A modified silent substitution electroretinography protocol to separate photoreceptor subclass function in lightly sedated dogs.

Abstract

Objective.

A previously published study successfully isolated photoreceptor responses from canine rods, long/medium wavelength (L/M) cones and short wavelength (S) cones using silent substitution electroretinography (ERG) performed under general anesthesia. We hypothesized that responses would be similar in dogs under sedation and that a curtailed protocol suitable for use in clinical patients could effectively isolate responses from all three photoreceptor subtypes.

Animals studied

Three normal adult purpose-bred beagles (2 females,1 male)

Methods

Dogs were dark adapted for 1 hour. Sine wave color stimuli were delivered using LEDs in a Ganzfeld dome. The ERG protocol under anesthesia was performed as previously published; sedated ERG protocols were performed after a 3 day washout period. Intravenous sedation (dexmedetomidine 1.25mcg/kg, butorphanol 0.1mg/kg) was administered for sedation. Statistical analysis was performed using two-way repeated measures ANOVA and linear regression.

Results

In both anesthetized and sedated animals, rod-derived responses peaked at low frequency (4–12Hz), L/M-cone responses peaked at high frequency (32–38Hz), S-cone responses peaked at low frequency (4–12Hz). The frequencies eliciting maximal responses were similar in sedated and anesthetized protocols, although rod amplitudes were significantly higher in the sedated protocols compared with anesthetized (P < 0.001).

Conclusion

We present a clinically applicable method to consistently isolate rod, and cone subclass function in sedated dogs. This may allow detailed evaluation of photoreceptor function in clinical patients with rod or cone subclass deficits without the need for general anesthesia or protracted adaptation times.

Introduction

The dog retina is rod dominant [1] and color vision is dichromatic, with the retina containing rods with rhodopsin maximally sensitive to 508 nm wavelength light and cone subtypes with opsins sensitive to long/medium wavelength light (555nm; L/M-cones) and short-wavelength light (429–435 nm; S-cones). (Figure 1A) [2, 3] Traditional electroretinography (ERG) uses white light of different intensities and with dark- or light-adaptation states to differentiate rod photoreceptors (sensitive to low luminance, maximal responses in the dark-adapted state, minimized or ablated responses in the light-adapted state) from cone photoreceptors (sensitive to higher luminance, isolated responses in the light-adapted state). However, separation of L/M- and S-cone subclass function using ERG in dogs has only recently been proven. [4]

Figure 1.

The principle of univariance and silent substitution. Canine photopigment spectral sensitivities are shown as curves (A) with S-cone opsin peak sensitivity at 430 nm, rhodopsin at 508 nm and L/M-cone opsin at 555nm (data extrapolated from [2, 3], with consideration for preretinal absorbtion of for example UV light absorbed by the cornea and lens). A example single wavelength of light stimulation of maximum power is shown at 480 nm (gray vertical line). In this example, the stimulus would elicit responses in rhodopsin, S-cone opsin and L/M opsin that are 78%, 38% and 18% of the maximally possible response respectively (represented by horizontal dashed colored lines). In the example in (B), the principle of univariance is demonstrated. The principle of univariance states that the responses of a photoreceptor after absorption of a single photon is always the same, independent of the wavelength of that photon. At the peak sensitivity wavelength of rhodopsin at 508 nm, a 50% response is elicited using a stimulus of 50% power (a), but when the wavelength of light is 565nm, a stimulus of 100% is required to elicit a similar 50% response, because the relative sensitivity of rhodopsin at that wavelength is 50% (b). Silent substitution leverages the principle of univariance by providing alternating (sine wave in different phase) lights of differing wavelength and intensity (contrast). In (C), two 100% power stimuli are administered: (a) – at 457 nm and (b) at 565 nm, which would each result in 50% stimulation of rhodopsin (black horizontal dashed line), but (a) would elicit a 77% S-cone opsin stimulation, whereas (b) would elicit 0% S-cone opsin stimulation. Rhodopsin contained in rods would not respond to an exchange between these two lights – the substitution is “silent” for this photoreceptor. If these lights (a) and (b) were presented in a sine wave stimulus with phases offset by 180° (D, upper 2 panels), rod (rhodopsin) responses would remain the same throughout the sine wave, whereas S-cone responses would oscillate (lower 2 panels). This would result in a single photopigment (rhodopsin) single silent substitution. Double silent substitution (and subsequent single photopigment response isolation) can be achieved by the phase offset admixture of lights of 3 different wavelengths and intensities, as shown in Table 1.

This ERG differentiation was achieved using silent substitution, which operates on the principle of univariance: the hyperpolarization of a photoreceptor after a photopigment is isomerized is identical, irrespective of the wavelength of the photon of light that causes the isomerization. (Figure 1B) [5] This means that if two different wavelengths of light are exchanged (e.g. in a sine wave pattern) without altering the number of photopigment isomerizations, this exchange (or substitution) does not result in a change in photoreceptor hyperpolarization (i.e. the substitution is silent). (Figure 1C–D) The investigator can isolate the response of one photoreceptor type by silent substitution of the other classes of photoreceptors whilst isolating the response of the photoreceptor of interest. This method also has the advantage that isolation can be achieved without the need for substantial adaptation times.

Although silent substitution photoreceptor subclass isolation using ERG has been proven in dogs, the method published was lengthy and performed under general anesthesia, limiting its potential applicability in clinical patients with retinal disorders.[4] This is an important limitation, considering the wealth of inherited retinal disorders described in dogs [6, 7] a number of which have cone or rod-specific abnormalities. More optimal phenotypic characterization of clinical disorders in client owned animals may help to facilitate more accurate diagnosis, and may help to quickly identify disorders with relevance to human diseases. In this study, we sought to determine if ERG silent substitution methods could be adapted for use under sedation, to determine if this method could be utilized in client-owned dogs.

Methods

The study was performed with University Institutional Animal Care and Use Committee approval (protocol 18–069-B). Animals were not euthanized for the purpose of the study. Three normal adult purpose bred beagles (1 male aged 31.3 months, 2 females aged 29.6 months and 37.3 months) were studied: a subset of animals previously published using the more extensive ERG protocol performed under general anesthesia (propofol induction, isoflurane maintenance). [4] Animals were dark adapted for 1 hour. Topical mydriatics (0.5% tropicamide and 2.5% phenylephrine, Akorn Pharmaceuticals, Lake Forest, IL, USA) were applied prior to systemic drug administration. Anesthesia protocols have previously been published in detail. [4]Anesthesia protocols were performed prior to sedated protocols with a minimum washout period of 3 days between protocols. In the sedation protocol, dexmedetomidine (1.25 mcg/kg) and butorphanol (0.01 mg/kg) were injected intravenously and following the ERG protocols, reversed using intramuscular atipamezole (10.25mcg/kg). Topical analgesia was applied (0.5% proparacaine, Akorn Pharmaceuticals, Lake Forest, IL, USA). Protocols and analysis of silent substitution ERG were as previously published using color LED light stimuli administered in a Ganzfeld dome (Q450SC, Roland Consult, Brandenburg an der Havel, Germany) [4]with the following exceptions: stay sutures were not used to maintain primary gaze; the rod protocol was performed at 3.25 cd/m² and 4–20 Hz, the L/M-cone protocol at 32.5 cd/m² luminance and 24–50 Hz, and the S-cone protocol at 130 cd/m² and 4–20 Hz (performed in that order). These adjustments were made based on the findings from the published study, [4] reflecting the optimal isolation conditions for each photoreceptor class. The entire protocol lasted approximately 20 minutes. Further details on the ERG conditions are provided in Table 1.

Table 1:

Phase and contrast settings for the LEDs used in sine-wave examinations

	Light intensity	Red LED (predicted 625nm)		Green LED (predicted 525nm)		Blue LED (predicted 470nm)		Frequencies measured (Hz). Bolded frequencies performed in both anesthetized and sedated animals
		Contrast (%)	Phase	Contrast (%)	Phase	Contrast (%)	Phase
Rod isolating conditions (29% rod contrast)	3.25 cd/m2	100	180	58	0	52	0	4, 8, 12, 16, 20, 24, 28, 32
L/M-cone isolating conditions (29% cone contrast)	32.5 cd/m2	91	0	12	0	59	180	4, 8, 12, 16, 20, 24, 28, 32, 38, 44, 50, 56
S-cone isolating conditions (59% cone contrast)	130 cd/m2	76	0	83	180	72	0	4, 8, 12, 16, 20, 24, 28, 32, 38, 44, 50, 56

Statistical analysis was performed using commercial software (GraphPad Prism 5.0a for Mac, San Diego, CA) and included two-way repeated measures ANOVA with Bonferroni post-test comparing sedated and anesthetized amplitudes and phases within the same animal at all shared stimulus frequencies. Linear regression analysis compared the slope of the phase of the responses between anesthetized and sedated protocols. Data are presented as mean ± standard error of the mean.

Results and Discussion

Amplitude of the responses peaked in the same stimulus frequency ranges in both anesthetized and sedated animals (Figure 2 A–C). Namely, rod-derived responses peaked at low stimulus frequency (4–12 Hz), L/M-cone responses peaked at high stimulus frequency (32–38Hz) and S-cone responses peaked at low stimulus frequency (4–12Hz). Although both rods and S-cones were responsive to the same frequency of stimuli (4–12Hz), the previously published study [4] demonstrated differences in the optimal luminances for rod- and S-cone-driven responses. Rod-driven responses had peak amplitude at low luminance (3.25 cd/m²), whereas S-cone-driven responses had peak amplitude at high luminance (130 cd/m²), consistent with the different rod and cone luminance response characteristics shown in humans. [8] Although L/M-cone-driven responses also peaked at high luminance, L/M-cones had maximal amplitudes recorded in response to higher frequencies of stimuli compared with S-cones (compare Figure 2B with 2C). Hence we showed that using these conditions, all three photoreceptor subclasses could be effectively separated in dogs by their specific stimulus response characteristics. In general, silent substitution peak amplitudes are substantially lower than standard white flash b wave amplitudes. [4]

Figure 2.

Response amplitudes (A-C) and phase of the response (D-F) for rods, L/M-cones and S-cones. Rod amplitudes (A) were significantly higher under sedation than anesthesia. In contrast, L/M-cone amplitudes (B) and S-cone amplitudes (C) were not significantly different between the two drug protocols. The phase of the response in relation to the stimulus was similar between sedated and anesthetized animals for rods (D), L/M-cones (E) and S-cones (F) at the frequencies that elicited peak response amplitudes. * P < 0.05, ** P < 0.01, *** P < 0.001, two-way repeated measures ANOVA with Bonferroni post-test. Data are presented as mean _ standard error of the mean. Note the difference in scale of the Y axis in A compared with B and C.

The amplitude of responses in rod isolating conditions were significantly higher using sedation compared with general anesthesia for shared stimulus frequencies (two-way repeated measures ANOVA overall P value for anesthesia type < 0.001, Figure 2A); variability was higher in sedated responses. The mean difference in amplitude at each frequency and significance using Bonferroni post-test at each frequency was 9.1 ± 2.3 μV (4Hz; P < 0.01), 12.5 ± 2.0 μV (8Hz, P < 0.001), 7.7 ± 1.8 μV (12Hz, P < 0.01), 5.5 ± 1.5 μV (16Hz, P < 0.05) and 4.2 ± 0.1 μV (20 Hz, difference not significant). Overall, the amplitudes of responses in L/M-cone isolating conditions were also higher using sedation compared with general anesthesia (two-way repeated measures ANOVA overall P value for anesthesia type < 0.05, Figure 2B). However, individual frequencies were not significantly different between drug protocols using Bonferroni post-test (P > 0.05 for all shared frequencies), although variability was higher in sedated responses. Two-way ANOVA did not identify differences in amplitudes of S-cone responses between anesthesia types (Figure 2C). These data suggest that in dogs, silent substitution rod-derived responses are most sensitive (and S-cone responses least sensitive) to the effects of general anesthesia. Another study performed in dogs [9] supports our findings; the investigators found that that different sedative combinations affected rod photoreceptor amplitudes more than cone responses, and that sedative combinations containing alpha-2 agonists (medetomidine or xylazine) had a lesser effect on scotopic rod amplitudes than isoflurane anesthesia. Although body temperature was maintained during each protocol using a heating pad, the more prolonged anesthesia time may also have impacted core body temperature to a greater extent than sedation with an associated greater inhibitory effect on rod ERG amplitude. [10]

The peak amplitude response characteristics were also supported by the findings of the phase of the response. The phase of the response (Figure 2 D–F) is an indicator of response latency (phase is described in relation to the stimulus, similar to implicit or peak time). Aspects of response phase were discussed in detail in the previously published article. [4] The phase value was significantly different by two-way ANOVA analysis (rod P < 0.05, L/M-cone P < 0.001, S-cone P < 0.001 for anesthesia type, see Figure 1 for Bonferroni post-test results); this was not unexpected as different sedative or anesthesia drugs affect peak times of ERG in dogs to different extents. [9, 11] The phase is expected to be the most linear in the range of frequencies where a photoreceptor response is optimally isolated. Although the extensive range of frequencies performed under anesthesia could not be tested under sedation, linearity was evident in rod responses from 4–16 Hz (sedated r² 0.97, P < 0.0001, anesthetized r² 0.97, P < 0.0001, Figure 2D), in L/M-cone responses from 28–44Hz (sedated r² 0.95, P < 0.0001, anesthetized r² 0.94, P < 0.0001, Figure 2E) and in S-cone responses from 4–16 Hz (sedated r² 0.95, P < 0.0001, anesthetized r² 0.96, P < 0.0001, Figure 2F). These frequency ranges corresponded with the peak amplitude of responses in both sedated and anesthetized animals.

In conclusion, we present a clinically applicable method to consistently differentiate photoreceptor subclass function in dogs that can be used to identify dysfunction without the need for general anesthesia or prolonged adaptation times. Future research could consider developing similar protocols to study blinding inherited retinal dystrophies in dogs, particularly those that might differentially affect one photoreceptor class.