Light-induced c-fos in melanopsin retinal ganglion cells of young and aged rodless/coneless (rd/rd cl) mice

Non‐rod, non‐cone ocular photoreceptors have been shown to mediate a range of irradiance detection tasks. The strongest candidates for these receptors are melanopsin‐positive retinal ganglion cells (RGCs). To provide a more complete understanding of these receptors in vivo, we have utilized a mouse that lacks rod and cone photoreceptors (rd/rd cl) and compared these animals to congenic wild‐types. Using real‐time polymerase chain reaction and immunohistochemistry, we address the following. (1) Is Fos expression within these RGCs driven by an input from the rods/cones or is it the product of the intrinsic photosensitivity of these neurons? We demonstrate that most Fos expression across the entire retina is due to the rods/cones, but in the absence of these photoreceptors, light will induce Fos within melanopsin RGCs. (2) Could the reported age‐related decline in circadian photosensitivity of rodents be linked to changes in the population of melanopsin RGCs? We show that old mice experience an ～ 40% reduction in melanopsin RGCs. (3) Does the loss of inner retinal neurons affect the responses of melanopsin RGCs? Aged (～ 700 days) rd/rd cl mice lose most of their inner retina but retain the retinal ganglion cell layer. In these mice, the proportion of melanopsin RGCs that express Fos in response to light is significantly reduced. Collectively, our data suggest that melanopsin RGCs form a heterogeneous population of neurons, and that most of the light‐induced c‐fos expression within these cells is associated with the endogenous photosensitivity of these neurons.     

Intrinsic light responses of retinal ganglion cells projecting to the circadian system

In mammals, light entrainment of the circadian clock, located in the suprachiasmatic nuclei (SCN), requires retinal input. Traditional rod and cone photoreceptors, however, are not required. Instead, the SCN-projecting retinal ganglion cells (RGCs) function as autonomous photoreceptors and exhibit light responses independent of rod- and cone-driven input. Using whole-cell patch-clamp recording techniques, we have investigated the morphological and electrophysiological properties of this unique class of RGCs. Although SCN-projecting RGCs resemble Type III cells in form, they display strikingly different physiological properties from these neurons. First, in response to the injection of a sustained depolarizing current, SCN-projecting cells fired in a transient fashion, in contrast to most RGCs which fired robust trains of action potentials. Second, in response to light, SCN-projecting RGCs exhibited an intensity-dependent transient depolarization in the absence of rod and cone input. This depolarization reached a peak within 5 s and generated increased spiking activity before decaying to a plateau. Voltage-clamp recordings were used to characterize the light-activated conductance which generated this depolarization. In response to varying light intensities, SCN-projecting RGCs exhibited a graded transient inward current which peaked within 5 s and decayed to a plateau. The voltage dependence of the light-activated current was obtained by subtracting currents elicited by a voltage ramp before and during illumination. The light-activated current displayed both inward and outward rectification and was largely unaffected by substitution of extracellular Na+ with choline. In both respects, the intrinsic light-activated current observed in SCN-projecting RGCs resembles currents carried by ion channels of the transient receptor potential (trp) family, which are known to mediate the light response of invertebrate photoreceptors.     

Synaptic inputs to retinal ganglion cells that set the circadian clock

Melanopsin-containing retinal ganglion cells (RGCs) project to the suprachiasmatic nuclei (SCN) and mediate photoentrainment of the circadian system. Melanopsin is a novel retinal-based photopigment that renders these cells intrinsically photosensitive (ip). Although genetic ablation of melanopsin abolishes the intrinsic light response, it has a surprisingly minor effect on circadian photoentrainment. This and other non-visual responses to light are lost only when the melanopsin deficiency is coupled with mutations that disable classical rod and cone photoreceptors, suggesting that melanopsin-containing RGCs also receive rod- and cone-driven synaptic inputs. Using whole-cell patch-clamp recording, we demonstrate that light triggers synaptic currents in ipRGCs via activation of ionotropic glutamate and gamma-aminobutyric acid (GABA) receptors. Miniature postsynaptic currents (mPSCs) were clearly observed in ipRGCs, although they were less robust and were seen less frequently than those seen in non-ip cells. Pharmacological treatments revealed that the majority of ipRGCs receive excitatory glutamatergic inputs that were blocked by DNQX and/or kynurenic acid, as well as inhibitory GABAergic inputs that were blocked by bicuculline. Other ipRGCs received either glutamatergic or GABAergic inputs nearly exclusively. Although strychnine (Strych)-sensitive mPSCs were evident on many non-ipRGCs, indicating the presence of glycinergic inputs, we saw no evidence of Strych-sensitive events in ipRGCs. Based on these results, it is clear that SCN-projecting RGCs can respond to light both via an intrinsic melanopsin-based signaling cascade and via a synaptic pathway driven by classical rod and/or cone photoreceptors. It remains to be determined how the ipRGCs integrate these temporally distinct inputs to generate the signals that mediate circadian photoentrainment and other non-visual responses to light.     

The development of melanopsin-containing retinal ganglion cells in mice with early retinal degeneration

In mammals, the neuronal pathways by which rod and cone photoreceptors mediate vision have been well documented. The roles that classical photoreceptors play in photoentrainment, however, have been less clear. In mammals, intrinsically photosensitive retinal ganglion cells (ipRGCs) that express the photopigment melanopsin project directly to the suprachiasmatic nucleus of the hypothalamus, the site of the circadian clock, and thereby contribute to non-image-forming responses to light. Classical photoreceptors are not necessary for photoentrainment as loss of rods and cones does not eliminate light entrainment. Conflicting evidence arose, however, when attenuated phase-shifting responses were observed in the retinal-degenerate CBA/J mouse. In this study, we examined the time course of retinal degeneration in CBA/J mice and used these animals to determine if maturation of the outer retina regulates the morphology, number and distribution of ipRGCs. We also examined whether degeneration during the early development of the outer retina can alter the function of the adult circadian system. We report that dendritic stratification and distribution of ipRGCs was unaltered in mice with early retinal degeneration, suggesting that normal development of the outer retina was not necessary for these processes. We found, however, that adult CBA/J mice have greater numbers of ipRGCs than controls, implicating a role for the outer retinal photoreceptors in regulating developmental cell death of ipRGCs.     

Aberrant synaptic input to retinal ganglion cells varies with morphology in a mouse model of retinal degeneration

Retinal degeneration describes a group of disorders which lead to progressive photoreceptor cell death, resulting in blindness. As this occurs, retinal ganglion cells (RGCs) begin to develop oscillatory physiological activity. Here we studied the morphological and physiological properties of RGCs in rd1 mice, aged 30–60 days, to determine how this aberrant activity correlates with morphology. Patch‐clamp recordings of excitatory and inhibitory currents were performed, then dendritic structures were visualized by infusion of fluorescent dye. Only RGCs with oscillatory activity were selected for further analysis. Oscillatory frequency and power were calculated using power spectral density analysis of recorded currents. Dendritic arbor stratification, total length, and area were measured from confocal microscope image stacks. These measurements were used to sort RGCs by cluster analysis using Ward's Method. This resulted in a total of 10 clusters, with monostratified and bistratified cells having five clusters each. Both populations exhibited correlations between arbor stratification and aberrant inhibitory input, while excitatory input did not vary with arbor distribution. These findings illustrate the relationship between aberrant activity and RGC morphology at early stages of retinal degeneration. J. Comp. Neurol. 522:4085–4099, 2014. © 2014 Wiley Periodicals, Inc.     

Blue light stress in retinal neuronal (R28) cells is dependent on wavelength range and irradiance

The aim of our study was to elucidate the role of wavelength and irradiance in blue light retinal damage. We investigated the impact of blue light emitted from light-emitting diode (LED) modules with peaks at either 411nm (half bandwidth 17nm) or 470nm (half bandwidth 25nm) at defined irradiances of 0.6, 1.5 and 4.5W/m(2) for 411nm and 4.5W/m(2) for 470nm on retinal neuronal (R28) cells in vitro. We observed a reduction in metabolic activity and transmembrane potential of mitochondria when cells were irradiated at 411nm at higher irradiances. Furthermore, production of mitochondrial superoxide radicals increased significantly when cells were irradiated with 411nm light at 4.5W/m(2) . In addition, such irradiation caused an activation of the antioxidative glutathion system. Using vital staining, flow cytometry and western blotting, we were able to show that apoptosis only took place when cells were exposed to 411nm blue light at higher irradiances; necrosis was not observed. Enhanced caspase-3 cleavage product levels confirmed that this effect was dependent on light irradiance. Significant alterations of the above-mentioned parameters were not observed when cells were irradiated with 471nm light despite a high irradiance of 4.5W/m(2) , indicating that the cytotoxic effect of blue light is highly dependent on wavelength. The observed phenomena in R28 cells at 411nm (4.5W/m(2) ) point to an apoptosis pathway elicited by direct mitochondrial damage and increased oxidative stress. Thus, light of 411nm should act via impairment of mitochondrial function by compromising the metabolic situation of these retinal neuronal cells.     

Morphological classification of retinal ganglion cells in mice

Mice have been used for extensive studies on optic nerves and retinal ganglion cells, but mouse retinal ganglion cells have not been classified morphologically. In the present study, normally placed retinal ganglion cells and displaced retinal ganglion cells in pigmented and albino mice were classified morphologically using horseradish peroxidase. These cells were classified into three types according to the sizes of the soma and the dendritic field: type I cells, large soma and large dendritic field; type II cells, small-to-medium soma and small dendritic field; and type III cells, small-to-medium soma and large dendritic field. Some ganglion cells had both symmetric and asymmetric cells. Each type was further subdivided according to the termination level of dendrites in the inner plexiform layer and the dendritic branching pattern. Except for type III displaced ganglion cells, dendrites of the normally placed ganglion cells and the displaced ganglion cells ramify in the outer two-fifths of the inner plexiform layer (sublamina a) or the inner three-fifths of the inner plexiform layer (sublamina b). Type III displaced ganglion cells ramify only in sublamina a. Dendrites of some normally placed type I ganglion cells ramify in both sublaminae. Displaced biplexiform cells were observed, the dendrites of which ramify in both the inner and the outer plexiform layers. All cell types were found in both mouse strains. © 1995 Wiley-Liss, Inc.     

Central projections of melanopsin-expressing retinal ganglion cells in the mouse

A rare type of ganglion cell in mammalian retina is directly photosensitive. These novel retinal photoreceptors express the photopigment melanopsin. They send axons directly to the suprachiasmatic nucleus (SCN), intergeniculate leaflet (IGL), and olivary pretectal nucleus (OPN), thereby contributing to photic synchronization of circadian rhythms and the pupillary light reflex. Here, we sought to characterize more fully the projections of these cells to the brain. By targeting tau-lacZ to the melanopsin gene locus in mice, ganglion cells that would normally express melanopsin were induced to express, instead, the marker enzyme beta-galactosidase. Their axons were visualized by X-gal histochemistry or anti-beta-galactosidase immunofluorescence. Established targets were confirmed, including the SCN, IGL, OPN, ventral division of the lateral geniculate nucleus (LGv), and preoptic area, but the overall projections were more widespread than previously recognized. Targets included the lateral nucleus, peri-supraoptic nucleus, and subparaventricular zone of the hypothalamus, medial amygdala, margin of the lateral habenula, posterior limitans nucleus, superior colliculus, and periaqueductal gray. There were also weak projections to the margins of the dorsal lateral geniculate nucleus. Co-staining with the cholera toxin B subunit to label all retinal afferents showed that melanopsin ganglion cells provide most of the retinal input to the SCN, IGL, and lateral habenula and much of that to the OPN, but that other ganglion cells do contribute at least some retinal input to these targets. Staining patterns after monocular enucleation revealed that the projections of these cells are overwhelmingly crossed except for the projection to the SCN, which is bilaterally symmetrical.     

Influence of the visual cortex on responses of retinal ganglion cells in the rat

The objective of the present investigation was to answer the following question: Does the visual cortex affect the neuronal firing of retinal ganglion cells in the rat? To test this hypothesis, the visual cortex was inactivated by a reversible cryoblockade. Action potentials of a ganglion cell were recorded from its axon at the optic tract level prior to, during, and following cortical blockade. The results indicated that indeed the visual cortex influenced the retinal output since its inactivation led to a modification of the firing pattern evoked in response to a flash of light. In most cases the modification was an increase of the bursting pattern of the evoked discharges. By contrast cooling nonvisual areas failed to modify ganglion cells' discharge. A comparison between cortico-geniculate and cortico-retinal feedback loops seems to suggest that the first path is involved mostly with the spatial organization of center-surround receptive fields, whereas the second path is associated with temporal aspects of the retinal responses in the rat.     

Characterization of multiple bistratified retinal ganglion cells in a purkinje cell protein 2‐Cre transgenic mouse line

Retinal ganglion cells are categorized into multiple classes, including multiple types of bistratified ganglion cells (BGCs). The recent use of transgenic mouse lines with specific type(s) of ganglion cells that are labeled by fluorescent markers has facilitated the morphological and physiological studies of BGCs, particularly the directional‐selective BGCs. The most important benefit from using transgenic animals is the capability to perform in vivo gene manipulation. In particular, the Cre/LoxP recombination system has become a powerful tool, allowing gene deletion, overexpression, and ectopic expression in a cell type‐specific and temporally controlled fashion. The key to this tool is the availability of Cre mouse lines with cell or tissue type‐specific expression of Cre recombinase. In this study we characterized the Cre‐positive retinal ganglion cells in a PCP2 (Purkinje cell protein 2)‐cre mouse line. We found that all of the Cre‐positive retinal ganglion cells were BGCs. Based on morphological criteria, we determined that they can be grouped into five types. The On‐ and Off‐dendrites of three of these types stratified outside of the cholinergic bands and differed from directional selective ganglion cells (DSGCs) morphologically. These cells were negative for Brn‐3b and positive for both calretinin and CART retina markers. The remaining two types were identified as putative On‐Off and On‐DSGCs. This Cre mouse line could be useful for further studies of the molecular and functional properties of BGCs in mice. J. Comp. Neurol. 521:2165–2180, 2013. © 2012 Wiley Periodicals, Inc.     

Structure/function relationships of retinal ganglion cells in the cat

Intracellular recording and horseradish peroxidase (HRP) iontophoresis was used to define structure/function relationships for single retinal ganglion cells in the intact cat eye. Fifteen physiologically characterized cells were labeled as follows. Five W-cells had gamma morphology, 6 X-cells had beta morphology, and 1 Y-cell had alpha morphology, and these relationships support earlier conclusions. However, one cell could not be physiologically classified despite beta morphology, one X-cell was not a beta cell, and one Y-cell was not an alpha cell. Whether these unusual structure/function relationships represent an artifact of methodology or complications to be added to prevailing notions requires further study.     

Cell transplantation to replace retinal ganglion cells faces challenges – the Switchboard Dilemma

The mammalian retina displays incomplete intrinsic regenerative capacities; therefore, retina degeneration is a major cause of irreversible blindness such as glaucoma, age-related macular degeneration and diabetic retinopathy. These diseases lead to the loss of retinal cells and serious vision loss in the late stage. Stem cell transplantation is a great promising novel treatment for these incurable retinal degenerative diseases and represents an exciting area of regenerative neurotherapy. Several suitable stem cell sources for transplantation including human embryonic stem cells, induced pluripotent stem cells and adult stem cells have been identified as promising target populations. However, the retina is an elegant neuronal complex composed of various types of cells with different functions. The replacement of these different types of cells by transplantation should be addressed separately. So far, retinal pigment epithelium transplantation has achieved the most advanced stage of clinical trials, while transplantation of retinal neurons such as retinal ganglion cells and photoreceptors has been mostly studied in pre-clinical animal models. In this review, we opine on the key problems that need to be addressed before stem cells transplantation, especially for replacing injured retinal ganglion cells, may be used practically for treatment. A key problem we have called the Switchboard Dilemma is a major block to have functional retinal ganglion cell replacement. We use the public switchboard telephone network as an example to illustrate different difficulties for replacing damaged components in the retina that allow for visual signaling. Retinal ganglion cell transplantation is confronted by significant hurdles, because retinal ganglion cells receive signals from different interneurons, integrate and send signals to the correct targets of the visual system, which functions similar to the switchboard in a telephone network – therefore the Switchboard Dilemma.     

Profound defects in pupillary responses to light in TRPM-channel null mice: a role for TRPM channels in non-image-forming photoreception

TRPM1 is a spontaneously active non-selective cation channel that has recently been shown to play an important role in the depolarizing light responses of ON bipolar cells. Consistent with this role, mutations in the TRPM1 gene have been identified as a principal cause of congenital stationary night blindness. However, previous microarray studies have shown that Trpm1 and Trpm3 are acutely regulated by light in the eyes of mice lacking rods and cones (rd/rd cl), a finding consistent with a role in non-image-forming photoreception. In this study we show that pupillary light responses are significantly attenuated in both Trpm1(-/-) and Trpm3(-/-) animals. Trpm1(-/-) mice exhibit a profound deficit in the pupillary response that is far in excess of that observed in mice lacking rods and cones (rd/rd cl) or melanopsin, and cannot be explained by defects in bipolar cell function alone. Immunolocalization studies suggest that TRPM1 is expressed in ON bipolar cells and also a subset of cells in the ganglion cell layer, including melanopsin-expressing photosensitive retinal ganglion cells (pRGCs). We conclude that, in addition to its role in bipolar cell signalling, TRPM1 is involved in non-image-forming responses to light and may perform a functional role within pRGCs. By contrast, TRPM3(-/-) mice display a more subtle pupillary phenotype with attenuated responses under bright light and dim light conditions. Expression of TRPM3 is detected in Muller cells and the ciliary body but is absent from pRGCs, and thus our data support an indirect role for TRPM3 in pupillary light responses.     

Early remodeling of müller cells in the rd/rd mouse model of retinal dystrophy

We studied the anatomical remodeling and gliosis of retinal Müller cells in the rd/rd mouse model of photoreceptor degeneration. A computational calculation of glutamine synthetase immunoreactivity was developed so we could specifically quantify changes in Müller cell anatomy between control mice (C57Bl/6) and the dystrophic strain. We found no change in the number of Müller cell somata between mice strains, indicating no cell proliferation as a function of development and degeneration. The retinal area occupied by the total Müller cell body (soma and processes) was significantly less in the rd/rd mouse retina compared with control mice. When only the outer retina was considered, we found rd/rd Müller cell processes were dramatically reduced during the cone phase of photoreceptor degeneration. However, at older ages an increase in Müller cell processes was seen. Conversely, glial fibrillary acidic protein (GFAP) expression showed a significant increase during cone degeneration followed by a reduction in older ages. Müller cell electrophysiology, particularly K⁺ currents and membrane potential, was similar between rd/rd and control Müller cells during cone degeneration. Together, these results show that glial remodeling in the rd/rd retina follows separate phases—an initial conservative glial response involving the loss of Müller cells processes, hyperexpression of GFAP, and preservation of normal electrophysiology followed by an active growth of Müller cell processes, glial seal formation, and attenuation of GFAP expression after complete photoreceptor loss. J. Comp. Neurol. 521:2439–2453, 2013. © 2013 Wiley Periodicals, Inc.     

Contributions of retinal direction‐selective ganglion cells to optokinetic responses in mice

In the mouse retina, there are two distinct groups of direction‐selective ganglion cells, ON and ON–OFF, that detect movement of visual images. To understand the roles of these cells in controlling eye movements, we studied the optokinetic responses (OKRs) of mutant mice with dysfunctional ON‐bipolar cells that have a functional obstruction of transmission to ON direction‐selective ganglion cells. Experiments were carried out to examine the initial and late phases of OKRs. The initial phase was examined by measurement of eye velocity using stimuli of sinusoidal grating patterns of various spatiotemporal frequencies that moved for 0.5 s. The mutant mice showed significant initial OKRs, although the range of spatiotemporal frequencies that elicited these OKRs was limited and the response magnitude was weaker than that in wild‐type mice. To examine the late phase of the OKRs, the same visual patterns were moved for 30 s to induce alternating slow and quick eye movements (optokinetic nystagmus) and the slow‐phase eye velocity was measured. Wild‐type mice showed significant late OKRs with a stimulus in an appropriate range of spatiotemporal frequencies (0.0625–0.25 cycles/°, 0.75–3.0 Hz, 3–48°/s), but mutant mice did not show late OKRs in response to the same visual stimuli. The results suggest that two groups of direction‐selective ganglion cells play different roles in OKRs: ON direction‐selective ganglion cells contribute to both initial and late OKRs, whereas ON–OFF direction‐selective ganglion cells contribute to OKRs only transiently.     

Melanopsin ganglion cell outer retinal dendrites: Morphologically distinct and asymmetrically distributed in the mouse retina

A small population of retinal ganglion cells expresses the photopigment melanopsin and function as autonomous photoreceptors. They encode global luminance levels critical for light‐mediated non‐image forming visual processes including circadian rhythms and the pupillary light reflex. There are five melanopsin ganglion cell subtypes (M1–M5). M1 and displaced M1 (M1d) cells have dendrites that ramify within the outermost layer of the inner plexiform layer. It was recently discovered that some melanopsin ganglion cells extend dendrites into the outer retina. Outer Retinal Dendrites (ORDs) either ramify within the outer plexiform layer (OPL) or the inner nuclear layer, and while present in the mature retina, are most abundant postnatally. Anatomical evidence for synaptic transmission between cone photoreceptor terminals and ORDs suggests a novel photoreceptor to ganglion cell connection in the mammalian retina. While it is known that the number of ORDs in the retina is developmentally regulated, little is known about the morphology, the cells from which they originate, or their spatial distribution throughout the retina. We analyzed the morphology of melanopsin‐immunopositive ORDs in the OPL at different developmental time points in the mouse retina and identified five types of ORDs originating from either M1 or M1d cells. However, a pattern emerges within these: ORDs from M1d cells are generally longer and more highly branched than ORDs from conventional M1 cells. Additionally, we found ORDs asymmetrically distributed to the dorsal retina. This morphological analysis provides the first step in identifying a potential role for biplexiform melanopsin ganglion cell ORDs.     

The course of axons of retinal ganglion cells within the optic nerve and tract of the chick (Gallus gallus)

Small laser lesions placed in the posthatch chicken retina resulted in axotomy and then death of all ganglion cells located in a sector peripheral to the primary damage. With the use of silver techniques, the patterns of degenerating retinal fibers in the optic nerve, chiasm, and optic tract were examined. In the proximal part of the optic nerve, radial retinal lesions resulted in a sheet of degenerating axons along the rostrocaudal extent of the nerve. The position of degenerating axons was related to the site of their entry in the optic nerve head with an overlapping distribution of degenerating fibers entering the optic nerve head from equivalent points from the temporal and nasal sides. In the optic chiasm, the distribution of fibers was similar to that seen in the proximal part of the optic nerve. In the optic tract there was a similar mixing of fibers from opposite sides of the retina. The ventral, nasal and temporal retinal fibers lay in the superficial part of the tract whereas the fibers from the nasal and temporal dorsal retina ran in the deeper, medial aspect of the tract. The central-to-peripheral axes of the retina were mapped along the rostrocaudal axis of the tract. As the tract approached the tectum degenerating fibers from single retinal lesions did not always remain together. In the case of a lesion in the ventral nasal retina, degenerating fibers split into two bundles located at opposite ends of the tract only to reunite at their terminal regional at the caudal pole of the tectum.(ABSTRACT TRUNCATED AT 250 WORDS)     

Periodic and nonperiodic burst responses of frog (Rana pipiens) retinal ganglion cells

Neural activity of class 3 retinal ganglion cells was recorded in frog optic tectum, using extracellular microelectrodes. The stimuli were rectangular patches of contrast (light-on-dark or dark-on-light), applied within the previously determined receptive fields, for periods ranging from a few milliseconds to several seconds. ON and OFF responses were recorded for as long as 1 s following stimulation. Poststimulus time histograms revealed two types of responses, labeled periodic and nonperiodic bursters. The periodic bursters were characterized by periods of high activity separated by silent or near-silent intervals. The bursts occurred rhythmically with frequencies roughly between 15 and 50 Hz. Nonperiodic bursters generally showed both broad and sharp peaks in activity, but no regular periodicities. Activity profiles were flat initially, with silent periods appearing after the first few stimulus presentations, suggesting an inhibitory nature of the bursting process. The records were shown to combine the activities of several neurons. Analysis of the waveforms in real time made possible isolation of some units. In these cases, neurons exhibited a high degree of selective synchrony, i.e., the sharing of a portion of the activity profile, and notable differences at other times. These data have implications for the processing of visual information.     

Distribution and morphology of retinal ganglion cells in the Japanese quail

A ganglion cell density map was produced from the Nissl-stained retinal whole mount of the Japanese quail. Ganglion cell density diminished nearly concentrically from the central area toward the retinal periphery. The mean soma area of ganglion cells in isodensity zones increased as the cell density decreased. The histograms of soma areas in each zone indicated that a population of small-sized ganglion cells persists into the peripheral retina. The total number of ganglion cells was estimated at about 2.0 million. Electron microscopic examination of the optic nerve revealed thin unmyelinated axons to comprise 69% of the total fiber count (about 2.0 million). Since there was no discrepancy between both the total numbers of neurons in the ganglion cell layer and optic nerve fibers, it is inferred that displaced amacrine cells are few, if any. The spectrum in optic nerve fiber diameter showed a unimodal skewed distribution quite similar to the histogram of soma areas of ganglion cells in the whole retina. This suggests a close correlation between soma areas and axon diameters. Retinal ganglion cells filled from the optic nerve with horseradish peroxidase were classified into 7 types according to such morphological characteristics as size, shape and location of the soma, as well as dendritic arborization pattern. Taking into account areal ranges of somata of each cell type, it can be assumed that most of the ganglion cells in the whole retinal ganglion cell layer are composed of type I, II and III cells, and that the population of uniformly small-sized ganglion cells corresponds to type I cells and is an origin of unmyelinated axons in the optic nerve.