This work presents an alternative flatmount retina preparation in which the removal of photoreceptor cell bodies enables faster antibody diffusion and improved patch pipette access to inner retinal neurons for immunohistochemistry, in situ hybridization, and electrophysiology experiments.
Bipolar cells and horizontal cells of the vertebrate retina are the first neurons to process visual information after photons are detected by photoreceptors. They perform fundamental operations such as light adaptation, contrast sensitivity, and spatial and color opponency. A complete understanding of the precise circuitry and biochemical mechanisms that govern their behavior will advance visual neuroscience research and ophthalmological medicine. However, current preparations for examining bipolar and horizontal cells (retinal whole mounts and vertical slices) are limited in their capacity to capture the anatomy and physiology of these cells. In this work, we present a method for removing photoreceptor cell bodies from live, flatmount mouse retinas, providing enhanced access to bipolar and horizontal cells for efficient patch clamping and rapid immunolabeling. Split retinas are prepared by sandwiching an isolated mouse retina between two pieces of nitrocellulose, then gently peeling them apart. The separation splits the retina just above the outer plexiform layer to yield two pieces of nitrocellulose, one containing the photoreceptor cell bodies and another containing the remaining inner retina. Unlike vertical retina slices, the split retina preparation does not sever the dendritic processes of inner retinal neurons, allowing for recordings from bipolar and horizontal cells that integrate the contributions of gap junction-coupled networks and wide-field amacrine cells. This work demonstrates the versatility of this preparation for the study of horizontal and bipolar cells in electrophysiology, immunohistochemistry, and in situ hybridization experiments.
The retina is a thin neural tissue located in the posterior eye where light is intercepted and processed into an electrochemical signal that can be interpreted by the brain. At the back of the retina, rod and cone photoreceptors are stimulated by light, which reduces the tonic release rate of the neurotransmitter, glutamate1. The first neurons to experience and respond to this light-induced change in glutamate concentration are the bipolar cells (BCs) and horizontal cells (HCs), whose somas reside in the outermost region of the inner nuclear layer (INL). These second-order neurons perform the first stage of signal processing in the retina and shape critical features of vision such as light adaptation, contrast sensitivity, and spatial/color opponency2. While these functions have been ascribed to BCs and HCs, the circuitry and biochemical mechanisms underlying these processes are not fully understood3. Therefore, the advancement of tools and methods for exploring BC and HC physiology is of paramount importance.
Vertical (transverse) retina sections have long proven the most practical model for studying BCs and HCs; however, certain aspects of BC and HC physiology are inaccessible to the experimenter under this model. Direct recordings from HCs or indirect measurements of their effects on BCs do not reflect the retina's endogenous connectivity since the lateral processes of these cells are severed during slicing. Whole mount retina preparations circumvent this issue by preserving these lateral processes, but the surrounding retinal layers pose a challenge for accessing these cells4. While there are abundant examples of immunostaining5,6,7,8 and patch clamp recordings9 from INL neurons in whole mount retinas, there is an opportunity to expedite and simplify the collection of these data. The inherent limitations of transverse sections and the challenges of the whole mount model thus inspired the development of this alternative flatmount retina preparation.
The following work describes a protocol for easily removing the photoreceptor layer from live, flatmount retinas to enhance access to BCs and HCs for simplified patch clamping and faster, more efficient immunolabeling. Peeling apart two pieces of nitrocellulose membrane attached to either side of an isolated retina tears the tissue through the photoreceptor axons, leaving a split retina that retains the outer plexiform layer (OPL) and all inner retinal layers. While others have described protocols for mechanically separating layers of the retina, these methods are either poorly suited to patch clamping and microscopy applications or require tedious manipulation of the tissue. Several of these methods require frozen or lyophilized tissue for layer separation, making them incompatible with electrophysiology experiments10,11,12. Others are designed for live tissue, but necessitate 5-15 sequential peels with filter paper4,11 or treatment with trypsin13 to remove the photoreceptors. The technique described here improves upon its predecessors by simplifying the photoreceptor removal procedure and expanding the repertoire of downstream applications.
After photoreceptors transduce photon absorption into neurotransmitter release, BCs and HCs are the first retinal neurons to process the visual signal23. While the importance of these neurons is well-appreciated, many of their functions are incompletely understood or unexplored altogether. Many BC and HC physiology studies would likely benefit from a flatmount retina preparation that improves access to INL neurons while preserving lateral connectivity. The development of the split retina method represents an effort to provide an easy protocol for acquiring high quality electrophysiological recordings and microscopy data from BCs and HCs in a flatmount orientation. The split retina preparation described here can be performed in about 20 min per mouse (10 min per retina) following retina isolation, without the use of specialized equipment. The method draws inspiration from existing photoreceptor removal procedures but offers significant improvements in simplicity, speed, and versatility4,10,11,12,13. Unlike previous methods for separating retinal layers, retina splitting does not require freezing, lyophilization, or repeated application of adhesives to the retina. With practice, nearly all photoreceptors can be removed in a single tear with the nitrocellulose membrane. The speed and ease of this approach allows one to minimize the time the retina spends out of carbogenated Ames, enabling high cell viability for long periods; split retinas can be maintained in carbogenated Ames media for several hours post-split. As a testament to the health of INL neurons in this preparation, a live/dead cell stain (Figure 4) and patch-clamp electrophysiology (Figure 6) confirm the viability of RBCs and HCs following a split.
The removal of the photoreceptor layer in split retinas conveys a significant advantage during immunolabeling by dramatically reducing the diffusion time for antibodies into the INL. Primary and secondary antibody labeling can be completed within 2 h, a substantial improvement to conventional flatmount staining which can take 72 h or longer depending on the target5,6,7,8,20,22. As a result, microscopy data can be acquired on the same day as tissue preparation, drastically accelerating the pace of immunofluorescence experiments. To facilitate mRNA probe annealing, FISH experiments typically recommend much longer fixation times (~24 h) than immunolabeling18. However, the experiments presented here demonstrate that a 2 h fixation still produces exceptional FISH labeling (Figure 5). Despite extending the fixation time from 30 min to 2 h, it was not necessary to perform antigen retrieval steps to obtain excellent immunolabeling, but this may vary with the antibody or antigen. The protease treatment in the FISH protocol may interfere with antibody labeling, likely due to the destruction of target epitopes. This issue was circumvented by using polyclonal antibodies that target multiple epitopes, decreasing the likelihood that epitope destruction would hinder immunolabeling. Additionally, a moderate protease treatment (ACD protease III) was used to prevent excessive epitope alteration while still providing sufficient tissue penetration.
Occasionally, the retina will instead split through the outer nuclear layer (ONL), leaving behind layers of photoreceptor somas with no INL cells visible. To prevent this, one should ensure that the retina lies completely flat on the glass, and that any residual liquid from around the retina has been removed. Pressing more firmly onto the nitrocellulose with the paintbrush may also help prevent splitting through the ONL. If the membrane becomes too wet or the retina is folded over itself, the chances of a successful split will be greatly diminished. Using DAPI to stain cell nuclei is useful for assessing the quality of the split and for determining the coverage of remaining photoreceptors. Photoreceptor nuclei are smaller, brighter, and more superficial (Supplementary Figure 3A), whereas BC nuclei are larger, dimmer, and deeper (Supplementary Figure 3B). In some cases, the plane of the tear will vary slightly across the piece of retina, resulting in patches where photoreceptors cell bodies have not been completely removed (Supplementary Figure 3C). For applications in microscopy and electrophysiology, this does not hinder the ability to collect quality data from regions where photoreceptors have been properly removed; large fields of exposed inner retina can easily be found when imaging or recording with a patch pipette. If more complete photoreceptor removal is desired, a second tear can be performed with an additional piece of nitrocellulose membrane, although 100% photoreceptor removal is not guaranteed. Caution is therefore advised when using split retinas in gene expression or proteomics studies where residual photoreceptor material could influence results. For single cell applications, this concern is unwarranted, as data from photoreceptors can be excluded from analysis.
The advantages of the split retina preparation are perhaps most salient in electrophysiological recordings of wide-field interneurons. Whereas traditional vertical slices sever the extensive processes of wide-field cells, the split retina preparation leaves the OPL and IPL intact, allowing one to capture input from wide-field cells such as HCs24, A17s25, TH ACs26, and NOS-1 ACs27 that would otherwise be overlooked in vertical slices. Therefore, interpretation of results and comparison with previous data collected from retinal slices requires careful thought. Nonetheless, in experiments using pharmacological mimics of light stimulation, these results resemble data recorded from retinal slices19. By expressing ChR2 under cell-specific promoters, one can stimulate a desired cell population while recording from BCs in the INL to investigate the desired cell's impact on the vertical information pathway. Recording directly from deeper INL neurons, such as amacrine cells, is also feasible in the split retina. While in this case the patch electrode must first travel through the more superficial INL neurons, there is considerably less tissue obstructing its path when compared to a traditional whole mount preparation.
In addition to measuring the influence of wide-field cells on other neurons, this method enables direct single-cell patch clamping from HCs, whose dendrites form an extensive gap-junction coupled network in the OPL28. Horizontal cells send critical feedback to photoreceptors which shapes the transmission of vertical information through the retina. However, since the dendritic fields of HCs are truncated in vertical slices, single cell recording data are lacking. This work presents anatomically and physiologically intact HCs from which ChR2-evoked currents are recorded in a triple transgenic mouse line (Figure 6 C-E). Outside of ChR2 stimulation, the split retina can be used to study endogenous HC currents and gap junction coupling28. While the split retina provides a convenient model for studying synaptic connectivity and neuronal activity induced by chemical application or ChR2 stimulation, the lack of photoreceptors precludes any direct exploration of natural light responses or light adaptation mechanisms.
In situ imaging in the retina has made admirable progress in recent years. However, the majority of imaging studies are limited to the ganglion cell layer in whole mount retina preparations29. The authors envision that the absence of photoreceptors in the split retina will make it an ideal model for live calcium imaging in the OPL and INL. Beyond calcium imaging, this model has great potential for use with genetically encoded biosensors such as iGluSnFR30,31, iGABASnFR32, and pHluorin33. Combined with the split retina preparation, these powerful tools may offer an efficient approach to exploring the synaptic interactions and biophysical properties of BCs and HCs that contribute to light processing in the retina.
The authors have nothing to disclose.
This work was supported by the following NIH grants: NIH grant R01EY031596 (to C.M.); NIH grant R01EY029985 (to C.M.); NIH grant P30EY010572 (to C.M.); NIH grant R01EY032564 (to B.S.). We thank Tammie Haley for her technical support in preparing retina sections, and Dr. Charles Allen for generously contributing the mRNA FISH probes used in this work.
#1.5 glass coverslips | Fisherbrand | 12544E | |
2 pairs of Dumont #5 forceps | Ted Pella | 38125 | |
25 gauge needle | Becton Dickenson | 305122 | |
470 nm LED | THORLABS | M470L2 | |
5-306 curved scissors | Miltex | 5-306 | |
9" disposable pasteur pipetes | Fisherbrand | 13-678-20D | for constructing custom transfer pipette |
Ai80d mouse | Jackson Laboratories | 25109 | RRID: IMSR_JAX:025109 |
Ames Medium w/L-Glutamate | US Biological | A1372-25 | |
amplifier control software | Molecular Devices | Clampex 10.3 software | |
anti-calbindin D28K antibody | Invitrogen | PA-5 85669 | RRID: AB_2792808, host species = rabbit; 1:100 dilution |
anti-CtBP2 antibody | BD Biosciences | 612044 | RRID: AB_399431, host species = mouse; 1:5000 dilution |
anti-GPR179 antibody | NA | NA | gift from Kirill Martemyanov; Scripps Research Institute, Jupiter, FL; host species = sheep; 1:1000 dilution |
anti-PKC alpha antibody | Sigma-Aldrich | P4334 | RRID: AB_477345, host species = rabbit; 1:5000 dilution |
anti-PKC alpha antibody | Santa Cruz Biotechnology | sc8393 AF594 | RRID: AB_628142, host species = mouse; 1:1000 dilution |
anti-PSD95 antibody | BD Transduction Laboratories | 610495 | RRID: AB_397862, host species = mouse; 1:1000 dilution |
anti-RGS11 antibody | NA | NA | gift from Ted Wensel; Baylor College of Medicine, Houston, TX; host species = rabbit; between 1:1000 and 1:5000 dilution |
anti-Synaptophysin P38 antibody | Sigma | S-S5768 | RRID: AB_477523, host species = mouse; 1:1000 dilution |
Aquamount mounting media | Epredia | 13800 | slide mounting media |
C57BL/6J mouse | Jackson Laboratories | 000664 | RRID: IMSR_JAX:000664 |
carbogen tank | Matheson | NA | 95% O2 and 5% CO2 |
custom transfer pipette | custom build | NA | Instructions: use scissors to cut off the tip of a plasitc transfer pipette at the point it begins to taper. Use pliers to safely break off the last 2-3 inches of a glass pasteur pipette. Fit the narrow end of the glass pasteur pipette into the wide tip of the plastic transfer pipete. Wrap parafilm around the joint of the two pieces to enhance the seal. |
Digitical optical power meter | THORLABS | PM100D | |
dissection microscope | Zeiss | Stemi 2000 | |
electrophysiology amplifier | Molecular Devices | Axopatch 200B | |
electrophysiology microscope | Olympus | OLYMPUS, BX50WI | Dodt gradient contrast microscopy |
Fluoromount-G | SouthernBiotech | 0100-01 | |
HC PL APO CS2 40x/1.3 | Leica | 506358 | |
HC PL APO CS2 63x/1.40 | Leica | 15506350 | |
Hybridization oven | Robbins Scientific | Model 1000 | for RNAscope protocol only |
Immedge hydrophobic barrier pen | Vector Laboratories | H-4000 | |
isoflurane | Piramal Critical Care | 66794-017-25 | |
Kimwipe (delicate task wipe) | Kimtech Science | 34155 | |
Leica HC PL APO CS2 40x/1.3 oil immersion objective | Leica | 506358 | |
Leica HC PL APO CS2 63x/1.40 oil immersion objective | Leica | 15506350 | |
Leica TCS SP8 X confocal microscope | Leica | discontinued | |
medium 15 mm petri dish | Corning | 25060-60 | eyes are kept here during retina dissection |
Merit 97-275 steel scissors | Merit | 97-275 | |
Micropipette Puller | Sutter Instrument | p-97 | |
Mm-Gabrd-C2 mRNA probe | ACD | 459481-C2 | |
mouse euthanasia chamber | NA | NA | custom build; glass petri dish covering a small glass jar. |
nitrocellulose membrane filters | GE Healthcare Life Sciences; Whatman | 7184-005 | 0.45 µm pore size |
Picospritzer | General Valve Corporation | Picospritzer II | referred to in the text as microcellular injection unit |
plastic transfer pipets | Fisherbrand | 13-711-7M | for constructing custom transfer pipette |
Plastic tubing | Tygon | R-603 | for connection to carbogen tank |
platinum harp | custom build | NA | for anchoring split retinas within the electrophysiology recording chamber. |
size 0 paint brush | generic | NA | for flattening retina during splitting. |
SlowFade Gold antifade reagent | Molecular Probes | S36937 | referred to in the text as anti-fade mounting media |
small 10 mm petri dish | Falcon | 353001 | eyes are placed here following enucleation |
small glass pane (7.5 cm x 5 cm) | generic | NA | isolatd retina pieces are placed onto this for the splitting procedure |
Superfrost plus microscope slides | Fisherbrand | 12-550-15 | electrostatically-charged glass microscope slides |
Thick-walled borosilicate glass pipettes with filament | Sutter Instrument | BF150-86-10HP | |
Vannas Scissors; straight | Titan Medical | TMS121 | not brand specific; any comparable scissors will work |
vGATFLPo mouse | Jackson Laboratories | 29591 | RRID: IMSR_JAX:029591 |
vGlut2Cre mouse | Jackson Laboratories | 28863, 016963 | RRID: IMSR_JAX:028863, RRID: IMSR_JAX:016963 |
Zombie NIR Fixable Viability Kit | BioLegend | 423105 | referred to in the text as MI-NIR |