of ketamine and xylazine. For topical

anesthesia, 0.5% Alcaine (Alcon, Fort Worth, TX, USA) was applied to the ocular surface. The mouse ocular surface was treated with 40% alcohol (Honeywell, Charlotte, NC, USA) using a 3 mm

corneal trephine (Miltex, York, PA, USA) for 35 s. The ocular surface was irrigated with 5 mL of phosphate-buffered saline (PBS; Welgene) to wash off the remaining alcohol. The treated area

of the ocular surface epithelium was removed with a spatula, and the surface was irrigated with PBS. Following wound formation, the ocular surface was treated with 300 µM CHIR or DMSO four

times per day. The wound area was monitored by staining with sodium fluorescein to assess healing.

The eyes and lids of mice were removed by enucleation and immersed in 4% paraformaldehyde (PFA) overnight at 4 °C. The airlift cultures of CjSCs were fixed for 30 min with ice-cold 4% PFA.

After fixation, the tissues and airlift cultures were washed, cryoprotected by immersion in 30% sucrose, and embedded in optimal cutting temperature compound. Then, they were flash-frozen in

liquid nitrogen and stored at –80 °C. Cryosections were sliced at a thickness of 10 μm with a cryostat (Leica, Wetzlar, Germany) and mounted on glass slides. The sections were thawed and

postfixed with 4% PFA for 10 min. For histological assessment, sections were permeabilized with 0.5% Triton X-100 and stained with PAS (Sigma) and Alcian blue (Abcam) staining kits according

to the manufacturer's instructions. Images were acquired using a microscope (Leica).

For immunostaining, tissue and airlift culture sections were washed with PBS and incubated for 1 h with 2% bovine serum albumin (Gibco, Dublin, Ireland) and 5% normal goat serum (Jackson

ImmunoResearch, West Grove, PA, USA) diluted in PBS. The sections were sequentially incubated overnight at 4 °C with primary antibodies against rabbit K13 (Abcam), rabbit K7 (Abcam), rabbit

ΔNp63α (BioLegend, San Diego, CA, USA), mouse ABCG2 (Santa Cruz Biotechnology, Dallas, Tx, USA), rabbit PAX6 (BioLegend), mouse Cyclin D1 (Abcam), rabbit β-catenin (Sigma), rabbit active

β-catenin (Cell Signaling Technology, Danvers, MA, USA), and mouse Ki67 (Beckman Coulter, Brea, CA, USA). After washing with PBS, the sections were incubated for 1 h at room temperature with

Alexa Fluor 488- or Cy3-conjugated secondary antibodies (Thermo Fisher Scientific, Waltham, MA, USA) and counterstained with Hoechst 33258 (Sigma). Images were acquired using a fluorescence

microscope with an Axiocam camera (Zeiss, Jena, Germany). Cells grown on coverslips were fixed with ice-cold 4% PFA for 15 min at room temperature, permeabilized with 0.2% Triton X-100, and

immunostained as described above. The number of antibody-labeled cells was quantified in eight randomly selected fields per section or coverslip. Data obtained from at least three

independent experiments were averaged and are presented as the mean ± S.D. values. Two-tailed Student’s t test was used to compare two experimental groups.

CjSC proliferation was assessed by injecting 1 mg of BrdU (Sigma) in 200 µL of PBS (Welgene) after LSC removal or on the day after wounding. Tissue sections were postfixed with 4% PFA,

permeabilized, and incubated with an anti-ΔNp63α antibody at 4 °C overnight. Then, the sections were incubated in 1 M HCl for 30 min at 37 °C and washed three times for 10 min each with 0.1 

mM borate buffer. After the tissue sections were washed with PBS, a rat anti-BrdU antibody (Abcam) was applied and incubated overnight at 4 °C. Then, the slides were incubated with the

appropriate secondary antibody for 1 h at room temperature.

For double immunostaining of EdU and ABCG2, cells grown on coverslips were treated with 5 µM EdU (Thermo Fisher Scientific) 6 h before fixation. Then, the cells were fixed with 4% PFA,

permeabilized with 0.2% Triton X-100, and subjected to immunostaining with an anti-ABCG2 antibody, as described above. To detect EdU, Click-iT EdU Imaging Kits (Thermo Fisher Scientific)

were used according to the manufacturer’s instructions.

Total RNA was isolated from conjunctival cells and conjunctival tissues using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s instructions. cDNA was

synthesized from RNA by reverse transcription with a Superscript III Kit (Invitrogen). PCR amplification was carried out using specific primer pairs and an nTaq-HOT Kit (Enzynomics, Republic

of Korea). Quantitative real-time PCR was performed using SmartGeneTM Q-PCR Master Mix (SmartGene, Republic of Korea), and expression in the samples was quantified by amplifying GAPDH as

the internal control for each sample. All experiments were performed at least three independent times.

We used chemical and mechanical epithelial debridement to induce LSCD in mice. Removal of the entire corneal and limbal epithelia was confirmed by fluorescein staining (Fig. 1a). On Day 7

after epithelial debridement, fluorescein staining was not observed, indicating that the wound area was almost completely resurfaced. Corneal opacity was also observed with vessel formation

after 2 weeks. On Day 7 post wounding, immunostaining for K7 and K13 (markers of conjunctival epithelial cells) revealed the expression of these proteins in the suprabasal and superficial

cell layers of the corneal surface as well as in the conjunctival epithelium (Fig. 1b)24,25. Moreover, Alcian blue-positive goblet cells were detected in the corneal periphery (Fig. 1c and

Supplementary Fig. 1). These results indicated LSCD-like pathological features such as corneal conjunctivalization and neovascularization.

a Induction of LSCD in C57BL/6 mice using complete corneal and limbal epithelial debridement. The corneas were stained with sodium fluorescein immediately (0 days) and 7 days after

epithelial debridement. The defect was completely resurfaced at 7 days. Light microscopy 2 weeks post-wounding showed blood vessels penetrating into the cornea and severe corneal opacity. b,

c The wound inflicted by epithelial debridement resulted in corneal conjunctivalization. Sagittal sections from unwounded and wounded ocular surface tissues were immunostained with

antibodies against the conjunctival epithelial cell markers K7 and K13 (b). Hoechst dye was used as a counterstain. K7 and K13 expression were detectable at the corneal surface on Day 7

post-wounding. Goblet cells appeared at the periphery of the resurfaced cornea 7 days after wounding (c). Sagittal sections from unwounded and wounded ocular surface tissues were stained

with Alcian blue. The sections were counterstained with hematoxylin and eosin. The arrowheads indicate Alcian blue-positive goblet cells. d The proliferative activity of CjSCs was increased

in the bulbar and forniceal conjunctivae of mice with LSCD. BrdU was applied on Day 2 post-wounding. On Day 3 after wounding, ocular surface tissue sections were immunostained with

antibodies against BrdU and ΔNp63α (a putative marker of CjSCs). e Quantitative analysis of ΔNp63α/BrdU double-positive CjSCs in the conjunctivae of mice with LSCD (wounded) and on a normal

ocular surface (unwounded). The error bars indicate the mean ± S.D. of three independent experiments. ***P