The Pattern Electroretinogram as a Tool to Monitor Progressive Retinal Ganglion Cell Dysfunction in the DBA/2J Mouse Model of Glaucoma

Abstract

purpose. To determine the baseline characteristics, reliability, and dynamic range of the pattern electroretinogram (PERG) as a tool to monitor progressive RGC dysfunction in the DBA/2J mouse model of glaucoma with spontaneously elevated intraocular pressure (IOP).

methods. PERGs were recorded from 56 undilated eyes of 28 anesthetized (ketamine-xylazine-acepromazine) DBA/2J mice of different ages (2–4 months, n = 44 eyes; 12–14 months, n = 12 eyes) in response to contrast reversal of gratings that maximize PERG amplitude (95% contrast, 1-Hz reversal, 0.05 cyc/deg spatial frequency, 50° × 56° field size). Robust averaging (1800 sweeps) was used to isolate PERG from background noise. Cone-driven ERGs in response to diffuse light flashes superimposed on a rod-adapting background (FERG) were also recorded.

results. PERGs had consistent waveforms and were reproducible across batches of mice and operators. In 2- to 4-month-old mice (prehypertensive stage), the PERG amplitude (mean, 8.15 ± 0.4 μV [SEM]) was considerably larger than the noise (mean 1.18 ± 0.1 μV). The test–retest variability (two different sessions 1 week apart) and interocular asymmetry of PERG amplitude was approximately 30%, and that of PERG latency was approximately 17%. In 12- to 14-month-old mice (advanced hypertensive stage) the PERG amplitude (mean, 1.29 ± 0.12 μV) was close to that of noise. In 12- to 14-month-old mice the FERG was reduced to a lesser extent compared with the PERG.

conclusions. The PERG has an adequate signal-to-noise ratio, reproducibility, and dynamic range to monitor the progression of functional changes in the inner retina in DBA/2J mice.

The use of mouse models of glaucoma is currently limited by the paucity of noninvasive methods for monitoring retinal ganglion cell (RGC) dysfunction and its progression. Electrical activity in the retina associated with RGC function can be evaluated by means of the pattern electroretinogram (PERG). The PERG is obtained in response to contrast reversal of patterned visual stimuli (gratings, checkerboards), rather than uniform flashes of light. ¹ RGCs are necessary for generating the PERG, because their selective loss after optic nerve transection in mammals, ^{2 3} including mice, ⁴ abolishes the response. In experimental primate models of optic nerve transection ³ and glaucoma, ⁵ the amount of PERG amplitude reduction is consistent with the degree of damage apparent by counting either RGCs or optic nerve fibers. In the same experimental animals, the a- and b-waves of the conventional bright flash-ERG are only slightly or not affected. An important distinguishing characteristic of the PERG is that it requires the physiological integrity of viable RGCs to be generated. The PERG amplitude can be reversibly reduced by intravitreal injections of tetrodotoxin which block the spiking activity in the inner retina. ^{6 7} Short-term elevation of the intraocular pressure (IOP) has the effect of reversibly reducing the PERG amplitude but not the flash ERG. ^{8 9} Altogether, these findings indicate that reduction in PERG amplitude may reflect both the reduced activity of dysfunctional, yet viable, RGCs as well as the lack of activity of lost RGCs. The PERG, therefore, may represent an important tool for detecting and monitoring the onset and the progression of RGC dysfunction in mouse models of glaucoma. In addition, the PERG allows the establishment of retinal resolution and contrast threshold, ^{4 10} which have a counterpart in cortical visual acuity and contrast sensitivity. ¹¹ Finally, the PERG stimulus can be precisely located on the retina, thereby allowing quantitative structure–function correlations for specific retinal regions, ¹² as well as topological assessment of RGC losses in glaucomatous DBA/2J mice. ¹³ These PERG properties may provide important information to characterize glaucomatous disease and the effect of neuroprotective agents on retinal function. For these reasons, we focused on the PERG rather than other potentially useful techniques in rodents, such as the scotopic threshold response (STR) of the full-field flash ERG. ^{14 15 16}  

Despite its great potential for monitoring RGC dysfunction in glaucoma models, the PERG has not been widely adopted, probably because it is considered technically difficult to record. ¹⁴ The purpose of this study was to determine the baseline characteristics, reliability, and dynamic range ¹¹ of an optimized version of the PERG as a tool for the systematic monitoring of progressive RGC dysfunction in the DBA/2J mouse model of glaucoma with spontaneously elevated IOP. ^{17 18} DBA/2J mice develop an iris disease characterized by iris atrophy and pigment dispersion due to mutations of two genes: Gpnmb and Tyrp1. ¹⁹ The iris disease is apparent at 6 months and progresses with age, resulting in elevated IOP. ¹⁸ Young (2- to 4-month-old) DBA/2J mice have normal IOP and normal histologic appearance of RGCs and optic nerve. ^{17 18} Glaucomatous damage in the optic nerve is apparent in approximately 75% of eyes by 13 months ¹⁸ and in approximately 90% of eyes by 18 months, ¹⁸ and appears to be IOP dependent. ²⁰ By 16 months, cataract may develop. ¹⁷ Preliminary results of this study have been shown in abstract form (Porciatti V et al. IOVS 2006;47:ARVO E-Abstract 4011). 

Methods

A total of 28 DBA/2J mice (n= 56 eyes) of both sexes and different ages obtained from Jackson Laboratories (Bar Harbor, ME) were recorded. All animals were treated according to protocols established by the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. The experimental protocol was approved by the Animal Care and Use Committee of the University of Miami. Mice were anesthetized with intraperitoneal injections (0.5–0.7 mL/kg) of a mixture of ketamine, 42.8 mg/mL; xylazine, 8.6 mg/mL; and acepromazine, 1.4 mg/mL. The head of anesthetized mice was gently restrained in a modified stereotaxic apparatus, ¹¹ in which the tips of the ear bars are covered with flexible plastic material. A mouth bite and a small nose holder, which allows unobstructed vision, were used to restrain the anterior part of the head. The body of the mouse rested on a feedback-controlled heating pad that maintained the body temperature at 37°C. Under these conditions, the eyes of the mice were naturally proptosed, with undilated pupils pointing laterally and upward. 

PERG Recording

The recording electrode was made of a thin silver wire (0.25 mm diameter) configured to a semicircular loop of 2 mm radius (Fig. 1) . Electrodes were simply obtained by bending the silver wire around a screwdriver with 2-mm diameter. Electrodes for the right eye and left eye had mirror geometry, with the vertical portion of the electrode connected with the electrode holder being on the temporal side of the eye in order not to interfere with vision. The electrodes were gently leaned on the corneal surface by means of a micromanipulator under microscopic control and positioned in such a way as to encircle the undilated pupil without limiting the field of view (Fig. 1) . Electrode positioning entailed minimal manipulation of the eye. The thin diameter of the silver wire allowed the electrode to follow small forward–backward eye movements associated with breathing that might otherwise increase the noise and even cause corneal abrasion. Corneal stimulation is known to induce cataract in the mouse, which precludes further PERG testing. ²¹ This electrode placement enabled us to check for possible development of cataract, as well as controlling that the pupil projection was centered on the pattern stimulus. Instillations of balanced saline every 30 minutes were sufficient to maintain the cornea and lens in excellent condition for the duration of the recording (∼1 hour). Extensive experience with mouse PERGs ⁴ has shown this simple electrode geometry and placement to give the most consistent results. Reference and ground electrodes were small stainless-steel needles inserted in the skin of the back of head and the back of the body, respectively. 

Previous studies in C57BL6/J mice have shown that the PERG amplitude is maximal at 0.05 cyc/deg and low reversal frequency (1 Hz) (Porciatti V et al. IOVS 2003;44:ARVO E-Abstract 2705). The PERG amplitude progressively decreases with increasing temporal frequency (temporal resolution approximately 13 Hz) (Porciatti V et al. IOVS 2003;44:ARVO E-Abstract 2705), increasing spatial frequency (spatial resolution approximately 0.5 cyc/deg) (Porciatti V et al. IOVS 2003;44:ARVO E-Abstract 2705), ^{4 10} decreasing contrast (contrast threshold approximately 10%) (Porciatti V et al. IOVS 2003;44:ARVO E-Abstract 2705), ⁴ decreasing luminance (Porciatti V et al. IOVS 2003;44:ARVO E-Abstract 2705), and decreasing stimulus area (Porciatti V et al. IOVS 2003;44:ARVO E-Abstract 2705). The spatiotemporal properties of the 2- to 4-month-old DBA/2J mouse PERG closely corresponded to those reported in C57BL6/J mice (not shown in figures). To maximize PERG amplitude, patterned visual stimuli were therefore high-contrast (98%), horizontal gratings of 0.05 cyc/deg spatial frequency with square-wave profile, abruptly reversing in spatial phase at 1 Hz (two contrast reversals/s) and 50 cd/m² mean luminance. Stimuli were generated by a video card (VSG3; Cambridge Research Systems Ltd, Kent, UK) and displayed on a 21-in. TV monitor (Multiscan G500; Sony, Tokyo, Japan). To further maximize PERG amplitude, a large stimulus area (50.2° × 58.0°) was obtained by presenting pattern stimulus from a short distance (20 cm). The stimulus display contained three full stimulus cycles. Each bar subtended a visual angle of 8.4°, which approximately corresponds to the average size of the receptive field centers of mouse retinal ganglion cells. ^{22 23} The center of the visual stimulus was aligned with the projection of the pupil. The PERG is a light-adapted response. To have a corresponding index of outer retinal function, a light-adapted ERG (FERG) was also recorded with undilated pupils. Uniform stimuli for FERG recording consisted of strobe flash stimuli of 20 cd/m² per second superimposed on a steady background light (12 cd/m²) and presented within a Ganzfeld bowl. Under these conditions, rod activity is largely suppressed, while cone activity is minimally suppressed. ²⁴  

Recording Protocol

Anesthetized mice were mounted on the holder, a corneal electrode was placed on the right eye, a small drop of balanced saline was added to prevent corneal dryness, and body temperature was allowed to stabilize at 37°C. Typical protocol consisted of a series of three consecutive responses to 600 contrast reversals each (PERG recording). The responses were superimposed to check for consistency and then averaged. Overall, 1800 sweeps were averaged, thereby reducing the uncorrelated electrical activity by a factor of \(\sqrt{1800{=}}\) 42. To have an estimate of the residual noise, the PERG recording was preceded by an additional recording of 600 sweeps, in which the contrast was set to zero. Immediately after PERG recording, the stimulus was switched to the strobe flash. Three consecutive responses to 30 flashes each (FERG) were recorded. After visual inspection for consistency, responses were averaged (the uncorrelated electrical activity was reduced by a factor of \(\sqrt{90\ {=}}\) 9.5). After PERG and FERG recording from the right eye, additional anesthesia was given as needed, a new electrode was placed on the left eye, and the procedure was repeated. The entire protocol required approximately 1 hour. Averaged PERGs and FERGs were automatically analyzed to evaluate the major positive and negative waves (Fig. 1) . For both PERG and FERG, response amplitude was evaluated peak-to-trough, and the latency was the time-to-peak of the major positive deflection. 

Results

PERG and FERG waveforms obtained in 44 eyes of 22 DBA/2J mice aged 2 to 4 months (preglaucoma stage) are displayed in Figure 2 . Thin lines correspond to individual waveforms, and thick lines correspond to the averaged waveform. PERG and FERG waveforms were reasonably reproducible in different eyes. Note that consistent waveforms were obtained from all tested eyes. Also note that two different batches of mice, recorded by two different operators, had very similar PERG and FERG average waveforms. 

Amplitudes and latencies of individual waveforms are displayed in Figure 3as amplitude–latency scattergrams. Because PERGs and FERGs have markedly different amplitudes and latencies, log₁₀ scales were used to compare them on the same plot. Note in Figure 3that, on average, PERGs have smaller amplitudes and longer latencies than do cone-driven FERGs. However, the spread of data is similar for both kinds of responses. 

Interindividual Variability

Mean PERG and FERG amplitudes are reported in Table 1 . Note that the coefficients of variation (SD/mean) are comparable between the responses in the two eyes and that the coefficient of variation (CV) of PERG amplitude and latency are of the same order of magnitude as those of the FERG. For both PERG and FERG, the CV of latency is smaller than that of amplitude. As shown in Figure 3 , PERG and FERG amplitudes can vary over a wide range among individuals. If an eye with a large PERG also had a large FERG, then normalizing the PERG amplitude to the FERG amplitude (computing a PERG/FERG amplitude ratio) would in principle reduce interindividual variability. However the Pearson correlation between PERG amplitude and FERG amplitude was weak (R= 0.23, P = 0.15). The CV of the PERG/FERG ratio (Table 1)was of the same order of magnitude or even larger than that of raw PERG and FERG amplitudes. Therefore there is no apparent advantage in computing a PERG/FERG amplitude ratio. 

Reproducibility of PERG

A subset of eight mice aged 2 to 4 months has been tested twice in two different sessions with a 1-week interval. The results of the first session have been included in the core sample (Table 1) . We compared the variation occurring between the responses of the same eyes recorded in the two different sessions (test retest variability) with that occurring between the right and the left eyes recorded in sequence (about one-half hour apart) during the same session (interocular asymmetry). Variation was calculated according to the formula [variation = |test1 − test 2|/½(test 1 + test 2)]. The results are reported in Table 2 . For both PERG and FERG amplitudes, the test-retest variability and interocular asymmetry were on the order of 0.3 (30%). The variability of PERG and FERG latency was typically smaller than that of PERG and FERG amplitude. As for interindividual variability, the within-individual variability of the PERG/FERG amplitude ratio was larger than that of PERG and FERG amplitude. 

Dynamic Range of PERG

The dynamic range of PERG amplitude can be defined as the interval between the level of signal when a normal number of functioning RGCs are present (i.e., young DBA/2J mice) and the level of noise (the PERG amplitude in response to a pattern stimulus with zero contrast). 

PERG and FERG waveforms obtained in 12 eyes of 6 DBA/2J mice aged 12 to 14 months (advanced glaucoma stage) are displayed in Figure 4 . Thin lines correspond to individual waveforms, and thick lines correspond to the averaged waveform. Note that PERGs of old mice have dramatically reduced amplitude compared with those of young mice depicted in Figures 2A and 2B . By contrast, the FERG is still recordable, although with reduced amplitude. The average latency and latency scatter are approximately similar to those of young DBA/2J mice. 

Average PERG and FERG amplitudes of young and old mice evaluated from individual waveforms are compared in Figure 5 . The PERG amplitude in 2- to 4-month-old mice (mean, 8.15 ± 0.4 μV [SEM]; n = 44 eyes) was higher than the noise amplitude (mean, 1.18 ± 0.1 μV; n =12) by a factor of 6.9. The PERG amplitude in older mice (1.29 ± 0.12 μV; n = 12) was close to that of the noise. Overall, the PERG amplitude was reduced by 6.8 μV (84%) between 3 and 13 months of age in DBA/2J mice. If the percentage of amplitude loss is rescaled on the response dynamic range (difference between average amplitude of young mice and the average noise amplitude), then the PERG amplitude of 13-month-old mice appears to have lost approximately 98% of its dynamic range. Older mice presented typical iris stromal atrophy, slightly enlarged pupil, and moderate corneal precipitates. ¹⁸ Corneal precipitates involved only minimally the pupil area, however, and no apparent cataract was present. It is unlikely that reduced PERG amplitude in older mice is due optical factors that degrade the stimulus contrast. At 13 months of age, the FERG is also significantly (P < 0.01) reduced in amplitude by 38%. The amount of FERG amplitude reduction is less than that of the PERG, indicating that outer retina function may be relatively spared in advanced stages of glaucoma. A previous study reported changes of comparable magnitude in the diffuse-flash ERG in old DBA/2NNia mice. ²⁵  

Discussion

Functional outcome measures are sorely needed for following the progression of glaucoma in mice and for assessing therapeutic effects of potential neuroprotective treatments. In this study, we determined the baseline characteristics, reproducibility, and dynamic range of an optimized version of the PERG as a tool to monitor progressive RGC dysfunction in the DBA/2J mouse model of glaucoma. The PERG technique may have the potential to detect RGC dysfunction or loss specifically, because RGC death induced by optic nerve transection in the C57BL/6J mouse abolishes the PERG response, whereas it leaves the luminance-flicker ERG unaltered. ⁴ The present study indicates that by approximately 13 months of age, the PERG is virtually abolished in the DBA/2J mouse glaucoma model. PERG alterations are unlikely to depend on optical factors such as corneal and lens opacities or defocus. A detailed histologic study of the DBA/2J mouse retina with advanced glaucoma revealed no cell loss in any retinal neurons other than ganglion cells. ¹³ In that study, no group of RGCs was found to be especially vulnerable or resistant to degeneration. Axonal atrophy, dendritic remodeling, and somal shrinkage seemed to precede ganglion cell death. It is conceivable that PERG losses found in our study are associated with RGC loss and/or abnormalities of RGC architecture. We cannot exclude, however, a possible association between PERG losses and ischemic effects as well as glial activation reported for some old DBA/2J mice obtained from Charles River (Sulzfeld, Germany). ²⁶  

Optimization of the PERG Technique

To provide a robust index of RGC function for systematic use in neuroprotection studies, in this study we used a PERG protocol that maximizes the response amplitude and minimizes the residual noise. Under the present recording conditions, the PERG is recordable in preglaucomatous DBA/2J mice with high signal-to-noise ratio (∼7:1, on average). This level of signal to noise is considerably higher than that reported in a previous pilot PERG study of DBA/NNia mice, in which different experimental conditions were used. ²⁷ The absolute level of PERG amplitude in preglaucomatous DBA/2J mice is on the same order of magnitude as that reported for the rat STR. ¹⁴  

Response Reproducibility

Comparable PERG waveforms were obtained in all tested eyes of 2- to 4-month-old mice, independent of whether they belonged to different batches or were recorded from different operators. In addition, PERG waveforms were comparable between eyes of the same animals during the same session and between the same eyes recorded in different sessions. 

Interindividual Variability

There was a wide range of PERG amplitudes among eyes of 2- to 4-month-old mice, resulting in a CV (SD/mean) of approximately 30%. This value is at least twice as large as the interindividual variability of RGC number reported in young DBA/2J ²⁸ and young DBA/2NNia ²⁹ and suggests that there may be other sources of PERG variability, such as anesthesia, electrode placement, and residual noise. A CV of 30%, however, is on the same order of magnitude as that reported for human PERG ^{30 31 32} as well as for human ordinary flash-evoked ERG. ^{33 34} In our own study, the interindividual variability of cone-driven flash ERG amplitude is comparable to that of the PERG amplitude. Similar variability for FERG and PERG would exclude optical factors (incorrect stimulus alignment, reduction in stimulus contrast due to image blurring) as a major source of PERG variability, because the flash-ERG is virtually independent from these optical factors. 

Test–Retest Variability

The intersession variability of PERG amplitude is approximately 30%. These values are somewhat larger than those reported for the monkey PERG ³⁵ and some studies of human PERG. ^{36 37} However, other human PERG studies report similar ranges of variability. ^{38 39 40}  

Interocular Asymmetry

The progression of glaucoma is often asymmetrical in both humans ⁴¹ and DBA/2J mice. ¹⁸ Differences in PERG amplitude and latency between the two eyes may therefore represent an important criterion in the evaluation of changes in RGC function. In future neuroprotection studies, the level of interocular asymmetry may set the sensitivity limit for comparing the effect of treatment between the treated- and the sham-treated eye. Our results show that the variability due to interocular asymmetry (within session) is comparable to the test retest variability (between sessions). It has to be taken into account that in the present protocol, the two eyes are recorded in sequence during the same session. Therefore, it is possible that animal conditions (primarily the level of anesthesia) vary during the same session to an extent comparable to that occurring when testing the same eyes in two different sessions. This may limit the precise measurement of interocular asymmetry. We are currently testing a new system in which the PERG is simultaneously recorded from both eyes. This method would allow recording both eyes under identical conditions and would shorten the recording time. Preliminary results obtained by simultaneous PERG recording of both eyes indicated that PERG interocular asymmetry in young DBA/2J mice is substantially smaller than that obtained using the present protocol. 

Dynamic Range

The PERG amplitude of 12- to 14-month-old mice is much reduced compared with that in 2- to 4-month-old DBA/2J mice, having lost approximately 98% of its dynamic range. This finding suggests that, after 1 year of age, the residual activity of RGCs and/or the inner retina circuitry impinging on them is small or absent. At 13 months of age, histologic damage to the optic nerve is apparent in approximately 75% of eyes. ¹⁸ Preliminary results of our group (Porciatti V et al. IOVS 2006;47:ARVO E-Abstract 4011; Porciatti V et al. IOVS 2003;44:ARVO E-Abstract 2705) indicate that in DBA/2J mice the PERG is already much reduced between 10 and 11 months of age, when histologic damage is evident in a minority (40%) of eyes. This observation suggests that RGC dysfunction may precede histologically visible damage of RGCs and the optic nerve. 

Conclusions

The PERG can be reliably recorded in DBA/2J mice. With our optimized protocol, the PERG has adequate signal-to-noise ratio, reproducibility, and dynamic range. In addition, the PERG is dramatically reduced in glaucomatous mice. These represent adequate conditions to monitor progression of functional changes of inner retina in DBA/2J mice. We are currently performing a longitudinal PERG study indicating that monitoring progression of RGC dysfunction is indeed possible in individual DBA/2J mice. Success in this study would open the possibility of using the DBA/2J mouse model as a template for systematic studies of neuroprotection, including functional endpoints. It remains to be established whether the PERG represents an adequate technique for monitoring RGC dysfunction in other mouse glaucoma models. ⁴²  

 Supported by National Eye Institute Grant R03 EY016322 and Center Grant P30-EY14801 and an unrestricted grant to the University of Miami from Research to Prevent Blindness, Inc. MS was supported by Allergan, Inc.

 Submitted for publication June 29, 2006; revised August 31 and September 29, 2006; accepted December 21, 2006.

 Disclosure: V. Porciatti, None; M. Saleh, Allergan (F); M. Nagaraju, None

 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.

 

Figure 1.

 

Geometry and placement of the PERG electrode.

Figure 2.

 

PERG and FERG waveforms recorded in 22 young (3- to 4-month-old) DBA/2J mice (n= 44 tested eyes). Waveforms shown in (A) and (C) and in (B) and (D) were obtained by different operators from independent batches of mice. Thin lines: individual waveforms. Thick lines: average waveform.

Figure 3.

 

Amplitude-latency scattergrams of PERGs and FERGs recorded from 44 eyes of 22 DBA/2J mice of 2 to 4 months of age.

Table 1.

 

Average PERG and FERG Amplitudes, Latencies, and PERG/FERG Amplitude Ratios Evaluated in 22 DBA/2J Mice Aged 2 to 4 Months

Right Eye (n = 22)	Mean	SD	CV	Left Eye (n = 22)	Mean	SD	CV
PERG amplitude	8.25	2.51	0.30	PERG amplitude	8.05	3.06	0.38
PERG latency	154.18	32.68	0.21	PERG latency	159.00	30.75	0.19
FERG amplitude	67.37	27.85	0.41	FERG amplitude	49.56	16.90	0.34
FERG latency	74.25	17.00	0.23	FERG latency	86.76	12.90	0.15
PERG/FERG amplitude ratio	0.14	0.05	0.36	PERG/FERG amplitude ratio	0.18	0.08	0.44

Table 2.

 

Variation between the Two Eyes within the Same Session (Interocular Asymmetry) and between the Same Eyes in Two Sessions (Test-Retest Variability)

	Interocular Asymmetry	SD	Test-Retest Variability	SD
PERG amplitude	0.30	0.23	0.29	0.21
PERG latency	0.16	0.15	0.17	0.20
FERG amplitude	0.25	0.15	0.34	0.23
FERG latency	0.20	0.12	0.21	0.16
PERG/FERG amplitude ratio	0.4	0.31	0.49	0.27
Number of observations	16		16

 The variation is determined by the calculation [|test 1 − test 2|/½ (test1 + test 2)].

Figure 4.

 

PERG (A) and FERG (B) waveforms recorded in six old (12- to 14-month-old) DBA/2J mice (n= 12 tested eyes). Thin lines: individual waveforms. Thick lines: average waveform.

Figure 5.

 

Mean PERG (A) and FERG (B) amplitudes recorded in young (2- to 4-month-old; n= 44 eyes) and aged (12- to 14-month-old, n= 12 eyes) DBA/2J mice. Noise represents the PERG amplitude recorded in young mice (n= 12 eyes) in response to a stimulus of zero contrast. Error bars, SEM.