Effects of 1,25 and 24,25 Vitamin D on Corneal Epithelial Proliferation, Migration and Vitamin D Metabolizing and Catabolizing Enzymes

This study investigated the effects of 1,25(OH)2D3 and 24R,25(OH)2D3 on corneal epithelial cell proliferation, migration, and on the vitamin D activating enzyme CYP27B1 (produces 1,25(OH)2D3) and inactivating enzyme CYP24A1 (produces 24R,25(OH)2D3). The role of the vitamin D receptor (VDR) was also examined. In VDR wildtype mouse corneal epithelial cells (WT), 1,25(OH)2D3 increased CYP24A1 protein expression and decreased CYP27B1 expression. In VDR knockout mouse epithelial cells (KO), 1,25(OH)2D3 increased CYP24A1 and CYP27B1 protein expression. 1,25(OH)2D3 did not affect WT cell proliferation, but did stimulate VDR KO cell proliferation. In a human corneal epithelial cell line (HCEC), 1,25(OH)2D3 increased CYP24A1 mRNA and protein expression. 1,25(OH)2D3 increased CYP27B1 mRNA levels in HCEC, but had no effect on CYP27B1 protein levels. 1,25(OH)2D3 inhibited HCEC proliferation and stimulated cell migration in primary human epithelial cells. 24,25(OH)2D3, on the other hand, increased both CYP24A1 and CYP27B1 protein expression in WT and VDR KO cells, and stimulated cell proliferation in both WT and KO cells. In HCEC, 24,25(OH)2D3 increased CYP24A1 and CYP27B1 mRNA and protein expression, and stimulated cell migration. In human primary corneal epithelial cells, 24,25(OH)2D3 stimulated migration. We conclude that 24R,25(OH)2D3 is likely involved in corneal epithelial cell regulation independent of 1,25(OH)2D3 or VDR.

have examined the regulation of CYP24A1 and CYP27B1 or the physiological roles of vitamin D3 in the anterior segment of the eye 7,9 , particularly as they pertain to 24R,25(OH) 2 D3.
1,25(OH) 2 D3 is considered to be the most biologically active vitamin D metabolite. The classical nuclear vitamin D receptor (VDR) is considered the primary receptor for 1,25(OH) 2 D3. Membrane-associated protein disulfide isomerase family A member 3 (Pdia3) has been shown to be a secondary receptor for 1,25(OH) 2 D3, with each receptor separately activating its own downstream mediators 10,11 . Pdia3 binding of 1,25(OH) 2 D3 has been shown to control extracellular Ca 2+ homeostasis and regulate bone growth and metabolism [11][12][13] . 1,25(OH) 2 D3 binding with VDR was first found to promote anti-proliferative and pro-differentiation responses in cancer cells [14][15][16] . Since then, many groups have reported similar anti-proliferative effects of 1,25(OH) 2 D3 and VDR in vitro and in vivo 17 . Another important pronounced effect of 1,25(OH) 2 D3 is promotion of increased CYP24A1 synthesis, accelerating the catabolism of 1,25(OH) 2 D3 18 . Immunomodulatory activity of 1,25(OH) 2 D 3 was recently demonstrated in corneal epithelial cells where it was shown to attenuate proinflammatory mediators while increasing antimicrobial peptides and antipseudomonas activity. In this same study, the expression of CYP24A1 was found to be increased and regulated by VDR in corneal epithelial cells cultured with 1,25(OH) 2 D3 9 .
The aims of this study were to investigate the effects of 1,25(OH) 2 D3 and 24R,25(OH) 2 D3 on corneal epithelial cell proliferation, migration, and on the expression of the vitamin D activating enzyme CYP27B1 and the vitamin D metabolizing enzyme CYP24A1 in human and mouse corneal epithelial cells. VDR knockdown and knockout corneal epithelial cells were utilized to examine the role of the VDR in these responses.
Human Corneal Epithelial Cell Line and Primary Epithelial Cells. The immortalized human corneal epithelial cell line (HCEC) utilized in this study has been previously described 7,26,27 . HCEC were grown on standard culture plates until confluent. All cells were grown in Dulbecco's modified Eagle's medium (Invitrogen, Carlsbad, CA) supplemented with FBS (3%), 1% ITS (BD Biosciences, Bedford, MA), and 40 μg/mL gentamicin (Life technologies, NY). The cells were subpassaged using trypsin (Sigma, Ann Arbor, MI) digestion, seeded in 35 mm dishes (Fisher Scientific, PA), and cultured in a humidified incubator at 37 °C with 5% CO 2 . The culture medium was replaced with fresh DMEM medium plus 3% serum every 2 days.
Primary human epithelial cells were obtained from de-identified donor corneal rims courtesy of Dr. Amy Estes from the Department of Ophthalmolgy, Medical College of Georgia, Augusta University. Corneas from 3 donors were used, and all cells were obtained from transplant corneal rims within 24 hours after the surgeries were performed. The donors were from: 32 year old male, 33 year old male, and 58 year old male. Primary human epithelial cells were cultured using the same methods as for HCEC. Cells from passage three to six were used for experiments. The presence of keratin 12 and mucin 1 mRNA was tested by PCR to confirm immortalized and primary cell types (data shown in Supplemental Figure S1). Mouse primary corneal epithelial cells (MPCEC). Wildtype (WT), heterozygous (HET) and VDR knockout (KO) mice were obtained and bred from the Jackson Labs (Strain: B6.129S4-Vdr tm1Mbd ). All animal studies were approved by the Augusta University IACUC, and methods were performed and animals treated in accordance with the Augusta University IACUC guidelines and regulations as well as the ARVO statement for the Use of Animals in Ophthalmic and Visual Research. Murine cells gave us the opportunity to study the effects of full VDR knockout, as opposed to only siRNA knockdown, on vitamin D signaling in corneal epithelium. Primary mouse corneal epithelial cell cultures were established using a modification of an established explant culture method 28,29 . Many VDR knockout mice die before 12 weeks of age, and the phenotype is usually expressed between 4 and 5 weeks of age (rickets type symptoms, alopecia, etc,), thus we collected corneas from mice as early as 4 weeks of age. Briefly, 4-week-old mice were killed, the eyes were enucleated, and cornea buttons were obtained using a pathogen-free disposable punch (2.5 mm) and washed with Ca 2+ free PBS (pH 7.2). Each cornea was cut in half and placed in a 35 mm dish (Fisher Scientific, PA) with the epithelial side up. After approximately 5 minutes in the laminar air-flow culture hood, the half corneal button attached itself to the bottom of the plate, and 1.5 mL of DMEM (Life technology, NY) with 10% serum containing 40 μg/mL gentamicin (Life technologies, NY), 1% ITS (BD Biosciences, Bedford, MA), and 100 ng/mL cholera toxin (LIST Biological Laboratories, Inc., Campbell, CA) were added and the tissue was placed in a humidified incubator at 37 °C with 5% CO 2 . Culture medium was replaced every 2 days. After 7 to 10 days cells were nearly 100% confluent. Cells were passaged using 2.5% trypsin (Sigma, Ann Arbor, MI) with gentle pipetting, centrifuged at 500 g for 5 minutes, and subcultured in DMEM with 3% serum containing 40 μg/mL gentamicin, 1% ITS, and 100 ng/mL cholera toxin.
Cell Migration Assay. Migration assays were performed using a scratch-wound protocol. Primary human corneal epithelial cells were seeded onto 35 mm cell culture dishes (3.5 × 10 4 cells/dish), incubated at 37 °C until they attained 80-90% confluence, and then starved in DMEM plus 0.1% fetal bovine serum overnight at 37 °C. Cells were scratch-wounded using a sterile 10 μL pipette tip, washed two to three times with medium to remove all loose or dead cells, photographed (0 hours), and 1,25(OH) 2 D3 (10 nM) or 24R,25(OH) 2 D3 (100 nM) was added. To avoid cell proliferation in the scratch assay, serum free medium was used for the assay, and the assay was limited to 18 hours duration. Control dishes were similarly scratch-wounded and incubated without addition of vitamin D metabolites. Cells that had migrated across marked reference lines were photographed using a Hoffman contrast-equipped microscope (Olympus, Tokyo, Japan). The extent of healing over time was defined as the ratio of the difference between the original wound area and the remaining wound area after 18 hours. Data was analyzed from 10 scratches, using cells from 3 independent donors. The area of migrating cells was analyzed using CellSens Dimension software (Olympus, Tokyo, Japan). Cell Proliferation. HCEC and MPCEC were seeded onto 24-well cell culture plates (5 × 10 4 cells/dish), incubated at 37 °C until plates reached 55-65% confluence, and then starved in DMEM plus 0.1% fetal bovine serum overnight at 37 °C. Cells were stimulated with 1,25(OH) 2 D3 (10 nM) or 24R,25(OH) 2 D3 (100 nM). The MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide] reduction assay was used as an index of viable cell density and was performed according to the manufacturer's instructions (Fisher Scientific, PA). Control and vitamin D-treated cells were incubated in serum-free medium containing 0.4 mg/mL MTT. During this time, mitochondrial and cytosolic dehydrogenases of living cells reduced the yellow tetrazolium salt (MTT) to a purple formazan dye capable of spectrophotometric detection. After 2 to 2.5 hours, the MTT solution was aspirated and dimethylsulfoxide (0.3 mL/well) was added. Optical densities of the supernatant were read at 540 nm using a microplate spectrophotometer (BioTek, Winooski, VT). Absorbances were normalized to the control cultures, which represented 100% viability. Growth ratios were calculated as the ratio of the absorbance of vitamin D-treated cells versus control cells.
Real Time PCR. Total RNA was obtained from HCEC, primary mouse epithelial cells, and primary human epithelial cells. Real time PCR was used to quantify CYP24A1 and CYP27B1 mRNA levels. GAPDH was used as the internal RNA control for human corneal epithelial cells, and TATA box binding protein (TBP) was used as the mouse corneal epithelial cell RNA control. The RT-PCR primers for human and mouse CYP24A1, mouse CYP27B1, human GAPDH, and human VDR were generated from the PrimerBank database 30,31 using National Center for Biothechnology Information (NCBI) sequence identification numbers (NM_009996, NM_010009, NM_001017536, NM_001289746.1, NM_001128915 for mouse CYP24A1, mouse CYP27B1, human VDR, human GAPDH, and human CYP24A1, respectively). The primers for CYP27B1 were designed using data from human decidual cells 2,32 . The primers for mouse TATA box binding protein (TBP) were generated from Universal ProbeLibrary of Roche Life Science. The primer specificity was validated with melting profiles and the electrophoresis gels of RT-PCR products (data shown in supplementary Figures S2-S4). Primers are listed in Table 2.
mRNA was isolated and cDNA was synthesized using the Bio-Rad RT-PCR system. First-strand synthesis was done at 42 °C for 60 minutes, and inactivated at 85 °C for 5 minutes. Equal amounts of cDNA were applied for PCR amplification in triplicate using the Bio-Rad system and SYBR probes. Amplification was performed at 95 °C for 10 minutes, followed by 40 cycles at 95 °C for 10 seconds and 60 °C for 30 seconds. Quantitative values were obtained from the quantification cycle (C q ). Each sample was normalized on the basis of its gene content (∆C q ). The formula 2 −(∆∆Cq) was used to analyze the results. Bio-Rad CFX Manager 3.1 software was used for RT-PCR data analysis Protein Extraction and Western Blot Analysis. Protein was isolated from confluent cells grown on 35 mm dishes by washing in PBS at 4 °C and adding lysis buffer (50 mM Tris-HCl [pH 8], 150 mM NaCl, 0.02% N3Na, 100 μg/mL phenylmethylsulfonyl fluoride, 1% NP-40, 50 mM NaF, 2 mM EDTA, and protease inhibitor [Sigma, Ann Arbor, MI]). Cell lysates were collected and Western blotting was performed as previously described 7 . Blots were labeled with CYP24A1 or CYP27B1 antibody at a dilution of 1:1000. Membranes were washed and then incubated with anti-rabbit horseradish peroxidase-conjugated secondary antibody (1:3000). Detection was performed using the enhanced chemiluminescence method (Pierce Biotechnology, Rockford, IL). To ensure equal protein loading, nitrocellulose membranes were labeled with GAPDH antibody at a dilution of 1:6000 (Cell Signaling Technology, Danvers, MA). Both X-ray film and the ChemiDoc ™ XRS + Imaging Systems (BIO-RAD, Hercules, CA) were used for chemiluminescence detection. For ChemiDoc ™ blots, stain-free gels were used. While GAPDH was always run on the same gel as the experimental protein, GAPDH required substantially lower exposure times on the stain-free gels, thus the GAPDH section of the gels were cut out from the full gel and imaged separately. For this reason, all stain-free gel results show separated GAPDH sections.
Statistical Analysis. All data are given as the mean ± SD of at least three experiments. Where applicable, differences between two groups were compared using the unpaired Student's t-test. For multigroup comparisons, ANOVA followed by Student-Newman-Keuls test were performed. P < 0.05 was considered statistically significant.
Data Availability. The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request.
Effects of VDR knockout. Compared to wild type mice, there was a significiant increase in CYP24A1 and CYP27B1 mRNA expression in fresh corneal epithelial cells pooled from VDR HET and KO mice (Fig. 7a). However, CYP24A1 and CYP27B1 protein expression was significantly lower in VDR KO mice (Fig. 7b,c). Protein levels were unaffected in cells from HET mice. To further explore the role of VDR in the vitamin D enzyme feedback loop, primary corneal epithelial cells from VDR knockout mice were cultured and exposed to 1,25(OH) 2 D3 and 24R,25(OH) 2 D3. Both 1,25(OH) 2 D3 and 24R,25(OH) 2 D3 treatment resulted in significantly increased CYP24A1 and CYP27B1 protein expression (Fig. 8). Figures S5-S9.

Discussion
The corneal epithelium has unique structural and physiologic properties that allow it to fulfill its functions in the eye. Using mass spectroscopy, our group previously examined vitamin D metabolites present in the ocular fluids of New Zealand white rabbits and also examined the influence of dietary vitamin D supplementation on these fluids. We detected relatively high concentrations of the metabolite 1,25(OH) 2    following Vit D supplementation. Furthermore, our lab previously reported vitamin D enhancement of tight junction resistance and gap junction cell to cell diffusion coefficients 7 . We also demonstrated attenuated corneal epithelial wound healing in 10-week-old VDR knockout mice 34 , which is similar to the effects of deletion of VDR deletion in mouse skin which resulted in retarding wound healing 35,36 . Deletion of VDR has also been shown to reduce β-catenin transcriptional activity and proliferation of cells at the leading edge of wound closure, and vitamin D was found to be required for a normal regenerative response of the skin to wounding 37,38 . Vitamin D has been found to have anti-inflammatory and immunoregulatory roles, and deficiency could result in dry eye disease 39,40 . The relationship of vitamin D to other ocular pathologies has also been studied, and the therapeutic potential of vitamin D has been examined 41 . In our current study, both 1,25(OH) 2 D3 and 24R,25(OH) 2 D3 significantly increased primary human epithelial cell migration. Therefore, anterior segment hypovitaminosis D could be detrimental after corneal injury, where vitamin D3 appears to play a role in the cell processes that occur during the corneal epithelial wound healing process as the cells migrate to cover the wound and differentiate into a multi-layered, mature corneal epithelium. Vitamin D has been shown to influence cell differentiation and proliferation 6,42 . Most of the literature describes an inhibitory effect of 1α,25(OH) 2 D3 on cell proliferation in various cell types, often associated with stimulation of cell differentiation. Our current study found that while both 1α,25(OH) 2 D3 and 24,25(OH) 2 D3 stimulated migration, 50 nM 1,25(OH)2D3 inhibited HCEC proliferation (Fig. 1A). In agreement with this, Reins et al. recently found a small but significant inhibitory effect of topically applied 100 nM 1α,25(OH) 2 D3 on mouse corneal epithelial wound healing 43 . It is possible that while proliferation is inhibited, differentiation may be stimulated. Interestingly, we found that both 1α,25(OH) 2 D3 and 24,25(OH) 2 D3 significantly increase epithelial proliferation in VDR KO MPCEC. Taken together with the data in Fig. 2 showing that 1α,25(OH) 2 D3 inhibits proliferation in HCEC, the data indicates that that VDR stimulation is likely inhibitory, while 1α,25(OH) 2 D3 and 24,25(OH) 2 D3 can act through a non-VDR mediated pathway to stimualte proliferation.
It is generally believed that 24,25(OH) 2 D3 is an inactive form of vitamin D3. This belief has been changing as of late, in that 24,25(OH) 2 D3 has been shown to effect bone mass and the differentiation and maturation of growth plate chondrocytes in vitro [44][45][46] . Moreover, a recent study of multiple sclerosis patients demonstrated that the ratio of serum 25(OH)D3 to 24R,25(OH) 2 D3 and not the ratio of 25(OH)D3 to 1,25(OH)2D3 correlated with higher disability and increased disease progression and brain atrophy 47 . In addition, 24R,25(OH) 2 D3 has been shown to increase bone healing and fracture healing 23 . 24R,25(OH) 2 D3 was also demonstrated to be important during human mesenchymal stem cell maturation, and has an essential role in Ca 2+ mineralization, gene expression, and the regulation of cytochrome P450 expression 24 . There have been no previous studies examining the possible function(s) of 24R,25(OH) 2 D3 in the cornea. The current study demonstrates that 24R,25(OH) 2 D3 stimulates both HCEC cell proliferation and migration. As both of these actions are crucial for corneal epithelial wound healing, these results provide evidence that 24R,25(OH) 2 D3 is likely involved in, and beneficial for corneal wound healing, as it is in bone and fracture healing.
CYP27B1 is the only enzyme known to catalyze 1α-hydroxylation of 25(OH)D3, and CYP24A1 is the only known enzyme capable of catalyzing 24-hydroxylation of 25(OH)D 48 . The action of CYP24A1 on 25-hydroxyvitamin D results in decreased 1,25(OH) 2 D3 levels and increased 24R,25(OH) 2 D3, while CYP27B1 increases 1,25(OH) 2 D3 levels 49 . In our studies, the CYP24A1 gene and protein expression level are highly inducible by 1,25(OH) 2 D3 in HCEC and primary mouse corneal primary epithelial cells, as in several other tissues, thus acting as a local control mechanism to prevent tissue-level 1,25(OH) 2 D3 intoxication 44,50 . In addition, our results indicate that stimulation of CYP24A1and the resultant increase in 24R,25(OH) 2 D3 levels will likely lead to stimulation of epithelial cell migration and proliferation as described above. Reins et al. recently described that the expression of CYP24A1 mRNA in HCEC and human primary corneal epithelial cells was increased with 1,25(OH) 2 D3 or 25(OH)D3 through a functional VDR, as silencing of the receptor blocked this response 9 . While we did not examine the influence of 1,25(OH) 2 D3 or 25(OH)D3 on CYP24A1 mRNA, we did find that CYP24A1 and CYP27B1 mRNA and protein were increased in VDR-silenced HCEC cultured with 24R,25(OH) 2 D3. Interestingly, CYP24A1 protein expression was not affected by 1,25(OH) 2 D3 or 24R,25(OH) 2 D3 in VDR-silenced HCEC. This would be a not so uncommon example of protein expression not mimicking mRNA expression. Previous studies have demonstrated that while 1,25(OH) 2 D3-dependent regulation of DNA synthesis in cartilage cells requires VDR 18,51,52 , other physiological responses to 1,25(OH) 2 D3 involve regulation via the Pdia3 membrane receptor (also called protein disulfide isomerase, ERp57,GR58, and 1,25D3-MARRS) [53][54][55] .
The current study also used cells obtained and cultured from VDR KO mice to further explore the role of VDR in corneal epithelial cell function. Unlike the VDR silencing studies, 1,25(OH) 2 D3 and 24R,25(OH) 2 D3 were both found to promote CYP24A1 and CYP27B1 protein expression in VDR KO MPCEC. The differing protein CYP24A1 and CYP27B1 expression levels in VDR silencing versus knockout studies may be a species specific result or a side-effect of either the gene knockout or silencing procedures. The differences are unlikely due to incomplete silencing of VDR in the siRNA study as gene silencing typically results in continued (although reduced) gene function in silenced cells versus total loss of function as in KO cells, which is not the response found in this study. Interestingly, significantly greater cell growth was observed in VDR KO mouse cells exposed to 1,25(OH)2D3 or 24R,25(OH)2D3 than was observed for WT mouse corneal epithelial cells (Fig. 2). This data indicates that there is a Vit D signaling pathway in corneal epithelium that is distinct from the traditional VDR pathway, and that VDR may be a negative regulator of corneal epithelial cell growth.
We report, for the first time, 24R,25(OH)2D3-initiated feedback control of the key vitamin D enzymes CYP24A1 and CYP27B1 through VDR-independent signaling in the corneal epithelium. A 24R,25(OH) 2 D 3 positive feedback loop has been reported in osteoblasts, where 24R,25(OH) 2 D3 markedly enhanced the CYP24A1 mRNA level while not effecting CYP27B1 mRNA, which may result in a higher production of 24R,25(OH) 2 D 3 25 . Our findings demonstrate that 24R,25(OH) 2 D3 significantly increases CYP24A1 and CYP27B1 expression in HCEC, VDR WT MPCEC, and VDR KO MPCEC. 1,25(OH)2D3 also had effects on CYP24A1 and CYP27B1 in VDR KO MPCEC. Additional functions of 24R,25(OH) 2 D3 in Vitamin D metabolism and catabolism need to be further exploreed in corneal epithelial cells.