Imaging retinal tissue can provide single-cell information that cannot be gathered from traditional biochemical methods. This protocol describes preparation of retinal slices from zebrafish for confocal imaging. Fluorescent genetically encoded sensors or indicator dyes allow visualization of numerous biological processes in distinct retinal cell types.
The retina is a complex tissue that initiates and integrates the first steps of vision. Dysfunction of retinal cells is a hallmark of many blinding diseases, and future therapies hinge on fundamental understandings about how different retinal cells function normally. Gaining such information with biochemical methods has proven difficult because contributions of particular cell types are diminished in the retinal cell milieu. Live retinal imaging can provide a view of numerous biological processes on a subcellular level, thanks to a growing number of genetically encoded fluorescent biosensors. However, this technique has thus far been limited to tadpoles and zebrafish larvae, the outermost retinal layers of isolated retinas, or lower resolution imaging of retinas in live animals. Here we present a method for generating live ex vivo retinal slices from adult zebrafish for live imaging via confocal microscopy. This preparation yields transverse slices with all retinal layers and most cell types visible for performing confocal imaging experiments using perfusion. Transgenic zebrafish expressing fluorescent proteins or biosensors in specific retinal cell types or organelles are used to extract single-cell information from an intact retina. Additionally, retinal slices can be loaded with fluorescent indicator dyes, adding to the method's versatility. This protocol was developed for imaging Ca2+ within zebrafish cone photoreceptors, but with proper markers it could be adapted to measure Ca2+ or metabolites in Müller cells, bipolar and horizontal cells, microglia, amacrine cells, or retinal ganglion cells. The retinal pigment epithelium is removed from slices so this method is not suitable for studying that cell type. With practice, it is possible to generate serial slices from one animal for multiple experiments. This adaptable technique provides a powerful tool for answering many questions about retinal cell biology, Ca2+, and energy homeostasis.
The zebrafish (Danio rerio) has become widely used in medical and basic scientific research1, owing to its small size, rapid development and vertebrate organ systems. The natural transparency of zebrafish larvae combined with established methods for transgenesis have enabled detailed visualization of cellular processes in a living animal. A number of genetically encoded fluorescent biosensors have been targeted to specific zebrafish cells to detect Ca2+ 2, hydrogen peroxide3, apoptotic activation4 and ATP5.
In vivo imaging of zebrafish larvae has led to breakthroughs in the field of neuroscience, including mapping of brain circuitry6 and drug development for central nervous system disorders7. Zebrafish are well suited for vision research because their retinas feature the laminar structure and neuron types of higher vertebrates, and they display robust visual behaviors8,9. Several types of retinal degenerations analogous to human disease have been modeled successfully and studied in zebrafish10,11, including live imaging of individual photoreceptors degenerating within a retina2,12.
While in vivo larval zebrafish imaging is a valuable tool, it becomes more challenging as fish grow and develop pigmentation, and some pharmacological treatments cannot permeate an entire animal. Further, certain cellular processes change with development and age, making later time points critical for understanding function and the progression of disease in adult animals. Biochemical methods such as immunoblot, quantitiative PCR, O2 consumption, and metabolomic analyses can provide important clues about biology of the retina as a whole, but it is difficult to discern contributions of individual cell types affected by disease. Imaging isolated retinal tissue ex vivo bypasses these issues, and while imaging flat mounted retinas affords a view of the outer retina13, deeper inner retinal features are obscured. Transverse retinal slices, such as those presented in fixed immunohistochemical analyses, enable a clear view of all layers and cell types but only offer a single snapshot of the dynamic processes involved in normal function and disease.
Here, we present a method for generating ex vivo transverse retinal slices from adult zebrafish for imaging. It is similar to methods for preparing amphibian and zebrafish retinal slices for electrophysiological and morphological studies14,15, with important modifications for time lapse imaging ex vivo using confocal microscopy. Fluorescence responses of biosensors or dyes in slices are monitored in real time with a confocal microscope while delivering pharmacological agents using perfusion. While the method was developed for imaging photoreceptors, it may be feasible to use it for visualizing Müller cells, bipolar cells, horizontal cells, amacrine cells, or retinal ganglion cells with appropriate fluorescent markers. Additionally, slices can be loaded with fluorescent cell-permeable dyes to report cell viability, vesicular transport, mitochondrial function, or redox state. This versatile preparation allows visualization of a wide range of subcellular processes throughout the retina, including Ca2+ dynamics, signal transduction and metabolic state.
All animal experiments were approved by the University of Washington Institutional Animal Care and Use Committee.
1. Preparing Animals and Equipment
NOTE: The retinal pigment epithelium (RPE) is a dark sheet of tissue surrounding the outside of the retina whose pigmentation can obscure retinal features and damage the tissue when confocal imaging ex vivo. In darkness, the RPE of zebrafish is retracted away from the retina; dark adapt fish to facilitate future removal of the RPE from the retina before slicing and imaging.
2. Preparing retinal slices (see Figure 1)
3. Imaging retinal slices
4. Imaging retinal slices during perfusion experiments where solutions are changed or flowed continuously
NOTE: Imaging retinal slices during perfusion experiments where solutions are changed or flowed continuously is similar to setup for static imaging or injection experiments, with the following modifications.
Stable positioning and transverse orientation of slices are key to successful imaging with injection or perfusion of pharmacological agents. Carefully examine and reposition slices prior to confocal imaging as needed to ensure all retinal layers are visible (Figure 2A, slice ii). If a slice is rotated slightly forward (Figure 2A, slice iii), bundles of outer segments will be visible and small adjustments can be made with forceps to bring the desired retinal layers into focus. Slices poorly adhered to the filter paper (Figure 2A, slice i) or retaining RPE (Figure 2A, slice iv) should not be used for time lapse imaging.
Cell viability is paramount to observing physiological processes ex vivo; a cell viability stain such as propidium iodide (PI) is recommended to assay cell health while practicing retina slicing. Dead cells near the cut edge of all slices will accumulate PI in their nuclei (Figure 2B, left panels), while 5-10 µm deeper into the slice, healthy cells with normal morphology and no PI staining can be imaged. For example, in photoreceptors, PI negative cells below the cut edge commonly display a stereotypical polarized, elongated morphology (Figure 2B, right panels). Photoreceptors remain viable in retinal slices for at least 4 h when stored in oxygenated Ringer's solution.
Many retinal cell structures can be visualized using combinations of transgenic markers and dyes; example confocal images of fresh retinal slices with double and triple fluorescent labeling are presented in Figure 3. To visualize dynamics of cone photoreceptor endoplasmic reticulum (ER) relative to the nucleus, retinal slices from transgenic fish expressing GFP targeted to cone ER17 can be counterstained with a nuclear dye (Figure 3A). A similar strategy can be employed to view the actin cytoskeleton using retinal slices from transgenic zebrafish expressing GFP-fused actin18 (Figure 3B). With another cell-specific promoter, Müller cells can be labeled with the red fluorescent protein tdTomato19 and visualized in retinal slices (Figure 3C, top panel). Glucose uptake into the retina can be assayed in vivo by orally gavaging adult zebrafish with a fluorescent glucose analog (NBDG), which becomes incorporated into cells visible in a retinal slice21 (Figure 3C, bottom panel). Double transgenic zebrafish can be employed to monitor multiple cell processes or cell types in tandem, such as Ca2+ dynamics in a subtype of cones. Figure 3D shows a triple labeling scheme for this type of experiment, with tdTomato expressed in long-wavelength cones2 together with mitochondrially-targeted Ca2+ sensor (mito-GCaMP) expressed in all cones22 and a nuclear stain.
Time lapse imaging of retinal cells is a key advantage of this slice prep. For instance, Ca2+ fluctuations in cone photoreceptor cytosol can be observed and later quantified during pharmacological manipulation, an experiment depicted in Figure 4. Figure 4A shows retinal slices expressing the Ca2+ sensor GCaMP2 (top) or control eGFP16 (bottom) in cone photoreceptors counterstained with a red lipophilic dye. In this experiment Na+ is isotonically depleted from the imaging chamber using perfusion, evoking large increases in cytosolic Ca2+ for the outer segment and cell body as reflected by GCaMP (Figure 4A, top panel). Fluorescence of retinal slices expressing Ca2+-insensitive eGFP is unaffected by this treatment (Figure 4A, bottom panel). Time lapse movies can be analyzed using ImageJ software to isolate and quantify fluorescence responses from single cone outer segments, cell bodies, and synapses (Figure 4B).
Supplement Solution | ||
50X stock* concentration (mM) | working concentration (mM) | |
D-glucose | 500 | 10 |
sodium lactate | 50 | 1 |
sodium pyruvate | 25 | 0.5 |
L-glutamine | 25 | 0.5 |
reduced glutathione | 25 | 0.5 |
ascorbic acid | 15 | 0.3 |
* store aliquots at -20 ºC < 6 months; add fresh to Ringer's solution |
Table 1. Components of 50X supplement stock solution and final working concentrations.
Ringer's Solution | |
concentration (mM) | |
NaCl | 133 |
KCl | 2.5 |
CaCl2 · 2H2O | 2 |
NaH2PO4 | 1.5 |
MgCl2 · 6H2O | 1.5 |
HEPES | 10 |
pH to 7.4 using NaOH | |
store at 4 ºC in sterile bottle < 1 month |
Table 2. Components of standard Ringer's solution.
Na+-Free Ringer's Solution | |
concentration (mM) | |
TRIS | 147 |
HCl | 120 |
KCl | 1 |
CaCl2 · 2H2O | 2 |
KH2PO4 | 1.5 |
MgCl2 · 6H2O | 1.5 |
HEPES | 10 |
pH to 7.4 using HCl | |
store at 4 ºC in sterile bottle < 1 month |
Table 3. Components of Na+-free Ringer's solution.
Confocal Imaging Settings | |||
Fluorophore | Excitation | Emission Filter | |
Wavelength | Laser Intensity | ||
GFP (including GCaMP) | 488 nm | 2 – 5% | eGFP or AlexaFluor 488 |
tdTomato | 559 nm | 5% | AlexaFluor 594 |
PI | 559 nm | 2% | PI |
Hoechst 33342 | 405 nm | 1% | DAPI |
NBDG | 488 nm | 10% | eGFP or AlexaFluor 488 |
C12 558/568 BODIPY | 559 nm | 1% | AlexaFluor 594 |
TMRM | 559 nm | 3% | RFP |
Table 4. Imaging conditions for fluorescent markers and dyes presented in Figures 2-4.
Figure 1. Schematic for preparing fresh zebrafish retinal slices. (A) Dissect away the eyecup, and discard lens and sclera (step 2.4.1.). (B) Cut the eyecup into three pieces; discard small edge piece (step 2.4.2.). (C) Drag filter paper under eyecup pieces with the inner retina facing toward the filter paper (step 2.5.1.-2.5.2.). (D) Flatten retina by using a paper towel to wick Ringer's solution downward through the filter paper (step 2.5.3.), then gently peel away remaining RPE with forceps (step 2.6). Move the filter paper to the slicing chamber on the tissue slicer stage and cut 400 µm slices (steps 2.8., 2.9.). (E) Transfer slicing chamber with slices and a wax or petroleum jelly ladder to a Petri dish of Ringer's solution (step 2.10.1.). (F) Use fine forceps to slide single slices from the slicing chamber to the ladder while keeping slices submerged. (G) Rotate filter paper strips (black) 90° and bury the edges of the filter paper in wax or petroleum jelly (steps 2.10.2.-2.10.3.). (H) Schematic of the final imaging chamber loaded with a ladder and slices. Please click here to view a larger version of this figure.
Figure 2. Examples of fresh zebrafish retinal slices displaying proper adhesion to the filter paper and cell viability. (A) Brightfield image of fresh retinal slices in a petroleum jelly ladder. Slices ii and iii display good adhesion and transverse retinal layers. Slices i and iv have retained substantial RPE, or are highly curved and not well-adhered to the filter paper, and should not be imaged. Scale bar = 200 µm. (B) Top-down Z montage of a fresh retinal slice from transgenic zebrafish expressing eGFP in cone photoreceptors (Tg(gnat2:eGFP)16). Hoechst dye labels all nuclei; propidium iodide (PI) counterstaining labels nuclei of dead cells, which appear near the cut edge of the slice (left). Z-stack step size = 1 µm; scale bar = 20 µm. Fluorescent imaging conditions are outlined in Table 4. Please click here to view a larger version of this figure.
Figure 3. Sample confocal images of double- and triple-labeled ex vivo retinal slices from transgenic adult zebrafish. (A) Two-color imaging scheme employing transgenic GFP tagged endoplasmic reticulum in cones (Tg(gnat2:calr-GFP)17, top) with Hoechst nuclear counterstain (blue, bottom). (B) GFP tagged actin in cones (Tg(gnat2:LifeAct-GFP)18, top) with Hoechst nuclear counterstain (blue, bottom). (C) RFP labeled Müller cells (Tg(GFAP:tdTomato)19, top) from a zebrafish fed fluorescent glucose (NBDG) demonstrating glucose uptake into cones20,21 (green, bottom). (This image is from Figure 2B of Kanow, et al.21) (D) Example of three-color imaging using double transgenic zebrafish. Left, RFP targeted to long-wavelength cone photoreceptors (Tg(trβ2:tdTomato)2). Center, calcium biosensor GCaMP targeted to cone photoreceptor mitochondria (Tg(gnat2:mito-GCaMP3)22). Right, overlaid images of tdTomato (magenta), mito-GCaMP (green), and Hoechst nuclear counterstain (blue). Images are maximum intensity Z-projections of 9 frames over a 7 µm tissue depth; scale bars represent 10 µm. Fluorescent imaging conditions are outlined in Table 4. Please click here to view a larger version of this figure.
Figure 4. Ca2+ imaging with GCaMP and control eGFP. (A) Representative images of fresh retinal slices expressing the fluorescent Ca2+ biosensor GCaMP (gnat2:GCaMP32, top) or eGFP (bottom) in cone photoreceptors. Left, slices at baseline; right, slices 2 min after Na+ was isotonically depleted from the imaging chamber using perfusion of a Tris-based Ringer's solution, which traps Ca2+ in photoreceptors. Scale bars = 10 µm. Fluorescent imaging conditions are outlined in Table 4. (B) Mean fluorescence changes of single photoreceptor compartments (outer segments, cell bodies, and synapses) during time lapse imaging of Na+ depletion. Slices were imaged every 10 s, stacks were processed using ImageJ, and fluorescence of GCaMP or eGFP was normalized to signal from the membrane dye. Solid lines, GCaMP (n for outer segments = 15, cell bodies = 24, synapses = 26); dashed lines, eGFP (n for outer segments = 26, cell bodies = 20, synapses = 31). Error bars represent standard error of the mean. Please click here to view a larger version of this figure.
Ex vivo imaging of fresh zebrafish retinal slices has proven to be a versatile tool for studying photoreceptor biology20,21,22, and is unique in that it enables analysis of single cells in a mature, fully differentiated retina. With practice, it is possible to conduct multiple experiments with tissue from a single fish, even using serial slices from the same part of the retina. In addition to the challenges and suggestions regarding preparation of amphibian retinal slices for electrophysiology studies14, there are important considerations for imaging experiments.
For photoreceptors, cell viability generally correlates with cell morphology, so it is important to handle the delicate retina minimally, particularly after the RPE has been removed. After slicing, use fine forceps to carefully slide slices horizontally away from the slicing chamber (Figure 1F) to transfer them to the ladder rather than lifting slices straight up, and always keep slices submerged in Ringer's solution. If possible, assemble the ladder of retinal slices near the confocal microscope to minimize damage to slices from shaking during transport.
Strong adhesion of retinal tissue to the filter paper is critical for creating slices that will remain stable during injection or perfusion experiments. It is necessary to carefully inspect each slice for morphology and stability just prior to imaging (Figure 2A); gently tapping the microscope table while viewing slice movement through the confocal ocular lens reveals stability of a particular slice. Despite precautions, drift in the Z-direction during injection or perfusion can present challenges for analysis, so it is useful to load a control dye, such as lipophilic dyes for membranes and mitochondria, to stably label an identifiable cell structure for normalization (Figure 4A, magenta). If slices drift > 10 µm in the Z-direction it may not be possible to extract usable single-cell data. Moderate drift in the X-Y direction can be corrected in post-processing using registration software such as MultiStackReg for ImageJ (RRID:SCR_002285).
Under static conditions, fluorescence of markers in retinal slices should remain stable, although photobleaching can occur during imaging. Care should be taken to minimize laser exposure during time lapses by refining parameters such as laser intensity, scan speed, and frame rate. Imaging controls are recommended for each fluorescent marker used. Controls include injecting or perfusing Ringer's solution without pharmacological agents in separate experiments, or conducting imaging experiments with non-biosensor markers that fluoresce constitutively.
Depending on the confocal excitation laser being used, pigments in the RPE can contribute to autofluorescence and even generate heat during imaging, so it is important to remove most of this tissue prior to slicing. Dark-adapting animals aids in removal of the RPE while minimizing damage to the underlying retina. Inability to image the RPE is a limitation of this slice preparation, though this may be overcome by using albino animals with a transparent RPE. Another limitation is that retinal ganglion cells and end feet of Müller cells may become obscured from view by the filter paper; as an alternative preparation retinas may be flat mounted with the photoreceptor side down on the filter paper and then sliced to provide a clearer view of the innermost retina. It is also important to note that long horizontal projections of some cells, such as wide field amacrine cells, may be severed during dissection or slicing. A final limitation is that many fluorescent dyes accumulate nonspecifically in photoreceptor outer segments, so for this part of the retina it is advisable to use genetically encoded biosensors rather than indicator dyes.
Given the wide range of available fluorescent cell reporter dyes23,24 for tissue and genetically encoded fluorescent biosensors used successfully in zebrafish2,3,4,5, this slice preparation could be used to study numerous biological processes in several retinal cell types (see examples in Figure 3, Figure 4). Imaging fresh retinal slices using live cell super-resolution microscopy could also provide exciting new insights to retinal function and health on a subcellular level. Further, this method can be adapted for ex vivo imaging of mouse retinal slices25. While successful preparation of fresh retinal slices requires practice, it is a powerful tool that is useful for addressing a wide range of cell-specific biological questions in a mature retina.
The authors have nothing to disclose.
We thank Ralph Nelson and Daniel Possin for thoughtful guidance while developing this protocol, and Eva Ma, Ashley George and Gail Stanton for generation of stable transgenic zebrafish lines. The work was supported by NSF GRFP 2013158531 to M.G., NIH NEI 5T32EY007031 to W.C. and M.G., and EY026020 to J.H. and S.B.
zebrafish | Univeristy of Washington South Lake Union Aquatics Facility | stocks maintained in-house as stable transgenic lines | |
petroleum jelly | Fisher Scientific | 19-090-843 | for petroleum jelly syringe |
3-mL slip tip syringe | Fisher Scientific | 14-823-436 | for petroleum jelly syringe |
20g 3.8 cm slip tip needle | Fisher Scientific | 14-826-5B | for petroleum jelly syringe |
plain 7 cm X 2.5 cm microscope slide | Fisher Scientific | 12-550-A3 | for eyecup dissection, slicing chamber |
Seche Vite clear nail polish | Amazon | B00150LT40 | for slicing chamber |
18 mm X 18 mm #1 glass coverslips | Fisher Scientific | 12-542A | for imaging ladders |
unflavored dental wax | Amazon | B01K8WNL5A | for imaging ladders |
double edge razor blades | Stoelting | 51427 | for tissue slicing |
tissue slicer with digital micrometer | Stoelting | 51415 | for tissue slicing |
filter paper – white gridded mixed cellulose, 13 mm diameter, 0.45 µm pore size | EMD Millipore | HAWG01300 | filter paper for mounting retinas |
10 cm petri dish | Fisher Scientific | FB0875712 | for fish euthanasia, dissection, imaging ladder assembly |
15 cm plain-tipped wood applicator stick | Fisher Scientific | 23-400-112 | for wire eye loop tool |
30g (0.25 mm diameter) tungsten wire | Fisher Scientific | AA10408G6 | for wire eye loop tool |
D-glucose | Sigma Aldrich | G8270 | component of supplement stock solution |
sodium L-lactate | Sigma Aldrich | L7022 | component of supplement stock solution |
sodium pyruvate | Sigma Aldrich | P2256 | component of supplement stock solution |
L-glutamine | Sigma Aldrich | G3126 | component of supplement stock solution |
L-glutathione, reduced | Sigma Aldrich | G4251 | component of supplement stock solution |
L-ascorbic acid | Sigma Aldrich | A5960 | component of supplement stock solution |
NaCl | Sigma Aldrich | S7653 | component of Ringer's solution |
KCl | Sigma Aldrich | P9333 | component of Ringer's solution |
CaCl2 · 2H2O | Sigma Aldrich | C3881 | component of Ringer's solution |
NaH2PO4 | Sigma Aldrich | S8282 | component of Ringer's solution |
MgCl2 · 6H2O | Sigma Aldrich | M0250 | component of Ringer's solution |
HEPES | Sigma Aldrich | H3375 | component of Ringer's solution |
Tris base | Fisher Scientific | BP152 | component of Na+-free Ringer's solution |
6 N HCl | Fisher Scientific | 02-003-063 | component of Na+-free Ringer's solution |
KH2PO4 | Sigma Aldrich | P5655 | component of Na+-free Ringer's solution |
50 mL conical centrifuge tube | Denville Scientific | C1062-P | container for Ringer's solution |
Vannas scissors – 8 cm, angled 5 mm blades | World Precision Instruments | 501790 | micro-scissors for eyecup dissection |
Swiss tweezers – #5, 11 cm, straight, 0.06 X 0.07 mm tips | World Precision Instruments | 504510 | fine forceps for eyecup dissection and slice manipulation |
single edge razor blades | Fisher Scientific | 12-640 | for eyecup dissection and trimming filter paper |
EMD Millipore filter forceps | Fisher Scientific | XX6200006P | flat forceps for handling wet filter paper |
C12 558/568 BODIPY | Fisher Scientific | D3835 | stains live cell nuclei; incubate 5 µg/mL for 15 min at room temperature |
propidium iodide (PI) | Fisher Scientific | P3566 | stains dead cell nuclei; incubate 5 µg/mL for 20 min at room temperature |
Hoechst 33342 | Fisher Scientific | 62249 | stains live cell nuclei; incubate 5 µg/mL for 20 min at room temperature |
Tetramethylrhodamine, methyl ester (TMRM) | Fisher Scientific | T668 | stains functional, negatively-charged mitochondria; incubate 1 nM for 30 min at room temperature |
tissue perfusion chamber | Cell MicroControls | BT-1-18/BT-1-18BV [-SY] | imaging chamber for injection or perfusion |
2-(N-(7-Nitrobenz-2-oxa-1,3-diazol-4-yl)Amino)-2-Deoxyglucose (NBDG) | Fisher Scientific | N13195 | fluorescent glucose analog adminitered orally to zebrafish 30 min prior to euthanasia |
Olympus laser scanning confocal microscope | Olympus | FV1000 | confocal microscope for visualizing fluorescence of slices at single-cell resolution |
Carbonyl cyanide 3-chlorophenylhydrazone (CCCP) | Sigma Aldrich | C2759 | experimental reagent which ablates mitochondrial respiration; treat slices to a final concentration of 1 µM |
miniature aspirator positioner | Cell MicroControls | FL-1 | for perfusion |
perfusion manifold, gas bubbler manifold, flow valve, 60cc syringe holder | Warner Instruments | various | for perfusion |