Live-Cell Imaging of Intact Ex Vivo Globes Using a Novel 3D Printed Holder

Summary

The present work describes a novel experimental protocol that utilizes a 3D printed holder to enable high-resolution live cell imaging of enucleated globes. Through this protocol, the cellular calcium signaling activity in wounded corneal epithelium from ex vivo globes can be observed in real time.

Abstract

Corneal epithelial wound healing is a migratory process initiated by the activation of purinergic receptors expressed on epithelial cells. This activation results in calcium mobilization events that propagate from cell to cell, which are essential for initiating cellular motility into the wound bed, promoting efficient wound healing. The Trinkaus-Randall lab has developed a methodology for imaging the corneal wound healing response in ex vivo murine globes in real time. This approach involves enucleating an intact globe from a mouse that has been euthanized per established protocols and immediately incubating the globe with a calcium indicator dye. A counterstain that stains other features of the cell can be applied at this stage to assist with imaging and show cellular landmarks. The protocol worked well with several different live cell dyes used for counterstaining, including SiR actin to stain actin and deep red plasma membrane stain to stain the cell membrane. To examine the response to a wound, the corneal epithelium is injured using a 25 G needle, and the globes are placed in a 3D printed holder. The dimensions of the 3D printed holder are calibrated to ensure immobilization of the globe throughout the duration of the experiment and can be modified to accommodate eyes of different sizes. Live cell imaging of the wound response is performed continuously at various depths throughout the tissue over time using confocal microscopy. This protocol allows us to generate high-resolution, publication-quality images using a 20x air objective on a confocal microscope. Other objectives can also be used for this protocol. It represents a significant improvement in the quality of live cell imaging in ex vivo murine globes and permits the identification of nerves and epithelium.

Introduction

Cornea
The cornea is a clear, avascular structure covering the anterior surface of the eye that refracts light to enable vision and protects the interior of the eye from damage. As the cornea is exposed to the environment, it is susceptible to damage from both mechanical causes (scratching) and from infection. A corneal injury in an otherwise healthy patient typically heals within 1-3 days. However, in patients with underlying conditions including limbal stem cell deficiency and type II diabetes, the corneal wound healing process can be greatly prolonged¹. As the cornea is highly innervated, these non-healing corneal ulcers and recurrent corneal erosions are very painful and greatly diminish the quality of life of patients experiencing them¹.

Cell signaling
When an otherwise healthy cornea is injured, calcium signaling events in the cells adjacent to the wound precede and prompt cellular migration into the wound bed, where they close the injury without the risk of scarring^2,3. These signaling events have been well-characterized in corneal epithelial cell culture models using live cell imaging². Preliminary experiments demonstrate significantly more calcium signaling after injury in non-diabetic cells compared to diabetic cells. However, characterization of the cell signaling events in ex vivo globes has proved to be a technical challenge.

Live cell imaging
Previous studies have successfully recorded calcium signaling events from in vitro cell culture models of corneal wound healing^4,5,6. Developing a methodology to produce high-quality images of these signaling events in ex vivo tissue is of great interest because it would permit the study of these events in a more complex and true-to-life system. Previous approaches have involved dissection of the cornea followed by immobilization in a UV-induced PEG gel^7,8,9. Immobilization is an essential yet challenging step when working with live tissue, as it must remain viable and hydrated throughout the course of the experiment. Furthermore, immobilization must not damage the tissue. While the PEG solution immobilized the tissue, the resolution and quality of the images produced were not consistent. Therefore, 3D printed holders were developed to immobilize intact globes to produce higher-quality images with less risk of tissue damage.

The approach
A unique 3D printed holder was developed to immobilize intact ex vivo globes for live cell imaging. This holder prevents damage from two major sources: it allows for imaging of an enucleated globe without the need to dissect the cornea, and it eliminates exposure to UV light. Without these sources of damage, the images obtained more accurately represented the response to the scratch injuries made experimentally. Furthermore, the 3D printed holder was calibrated to the precise dimensions of the murine eye. This provided a much better fit than immobilization in PEG solution, leading to a higher-quality image at lower-powered objectives due to decreased tissue movement. A cover bar attached to the top of the holder ensures that the globe remains immobile throughout the duration of the experiment and that there is no displacement of the globe when growth media is applied to maintain hydration and viability. The ability to print the holder to precise dimensions also allows us to generate an optimal fit for murine eyes of different sizes due to the age or disease status. This technology can be applied more broadly to develop holders for the eyes of different species based on their dimensions.

Protocol

The procedures involving animal subjects were approved by the Association for Research in Vision and Ophthalmology for the Use of Animals in Ophthalmic Care and Vision Research and the Boston University IACUC protocol (201800302).

1. Designing the 3D printed holders and cover bar

1. Design the 3D printed holders and cover bar accounting for the average diameter of mouse globes and 3D print the design (Figure 1A, B).
2. Keep the diameter of the inner wall of the holder slightly larger than the average diameter of the globe to account for the different sizes of individual mice. Keep the height of the holder at around half of the globe diameter, ensuring a tight fit of the globe when secured by the 3D printed cover bar.
3. Size the holder cover bar to the length of the holder's outer diameter with a width that is 1/4 to 1/2 of the holder diameter. The cover bar is sized to allow for access to the globe when secured in the holder for hydration and the removal of the eye at the conclusion of the experiment.
4. Print the holder and cover bar.

2. Sample collection

1. Euthanize mice (male C57BL/6 mice aged 9-12 weeks and 27 weeks old were used for this study) using established protocols in compliance with institutional guidelines. For this protocol, perform euthanasia with carbon dioxide followed by decapitation.
2. Remove the mouse head and place it immediately on ice to preserve the viability of the tissue. Enucleate the globes using dissection tools while preventing tissue damage.
3. Proptose the globe using tweezers. Clip the optic nerve using dissection scissors just below where it is held by the tweezers.
   NOTE: For further precautions, perform the following steps in a laminar flow hood.
4. Incubate the globes in 2 mL of medium in a p35 cell culture dish including a calcium indicator and/or cell membrane stain for 1 h in a 37 °C, 5% CO₂ incubator with low light conditions. Ensure the globes are submerged in the staining medium for uniform staining.
   1. For the experiments performed here, use the calcium indicator, Fluo4-AM (1:100)², and cell membrane counter stain, deep red plasma membrane stain (1:10,000)², with a final concentration of 1% (v/v) DMSO and 0.1% (w/v) pluronic acid in 2 mL of keratinocyte serum-free medium (KSFM) with the following growth supplements: 25 µg/mL bovine pituitary extract, 0.02 nM epidermal growth factor, 0.3 mM CaCl₂, and penicillin-streptomycin (100 units/mL and 100 µg/mL, respectively).
      ​NOTE: The incubation conditions and times are variable depending on the calcium indicator, tissue type, and sample volume. When using pluronic acid, caution is advised as it renders tissue permeable. This protocol calls for 10% pluronic acid. Lower concentrations of pluronic acid have been determined experimentally to be ineffective, and higher concentrations risk damage to the tissue.

3. Preparation of sample holders

1. Adhere the holders to a clean glass-bottom coverslip with glue that has not been used previously. The glue used in this protocol comes from single-use containers to ensure sterility and a new, unopened container is used every time.
2. Wash the holder in 70% ethanol. Place glue onto the holder and adhere the holder to the glass bottom coverslip. Ensure no glue is within the inner area of the holder as glue can fluoresce, complicating imaging.
3. Wait until the glue solidifies. Confirm the holder is secure against the coverslip.
   NOTE: P35 cell plates with glass-bottom coverslips were used for the experiments presented in this manuscript. Other glass-bottom slides and/or plates can be substituted based on the needs of the experiment.

4. Wounding of the eye globes

1. Remove the globes from the staining solution using sterile eye droppers, taking care to prevent tissue damage to the region of interest. Wash the globes for 5 min at room temperature using sterile phosphate-buffered saline to remove excess stain, and place the globes in the medium for transport to the microscope.
2. Wound the globes using a sterile 25 G needle in the region of interest.
   1. Using a sterile eye dropper, pick up and hold the globe from the back of the eye. This will keep the globe stable, prevent it from rolling, and allow consistent wounds to be made. Using this setup, the optic nerve will be inside the eye dropper nozzle, and the cornea will be facing outward.
   2. For a scratch wound, gently move a sterile 25 G needle across the exposed cornea. For a puncture wound, gently press the needle directly into the central cornea. Ensure the wound does not puncture the cornea.
      ​NOTE: Skip this step if a wound response or wounded environment is not required for the experiment. Previous studies have shown that both scratch wounds and puncture wounds to murine corneas made using this method are consistent in both diameter and depth¹⁰. Confirmation of the wound dimensions between independent globes was performed using a region of interest analysis.

5. Sample placement on the holder

1. Place the cornea or limbal region onto the coverslip in the inner area of the holder and stabilize using the 3D printed cover (Figure 1C-H).
2. Confirm that the globe is positioned correctly and that the site of interest is in contact with the glass coverslip. Once the globe has been placed into the holder, do not try to remove the globe as this may cause tissue damage.
3. Adhere the 3D printed cover to the holder using glue, ensuring stabilization. Ensure the cover bar adheres to the holder and not the globe.
   ​NOTE: The area to be imaged is placed down because the protocol is written for use on an inverted microscope. The protocol can be adapted for upright microscopes using holders with a smaller inner radius and the removal of the cover bar. This will result in less globe stabilization.

6. Sample imaging

1. Turn on the microscope and environmental chamber and verify that the chamber is humidified. Set the environmental chamber to 35 °C and 5% CO₂ for the duration of the experiment.
   NOTE: Microscopes with environmental chambers are preferable for this procedure to prevent dehydration and to keep the globe at optimal temperatures but are not required.
2. Place the coverslip, holder, and stabilized globe on the microscope stage within the environmental chamber and image using live cell imaging techniques⁹.
3. Pipette additional growth media onto the coverslip to prevent dehydration and maintain tissue viability. Ensure there is enough medium in the well to cover the globe in the holder. Depending on the duration of imaging, add fresh medium when needed throughout the experiment.
4. Begin experiments using live cell imaging techniques and protocols. Use low power laser settings to preserve the tissue and prevent tissue damage in long-duration experiments. Use appropriate objectives for long working distances. The experiments in this manuscript were performed using a 20x objective.
   NOTE: The laser power and gain, experimental duration, location, and plane of imaging are all variables depending on the experimental parameters. Imaging experiments on intact globes ranging from 1 h to 4 h in duration have been performed successfully in past publications¹⁰.
5. Record and save data in the preferred file format. The software used by the microscope produces .czi files for data recording.
6. Dispose of the globes as per the standard institutional protocols at the end of the protocol.

Representative Results

This protocol has been used to consistently produce publication-quality data and images¹⁰. The images obtained represent a significant improvement when compared to previous approaches (Figure 2). Using the 3D printed holder, images can be captured throughout the layers of the cornea, and calcium mobilization in different z-planes can be observed (Figure 3). This approach has been used to compare cell-cell signaling between apical and basal cell layers at a wound in young and old mice¹⁰.

The improved stability has also allowed for imaging of globes for significantly longer than previously possible. This has allowed studies of cellular migration into the wound bed to be extended by several hours beyond previous capabilities. With this extended timeline, substantial differences in cellular behavior can be observed during the wound healing response between young and old mice after corneal injury¹⁰. For this protocol, data were collected every 5 min for 4 h. When using the holder, no significant drift was observed in the x, y, or z planes throughout the course of the experiment (Figure 4, Video 1).

An advantage of this protocol is that it allows different regions of the cornea to be imaged based on the placement of the globe in the holder. In a recent publication, this feature was taken advantage of to gather images both at the central cornea and in the corneal-limbal region (Figure 5)¹⁰. An injury to the central cornea was found to produce calcium signaling events in cells adjacent to the nerves in the limbal region¹⁰. The versatility of the holder goes beyond stabilizing the globe for imaging experiments and has been used for several different applications. By placing the cornea face-up in the holder, nanoindentation experiments were performed to measure the stiffness of the corneal epithelium, basement membrane, and stroma in young and old mice^10,11.

Figure 1: Schematics and setup of the 3D printed holder. (A) Design of the 3D printed holder with annotated height and width dimensions. (B) Representative CAD file image from the 3D printer software. (C) Representative image of the holder with the attached cover containing a murine globe. (D) Sterile, single-use glue is applied to the bottom of the 3D printed holder. (E) The 3D printed holder is adhered to a glass-bottom p35 cell culture dish. (F) An enucleated globe is placed cornea-down into the holder using a sterile eye dropper. (G) The cover bar is adhered to the top of the 3D printed holder to ensure globe immobilization. (H) The glass-bottom plate with an adhered holder and globe is placed on the microscope stage. Please click here to view a larger version of this figure.

Figure 2: Using specialized 3D printed holders to stabilize the globe and yield higher-quality imaging data of multi-layered structures. (A) and (B) represent typical live imaging data of a wounded ex vivo cornea stabilized without and with a 3D printed holder, respectively. Calcium signaling events (green) and cell counterstain (deep red plasma membrane stain) can clearly be identified in the apical, basal, and stromal layers of the cornea in (B) but not in (A). Please click here to view a larger version of this figure.

Figure 3: Z-stack of a cornea immobilized in the 3D printed holder. Representative images of a z-stack taken through the layers of the cornea at a scratch wound (denoted with a white asterisk). The sample is stained with deep red plasma membrane stain to visualize the cell membranes and Fluo4-AM (green) to visualize calcium signaling. (A) The apical cell layer, (B) apical and basal cells, (C) basal cell layer, and (D) stroma can be seen in the wound bed. Please click here to view a larger version of this figure.

Figure 4: Representative 4 h time series experiment of a cornea revealing little movement of the globe in the x, y, or z directions. The image is of a scratch wound injury (denoted with a white asterisk) at the central cornea of an ex vivo murine globe. The globe is stained with deep red plasma membrane stain to visualize the cell membranes and Fluo4-AM (green) to visualize calcium signaling events. The globe is immobilized using a 3D printed holder. The images were taken 1 h apart, beginning 5 min after injury. Little drift in the x, y, or z directions was observed using this imaging setup throughout the course of the experiment. Please click here to view a larger version of this figure.

Figure 5: Diagram of imaging locations on an intact globe. The 3D printed holder permits imaging of an intact ex vivo globe at various locations. The holder was used to collect images from both the central and limbal regions of the cornea. Globes are stained with deep red plasma membrane stain to detect the cell membranes and Fluo4-AM (green) to detect calcium signaling. (A) Representative image of the central cornea at a scratch wound. (B) Representative image of the limbal region of the cornea after an injury to the central cornea. Please click here to view a larger version of this figure.

Video 1: Movie of a globe immobilized in a 3D printed holder for 4 h. A globe was immobilized with the 3D printed holder and stained with deep red plasma membrane stain (red) to visualize the cell membranes and Fluo4-AM (green) to visualize calcium signaling events. An image was taken every 5 min for 4 h, beginning 5 min after injury. Little drift in the x, y, or z directions was observed using this imaging setup throughout the course of the experiment. The playback speed is 30 frames per second. Please click here to download this Video.

Discussion

This protocol describes a live cell imaging technique that uses a 3D printed holder to stabilize and immobilize intact animal eyes. It is designed to circumvent several significant disadvantages recognized with previous live cell imaging protocols of ex vivo corneal tissue. This protocol offers many advantages for the live cell imaging of intact globes. It significantly reduces unnecessary tissue damage that could interfere with the wound healing response to experimentally induced scratch wounds. This includes damage to the nerves and epithelium from dissection and exposure to UV light. Furthermore, this protocol facilitates tissue hydration and viability by immobilizing the tissue in a way that allows growth medium to be applied periodically without disruption of the experiments. By changing the orientation of the eye in the holder, the protocol allows for the imaging of different regions on the globe. The use of confocal microscopy along with this protocol permits the imaging of different planes of the globe, allowing the observation of interactions between tissue structures. The protocol is highly versatile and can be adapted to globes of various sizes from mice and other species. The holders and covers can be easily removed from the imaging wells, sterilized, and reused. The 3D printing protocol is efficient and time-effective, allowing for the convenient creation of many holders at one time. The globes may be fixed within the holders if the samples need to be kept and imaged again at a later date. Fixed tissue retains deep red plasma membrane stain and Fluo4-AM stains for several weeks after fixation, which can be used as markers when staining for a specific protein(s).

Previous protocols called for dissection of the eye and the subsequent use of a UV-activated PEG gel to immobilize the cornea for imaging^7,8,9. Damage to the tissue incurred through these techniques could confound experimental results. This is particularly important to consider when studying the cellular and tissue processes involved in wound healing, as this additional damage beyond the experimentally applied wounds may alter the wound healing response^4,12. The input of sensory corneal nerves will be affected by the previous protocol, as the dissection process severs the nerves from their cell bodies in the trigeminal ganglion¹³. Furthermore, the dissection of the cornea would itself causes an injury response, as soluble factors released from an injury are responsible for the wound healing response⁴. This new procedure addresses these weaknesses by imaging the globes intact. Using this methodology, damage to the eye is limited to cutting of the optic nerve upon enucleation and maintains cell viability within the cornea for extended periods of time. Together, these improvements create a better simulation of the wound response.

Immobilization is an essential yet challenging step when working with live tissue, as it must remain viable and hydrated throughout the course of the experiment. While the corneas and globes could be secured with UV-induced PEG gel, the procedure requires UV irradiation after placing the tissue in the holder⁸. It was observed that the irradiation required to polymerize the PEG decreased cellular responsiveness. Furthermore, the PEG decreased the resolution of the images, and the AiryScan function was required to observe cell signaling. In this new protocol, the dimensions of the holder are optimized to precisely fit the globes, and the cover bar provides an additional layer of immobilization. The holder and the cover bar eliminate the need to use PEG, which produces higher-quality, higher-resolution images with real-time footage of the wound healing process. One solution to the above-mentioned issues would be to forego live cell imaging entirely and fix the globes in paraformaldehyde (4%) at predetermined times after injury. However, with fixed tissue, ongoing cellular processes such as calcium signaling and the effects of these processes on physical changes in the tissue cannot be observed. For groups interested in recording and quantifying communication events between cells in the corneal epithelium, the limits imposed by tissue fixation make this technique impractical for such purposes.

This new protocol has two key steps: (1) euthanasia and immediate globe enucleation, and (2) positioning and orientation of the globe within the holder. The cutting of the optic nerve and enucleation of the globe must be performed carefully by proptosing the eye without placing a brace to preserve the internal and external structures. Positioning and orientation are necessary for the imaging of the sites of interest and ensuring the largest possible field of view is in focus with adequate detail. Correct placement of the globe is necessary prior to applying the cover bar to the holder.

Several variables of the protocol can be adjusted depending on the requirements of the experimental parameters. The holders can be printed in different sizes to accommodate globes of different volumes. The parameters regarding holder diameter and height relative to the size of the globe remain the same across size ranges. If the user experiences drift in the z-plane, the potential reason might be that the tissue is not immobilized, and the holder should be re-manufactured. The staining of the intact globes can be adjusted depending on the experimental parameters, the stains used, and the optimal environment and concentrations for the tissue of interest. The experimental duration can be adjusted as long as tissue viability is maintained. Short-term experiments involving 1 h of constant imaging of the tissue and long-term experiments involving imaging of the tissue at regular intervals over the course of 4 h have both been successfully performed. These experiments were performed on an inverted microscope, and this protocol was designed for optimal use with a similar imaging setup. Once the globe is stabilized and positioned within the holder, removal of the globe can result in tissue damage; thus, optimal positioning and orientation of the globe before attaching the cover bar are essential. The globes can be fixed in 4% paraformaldehyde while still within the holders for further analysis of protein localization. Preliminary data have shown that deep red plasma membrane stain and Fluo4-AM staining are both retained through the fixation process.

Although this protocol has many advantages over previous protocols, there are a few limitations to this experimental approach. One disadvantage is that it is difficult to remove the globe from the holder without damaging the corneal epithelium. Thus, the globe must be fixed in the holder for re-imaging using protein or RNA localization methods at a later date. This can limit follow-up studies. Another disadvantage of this protocol during prolonged imaging protocols is that hydration must be maintained. The need to manually reapply the media at regular intervals requires somebody to actively monitor the imaging process for the extent of the experiment. This may prove logistically challenging for extended imaging protocols.

Materials

Name	Company	Catalog Number	Comments
1.75 blue polylactic acid (PLA) plastic	Creality (Shenzen, China)	N/A	Material for holder
35 mm Dish, No. 1.5 Coverslip, 14 mm glass diameter, Poly-D-Lysine Coated	MatTek Corporation (Ashland, MA)	P35GC-1.5-14-C	Well for imaging.
Autodesk Fusion 360 software	Autodesk (San Rafael, CA).	N/A	Software used for printing the holders.
BD 25 G 7/8 sterile needles single use 100 needles/box	Thermo Fisher Scientific (Waltham, MA)	305124	For experimentally-induced wounds to the globes
CellMask DeepRed	Invitrogen (Carlsbad, CA)	C10046	Cell membrane counterstain. Calcium indicator. 1:10,000 concentration with a final concentration of 1%(v/v) DMSO and 0.1% (w/v) pluronic acid
Complete Home Super Glue	Walgreens (Deerfield, IL)	N/A	For attaching the holder to the imaging well
Ender 3 Pro 3D printer	Creality (Shenzen, China)	N/A	For printing the holder
FIJI/ImageJ	ImageJ (Bethesda, MD)	License Number: GPL2	Softwareused for confirming consistency of wound depth and diameter between independent globes using Region of Interest analysis
Fluo-4	Invitrogen (Carlsbad, CA)	F14201	Calcium indicator. 1:100 concentration with a final concentration of 1%(v/v) DMSO and 0.1% (w/v) pluronic acid
Keratinocyte Serum-Free Medium	Gibco (Waltham, MA)	17005042	25 mg/mL bovine pituitary extract, 0.02 nM EGF, 0.3 mM CaCl2, and penicillin-streptomycin (100 units/mL, 100 µg/mL, respectively) added to medium
Phophate-Buffered Saline (PBS)	Corning, Medlabtech (Manassas, VA)	21-040-CV	Used to wash excess stain off of corneas before imaging
Zeiss Confocal 880 Microscope with AiryScan	Zeiss (Thornwood, NY)	N/A	20x magnification objective was used