Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 Jun 15.
Published in final edited form as: J Comp Neurol. 2019 Dec 27;528(9):1588–1598. doi: 10.1002/cne.24840

Wide-field amacrine cell inputs to ON parasol ganglion cells in macaque retina

Sara S Patterson 1,2,*, Andrea S Bordt 3,*, Rebecca J Girresch 4, Conor M Linehan 1, Jacob Bauss 4, Eunice Yeo 4, Diego Perez 3, Luke Tseng 3, Sriram Navuluri 3, Nicole Barnes Harris 3, Chaiss Matthews 3, James R Anderson 5, James A Kuchenbecker 1, Michael B Manookin 1, Judith Mosinger Ogilvie 4, Jay Neitz 1, David W Marshak 3
PMCID: PMC7153979  NIHMSID: NIHMS1546334  PMID: 31845339

Abstract

Parasol cells are one of the major types of primate retinal ganglion cells. The goal of this study was to describe the synaptic inputs that shape the light responses of the ON type of parasol cells, which are excited by increments in light intensity. A connectome from central macaque retina was generated by serial blockface scanning electron microscopy. Six neighboring ON parasol cells were reconstructed, and their synaptic inputs were analyzed. On average, they received 21% of their input from bipolar cells, excitatory local circuit neurons receiving input from cones. The majority of their input was from amacrine cells, local circuit neurons of the inner retina that are typically inhibitory. Their contributions to the neural circuit providing input to parasol cells are not well-understood, and the focus of this study was on the presynaptic wide-field amacrine cells, which provided 17% of the input to ON parasol cells. These are GABAergic amacrine cells with long, relatively straight dendrites, and sometimes also axons, that run in a single, narrow stratum of the inner plexiform layer. The presynaptic wide-field amacrine cells were reconstructed, and two types were identified based on their characteristic morphology. One presynaptic amacrine cell was identified as semilunar type 2, a polyaxonal cell that is electrically coupled to ON parasol cells. A second amacrine was identified as wiry type 2, a type known to be sensitive to motion. These inputs likely make ON parasol cells more sensitive to stimuli that are rapidly changing outside their classical receptive fields.

Keywords: vision, motion sensitivity, interneuron, primate, gamma amino butyric acid, electron microscopy, connectomics RRID:SCR_003584, RRID:SCR_005986, RRID: SCR_001622, RRID: SCR_017350, RRID:SCR_013654

Graphical Abstract

Parasol cells are one of the major types of retinal ganglion cells, and they give rise to the magnocellular stream of visual information processing. They respond well to moving and other rapidly changing stimuli in their receptive field. They also respond to stimuli outside of their classical receptive fields, and this paper is about the inhibitory local circuits that mediate those responses, wide-field amacrine cells. The parasol cells (orange, light blue, yellow and pink) were reconstructed from a connectome of central macaque retina, as were dendrites of the presynaptic wide field amacrine cells. Three wide field amacrine cells were wiry type 2 cells (dark blue, salmon and green), amacrine cells with active dendrites. The locations of their synapses onto parasol cells are circled.

graphic file with name nihms-1546334-f0001.jpg

Introduction

Parasol cells were named for their distinctive morphology, and they comprise approximately 10% of retinal ganglion cells. They are more sensitive to luminance contrast than the more numerous midget ganglion cells, and they respond more transiently to light stimuli at all levels of adaptation (reviewed by Rodieck, 1998). These properties have been proposed to enable parasol cells to respond preferentially to moving stimuli, and a recent study demonstrated directly that parasol cells are sensitive to small stimuli making cohesive movements within their receptive fields (Manookin, Patterson, & Linehan, 2018). Many of the differences between the light responses of parasol cells and midget ganglion cells are attributable to the types of local circuit neurons that provide their input (Abbott, Percival, Martin, & Grunert, 2012). The focus of this study was the synaptic inputs to ON type parasol cells. These respond with an increase in firing rate to increments of light in the center of their receptive fields, and they ramify in the fourth stratum (S4) of the inner plexiform layer (IPL) (Dacey & Lee, 1994).

The bipolar cells that provide excitatory input to ON parasol cells have recently been identified by electron microscopy. Diffuse bipolar (DB) cells, which receive input from multiple cones, provide the major input, but midget bipolar cells, which receive input from a single cone in central retina, also provide some input. Type DB4 cells provide the largest number of inputs, and there are also significant numbers of synapses from other bipolar cell types that costratify with ON parasol cells, including: DB5, giant and ON midget (Tsukamoto & Omi, 2016). Electrical coupling between the presynaptic bipolar cells promotes motion sensitivity in the receptive field centers of parasol cells. Having been depolarized in advance via electrical synapses with neighbors, the bipolar cell is more likely to release glutamate when a stimulus enters its receptive field center (Manookin et al., 2018).

Amacrine cells provide the majority of inputs to ON parasol cells (Jacoby, Stafford, Kouyama, & Marshak, 1996; Marshak, Yamada, Bordt, & Perryman, 2002). These are a diverse group of interneurons that are generally inhibitory, but they also include at least two types that release an excitatory co-transmitter (Marshak, 2016). Amacrine cells are clearly important for processing visual information in primate retina because they make the majority of the synapses in the IPL, and their most common targets are other amacrine cells (Koontz & Hendrickson, 1987). The specific types of amacrine cells that provide input to ON parasol cells have only been identified in light microscopic immunolabeling studies, and it is uncertain whether the contacts observed were actually synapses and not simply appositions. Moreover, there are amacrine cells known to costratify with ON parasol cells, but it is uncertain whether they make synapses onto the parasol cells.

The focus of many physiological studies of parasol cells has been on the bipolar cell input. The classical receptive field centers and antagonistic surrounds of parasol cells can be explained, in large part, by synaptic interactions in the outer plexiform layer (OPL) (Davenport, Detwiler, & Dacey, 2008; McMahon, Packer, & Dacey, 2004). However, more recent studies suggest that amacrine cells also contribute to the light responses of parasol cells (Cafaro & Rieke, 2010, 2013; Crook, Packer, & Dacey, 2014). Relatively simple models of the input pathway have been successful in accounting for the responses of parasol cells to artificial stimuli, but two recent studies have found that these do not account for the responses of parasol cells to natural stimuli (Heitman et al., 2016) (Turner & Rieke, 2016). They also do not account for the responses of parasol cells to stimuli outside their classical receptive fields (Greschner et al., 2016; Solomon, Lee, & Sun, 2006). Taken together, these findings suggest the light responses of parasol cells have major contributions from amacrine cells, but these are not well understood.

The goal of this study was to definitively identify the amacrine cells presynaptic to ON parasol cells using connectomics at electron microscopic resolution. The focus was on the wide - field amacrine cells, which convey signals from other regions of the retina and confer sensitivity to global changes in the visual field.

Material and Methods

Tissue Preparation

An eye was obtained from a terminally anesthetized adult male macaque (Macaca nemestrina) through the Tissue Distribution Program at the Washington National Primate Center. All procedures were approved by the Institutional Animal Care and Use Committee at the University of Washington. The anterior half of the eye and the vitreous humor were removed, and the eyecup was fixed and processed as previously described (Patterson et al., 2019). Briefly, the eyecup was fixed in 4% glutaraldehyde in 0.1M sodium cacodylate buffer pH 7.4 and, while in this solution, a 1 mm square of retina centered 2 mm temporal to the center of the fovea was cut out and then fixed overnight at 4°C. At this eccentricity, the edge of the foveal slope, the neurons were small enough to be studied in serial sections, but the displacement of retinal ganglion cells from cone pedicles was minimized.

The tissue was washed, postfixed in osmium ferrocyanide, rinsed and placed in 10% thiocarbohydrazide solution for 20 min at room temperature (RT).. The tissue was incubated in 2% aqueous OsO4 for 30 min at RT, washed again and stained in 1% aqueous uranyl acetate overnight at 4 °C. It was washed and stained with Walton’s lead aspartate at 60 °C for 30 min. After final washes, the tissue was dehydrated and embedded in Durcupan.

Microscopy

The tissue was sectioned in the horizontal plane and imaged using a Zeiss Sigma VP field emission scanning electron microscope equipped with a 3View system and a detector for backscattered electrons (Gatan, Inc.). In each 70 nm section, an area approximately 200 μm on a side was imaged at a resolution of 7.5 nm/pixel. The connectome contained 937 sections, spanning from the ganglion cell layer (GCL) to the OPL or inner nuclear layer (INL), depending on eccentricity. Image registration was performed using Nornir (http://nornir.github.io RRID:SCR_003584). There were some regions with poor image quality, likely due to charging of the surface of the block. There were also a few instances in which the section was not completely removed from the blockface, obscuring parts of the image, and approximately half of section 470 was not imaged. However, parasol cell dendrites and many processes of wide-field amacrine cells could be followed in spite of these artifacts.

Annotation

The major cell types were identified by their characteristic ultrastructure (Dowling and Boycott, 1966; Tsukamoto and Omi, 2015, 2016). Axon terminals of bipolar cells had abundant synaptic vesicles and synaptic ribbons. Processes of amacrine cells contained fewer synaptic vesicles, and they were typically clustered at synapses. Ganglion cell dendrites lacked presynaptic specializations. The criterion for identifying a synapse was that specialized membranes and clustered vesicles or synaptic ribbons were found in two consecutive sections. The serial EM volumes were annotated using the web-based, multiuser Viking software described previously (Anderson et al., 2011;http://connectomes.utah.edu RRID:SCR_005986.) Briefly, processes were annotated by placing a circular disc with the same diameter as the process at its center of mass and linking it to annotations on neighboring sections. Synapses were annotated with lines connected by 2–3 control points and linked to the parent neuron. The boundary between the INL and the IPL was designated as 0% and the IPL-GCL boundary as 100% depth.

The strategy was to first reconstruct the largest somas in the most sclerad row of the GCL, which were likely to belong to parasol cells, and, if the dendrites ramified in S4, to annotate the cells as completely as possible. Somas near the middle of the connectome were selected so that most or all of the dendrites could be annotated. The working hypothesis was that the responses of ON parasol cells to stimuli outside the classical receptive field were mediated by multiple types of wide-field amacrine cells. Accordingly,the smallest somas in the same row of the GCL, likely to be those of wide-field amacrine cells, were annotated, and those with processes in S4 were annotated as completely as possible. Finally, all of the amacrine cells and bipolar cells presynaptic to six ON parasol cells were annotated. The bipolar cell axons, identified by their synaptic ribbons, and dendrites of narrow-field amacrine cells, identified by their frequent branching and vertical orientation, were partially annotated. The presynaptic wide-field amacrine cells, identified by their long, narrowly-stratified dendrites, were annotated as completely as possible.

Data Analysis

Data analysis and 3D rendering were performed using an open-source Matlab (Mathworks, RRID: SCR_001622) program https://github.com/neitzlab/sbfsem-toolsRRID: SCR_017350. The image rendering was performed using the RenderApp function. A triangle mesh was built by rendering segments of connected annotations as rotated cylinders centered at each annotation’s X, Y and Z coordinates and scaled by their radii (Bordt et al., 2019). Using the synapseSphere function, synapses were rendered as unit spheres centered at each synapse annotation’s X, Y and Z coordinates then scaled to optimize visibility. For the two most completely annotated ON parasol cells, the distances between bipolar cell inputs and the nearest wide- and narrow-field amacrine cell input were calculated as the Euclidean distance between the coordinates of the post-ribbon synapse annotation and the post-conventional synapse annotation. Only synapses on the same branch of the ganglion cell dendrite were included in the analysis. The data and code used to generate the figures with SBFSEM-tools is available at https://github.com/neitzlab/SynapticInputsToOnParasolCells.

Dendrites of the wide-field amacrine cells and ON parasol cells were analyzed using the DendriteDiameter and IPLDepth functions. In total, 258 INL-IPL and 453 GCL-INL boundary markers were placed. INL-IPL, and GCL-IPL boundary surfaces were fit to the X, Y and Z coordinates of each boundary marker type using bicubic interpolation. Given an annotation’s X, Y coordinates, the surfaces supplied the Z coordinates of the IPL boundaries at that X,Y location. The annotation’s Z coordinate relative to the Z coordinates of each boundary could be calculated to determine percent IPL depth. In this way, IPL depth was calculated for each annotation individually. Dendritic diameter was calculated from the size of each annotation associated with a neuron. The code and data used to generate the figures in this study will be made available on GitHub upon publication.

Figures

Figures were prepared using Adobe Photoshop CS6, Tulip http://tulip.labri.fr RRID:SCR_013654 and SBFSEM-tools. The color palette was selected so that the cells and synapses could be distinguished by individuals with all of the common forms of color blindness http://mkweb.bcgsc.ca/colorblind/.

Results

Six ON parasol cells were reconstructed and identified by their large somas and dense dendritic arbors in S4 (Figure 1 and Table 1). In order to determine whether the cells were nearest neighbors, the ratio of dendritic field diameter to inter-cellular spacing was calculated. The mean spacing between the centers of the somas was 65.3 μm, and the two completely annotated ON parasol cells, 189 and 121, had dendritic arbors approximately 100 μm in diameter. The ratio of the two was 1:0.65, very close to the ratio determined in a study using tracer coupling to label neighboring ON parasol cells, 1:0.62 (Dacey & Brace, 1992).

Figure 1.

Figure 1.

Open in a new tab

Six neighboring ON parasol cells were reconstructed. Note the regular spacing of their somas and the overlap of their dendrites. In some areas, dendrites from three ON parasol cells were intermingled. ON parasol cells 121 and 189 were studied most extensively because they were located in the center of the group and completely annotated. Approximate boundaries of the connectome are indicated in white. Scale bar = 20 μm.

Table 1.

Stratification of ON parasol dendrites in the IPL. The boundaries of the IPL were annotated using Viking software. The IPLdepth function of SBFSEM-tools was used to characterize the distribution of dendrites from the six ON parasol cells. The INL-IPL boundary was designated 0% and the IPL-GCL border 100%. The depth profiles of the six cells were very similar and consistent with previous descriptions of ON parasol cell dendrites arborizing in stratum 4 of the IPL.

ON Parasol Cell Depth Mean Depth Median Depth Mode Standard Deviation
189 0.691 0.684 0.716 0.091
121 0.693 0.677 0.696 0.099
207 0.649 0.635 0.636 0.1
2234 0.647 0.635 0.654 0.094
5 0.664 0.646 0.654 0.086
126 0.678 0.66 0.678 0.074
Open in a new tab

Because smooth monostratified cells and parasol cells branch in the same stratum of the IPL, it was important to distinguish the two types. The dendritic arbor diameters of ON parasol cells 189 and 121 fell within the range reported in earlier studies of parasol cells at this eccentricity of the temporal macaque retina (Watanabe and Rodieck, 1989). The dendritic arbors of smooth monostratified cells are considerably larger than those of parasol cells (Crook et al., 2008; Yamada, Bordt, & Marshak, 2005).

The stratification patterns of the six sets of ON parasol cell dendrites in the IPL are summarized in Table 1. The mean values ranged from 65% to 69% depth, as reported previously (Jacoby et al., 1996). The standard deviations provided a rough indication of the breadth of the dendritic arbor; these ranged from 7–10%. The distributions (not illustrated) were also very similar in shape. They were skewed slightly to the right, as expected because the dendrites arise from the GCL (100%).

The resolution was sufficient to identify synaptic contacts with clustered vesicles or synaptic ribbons on the presynaptic side and postsynaptic densities. The synaptic inputs to the six ON parasol cells are summarized in Table 2. All six cells received a majority of their synaptic inputs from narrow-field or medium-field amacrine cells. The proportion of inputs from those amacrine cells ranged from 51% of the total for cell 207 to 70% for cell 126. There were always fewer inputs from dendrites of wide-field amacrine cells than from the other types; these ranged from 14% of the total for cell 189 to 22% for cell 5. The proportion of input from bipolar cells also varied considerably; it ranged from 9% for cell 5 to 30% for cell 207. When the entire sample of synapses onto the six ON parasol cells was considered together, 21% (599/2818) of the input was from bipolar cells, 17% (486/2818) was from wide-field cells, and 62% (1733/2818) was from other amacrine cells.

Table 2.

Synaptic inputs to ON parasol cells. Cells 189 and 121 were located in the center of the connectome, and all of their dendrites and synaptic inputs were annotated. The wide-field amacrine cells presynaptic to cells 189 and 121 were annotated as completely as possible. Each had greater than 100 annotations and typically many more, and they were all rendered using the RenderApp function in SBFSEM-tools to confirm their identity. Other amacrine and bipolar cells were annotated sufficiently to determine that they were unlikely to be wide-field amacrine cells. The dendrites of the other four ON parasol cells were truncated when they reached the edge of the connectome, and the criteria to distinguish the two types of amacrine cells providing their input were more subjective. Nevertheless, the results were very similar, and the two sets of data were analyzed together.

ON parasol cell Bipolar cell inputs Amacrine cell inputs Total
Wide-field Other
189 126 88 426 640
20% 14% 66%
121 149 117 342 608
25% 19% 56%
207 159 96 269 524
30% 18% 51%
2234 98 69 284 451
22% 15% 63%
5 28 69 214 311
9% 22% 69%
126 39 47 198 284
14% 16% 70%
Total 599 486 1733 2818
21% 17% 62%
Open in a new tab

The distributions of the synapses from wide-field amacrine cells and bipolar cells onto cells 121 and 189 are shown in Figure 2. The presynaptic wide-field amacrine cell dendrites were well-annotated, having at least 100 annotations and typically many more. The synapses from wide-field amacrine cells did not appear to be selectively distributed on the dendritic arbors of the parasol cells. Some synapses were found on spines or the distal tips of the dendrites, but the majority were found on more proximal dendrites. The spacing between synapses from wide-field amacrine cells was not regular. The inputs were close together in some instances, and there were some parts of the dendritic arbors that had little or no input from wide-field amacrine cells.

Figure 2.

Figure 2.

Open in a new tab

Synapses of wide-field amacrine cells (upper panel) and bipolar cells (middle panel) onto ON parasol ganglion cells 121 (a) and 189 (b). Their axons are labeled with arrows in the vertical sections (lower panel). Note that the axon of cell 121 meandered into the IPL before descending into the optic fiber layer. The wide-field amacrine cells made 116 synapses (white spheres) onto ON parasol cell 121 (a) and 85 synapses onto cell 189 (b). Scale bars = 20 μm.

There were no bipolar cell inputs onto the somas and very few onto the primary dendrites. The distributions of inputs onto the distal dendrites were more uniform than those of the wide-field amacrine cells. However, there were several instances where two or three inputs from bipolar cells were close together on a parasol cell dendrite. Some bipolar cell inputs were found on spines, but there were many spines without bipolar cell inputs. Using metadata from ON parasol cells 121 and 189, the spatial relationships between the two types of amacrine cell inputs and the nearest bipolar cell inputs to the same dendrite were compared. The data are listed in Supplemental Table 1. The narrow field amacrine cells frequently made synapses onto the ON parasol cell dendrites that were within 2 μm of a bipolar cell input. These comprised 61% of the inputs to cell 121 and 75% of the inputs to cell 189. Many of those synapses, 26% of the inputs to 121 and 35% of the inputs to 189, were within 1 μm of a bipolar cell input. On the other hand, synapses from wide field amacrine cells were typically located farther away from the synapses made by bipolar cells. Only 30% of the inputs from wide-field amacrine cells to cell 121 and 15% of the inputs to cell 189 were found within 2 μm of a bipolar cell input. Wide-field amacrine cells made 37 synapses onto dendritic branches of the two ON parasol cells that had no bipolar cell synapses at all.

Two types of types of wide-field amacrine cells presynaptic to the ON parasol cells were identified based on their morphology (Figure 3). Their somas and many of their dendrites were reconstructed, and they were compared with wide-field amacrine cells described previously. The first presynaptic cell, 4781, was a large polyaxonal amacrine cell with a large, ovoid soma in the GCL. Its dendrites were relatively thick and approximately 200 μm long, with very few branches. It also had several axons that typically arose from the distal tips of the dendrites, but one axon originated from the soma. Both axons and dendrites ramified in S4 of the IPL, most often at 60% of the IPL depth. Based on its morphology, cell 4781was identified as semilunar type 2 (Kolb, Linberg, & Fisher, 1992; Mariani, 1990). A typical synapse from semilunar cell 4781 onto an ON parasol cell dendrite is shown in Figure 4. The dendrites of the semilunar amacrine cell were larger than most, 0.5 μm or more in diameter, and they typically ran parallel to the horizontal sections. Their cytoplasm was less electron dense than that of the parasol cells, but there were other processes that were even more electron lucent.

Figure 3.

Figure 3.

Open in a new tab

Four amacrine cells with somas in the GCL were partially reconstructed and identified based on their morphology. Three wiry type 2 cells (182, 4315 and 4651) had relatively small somas and thin, straight dendrites that rarely branched. The semilunar type 2 cell had a larger soma and dendrites with a larger diameter. It also had axons (arrows) originating from the soma and the distal dendrites. Scale bar = 20 μm.

Figure 4.

Figure 4.

Open in a new tab

Semilunar amacrine cell 4781 (S) made a conventional synapse (arrowheads) onto ON parasol cell 207 (P). Note that the dendrite of 4781 was intermediate in electron density and oriented parallel to the plane of section. Sequential sections at right. Scale bars = 500 nm.

Three other wide-field amacrine cells were presynaptic to ON parasol cells. These had small somas in the GCL and very sparse dendritic arbors in S4 of the IPL. Their dendrites were also found most often at an IPL depth of 60%. They had seven to ten long, relatively straight, thin dendrites that branched only near the soma. Based on their morphology, they were classified as wiry type 2 cells (Kolb et al., 1992; Majumdar, Wassle, Jusuf, & Haverkamp, 2008; Mariani, 1990). A synapse from wiry amacrine cell 4651 onto an ON parasol cell dendrite is shown in Figure 5. Wiry amacrine cell dendrites were typically oriented parallel to the plane of section, but they were distinguishable from those of semilunar cells because they were smaller, typically less than 0.4 μm in diameter, and their cytoplasm was more electron-dense.

Figure 5.

Figure 5.

Open in a new tab

Wiry amacrine cell 4651 (W) was presynaptic to ON parasol cell 207 (P) at a conventional synapse (arrowheads). It was relatively electron dense and ran parallel to the plane of section. Sequential sections at right. Scale bars = 500 nm.

The synapses from the identified amacrine cells onto the ON parasol cells are illustrated in Figure 6. Semilunar type 2 amacrine cell 4781 and the ON parasol cells that received synapse from it are shown in Figure 6a. Many of the synapses onto the ON parasol cells (5/15) were made by dendrites close to the soma of this amacrine cell. One distal dendrite made more synapses (5/15) than the others, which made one or two synapses each. Although the dendrites of the amacrine cell and the ganglion cells overlapped and made frequent contacts, there were only 15 identified synapses. Two long dendrites of the amacrine cell that overlapped with dendrites of annotated ON parasol cells did not make synapses onto those ganglion cells at all. The axons of semilunar amacrine cell 4781 were found in regions of the connectome without dendrites of annotated ON parasol cells or else, presumably, outside the connectome, and they did not make any of the synapses illustrated in Figure 6a. However, they are expected to make inhibitory synapses onto ON parasol cells that were not annotated based on physiological evidence (Greschner et al., 2016).

Figure 6.

Figure 6.

Open in a new tab

a. Semilunar type 2 cell 4781 made 15 synapses (white circles) onto three of the six ON parasol cells. Five of the synapses were made by proximal dendrites, and the remainder were made by distal dendrites. The synapses did not appear to be randomly located; most were found in a relatively narrow swath running from the upper left to the lower right of the connectome. b. Three wiry type 2 amacrine cells made 18 synapses (white circles) onto five of the six ON parasol cells. Two of the wiry amacrine cells, 182 and 4315, made most of their synapses onto one of the ON parasol cells. Wiry cell 4651 did not show this tendency, however. Scale bar = 20 μm.

The synapses from three wiry type 2 amacrine cells onto five of the six annotated ON parasol cells are shown in Figure 6b. Two of the wiry amacrine cells directed a large proportion of their output to one of the parasol cells. Wiry cell 182 made five synapses onto parasol cell 189 and four synapses onto three other parasol cells. Likewise, wiry cell 4315 made four synapses onto parasol cell 5 and one synapse each onto two other parasol cells. However, wiry cell 4651 made two synapses onto parasol cell 207 and one synapse onto one other parasol cell. Given the very high spatial density of wiry amacrine cell somas in the central macaque retina and their wide dendritic arbors (Majumdar et al.,2008), there were likely many more synapses onto the annotated parasol cells from unidentified wiry type 2 cells.

The other synapses made and received by the amacrine cells presynaptic to the six parasol cells were not studied systematically. Nevertheless, there were some indications of synaptic interactions between neurons in the pathway providing input to the parasol cells. There were at least three instances where one of the identified amacrine cells received input from a bipolar cell at the same ribbon synapse as one of the identified parasol cells. Wiry type 2 amacrine cell 4651 received input at dyad synapses like this with ON parasol cells 207 and 121; it did not interact with those bipolar cells nearby. Semilunar type 2 amacrine cell 4781 received input at the same dyad synapse as ON parasol cell 126, and it made a synapse back onto the bipolar cell 6 sections (approximately 0.4 μm) later. Wide field amacrine cells presynaptic to the ON parasol cells also interacted with each other. Figure 7 shows an example, a synapse from wiry type 2 amacrine cell 182 onto an unidentified wide field amacrine cell that made a synapse onto ON parasol cell 5. Based on its ultrastructure, the postsynaptic amacrine cell appeared to be a different type.

Figure 7.

Figure 7.

Open in a new tab

Wiry amacrine cell 182 (W) made a synapse (arrowheads) onto an unidentified wide field amacrine cell. Note that the cytoplasm of the wiry amacrine cell is relatively electron dense, but the cytoplasm of the postsynaptic cell (A) is relatively electron lucent. Scale bar = 1 μm.

Discussion

Using a connectome from central macaque retina, six ON parasol cells were partially or completely reconstructed, and 2818 of their input synapses were identified. On average, ON parasol cells received 21% of their inputs from bipolar cells, but the proportion of inputs from bipolar cells was variable. This is very similar to results reported previously in a study using transmission electron microscopy to study inputs to five ON parasol cells from baboon retina. In that study, bipolar cells provided 20% of the input, on average, but the proportion varied from 13% to 24% (Jacoby et al., 1996). A later study of a peripheral macaque ON parasol cell partially-reconstructed from serial, ultrathin sections found that 13% of the input originated from bipolar cells (Marshak et al., 2002). An OFF parasol cell from peripheral macaque retina studied using similar techniques received 27% of its input from bipolar cells (Bordt, Hoshi, Yamada, Perryman-Stout, & Marshak, 2006). However, a study using light microscopic immunolabeling to identify synapses found that parasol cells from marmoset retina received a larger proportion of their input from bipolar cells (Abbott et al., 2012). The same was true of a parafoveal OFF parasol cell from macaque retina reconstructed from serial, ultrathin sections; it received 58% of its input from bipolar cells (Calkins & Sterling, 2007).

The results of this anatomical study are also consistent with previous studies of ON parasol cells using electrophysiology. ON parasol cells receive tonic, glycinergic inhibitory input from OFF amacrine cells, and as a result, there is both excitation and a decrease in inhibition to a spot of light in the receptive field center. This crossover inhibition from OFF amacrine cells is the predominant form of inhibitory input to ON parasol cells (Cafaro & Rieke, 2013; Crook et al., 2014). Crossover inhibition is typically mediated by narrow-field, glycinergic amacrine cells with dendrites in two or more strata of the IPL (Wassle et al., 2009). The results from this anatomical study indicate that, on average, 78% (1733/2219) of the inhibitory input to ON parasol cells originated from narrow field amacrine cells, and this value was very similar to the proportion of inhibitory input to ON parasol cells mediated by glycine (Crook et al., 2014). Narrow field cells also make those synapses closer to the nearest bipolar cell inputs, on average, and therefore may be more effective in modulating the effects of the excitatory input.

The remainder of the inhibitory input to the six on parasol cells originated from wide field amacrine cells. These are GABAergic amacrine cells, which often have somas in the GCL and have long processes in a single stratum of the IPL (Grunert & Wassle, 1990; Kalloniatis, Marc, & Murry, 1996; Wassle, Grunert, Rohrenbeck, & Boycott, 1989). Together, these wide-field amacrine cells provided 17% of all the synapses onto the ON parasol cells, or 22% of the synapses from amacrine cells. These findings indicate that a significant component of the inhibitory input to ON parasol cells originates from other areas of the retina.

The results are consistent with electrophysiological studies showing that responses of parasol cells are modulated by stimuli outside the classical receptive field, enabling them to convey information about global as well as local stimulus properties. The first of these to be reported was an excitatory “shift effect” in unidentified macaque retinal ganglion cells that, in retrospect, were likely to be parasol cells (Kruger, Fischer, & Barth, 1975). More recently, the effects of peripheral stimulation on macaque parasol cells were studied using extracellular recording (Solomon et al., 2006). An annulus with a large inner diameter produces no responses when presented alone, but it decreases the maintained activity and the responses to a centered, small spot. The effect is rapid and transient, a finding consistent with an origin from spiking amacrine cells. Gratings presented outside the classical receptive field also suppress spiking in parasol cells, possibly by a different mechanism. Other types of primate retinal ganglion cells respond to stimuli outside their classical receptive fields, and these effects are mediated by spiking, GABAergic amacrine cells (Huang & Protti, 2016; Protti et al., 2014).

In addition to the synapses from wide-field amacrine cells reported here, there is evidence for GABAergic synapses onto parasol cells from previous anatomical and physiological studies. Parasol cells in primate retina express GABAA receptors (Abbott et al., 2012; Macri, Martin, & Grunert, 2000), and currents generated by those receptors have been studied in macaque parasol cells using voltage clamp (Crook et al., 2014). The effects of a GABA antagonist, picrotoxin, on the light responses of parasol cells have been studied in macaque retina using intracellular recording (McMahon et al., 2004). Picrotoxin has very little effect on the classical receptive field surround, which is generated largely by horizontal cells in the OPL. However, the responses are altered in more subtle ways by picrotoxin. The latency of the response to the onset of a spot or an annulus is reduced, and the responses to stimulus offset are reduced in size. However, because GABA receptors are found on many types of primate retinal neurons (Grunert, 2000) and picrotoxin is a nonselective antagonist, it was uncertain whether these effects are mediated by direct GABAergic synapses onto parasol cells.

Two types of wide-field amacrine cells presynaptic to ON parasol ganglion cells were identified morphologically. In this respect, ON parasol cells are similar to ON alpha cells in the mouse retina, which receive input from at least two types of GABAergic amacrine cells (Park et al., 2018). Both express corticotropin releasing hormone and costratify with the ON alpha cells and have ON responses to light. One of these has graded responses, and the second is a polyaxonal cell that fires action potentials. Both types contribute to tonic, feedforward inhibition of the ON alpha cells.

The first identified amacrine cell presynaptic to the ON parasol cells was the semilunar type 2, a large polyaxonal amacrine cell that is tracer-coupled to ON parasol cells and contains immunoreactive cholecystokinin precursor (Dacey & Brace, 1992; Kolb et al., 1992; Mariani, 1990; Marshak, Aldrich, Del Valle, & Yamada, 1990). This amacrine cell has a large, ovoid soma in the GCL and relatively thick dendrites with very few branches. The semilunar cells also have several long axons that typically arise from the distal tips of the dendrites, but they also can arise from the soma (Jacoby et al., 1996). However, it was unclear whether semilunar cells also made chemical synapses onto ON parasol cells because antibodies to cholecystokinin precursor label two types of amacrine cells, and the earlier studies were done using light microscopy. This study provides the first direct evidence that dendrites of semilunar cells do, in fact, make chemical synapses onto ON parasol cells.

Light responses from semilunar type 2 amacrine cells have recently been reported (Greschner et al., 2016). They fire action potentials, likely as a result of depolarization via the gap junctions with ON parasol cells. Their action potentials radiate outward in multiple directions, and they make GABAergic inhibitory synapses onto other ON parasol cells. This lateral inhibitory network may increase the sensitivity of ON parasol cells to luminance contrast (Kenyon and Marshak, 1998). The gap junctions with spiking amacrine cells may also promote correlated firing of distant ON parasol cells responding to the same stimulus, as they do in ON alpha ganglion cells of mouse retina (Roy, Kumar, & Bloomfield, 2017).

Wiry type 2 amacrine cells were also presynaptic to ON parasol cells. These had small somas in the GCL and long, straight, thin dendrites that branched only near the soma. The dendritic arbors of wiry amacrine cells are very sparse, but, because the spatial densities of the somas are high and the dendrites are narrowly stratified, they form a very dense plexus of dendrites in the IPL. They ramify the same stratum of the IPL as the dendrites of ON parasol cells and are GABAergic (Majumdar et al., 2008), but synapses from wiry amacrine cells onto parasol cells have not been reported previously. The light responses of these amacrine cells have been recorded in macaque retina (Manookin, Puller, Rieke, Neitz, & Neitz, 2015). Because their dendrites generate N-methyl-D-aspartate spikes, their receptive fields are very large, and they transmit signals over long distances. Wiry type 2 amacrine cells depolarize at the onset of a spot of light, and their responses to full-field flashes have the same fast, transient kinetics as ON parasol cells. However, they respond equally well to moving light or dark objects. The inputs from wiry amacrine cells are expected to remove redundant visual information from the signals of ON parasol cells and to impart motion sensitivity to their responses.

There are many different morphological types of amacrine cells in primates (Mariani, 1990; Kolb et al., 2002), but only a few have been characterized at the level of synaptic interactions between identified populations of neurons, as was done here. However, a potential limitation of this study is that the connectome was built from a single piece of one retina of one animal. This does not take into account the regional differences or the individual variability in retinal anatomy. Nevertheless, this connectome provides a normative dataset of very high quality, and it is accessible to the public on a “Read Only” basis at http://v0152.host.s.uw.edu/Neitz/TemporalMonkey2/SliceToVolume.VikingXML. Because the retinas of humans and macaques are so similar, the data provided will be applicable to studies of human vision and its disorders.

Supplementary Material

1

Acknowledgments

This research was supported by NIH grants EY027859, EY027323, EY002576, NS099578, EY07031 and P30EY001730 (Vision Core), by the University of Texas BRAIN Seed Grant Program and by Research to Prevent Blindness.

Footnotes

Data Availability Statement: Access to the macaque EM volume dataset is available on request. Visualizing both the dataset and the annotations requires the open-source Viking Viewer developed in Bryan Jones’ lab at (http://connectomes.utah.edu). The 3D reconstructions from Viking Viewer annotations are visualized with SBFSEM-tools, an open-source MATLAB toolbox developed in the Neitz lab (https://github.com/neitzlab/sbfsem-tools). Access to the dataset through the Viking Viewer and SBFSEM-tools does not require that the user install the data locally.

References

  1. Abbott CJ, Percival KA, Martin PR, & Grunert U (2012). Amacrine and bipolar inputs to midget and parasol ganglion cells in marmoset retina. Vis Neurosci, 29(3), 157–168. doi: 10.1017/s095252381200017x [DOI] [PubMed] [Google Scholar]
  2. Bordt AS, Hoshi H, Yamada ES, Perryman-Stout WC, & Marshak DW (2006). Synaptic input to OFF parasol ganglion cells in macaque retina. J Comp Neurol, 498(1), 46–57. doi: 10.1002/cne.21040 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Bordt AS, Perez D, Tseng L, Liu WS, Neitz J, Patterson SS,… Marshak DW (2019). Synaptic inputs from identified bipolar and amacrine cells to a sparsely branched ganglion cell in rabbit retina. Vis Neurosci, 36, E004. doi: 10.1017/s0952523819000014 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Cafaro J, & Rieke F (2010). Noise correlations improve response fidelity and stimulus encoding. Nature, 468(7326), 964–967. doi: 10.1038/nature09570 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Cafaro J, & Rieke F (2013). Regulation of spatial selectivity by crossover inhibition. J Neurosci, 33(15), 6310–6320. doi: 10.1523/jneurosci.4964-12.2013 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Calkins DJ, & Sterling P (2007). Microcircuitry for two types of achromatic ganglion cell in primate fovea. J Neurosci, 27(10), 2646–2653. doi: 10.1523/jneurosci.4739-06.2007 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Crook JD, Packer OS, & Dacey DM (2014). A synaptic signature for ON- and OFF-center parasol ganglion cells of the primate retina. Vis Neurosci, 31(1), 57–84. doi: 10.1017/s0952523813000461 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Crook JD, Peterson BB, Packer OS, Robinson FR, Gamlin PD, Troy JB, & Dacey DM (2008). The smooth monostratified ganglion cell: evidence for spatial diversity in the Y-cell pathway to the lateral geniculate nucleus and superior colliculus in the macaque monkey. J Neurosci, 28(48), 12654–12671. doi: 10.1523/jneurosci.2986-08.2008 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Dacey DM, & Brace S (1992). A coupled network for parasol but not midget ganglion cells in the primate retina. Vis Neurosci, 9(3–4), 279–290. [DOI] [PubMed] [Google Scholar]
  10. Dacey DM, & Lee BB (1994). The ‘blue-on’ opponent pathway in primate retina originates from a distinct bistratified ganglion cell type. Nature, 367(6465), 731–735. doi: 10.1038/367731a0 [DOI] [PubMed] [Google Scholar]
  11. Davenport CM, Detwiler PB, & Dacey DM (2008). Effects of pH buffering on horizontal and ganglion cell light responses in primate retina: evidence for the proton hypothesis of surround formation. J Neurosci, 28(2), 456–464. doi: 10.1523/jneurosci.2735-07.2008 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Greschner M, Heitman AK, Field GD, Li PH, Ahn D, Sher A,… Chichilnisky EJ (2016). Identification of a Retinal Circuit for Recurrent Suppression Using Indirect Electrical Imaging. Curr Biol, 26(15), 1935–1942. doi: 10.1016/j.cub.2016.05.051 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Grunert U (2000). Distribution of GABA and glycine receptors on bipolar and ganglion cells in the mammalian retina. Microsc Res Tech, 50(2), 130–140. doi: [DOI] [PubMed] [Google Scholar]
  14. Grunert U, & Wassle H (1990). GABA-like immunoreactivity in the macaque monkey retina: a light and electron microscopic study. J Comp Neurol, 297(4), 509–524. doi: 10.1002/cne.902970405 [DOI] [PubMed] [Google Scholar]
  15. Huang JY, & Protti DA (2016). The impact of inhibitory mechanisms in the inner retina on spatial tuning of RGCs. Sci Rep, 6, 21966. doi: 10.1038/srep21966 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Jacoby R, Stafford D, Kouyama N, & Marshak D (1996). Synaptic inputs to ON parasol ganglion cells in the primate retina. J Neurosci, 16(24), 8041–8056. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Kalloniatis M, Marc RE, & Murry RF (1996). Amino acid signatures in the primate retina. J Neurosci, 16(21), 6807–6829. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Kolb H, Linberg KA, & Fisher SK (1992). Neurons of the human retina: a Golgi study. J Comp Neurol, 318(2), 147–187. doi: 10.1002/cne.903180204 [DOI] [PubMed] [Google Scholar]
  19. Koontz MA, & Hendrickson AE (1987). Stratified distribution of synapses in the inner plexiform layer of primate retina. J Comp Neurol, 263(4), 581–592. doi: 10.1002/cne.902630409 [DOI] [PubMed] [Google Scholar]
  20. Kruger J, Fischer B, & Barth R (1975). The shift-effect in retinal ganglion cells of the rhesus monkey. Exp Brain Res, 23(4), 443–446. [DOI] [PubMed] [Google Scholar]
  21. Macri J, Martin PR, & Grunert U (2000). Distribution of the alpha1 subunit of the GABA(A) receptor on midget and parasol ganglion cells in the retina of the common marmoset Callithrix jacchus. Vis Neurosci, 17(3), 437–448. [DOI] [PubMed] [Google Scholar]
  22. Majumdar S, Wassle H, Jusuf PR, & Haverkamp S (2008). Mirror-symmetrical populations of wide-field amacrine cells of the macaque monkey retina. J Comp Neurol, 508(1), 13–27. doi: 10.1002/cne.21666 [DOI] [PubMed] [Google Scholar]
  23. Manookin MB, Patterson SS, & Linehan CM (2018). Neural Mechanisms Mediating Motion Sensitivity in Parasol Ganglion Cells of the Primate Retina. Neuron, 97(6), 1327–1340 e1324. doi: 10.1016/j.neuron.2018.02.006 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Manookin MB, Puller C, Rieke F, Neitz J, & Neitz M (2015). Distinctive receptive field and physiological properties of a wide-field amacrine cell in the macaque monkey retina. J Neurophysiol, 114(3), 1606–1616. doi: 10.1152/jn.00484.2015 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Mariani AP (1990). Amacrine cells of the rhesus monkey retina. J Comp Neurol, 301(3), 382–400. doi: 10.1002/cne.903010305 [DOI] [PubMed] [Google Scholar]
  26. Marshak DW (2016). A tale of two neurotransmitters. Vis Neurosci, 33, E017. doi: 10.1017/s0952523816000146 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Marshak DW, Aldrich LB, Del Valle J, & Yamada T (1990). Localization of immunoreactive cholecystokinin precursor to amacrine cells and bipolar cells of the macaque monkey retina. J Neurosci, 10(9), 3045–3055. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Marshak DW, Yamada ES, Bordt AS, & Perryman WC (2002). Synaptic input to an ON parasol ganglion cell in the macaque retina: a serial section analysis. Vis Neurosci, 19(3), 299–305. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. McMahon MJ, Packer OS, & Dacey DM (2004). The classical receptive field surround of primate parasol ganglion cells is mediated primarily by a non-GABAergic pathway. J Neurosci, 24(15), 3736–3745. doi: 10.1523/jneurosci.5252-03.2004 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Park SJH, Pottackal J, Ke JB, Jun NY, Rahmani P, Kim IJ,… Demb JB (2018). Convergence and Divergence of CRH Amacrine Cells in Mouse Retinal Circuitry. J Neurosci, 38(15), 3753–3766. doi: 10.1523/jneurosci.2518-17.2018 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Patterson SS, Kuchenbecker JA, Anderson JR, Bordt AS, Marshak DW, Neitz M, & Neitz J (2019). An S-cone circuit for edge detection in the primate retina. Sci Rep, 9(1), 11913. doi: 10.1038/s41598-019-48042-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Patterson SS, Bordt AB, Girresch RJ, Linehan CM, Bauss J, Yeo E, Perez D, Tseng L, Navuluri S, Harris NB, Matthews C, Anderson JR, Kuchenbecker JA, Manookin MB, Ogilvie JM, Neitz J, & Marshak DW (2019). Wide-field amacrine cell inputs to ON parasol ganglion cells in macaque retina. University of Washington Library. doi: 10.6069/CNXW-5G77 [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Protti DA, Di Marco S, Huang JY, Vonhoff CR, Nguyen V, & Solomon SG (2014). Inner retinal inhibition shapes the receptive field of retinal ganglion cells in primate. J Physiol, 592(Pt 1), 49–65. doi: 10.1113/jphysiol.2013.257352 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Roy K, Kumar S, & Bloomfield SA (2017). Gap junctional coupling between retinal amacrine and ganglion cells underlies coherent activity integral to global object perception. Proc Natl Acad Sci U S A, 114(48), E10484–E10493. doi: 10.1073/pnas.1708261114 [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Solomon SG, Lee BB, & Sun H (2006). Suppressive surrounds and contrast gain in magnocellular-pathway retinal ganglion cells of macaque. J Neurosci, 26(34), 8715–8726. doi: 10.1523/jneurosci.0821-06.2006 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Tsukamoto Y, & Omi N (2016). ON Bipolar Cells in Macaque Retina: Type-Specific Synaptic Connectivity with Special Reference to OFF Counterparts. Front Neuroanat, 10, 104. doi: 10.3389/fnana.2016.00104 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Turner MH, & Rieke F (2016). Synaptic Rectification Controls Nonlinear Spatial Integration of Natural Visual Inputs. Neuron, 90(6), 1257–1271. doi: 10.1016/j.neuron.2016.05.006 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Wassle H, Grunert U, Rohrenbeck J, & Boycott BB (1989). Cortical magnification factor and the ganglion cell density of the primate retina. Nature, 341(6243), 643–646. doi: 10.1038/341643a0 [DOI] [PubMed] [Google Scholar]
  39. Wassle H, Heinze L, Ivanova E, Majumdar S, Weiss J, Harvey RJ, & Haverkamp S (2009). Glycinergic transmission in the Mammalian retina. Front Mol Neurosci, 2, 6. doi: 10.3389/neuro.02.006.2009 [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Yamada ES, Bordt AS, & Marshak DW (2005). Wide-field ganglion cells in macaque retinas. Vis Neurosci, 22(4), 383–393. doi: 10.1017/s095252380522401x [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1546334-supplement-1.pdf (186.3KB, pdf)

RESOURCES