In vivo optoretinography of phototransduction activation and energy metabolism in retinal photoreceptors

Abstract

The objective of this study is to verify the anatomic correlate of the second (2nd) outer retina band in optical coherence tomography (OCT), and to demonstrate the potential of using intrinsic optical signal (IOS) imaging for concurrent optoretinography (ORG) of phototransduction activation and energy metabolism in stimulus activated retinal photoreceptors. A custom-designed OCT was employed for depth-resolved IOS imaging in mouse retina activated by a visible light flicker stimulation. The spatiotemporal properties of the IOS changes at the photoreceptor outer segment (OS) and inner segment (IS) were quantitatively evaluated. Rapid IOS change was observed at the OS almost right away, and the IOS at the IS was relatively slow. Comparative analysis indicates that the OS-IOS reflects transient OS deformation caused by the phototransduction activation, and IS-IOS might reflect the energy metabolism caused by mitochondria activation in retinal photoreceptors. The consistency of the distribution of the IS-IOS and the 2nd OCT band supports the IS ellipsoid (ISe), which has abundant mitochondria, as the signal source of the 2nd OCT band of the outer retina.

Graphical Abstract

1 |. INTRODUCTION

By providing excellent depth-resolved imaging capability, optical coherence tomography (OCT) is revolutionizing retinal study and disease diagnosis. Previous studies have disclosed four hyper-reflective OCT bands at the outer retina, that is, photoreceptor side. Anatomic sources of these four OCT bands have been typically attributed as follows: 1st band at the external limiting membrane (ELM); 2nd band at the photoreceptor inner segment (IS) and outer segment (OS) junction; 3rd band at the posterior tip of the OS; and 4th band at the retinal pigment epithelium (RPE).¹ Anatomic correlates of the 2nd band remain controversial. Although cell biologists consider the 2nd outer retinal band as the connecting cilium (CC) between IS and OS, comparative alignment of OCT bands with an anatomically correct model of the outer retina suggested an alternative correlate to the 2nd outer retinal band, that is, the IS ellipsoid (ISe).² Experimental OCT study also provided the evidence to support the ISe as an anatomic correlate of the 2nd outer retinal band.³ Although the 2014 international OCT nomenclature meeting has officially affirmed the ISe as an anatomic correlate of the 2nd outer retinal band,⁴ a following adaptive optics (AO) OCT study suggested that ascription of the 2nd outer retinal band to the ISe is unjustified.⁵ Therefore, the actual source of the 2nd outer retinal band is still controversial.^{2, 6, 7} In principle, it is possible for both the IS/OS junction and ISe to affect the OCT signal. Further investigation is required to verify the effect of the ISe on the 2nd outer retinal band.

As the center of photoreceptor metabolism, the ISe consists of abundant mitochondria.⁸ Previous studies have provided evidence to support mitochondria as one signal source of the 2nd outer retinal band.⁷ However, direct in vivo verification of the location of the mitochondria is challenging. In principle, stimulus-evoked photoreceptor activation can cause mitochondria to be activated. It is known that the energy consumed by photoreceptors is predominantly provided by the mitochondria, 55–65% of which are located in ISe.^8–10 When the photoreceptors are stimulated by visual light, energy demand changes significantly, and mitochondria activities are accordingly modulated.^{11, 12} In different metabolic states, the morphological structure, motion dynamics, and fission or fusion configuration of mitochondria are different.^{13, 14} which might affect the optical property of the mitochondria. It has been observed that the refractive index of mitochondria increases by stimulation.^15–17 Therefore, we speculate that, if the 2nd outer retinal band attributes to the ISe, transient IOS change should be detectable in the ISe of stimulus activated photoreceptors.

Functional IOS imaging,^18–20 also termed as optophysiology²¹ or optoretinography (ORG),^22–24 has revealed a fast IOS response that occurs almost immediately after the stimulus onset in the outer retina. Relatively slow IOS changes were also observed in the inner retina with elongated stimulus period^{23, 25} or repeated flicker stimuli.¹⁸ Depth-resolved OCT has revealed the OS, which is the center of phototransduction, as the anatomic origin of fast OS-IOS.^25–27 The following studies support that fast OS-IOS is tightly correlated with the activation phase of phototransduction,^28–30 promising a unique biomarker for objective ORG of photoreceptor function.^{22, 29} Although the rapid OS-IOS has been actively investigated, transient IS-IOS is not well explored. In this study, we conducted functional OCT-IOS imaging to quantify the relationship between the OS-IOS and IS-IOS. Experimental data and comparative analysis suggest that the OS-IOS reflects transient OS deformation caused by the phototransduction activation, and the IS-IOS reflects the energy metabolism caused by mitochondria activation in retinal photoreceptors.

2 |. METHODS

2.1 |. Experimental setup

Figure 1 shows the schematic diagram of the custom-designed spectral-domain OCT (SD-OCT) system, which has been used for function IOS imaging of mouse retinas.²⁷ Briefly, a broadband near-infrared (NIR; λ_center = 840 nm, Δλ = 100 nm)superluminescent diode (SLD) was used as the light source. A fiber coupler with a splitting ratio of 75:25 divided the OCT light to the sample and reference arms. A 1200 line mm⁻¹ transmission grating was used to separate the beam into different wavelengths. A 70,000 Hz line CCD camera with 2048 pixels was used for recording OCT spectra. A light-emitting diode (λ_center = 505 nm) was used for 10 Hz retinal flicker stimulation with a 50% duty ratio and 5-second duration. The measured optical power of the incident stimulation beam on the cornea was 0.32 mW/mm². The ANSI maximum permissible exposure (MPE) limit of 505 nm ocular illumination light is 2.9 mW/mm².³¹ The axial and lateral resolutions were estimated as 3 and 12 μm, respectively.

FIGURE 1.

Schematic diagram of the functional OCT system. LED, light-emitting diode (λ_center = 505 nm); CL, collimation lens; PC, polarization controller; SLD, superluminescent diode (λ_center = 840 nm, Δλ = 100 nm)

2.2 |. Animal preparation

All animal care and experiments were performed under the Association for Research in Vision and Ophthalmology statement for the use of animals in ophthalmic and vision research. All experiments were conducted following the protocols approved by the Animal Care Committee at the University of Illinois at Chicago. Six adult C57BL/6J mice (2 months old, Jackson Laboratory, Bar Harbor, ME) were used in this study. Before OCT recording, the mice were anesthetized with a mixture of 100 mg kg⁻¹ ketamine and 5 mg kg⁻¹ xylazine injected intraperitoneally. A heating pad was used to maintain body temperature during the experiment. The pupil was fully dilated with 2.5% phenylephrine hydrochloride and 1% tropicamide. A cover glass, along with GenTeal eye gel (Alcon Laboratories, Fort Worth, Texas), was placed on the cornea to prevent drying as well as to serve as a contact concave lens to improve the image quality. All mice were dark-adapted for 2 hours before OCT imaging. It has been reported that healthy people would complete the adaptation within 30 min in a fully dark environment.³² For experiments with rod-dominant mouse retinas, a longer period of dark adaptation is generally required to make sure complete recovery of rhodopsin.^33–35 To keep it consistent with previous studies, we set the dark-adaptation time for 2 hours.^{27, 30}

2.3 |. Data acquisition

Experiments were performed in a dark room. The OCT B-scans were acquired after the mice were anesthetized. The recording retina areas are 1 mm away from the optic nerve head (ONH). For the scanning protocol used in Figure 2 and Figure 3, each OCT B-scan consists of 500 A-lines, corresponding to 2 mm at the retina, and the recording speed was 35 frames per second. For the high-speed scanning protocol used in Figure 4, each OCT B-scan consists of 100 A-lines, and the recording speed was 500 frames per second. The total recording time was 30 seconds; a 3-second prestimulation phase, a 5-second flicker stimulation phase, and a 22-second post-stimulation phase.

FIGURE 2.

(A) The original (A1), flattened (A2), and vessel-free (A3) OCT B-scans. (B) Representative prestimulus (B1) and poststimulus (B2, B3) IOS maps. Each IOS map is averaged over a 1 s interval. Corresponding IOS video is provided as supplemental material. (C) A diagram of rod photoreceptor. (D) Temporal courses of IOS changes at the outer (between ELM and RPE) and inner (between GCL and ELM) retinal regions. D1 corresponds to the experiment shown in B. D2 shows the early phase of IOS in D1. (E) Average IOS recording of 6 retinas. The average IOS in the outer retina shows a biphasic curve. The dark and gray arrowheads in D1 and E show the first rapid increasing phase and the second gradually increasing phase, respectively. GCL, ganglion cell layer; IPL, inner plexiform layer; INL, inner nuclear layer; OPL, outer plexiform layer; ONL, outer nuclear layer; ELM, external limiting membrane; ISe (2nd), inner segment ellipsoid (the 2nd hyper-reflective band); IZ, interdigitation zone; RPE, retinal pigment epithelium; Ch, Choroid; IS, photoreceptor inner segment; OS, photoreceptor outer segment

FIGURE 3.

(A) OCT intensity (A1) and IOS activity (A2) M-scan maps. Three white arrows, from top to bottom along the direction of the inner to outer retina, show the top boundary of ISe (2nd), the connection between the ISe (2nd) and OS, and the connection between OS and RPE, respectively. (B) OCT intensity (B1) and IOS activity (B2) waveforms of the M-scan at different time points. (B3) The early phase of B2. Green arrows show that the IOS change at the boundaries of the OS was stronger than the IOS change in the middle of the OS. The red arrow shows a peak response seen at ~3.5 seconds at the ISe (2nd) band

FIGURE 4.

(A) Temporal courses of average IS-IOS and OS-IOS correspond to Figure 3A2. (B) Average IS-IOS and OS-IOS from six mice. The average OS-IOS showed a biphasic curve. The dark and gray arrowheads show the first rapidly increasing phase and the second gradually increasing phase, respectively. (C) IOS M-scan within 100 ms, the white and red arrowheads show the onsets of IS-IOS and OS-IOS, respectively. (D) Temporal courses of average IS-IOS and OS-IOS correspond to C. (E) Average OCT intensity of IS and OS from 6 mice. (F) The comparison of the average OCT intensity of IS and OS at different time points before and after the stimulation

2.4 |. Data processing and statistical analysis

OCT B-scans were registered using a cross-correlation based subpixel registration algorithm.³⁶ The bulk motion between the sequential images is compensated by applying a rigid body transformation, which is commonly used in OCT image registration for IOS study.^{27, 37, 38} To investigate the IOS signal in different retinal layers, the OCT B-scan image was flattened by normalizing the thickness of each segmentation. The IOS value was defined as the pixel intensity change divided by the average pixel intensity of the prestimulus period (ΔI/I). To minimize the random noise, the following steps were applied to the IOS calculation.^{19, 22, 25, 27} (a) Prestimulus mean value and standard deviation calculation (background calculation). The mean value and standard deviation of each pixel were calculated from the prestimulation images, defined as Ī_pre(x,y) and σ(x, y). (b) Comparison of the intensity of each pixel with the background. If the intensity was larger than Ī_pre(x,y) + 3×σ(x,y) or smaller than Ī_pre(x,y) − 3×σ(x,y), the value was reserved. (c) Consecutive rule to minimize random noise. If the IOS response at a specific pixel did not recur at least three times continuously, the response was ignored. In this study, all the intensity changes, larger or smaller than the background, are considered as absolute values. The intensity Motion scan (M-scan) was calculated by averaging each OCT B-scan image toward the column direction and combined them. Similarly, the IOS differential M-scan was calculated by averaging each IOS B-scan image toward the column direction and combined them.

All the statistical analyses were performed using MATLAB (Mathworks, Natick, Massachusetts). The averaged IOS and image intensity at different layers are expressed as the mean ± standard deviation. One-way ANOVA was applied to compare the OCT intensity response at individual retinal layers before and after the stimulation.

3 |. RESULTS

Figure 2 shows the representative OCT images and IOS of the mouse retina. Figure 2(A) shows the B-scan image. The individual retinal layers including the ganglion cell layer (GCL), inner plexiform layer (IPL), inner nuclear layer (INL), outer plexiform layer (OPL), outer nuclear layer (ONL), external limiting membrane (ELM), ISe (2nd hyper-reflective band), interdigitation zone (IZ), retinal pigment epithelium (RPE), choroid (Ch) were clearly observed. To investigate the IOS in different retinal layers, the B-scan was flattened (Figure 2A2), and blood vessel areas were removed to avoid hemodynamic contamination (Figure 2A3). Figure 2(B) shows representative IOS maps before and after the stimulus. After the stimulation, rapid IOS change was observed in the outer retina (between ELM and RPE), and slow IOS changes were observed in the inner retina (between GCL and ELM). Figure 2(D) shows the temporal courses of IOS changes at the outer and inner retina regions, and Figure 2(E) shows the average IOS response from 6 mice. The IOS in the outer retina shows two phases: the first rapid increasing phase (dark arrowheads, Figure 2D1 and Figure 2(E)) right after the stimulus and the second gradually increasing phase (gray arrowheads, Figure 2D1 and Figure 2(E)).

To investigate the IOS change in different layers of the outer retina, M-scans of OCT intensity and IOS activity were generated, and average intensity and IOS changes of different layers were calculated (Figure 3). IOS activity in different layers showed clear differences, especially at the outer retina (Figure 3A2). Robust IOS was observed at the OS immediately after the stimulus, and a relatively slow and gradually increasing IOS was observed at the IS layer (Figure 3A2 and Figure 3B2). In the early phase, the rapid OS-IOS was not evenly distributed (green arrowheads, Figure 3B3). The IOS changes at the boundaries were stronger than those in the center area of the OS. On the other hand, the IS-IOS gradually developed a robust peak with spatial location coinciding with the ISe (2nd) band (red arrowhead, Figure 3B3).

The average IOS curves confirmed the differences between the OS-IOS and IS-IOS (Figure 4(A) and Figure 4 (B)). The OS-IOS showed a biphasic curve, a rapidly increasing phase (dark arrowheads, Figure 4(A) and Figure 4(B)), and a gradually increasing phase (gray arrowheads, Figure 4(A) and Figure 4(B)). The IS-IOS showed a gradually increasing curve. In order to verify the onset time of the OS-IOS and IS-IOS, the imaging speed was increased to 500 frames s⁻¹. Figure 4(C) and (D) show the IOS M-scan and the corresponding IOS waveforms. The OS-IOS first appeared at the OS boundaries within 2 ms after the stimulus. The IS-IOS appeared at the 2nd outer retinal band with a time delay of ~12 ms after the stimulation. Figure 4(E) shows the average OCT intensity of IS and OS layers from six mice. The average OCT intensity of the OS did not show detectable change by the stimulation, while the average OCT intensity of the IS showed gradually increased magnitude. The OCT intensity change differences between IS and OS are further confirmed by statistical analysis (Figure 4(F)).

4 |. DISCUSSION

In this study, depth-resolved OCT was used to characterize the light-evoked IOS changes in individual layers of the mouse retina. A rapid OS-IOS, which occurred within 2 ms, reflected the activation phase of the phototransduction,²⁹ and a relatively slow and gradually increasing IS-IOS might reflect the metabolic activity. Moreover, the spatial distribution of IS-IOS well matched the 2nd outer retinal band, providing evidence to support ISe as the anatomic source of the 2nd outer retinal band.

The rapid OS-IOS, which was observed almost immediately after the stimulus onset, is consistent with our previous study.^{22, 28} The time course of the rapid OS-IOS was highly correlated with transient OS deformation, caused by phototransduction activation.^{23, 30, 39} It has been well investigated that different types of cone photoreceptors show different responses after light stimulation.^{23, 24} ~98% of photoreceptors in the retinas of C57BL/6J mice are rods..^40–42 Thus, the IOS change in this study is predominantly caused by rods' deformation. It was demonstrated that the photoreceptor OS deformation can change the local intensity but did not affect the overall intensity of the whole OS region.²² Also, it has been reported that the IOS of mice with retinal degeneration disease showed significant signal attenuations compared to healthy mice.⁴³ Therefore, the rapid OS-IOS provides a unique biomarker for photoreceptor dysfunction due to retinal disease.²²

The IS-IOS revealed a delayed onset time, which might reflect the metabolic reaction of mitochondria, following the phototransduction in the OS. Mitochondria are the prominent light scatterers inside the photoreceptors,⁴⁴ and densely packed mitochondria presenting at the tip of the IS provide energy for the phototransduction processes through oxidative phosphorylation (Figure 2(C)).⁴⁵ Under metabolic stresses, mitochondria also undergo morphological changes, that is, fission and fusion, which could alter the refractive index of the mitochondria, resulting in IS-IOS changes as well as OCT intensity changes.⁴⁶ Tychinsky et al. demonstrated that the refractive index of mitochondria depends on their energy state.^{16, 17} Haseda et al. further demonstrated that the refractive index of a single mitochondrion was significantly changed regarding mitochondrial activity.¹⁵ Taken together, our data supports that the ISe, that is, highly packed mitochondria, should be attributed to the 2nd outer retinal band, and their metabolic activity might be measured by ORG.

There are some limitations in this study to be addressed. First, standardized ex vivo examinations, such as immunohistochemical study and electron microscopic (EM) observation, should be accompanied by ORG measurement to cross-validate the IS and OS changes in the future study. A previous study demonstrated the feasibility of EM observation of light-driven OS length change at the disc level.³⁹ Second, a relatively small sample size (6 mice) was used in this study as a proof-of-concept demonstration of functional OCT imaging of OS and IS changes. Further investigation, with increased sample size and variable stimulus conditions, will be essential to establish the relationship between the IOS changes and photoreceptor metabolic changes. Third, the mouse retina is known to be rod predominant. Following studies with human subjects under different light conditions will be required to evaluate the effect of rods and cones on the IOS imaging of the photoreceptor OS and IS changes.

5 |. CONCLUSION

In conclusion, we demonstrated the potential of functional IOS imaging of phototransduction activation and energy metabolism in retinal photoreceptors. Experimental results and comparative analysis indicate that the OS-IOS reflects transient OS deformation caused by the phototransduction activation, and the IS-IOS may reflect the energy metabolism caused by mitochondria activation in retinal photoreceptors. The consistency of the distribution of the IS-IOS and the 2nd OCT band supports the ISe, as the signal source of the 2nd OCT band of the outer retina. Functional OCT enables depth-resolved IOS imaging for concurrent ORG of phototransduction activation and energy metabolism of retinal photoreceptors, which can be valuable for early detection of photoreceptor dysfunction due to eye diseases.