Characterization of a Novel Human Organotypic Retinal Culture Technique

The retina is a highly specialized ocular structure responsible for transforming incoming light energy to electric signals, which are then processed by the brain for visual perception. The human retina contains a dynamic range of cell types, highly organized in a unique lamellar structure consisting of two synaptic and three nuclei layers¹ (Figure 1). Retinal homeostasis is sustained by the intricate connections between neuroretinal cells, blood vessels, nerves, connective tissues and the RPE¹. Due to the sophisticated retinal anatomy and physiology, mechanisms of many retinal diseases still remain poorly understood^2,3,4,5. To better study retinal diseases, HORC models have been developed^6,7,8,9. Compared to animal studies and in vitro cultures, HORC models are advantageous because they retain the dynamic cellular environment and complex neurovascular interactions in situ, providing a good model for clinical translation.

Figure 1: Posterior ocular structures of the human eye. Anterior to posterior, the retinal layers are: nerve fiber layer (NFL), ganglion cell layer (GCL), inner plexiform layer (IPL), inner nuclear layer (INL), outer plexiform layer (OPL), outer nuclear layer (ONL), photoreceptor inner segment (IS), and photoreceptor outer layer (OS). Cells within the retina include ganglion cells (blue), amacrine cells (yellow), bipolar cells (red), horizontal cells (purple), rod photoreceptors (pink) and cone photoreceptors (green). The vitreous is located anterior to the retina. The RPE, Bruch's membrane, choroid and sclera are located posterior to the retina. Note that the image shown is only a schematic representation of the retina and the ratio of cells/retinal connectivity within each layer may not be indicative of the in vivo setting. Please click here to view a larger version of this figure.

Previously characterized HORC protocols^6,7,8,9 have involved separating the retina from the underlying RPE-choroid and sclera using a surgical trephine. However, without the support provided by these underlying structures, the translucent retina becomes flimsy, difficult to handle and tools such as forceps can easily disrupt its integrity. Furthermore, isolating retina in culture without the RPE has been shown to cause ganglion cell apoptosis and photoreceptor degeneration^10,11,12. Thus, an alternative HORC protocol that minimizes the loss of retinal integrity and better mimics the in vivo environment would be useful. This is particularly important when studying retinal disease mechanisms, as physical injury during explant handling could introduce artifacts. Therefore, the aim of this study was to develop a novel HORC model that includes the RPE-choroid and sclera in order to protect retinal integrity during explant handling and culture.

In order to achieve this aim, retinal explants "sandwiched" between the residual vitreous and the underlying RPE-choroid and sclera were extracted. In the sandwich explants, the vitreous weighs down the retina to prevent retinal detachment and folding, whereas the tough, fibrous sclera acts as both a scaffold for structural support and a contact point for forceps. Moreover, animal models have shown that retaining the RPE in culture can prevent retinal degeneration and glial proliferation, a response of Müller cells to danger signals such as hypoxia and inflammation^10,11,12.

To characterize the model, sandwich retinal explants were stained with Hematoxylin and Eosin (H&E) to assess anatomical structures and immunohistochemistry (IHC) was performed, labeling explants with terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL, an apoptotic cell marker), glial fibrillary acidic protein (GFAP, a retinal inflammation and Müller cell activation marker), and vimentin, a marker of Müller cell integrity. To determine whether this model can be induced to develop molecular disease signs, the explants were exposed to high glucose (HG) with pro-inflammatory cytokines (Cyt), interleukin-1β (IL-1β) and tumor necrosis factor-α (TNF-α), a culturing environment that has been shown to mimic diabetic retinopathy (DR) in both cell and animal disease models^13,14,15. Luminex assays were used in the DR model to measure cytokines released into the culture medium.

Human donor eye cups were obtained from the New Zealand National Eye Bank following corneal excision for transplantation and as approved by the Northern B Health and Disability Ethics Committee (NTX/06/19/CPD/AM07).

NOTE: The culture should be done in a Class II biosafety cabinet to ensure sterile tissue culture conditions. The tissues must be cultured within 24 h post-mortem to avoid significant retinal integrity loss.

Figure 2: Images showing the procedure for collecting the sandwich retinal explants. For explant preparation, use dissecting scissors with one sharp tip and one blunt end, with the blunt end facing the inside of the globe to reduce tissue damage during incision. Also use forceps with blunt ends to avoid scratching the intraocular tissues during handling. Using surgical scissors, with the blunt end facing the inside, cut along the limbus to remove the iris and lens (A-C). The ONH can be located if looking straight into the opened eyecup (D). Using forceps with blunt tips, hold the sclera to stabilize the globe (E). Divide the globe vertically into two halves by making two deep incisions towards the ONH but do not cut through the ONH. Repeat this along the horizontal meridian (G). Apply blunt forceps at the sclera and gently open the globe to a clover shape (H-K). Use forceps to carefully manipulate the vitreous to smoothen the folded retina, as the vitreous tugs onto the retina, but do not touch the retina directly (I). Remove the vitreous if it is causing retinal detachment or folding (not shown in Figure 2 as the transparent vitreous is not well captured on photos). Locate areas where the retina is flat and extract retinal explants using a surgical trephine (L). Applying forceps at the sclera, carefully transfer the retinal explants into the pre-prepared culture medium (M,N). The entire sandwich explant should sink to the bottom of the well due to its weight, therefore each explant is submersed in the medium (O). Please click here to view a larger version of this figure.

1. Extraction of the sandwich retinal explants from the eyecup

1. Remove the iris and lens.
   1. Place the eye cup inside a Petri dish, with the iris and lens facing upwards and the optic nerve head (ONH) contacting the Petri dish (Figure 2A).
   2. Hold the eye cup steady at the limbus using forceps (Figure 2B).
   3. Detach the iris and lens by making small cuts circumferentially along the outer edge of the limbus (Figure 2A).
   4. Remove the iris and lens carefully. Avoid disturbing the retina during manipulation (Figure 2C).
      NOTE: The lens will not fall into the eye cup as it is attached, via zonule fibers, to the ciliary body.
2. Flatten the eye cup.
   1. With the eye cup still sitting upright, identify the ONH. This is easier with a bright white light source (Figure 2D).
   2. Incise at the four quadrants towards the ONH (Figure 2E). The Petri dish can be rotated for easier handling. Do NOT cut the ONH.
   3. Spread and flatten the eye cup carefully (Figure 2E).
   4. Apply the forceps on the sclera instead of the retina to avoid disrupting retinal integrity.
      NOTE: Peripheral retinal detachment and folding is unavoidable, as the weight of the overflowing vitreous will pull on the retina. In these areas, remove the vitreous to prevent further retinal folding. Remember to leave residual vitreous to stabilize the retina on top of the RPE-choroid and sclera.
3. Collect sandwich retinal explants.
   1. Place a surgical trephine on the retina in a region without retinal folds (Figure 2F).
   2. Press hard to penetrate the sclera, which should generate a cracking sound.
   3. Twist the trephine by 180° to ensure the sclera has been penetrated fully such that the retinal explant is now separated from the remaining tissue.
   4. Apply the forceps at the sclera and transfer the sandwich retinal explant to the culture medium (Figure 2G-I).
   5. Obtain 2-3 sandwich retinal explants from the peripheral retina of each quadrant.
      NOTE: The sandwich retinal explant can sometimes be trapped in the opening of the trephine. Using forceps, gently tease out a small section of the base of the sclera. This occurs more often when the blade is blunt and can cause loss of the retina (Figure 3A-C).
   6. Use a new trephine after cutting out 1-2 retinal explants as the blade becomes blunt easily (Figure 3C
      Figure 3: Troubleshooting. During the extraction process, the retinal explant may be trapped at the opening of the surgical trephine (A). Gently tease out the tissue from the base of the sclera without touching the retina (B). This can happen more often if the surgical trephine is used for extracting more than two retinal explants, as the blade can easily become blunt (C). Please click here to view a larger version of this figure.
4. Culture the retinal explants.
   1. Prepare the culture medium containing Dulbecco's Modified Eagle Medium nutrient mixture F-12 (DMEM-F12) and a 1x antibiotics and antimycotics mixture (AA, 100x stock).
   2. Prior to explant extraction, place 500 µL of medium in the wells of a 24-well plate and equilibrate in the incubator. This is important as adding the medium afterwards can dislodge the retina.
   3. Culture the sandwich retinal explants at 37 °C for up to 72 h in a humidified 5% CO₂ incubator in medium prepared in step 1.4.2.
   4. To induce DR-like changes, culture retinal explants in medium containing DMEM-F12 with a combination of 32.5 mM HG with Cyt, TNF-α (10 ng/mL) and IL-1β (10 ng/mL).

2. Paraffin-embedding of the sandwich retinal explants

1. Fix the sandwich retinal explants.
   1. Immerse the sandwich retinal explants in 10% formalin for at least 24 h.
   2. Remove the formalin solution and transfer the sandwich retinal explants gently into tissue pads and cassettes.
   3. Soak the sandwich retinal explants in 70% ethanol for at least 24 h.
   4. Paraffin-embed retinal explants into blocks.
   5. Cut paraffin-embedded retinal tissues into 5 µm thick sections using a microtome and mount onto glass slides. Store at room temperature until imaging.
2. Deparaffinize sections.
   1. Immerse the sections in 70% xylene for 5 min.
   2. Immerse the sections in 100% xylene for 5 min.
   3. Rehydrate the sections in 70% ethanol for 5 min.
   4. Rehydrate the sections in 100% ethanol for 5 min.
   5. Wash the sections under running tap water for 10 min.

3. Characterization using H&E, IHC and a Magnetic Luminex assay

1. H&E staining protocol
   1. Deparaffinize the sections using step 2.2.
   2. Hydrate the sections in tap water for 5 min.
   3. Stain the sections in Gill's 2 hematoxylin solution for 5 min.
   4. Wash the sections thoroughly under running tap water to remove excess stain.
   5. Differentiate by dipping the sections twice in 1% acid alcohol.
   6. Wash the sections quickly under running tap water.
   7. Stain the sections blue by dipping six times in 1% lithium carbonate (10 mg/mL).
   8. Wash the sections thoroughly under running tap water for 5 min to remove excess blue stain.
   9. Dip the sections in 1% eosin 10 times.
   10. Wash the sections quickly under tap water to remove excess stain.
   11. Dehydrate the sections by dipping them 10 times in 100% ethanol. Do this twice.
   12. Dip the sections in 70% xylene 10 times.
   13. Dip the sections in 100% xylene 10 times.
   14. Mount with a coverslip using dibutylphthalate polystyrene xylene (DPX) mounting medium.
   15. Take images using a light microscope.
2. IHC labeling protocol'
   1. Deparaffinize the sections using step 2.2.
   2. Place the slides into a solution containing 10 mM sodium citrate buffer with 0.05% Tween 20 at pH 6.0 and run an antigen retrieval in a pressure cooker automated at 121 °C for 2 min.
   3. Wash sections in phosphate buffered saline (PBS) for 5 min. Do this 3 times.
   4. Block the sections with PBS containing 0.1% Triton X-100 and 10% normal goat serum for 1 h at room temperature.
   5. Incubate sections overnight at 4 °C with primary antibodies conjugated to secondary antibodies (Table of Materials).
   6. Wash sections in PBS for 5 min. Do this 3 times.
   7. Stain nuclei using 4′,6-diamidino-2-phenylindole (DAPI) for 2 min.
   8. Wash and mount sections using an anti-fade reagent.
   9. Seal coverslips with nail polish.
   10. Take images using a confocal laser scanning microscope.
3. Magnetic Luminex assay
   1. Transfer 75 µL of media from each well to a 96-well-u-bottom plate at 24 and 72 h.
   2. Analyze the cell supernatant for IL-18, IL-6, IL-8 and vascular endothelial growth factor (VEGF) after 24 and 72 h using Luminex cytokine assay. Follow the manufacturer's instructions to conduct the assay¹⁶.

Retinal integrity was preserved in this HORC model. Retinal integrity was preserved in the cultured sandwich retinal explants but was lost in the retina cultured without adjacent structures. H&E was conducted to examine the structural integrity of sectioned sandwich retinal explants after 72 h in culture. The sandwich retinal explants showed preserved integrity and a distinct lamellae structure from GCL to ONL with compact nuclei in INL and ONL (Figure 4A). However, disrupted retinal integrity was found at the edges of the same sample, where the retina had detached from the underlying RPE-choroid and sclera, showing reduced retinal thickness and sever loss of nuclei in INL and ONL (Figure 4B).

Figure 4: H&E images of retinal explants cultured in medium for 72 h. A) Structural integrity maintained in areas attached to the underlying RPE-choroid and sclera, depicted by the distinct separation of each retinal layer from GCL to ONL and compact nuclei in INL and ONL. B) Severe loss in retinal integrity in regions detached from the RPE-choroid and sclera, shown by the overall retinal thickness reduction and decreased numbers of nuclei in INL and ONL. Scale bar = 50 µm. Please click here to view a larger version of this figure.

Characterization of retinal explants using IHC demonstrated well-preserved retinal integrity even after 72 h in culture. No TUNEL-positive cell nuclei, which mark apoptosis, were found in retinal explants cultured in basal conditions, demonstrating sustained cell viability even after 72 h (Figure 5A-C). GFAP expression remained localized in GCL and IPL, as seen in normal retinal tissues^17,18. There was no glial fibrosis, a sign of Müller cell activation and proliferation in response to retinal inflammation, which would otherwise be identified as upregulated GFAP expression that extends into the ONL^10,11,12(Figure 5D-F). Vimentin labelling (red) was highly expressed from GCL to ONL, showing preserved Müller cell integrity, as seen in normal retinal tissues (Figure 5G-I).

Figure 5: IHC images of retinal explants after culturing in basal media for 72 h. The lack of TUNEL-positive staining (green) suggests that no cellular apoptosis has occurred in the retinal layers from GCL to ONL (A-C). GFAP labelling (red) was restricted to GCL and IPL without glial fibrosis, which would be identified as extended expression of GFAP into the ONL (D-F). Vimentin (red) was highly expressed from GCL to ONL, locating the Müller cells that span through these retinal layers (G-I). Cell nuclei were stained with DAPI (blue). Scale bar = 50 µm. Please click here to view a larger version of this figure.

A DR model was developed by culturing sandwich retinal explants in HG and Cyt. HG and Cyt, IL-1β and TNF-α, increased the secretion of IL-18, VEGF, IL-6 and IL-8 relative to baseline levels. A Luminex assay was used to measure the cytokines IL-18, VEGF, IL-6 and IL-8 secreted by retinal explants after culturing for 24 and 72 h under HG + Cyt conditions (Figure 6). To minimize inter-donor differences, secreted cytokine levels were compared to baseline (indicated by the dotted red line), represented by cytokine levels secreted by the sandwich explants from the same donor but cultured in basal medium (prepared in step 1.4.1). After 24 h, secreted IL-18, VEGF, IL-6 and IL-8 levels increased significantly relative to baseline. After 72 h, only IL-18 and VEGF levels remained significantly elevated, whereas no significant difference was found in IL-6 and IL-8 levels compared to baseline.

Figure 6: Cytokines release by the sandwich retinal explants cultured in HG and Cyt, IL-1β and TNF-α, after 24 and 72 h. The secreted IL-18, VEGF, IL-6 and IL-8 levels increased significantly above baseline after 24 h (p < 0.05). After 72 h, IL-18 and VEGF levels remained significantly increased (p < 0.05), but no significant difference was found in IL-6 and IL-8 levels, compared to baseline. The dotted red line indicates the baseline level, which represents cytokine levels secreted by the explants from the same donor but cultured in normal medium. Please click here to view a larger version of this figure.

HORC is currently the most clinically translatable model in preclinical retinal research. Compared to in vitro cell culture models, HORC can better represent the anatomy of the human retina in situ, through retaining the dynamic retinal cell types and their connections with neurons, vasculatures and the extracellular environment¹⁹. Compared to animal models, HORC are more advantageous in studying the pathophysiology of and designing pharmaceutical treatments for human retinal diseases due to inter-species variations, such as different retinal cell types as well as varying photoreceptor density and proportions, which could contribute to unsuccessful clinical trials despite positive results in animal models¹⁹. However, in previous HORC protocols, the retina was intentionally detached from the RPE-choroid and sclera. Without the tough sclera, the flimsy retina is difficult to handle and retinal integrity can be compromised if directly handled by forceps^6,7,8,9. This is evidenced by the fact that previous studies using isolated retinas have reported increased Müller cell activation and TUNEL-positive nuclei by 72 h of culture. This is in contrast to our findings, suggesting that the sandwich method better preserves retinal integrity. Secondly, increased neuronal degeneration and glial proliferation were found in animal ORC models in retina cultured without RPE, compared to retina cultured with RPE^10,11,12.

To minimize the chances of compromising retinal integrity, sandwich retinal explants were utilized in this novel HORC protocol. In contrast to previous HORC protocols, the trephine cut through not only the retina, but also the RPE-vitreous and the sclera, such that the resultant explant is "sandwiched" and stabilized between the residual vitreous and the underlying structures. The residual vitreous weighs the retina down to prevent folding and detachment from the underlying structures while the sclera acts as a scaffold to support the retinal architecture to protect the retina from forceps-induced injury. Furthermore, culturing retina with RPE can prevent retinal degeneration and glial proliferation in culture^10,11,12.

Remember the following points when extracting sandwich retinal explants. When cutting along the limbus, the lens will weigh down the iris but will not fall inside the eye cup as it is attached, via zonule fibers, to the ciliary body. Apply the forceps on the sclera and avoid touching the retina to prevent disrupting retinal integrity. Fill the culturing wells with medium before placing the retinal explant inside the wells. Adding the medium in after placing the retinal explants inside the wells increases the risk of dislodging the retina from the RPE-choroid and sclera or flipping the entire explant upside down. When making deep incisions at the four quadrants, be mindful that the peripheral retina may detach due to the vitreous leaking out of the eye cup. To prevent detachment or folding of the retina, remove the vitreous in the area that is pulling on the retina. Although the previous point advises to remove the vitreous pulling on the retina, leave sufficient residual vitreous to weigh the retina on top of the RPE-choroid and sclera in order to increase the stability of the sandwich retinal explants. When cutting through the tough, fibrous sclera, the trephine can easily become blunt. As a result, excessive force may be required or the sclera may not be fully penetrated. Use a new trephine for every 1-2 extracted retinal explants to avoid incomplete penetration through the sclera. Experience in careful handling is crucial to protect the retina; therefore, training with porcine eyes, which are similar in size and structure to the human eye, is advisable before using the limited donor scleral eyecups available.

Retinal integrity was preserved in the explants sandwiched between residual vitreous and the underlying RPE-choroid and sclera. H&E staining showed better structural integrity when the cultured retinal explant was attached to RPE-choroid and sclera, compared to culturing the retina alone. Supporting this finding, the lack of TUNEL-positive cell nuclei suggested an absence of apoptosis within all retinal layers, demonstrating preserved cellular viability even after 72 h. Furthermore, vimentin labelling confirmed Müller cell integrity and the restricted GFAP expression in GCL and INL, found in normal retina, again verified the lack of inflammatory or stress signals in the retina.

The HORC model characterized in this study resulted in preserved retinal structure and Müller cell integrity while preventing retinal inflammation and cellular apoptosis. Furthermore, it can be used to model retinal diseases, which is particularly useful to identify disease pathways and pharmaceutical mechanisms. Previous animal and in vitro studies have shown that DR inflammation can be induced when animal retina or cells are exposed to HG and Cyt^13,14,15. In this HORC study, by applying HG and Cyt to the sandwich explants for 24 h in culture, secretion of IL-18, VEGF, IL-6 and IL-8 was significantly elevated above baseline. These cytokines are well known to increase in the vitreous and serum of patients with DR compared to non-diabetic controls, validating that this DR model is clinically translatable^{20,21,22,23,24,25}. Although this study has only characterized DR using Magnetic Luminex assay, other techniques such as Western Blotting, IHC and polymerized chain reaction (PCR) can also be performed²⁶.

Despite the clear advantages outlined above, this novel HORC model also has some limitations. A major limitation of this protocol is the low supply of tissue. Donor eye cups, which have all ocular structures but the cornea generally removed for transplantation, are rarely available for research due to the large demand for the sclera in surgical procedures such as glaucoma shunts, orbital implants, damage repair in retinal detachment and coverage of exposed sutures. In order to minimize this limitation, we suggest that all experimental groups are collected from the same donor. This removes the need to account for inter-donor-variabilities, such as age, duration of disease, baseline cytokine levels, and other factors that can contribute to confounding variables if comparing between different patients. Also, most donor tissues are from elderly people as it is harder to obtain samples from younger age groups, and even when available, the cause of death of young donors is often traumatic in nature leading to damaged tissues that are unsuitable for culture. Nonetheless, given that most chronic retinal diseases occur in aged patients, this model still retains usefulness in studying age-related chronic retinal diseases such as DR and age-related macular degeneration.