Fast and slow light-induced changes in murine outer retina optical coherence tomography: complementary high spatial resolution functional biomarkers

Abstract

Introduction

Light triggers a photochemical transduction cascade in photoreceptors that generates an electrical signal that is eventually transmitted to brain to elicit a visual percept. These electrical signals can be noninvasively recorded from the eye as an electroretinogram (ERG), which has long served as the sole means for noninvasive probing of retinal function. For the ERG responses recorded from the dark-adapted eye, the ERG a-wave is mainly linked to photocurrent changes from rod photoreceptors, and the ERG b-wave from activity of the second order neurons (1). The ERG c-wave, on the other hand, is produced by (retinal pigment epithelium) RPE and Muller cells as an indirect response to photoreceptor activity (2, 3). However, the full-field ERG reports only on the whole retinal response, so that local variations in dysfunction and focal treatment responses are often not detectable. In addition, ERG is known to only incompletely interrogate the retinal physiology. For example, it is hard to extract indices of mitochondria activity, a central aspect of retinal function, from the monovalent ion changes on a subsecond time scale measured by the ERG.

To address this problem, high spatial resolution measurement of optical changes following light stimulation using optical coherence tomography (OCT) in retinal tissue has been suggested (4–9). So far, both fast (seconds) and slow (minutes to hours) OCT changes show promise as retinal function biomarkers. Fast OCT responses have been designated as the optoretinogram (ORG), which can be revealed both as reflectance changes and as subtle (submicron) thickness alterations (5, 6, 10–13). In addition, a faster response peaked within 10 milliseconds has also been reported (6, 14). A prominent ORG response is the increase of outer retinal reflectivity of the inner segment ellipsoid zone (ISez) and the outer segment (OS) bands upon light stimulation (5, 7, 15). These fast responses are linked to light-induced rhodopsin isomerization and passive, osmotic alteration in water content as a consequence of phototransduction steps (5). However, exactly how these osmotic changes are produced and the direction of water movement after light stimulation are still under debate (5, 16). In addition, potential contributions from mitochondria function have been proposed to underlie ORG responses, but this hypothesis has not been fully tested (17, 18). Furthermore, although the fast OCT responses have been reported in human patients (19), it requires specialized custom-built OCT systems that are often not available in most eye clinics.

Slow responses have been measured by diffusion MRI and OCT in rat and mouse outer retina after long room light or dark adaptation (4, 20–22). For example, robust light–dark changes in mice retina include larger (4 to 6 μm) changes in outer retina thickness (from ELM to RPE), and large changes in hyporeflective band (HB) intensity between the photoreceptor tip and RPE layers (4, 22–24). The physiology of the ELM-RPE thickness change is well understood (for more details see the “Discussion” section) and closely linked to mitochondria-induced pH changes in the subretinal space (25, 26). A recent study found that the light–dark HB signal intensity changes to be correlated with those of the ELM-RPE thickness but in a pH-insensitive manner, suggesting a link to mitochondria activity as well (24). Intriguingly, the light–dark ELM-RPE thickness change measured with OCT measured can be used to detect oxidative stress in the outer retina (23, 25, 26).

In this study, we measured the following OCT biomarkers in the outer retina captured with a commercial OCT instrument: (1) light-induced changes in the mean intensity of the ISez and OS bands (i.e. the ORG response); (2) ELM-RPE thickness; and (3) HB magnitude. We determined the kinetics and intensity-response relations of the light stimulation induced reflectance increases for both ISez and OS bands in the outer retina. The light sensitivity and kinetics of these mouse ORG signals were compared to the outer retina ERG responses. In these same datasets, the ORG responses are compared with slow OCT biomarkers: ELM-RPE thickness and HB magnitude changes. We also asked how ATP synthesis efficacy (25, 27) and phototransduction modulated the ORG and OCT biomarkers. Our results support the notion that the fast ORG response is mediated by upstream events (e.g. the initial rod photoreceptor shape changes) and the passive light-induced osmotic water build up in the subretinal space is linked to phototransduction (5). In other words, the ORG is a consequence of the ion charge movement-based transduction mechanism that is measured by the ERG (5, 7, 18). On the other hand, the slow changes in outer retina thickness and HB intensity are consistent with downstream consequences of mitochondria-evoked shifts in water content in the dark compared to that in the light. Together, these different OCT modalities and biomarkers provide a complementary set of tools to noninvasively probe retinal function in vivo with high spatial resolution and sensitivity.

Results

Light-induced fast OCT intensity increases in the outer retina bands

Most of the study was performed with C57BL/6 J (B6J) mice, a common mouse strain used to study the visual system. Fig. 1 shows an example of OCT images captured from a dark-adapted eye (Fig. 1A, left panel) and the same retina location 2 minutes after light exposure (Fig. 1A, middle panel) using a commercial instrument (Bioptigen Envisu UHR2200), which constructs images based on logarithmic OCT intensity. OCT images were exported in 8-bit (256 level) tiff format. The intensity of one grey level is referred to as one a.u. throughout this study. A 2-minute light exposure was chosen as the ORG response metrics reached saturation at this time (see next section). Averaged intensity profiles of these two OCT images along retinal depth are shown on the right panel of Fig. 1A. Fig. 1B shows the light-induced intensity difference of the two images in Fig. A with its intensity profile plotted in the right panel. It is clear that the largest differences in OCT intensity induced by light stimulation are in the outer retinal regions, i.e. ISez and OS (5, 15, 17, 18). In this study, we focused on the outer retina regions for light-induced changes in the OCT image, as illustrated with a magnified view in Fig. 1C for OCT images and their intensity profile. Specifically, we measured light-induced OCT intensity changes for ISez and OS bands as indices ORG (5, 7), and ELM-RPE thickness and magnitude of HB as indices of slow OCT responses (4, 24).

Fig. 1.

Examples of the light-induced changes observed in the mouse retinal OCT image. (A) Averaged (20x) B-scan OCT images captured at baseline in the dark (left panel) and 2 minutes after light stimulation (middle panel). Mean depth intensity profiles across the retina are shown on the right panel. NFL: nerve fiber layer; IPL: inner plexiform layer; INL: inner nuclear layer; OPL: outer plexiform layer; ONL: outer nuclear layer; ELM: external limiting membrane. (B) Pseudocolor difference image of the images shown in panel A. Light-induced increases in the intensity of the two outer retinal layers are evident, whereas the other parts of retina showed relatively little change to light stimulation. Mean depth intensity profiles across the retina are shown on the right panel as black trace, with peaks for ISez and OS retinal layers. (C) Magnified view of the outer retina region of the OCT image, from the area marked by yellow square in panel A, with mean depth intensity profile differences shown on the right side. ELM-RPE thickness is measured as distance between ELM peak and basal side of RPE layer. HB is defined as intensity dip in the region between OS and RPE.

Open in new tabDownload slide

Kinetics of mouse ORG responses

To further characterize the ORG response, we then measured kinetics of the light-evoked signal intensity changes in the ORG for the ISez and OS bands to an 80-s light stimulation step (Fig. 2A). The ISez responses appeared to show slightly slower kinetics than the OS response. To quantitate the kinetics, the responses were fit with an exponential function (solid lines in Fig. 2A). The time constants for the average ON response of ISez and OS band reflectance increases are shown in Fig. 2B for both bright (4.0 log cd/m²) and dim (2.34 log cd/m²) light stimulations. There were significant differences observed (P < 0.01) for the time constants of ISez and OS responses elicited by bright light stimulation. The time constants for ISez response elicited by dim light stimulation may have been slightly larger than those for OS responses, but this could not be confirmed statistically. The amplitudes of ISez and OS responses for bright stimulation are shown in Fig. 2C. Although OS showed a larger response than ISez, the difference was not statistically significant (P = 0.08). Next, we asked if the signal intensity temporal trajectory changed in mice with higher (129/S6) or lower (B6J) ATP synthesis efficiency. Notably, light-induced reflectance changes in ISez and OS bands in S6 and B6J mice did not show significant differences in response amplitudes (P = 0.78 for OS and P = 0.80 for ISez, Fig. 2C). In contrast, S6 mice showed much smaller changes than B6J mice for slow OCT responses, i.e. long-duration light adaptation induced reductions in ELM-RPE thickness and HB magnitude (24,25).

Fig. 2.

Kinetics of the light-induced OCT responses. (A) Onset kinetics of ISez and OS band. Each data time point was measured from an averaged B-scan images after light exposure, and continuous curves are fit to an exponential equation. The ISez signal exhibited slightly slower development compared with OS signal. The light stimulus time course of is shown as the bottom trace. (B) The fitted time constants for ISez and OS responses show significant kinetic differences (p < 0.05) for bright light stimulation [bright: 1.0 × 10⁴ cd/m² (n = 6); dim: 2.2 × 10² cd/m² (n = 6)]. (C) Peak amplitudes for light-induced intensity increase of the ISez and OS bands elicited from C57B6J (n = 6) and 129S6 (n = 5) mice. (D) Time course of the Off kinetics of ISez and OS band to an 80-s pulse of light at two intensities [bright (n = 4): 1 × 10⁴ cd/m²; dim (n = 5): 219 cd/m²]. The duration of light stimulation is shown as the bottom trace.

Open in new tabDownload slide

The effects of light intensity on recovery kinetics are shown in Fig. 2D. After 80 s of bright light stimulation, both ISez and OS ORG signals persisted for several minutes. Ten minutes after bright light stimulation, their intensity had decayed ∼30% on average. On the other hand, after a dim light stimulation, ORG signals for both ISez and OS returned to baseline within 2 minutes.

Fast and slow light-induced changes of ELM-RPE thickness

The 80 s of light stimulation also induced a significant small extension of the outer retina (measured from ELM to RPE layer) thickness, as shown in Fig. 3A for C57BL/6 J mice. The ELM-RPE region expanded by about 0.7 µm (Fig.   3B), and similar magnitude of the response was also observed on S6 mice. The recovery time course of ELM-RPE thickness after an 80-s light stimulation also depended on the stimulus light intensity. ELM-RPE thicknesses were persistently extended over 10 minutes after termination of a bright (4.0 log cd/m²) light stimulation. On the other hand, for dim (2.34 log cd/m²) light stimulation, ELM-RPE thickness returned to baseline level within 5 minutes after the termination of the light stimulation (Fig. 3C).

Fig. 3.

Light-induced changes in ELM-RPE thickness. (A) ELM-RPE thickness measured for each mouse eye before (baseline in dark) and after 80-s bright light stimulation. There is a significant increase (P < 0.01, n = 7) of ELM-RPE thickness after light stimulation on C57Bl6J mice. (B) Comparison of mean amplitudes of light-induced ELM-RPE thickness changes elicited from C57BL/6 J and 129S6 mice. (C) Time course of off kinetics for ELM-RPE thickness changes from the baseline (values measured before onset of light stimulation). Duration of light stimulation [bright (n = 4): 1 × 10⁴ cd/m²; dim (n = 5): 219 cd/m²] is shown as the bottom trace. * significant (P < 0.05) difference for values measured after bright and dim light stimulation.

Open in new tabDownload slide

Using phase-sensitive measurement, the light-induced extension of photoreceptor OS length has been reliably detected (6, 14, 19). This change in OS length can be revealed on conventional OCT images (5). We also measured OS length changes in response to 80-s light stimulation, and the results are shown in supplementaary Material Fig. S1. However, the magnitude of OS length change is small, and none of light-induced OS length changes reached statistical significance. Consequently, we only focused on ELM-RPE thickness changes in this study.

Light-induced changes on the photoreceptor tip-HB magnitude

The HB magnitude, as illustrated in Fig. 1C, was increased from baseline after light stimulation. Fig. 4A showed HB magnitudes for each mouse eye before (baseline) and after an 80 s stimulation with bright light (4.0 log cd/m²). This increase of the HB magnitude was statistically significant. As S6 mice showed no light-induced HB responses (24), no comparison with B6J mice was made. Recovery time course of the HB also depends on the stimulation light intensity. A fast recovery was observed for dim light stimulation, whereas the HB magnitude persisted for several minutes after the termination of bright light stimulation (Fig. 4B).

Fig. 4.

Light-induced changes in the photoreceptor tip-HB magnitude of C57/Bl6J mice. (A) HB magnitudes measured for each mouse eye before (baseline in dark) and after 80-s bright light stimulation. There is a significant increase (P < 0.01, n = 7) of HB magnitudes after light stimulation. (B) Time course of off kinetics for HB magnitudes. Duration of light stimulation [bright (n = 4): 1 × 10⁴ cd/m²; dim (n = 5): 219 cd/m²] is shown as the bottom trace. *, significant (P < 0.05) difference for values measured after bright and dim light stimulation.

Open in new tabDownload slide

Effect of light-adaptation on OCT responses

Fig. 5 shows the effect of prolonged light adaptation (>5 hr to room light) on the mouse OCT responses. Compared with a fully dark-adapted (over-night) mice, ORG responses were significantly diminished from mice adapted to room light (Fig. 5A). With an 80-s bright light stimulation, the ISez response for dark-adapted mice averaged 10.13 ± 1.35 a.u. (mean ± SEM) but only 3.67 ± 1.35 a.u. for light-adapted animals. OS responses showed similar differences averaging 10.71 ± 1.12 a.u. for dark-adapted mice and 1.35 ± 0.88 a.u. of light-adapted mice. Similarly, light-adaptation also significantly diminished changes induced by 80-s bright light stimulation on ELM-RPE thickness (Fig. 5B, left panel) and HB magnitude (Fig. 5B, right panel).

Fig. 5.

Effect of light- and dark-adaptation on the light-induced OCT responses on C57BL/6 J mice. (A) Responses of ISez and OS bands to 80-s bright light stimulation for dark- and light-adapted mice. (B) Changes in ELM-RPE thickness and HB magnitude to 80-s bright light stimulation. (C) ELM-RPE thickness of dark- and light-adapted mice before (baseline) and after 80-s bright light stimulation. (D) HB magnitudes of dark- and light-adapted mice before (baseline) and after 80-s bright light stimulation. Dark-adapted: n = 5; light-adapted: n = 5. **, P < 0.01; ***, P < 0.001; and **** P < 0.0001.

Open in new tabDownload slide

There is a significant difference (P < 001) of ELM-RPE thickness at baseline for dark-adapted (51.86 ± 0.44 µm) and light-adapted (56.85 ± 0.67 µm) mice (Fig. 5C). The 80 s of light stimulation only slightly increased ELM-RPE thickness for dark-adapted mice, whereas no changes were observed for light-adapted mice. Consequently, there is still large significant difference of ELM-RPE thickness after light stimulation for these two groups of mice (P < 0.01). The difference in ELM-RPE thickness for these two groups of mice is still significant even after 15 minutes of continued light stimulation (data not shown).

Mice light-adapted to room light had a larger HB magnitude (15.92 ± 1.41 a.u.) than those dark-adapted overnight (2.74 ± 0.57 a.u., P < 0.001; Fig. 5D baseline). A short pulse of light stimulation elicited significant increases in the HB magnitude for dark-adapted mice (4.89 ± 0.99 a.u., P < 0.05), but had no effect on light-adapted mice (15.04 ± 1.14 a.u., P = 0.30).

ORG vs. ERG

The dark-adapted ERG a-wave is mainly mediated by the electrical responses of rod photoreceptors, whereas the b-wave has a large contribution from bipolar cells in the inner retina, and the c-wave reflects voltage changes in both RPE and Muller cells in response to ionic changes in the outer retina (1, 2). To compare the ORG and ERG light sensitivity curves, we determined the intensity-response relation of the OCT ISez and OS bands and compared the ERG a- and c-wave responses. Fig. 6A shows examples of ERG trances to a step of light stimulation. The ERG a-wave is the first negative peak followed onset of light stimulation, and the c-wave is a slow positive response usually peak at 3 to 5 sec after light onset. The time course of ERG responses is considerably faster than those observed for ORG light-on and -off kinetics (Fig. 2), indicating these two responses interrogate distinctly different functional pathways. Fig. 6B shows data normalized to the maximum values, with continuous curves as fit to a Naka–Rushton equation [Eq. (1) in "Materials and methods" section]. Measured by half-saturation constant K, the response of OS band also showed significant higher sensitivity (K = 1.48 ± 0.40 cd.s/m², n = 6) than the response for ISez band (K = 7.81 ± 1.65 cd.s/m², P < 0.01). Interestingly, the ORG showed intermediate sensitivity between ERG a-wave and c-wave.

Fig. 6.

Comparison of the intensity-response the OCT ISez and OS band, with the a- and c-wave of the ERG responses. (A) Example of ERG traces elicited by a light step stimulation. Intensities of light stimulation are shown on the left side panel. (B) Intensity-response relations for ERG a- and c-waves and OCT ISez and OS bands. The data were normalized to their respective peak amplitudes. Continuous cures are fitting to a Naka–Ruston equation (OCT responses: n = 6; ERG responses, n = 5).

Open in new tabDownload slide

Impact of phototransduction

To investigate the effects of phototransduction on the mouse ORG responses, we examined the OCT intensity of the dark-adapted retina and recorded the light-induced responses after administering sildenafil, a phosphodiesterase (PDE5/6) inhibitor (28–30). PDE6 activity in photoreceptor is inhibited in the dark, and light stimulation removes the inhibition and increases the enzymatic activity. Dark-adapted mice treated with sildenafil exhibited reduced OCT intensity changes of both ISez and OS bands (Fig. 7A), whereas sildenafil had no significant effect on dark-adapted ELM-RPE thickness or HB magnitude. Since sildenafil acts on the ERG (28, 29), reducing PDE6 activity with sildenafil also decreased the sensitivity of the ORG response for both the OS (Fig. 7B, P < 0.01) and the ISez (Fig. 7C, P < 0.01) bands.

Fig. 7.

Effect of sildenafil on the light-induced outer retinal OCT responses. Effect of sildenafil on baseline band intensity of ISez and OS on the dark-adapted mouse retina (A) (control: n = 5; Sil: n = 9; saline: n = 9). Effect of sildenafil on intensity-response relation for ISez band (B) and OS band (C). Continuous curves are fitting to a Naka–Ruston equation. Controls are replotted from data shown in Fig. 6.

Open in new tabDownload slide

Discussion

In this study, we identified for the first time with a commercial OCT system, differences between fast and slow OCT responses and the ERG from the same mouse eye. Light-induced fast responses were largely located at OS and ISez retinal bands, similar to those reported with noncommercial systems (5, 7). It has been postulated that ORG responses are mediated by light-induced passive osmotic changes that alters cell shape (5, 6, 11). Here, we provide additional support to link ORG signals with the complete phototransduction pathway. Light adaptation, which saturates the rod photoreceptor, significantly reduced the ORG responses (Fig. 5). Sildenafil, a PDE inhibitor that alters sensitivity of phototransduction (28, 29), also reduced the ORG photosensitivity for both the OS and ISez responses (Fig. 7).

The kinetics of the responses are also consistent with the ORG signals originating from phototransduction mechanisms. The phototransduction machinery is located mainly in the photoreceptor OS, and the ORG response was observed first in the OS region with relatively slower onset kinetics for the ISez band (Fig. 2) in agreement with a previous report (7). The ORG recovery kinetics for both OS and ISez depended on stimulus intensity (Fig. 2D), similar to the rod photoreceptor dark-adaptation time course recovery after different levels of light stimulation (31). Also, the ISez response is about 0.7 log unit less sensitive than the OS responses (Fig. 6B). Such differences in the time course and light sensitivity for ISez and OS band could reflect different subcellular organelle and protein composition in these two distinct photoreceptor regions. Comparison of the ORG with the ERG is summarized in Fig. 6. The dark-adapted ERG a-wave mainly reflects rod photoreceptor electrical activity (i.e. light-induced reduction of the dark current). The higher light sensitivity of the ORG than the ERG a-wave indicated a physiological mechanism for the ORG response subsequent to the ionic changes mediated by the dark-current. The ERG c-wave is generated by RPE and Muller cells in response to potassium concentration changes in the subretinal space produced by the rod photoreceptor light response (32, 33). The lower sensitivity and slower time course of the ORG OS and Ise responses compared with the ERG c-wave (Fig. 6) suggested a more delayed mechanism for the ORG than those produced by light-induced potassium changes in the subretinal space. It should be noted, however, that OCT-based ORG often only measures from selected regions of the retina, whereas full-field ERG measures summed activity from the whole retina. In this study, we did not compare ORG with the ERG b-wave, which is mainly mediated by bipolar cells in the inner retina.

As demonstrated in previous studies, slow responses can be detected in the outer retina on both ELM-RPE thickness and the photoreceptor tip-HB magnitude (4, 22, 24–26). As might be expected, light-induced changes on both ELM-RPE thickness and HB magnitude on short time scale (within minutes) were only a small fraction of the response amplitudes exhibited for fully (over hours) dark-adapted and light-adapted mouse retina (Fig. 5). Due to the low signal-to-noise ratio for ELM-RPE thickness and HB biomarkers, it is not possible to accurately measure the kinetics of these OCT signals for each mouse eye examined. However, the averaged signal from all recorded mice showed a similar time course as those of the ORG responses (Fig. 2 and Supplementary Material S2 and S3), suggesting a common mechanism (i.e. water redistribution between photoreceptor and subretinal space) for ORG and the early changes in these slow OCT biomarkers. Also, the different mechanisms mediating initial and long-term responses of slow OCT biomarkers are supported by a comparison of responses from B6J and S6 mice. These two strains of mice exhibit large differences in long-term responses that attribute to their difference in mitochondria activity (23–25, 34, 35). On the other hand, the phototransduction machinery is likely very similar in these two strains of mice. Consequently, both the ORG and the initial responses of these slow OCT biomarkers for these two mice were also very similar (Figs. 2C and 3B).

The mechanisms involved in generating the ORG response are an active area of investigation (5, 7, 11, 18). It has been postulated that the ORG responses are mediated by an osmolality increase of the photoreceptor initiated from light trigged dissociation of transducin from a trimer into α and βγ subunits (5). This hypothesis predicts water flow into the photoreceptor under light adaptation. However, this interpretation appears to contradict a large body of work since the mid-1990s involving a variety of techniques and model systems showing that light induces accumulation of water in subretinal space (20, 21, 36–38). Based on the common observation that excitation induces neurons to swell and decrease light scattering (39–43), we propose following signal pathway:

• Light → photoreceptor hyperpolarization → osmolality decreases → outer flow of water to subretinal space → photoreceptor shrink → OCT intensity (scattering) increase

• Pfäffle et al. provided a mathematic model of how reduction of photoreceptor volume (shrinking) leads to elongation of cell length (16).

On the other hand, as reviewed in a recent publication (26), long-term responses of slow OCT biomarkers (ELM-RPE thickness and HB) are correlated with metabolic demands in the outer retina region. Specifically, the slow responses of the ELM-RPE thickness and HB bands are contributed by water production from mitochondria metabolism and water removal by the fluid transporter on the RPE. The ORG responses (OS and Isez) reflect passive changes in the outer retina triggered by phototransduction, whereas the slow OCT biomarkers require active contribution of mitochondria metabolism and have a slower time course.

In sum, commercial OCT imaging can reveal light-induced change in the outer retina with both fast and slow kinetics. Passive cellular water movement seems to dominate the ORG response, whereas active modulation of the subretinal space hydration in the dark contributes to the slower OCT response. In any event, our results raise the possibility that the ORG and long-term OCT biomarkers are useful complementary noninvasive tools to probe direct photoreceptor activity and metabolic regulation, respectively; and a combination of the measurements would provide a more comprehensive view of outer retina function than currently available from ERG measurements, which could have clinical utility in evaluating visually impaired patients.

Materials and methods

Animal

All procedures involving animals were conducted under an approved National Institutes of Health (NIH; Bethesda, MD, USA) animal care protocol by the National Eye Institute Animal Care and Use Committee and ARRIVE guidelines. All methods were performed in accordance with the relevant guidelines and regulations. Adult wild-type mice (C57Bl/6 J and 129/S6), 3 to 6 months old, of either sex were used in this study. The mice were kept in regular animal housing under a 50 lux 14:10 hour light/dark cycle.

OCT imaging and analysis

Light stimulation

Light stimuli were provided by a LED ring light (World Precision Instruments, Sarasota, FL, USA) mounted on the OCT bore. A white plastic half-ping-pong ball diffuser was attached onto the OCT bore to provide full-field uniform light stimulation at the mouse's eye. The timing and duration of the light stimulation were controlled by a Raspberry Pi using a custom program. Neutral density films were added between the LED light source and pin-pong ball to adjust for light intensity. Light intensities were calibrated with Flame-S Spectrometer (Ocean Optics, Inc., Largo FL, USA) with the probe mounted at the same position as the mouse's eye.

ERG recording

ERG responses were recorded following published protocols with an Espion E2 Visual Electrophysiology System (Diagnosys, Lowell, MA, USA) (44, 45). After an over-night dark adaptation, mice were anesthetized with ketamine (100 mg/kg) and xylazine (6 mg/kg), and their eyes were dilated with a drop of tropicamide (1%) and phenylephrine (0.5%). The body temperature of the mouse was maintained at 37°C by a heating pad. Responses from both eyes were recorded with gold wire loop electrodes with band pass filtering from DC to 300 Hz. Steps of light with intensity ranged from 0. 01 to 1000 cd/m² were delivered to mouse eye by the attached ColorDome. Inter-stimulus intervals ranged from 1 to 5 minutes depending on stimulus intensity. Amplitudes of A-wave were measured from baseline to negative trough, while slowly developing c-wave amplitudes were measured from the baseline to the wave peak, usually occurring around 3 to 5 sec after light onset.

Effect of sildenafil

Mice were treated with sildenafil according to a previously published protocol (30). Briefly, dark-adapted mice were given an intraperitoneal injection of 29 mg/kg sildenafil, a dose that produces a transient impairment in retinal electrophysiology (29). OCT images were acquired 1 hour after sildenafil injection. Saline-injected mice were served as controls.

Funding

This work was supported by the intramural program of the National Eye Institute. S.G. was supported by the Overseas Research Project of Science and Technology, Talent Henan Health Commission of China (Grant No. 2018033), and Key Scientific Research Project of Higher Education of Henan Province (Grant No. 19A320078). B.A.B. was supported by the National Institutes of Health [EY026584, AG058171].

Authors' Contributions

E.D.C., B.A.B. and H.Q. designed the study. S.G., Y. Z., and Y.L. performed experiments. S.G., Y. Z., Y.L., and H.Q. analyzed data. H.Q. and B.A.B. wrote the manuscript and all coauthors critically commented and/or edited the manuscript. H.Q. supervised the project.

Data Availability

All data are included in the manuscript and/or Supplementary Material.

Notes

Competing Interest: Authors declare no competing interest

Disclaimer: The mention of commercial products, their sources, or their use in connection with the material reported herein is not to be construed as either an actual or implied endorsement of such products by the Department of Health and Human Services.

References