Preparing Porcine Eyes for Confocal Reflectance Microscopy to Visualize the Vitreous Collagen Fiber Network

Summary

This method visualizes the native structure of collagen fibers in the vitreous body using confocal reflectance microscopy.

Abstract

The vitreous is a clear gel that fills the space between the lens and retina. Its structure is supported by a network of collagen fibers, which likely influences viscoelastic properties and contributes to retinal disease. However, the organization of this collagen network remains poorly characterized because imaging the vitreous in its native state is challenging. Many imaging techniques applied to the vitreous cause dehydration and distortion, and dye-based methods can introduce artifacts or nonspecific labeling, making it difficult to study the collagen network. In this protocol, we outline: 1) a method for dissecting porcine eyes for imaging; 2) a strategy for mounting the eyes on an inverted microscope; 3) guidelines for imaging the vitreous collagen network in its native state using confocal reflectance microscopy; and 4) instructions for quantitative image analysis of network features. We highlight common pitfalls and emphasize key technical points. We report the application of confocal reflectance microscopy to the vitreous. Compared with prior vitreous imaging strategies, this approach preserves the native structure, orientation, and distribution of vitreous collagen fibers without exogenous dyes or dehydration artifacts. Confocal reflectance microscopy can probe the structural basis of vitreous gel stiffness in health and disease.

Introduction

The vitreous is the extracellular matrix between the lens and retina in the eye. Traction from the vitreous on the retina plays a major role in vision loss in diseases such as retinal detachment, macular hole, and vitreomacular traction^1,2,3,4. The ability of the vitreous to transmit tractional forces to the retina is determined by its gel stiffness, which in turn depends on the properties of its major components, collagen and hyaluronic acid⁵. The gel stiffness of the vitreous is likely influenced by features of the network of collagen fibers, since the stiffness of artificial collagen gels is driven by network characteristics such as density, alignment, and fiber width⁶.

The vitreous collagen fiber network is difficult to image due to high water content and low collagen concentration⁷, which leads to artifactual distortion during processing for microscopy or transmission electron microscopy. Traditional histology and transmission electron microscopy require processing to harden the vitreous for thin sectioning. The dehydration required for histological embedding causes vitreous fibers to aggregate into a honeycomb configuration⁸, and ice-crystal formation during freezing increases pore size in the extracellular matrix⁹. Plastic embedding⁸ and rapid freezing¹⁰ have been applied to the vitreous to minimize artifact, although some residual artifact likely remains. Phase contrast microscopy is a method that does not require sectioning and has been applied to the vitreous, but it requires removing a small piece of vitreous from the eye, thereby losing the original position and orientation of the network¹¹. More recently, quantitative polarized light microscopy has been used to map the average orientation of fibers across a grossly sectioned porcine eye, but individual fibers could not be resolved¹². Additionally, dye-based labeling approaches (e.g., trypan blue, picrosirius red, 5-(4,6-Dichlorotriazinyl) Aminofluorescein) can introduce artifacts or nonspecific labeling and were therefore avoided to preserve the native collagen network.

We adopted the sample-preparation method used by Filas et al. for polarized light microscopy study, i.e., gross dissection in an agarose block¹², and imaged the specimen using confocal reflectance microscopy, which has not previously been applied to the vitreous. Using this method, we have acquired images of the vitreous collagen fiber network in a hydrated, near-physiological state. This offers several advantages over prior methods, including reduced artifacts compared with freeze-fracture or dehydration-based approaches^10,13. This protocol requires an inverted confocal microscope with an immersion objective of at least 30x magnification, a numerical aperture greater than 0.9, and a working distance greater than 200 µm. We include key guidance for sample preparation and image acquisition to facilitate clear resolution of fibers. By maintaining a near-native state of the vitreous collagen matrix, this approach enables researchers to quantitatively analyze fiber orientation, density, and spatial distribution-parameters that are likely crucial to understanding vitreous stiffness and its pathological changes. This technique is particularly suitable for researchers interested in the vitreous collagen network, its structure, and the pathogenesis of vitreoretinal diseases.

Protocol

1. Tissue preparation and dissection

NOTE: This research was performed in compliance with institutional guidelines. At our institution, approval was not required from the Institutional Animal Care and Use Committee to perform experiments on ex vivo porcine eyes.

1. Obtain ex vivo porcine eyes from a meat processing plant. Use the eyes within 48 h of death.
   NOTE: Although the vitreous appearance does not change with longer postmortem time, the retina and choroid degrade, leading to dispersion of cellular and pigmented debris during dissection. This debris can interfere with fiber imaging and analysis.
2. Remove any excess muscle and connective tissue from the globe.
3. Place the three-prong pedestal on the base of a cube mold, such as an extra-large silicone ice cube mold (Figure 1A and Figure 2).
4. Place the globe on the three-prong pedestal, positioning it so that limbus rests on the prongs.
5. Make a 3% agarose gel by adding 3 g of general-purpose agarose to 100 mL of distilled water and then heating the solution until it is dissolved. Once the solution has cooled to around 37 °C, pour it into the mold until the eye is submerged in agarose.
   NOTE: This will embed the eye in agarose gel to stabilize the eye for dissection.
6. Chill the agarose cube for at least 4 h.
   NOTE: This increases the adhesion of the gel to the eye and will ensure the eye stays immobile during dissection
7. After chilling, remove the agarose cube from the mold. Then, remove the pedestal from the cube (Figure 1B).
8. Build a three-sided frame in the same dimensions as the agarose cube using a combination of 1 x 8 Lego bricks, plates, and tiles (Figure 3).
   NOTE: This will secure the cube in place and provide a guide for the blade during dissection. The height of the Lego frame will determine the location the eye will be cut. This can be visualized through the semitransparent agarose gel.
9. Place the cooled agarose cube in the Lego frame. Hold a high-profile microtome blade by the blunt edges and align the blade with the top of the Lego frame. Apply gentle downward pressure on the top of the cube and gently drag the microtome blade back and forth through the entire cube and the embedded eye (Figure 1C).
   NOTE: This will create a section of eye tissue that is embedded in the agarose, which will be referred to as the "eye cap".

2. Mounting sample for imaging

1. Using the microtome blade, trim approximately 1 mm of agarose from the entire cut surface of the cube (Figure 4A).
   NOTE: This will leave a small section of eye cap extending above the cut surface of the agarose cube
2. Use a tissue wipe or small triangular cellulose sponge to dry the cut surface of the agarose cube, especially near the border of the eye cap.
3. Add approximately one teaspoon of all-purpose silicone caulk to a small plastic sandwich bag and cut a 3 mm hole in one of the bottom corners of the bag to create a piping bag for the caulk.
4. Using the piping bag, apply a thin bead of caulk to the cut surface of the agarose cube, right along the edge of the eye cap, creating a waterproof border around the sample (Figure 4B). This will keep the vitreous gel from leaking out of the eye cap.
5. Place a large coverslip glass over the eye cap and press down so that the cut surface of the eye cap is in contact with the coverslip without any trapped air or fluid (Figure 4C). Confirm that the silicone bead makes a continuous seal with the coverslip.
   NOTE: If the silicone caulk does not form a continuous seal, vitreous fluid may leak out and allow air into the sample. Exposure to air will cause the vitreous to shrink, resulting in brighter and denser fibers, as well as a liquid area without fibers. If the vitreous is not in contact with the coverslip, it will be difficult to image.
6. While holding the agarose cube, invert the sample and place it on the large coverslip adapter for imaging on an inverted microscope (Figure 4D and Figure 5).

3. Image acquisition

1. Use an inverted confocal microscope that is not a spinning disk type. Use a water objective (with water or ocular lubricant gel as a coupler), or oil objective with at least a 0.90 numerical aperture, a magnification between 30x-60x, and a working distance of at least 200 µm.
2. Adjust the light path settings for the confocal microscope to obtain reflectance images. In place of the dichroic mirror used to redirect exciting light and transmit emitted light for fluorescence imaging, ensure that there is a beam splitting mirror that redirects a proportion of the illuminating laser and transmits a proportion of the reflected light.
   NOTE: The reflectance signal from vitreous collagen fibers is somewhat weak, so background noise can impact the ability to image fibers. An improved signal-to-noise ratio was observed with a wavelength of 473 nm, but we were able to image the vitreous collagen fibers with wavelengths ranging from 405 nm to 635 nm.
3. Set the detection wavelengths to include the wavelength of the illuminating laser.
4. To orient the Z position within the sample, decrease the laser power to the lowest possible setting and identify the flash of reflected light from the coverslip. This will happen at the bottom and the top of the coverslip (Figure 6).
   NOTE: Near the coverslip, there is a considerable reflective signal in the background. Away from the coverslip, the signal from the reflected light from the collagen fibers decreases.
5. Adjust the z position so that it is located above the top of the coverslip and within the sample.
6. Increase the laser power to a level higher than usually used for a fluorescence application.
   NOTE: There is a reflective ring artifact that will be in the center of the imaging field. A combination of zooming in and decentering the imaging field can help exclude this artifact (Figure 7).
7. Use line averaging and increase pixel dwell time in the image acquisition settings, which are standard methods for increasing signal-to-noise ratio in confocal microscopes.

4. Image analysis of fiber width using CT-FIRE ¹⁴

1. Install CT Fire following the instructions on this website: https://github.com/uw-loci/curvelets/wiki/Downloads-and-Installation-(CTF).
   NOTE: The versions used here were "CT-FIRE V3.0 BetaWindows64-2018b" and the MATLAB runtime, "MCR-2018b-win64". Both must be downloaded for the software to work.
2. Convert the images to 8-bit tif and open the 8-bit tif images in CT-FIRE.
3. Make sure the software is set to CT-Fire in the top right corner, and then select Run.
4. CT-FIRE will automatically upload the results of the analysis to a folder named "ctfireout" located in the same folder as the original image. Examine the quality of extraction by comparing the file starting with the name "OL_ctFIRE" to the original image. To follow this protocol, use the following default settings for image segmentation (i.e., identifying fibers): thresh_im2: 5; s_xlinkbox: 8; thresh_ext: 70; thresh_dang_L: 15; thresh_short_L: 15; s_fiberdir: 4; thresh_linkd: 15; thresh_linka: -150; thresh_flen: 15
   1. If fibers are not accurately identified, adjust the parameters by clicking the Update button under the parameters section in the top right corner.
      NOTE: Guidance on how to adjust the parameters can be found on page 20 in section "3 Notes" of the CT-FIRE manual, which is located in the CTFIRE application folder. Additional tips can be found in the article by Liu et al. in section 3.5.2¹⁵.
5. Adjust output settings if desired to limit the fibers that appear on the output images. To follow this protocol, use the following default settings: Minimum fiber length [pixels]: 30; Image Res. [dpi]: 300; Fiber line width [0-2]: 0.5; Max fiber width [pixels]: 15
6. To convert the units of width from pixel to µm, open the original image in Fiji. Use the dimensions of the image appearing in both pixels and µm in the upper left corner of the display window to calculate a ratio.

5. Image analysis of fiber density, curvature, alignment, lacunarity, hyphal growth unit, and fractal dimension using TWOMBLI¹⁶

1. Download TWOMBLI from https://github.com/wershofe/TWOMBLI.
2. Follow the instructions in the file: TwombliDocumentation under "Installation and dependencies".
3. Open the TWOMBLI_v1/programs folder and replace AnamorfProperties.xml file with AnamorfPropertiesEH.xml file (Supplemental File 1).
4. In the TWOMBLI_v1/programs folder, replace Twombli_v1.ijm file with Twombli_v2.ijm file (Supplemental File 2).
5. Create a folder with empty folders within it called "test HDM," "test masks," "HDM," and "masks".
6. Convert the images to 8-bit tif.
7. Put 3-5 images in a "test set input" folder.
8. Put all the images to be analyzed in a single folder without subfolders called "eligible".
9. In Fiji, click Plugins | Macros | Run | choose Twombli_v2.ijm in twombli/Programs.
10. Follow the prompts to run TWOMBLI. To follow this protocol, use the following parameters: Contrast Saturation, 0.35; Min Line Width, 3; Max Line Width, 3; Min Curvature Window, 40; Max Curvature Window, 40; Minimum Branch Length, 10; Maximum Display HDM, 100; Minimum Gap Diameter, 0
    NOTE: The parameters for the images may need to be adjusted. The software will guide the user through choosing the parameters, and additional instructions can be found in the file TwombliDocumentation.docx, which came with the program.

Results

One eye from each of three young adult pigs (a 4-month-old female, a 5-month-old male, and a 6-month-old female) was imaged after sectioning the eye in an axial plane through the inferior limbus. Six imaging locations were acquired in the X-Y sectioning plane, spaced 4 mm apart. At each imaging location, a contiguous 4 x 4 grid of 70 µm x 70 µm images was obtained (Figure 8). For each pig, each image was analyzed separately for several parameters (Table 1).

Discussion

Prior vitreous imaging methods showed artifactual fiber aggregation¹³, loss of fiber organization, orientation, and location¹⁷, or relied on much higher resolution with a smaller field of view, limiting insights into fiber distribution¹⁰. Confocal reflectance microscopy of ex vivo, grossly sectioned fresh tissue provides micron-scale images of vitreous collagen fibers while preserving natural fiber organization, orientation, and location. In our hand...

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
Agarose I	VWR	0710-500G
All-purpose silicone 1 caulk, clear	GE	2795576
Cellulose sponge (Weck-Cel Eye Spear)	Beaver -Visitec	000865
Custom 3 prong pedestal	NA	NA
Custom adapter for mounting large coverslip glass	NA	NA
Large coverslip glass, 1.42” x  2.36” no.1	Ted Pella	260460-100
Lego (classic baseplate, 1 x 8 tiles, 1 x 8 bricks, 1 x 6 plate)	Lego	11023, 3008, 3666, 4162
Microtome blades (high profile)	Lieca Biosystems	14035838926
Olympus UPlanSApo 30x silicone objective	Olympus	UPLSAPO30XS
Silicone large ice cube tray (Samuelworld)	Amazon	713243280060
Tissue wipes (Kimwipes)	Kimberly Clark	34120