Ferroptosis in the Lacrimal Gland Is Involved in Dry Eye Syndrome Induced by Corneal Nerve Severing

Abstract

Purpose: Dry eye syndrome (DES) is a prevalent postoperative complication after myopic corneal refractive surgeries and the main cause of postoperative dissatisfaction. Although great efforts have been made in recent decades, the molecular mechanism of postoperative DES remains poorly understood. Here, we used a series of bioinformatics approaches and experimental methods to investigate the potential mechanism involved in postoperative DES.

Methods: BALB/c mice were randomly divided into sham, unilateral corneal nerve cutting (UCNV) + saline, UCNV + vasoactive intestinal peptide (VIP), and UCNV + ferrostatin-1 (Fer-1, inhibitor of ferroptosis) groups. Corneal lissamine green dye and tear volume were measured before and two weeks after the surgery in all groups. Lacrimal glands were collected for secretory function testing, RNA sequencing, ferroptosis verification, and inflammatory factor detection.

Results: UCNV significantly induced bilateral decreases in tear secretion. Inhibition of the maturation and release of secretory vesicles was observed in bilateral lacrimal glands. More importantly, UCNV induced ferroptosis in bilateral lacrimal glands. Furthermore, UCNV significantly decreased VIP, a neural transmitter, in bilateral lacrimal glands, which increased Hif1a, the dominant transcription factor of transferrin receptor protein 1 (TfR1). Supplementary VIP inhibited ferroptosis, which decreased the inflammatory reaction and promoted the maturation and release of secretory vesicles. Supplementary VIP and Fer-1 improved tear secretion.

Conclusions: Our data suggest a novel mechanism by which UCNV induces bilateral ferroptosis through the VIP/Hif1a/TfR1 pathway, which might be a promising therapeutic target for DES-induced by corneal refractive surgeries.

Vision correcting refractive surgeries are becoming more popular globally, with annual surgical case numbers reaching over 7 million. Dry eye syndrome (DES) is a common postoperative complication after myopic corneal refractive surgeries, affecting 20% to 55% of those receiving surgery.¹ This leads to a greater burden on visual function, activities of daily living, professional work, and quality of life.^2,3 The clinical characteristics of DES mainly present as a decrease in tear breakup time, irritation, and ocular fatigue in almost all cases after surgery.⁴ These symptoms, in some cases, last six months or more. These patients have to use artificial eye drops for a long time, sometimes for over ten years.⁵ DES not only impairs the patient's functional visual acuity⁶ but may also induce epithelial hyperplasia and stromal remodeling. Both side effects may induce refractive regression.⁷ Hence, elucidating the molecular mechanism of DES induced by refractive surgery is very important. 

Currently, a growing body of literature reports that corneal nerve severing plays a major role in post refractive surgery dry eye. For example, the SMILE procedure performed manually via a 3-mm vertical side cut, which has less impact on corneal nerves, induces less postoperative dry eye than LASIK surgery.^8,9 Flap creation in LASIK surgery causes more damage to corneal innervation, which greatly decreases corneal sensitivity.⁴ Corneal sensitivity did not recover two years after surgery in some patients.¹⁰ Regarding its mechanism, some studies have suggested that corneal denervation may affect the corneal-lacrimal gland, corneal-blinking, and blinking-meibomian gland reflexes and decrease tear secretion.^11–13 The blink frequency is decreased, and the exposure time of the ocular surface is increased, which can also accelerate the evaporation of tears.¹⁴ The homeostasis of tear film was also damaged.^4,15 However, these studies only observed the phenotype and did not reveal the underlying mechanism involved in DES induced by refractive surgery. 

Recently, an intriguing study by Lee et al.¹⁶ demonstrated that unilateral corneal nerve severing not only diminished tear secretion of the surgical eye but also disrupted secretion of the contralateral eye. Moreover, they observed that immune cells increased in the bilateral cornea/conjunctiva on Day 14 after severing. Unilateral corneal nerve severing could alter the neuropeptide level of the lacrimal gland. Based on these findings, they postulated that unilateral corneal nerve damage might disrupt bilateral immune homeostasis and contribute to the development of DES. Nevertheless, the precise mechanisms by which corneal nerve severing impairs tear secretion in the lacrimal glands remain unclear. The underlying key molecular mechanism of DES induced by corneal denervation has yet to be addressed. 

Thus, to elucidate the molecular mechanisms of postoperative dry eye, we used a unilateral corneal nerve severing mouse model. By sequencing the mRNA expression profile of lacrimal glands and a series of bioinformatics methods, we sought to analyze the possible underlying mechanisms. Moreover, we demonstrated this mechanism based on in vivo experiments, focusing on structural changes in lacrimal glands, ferroptosis, neural transmitters, and the release of secretory vesicles. This study will provide new insight into preventive strategies for DES induced by refractive surgery. 

Materials and Methods

Animals

All animal were obtained from the Ophthalmic Animal Laboratory, Zhongshan Ophthalmic Center, Sun Yat-Sun University (Guangzhou, China). Animal care was performed in accordance with the Institution Animal Care and Use Committee of Zhongshan Ophthalmic Center (Permit Number: SYXK [YUE] 2019-195) and the Association for Research in Vision and Ophthalmology Statement for the Use of Animals in Ophthalmic and Vision Research. Animals were kept on a 12-hour/12-hour light/dark cycle and had free access to food and water. Eight weeks old BALB/c female mice were randomly divided into sham, unilateral corneal nerve-cutting (UCNV) + Saline solution, UCNV + vasoactive intestinal peptide (VIP), and UCNV + Ferrostatin-1 (Fer-1) groups (n = 30 per group). 

Mice Corneal Nerve-Cutting

All surgeries were performed on eight-week-old female mice under anesthesia with the aid of an operating microscope. Mice were anesthetized by intraperitoneal injection of 1% pentobarbital sodium (P-010; Sigma-Aldrich Corp., St. Louis, MO, USA). To mimic refractive surgery, a circular punch (Kai Europe GmbH, Solingen, Germany) was used to sever the stromal nerves of the right eye cornea. The punch was applied to the corneal surface and then twisted five times with slight pressure until the stroma was incised.¹⁶ At the end of the surgery, a drop of tobramycin eyedrops was applied to the nerve severing eye. The right eyes and left eye of the corneal nerve-cutting group were defined as test (T) and contralateral (CL) eyes, respectively. The right eye was lightly pressed with a cotton swab in sham group. 

In Vivo Drug Treatment

Mice received daily intraperitoneal injections of VIP (GC45146; GLPBIO, 5 nmol in 200 µL sterile PBS) in VIP treat group. We chose the intraperitoneal injection method for administering VIP based on previous similar studies, taking both the drug efficacy and toxicity into consideration.^17–20 The dose of VIP used was also selected based on a comprehensive review of previous literature on VIP administration in mouse models.^19–24 And the injections of Fer-1 (S7243; Selleck Chemicals, Houston, TX, USA) 1 mg/kg in sterile DMSO was diluted with PBS. The concentration of DMSO injected in vivo did not exceed 1% in the Fer-1 treated group, and the control animal similarly received sterile PBS. Two weeks after injection, the animals were sacrificed for testing. 

Corneal Epithelial Damage Staining

The damage of the corneal epithelium was evaluated by lissamine green dye (3% Lissamine Green B; Sigma-Aldrich). Mice were anesthetized, and 5 µL of lissamine green dye was dripped on the cornea. After the mice blinked three times, excess drops were absorbed with a soft tissue. Photographs were taken with a digital camera connected to a stereoscopic microscope (SL-D7/DC-3/IMAGEnet). The extent of the staining was assessed by a masked observer using a grading scale as follows: 0 for no punctuate staining, 1 for less than one third, 2 for one third to two thirds, and 3 for more than two thirds.²⁵ This evaluation was performed before the surgery and the last day of the observation period. 

Tear Volume Measurement

Tear secretion of the mice was measured by the phenol red thread test. A phenol red–impregnated cotton thread (FCI Ophthalmics, Pembroke, MA, USA) was placed at the lower lateral canthus for 60 seconds.¹⁶ Tear volume was determined by measuring the length of the wet thread in millimeters. The evaluation was performed in the same period of day before the surgery and the last day of the observation period. 

Corneal Whole-Mount Staining

The mice were euthanized according to the Institution Animal Care and Use Committee of Zhongshan Ophthalmic Center. The corneas were separated from the anterior segments after enucleation and were fixed for 60 minutes in 4% paraformaldehyde and then rinsed in PBS. The tissues were blocked (10% normal goat serum and 0.5% Triton X-100) and permeabilized for 60 minutes at room temperature. The rabbit anti-β III tubulin antibody (1:400; Abcam, Cambridge, MA, USA) was used to identify the corneal nerve fibers. After 72 hours (4°C) the cornea tissues were rinsed in PBS and then incubated with Alexa Fluor 555-conjugated anti-rabbit secondary antibody. Four radial incisions were made towar the center of the cornea after washing with PBS, and the flattened corneas were observed with a fluorescence microscope (Axio Imager Z1; Carl Zeiss, Inc., White Plains, NY, USA). 

RNA Library Construction and Sequencing of Lacrimal Glands

The extraorbital lacrimal glands were surgical removed and total RNA was isolated and purified using Trizol reagent (Invitrogen, Carlsbad, CA, USA) following the manufacturer's procedure. The RNA amount and purity of each sample was quantified using NanoDrop ND-1000 (Thermo Fisher Scientific, Waltham, MA, USA). The RNA integrity was assessed by Agilent 2100 (Agilent, Santa Clara, CA, USA) with RIN number >7.0. Approximately 5 µg of total RNA was used to deplete ribosomal RNA according to the manuscript of the Ribo-Zero rRNA Removal Kit (Illumina, San Diego, CA, USA). After removing ribosomal RNAs, the remaining RNAs were fragmented into small pieces using divalent cations under high temperature. Then the cleaved RNA fragments were reverse-transcribed to create the cDNA, which were next used to synthesis U-labeled second-stranded DNAs with Escherichia coli DNA polymerase I, RNase H, and dUTP. An A-base is then added to the blunt ends of each strand, preparing them for ligation to the indexed adapters. Each adapter contains a T-base overhang for ligating the adapter to the A-tailed fragmented DNA. Single-or dual-index adapters are ligated to the fragments, and size selection was performed with AMPure XP beads (Beckman Coulter Life Sciences, Indianapolis, IN, USA). After the heat-labile UDG (Uracil-DNA glycosylase) enzyme treatment of the U-labeled second-stranded DNAs, the ligated products are amplified with PCR by the following conditions: initial denaturation at 95°C for three minutes; eight cycles of denaturation at 98°C for 15 seconds, annealing at 60°C for 15 seconds, and extension at 72°C for 30 seconds; and then final extension at 72°C for five minutes. The average insert size for the final cDNA library was 300 bp (±50 bp). At last, we performed the paired-end sequencing on an Illumina Novaseq 6000 (LC-Bio Technologies Co., Ltd., Hangzhou, China) following the vendor's recommended protocol. 

Sequencing Data Preprocessing

R language (version 4.0.3) was used to conduct the bioinformatics analysis. After conducting the normalization and background correction of data, the probe level data were converted into gene expression values.²⁶ The average expression value was taken as the gene expression value for multiple probes corresponding to a single gene. Principal component analysis observed the distribution patterns of trephined, contralateral, and control groups. The “DESeq2" package in R was used for pairwise comparison among three groups and obtain respective differentially expressed genes (DEGs). DEGs with the fold change ≥ 2 and P value < 0.05 were chosen for the follow-up analyses. 

Pathway Enrichment Analysis

The Kyoto Encyclopedia of Genes and Genomes database is an open access informatic source from Japan for interpreting biological function and characteristics of the organic system. The cluster Profiler package (version 3.16.0) was used to perform the pathway enrichment analysis and filter the cancer-related pathways to eliminate possible confusion.²⁷ P value < 0.05 was considered as statistically significant. 

Hematoxylin and Eosin Stain

The exorbital lacrimal gland is the main lacrimal gland and in an adult mouse is located subcutaneously approximately 3 mm to the temporal side of the eye, anterior to the ear.²⁸ To evaluate lacrimal gland tissue changes, extraorbital lacrimal glands were removed and fixed in formalin, then embedded in paraffin, and 5 µm sections were cut on a microtome. Tissue slides were dried overnight at ambient temperature and heated at 60°C for one hour. Slides were dewaxed, rehydrated, and stained with hematoxylin and eosin, then dehydrated through graded alcohols, cleared with xylene, and mounted using a resinous mounting medium. 

Western Blot Assay

Tissues were homogenized in RIPA buffer (Beyotime Institute of Biotechnology, Jiangsu, China) supplemented with a protease inhibitor cocktail. Debris was removed by spinning in a centrifuge at 4°C for 15 minutes at a speed of 12,000g and then resolved by SDS-PAGE and transferred to PVDF (polyvinylidene fluoride) membranes. The membrane was blocked with 5% BSA for one hour at room temperature and then probed with antibodies: rabbit anti-VIP (1:100; Abcam), rabbit anti-Rab3D (1:1000; Proteintech, Rosemont, IL, USA), rabbit anti-α-SMA (1:1000; Proteintech), rabbit anti-β-tubulin(1:5000; Cell Signaling Technology, Danvers, MA, USA), rabbit anti-TfR1 (1:1000; Abcam), rabbit anti-Hif1a (1:1000; Cell Signaling Technology, Danvers, MA, USA), rabbit anti-GPX4 (1:1000; Proteintech), rabbit anti-AQP5 (1:5000; Abcam), Armenian hamster anti-IL1β (1:500; Santa Cruz Biotechnology, Dallas, TX, USA), mouse anti-IL6 (1:500; Santa Cruz Biotechnology), and mouse anti-TNF-α (1:500; Santa Cruz Biotechnology) overnight at 4°C. On the second day, the membranes were washed three times and incubated with horseradish peroxidase–conjugated goat anti-rabbit (1:10000; Cell Signaling Technology) or horseradish peroxidase–conjugated goat anti-mouse (1:10000; Cell Signaling Technology) for one hour at room temperature (RT). After washing, the membranes were placed into ECL (Enhanced Chemiluminescence) reagent (WBKLS0100; Millipore, Burlington, MA, USA) and visualized using the ChemiDoc Touch Imaging System (Bio-Rad Life Science, Hercules, CA, USA). 

RNA Extraction and Quantitative Real‑Time PCR

Total RNA of the mice lacrimal gland tissue was extracted by Tri Reagent (T9424; Sigma-Aldrich) and subjected to reverse transcription using a PrimeScript RT Reagent kit (RR037B; Takara Biotechnology Co., Kyoto, Japan). The concentration of RNA was measured by a Nano Drop 1000 (Thermo Scientific). Real-time PCR was performed using the SYBR Green system (RR064B, Takara Biotechnology Co.) on a Roche 480 (Roche, Basel, Switzerland). The relative quantification of target gene was calculated using the ΔΔCt method²⁹ and normalized to the endogenous expression of β-actin. The reaction of each sample was run in triplicate. Sequences for the primers were as follows: Ptgs2-F 5ʹ-CTGCGCCTTTTCAAGGATGG-3ʹ, Ptgs2-R 5ʹ-GGGGATACACCTCTCCACCA-3ʹ; β-Actin-F 5ʹ-AGGTCATCACTATTGGCAACG-3ʹ, β-Actin-R 5ʹ-ACGGATGTCAACGTCACACTT-3ʹ. 

Immunofluorescence Assay

Lacrimal glands were embedded in OCT compound and frozen at −80°C after 4% paraformaldehyde fixation and the gradient of dehydration. Sections were immersed in 0.1% Triton X-100 (Sigma-Aldrich) for 10 minutes and blocked with 10% BSA at RT for one hour. The slides were incubated with primary antibody, rabbit anti-AQP5 (1:500; Abcam), rabbit anti-TfR1 (1:400; Abcam), rabbit anti-Hif1a (1:100; Cell Signaling Technology), mouse anti- Cytokeratin 19 (1:200; Santa Cruz Biotechnology), rabbit anti-α-SMA (1:500; Proteintech) and rabbit anti-4HNE (1:500; Abcam) overnight at 4°C. After three washes, slides were incubated with Alexa-Fluor 555 conjugated anti-rabbit antibody (1:500; Cell Signaling Technology) and Alexa-Fluor 488 conjugated anti-rabbit antibody (1:500; Cell Signaling Technology) at RT for one hour. The nuclei were stained with 4ʹ,6-diamidino-2-phenylindole (DAPI; 1 µg/mL; Sigma-Aldrich). Slides stained without primary antibodies were used as negative controls. Images were captured using fluorescence microscopy (Axio Imager Z1; Carl Zeiss). ImageJ software was used to quantify the fluorescence intensity of the staining. Four regions of interest (ROIs) were randomly selected in each image, and at least three different layers of each tissue were analyzed. The mean gray value of the ROI was measured, and the background mean gray value was subtracted. The fluorescence intensity to the area of the ROI was normalized before obtaining the relative fluorescence intensity. 

TUNEL Staining

A TUNEL detection kit (Roche) was used for the TUNEL assay according to the manufacturer's instructions. In brief, frozen sections of lacrimal gland were fixed with fixation solution for 20 minutes and blocked with 5% BSA containing 0.3% TritonX-100 for two minutes. Then the sections were incubated with TUNEL reaction mix at 37°C for one hour. Fluorescence was observed with an fluorescence microscope (Zeiss, Oberkochen, Germany). TUNEL-positive cells were quantified with ImageJ software: TUNEL-positive rate (%) = (TUNEL positive nucleus/Total nucleus) × 100%. 

Histological Detection of Labile Fe (II)

Labile iron (Fe2⁺) in lacrimal gland sections were detected using FeRhoNox-1 fluorescent imaging probe (Goryo Chemical, Sapporo, Japan) following the manufacturer's procedure. Frozen sections of lacrimal gland were air dried for three minutes, fixed in 4% paraformaldehyde for one minute, and washed in deionized water briefly. Then, RhoNox-1 was placed on the sections and incubated for 30 minutes at 37°C in a dark chamber. The sections were then counterstained with DAPI 1 µg/mL nuclear stain and observed using fluorescence microscopy (Axio Imager Z1; Carl Zeiss). Fluorescence intensity is used for statistics. Of each image, four regions of interest, which included the lacrimal gland acini, were selected randomly, and the fluorescence intensity was quantified using the ImageJ as descried above. 

Statistical Analyses

Images were analyzed with ZEN (Carl Zeiss) and ImageJ software. Continuous variables are presented as the mean ± SD (n = sample size), and mRNA data are presented as the mean with minimum to maximum range. Categorical variables are presented as the median with minimum to maximum range. All statistical analyses were carried out with Prism software package (GraphPad Prism, ver. 7.0; GraphPad Software Inc, San Diego, CA, USA), and IBM SPSS software (version 23.0; IBM, Armonk, NY, USA). One-way ANOVA followed by the Tukey's multiple comparisons test or Kruskal-Wallis test with Dunn's multiple comparison test or Kruskal-Wallis one-way ANOVA with all pairwise multiple comparisons were used for multiple comparisons. Statistical significance was defined at *P < 0.05. 

Results

Unilateral Corneal Nerve Severing Induced Bilateral DES and Decreased AQP5 in Lacrimal Gland

To mimic refractive surgeries in clinical treatment, one cornea was incised with a 2-mm circular punch. The corneal subbasal and stromal nerves were severed, and neuronal damage was assessed by staining with an antibody against TuJ-1, a specific marker of corneal neural fibers.³⁰ As shown in Figure 1A, the corneal nerves were severed (white arrows). To evaluate whether severing in the subbasal and stromal nerves induces clinical manifestations of DES in both eyes, we performed corneal staining at 14 days after the procedure, which is the most commonly used clinical indicator.³¹ Our data indicated that corneal staining grades were significantly increased in both the severed eye (T) and the contralateral eye (CL) compared with the sham eye (S) [S: median (1), min-max (0-1); T: median (3), min-max (2-3); CL: median (2), min-max (1-3); **P < 0.01, *** P < 0.001] (Figs. 1B, 1C), which may be related to an aqueous tear deficiency. Similarly, after assaying tear secretion, we found that tear volume in the severed eye (T) and the contralateral eye (CL) was also significantly decreased compared with that in the sham eye (S: 4.5 ± 1.3; T: 2.3 ± 1.2; CL: 2.6 ± 1.3; *P < 0.05) (Fig. 1D). However, there was no difference between the contralateral eye and the severing eye (Figs. 1C, 1D). 

Figure 1.

 

Bilateral corneal epithelial damage and reduction of tear volume after unilateral corneal nerve severing. (A) To sever the corneal subbasal and stromal nerves, the cornea was partially incised with a circular punch. Immunofluorescence of the corneal nerve staining demonstrated that the corneal nerve were severed. Scale bar: 50 µm. (B) Representative photographs of corneal epithelial staining with lissamine green dye at day 14 after unilateral corneal nerve severing. (C) Corneal grade staining with lissamine green (n = 9 per group). **P < 0.01, *** P < 0.001 by the Kruskal-Wallis one-way ANOVA. (D) Tear volume measured with the cotton thread method (n = 7 per group) at day 14. *P < 0.05 by ANOVA. (E) AQP5 immunoreactivities in extraorbital lacrimal glands. Scale bar: 400 µm. (F) AQP5 signals in extraorbital lacrimal glands detected by Western blot analysis and expressed as semiquantified results (n = 6 per group). *P < 0.05 by the Kruskal‒Wallis test.

To further investigate the alterations in the lacrimal gland, we evaluated the protein level of AQP5, a water channel protein, in the lacrimal gland. AQP5 has been reported to be expressed in the apical plasma membranes of the acinar cells in lacrimal gland.³² Although Moore et al.³² reported that AQP5 KO mice do not exhibit an altered tear volume, other studies have demonstrated that AQP5 plays a role in facilitating osmotic equilibration to maintain intracellular homeostasis.³³ Our data showed that AQP5 expression in the bilateral lacrimal glands was decreased compared with that in the sham lacrimal glands (Fig. 1E). Western blot assays also demonstrated that unilateral severing could significantly induce AQP5 decrease in bilateral lacrimal glands (S: 2.082 ± 0.485; T: 1.123 ± 0.216; CL: 1.102 ± 0.153; *P < 0.05) (Fig. 1F). However, no difference was observed in the bilateral lacrimal glands of the unilateral corneal severing group. Thus these results suggest that unilateral corneal nerve damage could lead to bilateral DES. 

A Decrease in Secretory Vesicles and α-SMA was Observed in the Bilateral Lacrimal Glands After Unilateral Corneal Nerve Severing

In the unilateral corneal nerve severing group, dramatic histopathological alterations of the bilateral lacrimal glands occurred. The H&E staining of the lacrimal gland showed that there were some increased luminal spaces within acinar cells bilaterally in the nerve severing group, displaying a granular appearance compared with the sham group. More vacuoles were observed in the bilateral lacrimal glands (red arrows) (Fig. 2A). Moreover, Rab3D, a specific marker of secretory vesicle maturation,³⁴ was significantly downregulated in the bilateral glands with unilaterally severed eyes, as shown by Western blotting. The relative intensities of the bands were quantified by densitometry and normalized to β-tubulin levels (S: 1.442 ± 0.252; T: 0.564 ± 0.185; CL: 0.563 ± 0.190; ****P < 0.0001) (Fig. 2B). In addition, myoepithelial cells, which surround the basal side of acinar cells, identified by the presence of α-SMA, was significantly decreased in bilateral lacrimal glands after corneal nerve cutting. Myoepithelial cells could squeeze the acinar cell and thereby expelling the secretory products into the duct system.^35,36 Figure 2C vividly shows its location and change by immunofluorescence staining. In the sham section, α-SMA basically localized around the basal regions and formed a basket-like network around the acinar unit (white ring). However, after corneal nerve severing for two weeks, α-SMA was decreased in ipsilateral (T) and contralateral (CL) gland tissues. High-magnification images showed dissociation of the basket-like expression pattern of α-SMA in bilateral glands after corneal nerve-severing treatment (Fig. 2C). Moreover, Western blotting analysis also confirmed the downregulation of α-SMA in bilateral lacrimal glands (S: 1.417 ± 0.244; T: 1.021 ± 0.287; CL: 0.993 ± 0.219; *P < 0.05, **P < 0.01) (Fig. 2D). Thus these results indicate that corneal nerve severing may affect the structure in bilateral lacrimal glands and inhibit the maturation and release of secretory vesicles, subsequently decreasing tear drops. 

Figure 2.

 

Pathological changes in lacrimal acinus and myoepithelial cells. (A) H&E staining of the lacrimal gland 14 days after unilateral corneal nerve severing. The red arrows indicate increased large vacuoles within acinar cells bilaterally in the nerve severing group. Scale bar: 20 µm and 5µm. (B) Rab3D expression in extraorbital lacrimal glands detected by Western blot analysis and expressed as semiquantified results (n = 6 per group). ****P < 0.0001 by ANOVA. (C, D) Immunofluorescence and Western blot analysis of α-SMA from extraorbital lacrimal glands and the semiquantified results. The white circle shows α-SMA basically localized around the basal regions and formed a basket-like network around the acinar unit. Scale bar: 40 µm and 5 µm (n = 10 per group). *P < 0.05, **P < 0.01 by the Kruskal‒Wallis test.

Ferroptosis Occurs in Bilateral Lacrimal Glands After Unilateral Corneal Nerve Transection

To facilitate the discovery of lacrimal gland-derived genes and networks of DES etiology, we performed RNA sequencing-based profiling of T and CL lacrimal glands after corneal nerve severing, while the sham operation group (S) served as a control. We first compared the transcriptomic profiles of the T and S groups. As shown in Figure 3A, the analysis identified 410 DEGs between the sham and nerve-severing ipsilateral lacrimal glands, with 268 genes being upregulated and 142 showing downregulation (Data S1). The transcriptome differences between the CL and S groups, including 214 upregulated and 119 downregulated genes, were slightly smaller than those between the T and S groups (Fig. 3B; Data S2). Comparing the transcriptome differences between bilateral lacrimal glands after corneal nerve severing, we found that the transcriptome differences between the two groups were much less significant than their respective differences in the S group, with only 76 DEGs (Fig. 3C; Data S3). Based on the above results, we aimed to further determine whether any commonalities existed between bilateral lacrimal gland alterations. Comparing the altered transcriptional profiles of lacrimal glands on bilateral sides, we found that many DEGs (217 genes) were common between the two groups (T vs. S and CL vs. S), as presented in the Venn diagram (Fig. 3D). These results indicate that many similar pathological processes may occur in the bilateral lacrimal glands. 

Figure 3.

 

Ferroptosis may be involved in bilateral lacrimal glands after unilateral corneal nerve transection. Volcano plots of DEG distribution between the T and S groups (A), between the CL and S groups (B), and between the T and CL groups (C). The red and blue dots represent upregulated and downregulated genes, respectively. (D) Venn diagram showing the overlap of DEGs in different subgroups. (E) Top three common pathways in different subgroups based on common pathway enrichment analysis. (F) Fe²⁺ in extraorbital lacrimal glands was detected using a FeRhoNox-1 fluorescence imaging probe. Scale bar: 400 µm. (G) Fluorescence intensity of Fe²⁺ measurements (n = 7 per group). **P < 0.01, *** P < 0.001 by ANOVA. (H) Ptgs2 mRNA levels in extraorbital lacrimal glands (n = 9-10 per group). **P < 0.01, ****P < 0.0001 by ANOVA.

To better appreciate the biological significance of altered transcriptional profiles from bilateral lacrimal glands and explore the biological processes that coexist in both lacrimal glands after trephining, we performed pathway analysis of DEGs between the T and S groups (represented as triangles) and between the CL and S groups (represented as circles), respectively (Fig. 3E). We observed that three of the top ten enrichment pathways analyzed in the two altered transcriptional profiles were identical, including “PI3K-AKT signaling pathway” (T vs. S: P value < 0.001; CL vs. S: P value < 0.001), “Ferroptosis” (T vs. S: P value < 0.001; CL vs. S: P value = 0.009), and “Mineral absorption” (T vs. S: P value = 0.013; CL vs. S: P value = 0.014). Notably, we found that the “PI3K-AKT signaling pathway” and “mineral absorption” were both associated with ferroptosis by literature mining.^37,38 

Ferroptosis is a nonapoptotic and oxidative damage-related cell death accompanied by a large amount of iron accumulation.³⁹ Consistently, as shown in Figure 3F, increasing Fe²⁺ iron accumulation was observed in the bilateral lacrimal glands after unilateral corneal nerve severing. The significantly altered rates of labile iron are presented in Figure 3G (S: 44.158 ± 9.511; T: 72.606 ± 15.261; CL: 75.026 ± 13.998; **P < 0.01, *** P < 0.001). Furthermore, to verify ferroptosis of the lacrimal gland, we evaluated the mRNA level of Ptgs2, a specific marker of ferroptosis.³⁸ As shown in Figure 3H, the relative quantification of Ptgs2 was significantly increased in the bilateral lacrimal glands after corneal nerve severing in mice compared with sham operation mice (S: mean [1.000], min-max [0.459–1.539]; T: mean [3.206], min-max [1.692–5.613]; CL: mean [2.060], min-max [1.177–3.246]; **P < 0.01, ****P < 0.0001). Taken together, these data indicate that ferroptosis occurs in bilateral lacrimal glands after unilateral corneal nerve transection. 

In addition, TfR1, which imports iron from the extracellular environment into cells and plays a dominant role in ferroptosis,⁴⁰ was also significantly increased in ipsilateral and contralateral lacrimal glands in unilateral corneal nerve severed mice based on the Western blot assay. The relative intensities of the bands were quantified by densitometry and normalized to β-tubulin levels (S: 0.413 ± 0.183; T: 0.726 ± 0.223; CL: 0.744 ± 0.257; *P < 0.05) (Fig. 4A). Moreover, Hif1a, the dominant transcription factor of TfR1,^41,42 was also significantly increased in the bilateral lacrimal glands (S: 1.175 ± 0.262; T: 1.618 ± 0.212; CL: 1.598 ± 0.319; *P < 0.05) (Fig. 4A). Double staining indicated that Krt19, a marker of ductal cells,⁴³ did not colocalize with TfR1 or Hif1a (Figs. 4B, 4C). Thus these results suggest that ferroptosis occurs in acinar cells, not ductal cells. 

Figure 4.

 

TfR1 and Hif1a levels in extraorbital lacrimal glands. (A) TfR1 and Hif1a signals in extraorbital lacrimal glands 14 days after unilateral corneal nerve severing detected by Western blot analysis and expressed as semiquantified results (n = 6–8 per group). *P < 0.05 by ANOVA. (B) Representative images of TfR1 and Krt19 double staining in the lacrimal gland. Scale bar: 200 µm. (C) Representative images of Hif1a and Krt19 double staining in the lacrimal gland. Scale bar: 50 µm.

VIP/Hif1a/TfR1 Is Involved in Ferroptosis Induced by Unilateral Corneal Nerve Severing in Lacrimal Glands

How does corneal nerve severing induce ferroptosis in the lacrimal gland? Neural damage mainly decreases neural transmitters. Moreover, previous studies have demonstrated that parasympathetic nerves supplying lacrimal gland secretory elements release the neuropeptide vasoactive intestinal peptide (VIP)⁴⁴ and VIP is an important regulator of tear production in humans, stimulating tear secretion via an AC/cAMP/PKA cascade.^45,46 Thus we measured VIP expression levels in lacrimal glands by Western blotting. Consistent with our speculation, our data also showed that the level of VIP in lacrimal glands was significantly decreased in the nerve severed group bilaterally compared with that of the sham group (S: 0.751 ± 0.340; T: 0.365 ± 0.245; CL: 0.339 ± 0.125; *P < 0.05) (Fig. 5A). To verify whether VIP supplementation could promote lacrimal gland secretion by regulating ferroptosis, daily intraperitoneal injection of VIP after unilateral corneal nerve severing was performed for two weeks. As shown in Figure 5B, the increased mRNA level of Ptgs2 in lacrimal glands, the specific markers of ferroptosis, was reversed by Supplementary VIP (S: mean [1.000], min-max [0.369–1.636]; T: mean [2.259], min-max [1.385–3.110]; CL: mean [1.961], min-max [1.018–3.018]; T + VIP: mean [1.300], min-max [0.372–1.786]; CL + VIP: mean [0.866], min-max [0.186–1.502]; *P < 0.05, **P < 0.01). Similarly, the labile iron (Fe²⁺) levels demonstrated that Supplementary VIP notably decreased Fe²⁺ in lacrimal gland cells (Fig. 5C). After measuring the fluorescence intensity, Supplementary VIP significantly decreased Fe²⁺ iron accumulation bilaterally in the VIP-treated group (S: 28.801 ± 10.066; T: 51.679 ± 14.863; CL: 53.387 ± 19.452; T+VIP: 31.271 ± 5.072; CL+VIP: 32.806 ±9 .418; *P < 0.05, **P < 0.01) (Fig. 5D). Accordingly, glutathione peroxidase-4 (GPX4), a glutathione-dependent peroxidase that alleviates ferroptosis,⁴⁷ was also significantly increased in lacrimal glands treated with Supplementary VIP (S: 0.668 ± 0.347; T: 0.331 ± 0.144; CL: 0.256 ± 0.111; T+VIP: 0.693 ± 0.297; CL+VIP: 0.746 ± 0.317; *P < 0.05, **P < 0.01, *** P < 0.001) (Fig. 5E). 

Figure 5.

 

The neural transmitter VIP alleviated ferroptosis caused by unilateral corneal nerve severing in lacrimal glands. (A) Western blot of VIP from extraorbital lacrimal glands 14 days after the procedure and the semiquantified results (n = 7 per group). *P < 0.05 by ANOVA. (B) Ptgs2 mRNA levels in extraorbital lacrimal glands (n = 7 per group). *P < 0.05, **P < 0.01 by ANOVA. (C) Fe²⁺ in extraorbital lacrimal glands was detected using a FeRhoNox-1 fluorescence imaging probe. Scale bar: 100 µm. (D) Fluorescence intensity of Fe²⁺ measurements (n = 7-8 per group). *P < 0.05, **P < 0.01 by ANOVA. (E) GPX4 signals in extraorbital lacrimal glands detected by Western blot analysis and expressed as semiquantified results (n = 11 per group). *P < 0.05, **P < 0.01, *** P < 0.001 by ANOVA. (F) TfR1 and Hif1a levels in extraorbital lacrimal glands detected by Western blot analysis and expressed as semiquantified results (n = 9–16 per group). *P < 0.05, **P < 0.01, ****P < 0.0001 by ANOVA.

Accordingly, Western blotting analysis showed that the protein expression levels of TfR1 and Hif1a in the VIP-treated group were significantly lower than those in the unilateral corneal nerve resection group, whereas there was no significant difference compared with the sham group (Hif1a: S, 0.597 ± 0.152; T, 0.958 ± 0.291; CL, 0.942 ± 0.210; T + VIP, 0.663 ± 0.162; CL + VIP, 0.646 ± 0.085; *P < 0.05, **P < 0.01; TfR1: S, 0.696 ± 0.221; T, 1.123 ± 0.290; CL, 0.953 ± 0.269; T + VIP, 0.719 ± 0.181; CL + VIP, 0.692 ± 0.187; *P < 0.05, ****P < 0.0001) (Fig. 5F). Thus these results indicate that VIP/Hif1a/TfR1 is involved in ferroptosis in acinar cells of lacrimal glands after corneal neural severing. 

The Neural Transmitter VIP Alleviated Lipid Peroxidation and Cell Death in Lacrimal Glands Caused by Unilateral Corneal Nerve Severing

Ferroptosis is a form of regulating cell death that is dependent on iron and characterized by lipid peroxidation. The 4-HNE is considered a common byproduct of lipid peroxidation,^48,49 and immunofluorescence staining of 4-HNE was used to indicate the level of lipid peroxidation in the lacrimal gland after VIP supplementation. As shown in Figures 6A and 6B, Supplementary VIP significantly decreased 4-HNE accumulation bilaterally in the VIP-treated group (S: 13.394 ± 2.312; T: 25.094 ± 5.421; CL: 20.360 ± 3.724; T + VIP: 16.668 ± 2.687; CL + VIP: 14.076 ± 2.248; *P < 0.05, **P < 0.01, ****P < 0.0001). Recently, many studies have applied TUNEL to detect cell death caused by ferroptosis.^50,51 We found that cell death rate was significantly decreased in lacrimal glands treated with Supplementary VIP (S: 0.043 ± 0.023; T: 0.212 ± 0.047; CL: 0.157 ± 0.039; T + VIP: 0.065 ± 0.027; CL + VIP: 0.026 ± 0.014; ****P < 0.0001) (Figs. 6C, 6D). Thus these results suggest that VIP rescues lacrimal acinar cells from ferroptosis caused by unilateral corneal nerve severing. 

Figure 6.

 

The neural transmitter VIP alleviated lipid peroxidation and cell death caused by ferroptosis. (A) The 4-HNE immunofluorescence staining in lacrimal glands. Scale bar: 400 µm, 100 µm. (B) Fluorescence intensity of 4-HNE measurements (n = 6-7 per group). *P < 0.05, **P < 0.01, ****P < 0.0001 by ANOVA. (C) The percentage of TUNEL-positive cells was determined with ImageJ (n = 6 per group). ****P < 0.0001 by ANOVA. (D) Representative fluorescence images from the TUNEL assay of lacrimal glands (blue, DAPI-stained nuclei; green, TUNEL-stained dead cells; red, Rab3D-stained acinar cells). Scale bar: 200 µm.

Supplementary VIP Alleviated the Inflammatory Reaction Induced by Ferroptosis and Promoted the Maturation of Secretory Vesicles in the Bilateral Lacrimal Glands

A previous study demonstrated that ferroptosis also induces serious inflammatory reactions in experimental models of certain diseases.⁵² Therefore we examined immune cells in lacrimal glands by HE staining. As shown in Figure 7A, more inflammatory cells were localized in the ipsilateral and contralateral lacrimal glands with unilateral corneal nerve severing (black rings), which was not observed in the sham and VIP treatment groups. Accordingly, the expression of the immune factors IL1 β, IL6, and TNF-α was increased in bilateral lacrimal glands with unilateral corneal nerve severing. In contrast, Supplementary VIP significantly decreased the expression of IL1β, IL6, and TNF-α (IL1 β: S, 0.728 ± 0.222; T, 1.078 ± 0.055; CL, 1.017 ± 0.071; T + VIP, 0.775 ± 0.092; CL + VIP, 0.683 ± 0.161; **P < 0.01. IL6: S, 0.664 ± 0.214; T, 0.991 ± 0.087; CL, 1.010 ± 0.093; T + VIP, 0.674 ± 0.218; CL + VIP, 0.619 ± 0.228; *P < 0.05, **P < 0.01; TNF-α: S, 0.785 ± 0.160; T, 1.244 ± 0.317; CL, 1.085 ± 0.329; T + VIP, 0.726 ± 0.185; CL + VIP, 0.676 ± 0.177; *P < 0.05, **P < 0.01) (Figs. 7B–D). 

Figure 7.

 

Contribution of Supplementary VIP to the inflammatory reactions induced by unilateral corneal nerve severing in lacrimal glands. (A) H&E staining of the lacrimal gland 14 days after unilateral corneal nerve severing. Infiltration of leukocytes in black circle. Scale bar: 400 µm. (B–D) IL1 β, IL6, and TNF-α levels in extraorbital lacrimal glands detected by Western blot analysis and expressed as semiquantified results (n = 6–8 per group). *P < 0.05, **P < 0.01 by ANOVA.

To further confirm the discoveries above, we assessed the fine structure of the acinar component by HE staining. There were some large vacuoles in the surgery group (red arrows), whereas these large vacuoles were alleviated in the VIP treatment group (Fig. 8A). Moreover, downregulation of Rab3D was significantly reversed in bilateral glands with severing in unilateral eyes with VIP treatment (S: 1.470 ± 0.288; T: 0.632 ± 0.243; CL: 0.673 ± 0.211; T + VIP: 1.543 ± 0.471; CL + VIP: 2.052 ± 0.763; *P < 0.05, *** P < 0.001) (Fig. 8B). According to the decrease in Rab3D, a specific protein of maturation secretory vesicles, we speculated that large vacuoles might be induced by the decrease in secretory vesicles. Thus these numerous small vacuoles might be related to the increase in mature secretory vesicles after VIP treatment. Previous studies also support this speculation. Bhattacharya et al.⁵³ proposed that vacuoles are associated with a significant reduction in the expression of lysosomes, mitochondria, and secretory vesicles in lacrimal glands in a dry eye mouse model. Moreover, Noh et al.⁵⁴ reported that APX‑115A, an NADPH oxidase inhibitor, inhibited the lacrimal gland from forming large vacuoles. Because VIP could also reduce oxidative stress in pancreatic acini through the inhibition of NADPH oxidase,⁵⁵ it is possible that vacuolar formation in lacrimal acinar cells is alleviated by supplementation with VIP. Accordingly, we also found a significant increase in AQP5 protein in the VIP injection group after corneal nerve severing compared with control mice (S: 1.767 ± 0.397; T: 0.822 ± 0.291; CL: 0.809 ± 0.205; T + VIP: 1.610 ± 0.378; CL + VIP: 1.699 ± 0.318; *** P < 0.001, ****P < 0.0001) (Fig. 8B). Thus these results indicate that neural transmitters could repair immune damage induced by ferroptosis, restore immune homeostasis, and promote the maturation of secretory vesicles in the bilateral lacrimal glands. 

Figure 8.

 

VIP supplementation induced histological and functional changes in extraorbital lacrimal glands. (A) H&E staining of the lacrimal gland 14 days after the procedure. The red arrows indicate increased large vacuoles within acinar cells bilaterally in the nerve severing group. Scale bar: 50 µm. (B) Rab3D and AQP5 levels in extraorbital lacrimal glands detected by Western blot analysis and expressed as semiquantified results (n = 6-7 per group). *P < 0.05, *** P < 0.001, ****P < 0.0001 by ANOVA.

Supplementary VIP and an Inhibitor of Ferroptosis Improved Bilateral Dry Eye Symptoms Induced by Unilateral Corneal Nerve Severing

Tear secretion and corneal lissamine staining were detected two weeks later. Consistently, our data showed that bilateral corneal staining grades were significantly decreased in the VIP-treated group (S: median [1], min-max [0-2]; T: median [3], min-max [1-3]; CL: median [2], min-max [1-3]; T + VIP: median [1], min-max [0-2]; CL + VIP: median [1)], min-max [0-2]; *P < 0.05, **P < 0.01) (Figs. 9A, 9B). Tear secretion was also significantly increased after VIP injection (S: 6.4 ± 1.2; T: 3.4 ± 0.6; CL: 4.3 ± 0.8; T + VIP: 5.0 ± 1.0; CL + VIP: 5.7 ± 1.1; *P < 0.05; **P < 0.01, ****P < 0.0001) (Fig. 9C). More importantly, Fer-1, an inhibitor of ferroptosis,³⁹ could rescue the ocular surface defects and decrease the death of corneal epithelial cells in a dry eye mouse model.⁵⁶ Our result showed the decreasing bilateral corneal staining grades (S: median [1], min-max [0-2]; T: median [3], min-max [2-3]; CL: median [3], min-max [2-3]; T+ Fer-1: median [1], min-max [1-2]; CL+ Fer-1: median [1], min-max [0-2]; *P < 0.05, **P < 0.01) (Figs. 9D, 9E) and increasing tear secretion after Fer-1 treatment. (S: 9.6 ± 1.0; T: 3.9 ± 0.6; CL: 5.1 ± 0.5; T + Fer-1: 8.2 ± 2.3; CL + Fer-1: 9.0 ± 2.7; *P < 0.05, ****P < 0.0001) (Fig. 9F). Thus these results indicate that VIP supplementation and the inhibitor of ferroptosis after corneal nerve severing could ameliorate bilateral DES induced by unilateral corneal nerve damage. 

Figure 9.

 

The effect of Supplementary VIP and Fer-1 on corneal epithelial damage and tear volume recovery. (A, B) Representative photographs of corneal epithelial staining with lissamine green dye and corneal grade staining after VIP supplementation for two weeks (n = 10 per group). *P < 0.05, **P < 0.01 by the Kruskal-Wallis one-way ANOVA. (C) Tear volume measured with the cotton thread method after VIP supplementation for two weeks (n = 12-13 per group). *P < 0.05, **P < 0.01, ****P < 0.0001 by ANOVA. (D, E) Representative photographs of corneal epithelial staining with lissamine green dye and corneal grade staining after Fer-1 treatment for two weeks (n = 7 per group). *P < 0.05, **P < 0.01 by the Kruskal-Wallis one-way ANOVA. (F) Tear volume measured with the cotton thread method after Fer-1 treatment for two weeks (n = 7 per group). *P < 0.05, **P < 0.01, *** P < 0.001, ****P < 0.0001 by the Kruskal-Wallis test.

Discussion

Globally, the number of patients undergoing myopic refractive surgeries is increasing annually. DES, one of the most common postoperative complications, has become a major problem for these patients. According to previous research, DES after refractive surgeries tend to be caused by insufficient aqueous tear production rather than tear-film instability.⁵⁷ The aqueous layer of the tear film is mainly secreted by the lacrimal gland.⁵⁸ Therefore we focused on alterations in the lacrimal gland after corneal nerve injury. 

The strength of this study demonstrates that unilateral corneal nerve severing induced ferroptosis in lacrimal acinar cells, resulting in bilateral hypofunction of lacrimal glands. To our knowledge, our study revealed a novel molecular mechanism of corneal nerve severing-induced DES, although great efforts have been made in recent decades. According to a previous study, although Lee et al.¹⁶ discovered that unilateral corneal nerve damage may lead to bilateral corneal nerve degeneration, alter immune homeostasis, and mechanistically participate in the development of bilateral inflammatory disorders such as dry eye, they did not investigate alternations in the lacrimal glands. For the effects of unilateral corneal nerve severing on the bilateral lacrimal gland, we speculate that the anatomical structure of the neural system between the cornea and lacrimal gland could account for it. The sensory nerve of the cornea emanates from the ophthalmic branch of the trigeminal nerve, and its functional integrity plays an important role in the stimulation of tear secretion and the regulation of the blink reflex. There is evidence that the trigeminal nucleus may project nerve fibers bilaterally, and axons pass between the bilateral dorsal horns via the dorsal commissure.^59,60 These anatomical studies demonstrate the presence of central nervous system circuits. Therefore retrograde nerve signals from unilateral corneal afferent nerve injury may lead to bilateral corneal nerve degeneration and changes in the secretory function of the lacrimal glands through bilateral nerve projections in the central nervous system. This phenomenon was observed in other unilateral ocular diseases, such as ocular herpes zoster and herpes simplex keratitis.^61,62 Patients with unilateral herpes zoster ophthalmicus demonstrated profound and significant bilateral loss of the corneal nerve plexus, demonstrating bilateral changes in a clinically unilateral disease.⁶¹ 

Ferroptosis is a form of regulated cell death that is dependent on iron and characterized by lipid peroxidation.⁶³ In the present study, bioinformatics analyses and in vivo experiments demonstrated that ferroptosis occurs in bilateral lacrimal glands after unilateral corneal nerve severing. Moreover, after treatment with the inhibitor of ferroptosis, Fer-1, an increase in tear secretion was observed bilaterally after corneal nerve severing. Although there is no study about sensory nerve loss that leads to the induction of ferroptosis in other tissues, previous studies have reported that ferroptosis plays a role in the progression of myocardial ischemia‒reperfusion injury, liver fibrosis, acute renal tubular necrosis, and neurodegeneration.^64–67 Raven et al.⁶⁸ demonstrated that iron levels in Alzheimer's disease were significantly elevated. Abnormalities in iron homeostasis in brain tissue can induce massive production of ROS in brain cells, ultimately causing catastrophic oxidative damage to sensitive subcellular structures.⁶⁹ Moreover, other studies indirectly supported this discovery. Similar alterations in nerves, neurotransmitters and structures of the lacrimal gland were observed in the TSP-1^−/− mouse model of aqueous deficiency dry eye.⁵³ Nguyen et al.⁷⁰ showed downregulation of genes associated with the endoplasmic reticulum and Golgi in the lacrimal gland after the loss of parasympathetic control of secretion. These results suggest that the reduction in neurotransmitters might affect the expression of intracellular organelles, which is related to the formation of vacuoles in lacrimal acinar cells. 

Moreover, our study demonstrated that the neural transmitter, VIP, might play a role in ferroptosis induced by corneal nerve severing in lacrimal acinar cells. Supplementary VIP significantly alleviated ferroptosis by decreasing iron accumulation and lipid peroxidation in the lacrimal gland. VIP, functioning primarily as a neurotransmitter, has various effects on almost all epithelial tissues such as the heart, lung, liver, pancreas, digestive tract, and kidney.⁷¹ For example, VIP could dampen formyl-peptide–induced ROS production and inflammation by targeting a MAPK-p47phox phosphorylation pathway in monocytes.⁷² VIP could also reduce oxidative stress in pancreatic acini through the inhibition of NADPH oxidase.⁵⁵ In addition, Proneth et al.⁷³ suggested that ferroptosis could trigger inflammation by releasing the specific necrotic signaling pathway of ferroptosis, such as damage-associated molecular patterns. Our data also demonstrated that Supplementary VIP inhibited the inflammatory reaction induced by ferroptosis in lacrimal glands caused by corneal neural severing. Moreover, some neural transmitters, or neurotrophic factors, could alleviate inflammation and improve DES.⁷⁴ For example, VIP was decreased in a rabbit model of keratoconjunctivitis sicca.⁷⁵ Pigment epithelial-derived factor peptide with a neuroprotective domain plus docosahexaenoic acid could induce corneal nerve regeneration after experimental surgery and decrease DES by regulating inflammatory cells.⁷⁶ The 3-isobutyl-1-methylxanthine, which can elevate cAMP-like VIP, stimulates tear secretion in patients with dry eye disease.⁷⁷ Of course, it is possible that other neural transmitters might be involved in ferroptosis induced by corneal nerve severing, which needs further investigation. 

Furthermore, our study demonstrated that ferroptosis was regulated by the VIP/Hif1a/TfR1 pathway in lacrimal acinar cells, which is also partially supported by previous studies. For example, Maugeri et al.^78,79 and D'Amico et al.⁸⁰ indicated that VIP could decrease the expression of Hif1a in models of diabetic macular edema, lung adenocarcinoma, and neuroblastoma. Other articles have demonstrated that the expression of TfR1 can be transcriptionally regulated by Hif1a. For example, Bianchi et al.⁴¹ demonstrated HIF-1-mediated activation of TfR1 transcription by iron chelation. Wang et al.⁸¹ demonstrated that hypoxia regulates ferrous iron uptake and reactive oxygen species levels by Hif1a. Thus our data reveal the molecular mechanism underlying corneal nerve severing–induced ferroptosis in lacrimal acinar cells. 

In conclusion, this study elucidated the underlying molecular mechanism of dry eye induced by corneal nerve severing, as shown in Figure 10. Unilateral corneal nerve severing impairs the afferent nerves of the tearing reflex loop and subsequently leads to a significant reduction in VIP secretion from the efferent nerve fiber terminals innervating the lacrimal gland bilaterally. The reduction in VIP may lead to an increase in Hif1a in lacrimal acinar cells. Hif1a transcriptionally upregulates TfR1, which induces Fe2+ accumulation and ferroptosis. Lacrimal gland damage caused by ferroptosis induces a decrease in Rab3D, α-SMA, and AQP5, resulting in hypofunction of lacrimal glands bilaterally. Thus this study reveals a novel mechanism in DES induced by myopic corneal refractive surgeries. Focusing on the inhibition of ferroptosis may be a promising therapeutic strategy for this disease. 

Figure 10.

 

Schema of proposed mechanisms of dry eye induced by corneal nerve severing. Unilateral corneal nerve severing impairs the afferent nerves of the tearing reflex loop and subsequently leads to a significant reduction in VIP secretion from the efferent nerve fiber terminals innervating the lacrimal gland bilaterally. The reduction in VIP may lead to an increase in Hif1a in lacrimal acinar cells. Hif1a transcriptionally upregulates TfR1, which induces Fe²⁺ accumulation and ferroptosis. Lacrimal gland damage caused by ferroptosis induces a decrease in Rab3D, α-SMA and AQP5, resulting in bilateral hypofunction of lacrimal glands.

Acknowledgments

The authors thank the staff of the public experimental platform of Zhongshan Ophthalmology Center for assistance in our experiments. We also thank Guangzhou Huaxiang Medical Biotechnology Co., Ltd., for bioinformatics support. 

Supported by the National Natural Science Foundation (grant nos. 81900850, 82101134) and Guangzhou Science Technology and Innovation Commission (grant no. 201803010091). 

Disclosure: X. Liu, None; Z. Cui, None; X. Chen, None; Y. Li, None; J. Qiu, None; Y. Huang, None; X. Wang, None; S. Chen, None; Q. Luo, None; P. Chen, None; J. Zhuang, None; K. Yu, None