An Autotaxin-Induced Ocular Hypertension Mouse Model Reflecting Physiological Aqueous Biomarker

Abstract

Purpose: Animal models of ocular hypertension (OH) have been developed to understand the pathogenesis of glaucoma and facilitate drug discovery. However, many of these models are fraught with issues, including severe intraocular inflammation and technical challenges. Lysophosphatidic acid (LPA) is implicated in trabecular meshwork fibrosis and increased resistance of aqueous outflow, factors that contribute to high intraocular pressure (IOP) in human open-angle glaucoma. We aimed to elevate IOP by increasing expression of the LPA-producing enzyme autotaxin (ATX) in mouse eyes.

Methods: Tamoxifen-inducible ATX transgenic mice were developed. Tamoxifen was administered to six- to eight-week-old mice via eye drops to achieve ATX overexpression in the eye. IOP and retinal thickness were measured over time, and retinal flat-mount were evaluated to count retinal ganglion cells (RGCs) loss after three months.

Results: Persistent elevation of ATX expression in mouse eyes was confirmed through immunohistochemistry and LysoPLD activity measurement. ATX Tg mice exhibited significantly increased IOP for nearly two months following tamoxifen treatment, with no anterior segment changes or inflammation. Immunohistochemical analysis revealed enhanced expression of extracellular matrix near the angle after two weeks and three months of ATX induction. This correlated with reduced outflow facility, indicating that sustained ATX overexpression induces angle fibrosis, elevating IOP. Although inner retinal layer thickness remained stable, peripheral retina showed a notable reduction in RGC cell count.

Conclusions: These findings confirm the successful creation of an open-angle OH mouse model, in which ATX expression in the eye prompts fibrosis near the angle and maintains elevated IOP over extended periods.

Glaucoma is a neurodegenerative disease characterized by visual-field loss associated with optic nerve (ON) damage and progressive retinal ganglion cell (RGC) death.¹ Open-angle glaucoma (OAG), the most common subtype of glaucoma, is associated with elevated intraocular pressure (IOP), but the biological basis for this elevation remains largely unknown.² One challenge in glaucoma research is that the pathogenesis of IOP elevation and RGC death at the molecular level has not been elucidated in human eyes. A detailed etiological understanding is crucial for the treatment of OAG. In this context, animal models mimicking glaucoma pathogenesis play a pivotal role. Recently, morphological changes and RGC death have become detectable in human eyes in vivo using optical coherence tomography (OCT) and biomarkers of apoptosis.^3–5 Nonetheless, the cell-cell interactions and molecular mechanisms underlying glaucoma remain under-researched due to the absence of an ocular hypertension (OH) animal model based on the mechanism of IOP elevation. 

Historically, a variety of animal models has been used to investigate RGC death, mirroring glaucomatous ON damage. Despite considerable differences in ocular structures between mice and humans,⁶ mouse models serve as valuable tools in elucidating the intricate pathophysiology of glaucoma and in probing potential treatment strategies.⁷ Among these models, those based on ON crush, ischemia reperfusion, and drug-induced cell or axon damage do not effectively mirror clinical glaucoma. OH models, conversely, are more adept at replicating OAG, simulating the chronic IOP-induced axon damage. Multiple methods have been used to devise rodent glaucoma models, including introducing cells or particles into the anterior chamber to impede outflow by obstructing the trabecular meshwork (TM) or cauterizing the superior scleral vein, causing increased IOP and degeneration of ON axons.^8–10 However, these models face challenges such as high invasiveness, complications, inconsistently elevated IOP, and difficulties in maintaining prolonged IOP elevation.^10,11 The DBA2J strain, discovered by chance, is frequently used as an OH mouse model.¹² This strain carries two genetic mutations that result in chronic OH but eventually exhibits symptoms like iris atrophy, pigment dispersion, and cataract. Consequently, continuous observation of RGC death, ON changes, and IOP-induced modifications in the outflow pathway become untenable. Many other transgenic and inducible models of ocular hypertension and glaucoma were reported such as myocilin transgenic mice, constitutively active FN-EDA mice, and inducible viral models.^13–15 However, each mouse model has both advantages and challenges. For example, the Ad5.TGFβ2 model has challenges such as inflammation in the anterior segment of the eye. Given these limitations, there is a pressing need for animal models that effectively mimic OAG and are suitable for both pathological studies and drug-discovery research. Genetically modified animal models inducing OH, rooted in an in-depth understanding of molecular changes based on genetic information, are garnering growing interest. 

IOP homeostasis is sustained by the dynamics of the aqueous humor between its production in the ciliary body and its drainage, predominantly through the TM outflow pathway. The outer part of the TM forms the juxtacanalicular connective tissue composed mostly of extracellular matrix (ECM) with occasional cells which is located adjacent to the inner wall of Schlemm's canal. The juxtacanalicular connective tissue and the inner wall endothelium of Schlemm's canal have been identified as a key contributor to outflow resistance in the normal and glaucomatous eye.^16,17 Abnormal accumulation of ECM or TM/ECM stiffness have been also mentioned as the other major pathological changes related to the IOP elevation in the outflow pathway associated with glaucoma, yet the intricate molecular mechanism remains to be fully elucidated.^18,19 Autotaxin (ATX; also designated as ENPP2) is a secreted lysophospholipase D (LysoPLD) that is responsible for cleaving choline from lysophosphatidylcholine to yield lysophosphatidic acid (LPA).²⁰ LPA is a typical phospholipid that interacts with G-protein-coupled receptors and is thought to trigger various cellular responses, spanning cell adhesion, contraction, proliferation, migration, and cytokine and chemokine secretion.^21–25 Recent findings indicate augmented ATX expression in the tissues and aqueous humor of glaucoma patients, with ATX/LPA concentrations in the aqueous humor showing a correlation with IOP.²⁶ Moreover, cellular and animal research implies that the ATX-LPA axis is crucial in regulating aqueous humor outflow, as evidenced by alterations in TM or Schlemm's canal cell morphology, ECM accumulation, and tissue scarring, mediated through pathways like Rho GTPase, p38MAPK, and YAP/TAZ.^25,27–31 

This study aimed to determine whether sustained ATX upregulation in mouse eyes can elicit high IOP. Moreover, we assessed the degree to which it represents the human OAG pathogenesis. Previous studies highlighted that systemic ATX overexpression resulted in embryonic lethality, growth hindrance, and issues with neural-tube closure and vascular injury, whereas its conditional overexpression during the neonatal phase retarded retinal vascularization and curtailed vascular branching.³² Utilizing ATX conditional transgenic mice, this study endeavored to achieve heightened ATX expression in the eyeball through topical tamoxifen administration, thus minimizing systemic side effects in mature organisms.³³ Our results underscore the significance of the ATX-LPA pathway in the etiology of the IOP elevation seen in OAG. Additionally, our model offers a platform for probing the pathophysiological alterations that occur in ocular tissues because of OH. 

Methods

Preparation of Transgenic Mouse

All animal experiments adhered to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research, and approval was obtained from the Institutional Animal Research Committee of the University of Tokyo. Animals were maintained under a 12-hour light/dark cycle with unrestricted access to food and water. CAG-loxP-neor-loxP-ATX (LNL-ATX) transgenic mice with a C57BL/6J lineage were provided by Prof. Junken Aoki (The University of Tokyo, Tokyo, Japan).³² To produce ATX conditional Tg (ATX Tg) mice, homozygous LNL-ATX mice were bred with mice bearing tamoxifen-inducible Cre recombinase (Cre-ER) regulated by the CAG promoter (B6.Cg-Tg(CAG-cre/Esr1*)5Amc/J; Jackson Laboratories, Bar Harbor, ME, USA). The CAG-CreER mouse lineage was preserved through breeding with C57BL/6J mice (Japan SLC, Shizuoka, Japan). All experimental procedures commenced after six weeks, when the TM had reached full maturity.³⁴ Gene activation was initiated via topical administration of 5 mg/mL tamoxifen in corn oil as eye drops (10 µL/eye), three times daily at four-hour intervals for five days. CreER +/− LNL-ATX mice (ATX Tg) were used for the experiments while their littermates served as controls. Both male and female mice participated in all experiments. 

Genotyping

Tail samples from three-week-old pups were subjected to genotyping using specific primers. The efficiency and site specificity of topically administered tamoxifen with respect to gene excision were confirmed via PCR. One week after the termination of tamoxifen application, samples from the anterior and posterior eye, eyelids, tail, liver, and whole blood of ATX Tg mice were collected and instantly frozen using liquid nitrogen. DNA extraction was performed with NaOH. Subsequent PCR used the following primers, with the resulting products subjected to gel electrophoresis. After electrophoresis, DNA was stained with EtBr and visualized using a CCD camera (Printgraph AE-6932; ATTO, Tokyo, Japan). For genotyping the LNN-ATX and detecting DNA excision due to tamoxifen, the primers used were as follows: Fwd, 5′-CTTTTTCCTACAGCTCCTGGG-3′ and Rev - 5′-CCATTCGGCCCTCTTAATTCG-3′. In the case of CreER genotyping, the primers were Fwd, 5′-GCTAACCATGTTCATGCCTTC-3′ and Rev, 5′-AGGCAAATTTTGGTGTACGG-3′. 

IOP Measurement

After anesthetizing the mice with isoflurane, IOP was measured for up to three months after the initiation of topical tamoxifen using a Tonolab rebound tonometer (iCare, Helsinki, Finland) as per the manufacturer's guidelines. IOP measurement commenced at three minutes after anesthesia induction and concluded within five minutes to prevent any impact of anesthesia. All IOP readings were taken between 9:00 AM and 12:00 PM. Saline solution was used during the process to prevent corneal desiccation. 

LysoPLD Activity Measurement

Aqueous humor samples from both control and ATX Tg mice were collected using a glass needle before and at three to 94 days after tamoxifen injection. Immediately after collection, the samples were flash-frozen in liquid nitrogen and stored at −80°C until analysis. The ATX activity within the aqueous humor was ascertained by gauging the LysoPLD activity using methods previously described.^35,36 Samples were incubated with 2 mM 1-myristoyl (14:0) LPC (Avanti, Alabaster, AL, USA) in an assay buffer (comprising 5 mM MgCl₂, 5 mM CaCl₂, 500 mM NaCl, 0.005% Triton-X100, in 100 mM Tris-HCl at pH 8.0) at 37°C for four hours. Subsequently, free choline was detected through enzymatic photometry, using choline oxidase (Asahi Kasei, Tokyo, Japan), horseradish peroxidase (Toyobo, Osaka, Japan), and the TOOS reagent (N-ethyl-N-[2-hydroxy-3-sulfoprolyl]-3-methylaniline; Dojindo Molecular Technology Inc. Rockville, MD, USA) as the hydrogen donor.³⁷ 

Immunohistochemistry

At two weeks or three months after tamoxifen injection, the mice were euthanized under anesthesia by severing the cardiac coronary artery. Eyes were enucleated and subsequently fixed in 4% paraformaldehyde (PFA) at 4°C for two hours, followed by immersion in 30% sucrose at 4°C overnight. The eyes were then embedded in OCT compound (Sakura Finetek Japan, Tokyo, Japan) and stored at −80°C. Sections 10 µm in thickness were cut from these frozen blocks. The sections were first treated with a blocking solution (Nacalai Tesque Inc., Kyoto, Japan) at room temperature for 40 minutes to reduce non-specific binding. They were then incubated with primary antibodies overnight at 4°C, followed by PBS washes. Subsequent incubation with secondary antibodies was performed at room temperature for one hour. Finally, sections were mounted using VectaShield Mounting Media (Vector Laboratories, Burlingame, CA, USA). Immunofluorescence imaging was accomplished using a confocal laser microscope (LSM880 with Airyscan; Carl Zeiss, Oberkochen, Germany). For ATX expression analysis, Anti-autotaxin mAb (D323-3; MBL Life Science, Nagoya, Japan) was used as the primary antibody, and Alexa Fluor 488 chicken anti-rat IgG (H+L) (A21470; Invitrogen, Carlsbad, CA, USA) was used as the secondary antibody. For ECM accumulation analysis near the angle, anti-collagen I antibody (AB765P; Sigma-Aldrich, St. Louis, MO, USA), anti-Fibronectin antibody (sc-271089; Santa Cruz Biotechnology, Dallas, TX, USA), and Phalloidin; Fluorescent Derivatives (PHDR1; Acti-StainTM 535 (Cytoskeleton, Inc., Denver, CO, USA) were used as the primary antibody, and Alexa Fluor 488 goat anti-rabbit IgG (H+L) (A11008; Invitrogen) or Alexa Fluor 594 goat anti-mouse IgG (H+L) (A11032; Invitrogen) was used as the secondary antibodies. In addition, nuclear staining was performed using DAPI. The fluorescence intensity of ECMs was determined by measuring the signal intensity near the trabecular meshwork in the region of interest, and analyzed using ImageJ software (NIH, Bethesda, MD, USA). 

Transmission Electron Microscopy

The eyes of control or ATX Tg mice were enucleated, immediately cut into hemispheres and fixed in the solution containing 2% paraformaldehyde, 2.5% glutaraldehyde in 0.1M sodium phosphate buffer (pH 7.4) for 3 three hours at room temperature. The samples were then rinsed and post-fixed in the 2% osmium tetroxide in the same buffer solution on ice for two hours. They were then washed, dehydrated in the graded series of ethanol, and embedded in Epon 812 resin mixture (TAAB, Berkshire, UK). Semi-thin sections of about 0.7 µm thickness were cut on a Leica EM UC7 ultramicrotome, stained with 0.2% toluidine blue, and examined under a Nikon ECLIPSE Si microscope. Ultra-thin sections were cut, stained with uranyl acetate and lead citrate, and examined with a JEOL JEM-1400Flash electron microscope. 

OCT Measurement

Mice anesthetized via an intraperitoneal injection of 80 mg/kg ketamine and 7 mg/kg xylazine received tropicamide eye drops as treatment for mydriasis. Subsequently, posterior ocular OCT imaging was conducted using the OCT Bi-µ system (Kowa, Tokyo, Japan). Measurements of RNFL + GCL + IPL (ganglion cell complex [GCC]) and total layer thickness were taken 200 µm from the ON head manually using ImageJ software (NIH, Bethesda, MD, USA). 

RGC Counting

At three months after tamoxifen injection, the mouse eyes were initially fixed in 4% PFA at 4°C for 15 minutes. Subsequently, a circular incision was made to excise the cornea, followed by removal of the iris and lens. The eyes were then post-fixed in 4% PFA at 4°C for four hours. After fixation, posterior eye cups were prepared and immersed in 30% sucrose at 4°C overnight. The vitreous humor within these eye cups underwent digestion via treatment with hyaluronidase (Sigma-Aldrich) overnight at 37°C in a solution containing 0.1 M PBS (pH 5.3) and 2.5 mM EDTA. Visualization of the vitreous was facilitated using triamcinolone acetonide (FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan), and it was then carefully removed using tweezers and scissors. For immunostaining, a blocking solution containing 2% skim milk and 0.1% Triton-X100 in PBS was used. The primary antibody used was anti-RBPMS antibody (GeneTex Inc., Irvine, CA, USA), whereas the secondary antibody was Alexa Fluor 488 goat anti-rabbit IgG (Invitrogen). Fluorescence microscopy imaging of the central and peripheral regions of each eye quadrant was performed, positioned at 0.75 and 1.5 mm from the optic nerve hypoplasia (ONH), respectively. RBPMS-positive RGCs within these images were quantified using ImageJ software (NIH), with the researcher blinded to the sample groupings. 

Measurement of Outflow Facility

Outflow facility was determined using the two-level constant-pressure infusion technique, as described previously.^38,39 The anterior chambers underwent cannulation using borosilicate glass microneedles (1B100F-3; tip diameter, 84 µm and 1.0 mm in outer diameter, 25° angle bevel) connected to polyethylene tubing with an inner diameter of 0.35 mm and a reservoir bag. The cannulation into the eye was performed in a single, rapid insertion, and keeping the needle tip in the anterior chamber avoiding contact with the inside of the eye. The height of the reservoir bag was adjusted such that the IOP was either 15 or 25 mm Hg. IOP monitoring was achieved using a pressure transducer, as described previously.³⁸ Outflow facility (C) was calculated with the Goldman equation: C = (Fc1-Fc2)/(25-15)/10. Eyes in which the monitored values did not return to baseline when the needle was removed after measurement were excluded as the measurement was considered unsuccessful. The number of excluded eyes were as follows: before ATX Tg = 1 eye; two weeks ATX Tg = 1 eye; three-month control = 1 eye; three months ATX Tg = two eyes. 

Statistical Analysis

Data analysis was carried out using R software (version 4.1.2).⁴⁰ Results are expressed as mean ± standard deviation (SD). Box and whisker plots are presented, with the lower hinge representing the 25th percentile and the upper hinge representing the 75th percentile. The line inside the box represents the median and the cross represents the mean. When comparing two groups, a one-tailed Student's t-test was applied. To evaluate more than two groups, Dunnett's multiple comparisons test was used. In all analyses, P < 0.05 was considered statistically significant. 

Results

Expression of ATX in Mice After Tamoxifen Administration

To verify genetic modification, genotyping of anterior-eye, posterior-eye, eyelid, tail, liver, and whole-blood samples from ATX Tg mice treated with either topical tamoxifen or corn oil was performed (Fig. 1A). Bands approximately 1.6 kbp in size denoted preactivated fragments, whereas bands near 0.2 kbp represented activated fragments after excision of the segment between loxPs. High-intensity bands indicative of the active ATX form were prominent in the anterior eye and eyelid. In contrast, low-intensity bands were discerned in the posterior eye and tail. Notably, the liver and whole blood exhibited no detectable active forms. These findings imply that topical administration of tamoxifen triggered genetic modifications predominantly in tissues proximate to the eye. Using immunohistochemistry, protein expression of ATX in ocular tissues was evaluated (Fig. 1B). In control eyes, weak ATX signals were observed in the corneal epithelium, ON, ganglion cell layer, and inner and outer nucleal layers at both two weeks and three months after the ATX induction. In contrast, in ATX Tg eyes two weeks after the induction, conspicuous ATX signals were observed for the corneal epithelium, the angle region, and the ON. In addition, increased ATX signals were found in all layers of the retina, including the ganglion cell layer, relative to controls. There was no significant difference in signal strength between central retina and peripheral retina. These trends were also detected in ATX Tg eyes three months after induction, but the ATX signal increase in ATX Tg compared to control was slight for the angle region and the ON. To gauge ATX activity in the angle and retina, LysoPLD activity in the aqueous humor and combined retina + vitreous samples was quantified (Fig. 1C). The aqueous humor of ATX Tg mice exhibited significantly enhanced LysoPLD activity relative to controls (1.77-fold, P = 0.01). However, only a weak increase (1.16-fold) was observed in the vitreous and retina of these mice compared to controls (P = 0.04). Thus these observations imply that topical administration of tamoxifen to conditionally prepared ATX Tg mice evoked ATX expression mainly in the eye's anterior segment, without perturbing other systemic tissues. Moreover, ATX upregulation led to a marked surge in LysoPLD activity in the aqueous humor, although its effect was more limited in the vitreous and retina. 

Figure 1.

 

Induction of ATX expression. (A) PCR genotyping of ATX Tg mice with/without tamoxifen administration. A band of the active form (0.2 kbp) exhibited high intensity in the anterior eye and eyelid, and low intensity in the posterior eye and tail, but was absent in the liver and blood with topical tamoxifen. (B) These images showed merged images of ATX (red), and DAPI (white). Immunohistochemistry showed protein expression of ATX in the eyes of controls or ATX Tg mice two weeks or three months after topical tamoxifen treatment. Similar patterns of staining were obtained with three pairs of mice. (C) LysoPLD activity in aqueous humor (n = 8 eyes for control, n = 7 eyes for ATX Tg) and the retina + vitreous (n = 12 eyes for control, n = 8 eyes for ATX Tg). *P < 0.05, **P < 0.01 versus control.

Changes in LysoPLD Activity of the Aqueous Humor and IOP Caused by ATX Overexpression

To examine the prolonged effects of persistently augmented ATX expression, LysoPLD activity in the aqueous humor and IOP was periodically monitored for three months after topical application of tamoxifen. In ATX Tg mice, LysoPLD activity in the aqueous humor peaked at one to two weeks after tamoxifen treatment and decreased thereafter. Nevertheless, this activity consistently outpaced that of controls throughout the three-month observation period (Fig. 2). These results were consistent with the IHC for the angle area at two weeks and three months (Fig. 1B). This suggests that ATX overexpression or the resulting upregulation of lysoPLD activity continues at least in the aqueous humor until three months after the induction. IOP levels in ATX Tg mice exhibited a significant increase compared to controls at three days after tamoxifen cessation (Fig. 3). This increase of approximately 4 mm Hg above control values persisted for two months before gradually declining. In addition, to compare the baseline and IOP data at each time point, Dunnett's multiple comparisons test was used. In the control group, significantly higher IOP were observed only six to 14 days after tamoxifen administration but did not differ from baseline after two weeks of tamoxifen administration, when IOP had stabilized. In ATX Tg eyes, however, IOP was significantly higher five to 73 days after tamoxifen administration compared to before the administration. 

Figure 2.

 

LysoPLD activity in aqueous humor. Aqueous humor from ATX Tg mice with ATX expression demonstrated significantly higher lysoPLD activity than that of controls (n = 7–8 eyes for control, n = 8–10 eyes for ATX Tg). *P < 0.05, **P < 0.01, ***P < 0.001 versus control.

Figure 3.

 

IOP measurement. IOP in ATX Tg mice was significantly higher at one week after tamoxifen treatment, with an elevation of approximately 4 mm Hg that was sustained for approximately two months before a gradual decline (n = 20 eyes for control, n = 16 eyes for ATX Tg). *P < 0.05, **P < 0.01, ***P < 0.001 versus control.

Effect of Elevated Expression of ATX on the Aqueous Humor Outflow Pathway

To elucidate the underlying causes of IOP elevation, we examined the TM outflow pathway in both control and ATX Tg eyes. Initially, immunostaining and image analysis were conducted on the ECM surrounding the angle (Fig. 4AB). At two weeks after induction of ATX expression, collagen I and F-actin expression remained relatively unchanged. Conversely, higher fibronectin expression was evident compared to the control group (Fig. 4A, left panels, white arrowheads, Fig. 4B, left panels). At three months after the induction, statistical elevation of collagen I and fibronectin expression were observed. Although there was an upward trend in F-actin expression, it was not statistically significant (Fig. 4A, right panels, white arrowheads, Fig. 4B, right panels). To analyze the ultrastructural changes of conventional outflow tissues, transmission electron microscopy studies were used. Many giant vacuoles were shown in control eyes, but a few vacuoles were shown in ATX Tg eyes (Fig. 4C, arrows). After this, the aqueous humor outflow facility was quantified; before the ATX induction, 0.0113 µL/min/mm Hg 95% CI (0.0100, 0.0124) for control and µL/min/mm Hg 95% CI (0.0098, 0.0122) for ATX Tg, showing no significant difference, whereas at two weeks, the values were 0.0123 µL/min/mm Hg 95% CI (0.0103, 0.0141) for control and 0.0059 µL/min/mm Hg 95% CI (0.0042, 0.0073) for ATX Tg, which were significantly lower for ATX Tg. Similarly, at three months, the values were 0.0165 µL/min/mm Hg 95% CI (0.0146, 0.0184) for control and 0.0110 µL/min/mm Hg 95%CI (0.0087, 0.0130) for ATX Tg, with ATX Tg showing significantly lower outflow facility (Fig. 5). This reduction in outflow facility was particularly pronounced at two weeks, coinciding with the apex of LysoPLD activity and IOP elevation in ATX Tg mice compared to that seen at three months. These findings imply that after ATX induction in mice, ATX-mediated upregulation might spur fibrosis near the TM, potentially developing and contributing to increased resistance in aqueous humor outflow. This could account for the persistent IOP elevation observed. 

Figure 4.

 

(A) Collagen I expression in the angle of ATX Tg mice was not significantly increased at two weeks but exhibited a slight increase at three months. F-actin and fibronectin expression levels were elevated compared to controls at both two weeks and at three months after induction. (B) Fluorescence intensity in the trabecular meshwork region for collagen I, F-actin, and fibronectin was determined by image analysis (n = 5–6 eyes for each group). *P < 0.05, **P < 0.01 versus control. n.s., not significant. C, cornea; CB, ciliary body. (C) Ultrastructure of conventional outflow tissues in ATX Tg eyes compared with control eyes. Giant vacuoles in Schlemm's canal (SC) are indicated by black arrows. Red arrow heads indicate continuous basement membrane in ATX Tg eyes (representative images from n = 3 animals).

Figure 5.

 

Evaluation of aqueous humor outflow pathways. At three months after inducing ATX expression, ATX Tg mice exhibited notably reduced outflow facility relative to controls. Data are represented as box and whisker plots with individual data (n = 9–12 eyes for before, n = 9–10 eyes for two weeks, n = 7–8 eyes for three months). *P < 0.05, **P < 0.01, ***P < 0.001 versus control.

Evaluation of Retinal Disorders

To investigate the influence of continuous ATX expression and IOP elevation on retinal morphology, RGC layer thickness was assessed in live subjects using OCT (Fig. 6). Neither ONH structure nor retinal layer structure exhibited noticeable alterations (Figs. 6A, B). Posterior-eye OCT, a renowned clinical biomarker for RGC malfunction, is conventionally used to gauge the retinal nerve fiber layer (RNFL) thickness.^4,41 However, because of the mouse RNFL's relative thinness and the limited reliability of measurements thereof, the thickness of the GCC, which includes the RNFL, ganglion cell layer, and inner plexiform layer, was evaluated instead (Fig. 6C).⁴² After ATX induction, GCC thickness appeared consistent in both control and ATX Tg mouse eyes over a 3-month period (Fig. 6C). Furthermore, no structural changes in the ON head were evident in either group. 

Figure 6.

 

Examination of retinal thickness using OCT. Posterior OCT images acquired at up to three months after induction of ATX expression. (A) Retina and (B) ONH. (C) The thickness of the RNFL + GCL + IPL was measured at 200 µm from the ONH, representing inner retinal thickness (n = 20 eyes for control, n = 16 eyes for ATX Tg). *P < 0.05 versus control. GCL, ganglion cell layer; IPL, inner plexiform layer; INL, inner nuclear layer; OPL, outer plexiform layer; ELM, external limiting membrane; RPE, retinal pigment epithelium.

For a more comprehensive analysis of potential retinal damage, we produced retinal flat-mounts at three months after ATX induction and quantified RGCs (Figs. 7A, B). In the central region of the retina, there were no significant differences between the control and the ATX Tg mice. However, the ATX Tg group displayed a pronounced reduction in RGC counts within the retinal periphery compared to the controls. The OCT findings for the central retina aligned with the RGC counts of the flat mounts. It remains possible that although the RNFL was unquantifiable because of technical constraints, it might have exhibited a reduction in the ATX Tg group given the diminished peripheral RGC counts. Collectively, our data imply that mild but sustained IOP elevation can cause subtle RGC damage, resembling early-stage glaucomatous damage, with the peripheral retina being predominantly affected. 

Figure 7.

 

Immunofluorescence examination of whole-mounted retinas from ATX Tg mice. (A) Representative images of the RGC layer stained with RBPMS. (B) Quantitative analysis of RGCs in both central and peripheral areas revealed a significant reduction in the peripheral retina of ATX Tg mice (n = 10 mice for control, n = 8 mice for ATX Tg). *P < 0.05 versus control.

Discussion

In this study, we achieved mild IOP elevation through induction of consistently high ATX expression in mouse eyes. Previous studies have indicated that ATX can instigate cytoskeletal alterations and elevate ECM expression in human TM cells cultured in vitro. Furthermore, increased cell adhesion and barrier function have been observed in monkey Schlemm's canal endothelial cells cultured similarly, implying a role for the ATX-LPA pathway in increasing outflow resistance within the conventional outflow pathway.^26,31,43 In our research, induction of ATX expression led to increased lysoPLD activity in the aqueous humor, elevated expression of collagen I and fibronectin near the angle, and diminished aqueous humor outflow facility. Furthermore, ultrastructural evaluation by transmission electron microscopy revealed numerous giant vacuoles in control eyes but only a few vacuoles in ATX Tg eyes. It has been suggested that stiffness of Schlemm's canal endothelial cells is increased in primary open-angle glaucoma (POAG) patients,⁴⁴ impairing giant vacuole (GV) formation, and reduced density of GVs in Schlemm's canal endothelium have been reported in glaucoma patients and in animal models of glaucoma.^45–47 These observations imply that elevated lysoPLD activity in the aqueous humor may prompt cytoskeletal modifications, accelerate TM fibrosis, and potentially amplify the barrier function of Schlemm's canal, culminating in reduced aqueous humor outflow. 

During our study, IOP elevation in ATX Tg mice was markedly increased at one week after cessation of tamoxifen treatment. The increase in lysoPLD activity manifested at between one and two weeks but passed its peak after about one month. Previously, we noted that ATX and subsequent LPA induction can induce TM cell fibrosis and enhance the barrier function of cultured Schlemm's canal cells within only 24 hour. Therefore it stands to reason that ATX upregulation would rapidly influence the trabecular tissue and Schlemm's canal, increasing aqueous outflow resistance soon after tamoxifen treatment.⁴³ In human eyes, several mediators are implicated in glaucoma pathogenesis.^48–50 Notably, TGF-β2 concentrations are higher in eyes diagnosed with POAG, the most prevalent glaucoma variant.^51,52 TGF-β2 is thought to orchestrate POAG pathogenesis by regulating TM-cell fibrotic responses and ECM protein expression through canonical and alternative (noncanonical) signaling pathways.^51–53 Earlier, we documented differential upregulation of ATX and TGF-β2 levels in glaucomatous aqueous humor based on glaucoma subtypes, implying interplay between TGF-β and ATX.^54,55 In those studies, we found that trans-signaling between TGF-β2 and ATX could mutually regulate their expression. In the current study, ATX expression precipitated IOP elevation, outflow resistance reduction, and fibrosis in the angle area, persisting for up to three months. However, over time, IOP elevation and reduced outflow resistance diminished, even though angle fibrosis remained evident after three months. It is possible that secondary changes prompted by trans-signaling from other mediators may modulate ATX expression over this period; further investigation is necessary. 

Quantitative analysis demonstrated a significant reduction in the numbers of RGCs in the peripheral retina of the ATX Tg mice, whereas no substantial decline was observed in the central region. Such peripheral RGC degeneration has been noted in several mouse models of glaucoma, originating in the periphery.^8,56 Given the mild IOP elevation in our model to a level often observed in clinically identified normal-tension glaucoma, we might be able to discern more pronounced retinal damage with an extended timeframe. This damage could manifest as a decrease in inner-retina layer thickness and a lower RGC count in the central retina. 

So far, several glaucoma mouse models have been developed, although each of which has advantages and challenges. The most relevant and ideal models would be recapitulated as the mimic of human glaucomatous condition with similar phenotypes within the conventional outflow pathway including damage in the TM and SC which generates elevated IOP and outflow dysfunction leading to RGC and ON damage. The existing successful mouse models include genetically mice or models rely on inhibition of aqueous humor outflow from the angle using retention of substances in the angle such as microbeads or fibroblasts, however these models still possess challenges.^8,9,13,15 They often produce excessively strong temporary IOP elevation or struggle to maintain a consistently elevated IOP over extended periods. These models barely replicate the pathogenesis seen in clinical glaucoma. Furthermore, damage to the anterior segment, such as angle closure or inflammation caused by invasive procedures, complicates observations of the posterior segment. The DBA/2J mouse, a popular genetically modified model of glaucoma, experiences increased IOP because of iris pigment dispersion. However, this model has its drawbacks: IOP elevation can take up to 12 months to manifest, and the presence of corneal calcification makes reliable IOP measurement and evaluation of retinal damage challenging. Additionally, these mice often exhibit significant systemic complications.^12,57,58 More importantly, the DBA/2J mouse is more representative of secondary glaucoma, which is accompanied by ocular inflammation, rather than of POAG. Other genetically modified models of open-angle glaucoma have been reported, including mutant myocilin and TGF-β2 expressing mice.^13,56,59 Mutant myocilin-expressing mice overcome some of the same barriers as above. It has been reported that mutations in the myocilin gene are found in 3-4% of patients with POAG⁶⁰, and this model could be very useful in the study of several glaucoma related to genetic component, including Y437H myocilin.⁵⁶ As for the TGF-β2-over expressing mice, it can be said that the pathological changes of open-angle glaucoma are mimicked more closely in this model. TGF-β2–expressing mice have been shown to have long-term elevated IOP with Ad5 injection, and ECM accumulation in the angle and retinal damage have been investigated, which mimic human glaucoma. However, some reports indicate that the increase is only about one month after administration, suggesting that there still exists an issue with the stable sustained increase in IOP.⁵⁹ In addition, increased overall corneal thickness and anterior segment inflammation have been indicated as challenges in this model.^13,59 

In comparison, our mouse model offers several advantages. First, we observed mild and consistent IOP elevation that lasted at least 3 months, correlating with initial RGC damage in the peripheral retina. This model allows for exploration of primary changes in the RGC axon and cell body, because the chronic mild IOP elevation is similar to that seen in human OAG. Second, the structures of the cornea and lens remained unaffected and intact throughout our study, facilitating in vivo evaluations of structural and functional changes using OCT and electroretinography. Third, given the defined pathology of IOP elevation in this OH model, it is especially beneficial for understanding the mechanisms behind both current and future IOP-lowering medications. Fourth, our model enables the study of pathophysiological changes in the outflow pathways of the open angle, which includes both trabecular and uveoscleral pathways, in response to long-term accumulation of ECM and OH. This aspect is notably advantageous, distinguishing our model from previous animal glaucoma models that often inflicted mechanical or inflammatory damage on the angle tissue. In addition, animal models have already been reported using transgenes that are thought to be directly related to or downstream of ATX signals such as TGF-β2 and fibronectin extra domain A.^15,59 The ATX Tg mouse can also be a useful research tool to study the interrelation ship between these transgenes and ATX signaling. 

This study was not without its limitations. First, although the IOP elevation observed in our mouse model was both mild and stable, there was a noticeable decline in IOP after two months. This could be attributed to the diminishing lysoPLD activity in the aqueous humor, which reached its zenith at between one and two weeks. Such a decrease may have resulted from a feedback mechanism, although no detailed analysis confirmed this hypothesis. Another plausible explanation could be suboptimal recombination efficiency, as depicted in Figure 1, or the modest degree of fibrosis observed in the TM. A more in-depth exploration is warranted to elucidate the alterations that occur in the aqueous humor's chemical mediators over time. A method to maintain higher IOP consistently, perhaps by optimizing the efficiency of genetic modifications to achieve earlier initiation of expression, or by effecting alternative modifications, could greatly enhance the model's value and applicability. Second, induction of ATX was also identified in the corneal epithelium, retina, and ON. This might have influenced RGC death and the accuracy of IOP measurements; comprehensive assessment of this possibility is required. Previous research has established that ocular expression of ATX is distributed across the retina, vitreous, ciliary body, and retinal pigment epithelial cells.^61–63 Consequently, genotyping results combined with lysoPLD activity measurements imply that significant ATX induction occurs primarily in the anterior eye and eyelid (Fig. 1). The influence of ATX induction in the posterior eye might be limited at the systemic level. However, we cannot completely exclude the possibility that ATX overexpression in RGCs/ON may induce damage, apart from the effect of IOP elevation. In this study, a decrease in RGC number was observed only in the peripheral retina (Fig. 7), even though the increase in ATX expression in ATX Tg did not differ between the central and peripheral retina in IHC (Fig. 1B). This suggests that the damage of RGCs in this study is not a direct effect of ATX overexpression in the retina. In vitro analysis of RGCs/ON damage caused by ATX overexpression in RGCs or ATX exposure to RGCs would be useful for further evaluation of possibility of ATX-induced RGCs/ON damage. 

In conclusion, we have developed a mouse model for open-angle OH. By inducing ATX expression in the eye, fibrosis of the outflow pathway and sustained IOP elevation are achieved. This mouse model of POAG is also characterized by chronic RGC death, underscoring its significance for research into glaucomatous optic neuropathy and potential neuroprotective treatments. 

Acknowledgments

The authors thank Kuniyuki Kano and Junken Aoki for generously providing genetically modified mice needed for our experiments and for their support of our research. The authors also thank Hiroshi Sagara and Yuji Watanabe (Institute of Medical Science, the University of Tokyo) for the supports of transmission electron microscopy observations. 

Supported by grants from the Japan Society for the Promotion of Science (JSPS-KAKENHI 22K09807 and 23H03058). The English in this document has been checked by at least two professional editors, both native speakers of English. For a certificate, please see: http://www.textcheck.com/certificate/d01fqG. 

Disclosure: S. Shimizu, Senju Pharmaceutical Co., Ltd. (E); M. Honjo, None; M. Liu, None; M. Aihara, None