Functional Changes in the Retina during and after Acute Intraocular Pressure Elevation in Mice

Abstract

Purpose.: To examine retinal function using the full-field electroretinogram (ERG) during and after acute intraocular pressure (IOP) elevation in wild-type mice.

Methods.: IOP was elevated by anterior chamber cannulation in wild-type C57/BL6 mice. The pressure–function relationship was determined by IOP elevation in steps from baseline to 80 mm Hg. The rate of functional recovery was assessed for 60 minutes after an IOP spike of 50 mm Hg for 30 minutes. During and immediately after IOP elevation, scotopic ERG signals were recorded in response to dim and bright flashes (−4.54, −2.23, and 0.34 log cd · s · m⁻²) and analyzed for photoreceptoral (a-wave), ON-bipolar (b-wave), oscillatory potentials (OPs), and scotopic threshold responses (positive [p]STR/negative [n] STR). A full ERG protocol was collected 2 days before and 7 days after the single 50-mm Hg IOP spike.

Results.: The pSTR was most sensitive to IOP elevation with 50% amplitude loss (μ) at 41 mm Hg (μ, 95% confidence limits (CL): 37.7, 45.6) followed by nSTR at 45 mm Hg (95% CL: 41.0, 49.1). pSTR was significantly more sensitive than the b-wave (95% CL: 41.4, 49.1), a-wave (95% CL: 47.6, 55.3), and OPs (95% CL: 49.6, 59.2). pSTR showed slower recovery immediately after the 50 mm Hg spike compared with the b-wave (P = 0.02). One week after the 50-mm Hg spike, pSTR (−30% ± 6%, P < 0.001) and OP (−27% ± 2%, P < 0.001) amplitudes were reduced, whereas other components were unaffected.

Conclusions.: The STR in mice is more sensitive to acute IOP elevation and recovers slower than other ERG components. Reduction in pSTR and OP amplitude at 1 week suggests persistent impairment of inner retinal function can occur after a single IOP spike.

Experimental mouse models are increasingly important in studies of neurodegeneration and glaucoma, largely due to an improved understanding of mouse genetics, the relative ease of genetic modifications and an ever-increasing number of murine lines with well-defined genotypes and phenotypes. Several models of intraocular pressure (IOP) elevation, both induced and spontaneously occurring, have been used to study retinal neuronal injury in mice. ^1–5 Although retinal function has been assessed after IOP elevation in mice, detailed data of the retinal response during and after a well-controlled acute IOP elevation remain sparse. 

Investigators have examined the effect of acute IOP elevation on retinal function in humans and in some animal models. ^6–9 In general, their studies show that an IOP elevation of 30 to 35 mm Hg is needed to induce retinal dysfunction. Bui et al. ¹⁰ showed in rats that the inner retina derived scotopic threshold response (STR) was more sensitive to acute IOP elevation than ERG components arising from photoreceptors (a-wave) and bipolar cells (b-wave). ¹⁰  

The effect of a single acute IOP elevation on retinal function and its recovery has been investigated in the DBA/2J mice, a strain predisposed to spontaneous glaucomatous optic neuropathy. Nagaraju et al. ¹¹ showed that a 32% to 38% increase in IOP (∼5 mm Hg) could be achieved by a 60° head down position. Although this IOP change had no effect on the pattern ERG in 3-month-old DBA/2J mice, it produced a ∼65% amplitude reduction in 10-month-old mice. The pressure–function relationship to a moderate IOP challenge has yet to be assessed in wild-type mice. 

In this study, we examined in detail the pressure (IOP)–function relationship in adult wild-type mice during a step-wise increase in IOP. In particular, we examined whether ERG components arising from the inner, middle, and outer retina show different sensitivity to IOP elevation. We also considered whether there was a difference in the immediate (60 minutes) and medium-term (7 days) recovery of ERG components after a single IOP spike. 

Materials and Methods

Animals

All experimental methods and animal care procedures conformed to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and were approved by our Institutional Animal Experiment Ethics Committee (The University of Melbourne, 07087327.1). Wild-type C57/BL6 mice of 12 months of age (30–35 g, Royal Victorian Eye and Ear Hospital) were maintained in a 22°C, 12-hour light (∼40 lux)/12-hour dark environment with normal murine chow (WEHI breeder mix; Barastoc, VIC, Australia) and water available ad libitum. 

Electroretinography

Full-field ERGs were recorded simultaneously from both eyes. In brief, animals were dark-adapted overnight (>12 hours) and prepared for cannulation and ERG recording in the dark with head-mounted night-vision goggles (Scout2; Trivisio Prototyping GmbH, Dreieich, Germany). Animals were anesthetized with intraperitoneal injection of ketamine-xylazine (70 mg/kg and 7 mg/kg, respectively; Troy Laboratories Pty Ltd., Smithfield, NSW, Australia) followed by supplementation with 20% of the initial dose every 30 minutes as previously described. ^12,13 Mydriasis was achieved with 1 drop of tropicamide (0.5%, Alcon Laboratories, Pty, Ltd., Frenchs Forest, NSW, Australia) and phenylephrine (2.5%, Minims; Chauvin Pharmaceuticals, Surrey, UK). Corneal anesthesia was achieved with a single drop of proxymetacaine hydrochloride (0.5%; Alcon Laboratories, Inc.). Animals were lightly secured to a platform with wire loops across the upper back and nose. A circulating warm water heating pad was used to maintain body temperature (37–38°C). 

The pulse was monitored continuously throughout each experiment and systolic blood pressure was measured at regular intervals by using a noninvasive tail-cuff sphygmomanometer (ML125; ADInstruments, Bella Vista, NSW, Australia). This method was used to ensure constant blood pressure throughout the experiment, to avoid confounding readings induced by changes in ocular perfusion pressure. There was no significant change in blood pressure (average 105 ± 5 mm Hg) over the experimental period (ANOVA P = 0.63) consistent with similar observations made in rat. ¹⁴  

Signals were recorded using custom-made silver/silver-chloride electrodes (99.99% purity, 0.329 mm ¼ 29 G; A&E Metal Merchants, Sydney, NSW, Australia) adapted from those used for rats. ^14,15 The active electrode was centered on the cornea using micromanipulators (KITE; WPI). The inactive electrode comprises a loop (3 mm diameter) that fits around the equator of the eye, to rest behind the limbus. Both of these are referenced to a stainless steel ground (F-E2–60; Grass Technologies, West Warwick, RI) inserted into the tail. We found that this montage offered the low noise levels needed for scotopic threshold response (STR) recordings. ¹⁶  

After the mouse was placed on the heated platform, electrode placement and anterior chamber cannulation (described later) were performed in darkness with the assistance of monocular night-vision scopes (NVMT1; Yukon Advanced Optics, Mansfield, TX) fitted in place of the eye pieces of a dissecting microscope (MZ6; Leica, Wetzlar, Germany). Illumination was provided by an infrared light source (QC3650; Jaycar Electronics, Rydalmere, NSW, Australia). This allowed accurate cannulation and electrode positioning while maintaining dark adaptation. The eyes were lubricated after electrode placement and periodically throughout the session with 1.0% carboxymethylcellulose sodium (Celluvisc; Allergan, Irvine, CA). Signals were amplified 1000× and recorded with a band-pass setting of 0.3 to 1000 Hz (−3 dB) (P511 AC Preamplifier, Grass Telefactor) and 4 kHz acquisition (Powerlab 8SP; ADIntruments). 

Light stimuli were brief (1 ms) white flashes (5-W white LEDs, 5500°K; Luxeon Calgary, ALB, Canada) delivered via a Ganzfeld integrating sphere (Photometric Solutions International, Huntingdale, VIC, Australia). The integrated luminous energy was calibrated with a photometer (IL1700; International Light Research, Peabody, MA) with a scotopic luminosity filter (Z-CIE) in place. The scotopic luminosity filter was calibrated in terms of rodent photopigment absorbance (λ_max = 502 nm). ^17,18 Based on an average pupil area of 7 mm², our flash of 0 log cd · s · m⁻² gave 516 Φ/rod, consistent with previous calculations. ^12,19  

Before IOP elevation, all animals underwent baseline ERG recordings at a range of luminous energy. From −5.92 to −4.73 log cd · s · m⁻² 20 to 40 responses were averaged, with an interstimulus interval (ISI) of 1.5 to 3 seconds. Between −4.54 and −1.90 log cd · s · m⁻² 2 to 10 responses were averaged with an ISI of 3 to 10 seconds. From 1.32 and 2.22 log cd · s · m⁻² single responses were recorded with the ISI progressively lengthened from 10 to 120 seconds. Acute IOP elevation experiments were conducted on the same animal within 2 days of this baseline session. 

Acute IOP Elevation

After inactive electrode placement, the anterior chamber of the left eye was cannulated with a glass Pasteur pipette (∼50 μm, 1B100–6; WPI, Sarasota, FL) connected via polyethylene tubing (0.97 mm inner diameter; Microtube Extrusions, North Rocks, NSW, Australia) to a pressure transducer (Transpac IV; Abbott Critical Care Systems, Sligo, Ireland), which was in series with a sterile Hanks' balanced salt solution reservoir (JRH Biosciences, Lenexa, KS). IOP was controlled by adjusting the height of the reservoir to precalibrated levels and maintained within 1 mm Hg of the target pressure. IOP calibration was performed by reference to an aneroid sphygmomanometer (Livingstone, Rosebery, NSW, Australia) as previously described. ¹⁴ Cannulation was performed with the reservoir height set to eye level and the reservoir valve closed. Variation in needle diameter can lead to differences in flow rate, which may vary the time taken for IOP to stabilize. The anterior chamber was viewed under infrared illumination during valve opening, to ensure that these differences did not significantly affect IOP equilibration, and rapid filling was observed in all cases (∼5 seconds). On intraocular pressure equilibration, aqueous flow through normal outflow channels can result in a small difference between measured IOP and actual IOP due to resistance in the glass pipette. Our calibration of flow rate (0.23 ± 0.01 mm Hg per μL/min) through a glass needle found that within the range of IOP used in our experiments (up to 80 mm Hg) the overestimation of IOP due to pipette resistance (<0.2 mm Hg) was negligible (see Supplementary Material). This calculation for mice eyes assumes conventional outflow to be 0.0051 ± 0.0006 μL/min per mm Hg of IOP and uveoscleral flow to be 0.148 μL/min at a pressure of 80 mm Hg. ²⁰ As such, we believe that the resistance through the pipette did not alter the interpretation of our experimental results. 

Baseline ERG responses were obtained after 10 minutes of further dark adaptation, with IOP maintained manometrically at baseline (12 mm Hg). Baseline responses were recorded to luminous energies that elicited an STR (40 signals averaged, −4.54 log cd · s · m⁻²), a saturated rod ON-bipolar cell dominated b-wave (single flash, −2.23 log cd · s · m⁻²) and a photoreceptoral a-wave response (single flash, 0.34 log cd · s · m⁻²). The choice of luminous energy for the STR is consistent with that in previous work in mice. ^1,12 On completion of baseline ERG measurements, the saline reservoir was adjusted to give IOP levels according to one of the following protocols. 

Protocol 1: IOP-Response Function.

IOP was raised in a step-wise manner starting from 25 up to 50 mm Hg in 5-mm Hg increments and from there up to 80 in 10-mm Hg increments. At each step, IOP was stabilized for 10 minutes before the same ERG recordings (3 luminous energies) were made (n = 6, animals). 

Protocol 2: Time Course of Fixed IOP Spike.

IOP was raised to 50 mm Hg for 30 minutes (spike) and then decreased to baseline (∼12 mm Hg) for recovery. During both the insult and recovery (60 minutes) phases, ERG recordings were made to dim (−4.54 log cd · s · m⁻²) and intermediate luminous energies (−2.23 log cd · s · m⁻², rod b-wave response) at 10-minute intervals (n = 10, animals). 

On completion of the experiment, the cannula was removed, and animals were allowed to recover on an electric heating pad. Negative controls were established with sham experiments in which the eye was cannulated and IOP remained at baseline (n = 4), to control for the effect of experimental manipulation. A full scotopic series of ERG recordings was repeated in the same animals 7 days after IOP challenge. 

Data Analysis

Photoreceptor Response.

Analysis of rodent electroretinogram components in this laboratory have been described in detail elsewhere. ^10,14–16 Briefly, for the full scotopic ERG protocol the leading edge of the photoreceptor (P3) a-wave, which reflects photoreceptor activity, ²¹ was quantified by modeling the highest four luminous energies (1.29, 1.62, 1.92, and 2.22 log cd · s · m⁻²) as an ensemble using a delayed Gaussian function. ²² This was achieved by (see Weymouth and Vingrys ¹⁵ for details) floating the saturated amplitude(Rm_p3; μV), sensitivity (S; m² · cd⁻¹ · s⁻³), and delay (t _d; ms) to minimize the sum-of-squares merit function (Excel Solver; Microsoft, Redmond, WA). For the single flash responses measured during or after IOP challenge, P3 modeling is not viable and a-wave amplitude was quantified at a fixed time of 6 and 8 ms after stimulus onset, to minimize contamination from nonphotoreceptoral contributions. ^23,24 The data presented herein are for 8 ms, to improve signal-to-noise; however, the 6-ms analysis gave the same outcome. 

Bipolar Cell Response.

The postreceptoral P2 waveform reflects the activity of bipolar cell responses. ^22,25–27 The P3 derived from the delayed Gaussian model was digitally subtracted from the raw ERG to expose the P2–OP (oscillatory potential) complex, which was then digitally low-pass–filtered (47 Hz, −3 dB) to return the P2. ¹⁵ P2 amplitude was measured from the filtered waveform at a fixed time of 110 ms from stimulus onset, which corresponds to the average peak of control eyes (n = 10) at −2.23 log cd · s · m⁻². 

Scotopic Threshold Response.

The STR has been shown in mice ^12,28 and rats ¹⁶ to contain ganglion cell and some amacrine cell contributions. Response amplitudes of the positive (p)STR and negative (n)STR components were measured at fixed times of 120 and 220 ms after the stimulus flash, respectively. These fixed times were chosen to correspond to the STR peak and trough of control eyes (n = 10). 

Oscillatory Potentials (OPs).

OPs were isolated from the flash response to 0.34 log cd · s · m⁻². The effect of the a-wave on the OPs was reduced by removing the raw waveform up to the minimum of the first electronegative deflection. ²⁹ The resulting waveform was then digitally band-pass–filtered (fifth-order Butterworth, −3 dB at 55 and 210 Hz) to isolate the scotopic OPs. The OPs were then modeled using a Gabor function, as described previously. ²⁹ A Gabor comprises a Gaussian envelope that locates the oscillation as a function of time (m; milliseconds) with maximum amplitude (a; microvolt) and spread (s; milliseconds), and a resonating element that specifies the frequency (h; hertz) and phase (p; radians) of the oscillation. ²⁹ The parameters of the Gabor were optimized by minimizing the sum-of-squares merit function (Excel Solver; Microsoft). 

Retinal Function versus IOP.

IOP effects were expressed as relative amplitudes (treated/baseline, %) for the various components at each IOP step. Group data were shown as mean (±SEM) and plotted as a function of IOP for each ERG parameter. The IOP pressure-function relationship was modeled using an inverse cumulative normal. ¹⁰ The model returns the mean (μ), which indicates the IOP resulting in 50% amplitude reduction; and the SD (σ). A smaller SD equates to a steeper slope. ³⁰ The model was fitted to the relative amplitude at all pressure levels, with the maximum fixed to 100%. 

Sensitivity of various ERG components to acute IOP elevation was compared by contrasting parameters in terms of their 95% bootstrap confidence limits, with an effective nonparametric α of 0.05. The bootstrap method has been described elsewhere. ^10,14,29,31 In short, several synthetic populations of size n (bootstrap samples) were generated by random selection (with replacement) from the available data sets across IOP. ³² Each synthetic sample was parameterized with the inverse cumulative normal function to give an estimate of μ and σ for that particular bootstrap sample. After 1000 repetitions, the nonparametric 2.5% and 97.5% confidence limits for μ and σ could be derived. 

Statistical Methods.

One-way, repeated-measures analysis of variance (ANOVA; SPSS ver. 15.00; SPSS Inc., Chicago, IL) was applied to test the null hypothesis that there was no effect of IOP elevation on ERG component (α = 0.05). Two-way ANOVA was used to test the statistical difference between groups across time or luminous energy, and Tukey's test was used for post hoc analysis of subcategories. 

Results

Differential Sensitivity of ERG Components to IOP Challenge

Figure 1 shows the effect of step-wise IOP elevation (protocol 1) on the retinal function of a representative animal. The waveform elicited with the dim flash demonstrates the characteristic STR of mice with a positivity at approximately 120 ms and a negativity at 220 ms. ¹² The response to the brightest flash (0.34 log cd · s · m⁻²) showed an a-wave trough followed by a dominant corneal positive b-wave. Oscillatory potentials can be seen superimposed on the b-wave. 

Figure 1.

 

Representative ERG waveforms measured at dim (A, −4.54 log cd · s · m⁻²), medium (B, −2.23 log cd · s · m⁻²) and bright (C, 0.34 log cd · s · m⁻²) luminous energies. Responses from the IOP challenged eye (thick traces) at progressively increasing IOPs are compared with the waveform recorded at baseline (thin traces). Dotted vertical lines: represent the fixed times at which the amplitudes of pSTR (120 ms), nSTR (220 ms), b-wave (110 ms) and a-wave (8 ms) were measured.

Consistent with a study in rats, ¹⁰ increased IOP attenuated the ERG in a dose-dependent manner. The STR was affected at IOP levels that were lower than those needed to reduce other scotopic ERG components. In particular, the pSTR was reduced at 30 mm Hg, whereas the a- and b-wave remained largely unaffected. By 80 mm Hg, only a small slow negative a-wave remained with all other ERG components abolished. 

In addition to the reduction in amplitude, ERG components showed an increasing delay with higher IOPs. Delay in peak times of up to 13 ± 4 ms for pSTR, 38 ± 11 ms for nSTR, and 9 ± 4 ms for b-wave were found when IOP was elevated from baseline to 45 mm Hg (one-way ANOVA, P = 0.001, P = 0.04, P = 0.03, respectively). As such, ERG component amplitudes were analyzed at fixed times to better capture the combined amplitude and timing changes. 

To consider the relative sensitivity of ERG components to IOP elevation, data have been expressed relative to baseline. Figure 2A shows that IOP elevation had similar effects on both pSTR and nSTR (two-way ANOVA, P = 0.46). 

Figure 2.

 

Relative change in ERG amplitude at various IOP levels (mean ± SEM, n = 6). (A) Inner retinal components of the ERG (pSTR, nSTR, and OP amplitudes) in response to IOP. Pressure–function relationships for pSTR (thin solid line) and OP (thick solid line) are shown. (B) Comparison of pSTR, b- and a-wave responses to IOP. Pressure–function relationships for pSTR (thin solid line), b-wave (dashed line), and a-wave (thick solid line) were described by using a cumulative normal function.

Figure 2B shows that the pSTR pressure–function relationship was leftward of the b- and a-wave and had a steeper slope, suggesting greater pSTR sensitivity to IOP manipulation. This was confirmed with the pSTR pressure–function relationship returning a mean (μ) significantly below the lower confidence limit for the b- and a-wave (Table 1). In addition, the cumulative normal function fitted to the pSTR returned an SD (σ) that was significantly smaller than the lower limit for the a-wave σ, indicating a steeper slope. The nSTR was not significantly more sensitive than the b-wave (Table 1). The b-wave pressure–function relationship also had a significantly lower mean and steeper slope than did the a-wave (Table 1). 

Table 1.

 

Parameters and 95% Bootstrap Confidence Limits for the Cumulative Normal Function Used to Describe the Amplitude-Pressure Relationship

ERG Component	Mean (mm Hg) μ	95% CL	SD (σ)	95% CL
pSTR	41.2	37.7, 45.6	10.8	4.6, 16.1
nSTR	45.0	41.0, 49.1	12.2	7.0, 15.7
b-Wave	45.1	41.4, 49.1	12.2	6.5, 15.9
a-Wave	51.0	47.6, 55.3	19.1	14.5, 23.0
OP	55.1	49.6, 59.2	7.2	1.6, 11.0

Figure 2A shows that scotopic OP amplitudes were significantly less sensitive to IOP elevation compared to the pSTR (two-way ANOVA, P < 0.001) and nSTR (P < 0.001). Further analysis of IOP effects on OP characteristics was achieved by fitting the filtered OP waveforms with the Gabor model. Figure 3 shows that IOP-related reduction in OP amplitude was accompanied by a reduction of OP frequency (by −19 ± 4 Hz, ANOVA, P = 0.03), delay in OP peak time (from 34 ± 1 to 55 ± 11 ms, P = 0.05) and increased OP waveform spread (from 4 ± 3 to 16 ± 4 ms, P = 0.007). There was also a delay in phase for IOP levels of 35 mm Hg (phase change = −1.94 ± 0.55, t-test, P = 0.02) and 40 mm Hg (phase change = −1.63 ± 0.60, P = 0.04). IOPs of 70 and 80 mm Hg resulted in OP amplitudes that were within the noise of the system (±10 μV), therefore returning unreliable Gabor parameters; these recordings have been omitted from the figure. 

Figure 3.

 

Changes in scotopic OPs at various IOP levels (mean ± SEM, n = 6). Parameters are derived by fitting a Gabor model to the extracted OP waveform recorded to 0.34 log cd · s · m⁻². (A) Gabor function is composed of a Gaussian envelope and sine wave carrier. (B) Representative extracted OP waveforms obtained at baseline IOP, 35, 50, and 80 mm Hg. (C) Relative OP amplitude; (D) OP frequency h; (E) OP peak time m; (F) overall spread s of the oscillations in response to increasing IOP. The data for parameters m, h, s at IOP of 70 and 80 mm Hg have been omitted, as the small amplitudes of OPs at these IOP levels (see B) preclude a sensible fit of the Gabor model.

Rate of Functional Impairment and Recovery from an IOP Spike

To fully characterize the response of ERG components during and after an acute IOP spike, we subjected treated eyes to a fixed IOP of 50 mm Hg for 30 minutes followed by return to baseline for 60 minutes (protocol 2). Figure 4 shows the STR and scotopic bright-flash waveforms for a representative animal at baseline, during IOP challenge and at various times during recovery. Attenuation in the STR and b-wave amplitude was evident at 30 minutes after IOP spike onset. Functional recovery could be seen after IOP returned to baseline. 

Figure 4.

 

ERG waveforms during and immediately after a single IOP challenge. Representative dim flash (A, −4.54 log cd · s · m⁻²) and rod b-wave flash (B, −2.23 log cd s m⁻²) waveforms at baseline (thin trace), 30 minutes after 50 mm Hg IOP elevation (maximal dysfunction), and at 10, 30 and 60 minutes during recovery (thick traces). Vertical reference lines: fixed times used to quantify the amplitude change for pSTR (120 ms), nSTR (220 ms), and b-wave (110 ms).

Figure 5 shows group data (n = 10) for relative pSTR, nSTR, and b-wave amplitude during and subsequent to the IOP spike. During the IOP spike, all ERG components declined gradually. At 30 minutes, the pSTR, nSTR, and b-wave amplitudes were reduced by 56% ± 8%, 49% ± 13%, and 43% ± 6%, respectively, relative to baseline. Significant delays in peak time for pSTR and nSTR components were also observed during IOP elevation (Figs. 5B, 5C). When compared to the gradual step-wise IOP elevation to 50 mm Hg (protocol 1), the 50 mm Hg IOP spike (protocol 2) showed a similar degree of pSTR, nSTR, and b-wave amplitude loss (t-test; P = 0.32, P = 0.56, P = 0.22, respectively). 

Figure 5.

 

Functional changes during and immediately after a single IOP spike of 50 mm Hg for 30 minutes (mean ± SEM, n = 10). Hatched area in the box below each graph indicates the duration of IOP elevation. Open area in the box indicates 60 minutes of recovery, with IOP returned to baseline (∼12 mm Hg). (A) Relative amplitudes (treatment/baseline, %) for pSTR, nSTR, and b-wave during IOP insult and recovery. The x-axis for the nSTR and b-wave were jittered by −1 and +1 minute, respectively, to improve clarity. Relative b-wave amplitude for sham (n = 4) experiments is shown, with the shaded area indicating the 95% CL across the 90 minutes of experimentation. (B) Peak time delay for pSTR, (C) nSTR, and (D) b-wave. *Statistically significant delays (post hoc; P < 0.05).

After IOP restoration to baseline, pSTR amplitude recovery was significantly slower compared with b-wave recovery during the 60-minute period (two-way ANOVA, P = 0.02). There was no statistical significance between nSTR and pSTR recovery (two-way ANOVA, P = 0.13). Although nSTR amplitude was within normal limits by 60 minutes of recovery, there remained a significant nSTR peak time delay of 19 ± 5 ms (P < 0.01) (Fig. 5C). There was also significant pSTR delay of 7 ± 2 ms (P = 0.02) at 50 minutes recovery (Fig. 5B). 

Retinal Function 1 Week after IOP Spike

To examine the longer term effects of a single acute IOP spike on ERG components, we reassessed full scotopic ERG recordings 7 days after the single IOP spike used in protocol 2. There was no significant difference in IOP measured before or 7 days after the acute spike (13 ± 1 mm Hg vs. 14 ± 1 mm Hg; t-test, P = 0.54). These value represent the average of three repeated readings on lightly anesthetized animals under room lighting (∼100 lux) using a noninvasive rebound tonometer (iCare Finland, Espoo, Finland). Figure 6 shows that there was no significant change in b-wave (two-way ANOVA, P = 0.46), a-wave (two-way ANOVA, P = 0.09), or nSTR amplitude (two-way ANOVA, P = 0.15) at 7 days after IOP treatment compared with baseline. Phototransduction amplitude (Rm_P3) and sensitivity (S) were similar at baseline and 7 days after the IOP spike (Rm_P3: 355 ± 28 μV vs. 340 ± 36 μV, P = 0.75; S: 1429 ± 117 vs. 1390 ± 91 m² · cd⁻¹ · s⁻³, P = 0.80, data not shown). In contrast, the pSTR was significantly reduced by an average of 30% ± 6% (ANOVA; P < 0.001) compared to baseline (across luminous energies from −5.92 to −4.73 log cd · s · m⁻²). Likewise, OP amplitude was significantly reduced by an average of 27% ± 2% (ANOVA; P < 0.001 across luminous energies from −1.9 to 2.22 log cd · s · m⁻²). 

Figure 6.

 

ERG amplitudes (mean ± SEM, n = 7) versus luminous energy at baseline and 7 days after a single IOP spike for (A) peak amplitudes, (B) OP peak amplitudes, (C) nSTR amplitudes (at 220 ms); and (D) a-wave amplitude (at 8 ms). Relative amplitudes (treated/baseline %) are shown in the lower box of each panel.

Discussion

To our knowledge, this is the first study to describe in detail the pressure–function relationship of wild-type mouse ERG components, during and after acute IOP elevation. We show that the pSTR in mice was more sensitive to acute IOP elevation than those components arising from the middle (the b-wave) and more distal retinal layers (the a-wave, Fig. 2). Recovery of the pSTR after a single IOP spike was also slower compared with that of the b-wave (Fig. 5). In addition, the pSTR and the OPs showed persistent impairment 1 week after a single IOP spike (Fig. 6). 

Pressure–Function Relationship

The pSTR has been shown to have a significant ganglion cell contribution, ^{12,16,33–35} and is the full-field ERG component most sensitive to chronic IOP elevation in both rats and mice. ^1,36 The higher sensitivity of the STR compared with other ERG components during acute IOP elevation is consistent with a rat model of acute IOP challenge. ¹⁴ The pSTR amplitude attenuation was detectable at IOP levels as low as 30 mm Hg. This is similar to results in rats, where STR timing changes were noted at a similar pressure ¹⁰ and in mice subjected to chronic IOP elevation of 35 to 38 mm Hg. ¹ Given that our mice had systolic blood pressures of 105 ± 5 mm Hg, our data indicate that the pSTR becomes affected at an ocular perfusion pressure (blood pressure – IOP) ³⁷ of ∼75 ± 5 mm Hg. Previous studies have found that b-wave changes are seen when ocular perfusion pressure is reduced to 30 to 45 mm Hg. ^38,39 This is consistent with other reports that ocular perfusion pressures as low as ∼35 mm Hg are needed before optic nerve oxygen tension in porcine ⁴⁰ and cat eyes ⁴¹ and cytochrome activity in cat optic nerve becomes significantly reduced. ⁴² It is not clear whether ischemia is the sole mechanism underlying the functional changes we observed at an ocular perfusion pressure of 75 mm Hg. 

Of interest, the reduction in OP amplitudes occurred at IOPs that were higher than that required for nSTR and pSTR, even though OPs are also believed to be generated in the inner retina. ^43–45 During acute IOP elevation, OP amplitude was relatively unaffected at 50 mm Hg yet showed reduced frequency, increased waveform spread, and delayed peak time. We speculate that the lack of OP amplitude loss during mild acute IOP elevation may be related to altered inhibitory feedback circuits, which might enhance OP amplitude. For example, larger OPs are found after pharmacologic inhibition of GABAc receptors ⁴⁶ and in mice lacking inhibitory GABAc receptors. ⁴⁷  

It is worth noting that light-activated responses from inner retinal neurons depend on the integrity of distal retinal elements. As such, IOP-induced deficits in the a-wave, affect the b-wave, which in turn should affect the STR. The degree to which these downstream effects manifest depends on the gain relationships between the various cellular layers. It is known that a reduction in the ERG a-wave can result in a proportional b-wave attenuation. ⁴⁸ Our results show that the b-wave IOP-response function had a significantly lower mean and steeper slope than the a-wave, suggesting that bipolar cells are more sensitive to IOP challenge than photoreceptors. The gain relationship between the outer retinal ERG components and the STR is less well characterized. Should outer retinal deficits be amplified at the ganglion cell level under these circumstances, a more sensitive pSTR pressure–response function could be anticipated. However, our finding that there was selective, persistent pSTR impairment 7 days after IOP challenge in the presence of normal b-waves provides evidence that inner retinal function involving ganglion cells is more susceptible to IOP challenge. 

The mechanism of increased inner retinal sensitivity to IOP elevation is likely to involve many factors, including mechanical, vascular, and oxidative stress. Ganglion cells may be particularly susceptible as they receive insults both at their axons at the level of the optic nerve head ^49–51 and at their cell bodies ^2,52 during IOP elevation. Therefore, additional mechanical and vascular stress at the optic nerve head could overwhelm the bioenergetics required for normal ganglion cell function. ⁵³  

Recovery from IOP Spike

The time course of STR recovery in our study is consistent with previous studies in rats. ^14,54 This was much slower (60 minutes) than could be accounted for by the return of retinal reperfusion, as demonstrated in the porcine retina. ⁴⁰ The processes underlying functional loss and recovery are not well understood, and further work is needed to characterize the factors that influence the rate of this recovery. Of interest, although STR amplitude showed recovery during the 60 minutes after IOP challenge, the delay in peak time remained. A delay in timing of ERG components might reflect either a reduced efficacy in the transduction cascades mediating changes in membrane potential, or reduced efficacy of neurotransmission from distal neuronal components. ⁴⁸ These possibilities require further investigation. 

The attenuation of the pSTR and OPs at 7 days after an IOP spike shows that a single episode of stress can produce a selective inner retinal deficit in mice. As we have yet to assess ganglion cell anatomy at 7 days after the IOP challenge, it is not clear whether the functional reduction reflects cell loss or there was dysfunction in a normal cohort of inner retinal cells. 

Limitations

The approach used in this study does not attempt to model the chronic IOP elevation observed in glaucoma; however, it does highlight the impact of acute IOP-induced stress on retinal function. In rat and mouse models of chronic IOP elevation, IOP can regularly spike up to and above 30 mm Hg. ^1,5,36,55 Thus, in these chronic rodent models there are likely to be periods of acute ganglion cell dysfunction. It is clear that rodent eyes have different anatomic and biomechanical properties; thus, extrapolation of our outcomes to other species should be made cautiously. 

The full-field ERG is unable to identify region-specific sensitivity to IOP challenge. This would be particularly important in humans and animal models that have a fovea or area centralis. ⁴¹ A functional modality such as the multifocal ERG may be more useful as a function marker of regional stress. 

Summary

We showed that in mice the ganglion cell–dominated pSTR is more sensitive to acute IOP spikes than are ERG components arising from more distal retinal cell classes. This increased sensitivity manifests as leftward shifts as well as a steeper slope of the pSTR IOP-response function; a slower recovery immediately after a constant IOP challenge; and a selective pSTR and OP loss 7 days after a single 50-mm Hg IOP spike. These data also suggest that chronic IOP may not be needed to generate permanent dysfunction. The ability to simultaneously assess retinal function during and after acute IOP challenge provides a means to quantify the susceptibility of retinal neurons to IOP stress. 

Supplementary Materials

Supplementary Material - (PDF) 

Footnotes

 Supported by National Health and Medical Research Council Grants 475603 (JGC) and 400127 (BVB); Ophthalmic Research Institute of Australia grants; the Glaucoma Australia Fund; the Henry Greenfield Research Fund; and the Cranbourne Trust Fund.

Footnotes

 Disclosure: Y.X. Kong, None; J.G. Crowston, None; A.J. Vingrys, None; I.A. Trounce, None; B.V. Bui, None

Footnotes

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

The authors thank Karl A. J. Bromelow, Zheng He, and Vickie Wong for assistance in the daily care and monitoring of the animals.