A Mouse Model of Chronic Ocular Hypertension Induced by Circumlimbal Suture

Abstract

Purpose: To develop a chronic ocular hypertension mouse model by inducing intraocular pressure (IOP) elevation using a suture technique previously developed for rats.

Methods: C57BL/6 mice were given monocular circumlimbal suture (10/0) placement under anesthesia (ketamine/xylazine). The suture was left in place for 12 weeks (n = 10). A control group had the same treatment while it was removed after day 1 (n = 10). Intraocular pressure, electroretinogram (ERG), and optical coherence tomography (OCT) were measured in both groups for 12 weeks. At week 12, animals were euthanized and retina was harvested for histologic assessment.

Results: In the group with extended suture placement, IOP spiked from 14.3 ± 0.9 to 48.4 ± 4.8 mm Hg immediately after suture implantation. At day 1, it declined to 33.2 ± 3.7 mm Hg and remained elevated for 12 weeks. In the suture removal group, IOP normalized to baseline within a day following suture removal. In the group with prolonged IOP elevation, the retinal ganglion cell (RGC) mediated ERG responses continued to exacerbate and were significantly reduced by week 12 (P < 0.001). Progressive loss of retinal nerve fiber layer was found and it became significant from week 4 (P < 0.001). At week 12, significant loss of RGCs (P < 0.001) was noted. In the IOP normalization group, no alteration in ERG, OCT, and RGC counting was observed.

Conclusions: The circumlimbal suture approach produces a mild chronic IOP elevation in mice. Functional and structural changes under model induction are largely independent of the initial IOP spike.

Glaucoma is a chronic ocular disease, which is characterized by selective retinal ganglion cell (RGC) degeneration and progressive visual field loss. Elevated intraocular pressure (IOP) is a critical risk factor for disease development, and IOP lowering is the mainstay of disease management.^1–3 Over the past few decades, a variety of rodent models have been developed to simulate human glaucoma.^4–7 These models aim to induce either acute or chronic ocular hypertension leading to glaucomatous-like changes. By using such models, investigators have made significant insights into understanding the pathogenesis of the disease. 

Previously, a rat model of chronic IOP elevation using conjunctival circumlimbal suture was developed.^8,9 This minimally invasive technique employed the concept of oculopression to produce elevated IOP.^10–12 An advantage of the suture compression model is that it is not associated with ocular tissue penetration or traumatization, and therefore relatively preserves the immune privilege of the eye. It was reported that a mild, chronic IOP elevation was induced for at least 15 weeks following an initial IOP spike that appeared immediately after suture placement. Progressive loss of RGC function, retinal nerve fiber layer (RNFL) defect, and eventual cell death in the ganglion cell layer were observed under model induction, without evident complication. Moreover, the anterior chamber angle and depth were not affected by the circumlimbal suture compression. 

Although this model produced clinical features resembling human glaucoma, a limitation is the involvement of an initial IOP spike. Indeed, extremely high IOP challenge may cause acute retinal ischemia leading to permanent cell injury,^13–17 which may confound the effect of subsequent mild, chronic IOP insult in the suture model. However, a recent study by Crowston et al.¹⁸ showed that an acute IOP elevation (50 mm Hg) for 30 minutes induced reversible RGC dysfunction and minimal RGC death within 7 days in young healthy mice but that this was not the case for older mice. These results demonstrate that this magnitude and duration of IOP challenge did not cause permanent inner retinal cell damage in young mice. However, the critical IOP exposure that will induce permanent cell injury, analogous to an acute, ischemia model, needs further investigation. 

The aim of the current study is to develop a new, chronic, ocular hypertension model in young healthy mice, by replicating the previously published rat suture model of chronic ocular hypertension.⁸ The conjunctival circumlimbal suture was left in place to produce sustained IOP elevation over 12 weeks. To tease out the ambiguous role of the initial IOP spike, a second group of age-matched mice had the same suture implantation, but the suture was removed after the first day, reducing the IOP back to normal levels. This control group was used to reveal the specific impact of the initial IOP spike on the retinal cell physiology. Animals in this study exhibiting an IOP spike greater than 55 mm Hg were excluded from analysis, as it was likely that this level of IOP would cause acute ischemic injury in the retina and other potential unwanted side effects. In addition to IOP measurement, retinal function and structure were monitored using electroretinogram (ERG) and optical coherence tomography (OCT) respectively at baseline, week 2, 4, 8, and 12. At the end of the experiment, retinal tissues were harvested for RGC quantification. 

Materials and Methods

Animals

All animals were treated in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research, and all procedures were approved by the Animal Care and Use Committee of University of California, Berkeley. Twenty-seven 6-week-old male C57BL/6 mice were acquired from the Jackson Laboratory (Bar Harbor, ME, USA). Before experimentation, all mice were allowed to acclimatize to the housing facility for 2 weeks. Food and water were available ad libitum. 

IOP Induction and Measurement

All IOP measurements in this study were conducted in awake animals using a Tonolab rebound tonometer (iCare, Helsinki, Finland) without topical anesthesia except for the reading recorded immediately after suture implantation. During IOP measurement, awake mice were restrained by the scruff of the neck without applying pressure to the eyelid. Baseline IOP was measured 1 day before IOP induction. Following baseline IOP measurement, mice were anesthetized with a mixture of ketamine/xylazine (100 mg/kg and 10 mg/kg, respectively, intraperitoneal) on a veterinary heating pad. One drop of proparacaine hydrochloride (0.5%, Akorn, Lake Forest, IL, USA) was added to the cornea of a randomly selected eye for topical anesthesia. The suture procedure was performed as described previously in rats, but with a finer suture.⁸ A circumlimbal 10/0 suture (nylon, Fine Science Tools, Foster City, CA, USA) was secured around the equator of the eye, 0.4 to 0.5 mm behind the limbus, by five anchor points and double knots on the conjunctiva. In the process of knot tying, we stopped tightening the suture as soon as the loop was completely closed to form a knot. The residual suture was trimmed from the knots to reduce potential irritation. The suture only pierced through the subconjunctival space and therefore did not puncture the sclera. Suture compression on the major episcleral veins was avoided by piercing through the deeper subconjunctival space underneath those vessels. Figure 1 demonstrates the schematic (A), side (B), and top (C) views of the circumlimbal suture. In Figure 1B, an acute reversible cataract was observed as previously reported following ketamine/xylazine anesthesia, and fully disappeared within 24 hours.¹⁹ Following suture implantation, antibiotic ointment (Ak-Poly-Bac, Akorn) was applied to the eye. The fellow control eye was untreated and served as a within animal control. Lubricant eye gel (GenTeal, Alcon, Fort Worth, TX, USA) was used for the control eye during the suture procedure to maintain corneal moisture. In one group of mice, the suture was left in place for 12 weeks (n = 10). A second group of mice had suture removal after the day 1 IOP measurement, by simply cutting the suture and removing from the eye (n = 10). Intraocular pressure was measured immediately, 3 hours and 1 day after suture for both groups. After the first day, it was monitored every 3 days throughout the experimental period (10 AM to 12 PM under the same lighting conditions). An extra IOP recording at day 2 was made for the suture removal group to confirm that they had reverted to baseline. 

Figure 1

 

The illustration of circumlimbal suture in the mouse eye (A). A nylon 10/0 suture was placed on the conjunctiva, 0.4 to 0.5 mm behind the limbus and secured by five anchor points and two conjugate knots to compress the eyeball. The side view (B) and top view (C) of the suture.

Electroretinogram

Prior to ERG measurements, all mice were dark-adapted over night (>12 hours). Animals were anesthetized with ketamine/xylazine (as described previously) on a heating pad. A drop of proparacaine hydrochloride was applied for corneal anesthesia. A drop of tropicamide (0.5%, Akorn) and phenylephrine (2.5%, Paragon BioTeck, Portland, OR, USA) were used for pupil dilation. The full-field flash ERG was recorded with a VERIS system (Electro-Diagnostic Imaging, Redwood City, CA, USA) under dim red light. The electroretinogram stimulus intensity ranged from −5.90 to 2.25 log cd.s.m⁻² and custom-made chlorided silver corneal (positive) and scleral ring (negative) electrodes were used as described previously.⁸ A subdermal needle electrode was inserted in the tail to serve as ground. Signals were acquired from both eyes simultaneously. The positive scotopic threshold response (pSTR) was elicited with an intensity of −4.60 log cd.s.m⁻² (average of 20 repeats with an interstimulus interval of 2 seconds). The positive scotopic threshold response was reported to be RGC dominant and used in studies of IOP elevation in rodents.^8,20–22 It was quantified by taking the amplitude at a fixed time of 110 ms after the stimulus onset. The a-wave and b-wave were elicited using an intensity of 2.25 log cd.s.m⁻² without repetition. Modeling the a-wave with a delayed-Gaussian function (P3) returns the photoreceptor response (Rm_P3, μV).^23,24 To quantify the bipolar cell response, P3 was firstly removed from the raw waveform to return P2. The P2 peak response was then plotted as a function of intensity fit with a hyperbolic curve to return the maximum bipolar cell response (Vmax, μV).^25,26 Rm_P3 and Vmax were used for the photoreceptor and bipolar cell response, respectively. 

Optical Coherence Tomography

Following ERG measurements, animals were immediately transferred for OCT imaging under the same mydriasis and anesthesia. A spectral domain OCT (Bioptigen, Durham, NC, USA) was used to capture retinal cross-section images. The operation and scanning protocols were detailed previously.^27,28 Briefly, a 3 × 3 mm rectangular scanning sequence yielded a single en-face image of the retina with a customized mouse lens (50° field of view), which consisted of 100 B-scan images. Each B-scan comprised 1000 A-scans. The optic nerve head was centralized in the en-face image to allow B-scan image analysis at this location. Lubricant eye drops (Systane Ultra, Alcon) were used during imaging to prevent corneal desiccation. The B-scan images were analyzed with the InVivoVue Clinic software. Retinal layer thickness was measured using the manual “caliper” function in this software. The total retinal thickness was defined as the length from the RNFL to retinal pigment epithelium (RPE) with both layers included. Retinal nerve fiber layer and total retinal thicknesses were measured in four quadrants (nasal, temporal, superior, and inferior) of the en-face image by masked observers. Within each quadrant, three measurements (400, 500, and 600 μm from the center of the optic nerve head) were performed. Therefore, a total of 12 readings were averaged to return a single thickness value. 

Retinal Flat Mount and Quantification of RGCs

At week 12 after OCT measurements, animals were euthanized by carbon dioxide inhalation and cervical dislocation. Eyes were enucleated with retention of partial nasal and superior rectus muscles for orientation. The eye was fixed with 4% paraformaldehyde for 15 minutes at room temperature (RT) and then washed in phosphate-buffered saline (PBS) for 10 minutes. The dissection procedure was carried out as reported.²⁹ Briefly, an incision on the edge of cornea was made by a sharp blade and the eye was bisected. The cup-shaped retina was isolated and rinsed in PBS. It was then flattened by four radial cuts. After removal of residual PBS, the retina was fixed with methanol at −20°C overnight to facilitate permeabilization. Prior to immunostaining, it was rinsed with PBS at RT (30 minutes, three times) and incubated with a primary antibody, goat anti-Brn3a (brain-specific homeobox/POU domain protein 3A, Santa Cruz Biotechnology, Santa Cruz, CA, USA) at 4°C overnight with gentle shaking, diluted in blocking solution (1:100, PBS, 2% bovine serum albumin, 2% Triton). The retina was rinsed with PBS (30 minutes, three times) and incubated with a secondary antibody (Alexa Fluor donkey anti-goat IgG, Jackson ImmunoResearch Laboratories, West Grove, PA, USA) at RT for 2 hours, diluted (1:200) in blocking solution. It was then rinsed with PBS (30 minutes, three times), mounted on slides with the vitreal side up and coverslipped with an antifading medium (Prolong Gold, Invitrogen, Carlsbad, CA, USA). All samples were examined using a Zeiss Axioplan 2 Imaging epifluorescent microscope system (Carl Zeiss, Oberkochen, Germany). Brn3a is commonly used as a biomarker of RGCs, and it must be acknowledged that it can detect only 80% to 90% of RGCs and furthermore is likely to be down-regulated following the ocular hypertension insult.^30,31 The protocol for Brn3a-labeled RGC counting was adopted as previously described.³² Two equal-sized areas (450 × 320 μm, 20× magnification, 850 μm from the center of the optic nerve head) within each quadrant of the retina were imaged, and hence, eight areas were included for each retina (∼6.2% of retinal area). All images were transferred to ImageJ (National Institutes of Health, Bethesda, MD, USA), and the cell number in each area was determined in a masked manner, using the particle analysis function. Cell counting from all areas were averaged to return a reading for that retina. 

Statistical Analysis

Data analysis and statistics were performed using Prism 6 (GraphPad Software, La Jolla, CA, USA). Data were expressed as mean ± SD. For intraocular pressure, ERG, and OCT data, a 2-way repeated measures (RM) ANOVA was used with a Bonferroni post hoc test. For retinal ganglion cell counting data, a paired t-test was used. A value of 0.05 was set as a statistically significant level. 

Results

Figure 2 shows the IOP profile in the two groups. In the group with the suture left in place for 12 weeks (Fig. 2A, n = 10), IOP was elevated from 14.3 ± 0.9 to 48.4 ± 4.8 mm Hg immediately after suture placement. After 3 hours, it decreased to 38.9 ± 4.1 mm Hg and at week 1, it further decreased to 26.5 ± 1.8 mm Hg. In general, IOP after suture placement was gradually reduced over 12 weeks. At week 12, it was still significantly higher than the contralateral control eyes (17.9 ± 1.7 mm Hg versus 15.2 ± 0.9 mm Hg, P = 0.003, RM 2-way ANOVA with a Bonferroni post hoc analysis). In the second group of mice (FIg. 2B, n = 10), IOP was immediately increased from 14.4 ± 1.1 to 50.5 ± 3.6 mm Hg and reduced to 40.5 ± 5.5 mm Hg after 3 hours. These are similar to the data reported in Figure 2A. Following suture removal, IOP was reduced from 33.8 ± 5.1 (day 1) to 14.2 ± 1.4 mm Hg (day 2). The level of IOP at day 2 was similar to that of the contralateral control eyes (14.4 ± 0.5 mm Hg, P = 0.99), and subsequently it remained normalized to week 12. The intraocular pressure profile of control eyes was similar between the two groups (P = 0.22) and no elevation was found. The inset table in each panel shows the IOP of the sutured eyes at specific time points. Seven of initial 27 mice (26%) were excluded due to IOP spike over 55 mm Hg. 

Figure 2

 

Intraocular pressure profile of the two groups. The circumlimbal suture produced a mild, chronic IOP elevation following the initial spike (A). With suture removal after the first day, IOP recovered to normal level within 1 day (B). The inset table shows the IOP at specific time points. Mean ± SD, n = 10 for both groups.

Figure 3 demonstrates the effect of IOP elevation on retinal function as measured by the ERG. The average waveforms in Figure 3A show both outer (a-wave for photoreceptor; b-wave for bipolar cell) and inner (pSTR for ganglion cell) retinal function for both groups at week 12 (n = 10 for each group). For mice with prolonged IOP elevation, reduced waveform amplitudes were seen for all retinal cell classes. However, in mice with IOP normalization, there was no difference in waveform amplitudes between the two eyes. Figures 3B and 3C show the quantification of waveform amplitudes across 12 weeks, and the data were expressed as a percentage difference between the two eyes. In mice with prolonged IOP elevation (Fig. 3B), there was nonspecific functional deficit at week 2 (photoreceptor −13.7 ± 3.6%; bipolar cell −13.9 ± 3.1%; ganglion cell −16.2 ± 2.3%). From week 4 to week 12, photoreceptor and bipolar cell functions were stable, but ganglion cell function continued to deteriorate. More specifically, there was progressive and preferential loss of ganglion cell function and it became significant from week 4 when compared to photoreceptor and bipolar cell function. At week 12, ganglion cell function was significantly reduced by −26.0 ± 5.0% compared to −13.6 ± 3.6% for photoreceptor function (P < 0.001, RM 2-way ANOVA with a Bonferroni post hoc analysis) and −13.9 ± 2.9% for bipolar cell function (P < 0.001). In animals with IOP normalization (Fig. 3C), there was no significant difference among all retinal cell functions across 12 weeks. At week 12, ganglion cell dysfunction was −2.7 ± 2.0% compared to −1.0 ± 2.7% for photoreceptor (P = 0.44) and −0.8 ± 0.6% for bipolar cell (P = 0.30). In addition, there was no significant difference from baseline (P > 0.05 for all). 

Figure 3

 

Electroretinogram results. (A) Average waveforms at week 12 for the two groups. Scotopic a/b-waves were elicited using a stimulus intensity of 2.25 log cd.s.m⁻² and −4.60 log cd.s.m⁻² for pSTR. (B) Significant preferential loss of ganglion cell function was observed (compared to photoreceptor and bipolar cell dysfunction) from week 4 with IOP elevation for 12 weeks. However, with IOP normalization after day 1, there was no significant dysfunction of all retinal cells (C). Asterisks denote P < 0.001, mean ± SD, n = 10 for each group.

Retinal structural changes measured by OCT are shown in Figure 4 (n = 10 for each group). In Figure 4A, the yellow crosses on the en-face image show the locations for retinal layer measurements as described previously (see Materials and Methods). Figure 4B shows the B-scan image for RNFL and total retinal thickness measurements. A progressive loss of RNFL was found in eyes with prolonged IOP elevation (Fig. 4C). This structural defect became significant from week 4. At week 12, the RNFL thickness of sutured eyes was 15.0 ± 2.4 μm, which was significantly thinner than the control eyes (20.9 ± 3.0 μm, −28.2 ± 2.5%, P < 0.001, RM 2-way ANOVA with a Bonferroni post hoc analysis). A similar result was observed for total retinal thickness. In sutured eyes, the total retinal thickness was 197.6 ± 4.9 μm at week 12 compared to 203.6 ± 4.7 μm in control eyes (−3.0 ± 0.5%, P < 0.001). However, in the IOP normalization group (Fig. 4D), no change was found for RNFL or total retinal thickness across the 12 weeks. At week 12, the RNFL thickness was 20.6 ± 2.8 μm in sutured eyes and 20.7 ± 3.0 μm for control eyes (P = 0.09). The total retinal thickness was 203.3 ± 7.4 μm in sutured eyes and 203.3 ± 7.3 μm in control eyes (P = 0.99). 

Figure 4

 

Optical coherence tomography results. (A) The retinal layer thickness was measured at the location of the 12 yellow crosses (400, 500, and 600 μm from the center of optic nerve head in four quadrants) on the en-face image of the retina and averaged. (B) Retinal nerve fiber layer (RNFL) and total retinal (RNFL to RPE) thicknesses were measured. (C) Significant RNFL defect was found from week 4 in eyes with IOP elevation for 12 weeks and the same result was observed for total retinal thickness. (D) No difference in both RNFL and total retinal thicknesses was noted across 12 weeks in eyes with IOP normalization after day 1. Asterisks denote P < 0.001, error bars denote SD, n = 10 in each group.

At the end of week 12, the retina was collected for RGC density evaluation (Fig. 5, n = 10 for both groups). Figure 5A shows the locations and size of the regions on the retinal flat mount for cell counting. Representative images with labeled RGCs from both eyes are shown in Figure 5B for the prolonged IOP elevation group and Figure 5C for the IOP normalization group. The average cell density was significantly reduced in the prolonged IOP elevation group (Fig. 5D, control eyes 3088 ± 193 versus sutured eyes 2516 ± 374 RGC/mm², −18.6 ± 10.6%, P < 0.001). In the IOP normalization group, no difference was found between the two eyes (control eyes 3098 ± 189 versus sutured eyes 3073 ± 175 RGC/mm², P = 0.32). 

Figure 5

 

Cell counting results at the end of 12 weeks. (A) The region of interest in each quadrant on the retinal flat mount (450 × 320 μm, 850 μm from the center of the optic nerve head, 20× magnification) was photographed for RGC counting. Representative retinal flat mount images with Brn3a-labeled RGCs from both eyes in the mice with IOP elevation for 12 weeks (B) and IOP normalization after day 1 (C). (D) Retinal ganglion cell density was compared between the two eyes in both groups. C, control eye; S, sutured eye. Asterisk denotes P < 0.001, error bars denote SD, n = 10 in each group.

Figure 6 shows the linear correlation between OCT, ERG, cell counting results (all at week 12), IOP peak and IOP integral (12 weeks) in the prolonged IOP elevation group. A strong correlation was found between RNFL thickness and RGC (pSTR) amplitude (Fig. 6A, r² = 0.85, P < 0.001). A strong correlation was also found between RGC density and ERG amplitude (Fig 6B, r² = 0.82, P < 0.001). Additionally, RGC density was strongly correlated with RNFL thickness (Fig. 6C, r² = 0.76, P < 0.001). Figures 6D through 6F show the correlation between those parameters and IOP peak. No significant correlation was found in all of them (P > 0.05). On the contrary, there were strong correlations when compared to IOP integral (Figs. 6G–I, P < 0.01). 

Figure 6

 

The linear correlation between RNFL thickness (OCT at week 12), RGC amplitude (pSTR, ERG, at week 12), RGC counting results, IOP peak, and IOP integral in the mice with IOP elevation for 12 weeks.

Discussion

In the present study, we report a mouse model of chronic IOP elevation using the circumlimbal suture technique previously developed and optimized for rats.⁸ In general, results seen in rats can be reproduced in mice. Oculopression by circumlimbal suture produces prolonged and stable IOP elevation for at least 12 weeks, following an initial IOP spike. Furthermore, elevated IOP can be completely normalized within a day by cutting the suture and releasing the compression. After 4 weeks of elevated IOP, preferential inner retinal dysfunction manifested by a significant decrease in the RGC related ERG parameter, pSTR. Associated with this functional deficit, both the RNFL and total retinal thicknesses are progressively reduced and become significant from week 4. The structural defect of the RNFL is correlated with the functional deficit of the RGCs. At the end of the experimentation (week 12), a significant loss of Brn3a-labeled RGCs was found and it was correlated with both the functional (ERG) and structural (OCT) loss. Unlike the group with extended IOP elevation, when the IOP was normalized 1 day after the initial IOP spike, there was no evident functional, structural, or histologic deficit in the retina. 

To verify the applicability of a model for better understanding the events of chronic ocular hypertension, it is critical to examine the model induction, in particular, whether it is driven by chronic or acute IOP change. Given an initial IOP spike is present in this model (∼50 mm Hg), we added a second control group with no subsequent, postspike IOP elevation and measured the ERG, OCT, and histologic alterations over time. This was not previously reported in the rat model. The lack of progressive degeneration suggests that chronic, mild, ocular hypertension plays a key role in this model. Moreover, the correlation data (Fig. 6) also indicates that functional (ERG) and structural (OCT and cell counting) changes are significantly associated with IOP integral rather than IOP peak. The intraocular pressure spike appears to have little impact in this model, a result that is consistent with previous studies. It had been reported that in 3-month-old C57BL/6 mice with IOP elevation of 50 mm Hg for 30 minutes, near-complete recovery of ERG responses from RGCs, bipolar cells, and photoreceptors was found 7 days following challenge.¹⁶ Although the duration may be different, the level of IOP is similar in our study (50.5 ± 3.6 mm Hg, Fig. 2B). Moreover, this study also showed that no significant reduction in retinal blood flow at IOP levels lower than 70 mm Hg was found.¹⁶ The same findings are reported in rats.^15,33 Therefore, it is reasonable to assume that there will be no significant retinal ischemia following suture placement in our model. Although relatively minor by itself, it is still possible that the initial IOP spike plays a role in facilitating retinal injury caused by the subsequent chronic IOP elevation. For example, scleral biomechanical properties may be altered by the initial IOP spike.^34,35 This effect may make the eye more vulnerable following insult. Further investigation is required to understand the role of the IOP spike in this model. However, it is likely that similar spikes are induced in other commonly used “chronic” hypertension models. While no apparent functional deficit is found in our control group, there was acute nonspecific loss of all ERG parameters at week 2 in the group with extended suture compression (Fig. 3B). This phenomenon may arise from compromised choroidal blood flow.⁸ It is speculated that the suture compression may have resulted in stasis in the intrascleral veins and scleral venous plexus leading to reduced blood flow in the choroid. More importantly, despite acute functional loss at week 2, photoreceptor and bipolar cell functions maintain stability and do not further exacerbate after week 2, suggesting that the choroidal blood flow impairment is not substantial. Furthermore, OCT findings (Fig. 4C) indicated that the loss of the RNFL could account for the entire loss of the retina. This result implies that outer retinal layers remain intact in the animals with prolonged IOP elevation. 

The exact mechanism of IOP elevation induced by suture compression is uncertain. First of all, in the rat model, comparison between hydrostatic (by anterior chamber cannulation) and rebound tonometer IOP measurements had confirmed that, following suture placement, IOP readings by the rebound Tonolab reflected the true IOP.⁸ This comparison was not conducted in our study. However, we were able to measure changes of central corneal thickness and corneal curvature prior to and following suture placement using OCT (n = 4). As shown in the Supplementary Figure S1, the baseline group was with suture but with the knot loosely tied (Supplementary Fig. S1A). In the same eye, anterior chamber OCT was measured again immediately after the knot was tightened (Supplementary Fig. S1B). Supplementary Fig. S1C was at 2 weeks after suture placement. The data showed that there was no difference in central corneal thickness across the three time points (see Supplementary Table S1, P > 0.05, 1-way RM ANOVA). The same result was found for the corneal curvature. Given these two unchanged parameters, it is reasonable to assume that the rebound Tonolab would reflect the true changes in IOP. Given suture removal led to IOP reduction within 24 hours, it implies that ocular hypertension is not associated with chronic pathologic responses, such as inflammation of the anterior chamber outflow pathways.^36–38 In addition, the data in the rat suture model indicated that the anterior chamber angle is not affected by the suture compression.⁸ Furthermore, it is unlikely that we would have affected aqueous formation as it is associated with active transport mechanisms from the blood supply to the anterior eye structures.³⁹ Given that enhancement of this blood flow following suture compression is unlikely, increased aqueous formation leading to elevated IOP is also unlikely. In contrast, if the suture compression impairs this blood flow, we should observe a reduced aqueous production and would expect a lower IOP. Aqueous outflow from the anterior chamber involves two pathways, the conventional trabecular meshwork pathway and uveoscleral outflow pathway.⁴⁰ Aqueous outflow through the trabecular meshwork is unlikely to be impaired in that the anterior chamber angle was previously reported to be normal.⁸ Impairment of uveoscleral outflow has also been significantly associated with IOP elevation in a rat model of congenital glaucoma.⁴¹ Given the position of the suture, compression of the iris root and ciliary muscle as well as the choroidal and scleral space is likely. We propose that such compression would alter the normal physiology of these tissues. Compression likely compromises uveoscleral outflow facility. More specifically, reduced spacing in the ciliary muscle due to compression would lead to reduced outflow dynamics.⁴² In addition, compression of the ciliary muscle may also lead to reduced aqueous outflow as a result of compromised aqueous pumping capacity by the scleral spur.⁴³ Another element to determine IOP is the episcleral venous pressure. Intrascleral veins and the episcleral plexus may also have been compressed leading to elevated episcleral venous pressure.⁴⁴ Taken together, possible contributing factors include reduced aqueous outflow facility (mainly the uveoscleral pathway and trabecular meshwork pumping) and increased episcleral venous pressure. Further investigations are required to confirm these hypotheses, as techniques for the evaluation of aqueous humor dynamics in the mouse eye have been documented in recent studies.^45,46 It is also worth noting that the mechanism in this model may be similar to that leading to transient IOP elevation in patients following scleral buckling surgery. It is thought that the buckle obstructs venous outflow resulting in engorgement of the ciliary body. This change can cause the ciliary body to rotate anteriorly affecting the angle and thus causing IOP elevation. Given the different anatomy between human and mouse eyes, this etiology may be more pronounced in mice. 

There are advantages in the circumlimbal suture model. (1) It is less invasive as the immune privilege of the eye is not compromised during model induction. (2) Only a single intervention is needed to produce a sustained IOP elevation. Taken these two points together, some IOP models require multiple injections of foreign agents into the anterior chamber⁴⁷ or a single injection of hypertonic saline into aqueous drainage veins.⁴⁸ Laser models involve photocoagulation of the trabecular meshwork⁴⁹ and other models require cauterization of the episcleral veins.⁵⁰ These models are more invasive and in each case breach the integrity of the eye's immune privilege. Potential complications in these models such as trauma and inflammation may confound the effects of high IOP. (3) Another advantage of the circumlimbal suture model is the presence of clear optical media, which is of particular importance for noninvasive assessment of retinal function by the ERG and structure by the OCT. In the intracameral injection model, the induction procedure in the anterior chamber may compromise the clarity of the cornea and lens and hence reduce the capacity for ERG and OCT assessments. (4) While not performed in the study reported, we found that it was possible to control suture tension with an assistant continuously measuring instantaneous IOP during the knot tying process. It will provide information to help determine the termination of the knot tying process and achieve a desired IOP level. 

Although several advantages are shown in this model, some limitations are also found. (1) A limitation of the circumlimbal suture technique is that the level of IOP spike must be taken into careful consideration. In our pilot study (data not shown), age-matched mice (8 weeks) with extremely high IOP spikes (> 65 mm Hg) tended to develop severe complications within 2 to 3 weeks, such as hyphema, cataract, and tissue necrosis. This finding is not seen in the rat model. The rat eyes appear to have more robust tolerance for high IOP insult than the mouse eyes, possibly due to the eye size. In the mice reported in the present study with IOP spikes below 55 mm Hg (20 of total 27 mice), no marked complications were noted. (2) The location of the circumlimbal suture placement may determine the magnitude of the initial IOP spike. In the pilot study, we found that suture location further away from the limbus tended to induce higher IOP spikes (∼0.8 mm behind the limbus, IOP spikes ranging from 65 to 79 mm Hg, n = 4). The mechanism of this finding remains unclear. It is possible that “further back” suture compression physically pushes the lens. It moves forward and pushes the iris against the trabecular meshwork causing angle closure and a shallower anterior chamber. Therefore, when utilizing this model in mice, it is essential to avoid IOP spikes greater than 55 mm Hg, and careful suture placement is critical. (3) Another limitation is that there was ∼14% loss of both photoreceptor and bipolar cell function from week 2 to the end of 12 weeks (Fig. 3B). Outer retinal dysfunction is not typical in primary open-angle glaucoma. As discussed previously, it may be caused by compromised blood flow in the sclera and choroid. Given the suture model exerts compression exteriorly, it is difficult to avoid this side effect. (4) One more limitation in this study is the use of Brn3a staining for RGCs. Although we were able to show a significant relationship between RGC amplitude and density (Fig. 6B), more sophisticated investigations of cell loss and optic nerve damage are necessary, as the use of Brn3a staining protocol may overestimate the actual loss of RGCs in this model.⁵¹ 

In summary, a chronic mild IOP elevation can be induced in mice with the circumlimbal suture technique. This mild IOP elevation induced retinal changes largely independent of the initial IOP spike. We propose that this mouse model will be useful for studies of chronic ocular hypertension. 

Acknowledgments

The authors thank Liwei Zhang, Izhar Livne-Bar, and Paul Cullen for assistance in OCT and histologic assessment. 

Disclosure: H.-H. Liu, None; J.G. Flanagan, None