Spectral Domain Optical Coherence Tomography in Mouse Models of Retinal Degeneration

Abstract

Purpose.: Spectral domain optical coherence tomography (SD-OCT) allows cross-sectional visualization of retinal structures in vivo. Here, the authors report the efficacy of a commercially available SD-OCT device to study mouse models of retinal degeneration.

Methods.: C57BL/6 and BALB/c wild-type mice and three different mouse models of hereditary retinal degeneration (Rho ^−/−, rd1, RPE65 ^−/−) were investigated using confocal scanning laser ophthalmoscopy (cSLO) for en face visualization and SD-OCT for cross-sectional imaging of retinal structures. Histology was performed to correlate structural findings in SD-OCT with light microscopic data.

Results.: In C57BL/6 and BALB/c mice, cSLO and SD-OCT imaging provided structural details of frequently used control animals (central retinal thickness, CRT_C57BL/6 = 237 ± 2 μm and CRT_BALB/c = 211 ± 10 μm). RPE65 ^−/− mice at 11 months of age showed a significant reduction of retinal thickness (CRT _RPE65 = 193 ± 2 μm) with thinning of the outer nuclear layer. Rho ^−/− mice at P28 demonstrated degenerative changes mainly in the outer retinal layers (CRT _Rho = 193 ± 2 μm). Examining rd1 animals before and after the onset of retinal degeneration allowed monitoring of disease progression (CRT _{rd1 P11} = 246 ± 4 μm, CRT _{rd1 P28} = 143 ± 4 μm). Correlation of CRT assessed by histology and SD-OCT was high (r ² = 0.897).

Conclusions.: The authors demonstrated cross-sectional visualization of retinal structures in wild-type mice and mouse models for retinal degeneration in vivo using a commercially available SD-OCT device. This method will help to reduce numbers of animals needed per study by allowing longitudinal study designs and will facilitate characterization of disease dynamics and evaluation of putative therapeutic effects after experimental interventions.

Optical coherence tomography (OCT) has evolved over the past two decades to become an important diagnostic tool in clinical ophthalmology and other medical specialties. ¹ Widespread application of this powerful tool in animal research, however, was restricted because of poor image quality in commercially available first- and second-generation time domain OCT devices. ^2,3 Sufficiently high image quality could be achieved only with prototype devices with improved depth resolution that were specifically adapted for the respective animal visual system. ^4–7 Only recently, latest generation SD-OCT devices have become commercially available that provide scanning speed and depth resolution sufficient for small animal experimentation. ⁸ Given that SD-OCT is ideally suited for studying changes of retinal integrity, wild-type mice and mouse models of retinal degeneration were chosen to explore the efficacy of this method. 

Thus far, animal models of retinal degeneration have been studied extensively on both functional and structural levels. Morphologic changes were detected using either en face imaging methods such as funduscopy ⁹ and confocal scanning laser ophthalmoscopy (cSLO) ¹⁰ or, more commonly, ex vivo histologic approaches. Although light and electron microscopy provide (ultra-)high structural resolution, fixation procedures, dehydration preparatory to subsequent embedding, and staining, even the processes of cutting, flattening, and mounting histologic sections are potential sources for significant and variable alterations in dimensions. ^11,12 Indeed, retinal thickness measures of rodent retina are reported to be 95 μm in one study ¹³ and 170 to 217 μm in other studies. ^11,14 Taken together, ex vivo analysis of retinal tissue has certain limitations and should be interpreted accordingly. In vivo analysis of the retina has significant benefits as delicate, but functionally important changes such as retinal edema and focal neurosensory and pigment epithelial detachments can readily be detected by SD-OCT while potentially remaining undetected or masked by handling procedures in histologic analysis. Finally, monitoring of dynamic changes in individual animals requires a noninvasive study design. Here, SD-OCT has the potential to complement the existing in vivo methods in vision research by providing histology-analog structural details on retinal integrity. Because individual animals may be investigated at multiple time points, SD-OCT can also help to reduce the numbers of animals needed for such a study, which has both ethical and economic implications. To the best of our knowledge, this is the first study to demonstrate the efficacy of a commercially available SD-OCT device to obtain cross-sectional images from wild-type mice and mouse models of retinal degeneration. 

Materials and Methods

Animals

Animals were housed under standard white cyclic lighting (200 lux), had free access to food and water, and were used irrespective of gender. Rho ^{−/− 15} (n = 46), RPE65 ^{−/− 16} (n = 10), BALB/c (n = 15; Charles River Laboratories), C57/BL6/J (n = 37; Charles River Laboratories), C3H rd1/rd1 (n = 6, rd1), and control C3H wild-type (n = 6) mice ¹⁷ were used for imaging. All procedures were performed in accordance with the local ethics committee, German laws governing the use of experimental animals, and the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Because of the critical changes after postnatal day (P) 11, ^18,19 imaging in rd1 and corresponding wild-type were carried out at P11 and P28. Because of the later onset and slower course of degeneration in the Rho ^−/− and RPE65 ^−/− mouse models, imaging was performed at 1 and 11 months, respectively. 

Confocal Scanning Laser Ophthalmoscopy

For en face retinal imaging, we used the commercially available HRA 1 and HRA 2 (Heidelberg Engineering, Heidelberg, Germany) featuring up to two argon wavelengths (488/514 nm) in the short-wavelength range and two infrared diode lasers (HRA 1, 795/830 nm; HRA 2, 785/815 nm) in the long-wavelength range. A detailed protocol for anesthesia and imaging is described elsewhere. ¹⁰ Briefly, mice were anesthetized by subcutaneous injection of ketamine (66.7 mg/kg) and xylazine (11.7 mg/kg), and their pupils were dilated with tropicamide eye drops (Mydriaticum Stulln; Pharma Stulln GmbH, Stulln, Germany) before image acquisition. By applying hydroxypropyl methylcellulose (Methocel 2%; OmniVision, Puchheim, Germany) on the eye, the refractive power of the air corneal interface was effectively negated. A custom-made contact lens (focal length, 10 mm) was used to reduce the risk of corneal dehydration and edema and to act as a collimator. 

Spectral Domain Optical Coherence Tomography

SD-OCT imaging was performed in the same session as cSLO (i.e., animals remained anesthetized using identical preparatory steps). Mouse eyes were subjected to SD-OCT with a commercially available HRA+OCT device (Spectralis; Heidelberg Engineering, Heidelberg, Germany) featuring a broadband superluminescent diode at λ = 880 nm as a low coherent light source. Each two-dimensional B-scan recorded at 30° field-of-view consists of 1536 A-scans acquired at a speed of 40,000 scans per second. Optical depth resolution is approximately 7 μm, with digital resolution reaching 3.5 μm. ⁸  

To adapt for the optical qualities of the mouse eye, we mounted a commercially available 78-D double aspheric fundus lens (Volk Optical, Inc., Mentor, OH) directly in front of the camera unit. Imaging was performed with a proprietary software package (Eye Explorer, version 3.2.1.0; Heidelberg Engineering). Length of the reference pathway was adjusted manually according to manufacturer's instructions using the “OCT debug window” (press Ctrl/Shift/Alt/O simultaneously to open window in the active, calibrated OCT mode) to adjust for the optical length of the scanning pathway. The combination of scanning laser retinal imaging and SD-OCT allows for real-time tracking of eye movements and real-time averaging of OCT scans, reducing speckle noise in the OCT images considerably. ⁸ Resultant data were exported as 8-bit grayscale image files and were processed with a graphics editing program (Photoshop CS2; Adobe Systems, San Jose, CA). For quantification of central retinal thickness based on high-resolution volume scans, we used the proprietary software (Eye Explorer; Heidelberg Engineering). Briefly, each volume scan consisted of at least 70 B-scans recorded at 30° field-of-view centered on the optic disc, which were used to calculate an interpolated retinal thickness map across the scanned retinal area. Central retinal thickness was quantified using the circular OCT grid subfield at 3-mm diameter, with the center located on the optic disc. 

Histology

Three animals from each mouse line (Rho ^−/−, RPE65 ^−/−, BALB/c, and C57/BL6) were killed, and their eyes were enucleated for histologic analysis. After orientation was marked, the eyes were fixed overnight in 2.5% glutaraldehyde prepared in 0.1 M cacodylate buffer and were processed as described previously. ²⁰ Semithin sections (0.5 mm) of Epon-embedded tissue were prepared form the central retina, counterstained with methylene blue, and analyzed using a light microscope (Axiovision; Zeiss, Jena, Germany). 

For the timeline analysis of rd1 mice (n = 3 for each group: rd1 P11, rd1 P28, wild-type P11, and wild-type P28), eyes were embedded unfixed in Jung tissue freezing medium (Leica Microsystems, Wetzlar, Germany), frozen, and sectioned (14 μm) in a cryotome (HM560; Microm, Walldorf, Germany). Sections were then stained using hematoxylin/eosin staining. Morphologic observations and light microscopy were performed on a microscope (Imager.Z1 Apotome; Zeiss) equipped with a digital camera (Axiocam MrN; Zeiss). Images were captured using corresponding software (Axiovision 4.6; Zeiss). 

SD-OCT versus Histology

For correlation of retinal thickness measurements, inner retina was defined as ranging from the inner limiting membrane to the outer plexiform layer (OPL) and outer retina was defined as ranging from the outer nuclear layer (ONL) to the retinal pigment epithelium (RPE). In case of morphometric assessment of histologic sections, all measurements were taken at 1.4 mm eccentricity from the optic nerve head to mirror the SD-OCT-based thickness measurement along a circular ring scan (r = 1.4 mm) centered on the optic nerve head. Respective thickness was quantified by computer-assisted manual segmentation analysis using the proprietary software (Eye Explorer; Heidelberg Engineering) for SD-OCT data and the calibrated “line-tool” in Adobe (Photoshop CS2) for histologic micrographs. 

Student's t-test was used to analyze statistical significance between respective inner and outer retinal thickness measurements (C57/Bl6 vs. Rho ^−/−, RPE65 ^−/−, or BALB/c; C3H wild-type vs. C3H rd1/rd1 at P11 and P28). Correlation between thickness measurements by OCT versus histology was assessed by Pearson's correlation coefficient. 

Results

C57BL/6

A basic examination using cSLO en face imaging was performed in 4-week-old C57BL/6 mice to confirm the presence of regular retinal and vascular structures typical for normal wild-type animals. The examination included native red-free (RF; 513 nm), infrared (IR; 830 nm), and autofluorescence (AF) modes fluorescein angiography and indocyanine green angiography (Fig. 1). The cross-sectional SD-OCT imaging in these animals provided detailed in vivo data on retinal layer composition (Fig. 1) and retinal thickness. In a central circular area of approximately 3-mm diameter, total retinal thickness was 237 ± 2 μm (mean ± SD). The laminar organization of the murine retina, as determined in vivo by SD-OCT, correlated well with ex vivo light microscopy studies (Fig. 1). Notably, some wild-type mice feature Bergmeister's papilla, a structural remnant of the developmental hyaloid vascular system (HVS) at the optic disc. ²¹ Indeed, angiography showed characteristic perfusion of the HVS in mice until approximately P10 (data not shown), after which the hyaloid artery usually obliterates and leaves behind the channel of Cloquet. ²² In some animals, however, this development remains incomplete, leading to Bergmeister's papilla, as seen in Figure 1F. 

Figure 1.

 

Retinal SLO imaging and OCT in C57BL/6 mice (PW4) with a regular retinal structure. (A–D) Representative example of en face imaging using cSLO. (A) Native IR (830 nm), (B) RF (513 nm), and (C) AF mode. (D) Fluorescein angiography (FLA) confirming an intact retinal vasculature (arrowhead, artery; arrow, vein). (E–H) Corresponding OCT data. (E) Fundus picture with indicated orientation of cross-sectional SD-OCT scans. (F) Corresponding B-scan at the optic nerve head displaying a Bergmeister's papilla characterized by a structural remnant of the developmental hyaloid vascular system (asterisk), including enlarged image. (G) Light microscopic data of an age-matched C57BL/6 animal. (H) Representative B-scan with retinal layers labeled. GC/IPL, ganglion cell/inner plexiform layer; OLM, outer limiting membrane; RPE/ChC, RPE/choriocapillary complex.

BALB/c

En face imaging in BALB/c mice (postnatal week [PW]10) using cSLO showed nonpigmented retinal structures in the native RF (513 nm), IR (830 nm), and AF modes (Fig. 2). Given that BALB/c mice are highly susceptible to light-induced retinal degeneration, ^23,24 the hyperfluorescent dots in the AF mode may reflect lipid degradation products. Accordingly, overall retinal thickness in the central 3 mm, as determined by SD-OCT, was 211 ± 10 μm, considerably less than in the C57BL/6 strain, which would be hypothetically consistent with a mild light-induced photoreceptor loss. Virtual cross-sections displayed identical laminar organization in the inner retina compared with C57BL/6. However, because lack of pigmentation in BALB/c mice resulted in altered optical characteristics in the outer retina, signal composition of SD-OCT scans distal of the external limiting membrane (ELM) differed considerably (Fig. 2). Specifically, instead of two highly reflective layers (HRLs) thought to represent the inner/outer segment border (I/OS) and RPE (Fig. 1), SD-OCT scans in BALB/c mice demonstrated two additional HRLs possibly demarcating the nonpigmented choriocapillary and choroidal structures (Fig. 2). 

Figure 2.

 

Retinal SLO imaging and OCT in BALB/c mice (PW10). Typical results of cSLO en face imaging in the presence of nonpigmented retinal structures in (A) native IR (830 nm), (B) RF (513 nm), and (C) AF modes. (D) Fluorescein angiography in BALB/c mice displayed both retinal and choroidal vasculature (asterisk) because lack of pigment allowed light at λ = 488 nm to better penetrate the RPE/choriocapillary complex. (E) Fundus image with indicated orientation (F, H) of corresponding B-scans. (G) Representative histologic data of an age-matched BALB/c mouse. (H) Virtual cross-section with designation of the different retinal layers. Note the altered signal composition in the outer retina because of the absence of pigmentation in BALB/c mice (F). There are four HRLs possibly demarcating the I/OS, RPE, nonpigmented choriocapillary, and choroidal structures. GC/IPL, ganglion cell/inner plexiform layer; OLM, outer limiting membrane; RPE/ChC, RPE/choriocapillary complex; Cho, choroid.

Rho ^−/−

The visual pigment of the rod photoreceptors, rhodopsin, is an essential element of the phototransduction cascade and serves as a structural protein of the discs in the rod outer segments (ROS). Hence, in rhodopsin-deficient mice (Rho ^−/−), ROS are never formed, and rod-derived ERG signals cannot be generated. ¹⁵ En face imaging in Rho ^−/− mice (PW4) using cSLO showed characteristic signs of retinal degeneration with RPE irregularities visible in the native RF (513 nm, not shown), IR (830 nm), and AF modes (Fig. 3). SD-OCT imaging revealed a complete lack of ROS, whereas the ONL appeared only marginally thinner than the wild-type retina. Accordingly, central retinal thickness in Rho ^−/− mice was reduced to 193 ± 2 μm. The entire retina showed an absence of ROS, but the outer limiting membrane seemed not to be disturbed and could be distinguished from the RPE signal by a distance roughly equivalent to the extent of inner segment remnants (Fig. 3). Similarly, histologic sections showed an absence of ROS, whereas the outer nuclear layer appeared essentially intact with approximately six to eight rows of nuclei (Fig. 3). 

Figure 3.

 

Retinal degeneration and RPE irregularities in the Rho ^−/− mouse model (PW4). (A) Native IR (830 nm) and (B) AF mode cSLO data. (C) SD-OCT section and enlarged image, revealing a complete absence of ROS together with an apparent thinning of the ONL. (D) Ex vivo histology for comparison, confirming the observed reduction in ONL thickness and the lack of ROS. GC/IPL, ganglion cell/inner plexiform layer; OLM, outer limiting membrane; IS, photoreceptor inner segments; RPE/ChC, RPE/choriocapillary complex.

Rpe65 ^−/−

Mutations in the gene encoding RPE65 causes LCA2, a major form of Leber's congenital amaurosis, ^25–27 which is targeted in current clinical gene therapy trials. ^28,29 The protein RPE65 is expressed in the RPE, where it plays a pivotal role in maintaining normal vision by regenerating the visual pigment rhodopsin. ¹⁶ In Rpe65 ^−/− mice, the blocked visual cycle causes an accumulation of retinyl esters in RPE cells, where they form large lipid droplets (Fig. 4A). Cone photoreceptors degenerate rapidly, whereas the remaining rods are the exclusive source of electrophysiological response and start to degenerate slowly only after approximately 6 months of age. ^30,31 En face imaging at postnatal month (PM) 11 showed a characteristic pattern of hyperfluorescent flecks in the AF mode, which may again indicate metabolic remnants of photoreceptor outer segments (Fig. 4). Image resolution of SD-OCT in Rpe65 ^−/− mice was insufficient to resolve intracellular lipid accumulations in vivo. However, cross-sectional images revealed a reduction of ONL size, resulting in decreased total central retinal thickness of 193 ± 2 μm, which is more than the expected roughly 10% reduction with age and would be consistent with the observed AF. ³² Laminar organization in Rpe65 ^−/− mice was not as clearly delineated, perhaps because of a generalized reaction to the ongoing photoreceptor loss. Indeed, it has been shown that the genetic response in Rpe65 ^−/− mice includes modified expression of cytoskeletal elements and components of the extracellular matrix. ³³ Such a global response might reasonably affect the optical characteristics of the retinal sublayers. 

Figure 4.

 

Retinal degeneration and RPE irregularities in the Rpe65 ^−/− mouse model (PM11). (A) Histology showing lipid droplets from stored retinyl esters (arrows) in the RPE, together with a reduced retinal thickness. (B) Spots of hyperfluorescence in the AF mode, suggesting the presence of photoreceptor debris. (C) OCT cross-sectional image confirming a reduction of ONL thickness, though laminar organization (enlarged image) was not as clearly delineated as in wild-type mice. GC/IPL, ganglion cell/inner plexiform layer OLM, outer limiting membrane; RPE/ChC, RPE/choriocapillary complex.

rd1

The autosomal recessive retinal degeneration in the rd1 mouse is caused by a loss-of-function mutation in the gene encoding the β subunit of the rod cGMP phosphodiesterase 6 (PDE6β). ³⁴ Numerous mutations in the catalytic domain of the human homolog, PDE6β, have been found in human patients with autosomal recessive retinitis pigmentosa (arRP). ^35,36 Hence, the rd1 mouse and other mouse strains harboring PDE6β mutant alleles are considered relevant animal models of retinitis pigmentosa. 

First signs of rod photoreceptors cell loss in rd1 mice become evident at approximately, P10 when looking at TUNEL stainings, whereas ONL reduction become evident at age P12 and P13 and most rod photoreceptor nuclei have disappeared by P21. ^37,38 En face imaging in rd1 mice at P11 failed to detect signs of retinal degeneration. In contrast, rd1 mice at P28 revealed significant retinal degeneration both in the cSLO and SD-OCT imaging (Fig. 5). Cross-sectional images showed an intact inner retina, whereas the inner nuclear layer (INL) seemed to border the RPE almost directly with the OPL/ONL and photoreceptor segments virtually nonexistent. Although central retinal thickness in rd1 at P28 showed a marked reduction to only 145 ± 5 μm, rd1 mice at P11 featured a central retinal thickness at 246 ± 2 μm. The rapid progression of retinal degeneration in the rd1 model has been analyzed before using a custom-made SD-OCT device. ⁵ This study adds to the published data by presenting the onset of retinal degeneration in this model and resolving the retinal architecture before photoreceptor cell death (P11). 

Figure 5.

 

En face and SD-OCT imaging in rd1 and corresponding wild-type control mice at P11 and P28. At P11, no signs of retinal degeneration were evident in (A, D) cSLO or (B, E) SD-OCT imaging. (C, F) Histology confirmed normal ONL thickness. In contrast, rd1 mice at P28 revealed significant retinal degeneration (G). This pattern was also evident in the SD-OCT cross-sectional images (H) and corresponding histology (I), where inner retinal layers appeared intact, whereas the INL seemed to border the RPE almost directly. OPL/ONL and photoreceptor segments were virtually nonexistent. (J–L) Wild-type controls at P28 showed no signs of retinal degeneration. Scale bars, 100 μm.

SD-OCT versus Histology

Direct comparison of inner and outer retinal thickness assessed either by SD-OCT or histomorphometric analysis revealed excellent correlation (Figs. 6A–D). Consistent with existing data based largely on histologic assessment, ^15,31,37 SD-OCT data revealed a significant reduction of the outer retinal thickness in Rho ^−/−, RPE65 ^−/−, and BALB/c mice compared with C57/Bl6 animals (Fig. 6A). Similarly, timeline analysis of C3H wild-type versus C3H rd1/rd1 mice at P11 and P28 showed significant changes only in the rd1 mice at P28 (Fig. 6B). Traditional histomorphometric assessment mirrored the findings gained by SD-OCT data analysis (Fig. 6C), and direct comparison between the two data sets revealed excellent correlation coefficients for the whole retinal thickness (R ² = 0.897) and outer retinal thickness, respectively (R ² = 0.978). When analyzing the two data sets stemming from different protocols for histology separately (data not shown), the protocol using overnight fixation (Rho ^−/−, RPE65 ^−/−, BALB/c, and C57/Bl6) featured the lower correlation coefficient (r ² = 0.802) compared with the protocol using direct embedding, freezing, and sectioning (C3H wild-type and C3H rd1/rd1 mice at P11 and P28 [r ² = 0.954]). This could arguably be attributed to differential shrinkage during processing for histology, and it highlights the benefit of noninvasive assessment of retinal thickness by SD-OCT. 

Figure 6.

 

Evaluation of inner and outer retinal thickness in wild-type mice and animal models of hereditary retinal degeneration using either SD-OCT in vivo imaging or conventional morphometric assessment by histology. (A–C) Bar graphs indicate inner (dark) and outer (light) retinal thickness. (A) Note statistically significant reduction of outer but not inner retinal thickness in BALB/c, Rho ^−/− and RPE65 ^−/− compared with C57/BL6 mice. (B) There is no statistical significant difference between rd1 and wild-type mice at P11. Although outer retina remained unchanged in the wild-type mice at P28, there is complete loss of outer retina in the age-matched rd1 mice. (C) Conventional morphometric assessment of retinal thickness in histologic sections with similar changes as reported by SD-OCT (A, B). Significance was calculated using Student's t-test. ns, not significant (P > 0.01); *P ≤ 0.01; **P ≤ 0.001; *** P ≤ 0.0001. (D) Scatter plot shows the correlation of histologic versus OCT data on total (dots) and outer (error bars only) retinal thickness from C57/BL6, BALB/c, Rho ^−/−, RPE65 ^−/−, rd1-wt P11, rd1 P11, rd1-wt P28, and rd1 P28 mice. Pearson's correlation coefficients (r ²) are displayed for total retinal thickness (y = 0.975x) and outer retinal thickness (y = 1.018x) separately. All data are reported as mean ± SD (error bars).

Discussion

OCT has emerged as a valuable tool to analyze and monitor structural changes in the retina. Although noninvasive in character, the resolution of cross-sectional images obtained by third-generation OCT begins to approach the level of low-power micrographs gained from light microscopy. This allows for detailed in vivo structural analysis of retinal disorders and bears importance for a refined genotype/phenotype correlation in a clinical setting and for noninvasive longitudinal studies on animal models that mimic respective disease characteristics. Indeed, various prototype OCT setups have been reported to allow noninvasive, high-resolution imaging of rodent retina. ^6,7 Srinivasan et al. ⁷ used a custom build OCT setup to scan a normal C57/Bl6 mouse and to analyze the retinal thickness in a virtual cross-section obtained from a Long-Evans rat. Ruggeri et al. ⁶ demonstrated noninvasive retinal imaging (e.g., in a mouse model of retinal degeneration, Rho ^−/−) using an experimental OCT system. Both studies compare single virtual cross-sections with conventional histology without quantification, whereas others have used custom OCT systems to monitor the dynamic changes of natural disease progression in respective animal models. ^2,3,5 Yet the potential to evaluate therapeutic or adverse effects of experimental interventions in time-course experiments might prove to be even more important. Consequently, comparability between individual preclinical studies and preferably also between the preclinical and clinical setting will be of eminent importance. This is a challenge that is best addressed using commercially available equipment already approved for clinical use. In this study, we present data from frequently used wild-type mice and models of hereditary retinal degeneration using a commercially available third-generation OCT device approved for clinical diagnostic use. 

The identity of signal composition in linear A- and two-dimensional B-scans acquired by OCT imaging has been subject to debate. However, it is generally thought ^1,39,40 that—from vitreous to sclera—the first thin dark band resembles the nerve fiber layer followed by the lighter ganglion cell layer. The consecutive thick band of higher intensity illustrates the inner plexiform layer (IPL), approximated by a light band, the INL. The OPL appears as a thin stripe with signal intensity comparable to that of the IPL. The outer retina is composed of a thick light band, and ONL signal intensity is similar to that of INL. Apparently, both nuclear layers share the low signal intensity because they scatter or reflect light to a lesser extent than do both plexiform layers, which appear much darker. Located just distal to the ONL is the ELM, which is formed by adherence junctions between apical villi of Müller glia cells and distal photoreceptor cell bodies. Photoreceptor inner segments appear with the same signal intensity as the cell bodies, whereas the I/OS generates a strong signal possibly because of the high mitochondrial content in the ellipsoid region of the outer segment. ³⁹ The photoreceptor outer segments are visible between the I/OS border and the equally strong signal of the RPE/choriocapillary complex. In pigmented retinas, choroidal structures cannot be reliably detected because light is almost completely scattered or reflected as it travels through the more proximal layers. In nonpigmented retinas, the signal composition is considerably different distal of the ELM (Fig. 2), which argues for scatter/reflection attributed to melanin granules as playing the main part in this phenomenon. 

Earlier work exploring the correlation between OCT and histology in mice was performed primarily with second-generation, time-domain ² or custom-made, high-resolution OCT devices ⁵ and reported a substantial overestimation of retinal thickness by OCT. ³ In this study we used a commercially available SD-OCT device and obtained high overall correlation of total retinal thickness (R ² = 0.897) with only minor overestimation (2.5%) of retinal thickness by OCT. Interestingly, correlation of outer retinal thickness representing the photoreceptor cell layer separately showed an even higher Pearson's correlation coefficient (R ² = 0.978, Fig. 6D). Overall, the results obtained in vivo are in good agreement with earlier studies on the three animal models for retinal degeneration. ^15,31,37  

Although conventional histology features higher resolution, OCT delivers the advantage of producing morphologic data undistorted by handling, fixating, and staining procedures while being coregistered with en face topologic information. The latter aspect is particularly useful when assessing retinal degeneration types that do not show a uniform progression in the entire retina. OCT-based screening may also be useful for breeding purposes because it enables a noninvasive assessment of retinal health in mouse strains that are prone to develop sporadic retinal degeneration. 

Analogous to electroretinography, by which the use of identical hardware in the clinical setting and laboratory has led to new insights, ³¹ the application of a clinically approved OCT device for animal studies bears the potential to translate insights from bench to bedside in an efficient and timely manner. Here, we present evidence on the efficacy of a commercially available SD-OCT in small animal retinal imaging and provide in vivo structural data on mouse models of retinal degeneration. This should facilitate further studies on dynamic changes of retinal structure through the natural course of disease and should help to monitor putative therapeutic effects of novel interventional strategies. 

Footnotes

 Supported by Deutsche Forschungsgemeinschaft Grants Se837/5–2 & Se837/7–1 (KFO 134), Se837/6–1, and PA1751/1–1; German Ministry of Education and Research Grant 0314106; European Union Grants LSHG-CT-512036 and EU HEALTH-F2–2008-200234, Kerstan Foundation, EU (MEST-CT-2005–020235).

Footnotes

 Disclosure: G. Huber, None; S.C. Beck, None; C. Grimm, None; A. Sahaboglu-Tekgoz, None; F. Paquet-Durand, None; A. Wenzel, Novartis Pharma Schweiz (E); P. Humphries, None; T.M. Redmond, None; M.W. Seeliger, None; M.D. Fischer, None

Footnotes

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