THE JOURNAL OF COMPARATNE NEUROLOGY 301:382-400 (1990)
Amacrine Cells of the RIhesus
MonkeyRetina ANDRT3w P. MARIANI Laboratory of Neurophysiology, National Institute of Neurological Disorders and Stroke, National Institutes of Health, Bethesda, Maryland 20892
ABSTRACT Amacrine cells of the rhesus monkey, Macaca mulatta, were studied in 38 retinas Golgi-impregnatedas whole, flat preparations. By using criteria of dendritic morphology, span of arborization, and level of arborization in the inner plexiform layer, 26 types of amacrine cell ranging in size of dendritic span from 30 pm to 2 mm were identified and listed in increasing size of dendritic span. In some instances, different cell types could be grouped together due to similar morphological features. For example, 1 group, “knotty amacrine cells,” has small cell bodies and a profusion of small, varicose, intertwined processes that span up to 30 pm and are essentially monostratified, but each of the 3 types ends in different strata. Another group is 2 types with about 20 fine radiating processes spanning 1 mm that possess some prominent varicosities. One of these has all of its processes terminating in the innermost stratum of the inner plexiform layer (“spidery”-type 2 amacrine cells). The other with predominantly similarly ending processes has some that also terminate in the outermost stratum (“spidery”type 1 amacrines). These 2 cell types likely correspond to the type 1 and type 2 indolamineaccumulating amacrine cells in rabbit retina. Other types are individuals which cannot be grouped together but resemble familiar types in cat retina (AII and A13). Other types can be correlated with their putative neurotransmitter (type 1 CA-dopamine) or transmitteddrug receptor (“spiny”-benzodiazepine receptor) phenotype. Many types as yet have no known correlate from other Golgi studies or clues as to transmitter or receptor phenotype. This study provides evidence for an unprecedented number of amacrine cell types in the primate retina. The similar morphologies of different types of amacrine cell types within a group suggest other common features within these groups such as neurotransmitter phenotype. Key words: Golgi impregnation, retinal amacrine cells, neuronal morphology, Macaca mulatta
Amacrine cells are a heterogeneous and diverse class of retinal interneuron composed of many different and distinct morphological types (Cajal, ’33), each of which may subserve a different functional role. They are intrinsic retinal neurons with pre- and postsynaptic processes exclusively or primarily confined to the inner plexiform layer. Although generally considered to be axonless, there are examples of amacrine cells with axons (Mariani, ’82a; Catsicas et al., ’87; Dacey, ’89). Surprisingly, amacrine cells of the primate retina have not been well characterized, yet information on the number of types, their dendritic morphology, the span of their processes, and their level of arborization in the inner plexiform layer would be useful in understanding the functional organization of the primate retina. Previous studies with Golgi’s method have defined and identified a number of amacrine cell types in the monkey retina. The broadest of these works are the classic tome of Polyak (’41) and the well-cited study of Boycott and Dowl-
o 1990 WILEY-LISS, INC.
ing (’69). Both studies attempted comprehensive descriptions of all neuronal cell types in the primate retina, not just amacrine cells. Although these authors recognized the stratification of the inner plexiform layer, neither placed emphasis on this feature in coming to terms with the identification and classification of amacrine cells. More recently, others have identified individual types of macaque amacrine cells such as “spiny” (Mariani and Barker, ’87)or “starburst” (Rodieck, ’88) with Golgi impregnation, and there is a growing body of work on the identification of neurotransmitters in monkey amacrine cells (Hendrickson et al., ’81; Brecha et al., ’82; Tornqvist et al., ’82; NguyenLegros et al., ’84; Mosinger and Altschuler, ’85; Mariani and Hokoc, ’88;Marshak, ’89) and details of their synaptic connections in the inner plexiform layer (Mariani and Caserta, ’86; Hokoc and Mariani, ‘87; Hendrickson et al., ’88; Mariani and Hersh, ’88). Recent studies have also Accepted July 24,1990.
383
MONKEY RETINAL AMACRINE CELLS demonstrated a number of newly identified types of neurons and their synaptic connections in the outer plexiform layer (Boycott et al., ’75; Kolb et al., ’80; Mariani, ’81,’82a, ’83, ’84), showing this first synaptic layer of the primate retina to be richer in the number of cell types than previously thought (Kolb, ’70). Because of the limitations of earlier Golgi-impregnation studies on amacrine cells in the primate retina as outlined above, the information showing more cell types in the outer plexiform layer than previously thought, the wealth of new data on the likely neurotransmitter phenotypes of different amacrine cells, and the known differences in neuronal morphology of certain classes of retinal neurons between different species of mammals (Mariani, ’85), it seemed timely to examine the morphology of primate amacrine cells with the classic Golgi-impregnationmethod. The work reported here is an attempt to provide a comprehensive description of the number of types of amacrine cell and their morphology in the primate retina on the basis Golgi impregnation. Twenty-six different types are described. Where possible, comparisons have been made to morphologically similar types in other mammalian retinas. Correlations to putative neurotransmitter phenotype have been attempted as well.
MATERIALSANDMETHODS Thirty-eight retinas from the eyes of 20 rhesus monkeys, Macaca mulatta, were dissected from the sclera, choroid, and pigment epithelium in an aqueous solution of NaCl (0.9%w/v) and Gold impregnated as described previously (Mariani, ’82b, ’85).Laid flat on wax sheets with radial cuts in order to aid flattening, the retinas were fixed by covering with a solution of 2.5% glutaraldehyde in 0.1 M cacodylate buffer or 5% glutaraldehyde in an aqueous solution of 4% potassium dichromate. The retinas were then sandwiched between sheets of Whatman #50 hardened filter paper, Whatman #4 filter paper, and 1.5 x 1.5 in. glass microscope slides. These sandwiches were held together with rubber bands and placed in 50 ml of the dichromate-glutaraldehyde solution in a 50 ml ointment jar and placed in light-tight boxes. After 3-6 days, the sandwiches were removed from the glutaraldehyde-dichromate solution and rinsed by briefly dipping in 4 successive baths of aqueous 1%silver nitrate, blotting the edges on absorbent paper towels between dippings. Following the last rinse, sandwiches were placed in 50 ml of 1%aqueous silver nitrate in clean 50 ml ointment jars and returned to the light-tight boxes for 2-5 days. The sandwiches were then removed from the silver nitrate, rinsed in 10% ethanol and rapidly dehydrated in a graded series of ethanol. The sandwich was disassembled in 100% ethanol and the retinas were removed. The retinas, cleared in propylene oxide and infiltrated in a 1:1mixture of propylene oxide and Epon 812, were embedded by placing them ganglion cell side up on a 2 x 3 in. glass microscope slide in a drop of uncured Epon and coveringwith a sheet of polyethylene (plastic bag). A cover glass, held in place by a brass weight, was placed over the polyethylene and the Epon cured overnight at 60°C. After curing, the weight, cover glass, and polyethylene were removed. The preparations were studied by light microscopy, photographed, and drawn with the aid of a camera lucida. All drawings were made by utilizing a 40 x oil-immersionobjective (NA = l.O), and all drawings are presented to the same scale. Levels of dendritic branching in the inner plexiform layer were
determined by focusing on the innermost part of the inner nuclear layer, the outermost part of the ganglion cell layer, and the dendrites in question and noting the position of each of these landmarks with the focusing control knob of the microscope calibrated in pm, or from radial sections. Golgi-impregnatedamacrine cells were differentiated and classified on the bases of: 1)dendritic morphology, 2) size of dendritic span, and 3) level of dendritic arborization in the inner plexiform layer. They were then sorted, and are illustrated, on the basis of overall size (diameter of dendritic span). In the case of different amacrine cell types with both similar dendritic morphology and dendritic span, but differing in levels of arborization in the inner plexiform layer, preference in numbering (i.e., a lower number) was assigned to the type(s) that arborized or terminated in outer strata of the inner plexiform layer. The nomenclature for the 5 different strata of the inner plexiform layer, 7a-7e, is that introduced previously for the primate retina by Polyak (’41). All of the camera lucida drawings are of cells located 4-6 mm from the center of the fovea. Most of the amacrine cell types described here are literally found numbering in the hundreds in a single retina, particularly the ones with the smaller diameter dendritic fields. Therefore no attempt has been made to enumerate the individual types in the 38 retinas, except in the instances where very few of the particular cell types have been found, or where the numbers indicate a particular feature, such as the location of cell bodies of the subpopulations of a particular type.
RESULTS 3 typesof “Knotty”amacrine cells These cells are the 3 “knotty” amacrine cells illustrated by Polyak (’41: Fig. 57b) and redrawn from Polyak by Boycott and Dowling (’69: Fig. 54c). Types 1 through 3 “knotty” amacrines (Fig. 1)have small, 7-8 pm-diameter, spherical cell bodies exclusively located in the innermost row of somata of the inner nuclear layer. Usually a single 1-2 pm-thick process descends radially to the inner plexiform layer where it branches and arborizes profusely forming a rather compact, 30-45 pm in span, dendritic arborization composed of very fine processes (
Amacrine Cells of the RIhesus
MonkeyRetina ANDRT3w P. MARIANI Laboratory of Neurophysiology, National Institute of Neurological Disorders and Stroke, National Institutes of Health, Bethesda, Maryland 20892
ABSTRACT Amacrine cells of the rhesus monkey, Macaca mulatta, were studied in 38 retinas Golgi-impregnatedas whole, flat preparations. By using criteria of dendritic morphology, span of arborization, and level of arborization in the inner plexiform layer, 26 types of amacrine cell ranging in size of dendritic span from 30 pm to 2 mm were identified and listed in increasing size of dendritic span. In some instances, different cell types could be grouped together due to similar morphological features. For example, 1 group, “knotty amacrine cells,” has small cell bodies and a profusion of small, varicose, intertwined processes that span up to 30 pm and are essentially monostratified, but each of the 3 types ends in different strata. Another group is 2 types with about 20 fine radiating processes spanning 1 mm that possess some prominent varicosities. One of these has all of its processes terminating in the innermost stratum of the inner plexiform layer (“spidery”-type 2 amacrine cells). The other with predominantly similarly ending processes has some that also terminate in the outermost stratum (“spidery”type 1 amacrines). These 2 cell types likely correspond to the type 1 and type 2 indolamineaccumulating amacrine cells in rabbit retina. Other types are individuals which cannot be grouped together but resemble familiar types in cat retina (AII and A13). Other types can be correlated with their putative neurotransmitter (type 1 CA-dopamine) or transmitteddrug receptor (“spiny”-benzodiazepine receptor) phenotype. Many types as yet have no known correlate from other Golgi studies or clues as to transmitter or receptor phenotype. This study provides evidence for an unprecedented number of amacrine cell types in the primate retina. The similar morphologies of different types of amacrine cell types within a group suggest other common features within these groups such as neurotransmitter phenotype. Key words: Golgi impregnation, retinal amacrine cells, neuronal morphology, Macaca mulatta
Amacrine cells are a heterogeneous and diverse class of retinal interneuron composed of many different and distinct morphological types (Cajal, ’33), each of which may subserve a different functional role. They are intrinsic retinal neurons with pre- and postsynaptic processes exclusively or primarily confined to the inner plexiform layer. Although generally considered to be axonless, there are examples of amacrine cells with axons (Mariani, ’82a; Catsicas et al., ’87; Dacey, ’89). Surprisingly, amacrine cells of the primate retina have not been well characterized, yet information on the number of types, their dendritic morphology, the span of their processes, and their level of arborization in the inner plexiform layer would be useful in understanding the functional organization of the primate retina. Previous studies with Golgi’s method have defined and identified a number of amacrine cell types in the monkey retina. The broadest of these works are the classic tome of Polyak (’41) and the well-cited study of Boycott and Dowl-
o 1990 WILEY-LISS, INC.
ing (’69). Both studies attempted comprehensive descriptions of all neuronal cell types in the primate retina, not just amacrine cells. Although these authors recognized the stratification of the inner plexiform layer, neither placed emphasis on this feature in coming to terms with the identification and classification of amacrine cells. More recently, others have identified individual types of macaque amacrine cells such as “spiny” (Mariani and Barker, ’87)or “starburst” (Rodieck, ’88) with Golgi impregnation, and there is a growing body of work on the identification of neurotransmitters in monkey amacrine cells (Hendrickson et al., ’81; Brecha et al., ’82; Tornqvist et al., ’82; NguyenLegros et al., ’84; Mosinger and Altschuler, ’85; Mariani and Hokoc, ’88;Marshak, ’89) and details of their synaptic connections in the inner plexiform layer (Mariani and Caserta, ’86; Hokoc and Mariani, ‘87; Hendrickson et al., ’88; Mariani and Hersh, ’88). Recent studies have also Accepted July 24,1990.
383
MONKEY RETINAL AMACRINE CELLS demonstrated a number of newly identified types of neurons and their synaptic connections in the outer plexiform layer (Boycott et al., ’75; Kolb et al., ’80; Mariani, ’81,’82a, ’83, ’84), showing this first synaptic layer of the primate retina to be richer in the number of cell types than previously thought (Kolb, ’70). Because of the limitations of earlier Golgi-impregnation studies on amacrine cells in the primate retina as outlined above, the information showing more cell types in the outer plexiform layer than previously thought, the wealth of new data on the likely neurotransmitter phenotypes of different amacrine cells, and the known differences in neuronal morphology of certain classes of retinal neurons between different species of mammals (Mariani, ’85), it seemed timely to examine the morphology of primate amacrine cells with the classic Golgi-impregnationmethod. The work reported here is an attempt to provide a comprehensive description of the number of types of amacrine cell and their morphology in the primate retina on the basis Golgi impregnation. Twenty-six different types are described. Where possible, comparisons have been made to morphologically similar types in other mammalian retinas. Correlations to putative neurotransmitter phenotype have been attempted as well.
MATERIALSANDMETHODS Thirty-eight retinas from the eyes of 20 rhesus monkeys, Macaca mulatta, were dissected from the sclera, choroid, and pigment epithelium in an aqueous solution of NaCl (0.9%w/v) and Gold impregnated as described previously (Mariani, ’82b, ’85).Laid flat on wax sheets with radial cuts in order to aid flattening, the retinas were fixed by covering with a solution of 2.5% glutaraldehyde in 0.1 M cacodylate buffer or 5% glutaraldehyde in an aqueous solution of 4% potassium dichromate. The retinas were then sandwiched between sheets of Whatman #50 hardened filter paper, Whatman #4 filter paper, and 1.5 x 1.5 in. glass microscope slides. These sandwiches were held together with rubber bands and placed in 50 ml of the dichromate-glutaraldehyde solution in a 50 ml ointment jar and placed in light-tight boxes. After 3-6 days, the sandwiches were removed from the glutaraldehyde-dichromate solution and rinsed by briefly dipping in 4 successive baths of aqueous 1%silver nitrate, blotting the edges on absorbent paper towels between dippings. Following the last rinse, sandwiches were placed in 50 ml of 1%aqueous silver nitrate in clean 50 ml ointment jars and returned to the light-tight boxes for 2-5 days. The sandwiches were then removed from the silver nitrate, rinsed in 10% ethanol and rapidly dehydrated in a graded series of ethanol. The sandwich was disassembled in 100% ethanol and the retinas were removed. The retinas, cleared in propylene oxide and infiltrated in a 1:1mixture of propylene oxide and Epon 812, were embedded by placing them ganglion cell side up on a 2 x 3 in. glass microscope slide in a drop of uncured Epon and coveringwith a sheet of polyethylene (plastic bag). A cover glass, held in place by a brass weight, was placed over the polyethylene and the Epon cured overnight at 60°C. After curing, the weight, cover glass, and polyethylene were removed. The preparations were studied by light microscopy, photographed, and drawn with the aid of a camera lucida. All drawings were made by utilizing a 40 x oil-immersionobjective (NA = l.O), and all drawings are presented to the same scale. Levels of dendritic branching in the inner plexiform layer were
determined by focusing on the innermost part of the inner nuclear layer, the outermost part of the ganglion cell layer, and the dendrites in question and noting the position of each of these landmarks with the focusing control knob of the microscope calibrated in pm, or from radial sections. Golgi-impregnatedamacrine cells were differentiated and classified on the bases of: 1)dendritic morphology, 2) size of dendritic span, and 3) level of dendritic arborization in the inner plexiform layer. They were then sorted, and are illustrated, on the basis of overall size (diameter of dendritic span). In the case of different amacrine cell types with both similar dendritic morphology and dendritic span, but differing in levels of arborization in the inner plexiform layer, preference in numbering (i.e., a lower number) was assigned to the type(s) that arborized or terminated in outer strata of the inner plexiform layer. The nomenclature for the 5 different strata of the inner plexiform layer, 7a-7e, is that introduced previously for the primate retina by Polyak (’41). All of the camera lucida drawings are of cells located 4-6 mm from the center of the fovea. Most of the amacrine cell types described here are literally found numbering in the hundreds in a single retina, particularly the ones with the smaller diameter dendritic fields. Therefore no attempt has been made to enumerate the individual types in the 38 retinas, except in the instances where very few of the particular cell types have been found, or where the numbers indicate a particular feature, such as the location of cell bodies of the subpopulations of a particular type.
RESULTS 3 typesof “Knotty”amacrine cells These cells are the 3 “knotty” amacrine cells illustrated by Polyak (’41: Fig. 57b) and redrawn from Polyak by Boycott and Dowling (’69: Fig. 54c). Types 1 through 3 “knotty” amacrines (Fig. 1)have small, 7-8 pm-diameter, spherical cell bodies exclusively located in the innermost row of somata of the inner nuclear layer. Usually a single 1-2 pm-thick process descends radially to the inner plexiform layer where it branches and arborizes profusely forming a rather compact, 30-45 pm in span, dendritic arborization composed of very fine processes (
