Obstruction of the Tear Drainage Altered Lacrimal Gland Structure and Function

Abstract

Purpose: Orbital glands and drainage conduits are two distinct entities that constitute the lacrimal apparatus system, the malfunction of which leads to a range of ocular surface disorders. Despite the close functional relationship, how the two parts interact under pathophysiological conditions has not been directly tested. The study aims to investigate the lacrimal gland (LG) structural and functional changes upon the drainage system obstruction, thus, testing their function link.

Methods: Dacryocystectomy was performed in C57BL/6 mice to create a surgical model for tear duct (TD) obstruction (STDOB). Prickle1 mutant line with congenital nasolacrimal duct dysplasia serves as a genetic model for TD obstruction (GTDOB). Alterations of the LG and the ocular surface in tear duct obstruction mice were examined.

Results: STDOB and GTDOB mice showed similar ocular surface phenotypes, including epiphora, corneal epithelial defects, and conjunctival goblet cell abnormalities. At the molecular and cellular levels, aberrant secretory vesicle fusion of the LG acinar cells was observed with altered expression and localization of Rab3d, Vamp8, and Snap23, which function in membrane fusion. LG secretion was also altered in that lactoferrin, lipocalin2, and lysozyme expression were increased in both LG and tears. Furthermore, STDOB and GTDOB mice exhibited similar LG transcription profiles.

Conclusions: Physical obstruction of tear drainage in STDOB or GTDOB mice leads to LG dysfunction, suggesting a long-distance interaction between the tear drainage conduits and the LG. We propose that various components of the lacrimal apparatus should be considered an integral unit in diagnosing and treating ocular surface diseases.

The lacrimal apparatus is a complex system encompassing glands for tear production and sacs and ducts for tear drainage.¹ It plays an important role in maintaining tear film stability, corneal transparency, and ocular surface health.² The lacrimal gland (LG) is the main contributor to the aqueous layer of the tear film, which consists of water, electrolytes, and proteins.³ The drainage or tear duct (TD) system consists of upper and lower puncta, lacrimal canaliculi, lacrimal sacs, and nasolacrimal/tear ducts, providing a route for tears draining into the nasal cavity. Besides draining function, the TD regulates tear dynamics and maintains tear film homeostasis.^4–8 Stimulation of the nasal mucosa also leads to the reflex secretion of the LG.^9,10 

Normal tear quantity and constituents are maintained from the balance of tear secretion and drainage, which are crucial for ocular surface protection. Defective LG secretion is well known to be a major cause of dry eye disease.¹¹ Besides, a plethora of studies indicate that obstruction of the tear drainage also leads to reduced tear production. For example, a complete cessation of tear secretion was reported in rare ectodermal dysplasia of the Acro-Dermato-Ungual-Lacrimal-Tooth syndrome with congenital TD obstruction,¹² and reduced tear secretion is observed in the case of lower lid ectropion with obstructed tear outflow¹³ and in experimental subjects (normal individuals) with unilateral or bilateral punctal occlusion.^14,15 Consequently, punctal occlusion (punctual cautery/punctual plug), recommended as an option for treating moderate to severe dry eye disease,¹⁶ such as the Sjögren 's syndrome,^17,18 Stevens-Johnson syndrome,¹⁹ and the post-LASIK subjects, has no improvement in tear secretion, despite the reported improved symptom scores.²⁰ Furthermore, patients with pemphigus of the conjunctiva with blocked punctum also have poor tear production.¹³ 

The lacrimal sac is continuous with the nasolacrimal duct. Several ocular diseases, such as chronic dacryocystitis and lacrimal sac tumors, require surgical operations on the lacrimal sac, often causing ocular surface irritations similar to the punctual occlusion or lacrimal gland malfunction. Together with many other observations related above, they imply a functional link between the LG and the TD. However, this link has not been directly tested due to a lack of animal models. A rabbit surgical model was created previously only to investigate structural disorder and dysfunction of the tear duct but not the LG.²¹ Besides, the rabbit nasolacrimal anatomy differs greatly from the humans in that it has only a single punctum located in the lower eyelid,²² questioning its suitability for studying the human tear duct system. On the other hand, although mice are widely used in modeling human diseases, few studies detail the anatomy of the mouse lacrimal drainage system.^23,24 Recently, we described the mouse TD ontogenesis and anatomy, demonstrating its similarity to humans using a Prickle1 mutant mouse line.²⁵ This is, to our knowledge, the first genetic model having incompletely developed TD (GTDOB) by disruption of Prickle1 gene.^26–28 

In this study, we used two types of mouse models to address the functional link between LG and TD. We first created a surgical mouse model (surgical tear drainage obstruction [STDOB]) with the lacrimal sac removed to study the impact of obstructed tear drainage on LG. We then examined the cellular and structural alterations of LG in both STDOB and GTDOB mice and compared their LG phenotypes in parallel. The results of the two models point to the same conclusion that a long-distance interaction between LG and TD plays an essential role in ocular surface health. 

Materials and Methods

Animals

Animal husbandry and experimentation were conducted strictly according to the Association for Research in Vision and Ophthalmology. All experimental procedures were reviewed and approved by the Animal Care and Use Committee (ACUC) at Zhongshan Ophthalmic Center (protocol No. 2021-034). Female wild-type C57BL/6J mice aged 6 to 8 weeks (n = 50) were purchased from the Vital River Laboratory Animal Technology (Beijing, China). Prickle1 mutant mice (the Prickle1^a/b group, n = 10, aged 6–8 weeks) were generated by crossbreeding a Prickle1 gene-trap mutant allele (Prickle1^a/+) to a straight knockout allele (Prickle1^b/+).²⁹ Female adult wild-type C57BL/6 mice were randomly divided into three groups, one receiving bilateral dacryocystectomy (the STDOB group, n = 20), one receiving bilateral sham surgery (the sham group, n = 20), and the other did not undergo any surgery (the wild-type group, n = 10). All mice were euthanized separately at 8 weeks postoperatively, and Prickle1 mutant mice were also euthanized on the same day. 

Dacryocystectomy

All animals were kept at room temperature of 23 ± 2°C with relative humidity of 60 ± 10%, 12 hours of light and 12 hours of darkness, with adequate food and water for 7 days before the lacrimal sac removal surgery. Mice were deeply anesthetized with intraperitoneally injected 1% sodium pentobarbital (50 mg/kg), and 0.5% fast green dye (ab146267) was instilled onto the ocular surface to trace the nasal tear fluid path. The skin of the lacrimal sac area was carefully opened with small straight scissors, and the stained lacrimal sac was found and removed. The skin incision was then closed with a 10-0 nylon suture (the STDOB group). A similar operation was performed for the sham group but with the lacrimal sac left intact. All surgeries were performed by the same surgeon. After surgery, tobramycin eye ointment was applied to both eyes and skin wounds to prevent infection. 

Measurement of the Tear Flow

Tear flow was examined using a phenol red thread (Jingming, Tianjin, China).³⁰ In brief, the thread was applied to the lateral canthus of the eye in the anesthetized mice for 15 seconds. The length of the wet portion was measured in millimeters. Eyes were tested one at a time, first the right eye followed by the left eye. The average tear flow in both eyes was recorded. 

Corneal Fluorescein Staining

Fluorescein staining was used to assess the barrier damage of the corneal epithelium. The staining was performed by instilling 0.25% fluorescein sodium (Jingming, Tianjin, China) onto the cornea and photographed with a digital camera under cobalt blue light. The measurement of fluorescein-stained area was conducted using ImageJ software (version 1.52; National Institutes of Health, Bethesda, MD, USA).³¹ Percentage of the corneal fluorescein staining (CFS) area was calculated. 

Gross Morphological and Morphometric Measurement

The body weights of mice were recorded 8 weeks after surgery. The extra orbital LGs of each animal were dissected immediately after euthanasia. The LGs were placed on a grid paper sheet, and photographs were taken. The absolute wet weight of each LG was measured using a laboratory precision balance (Sartorius, BSA124S, Germany), and the weight indices were calculated as the LG weight divided by the body mass (LG/body, mg/g).³² 

The average acinus area measurement on tissue sections was performed on 3 random fields (10 random acini from each field) of each LG at 400× magnification. Five mice (LGs) were measured for each group. The acini area per imaging field was calculated as the acini percent area, three mice (LGs) were measured for each group. For each sample, three random fields were used for analysis. 

Tissue Histology

For Periodic Acid Schiff (PAS) staining, eyeballs were fixed in 10% formalin and embedded in paraffin and sectioned at 5 µm using a paraffin microtome (RM2016, Leica, German). The sections were stained with a PAS staining kit (G1008-20ML; Servicebio, Wuhan, China). Representative images of the conjunctiva were captured with a light microscope (Tissue FAXS Q+, TissueGnostics, Vienna, Austria). Three sections of each nasal, central, and lateral part of an eye from five animals were studied. The number of PAS-positive cells per section was counted.³³ 

For hematoxylin and eosin (H&E) staining, the LGs were fixed in 4% paraformaldehyde (PFA) for 24 hours immediately after dissection. The specimens were then washed and dehydrated in ascending grades of ethanol solutions, and were cleared in xylene for 2 hours after the last wash with pure ethanol. Impregnation and embedding were done first in soft paraffin wax at 59°C for 3 hours and then in hard paraffin at 63°C for 3 hours. The 5  µm sections were prepared every 50  µm for H&E staining. Representative images of the LG were captured with a light microscope (Tissue FAXS Q+, TissueGnostics, Austria). 

Transmission Electron Microscopy

Transmission electron microscopy (TEM) samples were prepared as described previously.³⁴ Briefly, LGs were dissected from the carcass quickly and continued fixation in 2.5% glutaraldehyde solution at 4°C for 24 hours. Sections were prepared and collected on mesh grids, stained with uranyl acetate and lead citrate, and examined under an electron microscope (JEM 1400; JEOL, Tokyo, Japan). All photographs were taken with a bio scan camera (Morada G3; EMSIS, Munster, Germany). 

Immunohistochemistry

As described previously, LG samples were fixed in 4% PFA overnight at 4°C. The tissue blocks were washed, dehydrated, embedded in paraffin, and cut in 5  µm sections. Antigen retrieval was performed by incubation with a 0.01M citrate solution (pH 6.0) for 30 minutes at 98°C. The sections were then blocked with 0.3% BSA and followed by incubation with primary antibodies (Rabbit anti-RAB3d 1:200, Rabbit anti-Vamp8 1:100, and Rabbit anti-Snap23 1:100) and then with Alexa-Fluor 488 goat anti-rabbit secondary antibody (1:1000). Staining was visualized with a confocal Zeiss microscope (LSM-880, Zeiss, Oberkochen, Germany). High-resolution digital images were captured and stored in TIFF format. Antibodies used are rabbit anti-Rab3d polyclonal antibody (Catalog No. 12320-1-AP; Proteintech, Rosemont, IL, USA); rabbit anti-Vamp8 monoclonal antibody (Catalog No. MA5-32502; Invitrogen, Grand Island, NY, USA); rabbit anti-Snap23 polyclonal antibody (Catalog No. DF13314; Affinity Biosciences, Cincinnati, OH, USA); and Alexa-Fluor 488 goat anti-rabbit secondary antibody (Catalog No. 111-545-003; Jackson ImmunoResearch Laboratories, West Grove, PA, USA). 

Protein Sample Preparation and Western Blotting

To prepare tear proteins, pilocarpine (300  µg/kg body weight) was subcutaneously injected to stimulate lacrimal gland secretion.³⁵ Tears were collected from the eyelid margin into Eppendorf tubes using a 0.5  µl micropipette 5 minutes after injection and stored at −80°C. 

To prepare LG proteins, unstimulated LGs were isolated and proteins were extracted with RIPA buffer (R0278; Sigma-Aldrich, Allentown, PA, USA) and 1 mM phenylmethyl sulfonyl fluoride (329-98-6; Sigma-Aldrich, USA), and a protease inhibitor cocktail (P8340; Sigma-Aldrich, USA). After centrifugation, the supernatant was collected and stored at −80°C. 

The bicinchoninic acid (BCA) method was used to determine the tear and LG protein concentration (BCA Protein Assay Kit, P0012; Beyotime, Haimen, Jiangsu, China). Western blotting analysis was then performed to determine relative protein expression levels in lacrimal gland tissue and tear fluid. Briefly, proteins were separated by electrophoresis on approximately 4% to 20% sodium dodecyl sulfate-polyacrylamide gels and transferred to polyvinylidene difluoride membrane (PVDF) membrane. After blocking in 5% bovine serum albumin in Tris-buffered saline containing 0.05% Tween-20 (TBST) for 2 hours, membranes were incubated with primary antibodies in TBST at 4°C overnight. After 3 times washing with TBST for 10 minutes each, membranes were incubated with horseradish peroxidase–conjugated secondary antibody (7074S; Cell Signaling Technology, Danvers, MA, USA) for 1 hour. After washing three times with TBST solution, protein signals were developed by enhanced chemiluminescence reagents (ECL; Catalog No. WBKLS0100; Millipore, Chicago, IL, USA). Signal density on membranes was digitalized using ImageJ software. For LG samples, membranes were probed with anti-Gapdh as a loading control. 

All antibodies for Western blotting were at 1:1000 dilution. Antibodies are rabbit anti-lysozyme (ST50-02; Invitrogen, Grand Island, NY, USA), rabbit anti-Lactoferrin (GXP297214; GenXspan, Shenzhen, China), rabbit anti-lipocalin2 antibody (ab63929; Abcam, Cambridge, MA, USA), and rabbit anti-GAPDH (14C10; Cell Signaling Technology, Danvers, MA, USA). The horseradish peroxidase-conjugated secondary antibody was bought from Cell Signaling Technology (7074S). The rest of the antibodies used for Western blotting were described in the previous section. 

RNA Sequencing

The LGs were dissected from age-matched 8-week-old control, surgical, and Prickle1 mutant mice. Three biological replicates from each group were subjected to RNA sequencing (RNAseq) analysis. Total RNA was extracted using a Trizol reagent kit (15596026, Invitrogen) according to the manufacturer's protocol. RNA quality was assessed on an Agilent 2100 Bioanalyzer (Agilent 2100; Agilent Technologies, Palo Alto, CA, USA) and agarose gel electrophoresis. mRNA was enriched by oligo (dT) beads (NEB #7335; New England Biolabs, Ipswich, MA, USA). The enriched mRNA was fragmented and reversely transcribed into cDNA with random primers (NEB#E 7530; New England Biolabs, Ipswich, MA, USA), followed by second-strand cDNA synthesis using DNA polymerase I (NEB#E 7530; New England Biolabs, Ipswich, MA, USA). The cDNA fragments were purified with a QiaQuick PCR purification kit (Catalog No. 28104; Qiagen, Hilden, Germany) and ligated to Illumina sequencing adapters. The size-selected ligation products from agarose gel electrophoresis were PCR amplified and sequenced using Illumina HiSeq2500 in Gene Denovo Biotechnology Co. (Guangzhou, China). 

Raw data (raw reads) of fastq format were firstly processed through in-house perl scripts to remove reads containing adapters, ploy-N, and low-quality reads. Q20, Q30, and GC content of the clean data were calculated. Reference genome and gene model annotation files were downloaded from the genome website (http://asia.ensembl.org/Mus_musculus/Info/Index) directly. Index of the reference genome was built using Bowtie version 2.0.6, and paired-end clean reads were aligned to the reference genome as FPKM (fragments per kilobase of transcript per million fragments mapped) using Top Hat version 2.0.9. 

RNA differential expression analysis was performed by DESeq2 software between two different groups. The genes with a false discovery rate (FDR) below 0.05 and absolute fold change ≥2 were considered differentially expressed genes (DEGs). The data were subjected to Gene Ontology (GO) Enrichment Analysis, which included molecular function, cellular components, and biological process. Pathway Enrichment Analysis based on the Kyoto Encyclopedia of Genes and Genomes (KEGG) database was performed to understand the significantly altered metabolic pathways in STDOB surgical or Prickle1 mutant mice. GO enrichment analysis and pathway analysis were performed with the given criteria of P < 0.05 and FDR <0.05. 

Real-Time RT-qPCR

For measuring gene expression, total RNA was isolated from whole LG with cooled TRIzol reagent (15596026; Invitrogen) and reverse transcribed into cDNA with HiScript III All-in-one RT SuperMix Perfect for qPCR (R333; Vazyme). Real-time quantitative PCR (RT-qPCR) was performed with the ChamQ SYBR Color qPCR Master Mix (Q431; Vazyme) on the Applied Biosystems 7500 system. The PCR program was as follows: 40 cycles of 95°C for 10 seconds and 60°C for 30 seconds after an initial denaturing at 95°C for 30 seconds. The primers for Prickle1 were: 5′-GTATGCTGGCACCCGTCCTG-3′ (forward) and 5′-GCACCGAGGCTTGAGCAGTT-3′ (reverse). The primers for Gapdh were 5′-GGAGAGTGTTTCCTCGTCCC-3′ (forward) and 5′- ATGAAGGGGTCGTTGATGGC-3′ (reverse). The relative expression levels were calculated using the CT method on the default ABI software SDS 2.3 (ThermoFisher Scientific, Waltham, MA, USA), as described previously.³⁶ 

Statistical Analysis

Differences in phenol red thread test values, percentages of CFS area, and goblet cell numbers between two individual groups were analyzed using GraphPad with an unpaired Student t-test to detect statistical powers. Comparisons of LG body weights, average acinus area, percent areas of acini, the relative mRNA levels, and Western blotting signal intensities were conducted with one-way analysis of variance (ANOVA) with Tukey post hoc testing. Differences between the measurements were considered significant if the P value was 0.05 or less. 

Results

Creation of a Surgical Tear Drainage Obstruction Mouse Model by Dacryocystectomy

To understand whether the tear drainage system would impact LG function, we performed lacrimal sac removal experiment in mice. We developed a surgical protocol (see Materials and Methods) in which fast green dye was instilled on ocular surface to trace the nasal tear fluid path (Figs. 1A-C). A successful operation will block tear flow, thus, the fast green dye through the nasolacrimal duct. Five out of 20 STDOB mice or sham-operation control mice were randomly picked for tear-outflow examination. None of the STDOB mice showed nasal tear duct staining of fast green, indicating successful lacrimal sac removal. In contrast, all sham operation controls showed green dyes in the tear duct (Figs. 1D, 1E). 

Figure 1.

 

Schematic illustration of dacryocystectomy to create a surgical tear drainage obstruction (STDOB) mouse model. (A-C) Schematic diagrams of the lacrimal sac removal procedure. (A) Fast green dye was instilled onto the ocular surface to track the tear path through the drainage duct. (B) The lateral nasal skin was carefully opened to expose the lacrimal sac marked by the fast green dye. (C) The lacrimal sac (L.S.) was removed, and the skin was sutured. (D, E) Representative pictures demonstrating dye-filled lacrimal sac of the sham (left) and experiment (right) eyes 8 weeks post operations. White arrows indicate the position of the lacrimal sac.

Ocular Surface Abnormalities of the STDOB Mice

We first examined whether the obstructed tear drainage in STDOB mice increased tear accumulation on the ocular surface. It was first noticed that abundant white eye discharge was present in the inner canthus of the STDOB but not in the control mice (Figs. 2A, 2B). The tear volume of the STDOB mice measured by the phenol red thread³⁰ also significantly increased in STDOB mice (6.30 ± 0.94 mm for the STDOB group vs. 1.85 ± 0.85 mm for the sham group; P < 0.0001; Fig. 2C). We speculated that the ocular surface directly covered by the tear film might also be altered in STDOB mice. Indeed, when examined by fluorescein staining, the cornea epithelium of STDOB mice also showed larger stained areas than the sham (Figs. 2D, 2E) (10.75 ± 4.55% in the STDOB group vs. 3.41 ± 1.75% in the sham control mice group; P = 0.0002; Fig. 2F). 

Figure 2.

 

STDOB mice exhibited epiphora, increased corneal fluorescein staining, and increased conjunctival goblet cells. (A) Sham operation: a representative image of the normal ocular surface. (B) Epiphora with white discharge in the inner canthus of the STDOB mice. Arrows point to the inner canthus. (C) Tear flow test using the phenol red thread. (D) Corneal fluorescein staining of the sham-operated mice showing integral surface. (E) Punctate fluorescein staining of the STDOB mouse cornea. (F) Quantifications of the fluorescein-stained corneal area of the STDOB and sham mice. (G) PAS-stained goblet cells in the sham-operated mice had a round shape and uniform size. (H) More conjunctival goblet cells of the STDOB mice stained by the PAS. (I, J) Magnified images of the goblet cells from boxed areas in (G) and (H), respectively. (K) Quantification of the number of goblet cells of the STDOB and sham mice; n = 10 mice for (C) and (F); and n = 5 mice for (K).

We next examine whether conjunctival epithelium, another major component of the ocular surface, was also altered in the STDOB mice. By PAS staining, the conjunctival epithelium appeared thickened with increased goblet cells (Figs. 2G-J; 42.47 ± 7.39 cells/section in the STDOB mice group vs. 13.33 ± 3.53 cells/section in the sham controls group; P < 0.0001; Fig. 2K). Additionally, goblet cell secretion revealed by PAS staining appeared remarkably more active in the STDOB than in the sham control mice (see Figs. 2G-J). Thus, the results suggested a systemic alteration of the ocular surface epithelium and secretion upon obstructed tear drainage. 

Ocular Surface Phenotypes of the STDOB Mice Resemble Those of Prickle1 Mutants

We next investigated whether the ocular phenotypes manifested in the STDOB mice were also present in mice with congenital tear drainage obstruction. The only reported genetically engineered mice with tear duct dysplasia are the Prickle1 gene knockouts.^27,35 Similar to the STDOB animals, Prickle1 mutant mice showed epiphora with white discharge (Figs. 3A, 3B). The amount of tear secretion measured by the phenol red thread was also increased (7.45 ± 0.49 mm for the Prickle1^a/b group vs. 1.9 ± 0.61 mm for the wild type mice group; P < 0.0001; Fig. 3C). Additionally, Prickle1 mutant corneal epithelium showed punctate fluorescein staining (Figs. 3D, 3E), with the mean area of epithelial staining significantly increased (13.17 ± 3.07% for the Prickle1^a/b group vs. 1.54 ± 1.63% for the control mice group; P < 0.0001; Fig. 3F). Conjunctiva goblet cells of the Prickle1 mutant mice showed heavily apical staining compared with the controls (Figs. 3G-J). Furthermore, the number of goblet cells in Prickle1 mutant mice was significantly increased as well (40.00 ± 6.78 cells/section for the Prickle1^a/b group vs. 9.66 ± 1.79 cells/section for the control mice group; P< 0.0001; Fig. 3K). Thus, Prickle1 mutants fully recapitulate the ocular surface phenotype of STDOB mice, indicating that obstruction of tear duct does cause tear-related ocular surface disorders. 

Figure 3.

 

STDOB mice showed similar phenotypes to those observed in Prickle 1 mutant mice. (A, D) Normal ocular surface with intact corneal epithelium showed in the wild-type mice. (B, E) Epiphora with white discharge and increased fluorescein stain on the cornea of the Prickle1 mutant mice. The arrow indicates the white discharge in the inner canthus. (C) Tear flow was significantly increased in the Prickle1 mutant mice using the phenol red thread test. (F) Corneal fluorescein staining score revealed more severe staining of the Prickle1 mutant mice. (G) Periodic acid-Schiff (PAS) staining of the conjunctival epithelium of the wild-type mice showed rounded and uniform staining patches. (H) The Prickle1^a/b group showed an increase in goblet cells with enhanced staining in the apical areas. (I) and (J), Magnified images from boxed areas in (G) and (H), respectively. (K) The number of goblet cells was significantly increased in the Prickle1^a/b group; n = 10 mice for (C) and (F); and n = 5 mice for (K).

Tear Components and Lacrimal Gland-Secreted Proteins are Similarly Altered in STDOB and Prickle1^a/b Mice

The tear fluid is mainly secreted by the LG, playing an important role in the maintenance of ocular health. The observed epiphora and ocular surface abnormalities prompt us to investigate whether major tear proteins were also altered. Lysozyme, tear lipocalin, and lactoferrin of the tear film have diverse roles in the protection of the ocular surface. An examination of the levels of these proteins in the tear fluid by Western blotting revealed increased expression of all three proteins in both STDOB and Prickle1 mutant animals (Fig. 4A). Consistent with the tear fluid, the LG secretion of these proteins was also increased (Fig. 4B). Further quantification of Western blotting showed that the LG lactoferrin, lipocalin, and lysozyme were increased by 1.7-fold (P = 0.0686), 1.7-fold (P = 0.0218), and 1.8-fold (P = 0.0012), respectively, in STDOB mice (Fig. 4C). Prickle1 mutant mice showed similar increasing trends of LG lactoferrin (1.6-fold, P = 0.0113), lipocalin (2.1-fold, P = 0.0744), and lysozyme (2.7-fold, P < 0.0001; see Fig. 4C). Therefore, the LG secretory function is altered upon tear duct obstruction. 

Figure 4.

 

Alterations of tear component and lacrimal gland-secreted proteins in STDOB and Prickle1 mutant mice. (A, B) Representative images of Western blotting of lactoferrin, lipocalin, and lysozyme in the tear fluid (A) and LG (B). The same amount of proteins (see Methods and materials) were loaded for polyacrylamide gel electrophoresis, with the Gapdh serving as a control for LG protein loading (B). (C), Quantification of Western blotting analysis showed increased lactoferrin, lipocalin, and lysozyme from the LG of the STDOB and Prickle1 mutant mice. Protein levels were normalized to Gapdh, and relative expression was compared between the sham and the treated groups (n = 3 LGs/mice).

Aberrant LG Secretory Vesicle Fusion in STDOB and Prickle1 Mutant Mice

The altered LG secretory function led us to examine whether the LG size and histology also changes. On stereomicroscope, no obvious differences in LG size or general appearance were observed (Figs. 5A-C). The wet weight of individual LG did not significantly alter when normalized to the body weight (LG mass/body mass: STDOB mice 0.55 ± 0.03 mg/g, Prickle1^a/b mice 0.56 ± 0.04 mg/g, and sham mice 0.56 ± 0.03 mg/g; P = 0.89; Fig. 5D). 

Figure 5.

 

Morphologic alterations of lacrimal glands in the STDOB and Prickle1 mutant mice. (A-C) Stereomicroscopy of the LGs from the sham (A), STDOB (B), and Prickle1 mutant mice (C). (D) Quantification of relative LG weights. The wet LG weights were normalized to the body weights presented as ratio indices (LG/body, mg/g); N =10 LGs. (E-J) H&E staining of the LGs from the sham controls (E, H), STDOB (F, I), and Prickle1 mice (G, J). (H), (I), and (J) are magnified images from the boxed areas of (E), (F), and (G), respectively. (K) Average individual acinus areas. Ten acini from each imaging field are roughly randomly chosen for measuring individual acinus areas. The results are presented as average area/acinus. The n = 5 mice/group. (L), The total average acinar area per imaging field was presented as percentages (n = 3 mice/group).

We then performed H&E staining to determine whether pathological changes occurred in the LG among the three groups. LG in all groups showed similar acinus histology (Figs. 5E-J). Additionally, the STDOB (see Figs. 5F, 5I), Prickle1 mutant (see Figs. 5G, 5J), and sham control mice showed similar individual (STDOB 258.4 ± 9.834  µm²; Prickle1^a/b 258.0 ± 8.377  µm²; sham 252.7 ± 7.134  µm²; P = 0.1748; Fig. 5K) and total percent (STDOB 78.6 ± 4.643 %; Prickle1^a/b 77.12 ± 2.964 %; sham 75.39 ± 3.359 %; P = 0.1579; Fig. 5L) acinar areas. 

We next investigated whether the ultrastructure of the LG may have been altered corresponding to the secretory function changes. On the TEM of the sham group, LG secretory vesicles had uniform sizes filling in the apical and central areas of the acinar cells with moderate electron density, which were membrane-bound (Figs. 6A, 6B). In contrast, the STDOB and Prickle1 mutant LGs reduced electron density in acinar cells (Figs. 6C-F). Furthermore, the vesicle membrane boundary was blurred, and the fusion of vesicles was noticed (see Figs. 6D, 6F). Thus, despite the relatively normal H&E histology, the LG acinar ultrastructure was altered significantly, consistent with the abnormal tear components and LG secretion. 

Figure 6.

 

LG ultrastructure revealed by transmission electron microscopy (TEM). (A, B) LG showed uniform electron density of the acinus secretory vesicles in the sham-operated mice. (C-F), Altered electron density of the LG acinar vesicles in the STDOB and Prickle1 mutant groups. (B), (D), (F) Higher magnification of the boxed areas from (A), (C), and (E), respectively. The green arrow in (B) indicates the clear membrane-bound secretory vesicles in the sham mice. The red arrow in (D) (STDOB) and the purple arrow in (F) (Prickle1 mutant) point to the enlarged fused vesicles with lighter electron density and heterogenous membrane curvatures.

Altered Expression and Localization of Proteins Related to LG Vesicle Fusion and Secretion

To further evaluate how the LG secretory function was altered, we examined protein expression and localization of Rab3d (a vesicle-trafficking GTPase), Vamp8 (a v-SNARE vesicle fusion receptor), and Snap23 (a membrane fusion targeting t-Snare protein), which have been shown to be involved in LG secretion. Immunohistochemistry revealed basally localized Rab3d in the STDOB and Prickle1 mice, contrasting to its normal apical distribution in the sham controls (Fig. 7A). Moreover, Western blotting showed a remarkable decrease of Rab3d expression (0.65-fold, P = 0.025 and 0.69-fold, P = 0.024 for STDOB and Prickle1 mice, respectively; Fig. 7B). Vamp8 and Snap23 were also basally mislocalized (Figs. 7C, 7E), but their expression levels were enhanced (Fig. 7D: Vamp8:1.24-fold for STDOB mice, P = 0.182 and 2.46-fold for Prickle1 mice, P = 0.004; Fig. 7F: Snap23: 1.29-fold for STDOB mice, P = 0.122, and 2.16-fold Prickle1 mice, P = 0.029). Collectively, our findings suggested that apical trafficking or fusion of the tear-containing vesicles might be impaired in the STDOB and Prickle1 mutant mice. 

Figure 7.

 

Altered expression and localization of proteins relevant to lacrimal gland fusion and secretion. (A) Rab3d immunofluorescence was abundant in the apical lumen side of the acinar cells in the sham LGs (left panels), but uniformly distributed in the STDOB (middle panels) and more basally localized in the Prickle1 mutant acini. The bottom panels are magnified images from the top panels. Apical and basal positions of the acinus lumen are illustrated in the right corner of the first panel. (B) Western blotting demonstrated reduced Rab3d expression in the STDOB and Prickle1 mutant LGs compared with the sham. (C) Vamp8 immunofluorescence showed a similar altered pattern as the Rab3d in the STDOB and Prickle1 mice. (D) Western blotting showed increased LG expression of Vamp8 in the STDOB and Prickle1 mice. (E) Expression of Snap23 in LG acini was also basally shifted and showed increased levels in STDOB and Prickle1 mice compared with the sham (F). The white dotted lines indicate acini boundaries. For Western blotting and quantification, n = 4 LGs/mice.

Lacrimal Gland Expression Profiles of STDOB and Prickle1 Mutant Mice

The targeted examination of markers relevant to the LG secretion suggested that STDOB mice and the Prickle1 mutant mice shared similar cellular alterations. To further grasp their similarities at a larger scale, we performed RNAseq analysis to examine the LG transcriptomic profiles of the two models. The RNAseq analysis found 682 and 1351 differentially expressed genes (DEGs) in the STDOB and Prickle1 mouse LGs, respectively (Supplementary Figs. 1A, 1B). Three hundred seventy-five altered genes were found in common between the STDOB mice and Prickle1 mice mutant LGs, which was more than half of the STDOB and about one-third of the Prickle1 mutant DEGs (Figs. 8A, 8B). GO analysis demonstrated a remarkable similarity in biological processes, molecular functions, and cellular components between the two models (Supplementary Figs. 1C, 1D). Strikingly, the top 6 terms shared by the total 682 STDOB and 1351 Prickle DEGs are essentially the same in KEGG pathways (Figs. 8C, 8D). We further performed the KEGG enrichment of the common 375 DEGs shared by the two models and found that pathways, including PI3K-Akt, calcium, phospholipase D, and NF-kappa B signaling (Fig. 8E), were all reported to be involved in glandular secretion.^37–40 Thus, the STDOB and Prickle1 mutant LGs are generally similar in transcriptomic landscapes. 

Figure 8.

 

RNA sequencing and KEGG pathway analysis. (A) Differentially expressed genes (DEGs) of LGs from respective STDOB and Prickle1 mutant mice compared with the sham. Three hundred seventy-five common genes were shared by STDOB and Prickle1 mutant groups. (B) A Venn diagram of intersections between up- and downregulated DEGs. (C) Top 6 KEGG pathways enriched for the total 682 STDOB DEGs. (D) Top 6 KEGG pathways enriched for the total 1351 Prickle 1^a/b DEGs. (E) The top 6 KEGG pathways enriched for the shared 375 DEGs between STDOB and Prickle1 mutant mice.

Prickle1 is not Expressed in the Lacrimal Gland

The LG changes of the Prickle1 mutant mice could be directly caused by the loss of Prickle1 in the LG. We, therefore, examined whether Prickle1 is expressed in the lacrimal gland. We first checked the average FPKMs of Prickle1 transcripts in the sham, STDOB, and Prickle1 mice groups, which all appeared very low (1.59, 0.97, and 0.22, respectively; Fig. 9A). We then examined Prickle1 mRNA expression in the LG tissue by RT-qPCR analysis. Compared with the control Gapdh expression, which was detectable at about 20 PCR cycles (Ct20), the Prickle1 expression required more than 35 cycles (Ct35) to be detected in the sham or STDOB LGs, and even more cycles in the Prickle1 mutants (above 40 cycles), as anticipated (Fig. 9B). Last, we took advantage of the Gfp knock-in reporter in the Prickle1 locus to surrogate Prickle1 expression using Prickle1^b/+ heterozygous mutants. Immunostained-GFP was found strongly expressed in the hair follicles (Fig. 9C) but undetectable in the LG (Fig. 9D). Thus, the data suggested that obstruction of the tear duct is likely the cause of the abnormalities of LG. 

Figure 9.

 

Prickle1 is not expressed in the lacrimal gland. (A) The mRNA expression amplitudes of the Prickle1 gene in RNA-seq data from the three experimental groups. FPKM, fragments per kilobase of transcript per million mapped reads. (B) RT-qPCR to determine Prickle1 mRNA levels in the lacrimal gland; CT, cycle threshold. (C) GFP reporter protein was detected by immunohistochemistry in hair follicles (serving as a positive control) but not in the LG acini of the Prickle1^b/+ mice.

Discussion

The tear apparatus consists of many components with diverse yet coordinated functions. Tears are secreted from the LG, distributed by blinking and evaporated from the ocular surface, and drained through the nasolacrimal duct.⁴¹ Anatomically, the nasolacrimal duct, conjunctiva, and LG epithelia are continuous and share the same origin from the embryonic conjunctival ectoderm.^25,42 Functionally, the lymphoid tissues of the three parts of the tear apparatus are also connected,⁴³ and, as a result, the nasolacrimal system plays an important role in ocular surface innate immune defense.⁴⁴ Despite the intimate relationships of the three major components of the tear apparatus, the functional connection between either of the two has not been established. In this study, we demonstrated, for the first time, that tear sac/duct impacts LG structure and tear secretion function in both surgical and genetic mouse models with tear duct partially removed. 

Several previous studies using different animal models demonstrated that ligation of the LG ducts causes reduced tear secretion and LG weights, accompanied by inflammation, increased cell proliferation, apoptosis, and lipid accumulation.³⁰ Our finding that obstruction of the tear duct also led to LG structural and functional changes is striking. Unlike the LG ducts, which physically connect to LG in proximity, tear ducts are separated from LG by a wide-open territory of the ocular surface. Thus, our results hint that a remote signaling system exists between the tear duct and LG to coordinate their physiological activities. It further supports the notion that LG possesses a great remodeling plasticity and capacity illustrated by several studies.^45,46 

Management of a few ocular surface diseases requires removal or obstruction of the tear drainage components. However, the surgical side effects on the ocular surface have not been systemically evaluated due to a lack of animal models. We created a mouse model with the lacrimal sac surgically removed. Surprisingly, we observed corneal injuries and aberrant secretion of the conjunctival goblet cells accompanied by epiphora phenotypes. The ocular surface phenotypes resemble those seen in Prickle1 mutant mice we previously reported²⁷ and in the current study. These results strongly suggested an inherent relevancy existing between the tear duct and ocular surface. 

It is intuitive to think that the observed epiphora in our animal models is due to obstructed tear drainage leading to tear accumulation on the ocular surface. However, the tear appearance does not look normal in that white discharges were observed. Therefore, we examined whether LG malfunction is involved. Indeed, examination of LG histology at different levels reveals striking ultrastructural differences by TEM. We further performed immunohistochemistry with molecular markers involved in vesicle secretion, including Rab3d, Vamp8, and Snap23, and found remarkable alterations in their expression and apicobasal distribution in our animal models. Moreover, tear components were also changed by Western blotting. The results suggest that, on the one hand, the epiphora could directly result from the obstruction of the tear duct, but most importantly, LG dysfunction might also be involved. 

The Rab3d, Vamp8, and Snap23 are crucial for membrane fusion of the secretory vesicles as well as their trafficking. These proteins are usually polarized in their cellular distribution.^47–51 Altered distribution pattern of these proteins has been documented as pathological characteristics of exocrine glands in patients with dry eye disease.^52,53 The altered expression and subcellular localization of these proteins in acinar cells of our animal models recapitulate these in dry eye disease. For instance, the vesicle-vesicle (instead of vesicle-plasma membrane) fusion of the LG acini in our models is a sign of disruption of the exocytic machinery appeared in the salivary gland of patients with Sjögren's syndrome.⁵⁴ Similarly, the loss of Rab3d apical localization of the LG was also observed in the mouse model for Sjögren's syndrome^55,56 and in the salivary gland of patients with Sjögren's syndrome.⁵² The loss of apical domains of Vamp8 and Snap23 localization observed in this study (mislocalized or shifted basolaterally) has been reported in the salivary gland acinar cells of Sjögren's syndrome as well.⁵³ Consequently, the lipocalin, lactoferrin, and lysozyme in tears or LG were all altered in our animal models. Altogether, these data indicate that tear duct blockage leads to LG dysfunction, which might also be a suitable dry eye disease model system. 

Several clinical investigations showed reduced lacrimal fluid drainage was followed by decreased tear production with epiphora.⁵⁷ For instance, reduced tear production was observed in patients with eversion of the lower punctum or compression of the inferior lacrimal canaliculus.⁵⁸ Similarly, the dacryocystectomy also leads to a compensatory reduced tear production.¹³ Although the LG cannot be directly evaluated in these patients, the above results combined with our data suggest that the LGs of these patients may also be impaired, and tear production in our animal models likely decreased rather than overproduced. 

Several mechanisms could have contributed to the LG dysfunction upon blockage of the tear drainage. First, the excess tears on the ocular surface may send a signal to LG through the conjunctiva openings of lacrimal gland ducts,¹³ which directly feedbacks tear secretion. Second, a communicative path from the tear duct to LG through regional circulation was interrupted. It has been proposed that the tear components are absorbed into the blood vessels of the cavernous body around the lacrimal sac and nasolacrimal duct, which connect with the extraocular LG.⁴⁴ Third, neural communications through the “nose (or tear duct)-superior salivatory nucleus-LG” axis⁵⁹ was impaired leading to LG dysfunctions. Some studies demonstrate higher levels of neuropeptides in the nasal and tear secretions of allergic rhinoconjunctivitis,^60,61 which may contribute to this neural axis. Additionally, this neural axis has facilitated intranasal tear neurostimulators to treat dry eye disease in some circumstances.⁶² Nonetheless, these possible mechanisms require further in-depth investigations in the future. 

In conclusion, this study offers a new perspective on the importance of the integrity of the nasolacrimal system on the LG function. The long-distance interaction between the nasolacrimal duct and LG may underlie many tear-related ocular surface diseases, including the infamous dry eye disease. As such, surgeries to remove tear duct components should be carefully considered in clinics, and the tear duct may also facilitate the treatment of the dry eye disease. Last, we created both STDOB and genetic mouse models, which could be utilized to investigate a broad range of nasolacrimal-ocular surface diseases. 

Acknowledgments

Supported by the National Natural Science Foundation of China (82070922) and the Science and Technology Program of Guangzhou (202201020544). 

Disclosure: B. Xiao, None; D. Guo, None; R. Liu, None; M. Tu, None; Z. Chen, None; Y. Zheng, None; C. Liu, None; L. Liang, None