Intracameral Injection of a Chemically Cross-Linked Hydrogel to Study Chronic Neurodegeneration in Glaucoma

Abstract

Investigation of neurodegeneration in glaucoma, a leading cause of irreversible blindness worldwide, has been obfuscated by the lack of an efficient model that provides chronic, mild to moderate elevation of intraocular pressure (IOP) with preservation of optical media clarity for long term, in vivo interrogation of the structural and functional integrity of the retinal ganglion cells (RGCs). Here, we designed and formulated an injectable hydrogel based on in situ cross-linking of hyaluronic acid functionalized with vinylsulfone (HA-VS) and thiol groups (HA-SH). Intracameral injection of HA-VS and HA-SH in C57BL/6J mice exhibited mild to moderate elevation of IOP with daily mean IOP ranged between 14±3 and 24±3 mmHg, which led to progressive, regional loss of RGCs evaluated with in vivo, time-lapse, confocal scanning laser ophthalmoscopy; a reduction in fractional anisotropy in the optic nerve and the optic tract projected from the eye with increased IOP in diffusion tensor magnetic resonance imaging; a decrease in positive scotopic threshold response in electroretinography; and a decline in visual acuity measured with an optokinetic virtual reality system. The proportion of RGC loss was positively associated with the age of the animals, and the levels and the duration of IOP elevation. The new glaucoma model recapitulates key characteristics of human glaucoma which is pertinent to the development and pre-clinical testing of neuroprotective and neuroregenerative therapies.

Graphical Abstract

1. Introduction

Glaucoma, a leading cause of irreversible blindness, is the most common neurodegenerative disorder of the optic nerve characterized by chronic, progressive loss of retinal ganglion cells (RGCs) [1]. Similar to other common neurodegenerative diseases of the central nervous systems (CNS) such as Alzheimer’s disease and Parkinson’s disease, the course of disease deterioration in glaucoma patients is chronic and endures for years, typically beginning with loss in the peripheral visual field and contrast sensitivity before compromising the central vision [2]. While intraocular pressure (IOP) elevation is a major risk factor for the development and progression of glaucoma, most glaucoma patients have mild to moderate elevation of IOP [3–5]. Generating mild to moderate IOP elevation with preservation of the optical media clarity in experimental models to recapitulate the IOP profiles of human glaucoma, however, has been technically challenging. An ideal model of experimental glaucoma should be able to: (1) develop mild to moderate elevation of IOP; (2) exhibit chronic, progressive loss of retinal ganglion cells (RGCs) and the degree of RGC loss is related to the levels and the duration of IOP elevation; and (3) preserve the optical media clarity without the development of cataract and corneal opacities induced by the experimental procedures. These features are critical to the investigation of chronic, not acute, degeneration of RGCs and the development of neuroprotective and neuroregenerative therapies. Many of the existing models of experimental glaucoma, however, fall short of demonstrating these attributes.

We designed an in-situ hydrogel based on chemical cross-linking of hyaluronic acid functionalized with vinylsulfone (HA-VS) and thiol groups (HA-SH) for injection into the anterior chamber to induce chronic elevation of IOP. Intracameral injection of hyaluronic acid per se is not effective to induce chronic IOP elevation as the IOP typically returns to the normal levels in 24 hours [6,7]. A unique design of the HA-VS and HA-SH polymers is that the mixture remains injectable in room temperature, and yet it rapidly forms a macroscopic, transparent in situ hydrogel once it is in the anterior chamber where the temperature rise triggers chemical cross-linking of the polymers. Here, we reported for the first time the longitudinal profiles of RGC loss in relation to the levels and the duration of IOP elevation in an experimental model of glaucoma using time-lapse, in vivo confocal scanning laser ophthalmoscopy, and characterized the structural and functional integrities of the visual pathway with diffusion tensor magnetic resonance imaging, electrophysiology, and behavioral testing.

2. Materials and Methods

2.1. Formulation of hydrogel

All chemicals were purchased from Sigma unless specified. Hyaluronic acid (HA) of various molecular weights (MW) was purchased from Bloomage Freda Biopharm Co., Ltd. (Jinan, China). The weight-average molecular weights (Mw) were provided by the supplier. Hyaluronic acid was modified with pedant vinylsulfone (VS) and thiol (SH) groups by methods previously described [8]. In brief, HA was dissolved in water and 5M NaOH was added drop wise to the polymer solution to a final concentration at 0.1M. Divinyl sulfone (DVS) were added instantly into the vigorously mixing polymers in excess at a molar ratio of 1.25× the hydroxyl groups of HA. The degree of modification was adjusted by varying the HA concentration and reaction time [8]. HA-SH was prepared by reacting HA-VS with excess dithiothreitol (DTT) [9]. The polymers were purified by tangential flow filtration (mPES MidiKros TFF Filter, NMWL=3 kDa, D02-E003–05-N, Spectrum Laboratory inc., USA) or dialysis (Spectrum Laboratory inc., USA), with pH adjusted to 5.0 using 0.1M NaOH, sterilized by filtering through a 0.22 μm syringe filter (Millex GP, Merk Millipore Ltd., USA) and freeze dried. The dried powered were stored at −20°C. The degree of modification was determined by HNMR and Ellmans’ assay for HA-VS and HA-SH respectively [9].

2.2. Measurement of gelation time

The gelation time was determined by pipetting test. The two polymers HA-VS and HA-SH were mixed at equal volume and the gelation time was determined when the gel could no longer be aspirated by a 10 μl pipette set to aspirate 10 μl liquid. This test examined the deformation of a hydrogel under a constant pressure and was modified from the conventional tube inversion test for gelation. In contrast to the dynamic time sweep test on dynamic mechanical analysis (DMA), pipetting test allows determination of gelation time for hydrogel formulations with a relatively fast gelation time. A water bath with temperature control was used in temperature-controlled experiments. At least three replicas were conducted for each formulation.

2.3. Measurement of mechanical properties of hydrogel

The mechanical properties of the hydrogels were measured by dynamic mechanical analysis (DMA) as previously described [9]. DMA determines the mechanical properties of viscoelastic material by measuring oscillatory shear deformation (with shear strain controlled) and the corresponding shear stress. Hydrogel is viscoelastic material with both elastic and viscous properties. By measuring both the shear stress response and the phase shift between applied strain and the responded stress, the storage modulus (G’, the elastic property of the gel) and loss modulus (G”, the viscous property of the gel) were determined [6,7]. For G’ measurements of formed hydrogel, molded hydrogel of 8 mm (D) ×1 mm (H) was placed on an 8 mm parallel plate fixture on a DMA machine (ARES Rheometer, TA Instruments, New Castle, DE). G’ was measured at 5 rad/s and 1% strain. For time sweep test, 50μl of polymer mixture was placed on an 8 mm parallel plate fixture loaded to DMA machine. G’ and G” were measured at 15 rad/s, 25% strain. For viscosity measurement, polymer solution was loaded to 50 mm cone-plate (cone angle=0.04 rad, gap=0.05mm) fixture on a DMA machine and measured in steady rate sweep test.

2.4. Animals

Two to 11-month-old C57BL/6J and Tg(Thy1-YFP)GJrs/GfngJ (or Thy-1-YFP-G) transgenic mice with Thy-1 promoter sequences driving the expression of the enhanced YFP were acquired from Jackson Laboratory (Bar Harbor, Maine). All mice were fed ad libitum, the environment was kept at 21°C with a 12-hour light and 12-hour dark cycle, and were in compliance with the NIH Guide for Care and Use of Laboratory Animals. For the MRI and optokinetic studies, the animal experimental protocols were reviewed and approved by the University of Pittsburgh’s Institutional Animal Care and Use Committee. All other experimental procedures were approved by the Chinese University of Hong Kong University Animal Experimentation Ethics Committee and Hong Kong Department of Health, which adheres to The International Guiding Principles for Biomedical Research Involving Animals and The Hong Kong Code of Practice for Care and Use of Animals for Experimental Purposes.

2.5. Hydrogel preparation and in vivo injection

2.6 MDa HA was first dissolved in Balanced Salt Solution Plus component I (Alcon Laboratories, Inc., Fort Worth, TX, United States) and Balanced Salt Solution (Alcon Laboratories, Inc., Fort Worth, TX, United States) at 0.4% w/v concentration overnight and filter sterilized using 0.22 μm syringe filter (Sartorius, Germany). Right before dissolving the polymers, Balanced Salt Solution Plus component II was added to the viscosity enhanced Balanced Salt Solution Plus component I solution at 1:24 volume ratio. Afterwards HA-VS and HA-SH of MW 29 kDa, 20% DM were dissolved in Balanced Salt Solution Plus (components I and II) and Balanced Salt Solution, respectively, at 6% w/v concentration for about 20 minutes. The polymers were then placed on an ice-cold metal plate. Before injection, 5μl of HA-VS and 5μl of HA-SH polymers were mixed on an ice-cold metal plate. 1.5μl of the mixture was aspirated by a pulled glass micropipette needle and injected into the anterior chamber targeted at the anterior chamber angle. It took approximately 3 minutes to complete the intracameral injection.

2.6. Measurement of intraocular pressure

The animals were anesthetized with inhaled 2.5% isoflurane in oxygen via a sealed immersion box. Intraocular pressure was measured with a rebound tonometer (Tonolab; Icare, Finland) with the animals positioned to ensure the probe of the tonometer engaged perpendicularly to the central cornea. The average of 6 readings was calculated for each IOP measurement. All IOP measurements were taken in the afternoon between 4–6pm.

2.7. In vivo imaging of retinal ganglion cells

A confocal scanning laser ophthalmoscope (CSLO) (HRA2; Heidelberg Engineering, Heidelberg, Germany) with illuminating wavelength of 488 nm was used to perform in vivo imaging of the retinas of the Thy-1-YFP-G transgenic mice. The animals were gently and steadily positioned by a technician over a fixed stage next to the objective of the CSLO without the need of general anesthesia. A 10-second rest interval was given for every 15 seconds of imaging to allow eye blinking and keep the corneal surface moist. The scan rate of the CSLO was 5 frames/sec. Each retinal image subtends 30° at a resolution of 768×768 pixels. The software locks the same retinal location to reduce the impact of motion artifact during the imaging. Fifteen retinal images at the same location were averaged to increase the signal-to-noise ratio. In vivo imaging was performed prior to, and at 3, 7, 14, 21, and 28 days, after intracameral injection of hydrogel. Retinal images with YFP-expressing RGCs optimally focused at the same retinal location over the 28-day study period were selected for RGC counting.

2.8. Magnetic Resonance Imaging

All MRI experiments were performed using a 9.4-Tesla/31-cm Varian/Agilent horizontal bore scanner (Santa Clara, CA, USA) with a volume transmit and receive coil designed for mouse studies. Animals were anesthetized with a mixture of air and isoflurane (3% for induction and 1.5% for maintenance) and were kept warm under circulating water during MRI experiments. To ensure reproducible slice orientation and positioning, scout T2-weighted images were first acquired in the coronal, transverse and sagittal planes with a spin-echo pulse sequence. Diffusion tensor imaging (DTI) was performed at 3, 7 and 28 days after intracameral injection of hydrogel. DTI slices were oriented orthogonal to the intracranial optic nerves and were acquired using a fast spin-echo sequence, with 12 diffusion gradient directions at b=1.0ms/μm² and two additional B0 images at b=0ms/μm² (b₀). Other imaging parameters included: Repetition time/echo time=2300/27.8ms; echo train length=8; duration of diffusion gradient pulses (δ)/time between diffusion gradient pulses (Δ)=5/17ms; number of averages=4; field of view= 20×20mm²; acquisition matrix 192×192 (zero-filled to 256×256) and slice thickness= 0.5mm.

2.9. Electroretinography

ERG was performed at baseline and at 28 days after intracameral injection of hydrogel. Mice were dark-adapted for at least 12 hours before the ERG. All subsequent procedures were performed in a dark room under dim red light. Mice were anesthetized by intraperitoneal injection of ketamine (75 mg/kg) and xylazine (10 mg/kg). Pupils were fully dilated with 1% cyclopentolate, 0.5% phenylephrine, and 0.5% tropicamide eye drops. Body temperature was maintained at 37°C with a heated pad during the recording. A gold ring electrode was placed on the cornea as an active electrode and moistened with 0.3% hypromellose eye gel. The reference electrode was placed in the mouth, and ground electrode was inserted subcutaneously into the hind leg. The scotopic ERG measurements were obtained by the Diagnosys Espion system with light emitting diodes (LED) ColorDome Ganzfeld stimulator (Diagnosys LLC, Lowell, MA, USA). The scotopic ERGs were recorded under stimulus intensities from 2.24 × 10⁻⁵ to 0.1 cd·s·m⁻². ERG responses were averaged from 25 flashes with 5-second intervals for the dimmest flash to 10 flashes with 10-second intervals for brightest flash. The scotopic a-wave and b-wave amplitudes were measured from baseline to maximum of first negative trough, and from the maximum of first negative trough to maximum of first positive peak, respectively, at a flash intensity of 0.1 cd·s·m⁻². The pSTR was evaluated at the stimulus intensity of 2.24 × 10⁻⁵ cd·s·m⁻² in the scotopic ERG and measured from the baseline to the first positive peak.

2.10. Optokinetic response

An OptoMotry virtual-reality system (CerebralMechanics, Lethbride, Alberta, Canada) was used to assess visuomotor behavior in awake, freely moving mice by quantifying the visual acuity of each eye before and at 7, 14 and 28 days after the onset of chronic IOP elevation. Spatial frequency was between 0.042 and 0.750 cycles/degree using a simple staircase method with constant rotation speed (0.12 degrees/s) and 100% contrast.

2.11. Immunohistochemical staining of retinal whole mounts and retinal sections

Mice were anesthetized and perfused transcardially with saline followed by 4% paraformaldehyde in 0.1M phosphate buffer saline (PBS) at pH 7.4. Retinas were dissected after identifying the inferior area of the retina by the inferior position of the ophthalmic artery and post-fixed for 3-hours, and then rinsed in 0.1M phosphate buffer. The retinas were incubated in a blocking buffer containing 10% normal goat serum (NGS), 2% bovine serum albumin, and 0.5% Triton X-100 in phosphate-buffered saline. The primary antibodies used were as follows: mouse anti-TUJ1 (Biolegend), rabbit anti-RBPMS (GeneTex), and mouse anti-RHO (Abcam). The primary antibodies were diluted in 5% NGS, 1% BSA, 0.3% Triton X-100 in PBS and incubated with retinas at 4°C overnight. After washing with PBS, fluorophore-conjugated secondary antibodies were added and incubated for 2 hours at room temperature. The retinas were mounted with GB-mount (Golden Bridge Life Science). The retinal whole mounts with TUJ1 or RHO staining were imaged at 40X magnification by a confocal laser scanning microscope (Nikon C1 Confocoal system, Tokyo, Japan). The retinal whole mounts with RBPMS staining for cell counting were imaged at 20X magnification by a fluorescent microscope (Eclipse, Nikon, Tokyo, Japan). A total of sixteen 440μm by 330μm images were captured from the central (4 images from 4 quadrants), middle (4 images from 4 quadrants), and peripheral (8 images from 4 quadrants) retina for cell counting.

2.12. Statistical methods

All statistical analyses were performed using Stata 15.0 (StataCorp; College Station, TX). Intraocular pressure measurements between the hydrogel-injected eye and the fellow control eye of an individual animal collected over 28 days after intracameral injection of hydrogel were compared with a one-tailed paired t-test per mouse to determine how many animals were successfully induced with IOP elevation. Comparisons of IOP, MRI, visual acuity, and RGC counts collected over multiple time points between the hydrogel-injected eyes and the fellow eyes (or between the visual pathways projected from the left and the right eyes for MRI measurements) were compared with linear mixed models y = Xβ + Zu + ε, where y represents the parameter of interest (i.e. IOP, MRI, visual acuity, or RGC count); X represents the design matrix of the fixed effects (i.e. hydrogel-injected eye vs. fellow eye); β represents the coefficient of the fixed effects parameter; u represents the random effect parameter (i.e. random intercept of an individual animal); Z represents the design matrix of the random effects; ε represents the residual. The models adjusted for repeated measurements over time with random intercept nested at the mouse level. Multivariable linear mixed modeling was performed to investigate the association of the impact of baseline age, the levels of IOP elevation, and the duration of IOP elevation (fixed effects variables) on the proportion of surviving RGCs (dependent variable y), with adjustment for repeated measurements over time with random intercept nested at the mouse level. ERG measurements between the hydrogel-injected eyes and the fellow eyes were compared with paired t-test at the baseline and at 28 days. P<0.05 was considered to be statistically significant.

2.12. Data and materials availability

All data that support the findings of this study are made available in this article. Further requests for data and/or materials can be submitted to corresponding author CL.

3. Results

3.1. Formulation of cross-linking hydrogel

The chemical cross-linking is based on a vinylsulfone-thiol click reaction [10,11] which has been investigated to formulate in situ hydrogels for tissue repair and regeneration [9,12]. The vinylsulfone (VS) and thiol (SH) groups are functionalized on hyaluronic acid (HA) using a one-step “click” method we previously described [8,13]. The two polymers can be separately dissolved in a buffered solution. Upon mixing, they form chemical cross-links and transform from a solution to a gel. By changing the molecular weight, degree of modification, polymer concentration, pH, and temperature, the gelation time can be adjusted ranging from about 20 seconds to more than two hours (Fig. 1) and the storage modulus of the hydrogel can be modified (Supplemental Fig. S1). Successful induction of IOP elevation is predicated on the formation of an in-situ hydrogel in the anterior chamber. Precise timing for the gelation is therefore critical. The selected formulation for intracameral injection consists of HA-VS and HA-SH with a molecular weight of 29 kDa, 20% degree of modification at 6% weight/volume, dissolved in Balanced Salt Solution and Balanced Salt Solution Plus buffer, respectively, and mixed at 1:1 volume ratio. This formulation has a pH of 8.3 and forms a hydrogel in about 40 seconds at 37 °C. To increase the viscosity of the hydrogel which helps prevent aqueous humor reflux into the injection needle at the time of intracameral injection, we also added native HA of 2.6 MDa at 0.4% weight/volume to the Balanced Salt Solution and Balanced Salt Solution Plus buffer (Supplemental Fig.S2). We mixed the polymers on an ice-cold metal plate, aspirated the polymers at room temperature (i.e. 23°C), and performed intracameral injection slowly in approximately 3 minutes.

Fig. 1.

Gelation time of different formulations of hydrogel composed of vinyl sulfone modified hyaluronic acid (HA-VS) and thiolated hyaluronic acid (HA-SH). The gelation time can be adjusted by modifying (A) the molecular weight of hyaluronic acid (degree of modification: 20%; polymer concentration: 8%; pH 7.4; temperature: 23°C); (B) degree of modification of VS and SH (molecular weight: 29 kDa; polymer concentration: 4% or 8%; pH 7.4; temperature: 23°C); (C) polymer concentration (molecular weight: 29kDa; degree of modification: 20% pH 7.4; temperature: 23°C); (D) pH (molecular weight: 29kDa, degree of modification:20%; polymer concentration: 4% or 8%; temperature: 23°C); and (E) temperature (molecular weight: 29 kDa; degree of modification: 20%; polymer concentration: 6%; pH 8.3). Experiments were conducted in triplicate. All y-axis values are log transformed. The center line and whiskers represent mean ± s.e.m.

3.2. Longitudinal profile of IOP elevation after intracameral injection of cross-linking hydrogel

We mixed 5μl of HA-VS and 5μl of HA-SH, aspirated 1.5μl of the mixture, and injected into the anterior chamber of 19 Thy-1-YFP-G mice via a pulled glass micropipette needle (Fig. 2A). The Thy-1-YFP-G mice express YFP under the control of a Thy-1 promoter which permits in vivo investigation of progressive RGC degeneration [14–16]. One eye was randomly selected for intracameral injection in each animal. 89.5% (17/19) of the animals showed significant IOP elevation in the hydrogel-injected eyes compared with the fellow control eyes over 4 weeks of follow-up (comparisons of at least 19 pairs of IOP measurements between the right and left eyes collected at different days over 28 days for each animal) (Fig. 2B). There were no adverse effects in the anterior chamber following intracameral injection. Age of the animals at the time of intracameral injection was 4±1 months (range: 2–11 months) (Supplemental Fig. S3). The IOP (mean ± s.e.m) was similar between the hydrogel-injected eyes and the fellow control eyes before the intracameral injection (8.9±0.1 mmHg and 9.2±0.1 mmHg, respectively, p=0.174). Taking all the IOP measurements obtained throughout the 4-week study period into consideration (707 IOP measurements from 34 eyes of 17 animals), the mean IOP of the hydrogel-injected eyes and the fellow control eyes was 18.2±0.5 mmHg and 9.2±0.1 mmHg, respectively (p<0.001). The daily 75^th percentile of IOP was ≤30 mmHg (Fig. 2B), suggesting most animals had mild to moderate elevation of IOP. Intracameral injection of the same hydrogel formulation without functionalization with the vinylsulfone and thiol groups failed to induce significant IOP elevation (n=12) (Supplemental Fig. S4). The overall mean IOP was 9.6±0.1 mmHg in the hydrogel injected eyes, and 9.5±0.1 mmHg in the fellow control eyes (p=0.309). Cross-linking of the hydrogel is critical to the elevation of IOP.

Fig. 2.

Intracameral injection of cross-linking hydrogel for elevation of intraocular pressure in 17 C57BL/6J mice. (A) A schematic diagram of the intracameral injection procedure of cross-linking hydrogel for elevation of intraocular pressure (IOP) in C57BL/6J mice. (B) Intraocular pressure profiles of 17 Thy-1-YFP-G mice with significant IOP elevation after intracameral injection of cross-linking hydrogel. All boxes show the 25th, 50th (median), and 75th percentiles; + represents the data mean; whiskers follow a Turkey boxplot distribution and represent the lower and upper ranges of the data within 1.5× the interquartile range (IQR); outliers beyond the 1.5× IQR are marked by red or white dots. Red – hydrogel injected eyes; white – fellow control eyes. The mean IOP of the hydrogel-injected eyes and the fellow control eyes was 18.2±0.5 mmHg and 9.2±0.1 mmHg, respectively (p<0.001). Comparisons of intraocular pressure between the hydrogel-injected eyes and the fellow control eyes were performed with linear mixed modeling to adjust for multiple measurements within the same animal over the 4-week study period.

3.3. Time-lapse in vivo imaging of progressive retinal ganglion cell loss

We previously described the technique of time-lapse in vivo imaging of RGCs in a transgenic mice model that expresses fluorescent protein under the control of a Thy-1 promoter using a confocal scanning laser ophthalmoscope (CSLO) and characterized the longitudinal profiles of RGC degeneration following optic nerve crush and acute IOP elevation [14–16]. In this transgenic line, about 96% of the fluorescent retinal cells are RGCs [14]. We imaged the retinas of the 17 Thy-1-YFP-G transgenic mice with significant IOP elevation in the hydrogel-injected eyes using the CSLO at 3, 7, 14, 21 and 28 days following intracameral injection of the cross-linking hydrogel (Fig. 3A). The longitudinal profile of RGC loss followed a curvilinear pattern with 57±5%, 51±5%, 45±5%, 41±5% and 37±4% of RGCs surviving at day 3, 7, 14, 21, and 28, respectively (Fig. 3B). The level of IOP, the duration of IOP elevation, and the baseline age were negatively associated with the proportion of surviving RGCs (multivariable linear mixed modeling) (p≤0.036) (Supplemental Table S1). All the 17 eyes showed regional loss of RGCs (Fig. 3A). No significant change in the proportion of surviving RGCs between baseline and week 4 was detected in the fellow control eyes (p=0.672). Immunohistochemical staining with class III beta-tubulin (TUJ1) in retinal whole mounts from eyes with IOP elevation confirms that the YFP-positive retinal cells are RGCs (Fig. 4). Immunohistochemical staining with RBPMS, a pan-RGC marker, was performed in the retinas of 11 C57BL/6J mice dissected at 3 days (n=4 mice), 7 days (n=3 mice), and 28 days (n=4 mice) following intracameral injection of hydrogel in one randomly selected eye (Supplemental Fig. S5). There were significant losses of RGCs (p<0.001) and the proportion of surviving RBPMS-positive RGCs relative to the fellow control eyes were 94±5%, 91±7%, and 80±5%, respectively.

Fig. 3.

High intraocular pressure-induced retinal ganglion cell loss following intracameral injection of cross-linking hydrogel in 17 C57BL/6J mice. (A) Representative time-lapse, in vivo retinal images captured with confocal scanning laser ophthalmoscopy (CSLO) in a Thy-1-YFP-G mouse at the baseline, and 3 days, 7 days, 14 days, 21 days and 28 days following intraocular pressure (IOP) elevation induced by intracameral injection of cross-linking hydrogel. The fluorescence signals in the CSLO images represent the soma of individual retinal ganglion cells. Regional loss of retinal ganglion cells, most prominent over the superior sector (outlined by the dotted lines in the day 28 image), is observed. Scale bar: 500μm. (B) The proportions of YFP-positive retinal ganglion cells in the 17 Thy-1-YFP-G mice remaining over 4 weeks of IOP elevation (the IOP levels are shown in Fig. 2B). All boxes show the 25th, 50th (median), and 75th percentiles; + represents the data mean; whiskers follow a Turkey boxplot distribution and represent the lower and upper ranges of the data within 1.5× the interquartile range (IQR); outliers beyond the 1.5× IQR are marked by red dots.

Fig. 4.

Specificity of fluorescent YFP-G signals from in vivo CSLO imaging to retinal ganglion cells in C57BL/6 mice retina. (A) In vivo confocal scanning laser ophthalmoscopy (CSLO) image of the retina and (B) the corresponding confocal microscopy image of the retinal whole mount isolated from a Thy-1-YFP-G mouse after 4 weeks of IOP elevation. The florescence signals in the CSLO and confocal microscopy images represent the soma of individual retinal ganglion cells. Immunohistochemical staining with class III beta-tubulin (TUJ1) (red) shows colocalization with YFP-positive retinal ganglion cells (green). (C) Magnified images of the boxed areas in (A) and (B). Scale bar: 50μm

3.4. Diffusion tensor magnetic resonance imaging of the optic nerve and the optic tract

We performed longitudinal diffusion tensor magnetic resonance imaging (DTI) to characterize the degree of diffusion anisotropy in the optic nerve and the optic tract in six C57BL/6J mice at 3, 7, and 28 days following intracameral injection of the cross-linking hydrogel to the right eye (Fig. 5A). The IOP of the right eye was significantly higher than the left eye (Supplemental Fig. S6) and the fractional anisotropy (FA) of the right optic nerve was on average 13% (95% CI: 6% to 20%) smaller than that of the left optic nerve (p<0.001) whereas the FA of the left optic tract projected from the right, hypertensive eye was on average 11% (95% CI: 5% to 17%) smaller than that of the right optic tract projected from the left, normotensive eye (p=0.001) over the 4-week study period (Fig. 5B). Significant differences in FA between the right and left optic nerves, as well as between the right and left optic tracts, could be detected as early as at 3 days following IOP elevation (p≤0.019). To further dissect the directional diffusivities of DTI measures in relation to the optic nerve and the optic tract, longitudinal changes in axial diffusivity and radial diffusivity after induction of IOP elevation were analyzed. Axial diffusivity was on average 9% (95% CI: 2% to 16%) smaller in the right than the left optic nerves (p=0.009), but it was not significantly different in the optic tracts (p=0.380) (Supplemental Fig. S7A). By contrast, radial diffusivity was 27% (95% CI: 4% to 50 %) greater in the right than the left optic nerves (p=0.023), and 9% (95% CI: 4% to 15%) greater in the left than the right optic tracts (p=0.002) (Supplemental Fig. S7B). There were no differences in fractional anisotropy, axial diffusivity, or radial diffusivity of the optic nerve and optic tract measurements between the fellow control eyes on day 28 following hydrogel injection and eyes of age-matched naïve mice (p≥0.098) (Supplemental Fig. S8A–C).

Fig. 5.

Diffusion tensor magnetic resonance imaging measurements of C57BL/6J mice optic nerve and optic tract following intracameral injection of cross-linking hydrogel. (A) Diffusion tensor magnetic resonance imaging (DTI) at the levels of the pre-chiasmatic optic nerve (top row; Bregma = 0.1 mm) and the optic tract (bottom row; Bregma = −1.9 mm) at 4 weeks after chronic IOP elevation induced by intracameral injection of cross-linking hydrogel to the right eye. Regions of interests (ROI) were drawn on the optic nerve and optic tract of each hemisphere in the color-encoded fractional anisotropy map, and then overlaid on the DTI parametric maps for quantitation. Color representations for the principal diffusion directions in color-encoded fractional anisotropy map: blue, caudal-rostral; red, left-right; green, dorsal-ventral. Scale bar = 1 mm. (B) Relative difference in fractional anisotropy (expressed in % difference between the visual pathways projected from the hydrogel-injected eyes to the fellow controlled eyes) measured by DTI at 3 days, 7 days, and 28 days following intraocular pressure (IOP) elevation induced by intracameral injection of hydrogel (n=6 mice). Fractional anisotropy was significantly decreased in the optic nerve and the optic tract projected from the eyes with increased IOP (*p≤0.001). Comparisons of fractional anisotropy between the visual pathways projected from the left and right eyes were performed with linear mixed modeling after adjusting for multiple measurements in the same animal over the 4-week study period. The center line and whiskers represent mean ± s.e.m. The IOP measurements over the 4-week study period are shown in Fig. S6.

3.5. Electrophysiology

To evaluate the functional responses of the outer and inner retina, we measured the scotopic electroretinogram (ERG) response (a- and b-waves) and the positive scotopic threshold response (pSTR), respectively, before and at 28 days after IOP elevation via intracameral injection of cross-linking hydrogel in five C57BL/6J mice (Supplemental Fig. S9). The a- and b-waves of the scotopic ERG originate from the photoreceptors and Müller cells/bipolar cells, respectively, whereas the pSTR originates from the RGCs. At the baseline, the scotopic ERG a- and b-waves and the pSTR did not differ between the right and left eyes (p≥0.450) (Fig. 6). At 28 days following IOP elevation, no significant differences in the scotopic ERG a- and b-waves were detected between the right and left eyes (p≥0.206). The pSTR, by contrast, was significantly smaller in the eyes injected with the hydrogel (25±5μV) than the fellow eyes (38±3μV) (p=0.041), suggesting that the functional integrity of the inner retina, but not the outer retina, was compromised (Fig. 6). Immunohistochemical staining with RHO (a specific marker of outer segments of photoreceptors) confirmed the outer retinas were intact on day 3 and 28 after IOP elevation induced by intracameral injection of hydrogel (Supplemental Fig. S10).

Fig. 6.

Scotopic response of C57BL/6J mice retina using electroretinography following intracameral injection of cross-linking hydrogel. (A) Positive scotopic threshold response (pSTR) captured at a flash intensity of 2.24X10–5 cd·s·m−2 before (upper panel) and 4 weeks (lower panel) after intraocular pressure (IOP) elevation induced by intracameral injection of hydrogel. The amplitude of pSTR was significantly reduced in the hydrogel-injected eyes compared with the fellow control eyes at week 4 after IOP elevation (n=5 mice). Comparisons of pSTR between the hydrogel-injected eyes and the fellow control eyes per time point were performed with paired t-test. The center line and whiskers represent mean ± s.e.m. Intraocular pressure measurements of these 5 mice over the 4-week study period are shown in Fig. S8. *p<0.05; N.S.=not significant. (B) Scotopic electroretinogram response captured at a flash intensity of 0.1 cd·s·m−2 before (upper panel) and 4 weeks (lower panel) after IOP elevation induced by intracameral injection of cross-linking hydrogel. The amplitudes of a-wave and b-wave were not significantly different between the hydrogel-injected eyes and the control fellow eyes before and after 4 weeks of IOP elevation (n=5 mice, measured from the same group of animals). Comparisons of pSTR between the hydrogel-injected eyes and the fellow control eyes at baseline and on day 28 shown here were performed with paired t-test. The center line and whiskers represent mean ± s.e.m. Intraocular pressure measurements of these 5 mice over the 4-week study period are shown in Fig. S8. N.S.=not significant

3.6. Optokinetic response

Visual acuity (cycles/degree) was measured with an optokinetic virtual reality system in six C57BL/6J mice with IOP elevation following intracameral injection of hydrogel (Fig. 7A). Visual acuity measured with an optokinetic virtual reality system progressively declined in eyes with IOP elevation, decreasing from 0.39±0.01 before IOP elevation, to 0.33±0.01, 0.30±0.01, and 0.27±0.01 at week 1, 2 and 4 following IOP elevation (Fig. 7B) with an estimated rate of change of −1.0%/day (95% CI: −0.8 to −1.3%/day) (p<0.001). There was no difference in visual acuity between the fellow control eyes on day 28 following hydrogel injection and eyes of age-matched naïve mice (p=0.105) (Supplemental Fig. S8D).

Fig. 7.

Visual acuity measurements obtained from an optokinetic virtual reality system following intracameral injection of cross-linking hydrogel in 6 C57BL/6J mice. (A) Intraocular pressure (IOP) measurements before and after IOP elevation induced by injection of cross-linking hydrogel to the right eye of six C57BL/6J mice. IOP was higher in the hydrogel-injected eyes compared with the fellow control over the 4-week study period (p<0.001). Comparisons of intraocular pressure between the hydrogel-injected eyes and the fellow control eyes were performed with linear mixed modeling to adjust for multiple measurements within the same animal over the 4-week study period. The center line and whiskers represent mean ± s.e.m. *p<0.001 (B) Visual acuity measurements obtained from an optokinetic virtual reality system before and after IOP elevation to the right eye. Visual acuity progressively declined over time in the hydrogel-injected eyes compared with the fellow control eyes over the 4-week study period (p<0.001). Comparisons of visual acuity between the hydrogel-injected eyes and the fellow control eyes were performed with linear mixed modeling to adjust for multiple measurements within the same animal over the 4-week study period. The center line and whiskers represent mean ± s.e.m. *p<0.001

4. Discussion

4.1. Intracameral injection of cross-linking hydrogel supports a chronic model of neurodegeneration in retinal ganglion cells

We designed and formulated a cross-linking hydrogel for intracameral injection to study chronic degeneration of the RGCs and the optic nerve in experimental glaucoma. Intracameral injection of cross-linking hydrogel is able to provide mild to moderate elevation of IOP after a single injection for at least 4 weeks with daily mean IOP ranged between 14±3 and 24±3 mmHg and daily 75^th percentile ≤30mmHg. More important, our model preserves the clarity of the optical media, which allows time-lapse, in vivo interrogation of the structural and functional integrity of the RGCs. That the proportion of surviving RGCs was negatively associated with the age of the animals as well as the levels and the duration of IOP elevation (Supplemental Table S1) recapitulates the key features for development and progression of human glaucoma reported across population-based studies and clinical trials [3–5,17–20]. The investigation of RGC and optic nerve degeneration has been largely centered on acute injury models of optic nerve crush or ischemia, which lead to drastic loss of RGCs within days. Intracameral injection of the cross-linking hydrogel provides a chronic model of neurodegeneration, which will expedite the development of neuroprotective and neuroregenerative therapies for glaucoma and other chronic neurodegenerative diseases.

4.2. Optimization of the HA-VS and HA-SH formulation suitable for intracameral injection

We screened different hydrogel formulations (Fig. 1) and noted that timing for gelation is critical to the success of induction of IOP elevation. The selected formulation should be able to gel rapidly right after intracameral injection or it will be diluted by the aqueous humor. On the other hand, rapid gelation may compromise intracameral injection as solidified hydrogel would clog the injection needle. We tackled these problems by exploring the effect of pH and temperature on the hyaluronic acid (HA)-based VS-SH Michael Addition gelation reaction. The reaction between VS and SH depends on the ionization of the thiol anion (S⁻); one-unit increase in pH raises the concentration of S⁻ by ~10 times and shortens the gelation time by ~10 folds (Fig. 1D). The formulation we chose forms a gel after ~10 minutes at room temperature at physiological pH 7.4. We accelerated the gelation to ~3 minutes by raising the pH to 8.3, using a combination of clinical-grade eye irrigation solution – Balanced Salt Solution and Balanced Salt Solution Plus. This formulation forms a gel in ~40 seconds at physiological temperature. To prevent gelation before injection, the polymer solutions were mixed on ice, which prolonged the gelation time by ~20 folds compared with the physiological temperature at 37°C (Fig. 1E). Native higher molecular weight hyaluronic acid was added to increase the viscosity of the solution to prevent aqueous humor reflux into the injection needle and polymers dilution at the time of injection. The unique design of the cross-linking hydrogel facilitates aspiration and intracameral injection, which typically takes approximately 3 minutes. Due to the viscoelastic nature of this hydrogel with a polymer chain thickness within the angstrom range, the mesh size cannot be visualized using electron microscopy, which often requires sample dehydration. However, using the Amsden equation and the Blob model, the mesh size was estimated to be approximately 5 nm [8]. In this study, a high success rate of induction of IOP elevation is observed with 89.5% (17/19) of the hydrogel-injected eyes demonstrating significant IOP elevation compared with the fellow eyes.

4.3. Intracameral injection of hydrogel preserves optical clarity for long-term follow-up in vivo imaging

While a number of animal models for IOP elevation have been described, preserving the optical media clarity of the eye for investigation of the structural and functional integrities of the RGCs has been technically challenging. Hypertonic saline injection to the episcleral veins and photocoagulation of the trabecular meshwork can induce IOP elevation and RGC degeneration [21–25]. However, development of corneal opacities and hyphemia obscuring the visual axis is not uncommon. Whereas the DBA/2J transgenic mice engenders an inheritable form of glaucoma with IOP rise secondary to pigment dispersion, anterior chamber inflammation, and iris adhesion [26], these features are not observed in most clinical forms of glaucoma and the development of cataract and pigment dispersion inevitably obscures the optical media. Likewise, although intracameral injection of microbeads can induce mild to moderate elevation of IOP [27–32], the presence of microbeads in the anterior chamber, and the development of lens and corneal opacities would jeopardize in vivo imaging of RGCs and reliable measurement of visual function. The fact that we were able to perform time-lapse, in vivo imaging of the retina following intracameral injection of hydrogel (Fig. 3) and that there were no significant differences in the amplitudes of scotopic a- and b-waves between the hydrogel-injected eyes and the fellow control eyes (Fig. 6) support the notion that the optical media was well-preserved in our animals. The formula of the cross-linking hydrogel comprises a hyaluronic acid-based VS-SH reaction gelation method. Hyaluronic acid is highly biocompatible. An in vitro study shows that both gel incubation and in situ gelation of HA induces no toxicity to ocular cells, and an in vivo study in rabbit and monkey eyes demonstrated its long-term biocompatibility with no notable adverse reactions [33,34].

4.4. Hydrogel-mediated elevation of IOP induces chronic, progressive loss of retinal ganglion cells

To our knowledge, our data are the first to demonstrate progressive loss of RGCs in vivo in experimental glaucoma. In contrast to acute IOP elevation (e.g. elevating the IOP to 110 mmHg for 60–90 minutes) which induces RGC loss in the first week of injury but none thereafter [16], we detected progressive loss of RGCs over 4 weeks of chronic IOP elevation (Fig. 3). It is also notable that unlike the diffuse RGC loss following optic nerve crush [15], RGC loss in our glaucoma mouse model was regional, which is parallel to the localized retinal nerve fiber layer defects typically observed in patients with mild to moderate glaucoma. Although histological assessment of RGC loss evaluated with immunohistochemical staining of RBPMS (Supplemental Fig.S5) showed a lesser degree of RGC loss compared with in vivo CSLO imaging (Fig.3), it is worth noting that the proportions of surviving RGCs were referenced to the fellow control eyes. A major advantage of in vivo CSLO imaging is the ability to perform serial imaging of the same eye, which offers a greater sensitivity to detect change compared with histological assessment at multiple time points. Collectively, intracameral injection of cross-linking hydrogel recapitulates the key characteristics of RGC degeneration in glaucoma, which is progressive and regional, with increasing RGC loss in animals of older age (the proportion of surviving RGCs was 2.8% smaller for each month increase in baseline age after controlling for the duration and level of IOP elevation) and higher levels of IOP (the proportion of RGCs was 1.2% smaller for each mmHg increase in IOP after controlling for the duration of IOP elevation and baseline age) (Supplemental Table S1).

4.5. Eyes with intracameral injection of hydrogel demonstrate loss of inner retinal function

Preservation of clear optical media is relevant not only to in vivo imaging of RGCs, but also to the evaluation of visual function. Scotopic threshold response (STR) originates from the inner retina and the amplitude of positive scotopic threshold response (pSTR) has been shown to decrease after optic nerve crush [35]. By contrast, eyes with retinal ischemia following acute increases of IOP at 70mmHg or above show decreases in the amplitude of EGC scotopic a- and b- waves and disorganized outer retinal layer [36–42]. The significantly smaller amplitude of pSTR in the hydrogel-injected eyes compared with the fellow control eyes at week 4 of IOP elevation, together with the lack of significant differences in scotopic a-wave and b-wave amplitudes between the hydrogel-injected eyes and the fellow control eyes at the baseline and week 4 strongly support that IOP elevation following intracameral injection of hydrogel damaged the inner retina, not the outer retina (Fig. 6). Histological assessment on day 3 and 28 following hydrogel injection further confirms the absence of outer retinal damage over the study follow-up (Supplemental Fig. S10). Loss of the inner retinal function was accompanied with progressive decline in optokinetic response over the 4 weeks of IOP elevation (Fig. 7B). Together, intracameral injection of cross-linking hydrogel can be efficiently applied to study serial structural and functional degeneration of the RGCs.

4.6. Hydrogel-induced retinal ganglion cell loss causes concomitant damage to the optic nerve and optic tract

Degeneration of the intracranial optic nerve, optic tract, lateral geniculate nucleus, and visual cortex has also been reported in glaucoma patients [43–46]. To characterize white matter changes in the brain, we employed 9.4-Tesla DTI to measure fractional anisotropy – a summary measure of microstructural integrity – in the intracranial optic nerve and the optic tract in our glaucoma model. Surprisingly, significant changes in fractional anisotropy were detected not only in the optic nerve but also in the optic tract as early as 3 days after IOP elevation (Fig. 5), coinciding with the loss of RGCs detected by confocal scanning laser ophthalmoscopy (Fig. 3). Although the IOP was not significantly high in the hydrogel-injected eyes at day 3 (all below 20mmHg) (Supplemental Fig.S6), the early changes in the optic nerve and the optic tract likely reflect the impact of IOP elevation over 3 days. The fact that we did not have additional IOP measurements over the first three days may limit the interpretation of the DTI findings. Fractional anisotropy was reduced in the ipsilateral intracranial optic nerve (projected from the eye with intracameral injection of hydrogel) and the contralateral optic tract (97%−98% of the optic nerve fibers are crossed over to the contralateral eyes in mice) [47] throughout the 4 weeks of IOP elevation. Axial diffusivity and radial diffusivity represent two scalar vectors of fractional anisotropy. In general, axial diffusivity decreases in axonal injury, whereas radial diffusivity increases with demyelination and/or decreases in axonal diameters or density. The finding of a reduction in axial diffusivity in the optic nerve with a concomitant increase in radial diffusivity in the optic nerve and optic tract in our glaucoma model corroborates a recent study demonstrating astrocytic and inflammatory changes of the axons with fragmented myelin sheath in the DBA/2J mice model of glaucoma [48]. Although we did not investigate the association between RGC loss and DTI diffusivity changes, axon counts measured with histology have been shown to linearly decrease with DTI fractional anisotropy and diffusivity parameters in non-human primate models of glaucoma [49]. That no additional changes in the DTI parameters were detected from day 7 onwards may be connected to the fact that DTI imaging is not sensitive to detect additional but insubstantial loss in the optic nerve and optic tract diffusivity in mice. Additional MRI approaches, such as diffusion kurtosis imaging, diffusion-based spectrum imaging, and white matter tract integrity measurements [50–53], would be worthy of investigation to further uncover the structural changes of the optic nerve and optic tract. Our hydrogel-induced glaucoma mouse model underscores the notion that glaucoma is a neurodegenerative disease and that degenerative changes in the optic nerve and the optic tract can be detected in the early stages of glaucoma.

5. Conclusions

Current treatment modalities for glaucoma patients is centered on lowering the IOP. That irreversible progressive loss in vision can be observed in some glaucoma patients despite IOP lowering treatment indicates an incomplete understanding of the pathogenesis of glaucoma and the lack of a neuroprotective therapy to halt disease progression. Intracameral injection of the cross-linking hydrogel for induction of chronic IOP elevation has provided a simple and efficient model of chronic optic nerve injury to advance the investigation and development of neuroprotective and neuroregenerative therapies.

Statement of Significance.

A new model to study chronic neurodegeneration in glaucoma has been developed via intracameral injection of a specifically designed hyaluronic acid functionalized with vinylsulfone and thiol groups for cross-linking. Intracameral injection of the chemically cross-linked hydrogel generates mild to moderate IOP elevation, resulting in progressive degeneration of the retinal ganglion cells, optic nerve, and optic tract, and a decline in visual function. The model recapitulates the key features of neurodegeneration in human glaucoma, which will facilitate and expedite the development of neuroprotective and neuroregenerative therapies.