Peripapillary Rat Sclera Investigated In Vivo With Polarization-Sensitive Optical Coherence Tomography

Abstract

Purpose.: To demonstrate polarization-sensitive (PS) optical coherence tomography (OCT) for noninvasive, volumetric, and quantitative imaging of the birefringent properties of the peripapillary rat sclera; to compare the findings from PS-OCT images to state-of-the-art histomorphometric analysis of the same tissues.

Methods.: A high-speed PS-OCT prototype operating at 840 nm was modified for imaging the rat eye. Densely sampled PS-OCT raster scans covering an area of ~1.5 × 1.5 mm centered at the papilla were acquired in the eyes of anesthetized male Sprague-Dawley rats. Cross-sectional PS-OCT images were computed, and fundus maps displaying the birefringent properties of the sclera were analyzed. Postmortem histomorphologic analysis was performed.

Results.: Polarization-sensitive OCT enables visualization of the polarization properties of ocular tissues in vivo. The birefringent characteristics of the rat sclera were quantitatively assessed. Scleral birefringence formed a donut-shaped pattern around the papilla with significantly increased values of 0.703 ± 0.089°/μm (i.e., 1.64 × 10⁻³ ± 0.2 × 10⁻³; mean ± standard deviation) and 0.721 ± 0.084°/μm (i.e., 1.68 × 10⁻³ ± 0.2 × 10⁻³) at an eccentricity of 0.4 mm for the left and right eyes, respectively. Birefringent axis orientation maps revealed a ring-shaped distribution around the optic nerve. Postmortem PS-OCT micrographs provided access to retinal and scleral microstructure and were compared to standard histomorphologic analysis.

Conclusions.: Polarization-sensitive OCT enables quantitative imaging of tissue polarization properties in addition to conventional OCT imaging based on reflectivity. In the rat sclera, in vivo PS-OCT provides access to volumetric mapping of birefringence. Scleral birefringence is associated with microstructural tissue organization. Therefore, PS-OCT should prove a valuable tool for the in vivo investigation of peripapillary sclera in glaucoma.

Introduction

The sclera is the opaque and protective outer layer of the eye, containing collagen and elastic fiber. Recent work has suggested a major role of scleral biomechanics in the pathogenesis of glaucoma, one of the leading causes of blindness worldwide.¹ Glaucoma leads to vision loss by damaging the retinal ganglion cell axons and may be associated with increased intraocular pressure (IOP). The principal site of damage is the region around the optic nerve head (ONH). Recent studies not only demonstrated that the biomechanical properties of the human sclera change due to age and glaucoma, but also revealed a relation between scleral biomechanical behavior and susceptibility to retinal ganglion cell loss in experimental glaucoma.^2–7 The obvious impact on ganglion cell damage and thus on the development of glaucoma reinforced the necessity for proper tools for investigating scleral structure in situ and in vivo. 

Several methods have been proposed for exploring the sclera's structure in vitro. Ultrastructural techniques such as electron microscopy of histologic samples were used to visualize the fibrous microstructure in the sclera.⁸ Second harmonic generation (SHG) microscopy was used to specifically assess collagen compounds. Using small-angle light scattering and wide-angle X-ray scattering, the collagen fiber orientation was measured around the ONH.^9,10 Imaging techniques based on polarized light take advantage of the birefringence caused by collagen fibers. Polarization light microscopy gives access to quantitative information about the scleral collagen organization and enables imaging of differently oriented fibers.^11,12 

Polarization-sensitive (PS) optical coherence tomography (OCT) is an emerging functional extension of OCT, which—in addition to providing conventional OCT images based on the intensity of backscattered light—provides image contrast based on the light's polarization state.^13,14 In ophthalmology, PS-OCT has proven useful for improved segmentation and visualization of pigmented structures such as the retinal pigment epithelium in age-related macular degeneration^15–17 and may be of particular interest for the measurement of retinal nerve fiber layer birefringence in glaucoma.^18–20 In this article, we present PS-OCT as a noninvasive optical method for in vivo characterization, mapping, and quantification of the scleral collagen compounds' birefringence in the rat eye, which can readily be expanded to longitudinal studies. 

Materials and Methods

Animals

Male Sprague-Dawley rats (age 7–8 weeks) were purchased from the Medical University of Vienna breeding facility and kept under controlled lighting conditions (12 hours light, 12 hours dark). All experiments were performed in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and under a protocol approved by the ethics committee of the Medical University of Vienna and the Austrian Federal Ministry for Science and Research (protocol number GZ 66.009/0034-II/3b/2013). 

In Vivo Polarization-Sensitive Optical Coherence Tomography

A spectral-domain PS-OCT prototype based on polarization-maintaining (PM) fiber optics was modified for imaging the rat retina (Fig. 1A). A detailed description of the system has been presented previously.²¹ In brief, the instrument employed a superluminescent diode at 837-nm wavelength with a spectral bandwidth of 52 nm as a light source. Circularly polarized light was used for illuminating the eye. In the PS detection unit, the OCT signal was split into two orthogonal polarization components, which were detected by two identical spectrometer units. The spectrometer line scan cameras were operated at a line rate of 70 kHz. 

Figure 1

 

In vivo PS-OCT imaging in the rat eye. (A) Sketch of PS-OCT prototype. SLD, superluminescent diode, 2 × 2 fiber coupler; PBS, polarizing beam splitter; SP, spectrometer. (B) Volume rendering of PS-OCT reflectivity data set. RE, retina; SC, sclera; EOF, extraorbital fat. (C–E) PS-OCT B-scan images. (C) Reflectivity image. (D) Phase retardation image. (E) Axis orientation image.

Rats were anesthetized using ketamine (Ketasol [aniMedica GmbH, Senden-Bösensell, Germany]; 80 mg/kg body weight, intraperitoneal) and xylazine (Rompun [Bayer Austria GmbH, Vienna, Austria]; 8 mg/kg body weight, intraperitoneal) in order to immobilize the animal for PS-OCT imaging. The pupils were dilated using tropicamide (Mydriaticum [Agepha Pharmaceuticals, Vienna, Austria]; topical) and phenylephrine (2.5%, topical). The rats were placed in a custom-made mount allowing translation and rotation for aligning the eye with respect to the measurement beam. 

The retina was raster scanned using a pair of galvanometer scanners (Thorlabs, Inc., Newton, NJ, USA) and a demagnifying telescope. The beam diameter at the cornea was 0.85 mm. A scan field of 23° × 23°, corresponding to an area of 1.5 × 1.5 mm, around the ONH was covered by the OCT scan. The image range in beam direction was 3.6 mm. Data sets of 512 (x) × 400 (y) × 2048 (z) voxels were recorded in ~3 seconds. 

Ex Vivo Polarization-Sensitive Optical Coherence Microscopy

The PS-OCT system was also used in a scanning confocal microscopy configuration for imaging unstained histologic tissue sections of rat eyes. A magnifying telescope (magnification 1.8×) placed after the galvanometer scanner increased the beam diameter to 2.3 mm and relayed the scanned beam onto the objective lens. Microscope objectives (Owis 20×, Nachet 40×) were used to focus the beam onto the samples. Data sets covering areas of 480 × 100 and 240 × 240 μm were acquired with the 20× and 40× objectives, respectively. Histologic sections were covered with a thin film of phosphate-buffered saline (PBS) in order to reduce reflectivity at air–tissue interfaces. Gross overview images covering an area of 3.5 × 4.0 mm were acquired using a single 50-mm focal length achromatic lens in a telecentric scanning configuration. 

Image Processing

Polarization-sensitive OCT images were generated as previously described.²² In addition to standard OCT images based on reflectivity, PS-OCT images displaying phase retardation and fast optic axis orientation were computed. Polarization-sensitive OCT images were computed from the same raw data as the reflectivity images. Therefore, there is a pixel-to-pixel correspondence between reflectivity, retardation, and axis orientation images, which enables the transfer of segmentation locations for the analysis of ocular structures. 

For the in vivo PS-OCT images, the segmentation of the sclera was not straightforward and required the following processing steps. First, the vitreoretinal interface (inner limiting membrane, ILM) was detected in the reflectivity B-scan images based on the step-like increase of backscattered intensity. Second, the retinal pigment epithelium (RPE) was segmented as the brightest layer located >80 pixels (>140 μm) posterior to the ILM. Note that since albino Sprague-Dawley rats lack pigmentation in the RPE, depolarization in PS-OCT images cannot be exploited for RPE segmentation.²³ Third, the anterior border for analyzing the sclera was set by shifting the segmented RPE profile by 20 pixels (35 μm) in the posterior direction, that is, beyond the choroid. The posterior border for the scleral analysis was set 60 pixels (106 μm) posterior to the anterior border. 

Polarization effects of birefringent structures such as the cornea affecting the OCT beam have to be compensated in PS-OCT images in order to enable analysis of the polarization properties in the sclera. Hence, a map of polarization states at the segmented RPE level was generated and used for compensation analogous to our previously described approach.²⁴ Scleral birefringence was determined for each A-scan by computing the slope of a least-squares fit of phase retardation along the depth extension of the analysis window (Figs. 2A, 2B). The birefringence values for each A-scan were then color mapped into a two-dimensional image. For each eye, the location of the ONH was defined manually in the fundus maps. Radial birefringence profiles were automatically generated and averaged azimuthally around the ONH. Average scleral birefringence was computed for the nasal, superior, temporal, and inferior sectors. 

Figure 2

 

Mapping scleral birefringence. (A) PS-OCT B-scan images. Reflectivity (top) and phase retardation. A phase retardation A-scan is shown on the right. A gradual increase of retardation values can be observed in the sclera represented by the green band. Scleral birefringence, that is, the slope of the retardation increase over depth, was computed for all A-scans in the 3D data set. (B) Birefringence map. The location of the ONH is indicated by an asterisk. N, S, T, I: nasal, superior, temporal, inferior. Color bar range: 0.0 to 1.0°/μm. Azimuthally averaged birefringence values are shown for left and right eyes in (C) and (D), respectively. Each of the gray lines represents an individual rat. The average profiles of all left and right eyes are plotted as red dashed lines. (E) and (F) show the azimuthally averaged birefringence values at eccentricities of 200, 400, and 600 μm for left and right eye, respectively. Whiskers indicate the standard deviation. Statistically significant differences are denoted by * (P < 0.05) and *** (P < 0.001); n.s., not significant difference (P > 0.05). The mean birefringence in the N, S, T, and I sectors is shown for the left and right eyes in (G) and (H). Each gray bar represents an individual rat eye.

For the ex vivo PS-OCT images, the tissue surface was segmented similarly to the ILM in the in vivo images. Using the front surface as a backbone, the tissue back surface or the microscope glass slide surface beneath the tissue slab was segmented. Phase retardation and axis orientation images of the polarization states at the latter surface were generated. Further, en face projection images of backscattered intensity within the segmented volume were computed. 

Histologic Analysis

For histomorphologic analysis, rats were euthanized by an overdose of sodium pentobarbital (release ad usum veterinarium). The eyes were enucleated within minutes after death and placed in paraformaldehyde (PFA, 4%). In order to guarantee quick and thorough fixation of ocular tissues, a small incision was made into the ocular bulbus close to the corneal limbus, and PFA was injected into the eye. After 24 hours of fixation, the cornea and lens were removed. Following a further 24 to 48 hours of fixation at +4°C, the eyes were placed in solutions with increasing sucrose concentrations. Finally, the eyes were carefully embedded in optimal cutting temperature compound (O.C.T. Compound; Tissue-Tek, Sakura Finetek, Alphen aan den Rijn, The Netherlands) mixed with 20% sucrose and frozen for cryosectioning. Using a cryotome (Leica, Wetzlar, Germany), the frozen tissue blocks were sectioned into slabs with thicknesses between 6 and 45 μm. Unstained histologic sections were imaged with PS-OCT as described above and with standard bright-field microscopy and laser scanning confocal microscopy. Selected histologic sections were stained with hematoxylin and eosin (H&E) and imaged with a bright-field microscope (Axioimager; Carl Zeiss, Oberkochen, Germany). In order to enhance contrast of the sclera, polarization contrast micrographs were recorded. 

For the visualization of scleral collagen fibers, cryosections of 30-μm thickness were rinsed with distilled water and incubated with 0.1% sirius red F3B (Sigma-Aldrich Corp., St. Louis, MO, USA) in saturated picric acid for 1 hour at room temperature. Sections were rinsed for 5 minutes with acidified water (0.5% acetic acid in distilled water). After dehydration in 85% and 96% ethanol, sections were cleared in xylene and mounted in Entellan mounting medium (Merck Millipore, Billerica, MA, USA). 

Single images and z-stacks were acquired from each picrosirius red–stained tissue section using confocal laser scanning microscope LSM780 (Carl Zeiss) with a 20× (NA 0.8) and 63× oil lens (NA 1.4). A rhodamine filter set was used with an excitation wavelength of 561 nm and emission at 590 nm to visualize the collagen fibers. 

Statistical Analysis

Statistical comparison between birefringence measurements in left and right eyes as well as in eye sectors was performed by means of paired-samples two-tailed Student's t-test. A P value < 0.05 was considered statistically significant for all analyses. 

Results

The left and right eyes of eight Sprague-Dawley rats were imaged with the PS-OCT prototype. Exemplary PS-OCT images are shown in Figure 1. A volume rendering of a three-dimensional (3D) PS-OCT data set is shown in Figure 1B. The sclera can be observed as a hyperscattering layer sandwiched between the retina and extraorbital tissue. Polarization-sensitive OCT B-scan images are shown in Figures 1C through 1E. The reflectivity B-scan (Fig. 1C) provides access to the layered structure of retina and sclera as known from standard SD-OCT. Due to the lack of pigmentation in the albino rat, light is enabled to penetrate all the way through the sclera. The phase retardation B-scan (Fig. 1D) provides access to the birefringent tissue properties of ocular structures. Most layers in the retina do not affect the polarization state and therefore appear in uniform blue color. Increasing retardation values can be observed in the birefringent sclera. Consequently, defined birefringent axis orientation values are visible only in the sclera (Fig. 1E). A fast axis orientation of 0° refers to the vertical direction with respect to the OCT scanner, that is, the inferior–superior axis in the animal eye. Each B-scan is recorded in horizontal direction, that is, along the nasal–temporal axis in the animal eye. Hence, an axis orientation of 90° corresponds to the left–right direction in the B-scan images, and an axis orientation of 0° corresponds to the direction perpendicular to the image plane. As can be observed in Figure 1E, axis orientation is not subject to strong variations with depth. Pixels with intensities ≤ 5 dB above the noise level are displayed in gray in both Figures 1D and 1E. 

Peripapillary Scleral Birefringence

Polarization-sensitive OCT imaging was performed in the left and right eyes of eight rats. The right eyes of two animals were affected by retinal degeneration visible in the OCT images, and the results of scleral birefringence measurements in these rats were excluded from the following analysis. Fundus maps of peripapillary scleral birefringence were generated (Figs. 2A, 2B). Birefringence values increase from the ONH outward, peak at 0.7°/μm at an eccentricity of ~0.4 mm in both left and right eyes, and decrease again in the periphery. Profiles of birefringence averaged azimuthally, that is, averaged over an annular ring, and plotted as a function of eccentricity from the ONH are shown in Figures 2C and 2D. Azimuthally averaged birefringence values in close vicinity of the ONH (0.2-mm eccentricity), at 0.4-mm eccentricity, and at 0.6-mm eccentricity were compared. For the left eye, birefringence (mean ± standard deviation) was 0.329 ± 0.068, 0.703 ± 0.089, and 0.438 ± 0.112°/μm at eccentricities of 0.2 mm, 0.4 mm, and 0.6 mm, respectively (Fig. 2E). For the right eyes, the corresponding average birefringence values were 0.322 ± 0.089, 0.721 ± 0.084, and 0.493 ± 0.049°/μm, respectively (Fig. 2F). When comparing the values for different eccentricities within the same eyes, we found statistically significant differences between birefringence values at 0.2 mm and 0.4 mm, and between values at 0.4-mm and 0.6-mm eccentricity (P < 0.001). The difference between birefringence values at 0.2-mm and 0.6 mm-eccentricity was not significant for the left eyes (P = 0.108) but weakly significant for the right eyes (P = 0.011). 

Mean birefringence values were computed in the nasal, superior, temporal, and inferior peripapillary sector. The results for the left and right rat eyes are listed in the Table and shown in Figures 2G and 2H, respectively. Scleral birefringence was analyzed statistically for differences between peripapillary sectors as well as between left and right eyes. There was no statistically significant difference between corresponding sectors in left and right eyes. Nor was the difference between different sectors in the same eye significant. 

Table

 

Average Birefringence (in °/μm) Measured in Nasal (N), Superior (S), Temporal (T), and Inferior (I) Sectors of Left (OS) and Right Eyes (OD)

Animal ID	1		2		3		4		5		6
Eye	OS	OD	OS	OD	OS	OD	OS	OD	OS	OD	OS	OD
N	0.891	0.677	0.649	0.725	0.613	0.635	0.590	0.685	0.629	0.785	0.743	0.842
S	0.761	0.807	0.664	0.532	0.538	0.798	0.591	0.655	0.594	0.659	0.791	0.698
T	0.629	0.783	0.717	0.685	0.631	0.769	0.641	0.618	0.715	0.480	0.741	0.701
I	0.745	0.790	0.719	0.662	0.557	0.774	0.682	0.627	0.641	0.587	0.740	0.746

Birefringent Axis Orientation

Figure 3 shows fast optic axis orientation maps in the sclera at ~160 μm posterior to the RPE. In collagen fibers, the fast axis is parallel to the fiber direction.²⁵ An azimuthally varying orientation can be observed around the ONH. Axis orientation values were averaged within each of 30 sectors centered at the ONH. The results are plotted in Figure 3D and show an approximately circular arrangement. 

Figure 3

 

Mapping birefringent axis orientation. (A) Intensity projection fundus image. (B) Axis orientation image in the sclera ~160 μm posterior to the RPE. Color scale: −90° to 90°. S, T, I, N: superior, temporal, inferior, nasal. (C) Alternative axis orientation plot around the ONH. (D) Azimuthal variation of the axis orientation values around the ONH. Azimuth orientation starts at 0° at noon and proceeds clockwise. The dashed lines indicate the orientation values for a circular field around the ONH.

Ex Vivo Imaging of Ocular Tissue

In order to compare the appearance and morphology of ocular structures in PS-OCT images with conventional histology, cryosections of thicknesses between 7 and 45 μm were prepared and imaged with bright-field microscopy and PS-OCM. Figure 4A shows a bright-field confocal microscopy image of an unstained 7-μm cryosection. The retinal layers can be clearly distinguished, and detail at the cellular level such as single photoreceptors or erythrocytes can be observed. In the PS-OCT reflectivity image acquired with the 20× objective in a 40-μm section of the same eye (Fig. 4B), the retinal layers show varying contrast based on their reflectivity. Even finer structural details can be observed in the montage of three data sets recorded with the 40× objective in a 7-μm-thick histologic section of the same eye (Fig. 4E). Varying granularity is visible in the retinal layers. The stratified structure of the sclera is discernible in the right tile of the image. 

Figure 4

 

Ex vivo imaging of ocular tissue. (A) Bright-field microscope image of a Sprague-Dawley rat retinal frozen section of 7-μm thickness. (B–D) Adjacent PS-OCM images of a 40-μm-thick section of the same eye. (B) Reflectivity image. (C) Retardation image. Color scale: 0° to 90°. (D) Axis orientation image. Color scale: −90° to 90°. (E) PS-OCM reflectivity images of ocular tissue. Retinal layers, choroid, sclera, and extraocular muscles can be observed. NFL, nerve fiber layer; GCL, ganglion cell layer; IPL, inner plexiform layer; INL, inner nuclear layer; OPL, outer plexiform layer; ONL, outer nuclear layer; PR IS, photoreceptor inner segments; PR OS, photoreceptor outer segments; RPE, retinal pigment epithelium; CH, choroid, SC, sclera; EOM, extraocular muscle.

Birefringent structures are revealed in the corresponding retardation image, Figure 4C. Birefringence caused by the collagen fibers gives rise to increased retardation values to be observed in greenish color in the sclera. Rather low retardation values can be observed in the extraorbital muscles at the lower image boundary. While muscles exhibit strong birefringence in PS-OCT images when imaged perpendicular to their fiber orientation,^26,27 the muscles in Figure 4C were oriented obliquely with respect to the section plane, therefore having only a small birefringent component along the projection angle. However, in the axis orientation image, Figure 4D, the birefringent muscle tissue—as well as the sclera—shows a well-defined orientation. 

Both macroscopic and microscopic images of the posterior eye are shown in Figure 5. A mosaic of bright-field microscopy images of an unstained tangential section anterior to the papilla is displayed in Figure 5A. Bright-field micrographs of H&E-stained tissue sections are shown in Figure 5B. Polarization contrast micrographs of the same samples (Fig. 5C) highlight the birefringent sclera in alternating blue and yellow colors. Images from a 3D PS-OCT data set of the 45-μm-thick tissue slab of Figure 5A are shown in the bottom row of Figure 5. Reflectivity, retardation, and axis orientation images analogous to the mosaic of micrographs are shown in Figures 5D through 5F. The sclera can be observed as a birefringent ring around the retina in the central circle. Cross-sectional PS-OCT B-scan images of the same data set are shown in Figures 5G through 5J. 

Figure 5

 

Ex vivo imaging of peripapillary tissue. (A) Mosaic of bright-field microscopy images of an unstained tangential section anterior to the papilla. (B) Bright-field micrograph of an H&E-stained retinal frozen tissue section. (C) Polarization contrast micrograph of (B). Scleral birefringence manifests in alternating blue and yellow colors. (D–F) En face OCT images of the tissue slab. (G–J) Cross-sectional OCT images at the location indicated by the dashed line in (D). (D, G) Reflectivity. (E, H) Retardation. (F, J) Axis orientation. PS-OCT color maps as in Figure 4.

Collagen fiber organization was visualized in tangential cryosections by means of picrosirius red staining. A 30-μm-thick tangential eye section is shown in Figure 6. Collagen fiber–rich structures—muscle tissue and sclera—appear in lush red color (Fig. 6A). A PS-OCT retardation image of the same section (Fig. 6B) features increased values at the location of the sclera. Note that the birefringent scleral ring is disrupted by blood vessels in some locations (cf. Fig. 6D). Axis orientation images (Figs. 6C, 6F) reveal the annular orientation. Confocal fluorescence micrographs were acquired across the scleral ring. Collagen fiber bundles of different sizes can be observed in the mosaic of Figure 6E. While the fiber bundles appear thin close to the optic nerve, the scleral structure looks rather dense with thicker bundles at larger eccentricity. An annular orientation around the optic nerve similar to that in the axis orientation images can be observed. The depth locations of scleral fiber bundles mapped in different colors can be appreciated in the confocal micrograph at the inner scleral border (Fig. 6G). 

Figure 6

 

Ex vivo imaging of scleral fiber organization. (A) Tangential section stained with picrosirius red highlighting collagen-rich tissues (stitched micrograph). (B) PS-OCT retardation image. Scleral birefringence leads to increased retardation values. Color map as in Figure 4. (C) Axis orientation image. Color scale from −90° to +90°. 0° (green) corresponds to the up–down and ±90° (dark red/blue) to the left–right direction in the image, respectively. (D) Magnification of the micrograph section showing the sclera around the optic nerve (ON). Other structures such as blood vessels (BV), extraorbital fat (EOF), and muscles (EOM) can be observed. High-resolution z-stacks were acquired across the scleral annulus at the locations indicated in yellow. (E) Confocal fluorescence micrographs reveal collagen bundles in the sclera and their annular orientation. (F) Alternative axis orientation plot shows annular orientation of fast birefringent axes in the sclera around the ON. (G) Scleral fiber bundles close to ON. Color-coded depth location reveals layered structure of the fiber bundles. Color scale: 0 to 14 μm.

Discussion

Optical coherence tomography is emerging as a noninvasive alternative for imaging small animals in preclinical research. Polarization-sensitive OCT adds tissue-intrinsic contrast and enables quantitative measurements. In this study, we demonstrate the feasibility of PS-OCT to perform high-resolution, noninvasive imaging as well as to assess the birefringent properties of ocular structures at the posterior pole in the rat eye. 

Quantitative measurements of scleral birefringence have been reported only recently. A very recent PS-OCT study of the sclera close to the corneoscleral limbus reported birefringence values of 0.9 to 1.4°/μm in humans (in vivo) and 1.5 to 1.8°/μm in ex vivo porcine sclera.²⁸ The PS-OCT measurements in that report were performed at 1310-nm wavelength, such that comparisons between these birefringence values with those in rats shown here have to be made with caution. Assuming a linear relation of tissue birefringence and wave number, these values have to be scaled with a factor of 1.56, that is, 1310/840 nm, and correspond to 1.4 to 2.3°/μm in humans and 2.5 to 3.0°/μm in porcine sclera at 840-nm wavelength. Alternatively, a dimensionless number is often used for comparison by expressing birefringence in nanometer wavelength per thickness (nm/nm). For the human and porcine sclera, birefringence values thus correspond to 3.2 × 10⁻³ to 5.0 × 10⁻³ and 5.4 × 10⁻³ to 6.5 × 10⁻³. In the rat sclera investigated in the current study, considerably lower birefringence values ranging from 0.2 to 0.8°/μm (0.5 × 10⁻³–1.8 × 10⁻³) were measured. Species-specific parameters such as IOP or scleral microstructure may influence birefringence. Moreover, scleral characteristics differ significantly around the ocular bulbus.^9,28 However, even though modern ophthalmic OCT prototypes operating at long 1050-nm wavelengths provide sufficiently deep penetration for scleral imaging at the posterior pole,^29–31 no quantitative birefringence measurements have yet been reported in the human sclera at the posterior pole. To the best of our knowledge, the results presented here represent the first in vivo measurement of scleral birefringence both in the rat eye and at the posterior pole, and may therefore serve as a baseline for investigations in rodent models of glaucoma. 

A consistent donut-shaped pattern of significantly increased birefringence was observed closely surrounding the ONH in rats (Fig. 2). This pattern can be explained by the preferential alignment of scleral fibers found recently by small-angle light scattering (SALS) in rat scleras and, similarly, by wide-angle X-ray scattering (WAXS) in human scleras in vitro.^9,10 Similar to the preferentially circular alignment observed with SALS and WAXS, we found an azimuthally varying birefringent axis orientation around the ONH (Fig. 3). Furthermore, locally increased birefringence was also observed in PS-OCT images at the rim of the scleral canal in the human eye.^22,30 

The appearance of anatomical structures provided by novel imaging methods such as PS-OCT requires calibration with respect to conventional approaches such as histology. Animal experiments provide the possibility to compare 3D PS-OCT images acquired in vivo and in situ to ex vivo tissues sections prepared from the exact same eyes using established histomorphologic procedures. For this study, cryosections of rat eyes were investigated with microscopy and polarization contrast microscopy using standard H&E staining protocols. In addition, a PS-OCT microscopy setup was used for imaging the tissue sections. The layered appearance of the retinal, choroidal, scleral, and extraorbital structures in PS-OCT reflectivity images was similar to that in micrographs. Similar details were observed in an earlier study by Grieve and coworkers,³² who imaged ex vivo rodent eyes with full-field OCT. Even finer details of the granular retinal microstructure and stratification in the sclera were visualized with high-magnification PS-OCM (Fig. 4). Note that the appearance of the layered structure is different from that of standard in vivo OCT images such as Figure 1C since the histologic sample is illuminated perpendicularly to the layers in the OCM images, rather than along the optical axis of the eye as with in vivo OCT imaging. For instance, the photoreceptors appear as two highly reflecting bands in Figure 4B, while Figure 1C exhibits the negative contrast, that is, two dark bands framed by three highly reflecting stripes. 

The polarization sensitivity of our PS-OCM setup also provided access to birefringent tissues including collagen fibers. In addition to extraorbital muscles, in particular, the stratified scleral structure was clearly observed (Figs. 4, 5), in good agreement with conventional polarization contrast micrographs (Fig. 5C). However, a direct comparison of quantitative birefringence values in ex vivo and in vivo tissue should be avoided, since formaldehyde fixation leads to microstructural alterations and tissue shrinkage³³ and hence potentially to a change of the birefringent properties. 

We also imaged collagen fibers in ex vivo sclera using picrosirius red staining (Fig. 6). In tangential cryosections, the scleral collagen fibers exhibited an annular arrangement around the optic nerve (Figs. 6C–G), similar to the in vivo observations using PS-OCT imaging (Fig. 3). The packing density of collagen fibers varied with increasing distance from the scleral canal. At the border to the optic nerve, fiber bundle packing was rather sparse, was increased at larger distance, and was decreased again at the outer rim of the sclera. The corresponding PS-OCT retardation image (Fig. 6B) showed higher values in the central part of the annulus and lower values at the inner and outer rim. While, as mentioned above, structural changes due to formol fixation may in general influence the birefringent properties, these images are in good agreement with the results published recently by Yamanari et al.^28,34 and suggest that PS-OCT measurements of scleral birefringence can be related to tissue microstructure. 

One limitation of the current PS-OCT imaging design is the limited penetration of 840-nm light into posterior ocular tissues in pigmented animals (Baumann B, et al. IOVS 2014;55:ARVO E-Abstract 2101). As such, PS-OCT imaging of the sclera would be restricted to glaucoma models based on albino rat strains such as Sprague-Dawley or Wistar.^35–40 However, long-wavelength light sources operating in the 1300-nm wavelength regimen enable OCT imaging with deep penetration into the sclera in the posterior rat eye even in pigmented animals, as demonstrated, for instance, by Cimalla et al.⁴¹ By combining such light sources with PS-OCT technology,^42–44 scleral PS-OCT imaging may also be performed in glaucoma models using pigmented rats. 

The recent interest in the role of the sclera was fueled by the development of new insights provided by animal models and the development of computational models focusing in particular on the scleral structure and its impact on glaucoma. Polarization-sensitive OCT, as indicated in this article, might facilitate preclinical glaucoma research (1) by providing in situ access to structural reflectivity images revealing retinal microstructure including the ganglion cell layer (GCL) and the retinal nerve fiber layer (RNFL), (2) by enabling quantitative measurements of scleral birefringence and fiber orientation, and (3) by enabling noninvasive in vivo imaging of the same eye in longitudinal studies. 

Acknowledgments

The authors thank Harald Sattmann, Siegfried Gollubits, Sandra Peiritsch, Alexandra Pernstich, and Roberto Plasenzotti, DVM, at Medical University of Vienna for excellent technical support, as well as Marco Bonesi, PhD, for assistance with image acquisition. They thank Andreas Baumann, MA MSc, at the University of Vienna for contributions to the statistical analysis. 

Presented in part at Photonics West 2013, San Francisco, California, United States, February 2013 (program number 8567-15), at the annual meeting of the Association for Research in Vision and Ophthalmology, Seattle, Washington, United States, May 2013, and at European Conferences of Biomedical Optics 2013, Munich, Germany, May 2013 (program number EW2B.1). 

Supported by the Austrian Science Fund (FWF Grants P19624-B02 and P25823-B24) and the European Union (FP7 HEALTH Program Grant 201880, FUN-OCT). 

Disclosure: B. Baumann, None; S. Rauscher, None; M. Glösmann, None; E. Götzinger, None; M. Pircher, None; S. Fialová, None; M. Gröger, None; C.K. Hitzenberger, None