1. Introduction
Knowledge regarding the pathogenesis of the ocular surface disease of vitamin A deficiency continues to evolve. In ancient times, Egyptian physicians found that treatment of night blindness with liver from an ass or ox improved this condition, and the nutritional component in the liver responsible for this was eventually identified as vitamin A. In 1816, François Magendie, a pioneering physiologist, found that dogs fed a diet containing sugar in distilled water without animal products lost weight, developed corneal ulcers, and subsequently died [
1]. Charles-Michel Billard, a French pediatrician, reported in 1828 that children with complete marasmus develop softening of the cornea, with an appearance similar to what was observed in the malnourished dogs [
1]. Lunin (1881) observed that mice had a greater survival rate of when whole dried milk was added to their diet [
2]. In the early 1900s, E. McCollum at the University of Wisconsin and Osborne and Mendel at Yale independently discovered that young rats fed a basic diet consisting of carbohydrate and protein sources had better growth and survival when their diet was supplemented with butter fats or extract either of butter fats, egg yolk, and/or liver. The unknown factor in these supplements was called “fat-soluble factor A”. Apart from defective growth and survival, the young rats fed the basic diet developed eye disease, which could be alleviated by adding butter fats [
1,
2]. Thus, by 1917, it was experimentally established that lack of a fat-soluble factor in the rat diet caused weight loss and xerophthalmia that could result in blindness. Subsequently, in 1931, the swiss chemist Paul Karrer extracted vitamin A from cod-liver-oil and described its chemical structure, and vitamin A was eventually chemically synthesized by Isler and his group in 1947 [
2].
It is now known that vitamin A is a group of nutritional unsaturated organic compounds that consist of an unsubstituted β-ionone ring and isoprenoid chain. These include alcohol (retinol), aldehyde (retinal) and ester (retinyl ester) forms, retinoic acid, and several provitamin A carotenoids—mostly notably the β-carotene.
Humans ingest vitamin A either as retinyl ester found in animal sources, including eggs, liver, fish, and dairy products, or in the carotenoid forms, α-carotene, β-carotene, and β-cryptoxanthin, from plants. During intestinal digestion, retinyl ester and provitamin A carotenoids are partially or completely hydrolyzed into free retinol, which enters enterocytes either by simple diffusion or a carrier-mediated process [
3]. Once inside enterocytes, retinol is re-esterified with long chain fatty acids, such as palmitate, and incorporated along with other dietary lipids into triacylglycerol-rich chylomicrons that are released into the lymphatics, while free retinol is also released into the portal circulation [
4]. In the blood, the triacylglycerol in chylomicrons is hydrolyzed by lipoprotein lipase, resulting in the production of “chylomicron remnants”, which are exclusively taken up by hepatocytes and stored predominantly as retinyl ester in the perisinusoidal stellate cells. Approximately 70% of the vitamin A in the body is stored in these hepatic stellate cells [
5]. Prior to release from the liver, retinyl ester is hydrolyzed into free retinol and complexed with retinol binding protein (RBP) [
6]. Plasma RBP is the high affinity specific retinol carrier protein in the blood that stabilizes retinol and delivers it to the target tissues [
7]. RBP, either through a cell-specific receptor or facilitated diffusion, delivers retinol to the lacrimal gland acinar cells, where it is stored as fatty acyl esters (retinyl linoleate, retinyl palmitate, retinyl stearate) [
8,
9,
10]. Retinol bound to RBP is secreted by the lacrimal gland into the tears [
10]. Retinol concentration in secreted lacrimal fluid was reported to fall to non-detectable levels in vitamin A deficient rabbits [
10]. Cellular and biochemical changes were observed in lacrimal gland acini of vitamin A deficient rats, including decreased secretory granules, endoplasmic reticulum atrophy, and reduction of histochemically detectable glycoproteins [
11].
These findings indicate that vitamin A is essential for maintaining a lubricated uninflamed ocular surface and clear cornea. While the vitamin A deficient phenotype is well characterized, the biologic activity of retinoids on the ocular surface is incompletely understood. The objective of this study was to investigate the lacrimal ocular surface retinoid axis, particularly the immunomodulatory effects of retinoic acid (RA) and change in conjunctival myeloid cell number and phenotype in VAD.
4. Materials and Methods
4.1. Animals and Vitamin A Deficiency Protocol
The animal protocol for this study was designed according to the Association for Research in Vision and Ophthalmology (ARVO) Statement for the use of Animals in Ophthalmic and Vision Research and was approved by the Institutional Animal Care and Use Committees at Baylor College of Medicine (AN2032, approved 30 July 2019). Female C57BL/6J (B6) mice aged 6–8 weeks were purchased from Jackson Laboratories (Bar Harbor, ME, USA) and allowed to rest in a humidified environment for a week before the experiment. Vitamin A deficiency was induced with a standard protocol [
22]. Briefly, beginning at 2 weeks of gestation, all pregnant females were fed a vitamin A-deficient green colored diet (58M1/NO Vit. A green); after birth, the dams either stayed on the A-deficient diet (green color) or were placed on a control yellow colored vitamin A supplemented diet (58m1 w/yellow dye). RXRα mutant (I273N)
Pinkie mice were bred internally and maintained in a specific-pathogen-free vivarium.
4.2. Measurement of Goblet Cell Density
Goblet cell density was measured as previously described [
23]; briefly, following euthanasia, eyes were excised (
n = 9/group) and fixed in 10% formalin. Paraffin embedded sections, 5 µm thick, were cut with a microtome (Microm HM 340E; Thermofisher Wilmington, DE, USA). Sections were stained with periodic acid Schiff (PAS) reagent and were examined and photographed with a microscope (Eclipse E400; Nikon, Garden City, NY, USA) equipped with a digital camera (DXM1200; Nikon). Using the NIS Elements software, goblet cells were manually counted. The length of the conjunctival goblet cell zone was measured by drawing a digital line on the surface of the conjunctiva image from the first to the last PAS
+ goblet cell. The results are presented as PAS
+ goblet cells/mm.
4.3. Immunofluorescent Staining
Immunofluorescent staining was performed to detect STRA6 (Cat# 22001-1-AP, 1:100 dilution, Proteintech Group, Rosemont, IL, USA) and RBP4 (Cat# 11774-1-AP, 1:100 dilution, Proteintech Group, Rosemont, IL, USA) on the ocular surface of B6 mice. Optimal cutting temperature (OCT) embedded frozen eye tissue sections were methanol-fixed for 5 min, washed, and permeabilized with 0.3% Triton X-100 in phosphate buffered saline (PBS). Nonspecific binding sites were blocked with 20% goat serum for 1 h and incubated with primary antibodies overnight and secondary goat anti-rabbit Alexa-Fluor 488/555 conjugated IgG antibodies for 1 h. Images were captured with a confocal microscope (Nikon AR-1, Garden City, NY, USA).
4.4. Real-Time PCR
Following euthanasia, the corneal epithelium was scraped, and total RNA was extracted using a QIAGEN RNeasy Plus Micro RNA isolation kit (Qiagen, Valencia, CA, USA) according to the manufacturer’s instruction. The cDNA was synthesized using the Ready-To-Go−You-Prime-First-Strand kit (GE Healthcare, Pittsburgh, PA, USA). Quantitative real-time PCR was performed with specific Taqman probes (Life Technologies, Grand Island, NY, USA) for STRA6 (Mm00486457_m1) and RBP-4 (Mm00803264_g1). The HPRT-1 gene was used as an endogenous reference for each reaction. The results of real-time PCR were analyzed by the comparative CT method.
4.5. Western Blot
Corneal epithelial cells from B6 and vitamin A deficient mouse were scrapped and placed in cell lysis buffer (Part no. 895347, R&D Minneapolis, MN, USA). Protein concentration was determined using a micro bicinchronic acid (BCA) protein assay (Cat# 23235, Thermo Fisher, Waltham, MA, USA). Here, 50 µg of corneal extract was resuspended in sodium-dodecyl sulphate (SDS) sample buffer, boiled for 5 min, and analyzed on 4–15% mini-protean TGX™ stain-free gels (Cat# 4568084, Bio-Rad, Hercules, CA, USA). The proteins were transferred to polyvinylidene difluoride membranes (Bio-Rad, Cat# 170-4157). The blots were incubated with an anti-STRA6 (Cat# 22001-1-AP, 1:100 dilution, Proteintech Group, Rosemont, IL, USA) or an anti-β actin antibody (Sigma Aldrich, St. Louis, MO, USA, Cat# A5441) overnight. After secondary antibody incubation, the signals from antigen–antibodies complexes were developed with ECL plus Western Blotting Detection kit (Cat# RPN2106, ECL, GE Healthcare, Chicago, IL, USA). Images were taken using ChemiDoc Touch Imaging System (ChemiDoc Touch Imaging System; Bio-Rad, Hercules, CA, USA), and band densities were measured by Bio-rad software (Image lab, v. 6.0; Bio-Rad).
4.6. Flow Cytometry
Conjunctivae from B6, vitamin A deficient, or Pinkie strains were excised, chopped with scissors into tiny pieces, and incubated with 0.1% type IV Collagenase for 1 h to yield single cell suspensions. Samples were incubated with anti-CD16/32 for 5 min at room temperature and subsequently stained with CD45 (clone 30-F11, Biolegend, San Diego, CA, USA), MHC II (clone I-A/I-E, BD Pharmingen; San Diego, CA, USA), CD11b (clone M1/70, Thermofisher Scientific, Waltham, MA, USA). Cells were stained with an infra-red fluorescent viability dye (Life Technologies, Grand Island, NY, USA). The gating strategy was as follows: lymphocytes were identified by forward -scatter area (FSC-A) and side scatter area (SSC-A) gates, followed by two singlets gates (FSC height vs. FSC-A and SSC-height vs. SSC-A), followed by live/dead identification using the infra-red fluorescent viability dye. Alive CD45+ cells were plotted for MHC II and CD11b expression. Negative controls consisted of fluorescence minus one splenocytes. Cells were acquired with either BD LSR II or BD Canto II Benchtop cytometers with BD Diva software v. 6.7 (BD Biosciences, San Diego, CA, USA), and 200,000 or more events were collected. Final data were analyzed using FlowJo software version 10 (Tree Star Inc., Ashland, OR, USA).
4.7. Measurement of NFκB p65 Activation
Corneal epithelium was scrapped using a dulled blade and nuclear protein extraction was performed according to the manufacturer’s instructions. NFκB p65 activation was measured by a TransAM NFκB p65 kit that specifically quantifies phosphorylated NFκB p65 (Cat# 40596, Active Motif, Carlsbad, CA, USA). Nuclear extracts from B6 mice maintained in a normal environment and from B6 mice with desiccating stress induced dry eye were added to wells of a 96-well plate with immobilized oligonucleotide containing an NFκB consensus binding site. The activated p65 in the nuclear extract binds to the oligonucleotide. After incubation with specific anti-p-p65 antibodies, horseradish peroxidase (HRP) conjugated secondary antibodies provided a sensitive colorimetric readout at 450 nm using a colorimetric plate reader (Tecan Infinite M200, Magellan v. 6.55 software; Tecan, Männedorf, Switzerland).
4.8. RNA Seq Data Analysis
Conjunctival epithelium was excised from B6 and Pinkie strains and total RNA was extracted using a QIAGEN RNeasy Plus Micro RNA isolation kit (Qiagen, Valencia, CA, USA) according to the manufacturer’s instructions. The concentration and purity of RNA were assessed using a NanoDrop 1000 (ThermoFisher Scientific, Waltham, MA, USA). RNA-Seq was performed by the Beijing Genomics Institute (BGI) using the BGISEQ500RS to generate 100 bp paired-end reads. The raw data were cleaned by removing reads containing adapter or poly-N sequences, and reads of low quality using SOAPnuke (v. 1.5.2, parameters: −l 15 −q 0.2 −n 0.05), and the expression levels of the resulting genes and transcripts were determined using RSEM (v. 2.2.5, default parameters). Detection of DEGs (differentially expressed genes) was performed with DEseq2 (parameters: fold change ≥ 2.00 and adjusted p-value ≤ 0.05). A total of 19,511 genes were obtained as raw data. Genes were passed through the Benjamini–Hochberg procedure to obtain the critical value for false discovery and a total of 1375 genes passed with a p-value > 0.0006. The selected genes were clustered in a heat map based on GSCA pathways.
4.9. Statistical Analysis
Based on normality, parametric student T or nonparametric Mann–Whitney U tests were performed for statistical comparisons with an alpha of 0.05 using GraphPad Prism 9.0 software (GraphPad Software, Inc., San Diego, CA, USA).