Bipolar Cell Type-Specific Expression and Conductance of Alpha-7 Nicotinic Acetylcholine Receptors in the Mouse Retina

Purpose Motion detection is performed by a unique neural network in the mouse retina. Starburst amacrine cells (SACs), which release acetylcholine and gamma-aminobutyric acid (GABA) into the network, are key neurons in the motion detection pathway. Although GABA contributions to the network have been extensively studied, the role of acetylcholine is minimally understood. Acetylcholine receptors are present in a subset of bipolar, amacrine, and ganglion cells. We focused on α7-nicotinic acetylcholine receptor (α7-nAChR) expression in bipolar cells, and investigated which types of bipolar cells possess α7-nAChRs. Methods Retinal slice sections were prepared from C57BL/6J and Gus8.4-GFP mice. Specific expression of α7-nAChRs in bipolar cells was examined using α-bungarotoxin (αBgTx)-conjugated Alexa dyes co-labeled with specific bipolar cell markers. Whole-cell recordings were conducted from bipolar cells in retinal slice sections. A selective α7-nAChR agonist, PNU282987, was applied by a puff and responses were recorded. Results αBgTx fluorescence was observed primarily in bipolar cell somas. We found that α7-nAChRs were expressed by the majority of type 1, 2, 4, and 7 bipolar cells. Whole-cell recordings revealed that type 2 and 7 bipolar cells depolarized by PNU application. In contrast, α7-nAChRs were not detected in most of type 3, 5, 6, and rod bipolar cells. Conclusions We found that α7-nAChRs are present in bipolar cells in a type-specific manner. Because these bipolar cells provide synaptic inputs to SACs and direction selective ganglion cells, α7-nAChRs may play a role in direction selectivity by modulating these bipolar cells' outputs.

V isual processing begins in the retina, where images are captured by photoreceptors and are processed by the following retinal second-and third-order neurons: bipolar, amacrine, and ganglion cells. Each of these neurons consists of multiple subsets, including approximately 15 types of bipolar cells, 1-5 more than 20 types of amacrine cells, 6,7 and 15 to more than 30 types of ganglion cells. [8][9][10][11] These diverse types of neurons are thought to form parallel neural networks that encode distinct image features.
Parallel neural networks are initiated by a variety of secondorder bipolar cell types. Different types of bipolar cells have unique signaling characteristics [12][13][14][15] and connect to different postsynaptic partners. 3 Additionally, the output of each bipolar cell can be further tuned by inhibition, allowing for a greater diversity of information output. 16,17 Such a wealth of factors underlies the phenomena of parallel processing, whereby unique circuits initiated by bipolar cells carry distinct information, such as color, contrast, or possibly even motion.
Visual motion detection is a vital function in animal survival. Certain retinal neurons are known contributors to retinal motion detection. One such example is the multiple types of direction selective ganglion cells (DSGCs), which are sensitive to objects moving in a particular direction. 18,19 DSGC motion tracking depends on synaptic inputs from starburst amacrine cells (SACs), 20,21 another neuronal class participating in retinal motion detection. SACs receive synaptic inputs from bipolar cells and release both gamma-aminobutyric acid (GABA) and acetylcholine to postsynaptic neurons, which also exhibit direction selectivity. [22][23][24] SACs release GABA on to DSGCs in response to light moving in the SAC's preferred direction, which induces unidirectional excitation in these ganglion cells. 21,[25][26][27] Although GABA release from SACs to DSGCs has been rigorously studied, the role of acetylcholine remains unclear.
Receptors for acetylcholine are broadly distributed in the central nervous system, including the retina. They are classified into two receptor classes as follows: nicotinic and muscarinic receptors. Nicotinic receptors are cationic channels that induce excitation in cells. In contrast, muscarinic receptors are G protein-coupled receptors that signal via G q or G i subunits. Nicotinic acetylcholine receptors are expressed by DSGCs, which receive cholinergic input directly from SACs. [28][29][30] Nicotinic and muscarinic receptors are also present in subsets of bipolar, amacrine, and other ganglion cells in the mammalian retina. 28,[31][32][33] Given that SACs are the only known source of acetylcholine in the retina, these acetylcholine receptorexpressing neurons may also contribute to motion computations.
We focused on a7 nicotinic acetylcholine receptors (a7-nAChRs) because these receptors distribute broadly in the central nervous system, induce depolarization with high calcium permeability, modulate presynaptic transmitter release, and exhibit a high sensitivity for certain toxins, such as abungarotoxin (aBgTx). [34][35][36][37][38] We examined a7-nAChR-specific expression among multiple bipolar cell types using immunostaining methods. Additionally, a7-nAChR conductance was examined using patch-clamp methods. In the present work, we demonstrate that a7-nAChRs are expressed by bipolar cells in a type-specific manner.

Retinal Preparation
Animal protocols were approved by the Institutional Animal Care and Use Committee at Wayne State University. Experiments were performed in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Visual Research. We used wild-type mice (C57BL/6J; Jackson Laboratory, Bar Harbor, ME, USA) and Gus8.4-GFP mice (gift from Robert Margolskee 39 ). The experimental techniques were similar to those described previously. 15,40 Briefly, mice (28-60 days of age) were euthanized using carbon dioxide and cervical dislocation. Using a dissecting microscope, the retina was isolated, the retinal slab placed onto a piece of filter membrane (HABG01300; Millipore, Burlington, MA, USA) and cut into slice preparations (250-lm thick) using a hand-made chopper. Retinal dissections were performed in HEPES-buffered solution containing the following (in mM): 115 NaCl, 2.5 KCl, 2.5 CaCl 2 , 1.0 MgCl 2 , 10 HEPES, 28 glucose, adjusted to pH 7.38 by NaOH. The dissection medium was cooled and continuously oxygenated. The retinal preparations were stored in an oxygenated dark box at room temperature.

Whole-Cell Recordings and Puff Application
Whole-cell patch recordings were made from bipolar cell somas in retinal slices by viewing them with an upright microscope (Slicescope Pro 2000; Scientifica, Uckfield, East Sussex, UK) equipped with a charged-couple device camera (Retiga-2000R; Q-Imaging, Surrey, BC, Canada). Recordings were performed in Ames' medium buffered with NaHCO 3 (Sigma, St. Louis, MO, USA), which was continuously bubbled with 95% O 2 and 5% CO 2 ; the pH was 7.4 at 308C. Electrodes were pulled from borosilicate glass (1B150F-4; World Precision Instruments, Sarasota, FL, USA) with a P1000 Micropipette Puller (Sutter Instruments, Novato, CA, USA) and had resistances of 8 to 12 MX. The intracellular solution contained the following (in mM): 111 K-gluconate, 1.0 CaCl 2 , 10 HEPES, 10 EGTA, 10 NaCl, 1.0 MgCl 2 , 5 ATP-Mg, and 1.0 GTP-Na, adjusted to pH 7.2 with CsOH. Liquid junction potentials were corrected after each recording. Clampex and Multi Clamp 700B (Molecular Devices, San Jose, CA, USA) were used to acquire data and to control puff application using a Picospritzer III (Parker Hannifin, Cleveland, OH, USA). A selective a7-nAChR agonist, PNU282987 (30 lM; Tocris, Bristol, UK) was puffed onto recording bipolar cell axon terminals. The data were digitized and stored with a computer using Axon Digidata 1440A (Molecular Devices). Responses were filtered at 1 kHz with the four-pole Bessel filter and sampled at 2 to 5 kHz.
A fluorescent dye, sulforhodamine B (0.005%; Sigma), and Neurobiotin (NB; 0.5%; Vector Lab, Burlingame, CA, USA) were included in the recording pipette; these dyes did not affect the physiological recordings. 15 Sulforhodamine B images were observed after physiological recordings. To visualize NB staining, the slice preparation was fixed with 4% paraformaldehyde for 30 minutes, incubated with streptavidin-conjugated Alexa 488 (1:200; Life Technologies, Carlsbad, CA, USA) and anti-ChAT antibody (1:200; Millipore) overnight, and then incubated with the secondary antibody for 2 hours at room temperature. The preparation was viewed with a confocal microscope (TCS SP8; Leica, Wetzlar, Hesse, Germany) using a water-immersion, 363 objective. Bipolar cell types were determined based on previous references. 1,15 Immunohistochemistry Using live retinal slice sections, either aBgTx -conjugated Alexa 488 or 555 (1:100) was applied for 1 hour, followed by several rinses with HEPES buffer solution, and fixation using 4% paraformaldehyde for 30 minutes. After several washes in 0.1 M PBS, sections were blocked in a solution containing 10% normal donkey serum (NDS) and 0.5% Triton X-100 in PBS for 1 hour at room temperature. Primary antibodies (Table 1) were diluted in 3% NDS and 0.5% Triton X-100 in PBS. Sections were incubated with the primary antibody overnight at room temperature, and then incubated with a secondary antibody conjugated with Alexa dyes (Thermo Fisher Scientific, Waltham, MA, USA) for 2 hours. The preparation was viewed with a confocal microscope (TCS SP8; Leica) using 363 water immersion objectives. The z-step for stack images was 0.3 lm.

Western Blot
Retinal tissue was lysed by sonication in radioimmunoprecipitation assay buffer with a protease inhibitor cocktail (P8340; Sigma) and was sonicated followed by centrifugation. Total protein samples (40 ug) were run on a 12% SDS-PAGE in Trisglycine-SDS buffer and then electroblotted onto a nitrocellulose membrane (BioRad, Hercules, CA, USA). After blocking with 5% BSA in Tris-buffered saline containing 0.05% Tween 20 and 5% milk (TBS-T) for 1 hour, the membrane was probed with the a7-nAChR primary antibody (1:1000) or primary antibody (1:1000) incubated with blocking peptide (1:200) for 30 minutes in 3% BSA in TBS-T overnight at 48C. Then, the membrane was incubated with a horseradish peroxidase secondary antibody (1:2000) diluted in TBS-T with 5% milk at room temperature for 2 hours. Bands were visualized using a Kodak 4000R Pro Molecular Imaging System (Carestream Health Inc., Rochester, NY, USA).

Data Analysis
Using AutoQuant X3 and Image-Pro Premier 3D software (Media Cybernetics, Rockville, MD, USA), we conducted image deconvolution and three-dimensional alignment. ChAT bands in stack images were realigned and NB-filled bipolar cells were analyzed. We analyzed colocalization of bipolar cell markers with aBgTx fluorescence or a7-nAChR antibody by examining each soma in each digital section (0.3 lm). A stained cell was categorized as a bipolar cell if the cell possesses dendrites.

RESULTS
We conducted immunohistochemistry to examine a7-nAChR expression in each type of bipolar cell, and patch clamp recordings to examine a7-nAChR conductance in bipolar cells.

a-BgTx Fluorescence and a7 Antibody Staining
Using retinal slice sections, we labeled a7-nAChR-expressing cells by aBgTx-conjugated Alexa 488 or Alexa 555 application. Many somas in the inner nuclear layer (INL) and ganglion cell layer were labeled (Figs. 1A, 1B), suggesting that a subset of bipolar, amacrine, and ganglion cells express a7-nAChRs. Both Alexa dyes labeled the same sets of cells ( Fig. 1C; 78/78 cells colocalized, n ¼ 3 mice). aBgTx is highly specific for a7-nAChRs in the mammalian retina 41,42 ; however, it may also bind to the b3 GABA A receptor subunit. 43,44 Therefore, we used an antibody against a7-nAChRs to investigate whether aBgTx-stained cells are specific to a7-nAChRs. We first authenticated the a7-nAChR antibody (Table 1)   58 kDa, which matches the known molecular weight of the splice variant a7-2. 46 When preabsorbing the antibody with the peptide antigen, no bands were observed, which authenticated the antibody (lane 2, Fig. 1D). Then, we co-applied aBgTx and the a7-nAChR antibody to retinal slice preparations (n ¼ 3 mice, Fig. 1E). Almost all cells demonstrating aBgTx fluorescence also expressed a7-nAChRs (192/194, 99.0%), confirming that aBgTx cells express a7-nAChRs. We used aBgTx conjugated with Alexa dyes for subsequent studies.

OFF Transient Bipolar Cells
Previously, we investigated the physiological signaling properties of each type of bipolar cell and found that some bipolar cells exhibit transient signaling while others exhibit slower, sustained signaling. 14 We thought that transient bipolar cells might be important for detecting motion stimuli and may express a7-nAChRs. We first examined type 2, 3a, and 3b bipolar cells, which constitute the kinetically transient neurons among the five types of OFF bipolar cells. 14 OFF bipolar cells are characterized by their axonal ramification processes close to the OFF-ChAT band. We used specific markers for these bipolar cells, including synaptotagmin 2 (Syt2) for type 2, HCN4 for type 3a, and PKAIIb for type 3b. 2,48 Of Syt2-expressing type 2 bipolar cells, 82% were labeled with aBgTx fluorescence ( Fig.  2A, n ¼ 5 mice, 32/39 cells were positive). We also conducted whole-cell recordings to examine whether these receptors depolarized the cell after application of a specific a7-nAChR agonist, PNU282987. Type 2 cells depolarized in response to a PNU282987 puff application onto their axon terminals (Fig. 2B, 3.6 6 0.5 mV, 5/6 cells depolarized). To verify these responses were attributable to direct activation of a7-nAChRs, we applied CoCl 2 (4 mM) in the bath solution, which eliminates synaptic transmission. After disrupting retinal network effects with CoCl 2, we observed that type 2 bipolar cells were still depolarized by PNU application. Taken together, these results indicate that type 2 bipolar cells possess active a7-nAChRs.

ON Transient Bipolar Cells
Similar to OFF bipolar cells, ON bipolar cells also exhibit either transient or sustained signaling kinetics in a type-dependent manner. 15 A majority of type 5 and 7 bipolar cells are transient and ramify close to the ON ChAT band. Given this relationship between type 5 and 7 bipolar cell axon terminals and ON-SAC processes, we hypothesized that type 5 and/or 7 cells would respond to a cholinergic stimulus. We used the Gus-GFP mice in order to identify type 7 bipolar cells. 2,49 Type 5 bipolar cells comprise multiple subsets, 2,3,13,50,51 and there are no appropriate markers available for type 5 subsets. Therefore, we conducted whole-cell recordings to examine the physiological properties of a7-nAChRs in these cells. A majority of type 5 bipolar cells did not respond to PNU282987 puff application, indicating that these type 5 cells do not possess a7-nAChRs (Fig.  3A). However, three of 22 cells depolarized in response to PNU (Fig. 3B). Axon terminals of these cells were relatively narrow (12.5 6 0.3 lm) and were not XBC-like. 3,13 Furthermore, two of three cells were tested with CoCl 2 application and remained depolarizing in response to PNU, suggesting that a subset of sustained type 5 bipolar cells responded to PNU puff. 15 In contrast, 89% of GFP-expressing type 7 bipolar cells exhibited aBgTx fluorescence (Fig. 3C, n ¼ 5 mice, 55/62 cells were positive). All type 7 bipolar cells depolarized in response to the PNU puff (Fig. 3D, 4.0 6 0.7 mV, 6/6 cells depolarized). These results indicate that type 7 bipolar cells possess functional a7-nAChRs.

Sustained Bipolar Cells
We examined four different types of sustained bipolar cells, type 1, 4, 6, and rod bipolar cells (RBCs). We used NK3R and calsenilin (Csen) antibodies to label type 1 and 4 OFF bipolar cells, respectively. 52,53 NK3R antibody labels both type 1 and 2 bipolar cells. 53 We used NK3R and Syt2, a specific type 2 marker, antibodies to identify type 1 cells. A majority of type 1 bipolar cells demonstrated aBgTx fluorescence (Fig. 4A,  Type 1 and 4 cells were rarely encountered during patch-clamp recording; therefore, we cannot demonstrate whether these cells are responsive to PNU puff. Syt2 is a marker for type 6 bipolar cells; however, it labels only type 6 axon terminals, not somas. 2,54 Syt2 is therefore not suitable for detecting a7-nAChR-expression in type 6 cells. Given this information, we performed patch-clamp recordings with NB staining to examine whether type 6 cells possess a7-nAChRs. A majority of type 6 cells did not respond to PNU puff (Fig. 4C, 6/7 cells were negative), suggesting that type 6 cells do not possess a7-nAChRs.
PKCa labels rod bipolar cells (RBCs), 55 which did not demonstrate aBgTx fluorescence (Fig. 4D, n ¼ 2 mice, 65/65 cells were negative). In addition, whole-cell recordings revealed that none of RBCs depolarized in response to PNU application (0/22 cells depolarized). In almost half of cases, RBCs hyperpolarized in response to PNU puff (10/22 hyperpolarized). This response was blocked with CoCl 2 , suggesting that the hyperpolarization was the consequence of surrounding amacrine cell excitation by PNU puff rather than direct responses of RBCs. 56 These results indicate that RBCs do not possess a7-nAChRs.
We could not investigate a7-nAChRs in type 8 and 9 bipolar cells because the fluorescent marker in clomeleon-1 mice, which highlight type 8 and 9 cells, 2 fades out over multiple generations. 54 Furthermore, we did not encounter these cells by patch-clamp recordings because of their low density compared with other bipolar cell subtypes. 2 Altogether, we found that a subset of bipolar cells express a7-nAChRs, which are summarized in Table 2.

DISCUSSION
We examined the a7-nAChR expression in nine types of bipolar cells by immunostaining, and measured physiological activity by measuring conductance of a7-nAChRs. As shown in Table 2, type 2 and 7 bipolar cells exhibited a7-nAChRs expression and conductance, as demonstrated through immunostaining and whole-cell recording, respectively. Additionally, immunostaining revealed that the majority of type 1 and 4 cells showed fluorescence when stained with aBgTx-conjugated Alexa dye, demonstrating that these cells express a7-nAChRs. We therefore conclude that four types of bipolar cells possess a7-nAChRs. In contrast, type 3a, 3b, 5, 6, and RBCs did not exhibit a7-nAChRs. Although a small portion of type 3a cells colocalized with aBgTx fluorescence (Fig. 2C), none of the type 3 cells responded to PNU puff application. The majority of type 5 cells did not respond to PNU puff. However, a small type 5 population responded to PNU puff (3/22 cells), which could represent a subset of type 5 cells. 2,3,13, 50 We ultimately concluded that the majority of type 3a, 3b, 5, 6, and rod bipolar cells do not possess a7-nAChRs.
aBgTx has a high affinity for a7-nAChRs; however, aBgTx also has an affinity for other homomeric a8 or a9 nAChRs. 57 Although the a8 nAChR isoform is uniquely expressed in avian retina 42,58-60 and a9 is mainly expressed in cochlea and vestibular organs, 42 we cannot rule out the possibility of a8 or a9 nAChR expression in the mouse retina. We therefore conducted PNU application to verify the a7 expression in a subset of bipolar cells.
We demonstrated that both aBgTx-conjugated Alexa dye and the a7 antibody visualize somas as well as axons of bipolar cells, consistent with previous findings from Keyser's group. 32,33 We used live-tissue preparation for aBgTx staining, indicating that a7 nAChRs are expressed from the axons to the somas of bipolar cells. This diffuse expression suggests that  either synaptic or paracrine transmission could occur from SACs to bipolar cells, consistent with previous reports. 29,30,32 This may explain the expression of a7-nAChRs in type 1 bipolar cells, which do not have an intimate relationship with the cholinergic plexus like other a7-expressing cells. 14 Expression of a7-nAChR as well as other nicotinic and muscarinic receptors have also been shown in subsets of bipolar cells. 4,32,33,61 In the present study, we examined the bipolar cell types that possess a7-nAChRs to elucidate the function of these receptors.

Significance of a7-nAChRs in Retinal Bipolar Cells
Located in the midretina, bipolar cells relay visual signals from photoreceptors to ganglion cells, the input and output neurons of the retina, respectively. Bipolar cell types are identified by their axon terminal ramification patterns in the inner plexiform layer (IPL), where synaptic interactions among bipolar, amacrine, and ganglion cells occur. Each type of bipolar cell is thought to extract different components of visual signals from photoreceptors, such as color and motion, and pass these components to discrete retinal circuits, forming parallel processing pathways. [62][63][64] Signal transmission from bipolar cells to retinal ganglion cells is then modulated either by feedback or feedforward effects of amacrine cells.
In addition to unique axonal ramification patterns, bipolar cell types exhibit molecular and physiological differences. Wässle et al. 2 established that each type of bipolar cell can be recognized by unique markers, such as antibodies or transgenic mouse lines. Furthermore, Shekhar et al. 4 used single-cell transcriptomics to show that each type of bipolar cell expresses different sets of genes. They further found that various kinds of voltage-gated channels, receptors, transcription factors, and adhesion molecules are uniquely expressed by each type of bipolar cell. Furthermore, we found that each type of cone bipolar cell exhibits unique temporal properties in response to sinusoidal light stimuli. 14,15 Similarly, Baden et al. 12 and Borghuis et al. 65 used fluorescence imaging tools to show that temporal responses and spike generation differ among bipolar cells ramifying in different layers of the IPL. These reports suggest that each bipolar cell responds differently to certain types of images, and thus, transfers diverse kinds of visual signals to different neural networks in the IPL.
Our observation that a subset of retinal bipolar cells express a7-nAChRs further adds to the diversity of factors modulating bipolar cell output. a7-nAChRs exhibit high calcium permeability, which excites cells and could regulate neurotransmitter release from axon terminals. 66 It has been demonstrated that glutamatergic neurons throughout the central nervous system express a7-nAChRs, which facilitate glutamate release from presynaptic axon terminals. [34][35][36]38,67 Furthermore, presynaptic a7-nAChRs facilitate other types of neurotransmitter release 37,68,69 through direct passage of calcium into the presynaptic terminal. 70,71 It may also be possible that these channels activate nearby voltage-gated calcium channels. 68 The expression of a7-nAChRs in glutamatergic bipolar cells therefore suggests a similar role in facilitating glutamate release to postsynaptic neurons.

Implications for the Role of a7-nAChR-Expressing Bipolar Cells in the Retinal Neural Network
GABAergic and glycinergic amacrine cells have been shown to modulate bipolar cell outputs via presynaptic inhibition, thereby increasing bipolar cell functional diversity. 17 Conversely, presynaptic bipolar cell expression of a7-nAChRs suggests another possible mechanism for directly altering bipolar cell outputs via feedback amacrine cell excitation. SACs, which release acetylcholine in response to preferred directional stimulation, are the only cholinergic neurons in the retina. 72 Thus, neurons that receive acetylcholine inputs, such as a7-nAChR-containing bipolar cells, may be important for retinal motion detection, especially given previous reports demonstrating that SACs and DSGCs receive synaptic inputs from particular types of bipolar cells, including type 5 bipolar cells. 3,73 Previous reports, however, demonstrate that bipolar cells themselves do not exhibit direction selectivity. [73][74][75][76] These results have led to a consensus that bipolar cells do not play a major role in retinal direction selectivity.
Nevertheless, this assumption was recently challenged via connectomic data that revealed how varying bipolar cell synaptic partners, as well as the arrangement of bipolar cell types within motion circuits, may contribute to retinal motion detection. These studies revealed that different types of bipolar cells not only carry visual signals to different IPL layers, but also make divergent synaptic connections to postsynaptic neurons. For example, multiple sets of type 5 bipolar cells ramify near the ON-ChAT band within sublamina 3 13 ; however, only one subset of type 5 provides synaptic inputs to DSGCs. Conversely, XBC-type 5 cells do not provide synaptic inputs to DSGCs. 3 In the case of SACs, different types of bipolar cells provide synaptic inputs to various locations along their dendrites. 51,77,78 Our results indicate that a7-nAChRs are expressed by type 1, 2, 4, and 7 bipolar cells. Interestingly, type 4 bipolar cells are the main ''OFF'' inputs to ON/OFF DSGCs. 3 Furthermore, the other three types of a7-nAChRexpressing bipolar cells provide synaptic inputs to SACs at proximal dendrites. 51,77,78 In conjunction with connectomic studies, our results imply that particular types of bipolar cells play a pivotal role in motion detection.
SACs release both GABA and acetylcholine in response to a moving object. The contribution of acetylcholine to retinal motion detection is highly debated, with some groups suggesting acetylcholine contributes almost half of DSGC excitation while other groups find that glutamate instead predominantly provides DSGC excitation. [28][29][30]75,[79][80][81] Although the functional outcome of a7-nAChRs in bipolar cells to direction selectivity in the third-order neurons will require future investigation, our new findings may contribute to the elucidation of acetylcholine involvement in motion detection circuit.