Three-dimensional arrangement of elastic fibers in the human corneal stroma.

Tissue specimens

Five human corneas were obtained from the CTS Eye Bank, Bristol, UK. Cornea 1 was from a 69-year–old female. The whole enucleated eye was fixed in 4% paraformaldehyde. The cornea was removed with a scleral rim and placed in modified Karnovsky's fixative (2.5% glutaraldehyde and 2% paraformaldehyde in 0.1M cacodylate buffer at pH 7.2) for 30 min, dissected and processed for serial block face scanning electron microscopy (SBF SEM) and TEM. Cornea 2 was from a 50-year-old male and was received from the Eye Bank in Eagle's minimum essential medium as a cornea with about 2 mm of the adjacent sclera. It was de-swelled with 8% dextran overnight, mounted in a Barron artificial anterior chamber to maintain a trans-corneal pressure and was fixed using the same modified Karnovsky's fixative for 3 h. It was then dissected and used for SBF SEM and TEM. Cornea 3 was from a 79-year-old male. It was collected and fixed in modified Karnovsky's fixative within two days of death and was processed for TEM using the orcein method (see below). Cornea 4 was from a 72-year old female. It was fixed in 4% paraformaldehyde then wax embedded for histology. Cornea 5 from a 67-year-old male was processed for non-linear microscopy as described below. Cornea 6 was from a 13-week old foetus obtained from the Human Developmental Biology Resource (HDBR). The whole globe was fixed in modified Karnovsky's fixative and was then processed for SBF SEM as below.

Institutional Ethics

Committee approval was obtained for this study and the research followed the Tenets of the Declaration of Helsinki.

Serial block face scanning electron microscopy Human

Corneas were fixed for 3 h in 2.5% glutaraldehyde/2% paraformaldehyde in 100 mM sodium cacodylate buffer pH 7.2 at room temperature (RT). The cornea was cut into thin segments and post fixed with 1% osmium tetroxide for 1 h. After a brief wash with fresh buffer followed by a 20 min wash in distilled water the samples were incubated in 0.5% low molecular weight (di-Gallic) C 14 H 10 O 9 tannic acid (mw 322.22) (AR Mallinckrodt, Dublin, Ireland) in distilled water for 2 h. The samples were then washed in distilled water for 30 min and placed in 1% aqueous uranyl acetate for 1 h in the dark at RT. Corneal samples were then dehydrated in an ethanol series from 70% to 100% ethanol for 1 h after which they were placed in 2% uranyl acetate in 100% ethanol for 2 h in the dark at RT. After washing in 100% ethanol for 40 min, the corneal samples were placed in a 1:1 mixture of 100% ethanol: 100% acetone for 20 min. The samples were then incubated in a saturated solution of lead acetate in a mixture of 1:1 100% ethanol 100% acetone for 2 h ( Kushida, 1966 ). After staining, the corneal samples were washed in two changes of 1:1 mixture of 100% ethanol and 100% acetone for 15 min and placed in 100% solution of acetone for 20 min. After washing the samples 3 times for 20 min in 100% acetone, the samples were finally embedded in CY212 (TAAB) epoxy resin and polymerised for 24 h at 60° C. The surfaces of polymerised resin blocks were then trimmed and attached to Gatan (PEP6590) specimen pins. The pins were then gold coated and transferred to a Zeiss Sigma VP FEG SEM equipped with a Gatan 3View2 system, where data sets of up to 1000 images were acquired of the block surface every 50 nm through automated sectioning. Each image was acquired at 4K × 4K pixels, at a pixel resolution of 4 nm and a pixel dwell time of 8 μs, using an accelerating voltage of 3.4 keV in low vacuum variable pressure mode (28 Pa). Imaging data was acquired from a 16.19 μm × 16.19 μm region of interest. Selected serial image sequences were extracted from the image data and 3D reconstructions were generated with Amira 6.0 software (FEI, Mérignac, France) using both manual hand tracing for the larger structures and automated thresholding for the fibers. For the full thickness quantification analysis, the peripheral cornea was imaged every 50 nm en face at 5000 × magnification, starting from the epithelium and ending at Descemet's membrane, obtaining over 21,000 images with resolutions of 2000 × 2000 pixels. The data were split into sets of ∼1000 images, fibers were segmented using a mixture of manual and automatic thresholding, resulting in a fiber voxel count that was subsequently converted to percentage using the total volume voxel count. Cells were blocked out to background level in order to remove their contribution to the images. The more detailed quantification above Descemet's membrane was carried out by measuring fiber volume every 200 images. Both were done using Amira 6.0 software with XImagePAQ extension. When sectioning the 16 μm × 16 μm block face through a thickness of almost 1 mm, it was very difficult to ensure that the en face sections were exactly parallel to the surface of the cornea. When viewing our series of images from different depths, it was evident that the blocks were being cut at a slight angle to the surface of the cornea. To correct for this, the number of 50 nm sections between where Descemet's membrane was just visible down to the depth where the stroma was just not visible (which ideally would be zero if the block was being cut parallel to the corneal surface) was used to make a geometrical calculation of the offset angle and hence correct the section thickness and the depth measurements to allow for this when plotting the % fiber volume as a function of depth. This led to an uncertainty of ±5 μm in each depth measurement.

Transmission electron microscopy

Embedded samples from Corneas 1 and 2 that had been en bloc stained for SBF SEM, were also used for TEM. A Leica UC6 ultra-microtome was used to cut 90 nm-thick gold sections that were floated on distilled water before being mounted on copper grids. Additionally, separate sections were cut from Cornea 3 and stained with the elastic fiber-specific stain 0.2% orcein in acid alcohol (1% HCl in 70% EtOH), then rinsed with 70% EtOH, air-dried, and counterstained with uranyl acetate and lead citrate. Both normal and orcein stained ultra-thin sections were visualised using a Jeol 1010 TEM (JEOL, Tokyo, Japan).

Non-linear microscopy

Corneal buttons 8 mm in diameter were dissected from the centre of Cornea 5, placed into dialysis tubing (molecular cut-off 14 kDa), and brought to physiological hydration by immersion in 2.5% polyethylene glycol (PEG) overnight. The buttons were then mounted on Superfrost glass slides, in 1:1 phosphate buffered saline (PBS) – glycerol solution, protected by a 0.16 mm thick glass coverslip. A modified confocal microscope (Olympus FluoView IX71 and F300) was used to obtain TPF and SHG images (see Bell et al. (2014) for a complete description of the system configuration). TPF and SHG images from the same area of cornea were overlaid using ImageJ software ( Schneider et al., 2012 ), resulting in a composite image.

Histology

Paraffin wax sections were dewaxed in xylene and rehydrated in a descending industrial methylated spirit (IMS) gradient to water. Tissue sections were incubated in Miller's elastic stain at 60 °C for 1 h, destained in 95% IMS and washed in water before 10 min of nuclear staining in Mayer's haematoxylin, followed by counterstaining of collagen fibrils with Van Gieson's stain for 5 min. Sections were then washed in water, dehydrated through an ascending IMS gradient, cleared in xylene and mounted under coverslips with DPX mountant. Elastic fibers and collagen appeared purple/black and pinkish/red, respectively.