Number and dendritic morphology of retinal ganglion cells that survived after axotomy in adult cats

Number and Dendritic Morphology of Retinal Ganglion Cells that Survived after Axotomy in Adult Cats

Masami Watanabe,’,* Hajime Sawai,*,+ and Yutaka Fukuda2

‘Department of Physiology, Institute for Developmental Research, Aichi Human Service Center, Kasugai, Aichi 480-03, and *Department of Physiology, Osaka University Medical School, Suita, Osaka 565, Japan

SUMMARY

Retinal ganglion cells (RGCs) of adult cats were labeled by injection of diI into the proximal stump of completely transected optic nerves. Approximately 2% to 5% of the RGC population appeared viable 2 months after these axotomies, based on diI retention. The morphological type and dendritic arbor of these surviving RGCs were examined after intracellular injections of Lucifer Yellow into diI-labeled RGCs. Postaxotomy survival rate was much higher for d i k e cells than for &like cells. How- ever, in one of four retinas examined, a large number of RGCs seemed to survive axotomy, and among these, #? cells survived at an unusually high rate. Dendritic arbors of surviving RGCs were also examined after intracellular

injection of horseradish peroxidase. Some dendrites of these RGCs lacked branches and were thin in caliber. Other dendrites displayed many spiny processes and bulbous swellings. Essentially, these results confirm the previous suggestion that a: cells survive axotomy longer than #? cells. The ability of a: cells to regenerate axons may thus be attributable to their relatively high resistance to axotomy. The atypical dendritic profiles seen after optic nerve transection may reflect either degeneration or regrowth of dendrites. o 1995 John Wiley &Sons, Inc.

Keywords: optic nerve section, retinal ganglion cells, alpha cell, beta cell, adult cat.

INTRODUCTION

Certain population of retinal ganglion cells (RGCs) survive in adult mammals for a relatively long time after complete transection of the optic nerve (ON). Normal-looking RGCs and optic fibers have been observed in the rabbit, for example, 2 years after ON transection (Scherer and Schnitzer, 199 1). In rats, a small population of RGCs persisted 20 months after the axotomy (Villegas-Ptrez et al.,

Received September 9, 1994; accepted January 11, 1995 Journal of Neurobiology, Vol. 27, No. 2, pp. 189-203 (1995) 0 1995 John Wiley & Sons, Inc. CCC 0022-3034/95/020189- 15

* To whom correspondence should be addressed. + Present address: Department of Welfare and Health Sci-

ence, Okayama Prefectural University, 1 1 1 Kuboki. Souja 7 19- 1 1, Japan.

1993). On the basis of electroretinogram recordings in cats, Maffei and Fiorentini (198 1, 1982) reported that X (that is, p) cells survived axotomy longer than other types of RGCs. In subsequent studies of the cat RGCs (Hollander et al., 1984, 1985), large RGCs with primary dendrites resembling those of a cells (Boycott and Wassle, 1974) persisted more than 15 months after ON transection, whereas very few medium-sized RGCs (that is, p cells) were detected. However, because soma size of RGCs changes after axotomy (Murray and Grafstein, 1969; Maffei et al., 1990; Bahr et al., 1992), soma size alone may not suffice to classify RGC cell types. The extent to which these earlier and later studies of axotomized cat RGCs differ thus remains unclear.

The present study was designed to measure in adult cats numbers of RGCs persisting after com-

189

190 Wuranahe et a/.

plete ON transection, proportions of morphological types among the surviving RGCs, and morphological changes in these cells. To compare the results with those of our studies of ON regeneration after peripheral nerve transplantation survival time was fixed at 2 months (Watanabe et al., 1991, 1993a, 1993b). RGCs surviving axotomy were counted after retrograde labeling with diI, a fluorescent carbocyanine dye that can be retained 2 months in the RGCs (Honig and Hume, 1986; Wa- tanabe et al., 1991). Relative numbers of RGCs were estimated from entire populations of RGCs injected intracellularly with Lucifer Yellow within small patches of retina. Finally, the morphology of dendrites of surviving RGCs was examined in cells injected intracellularly with horseradish peroxidase (HRP).

A preliminary report of the present study appeared as an abstract form (Watanabe and Fukuda, 1993).

MATERIALS AND METHODS

Eleven adult cats of either sex, 2 to 3.5 kg in weight and 1 to 2.5 years old, were used.

Optic Nerve Transection and Labeling of Retinal Ganglion Cells

Unilateral ON transections were performed on cats se- dated with an intramuscular injection of60 mg ketamine hydrochloride and anesthetized with a gas mixture of 1% to 2% halothane, nitrous oxide ( I L/min), and oxygen (1 L/min). The heart rate was monitored during anesthesia. The left ON of nine cats (6,7,8,9, 10, 1 I , 12, 18, and 19) was exposed and completely cut at 4 to 6 mm from the eyeball with fine scissors. Care was taken not to injure blood vessels, especially the ophthalmic artery that en- ters the sclera from the ventral margin of the ON. Com- pleteness of transection was ensured by trimming a seg- ment of the ON stump attaching to the eyeball approximately 1 mm off in length, and turning this stump upward. Ten microliters of a suspension of diI ( 1,l’-dioc- tadecyl-3,3,3’,3’-tetramethylindocarbocyanine perchlo- rate, Molecular Probes; sonicated at 10 mg/ml in saline containing 1% Triton X) was injected into the optic stump with a 10 p l Hamilton microsyringe (Watanabe et al., 199 I ). A small piece of Gelfoam soaked in 5 pl of diI suspension was placed in the space behind the optic stump. Penicillin powder (400 kU) was administered in the orbital space, and the skin was closed with sutures.

In two cats (15 and 16), RGCs were labeled with dil injected into the lateral geniculate nucleus (LGN) I week prior to ON transection (performed as just described).

The head of each animal was fixed on a stereotaxic head holder (Narishige SN-3A) under gas anesthesia (as just described). After the right LGN was located with a glass microelectrode by monitoring light-evoked field potentials, the electrode was replaced with a 10 p1 Hamilton syringe. A total of 5 pl of di1 suspension was injected in 1 to 2 pl shots in each of three to four tracks penetrating the LGN.

In Vifro Preparation for lntracellular Injection of Lucifer Yellow or HRP

At 54 to 67 days after ON transection, the cats were anesthetized by the procedure just described, and the left eye was enucleated. The cats were then killed by an overdose of sodium pentobarbital. The retina was dissected from the left eye in oxygenated (95% oxygen, 5% carbon diox- ide) Ames medium (Sigma, A 1420), and affixed on a plate of 30% gelatin in a chamber where oxygenated Ames medium was superfused. The right eyes of two cats (1 5 and 16) were also enucleated (before the cats were killed), and their retinas were dissected and fixed with 1 % paraformaldehyde in 0.1 M phosphate buffer.

lntracellular Injection of Lucifer Yellow

DiI-labeled RGCs were filled intracellularly with Lucifer Yellow in retinas of seven cats (6, 7,8, 9, 10, 15, and 16), using a procedure similar to that described by Tauchi and Masland (1 984). Microcapillary electrodes were filled with 3% Lucifer Yellow CH (Sigma, L 0259) in 50 mM Tris buffer (pH 7.2) and beveled to a tip resistance of 60 to 80 MQ with a K. T. Brown beveler (Sutter BV- 10). The diameter of each diI-labeled cell was measured with an ocular micrometer, and any cell with soma1 diameter exceeding 8 p m was regarded as an RGC (Hughes, 1981; Maffei et al., 1990). Each cell to be Luci- fer-filled was placed in the center of the field view under green excitation (excitation filter 5 10 to 560 nm and barrier filter 590 nm long pass, Nikon G-2A) and was pene- trated with a Lucifer Yellow-filled electrode; dye was ejected by passage of direct negative currents (3 to 10 nA) for 1 to 2 min. After dendritic arbors were visualized with Lucifer Yellow, the morphological type was determined under blue excitation (excitation filter 450 to 490 nm and barrier filter 520 nm long pass, Nikon B-2A), and the cell position was recorded in terms of horizontal and vertical coordinates of the microscope stage. These pro- cedures were repeated for all nearby cells with soma1 diameter larger than 8 Fm until the electrode tip clogged. After Lucifer Yellow was injected into cells in several patches, the retina was pealed off of the gelatin plate and fixed for 1 h with 1 % paraformaldehyde in 0.1 M phosphate buffer (pH 7.4). The fixed retinas were stored in phosphate buffer supplemented with sodium azide (0.1%) at 4°C for further examination.

Retinal Ganglion Cells Surviving Axotomy 191

lntracellular Injection of HRP

Retinas of four cats (1 1, 12, 18, and 19) were used for intracellular injections of HRP (Rodieck and Watanabe, 1986, 1993; Watanabe and Rodieck, 1989). Rhodamine isothiocyanate-conjugated HRP (Sigma, P 503 1) was dissolved in 50 mM 3-(N-morpholino)propanesulfonic acid (MOPS) solution, at final concentration of 1000 U in 130 pl. The MOPS solution was adjusted to pH 7.0 with lithium hydroxide and supplemented with 0.4% Lucifer Yellow CH. Microcapillary electrodes were filled with the HRP solution and were beveled to a tip resistance of 60 to 100 MO as already described. Under blue excitation (Nikon B-2A), diI-labeled cells were pene- trated with an electrode through which negative current (1 to 2 nA) was briefly applied to eject Lucifer Yellow. After observing that soma and primary dendrites were filled with Lucifer Yellow, positive current pulses (3 to 10 nA; 1 s on, 1 s off) were applied to eject HRP for 1 to 2 min. HRP was injected into 20 to 40 cells in one retina, then the retina was pealed off of the gelatin plate, fixed for 1 h with 1% glutaraldehyde in 0.1 Mphosphate buffer (pH 7.4). The retina was reacted with 3,3'-diaminobenzi- dine tetrahydrochloride (0.5 mg/ml) and hydrogen per- oxide (0.5 mg/ml) in 0.1 M phosphate buffer, whole- mounted on a gelatin-coated slide glass, dehydrated, and mounted in Entellan New (Merck, 7961) with the same procedure as described previously (Watanabe et al., 1993b).

Count of Surviving RGCs

After Lucifer Yellow injections and aldehyde fixation, four fixed retinas (of cats 6, 8,9, and 10) were mounted in 0.1 M phosphate buffer and cover slips were placed. RGCs labeled with diI were examined under green excitation (Nikon G-2A), and their positions were recorded with a microscanner (Sapporo Breweries) connected to a personal computer. Labeled cells were plotted with an X- Y plotting program developed by R. W. Rodieck (Wata- nabe and Rodieck, 1989; Rodieck and Watanabe, 1993). The microscanner reads the stage coordinates at 10 pm acuity with 20 wm reproducibility. For this acuity, labeled RGCs within 20 pm apart from the other RGCs, unless they were in a cluster, were regarded as being dou- bly counted, and their coordinates were deleted from the data.

RESULTS

Morphology of Dil-Labeled RGCs

RGC somata with visible levels of retrogradely transported diI were observed 2 months after complete transection of the ON. This diI was particu- lated in appearance and confined to the cytoplasm

and primary dendrites in most cells [Fig. l(A and inset)]. Some RGCs displayed brightly labeled dendrites [Fig. l(B)]. Several RGCs extended two processes in the nerve fiber layer. In the cell shown in Figure 1(C), one process projected toward the optic disc like a normal axon; the other projected in the opposite direction and ended in a swelling. In a very few cases, whole diI-labeled dendritic arbo- rizations were seen with no attached soma [Fig. l(D)].

After diI application to transected ONs, one might question whether diI-containing cells included displaced amacrine cells or non-neural cells that had taken up dye molecules released from degenerated RGCs. Lucifer Yellow injections (see later) showed that all cells with diI accumulations sufficient to produce bright soma1 fluorescence, without heavy nuclear fluorescence [Fig. 1(A, inset)], sent an axon through the nerve fiber layer toward the optic disc. We have therefore regarded cells with cytoplasmic diI accumulations, and somata larger than 8 pm in diameter under a fluorescence microscope (Hughes, 198 1; Maffei et al., 1990), as surviving RGCs.

Number and Distribution of Surviving RGCs

In three of five retinas examined, consistently few RGCs (1660 to 3287 per retina) survived axotomy (Table 1). The total number of RGCs surviving in the fourth retina examined (cat 7) could not be counted because a peripheral portion of the retina was damaged. The densities of surviving RGCs in several sectors of this retina, however, were close to those of one of the other retinas (9 in Table 1). This suggests that similar numbers of RGCs survived axotomy in these retinas (7 and 9). In the fifth retina examined (8 in Table l), diI-labeled RGCs were considerably more numerous (5585).

Maps of the RGCs surviving in two retinas (6 and 8) are shown in Figure 2. Because most RGCs were lost from the retinas examined, location of the area centralis was estimated from the blood vessel pattern [see Fig. 7(B) of Hughes, 19751. If this as- sumption is correct, then surviving RGCs were dis- tributed most densely in the vicinity of the optic disc of both retinas, and very sparsely in the peripheral retina. The distribution in one other retina (9, not shown) was similar to that seen in retina 6 of Figure 2.

792 Watanahe et al.

Figure 1 Photomicrographs of diI-labeled cells under 5 10 to 560 nm excitation filter and 590 nm long pass barrier filter (Nikon G-2A). (A) Population of diI-labeled RGCs in retina 7. Note that primary dendrites (arrows) of some cells were filled with diI particles. The soma to the left in the inset shows diI accumulation in the cytoplasm, with little or none in the nucleus. (B) An RGC in retina 8 extending labeled dendrites 300 pm long. (C) An RGC in retina 9 with two processes in the nerve fiber layer. The axon (left) is directed to the optic disc (large arrow); a second process (arrowheads) extends in the opposite direction and ends with a bulbous swelling. Thin dendrites were also labeled with diI (small arrows). (D) Brightly labeled dendrites in retina 9. The dendrites were clearly labeled and had many varicosities, but the soma was not found in the center of primary dendrites (arrows). Scale bars (A, B, C, D) = 50 pm and (A inset): 20 pm.

Retinal Ganglion Crlls Surviving Axot0m.y 193

Table 1

Experimental Survival Total No. of dil-

Number of Surviving Retinal Ganglion Cells

No. Sex Time (days) Labeled Cells

6' M 58 2472 8' F 61 5585 9 F 57 1660

10 F 58 3287 ~ ~~~~

' Distributions of ganglion cells were shown in Figure 2(A, B), respectively.

Dendritic Morphology of Lucifer Yellow-Injected RGCs

To identify cell types among surviving RGCs, Luc- ifer Yellow was injected into all cell bodies exceeding 8 pm diameter within small patches at various locations across the retina. Lucifer Yellow injections are most likely to fill thick- and medium-caliber dendrites, and least likely to fill long, thin dendrites at their full length. Unequivocal identifica- tion of morphological types was therefore possible only for the cell-types termed a and @ by Boycott and Wassle (1974). Other RGCs were classified as those neither a nor @ (NAB).

Primary dendrites of a cells were thick and straight, and multibranched over relatively wide fields [Fig. 3(A)]. Dendrites of most RGCs classified as @-type were bushy and branched frequently within relatively smaller fields. The dendritic arbor

of some of these cells seemed partial [Fig. 3(B)], in that similar arbor shapes are rarely observed in normal cat retinas. Despite this unusual morphology, the cell in Fig. 3(B) was categorized as a ,8 cell because its dendrites resembled those of normal /? cells, and those of @ cells with regenerated axons shown in Fig. 5(A) of Watanabe et al. (1993b).

A variety of NAB cells are shown in Figure 3(C through F). One of these [Fig. 3(C,D)] displayed dendrites that individually resembled those of normal @ cells. This cell was classified as NAB, however, because some of these dendrites ramified within the outer sublamina of the inner plexiform layer (IPL) [Fig. 3(C)], whereas other dendrites of the same cell ramified within the inner sublamina of the IPL [Fig. 3(D)]. The ,8 (or X) RGCs in the adult cat retina neither send dendrites into both sublaminae of the IPL, nor generate ON-OFF responses to light flashes (Fukuda and Stone, 1974; Stone and Fukuda, 1974; Fukuda et al., 1984). The two cells shown in Figure 3(E) were classified as NAB because each had a smaller number of primary dendrites and fewer branches than typical a cells. This was commonly observed in NAB cells. Two other NAB cells are shown in Figure 3(F): these extend thin dendrites off to one side of the soma, either in the inner or outer sublamina of the IPL (the upper and lower cell, respectively, in this micrograph). The dendrites of these RGCs resemble those of the OFF tonic W-cell in Figure 6(C) of Fukuda et al. (1,984).

Figure 2 Distribution of RGCs that survived 2 months after axotomy in retinas 6 (A) and 8 (B). Each dot represents a diI-labeled RGC. The area centralis (AC) is located at the projected intersection of each pair of arrows marked AC. *Optic disc.

194 Wutunahe ct u[.

Figure 3 Photomicrographs of Lucifer Yellow-injected RGCs. The photographs were taken under 450 to 490 nm excitation filter, 520 nm long pass barrier filter (Nikon B-2A) after fixation with paraformaldehyde. Arrows indicate axons. (A) An a-type cell in retina 9, which had thick, smoothly tapered primary dendrites. Dendritic field diameter was approximately 150 pm. The axon (large arrows) was out of focus and very dim. (B) A p cell in retina 6. Note that dendrites ramified frequently in a part of their field. (C, D) An NAB cell in retina 10. Although the dendritic arbor resembled that of 0 cells, some dendrites ramified in the outer sublamina of the IPL (C), whereas other dendrites of the same cell ramified in the inner sublamina of the IPL (D). The bistratified dendrites of this cell resemble those of early developing RGCs in the kitten retina (Bodnarenko and Chalupa, 1993). (E) Two NAB cells in retina 7 having fewer dendrites than typical N cells. Arborization ofthe primary dendrites of the cells resembled that of N cells. (F) Two neighboring NAB cells with thin dendrites in retina 10. Dendrites of the upper cell extended in one direction within the inner sublamina of the IPL, whereas those of the lower cell ramified in the outer sublamina of the IPL. Scale bars; (A, E, F): 50 pm; (B, C, D): 20 Nm.

(Y and @ Cell Types among Surviving RGCs

Lucifer Yellow was injected intracellularly into all 399 diI-labeled cells with somata exceeding 8 pm in diameter, in 23 localized areas in four retinas.

These areas included patches near the area centralis of each retina and patches in other retinal sectors. Table 2 summarizes proportions of three types (LY, p, and NAB). a cells comprised 8% to 24% of these cells, and thus were severalfold more numerous than a cells in the normal retinas (3% to


Figure 3 (Continued)

5%) (Hughes, 1981). cells were rarely encoun- tered in three retinas (2% to 4%), but unusually numerous (32%) in one retina (8). NAB cells were far more numerous than a and p cells in all five retinas in Table 2. No remarkable difference in the proportions of cell types was observed between different quadrants of a given retina.

Because some RGCs degenerate very rapidly after axotomy (Villegas-Pkrez et al., 1993; Takano and Horie, 1994), one might imagine that the cut

ends of axons of RGCs retract or degenerate before accumulating detectable amount of diI and that p RGCs consequently persist in the retina un- labeled. To test this possibility, RGCs were retrogradely labeled by injection of di1 into the contra- lateral lateral geniculate nucleus 1 week before unilateral O N transection. The axotomized retinas examined (1 5 and 16) contained relatively few, if any, p cells (Table 2). By contrast, the right (unop- erated) retinas contained many diI-labeled @-like

196 Watanabe et a1

Figure 4 Camera lucida drawings of HRP-injected RGCs. Axons (indicated with arrows, not drawn at full length), directing to the optic disc, could be followed up to 1100 pm (A), and 980 pm (C), respectively, to a faded terminal. The axon of the cell B ended at 480 pm length with a bulbous swelling. Eccentricities and retinal quadrants are: (A) 7.5 mm in the lower nasal, (B) 10.7 mm in the upper nasal, (C) 9.5 mm in the lower nasal. (A and B) a cells. Dendrites of the

Retinal Ganglion Cells Siirliiving Axotorn.v 197

Table 2 Proportion of Alpha and Beta Cells in Surviving Retinal Ganglion Cells

Proportion of Cell Types (%) No. Lucifer Experimental Yellow-

Sex Alpha Beta Not Alpha or Beta Injected Cells No.

Labeled with optic stump injection 6 M 21.4 4. I 7 M 8.3 I .7 8 F 8.6 32.3 9 F 24.1 1.7

10 F 19.8 2.5

Labeled with LGN injection 15 F 23.3 2.3 16 M 15.7 0

74.5 90.0 59.1 74.1 77.8

74.4 84.3

98 60 93 58 81

43 5 1

cells (not shown), that is, RGCs with the medium- sized round soma and concentric nucleus.

Dendritic Morphology of HRP-injected RGCs

To examine the morphology of dendrites in more detail, RGCs were filled with HRP (see Methods) after axotomy and diI-labeling. Although HRP was intracellularly injected into 117 RGCs in four retinas, the dendrites were clearly filled in only 29 of these cells. Of these, 15 cells displayed dendrites resembling those of a cells, two cells were classified as y cells (Boycott and Wassle, 1974), and 12 RGCs did not resemble any previously described types of RGCs. No p cells were recovered from these HRP injections.

Two examples of HRP-injected a cells are shown in Figure 4(A,B). The soma and dendrites of the cell in Figure 4(A) were endowed with many fine processes resembling dendritic spines. The dendrites of the RGC in Figure 4(B) were thick in caliber and smooth-surfaced, but dendritic branches in the lower right field seemed to have degenerated. Some distal dendrites of the cell in Fig- ure 4(B) ended in a bulbous swelling. Examples of

the swellings are depicted in Figure 6(A) at higher magnification.

The dendritic arbor of the cell in Figure 4(C) resembled that of y cells, except for the presence of many fine distal processes. The axon arose from the thinnest primary dendrite. This observation suggests that the axon may have emerged anew from the primary dendrite after axotomy.

Figure 5 shows three examples of RGCs that did not resemble the cell types a, p, or y. The cell in Figure 5(A) had four primary dendrites, plus some secondary dendrites that emerged at approximately a right angle from the shaft of the primary dendrites. Although the dendritic morphology of this cell resembled that of c cells projecting to the cat geniculate wing (Rodieck and Watanabe, 1986), the peripheral dendrites were bushier than those of normal E cells. Two unusual features of the cell in Figure 5(B) include dendrites of remarkably thicker caliber than its other dendrites and those of normal cy cells, and dendrites with many spinelike processes. One primary dendrite was thicker and more filled with spines than the others, yet did not branch [Figure 6(B)]. The dendrites of the cells in Figures 5(B,C) were endowed with spiny processes and their distal processes had lobulated swellings with spines, which are unusual features in den-

cell in (A) had many fine processes that resembled dendritic spines. Some dendrites of the cell in (B) had small bulbous swelling (triangles). The portion of this drawing labeled “6A” is seen at higher magnification in Figure 6(A). Dendrites in the lower right field ofthe cell (B) seem to have degenerated. (C) A 7-like cell. This RGC was endowed with long, straight primary dendrites that ramified fine processes. The axon emanated from a fine primary dendrite (triangles) that branched in the inner plexiform layer. This suggests that the axon grew anew from the primary dendrite, which was under degeneration. Scale bar = 100 pm.

198 Watanabe et al.

drites of a cells. Figures 6(C,D) are magnified views of dendrites of the cell in Figure 5(B), whereas Fig- ure 6(E) shows the apposed terminal swellings of two dendrites seen in Figure 5(C). Some dendrites of NAB cells ended with a bulbous swelling [Fig. 5(A)] resembling those in the a cell [Figs. 4(B) and 6(A)].

DISCUSSION

The main purpose of the present study was to de- termine which types of cat RGCs remain viable after ON transection and to examine the basic morphology of these surviving RGCs.

Number of Surviving RGCs

Numbers of RGCs in the normal cat retinas range from 120,000 to 190,000 (Hughes and Wassle, 1976; Stone, 1978; Hughes, 1981; Chalupa et al., 1984). The numbers of RGCs that retained diI after ON transection in the present study (1660 to 5585) thus constitute 1% to 4% of the normal RGC population. Because it is reported that more than 10% of the total RGCs survive 2 months after transection in the rat (Villegas-Ptrez et al., 1993), the cat RGCs may be less tolerant of axotomy than rat RGCs.

An unusually large number of viable RGCs were found in one of five retinas examined after ON transection (Table 1). We have no unique explanation for this datum. The number seen nears those found in retinas after axonal regeneration with peripheral nerve grafts (Watanabe et al., 1993b). Be- cause neurotrophic factors secreted by Schwann cells can enhance survival of axotomized RGCs (Assouline et al., 1987; Berry et al., 1988; Maffei et al., I990), one explanation for different RGC survival rates after ON transection could be due to different levels of such neurotrophic effect (such as after peripheral nerve damage during ON transection).

Morphological Changes of Surviving RGCs

Morphological similarities as well as differences between RGCs in normal and ON-sectioned retinas were revealed by intracellular HRP injections. The dendrites of typical a cells were smoothly tapered, contoured and large-calibered where they emerged

from the soma and formed dendritic fields as large as those of normal a cells (Boycott and Wassle, 1974; Fukuda et al., 1984; Rodieck and Watanabe, 1986). The dendritic morphology of some a cells thus appeared to have been well-preserved [Figs. 3(A), 4(A), 4(B)]. Dendrites of some p cells [Fig. 3(B)] degenerated in a part of the field but were bushy enough to be classified as the type.

Dendrites of some RGCs in ON-sectioned retinas were endowed with many spiny processes and bulbous swellings at distal ends, whereas some dendrites seemed to have degenerated in a part of the field. In addition to dendritic changes, excess, axon-like processes were observed in the nerve fiber layer when the morphology was visualized by Lucifer Yellow injections, as well as in RGCs labeled with diI [Fig. l(C)].

Similar morphological changes were observed in cat RGCs after axonal regeneration into peripheral nerve grafts (Watanabe et al., 1991, 1993b). The dendrites of some of these RGCs seem to have lost processes in a part of the field, whereas the distal dendrites of other RGCs were endowed with fine processes. The axon of one a cell emanated from a primary dendrite so much thinner than other dendrites that it seemed to be under degeneration [Fig. 3(A) in Watanabe et al., 19911. Thick, axon-like processes in the nerve fiber layer were also observed, as in rat (Thanos, 1988; Tabata and Fu- kuda, 1992), hamster (Cho and So, 1992), and cat (Watanabe et al., 1993b) RGCs with regenerated axons. The morphological changes observed in RGCs with regenerated axons probably resulted from the ON transection.

Proportion of a, ,8, and NAB Cells in Surviving RGCs

Lucifer Yellow-injected RGCs in the present study were classified according to the same criteria as those used in our previous study of RGCs with regenerated axons (Watanabe et al., 1993b). Lucifer Yellow injection into all RGCs within localized retinal areas enabled us to classify cells on the basis of dendritic morphology. Relative numbers of a-, p-, and NAB-type cells in retinas after ON-transection, ON-transection plus peripheral nerve trans- plants (Watanabe et al., 1993b), and in the normal retinas (Hughes, 198 1 ) are summarized in Figure 7. On average, a cells comprised 16% of the RGCs persisting after ON transection alone (upper bar). This is four times higher than the proportion in the


A

c

Figure 5 Camera lucida drawings of HRP-injected RGCs that could not be classified into any known types. Axons [indicated with arrows in (A and C), not drawn at full length], directing to the optic disc, could be followed up to 220 pm (A), and 2300 pm (C) to a faded terminal. The axon of the cell (B) (not drawn) emanated over one of the primary dendrites and ran up to 600 pm. The sectors labeled “6C,” “6D,” and “6E” are shown at higher magnification in Figure 6(C, D, and E), respectively. Eccentricities and retinal quadrants are: (A) 9.9 mrn in the lower nasal, (B) 6.5 mm in the upper nasal, (C) 6.7 mm in the upper temporal. Primary dendrites of the cell in A resembles those of the c cell projecting to the geniculate wing. Some ofthe dendrites also ended in bulbous swellings (triangles). Dendrites ofthe cell in (B) had many fine processes resembling dendritic spines. Dendntic calibers in the upper right field were thick, but those in the lower left of this panel were thin. One primary dendrite (indicated with an arrow and “6B’) was extremely thick and had many spiny processes. Several thick dendrites had branched bulbous swellings at distal ends. Dendrites of the cell in (C) were long and had bulbous swelling and curly processes at the distal end (indicated with rectangle 6E). Scale bar = 100 pm.

Figure 6 Photomicrographs of distal dendrites of HRP-filled RGCs indicated with the rec- tangles and the arrow in Figures 4 and 5.(A) Bulbous endings (arrows) at the distal dendrites of the (Y cell shown in Figure 4(B). (B) A thick and short primary dendrite with many spiny processes of the cell in Figure 5(B). The process did not give rise to secondary dendrites. (C, D) Irregular-shaped terminal processes of the cell in Figure 5(B). The processes had many bouton- like swellings and branches. (E) Two apposing terminal swellings of distal dendrites of the cell in Figure 5(C). Scale bar: (A, D, E) 20 pm; (B, C) = 10 pm.

normal retinas (middle bar). /3 cells comprised 9% of the RGCs after ON transection (upper bar), one- fifth the proportion seen in the normal retinas

(middle bar). Our results agree fundamentally with the recent demonstration by Silveira et al. ( 1994) that most P-type RGCs disappear in the cat retina

Retinal Gunglion Cells Stirviving Axotomy 201

Axotomized N = 390

Normal

Regenerated N=443

a P NAB Figure 7 Diagram of proportions of CY (dark shading), /3 (no shading), and NAB (light shading) types of RGCs surviving axotomy (Axotomized), RGCs in the normal retina (Normal), and RGCs with regenerated axons into peripheral nerve grafts (Regenerated). Proportions of axotomized RGCs are averages of the values in five retinas (6 through 10) shown in Table 2, those in the normal retina are averages of values in Figure 5 of Hughes (198 I), and those in RGCs with regenerated axons are values in Table 2 of Watanabe et al. (1993b). Note that proportions of CY cells in both axotomized and regenerated are higher than that in the normal retina, whereas proportion of p cells in the axotomized is the lowest in the three.

within 2 weeks after axotomy, but do not support that X (6) cells remain viable longer after axotomy (Maffei and Fiorentini, 198 1 , 1982). The types of RGCs that contribute to electroretinograms and field potentials elicited by sinusoidal gratings of high space frequency nevertheless remain to be identified.

After ON transection, a and p RGCs were out- numbered by RGCs that were clearly NAB. Except for the few NAB cells that resembled previously described cell types [Figs. 5(C) and 6(A)], we cannot deduce from our data whether NAB cells arose from a, 0, or other known cell types by modifica- tion of existing dendrites or by replacement of orig- inal dendrites by neurites of unclassifiable mor- phologies.

Cellular Difference in Resistance to Axotomy and Ability of Regeneration

Numerous studies have concluded that a RGCs in the cat and their counterparts in the mammalian retinas (Peichl et al., 1987) remain viable and are more capable of axonal regeneration than other types of RGCs (Hollander et al., 1984, 1985; Cottee et al., 1991; Silveira et al., 1994) (Fig. 7). Recent electrophysiological recordings from regenerated optic fibers teased from peripheral nerve grafts are consistent with this morphological result (Wata-

nabe et al., 1993a): in a sample of 47 fibers, 15% were Y-type (a) and 57% were X-type (p). Axot- omy-induced changes in axonal profiles observed with the electron microscope and in antidromically evoked field potentials in the rat ON also suggest that retrograde degeneration proceeds more slowly in the RGCs with thicker axons than in the RGCs with finer axons (Sugioka et al., 1993). The greater ability of a cell to survive axotomy may offer these cells a greater chance of successful axonal regeneration.

In conclusion, a and cells appear to differ not only in terms of their well-known morphological and physiological properties, but also in biological properties to react on axotomy. One of these differences may be differential sensitivity to neurotrophic factors (Carmignoto et al., 1989; Morse et al., 1993; Mansour-Robaey et al., 1994). Indeed, Mey and Thanos (1993) demonstrated that the per- centage of large RGCs surviving ON transection was higher in the retinas in which neurotrophic factors (one or both of brain-derived neurotrophic factor and ciliary neurotrophic factor) were injected into the vitreous humor than in the retinas without the injection. Differences in sensitivity to neurotrophic factor or factors remain to be detected among well-established RGC types of the cat retina.

We thank Dr. A. T. Ishida for comments on the manuscript. This study was supported by Grants-in- Aid for Scientific Research (03044097, 0367084 1, 05304040) from the Ministry of Education, Culture and Science of Japan, and a Grant from Uehara Foundation for Science Promotion.

REFERENCES

ASSOULINE, J. G., BOSCH, P., LIM, R., I O M , I. s., JEN- SEN, R., and PANTAZIS, N. P. (1987). Rat astrocyte and Schwann cells in culture synthesize nerve growth factor-like neurite-promoting factors. Dev. Bruin Res.

BAHR, M., WIZENMANN, A., and THANOS, S. (1992). Effect of bilateral tectum lesions on retinal ganglion cell morphology in rats. J . Comp. Neurol. 320:370- 380.

BERRY, M., REES, L., HALL, S., YIU, P., and SEVERS, J. (1988). Optic axons regenerate into sciatic nerve iso- grafts only in the presence of Schwann cells. Bruin Res.

BODNARENKO, S. R. and CHALUPA, L. M. (1 993). Strat- ification of ON and OFF ganglion cell dendrites de-

31:103-118.

Bull. 20:223-23 1.

202 Wutanahe et al

pends on glutamate-mediated afferent activity in the developing retina. Natitre 364: 144- 146.

BOYCOTT, B. B. and WASSLE, H. (1974). The morphological types of ganglion cells of the domestic cat’s retina. J. Physiol. (Lond.) 240:397-4 19.

CARMIGNOTO, G., MAFFEI, L., CANDEO, P., CANELLA, R., and COMELLI, C. (1989). Effect of NGF on the survival of rat retinal ganglion cells following optic nerve transection. J. Neitrosci. 9: 1263-1 272.

CHALUPA, L. M., WILLIAMS, R. W., and HENDERSON, Z. ( 1 984). Binocular interaction in the fetal cat regu- lates the size ofthe ganglion cell population. Neurosci- ence12:1139-1146.

CHO, E. Y. P. and So, K.-F. (1992). Characterization of the sprouting response of axon-like processes from retinal ganglion cells after axotomy in adult hamsters: a model using intravitreal implantation of a peripheral nerve. J. Nezirocytol. 21589-603.

COTTEE, L. J., FITZGIBBON, T., WESTLAND, K., and BURKE, W. ( I 99 1 ). Long survival of retinal ganglion cells in the cat after selective crush of the optic nerve. Euv. J. Neurosci. 3: 1245- 1254.

FUKUDA, Y. and STONE, J. ( 1 974). Retinal distribution and central projections of Y-, X- and W-cells of the cat’s retina. J . Neuvophysiol. 37:749-772.

FUKUDA, Y., HSIAO, C.-F., WATANABE, M., and ITO, H. (1 984). Morphological correlates of physiologically identified Y-, X-, and W-cells in the cat retina. J. Nezi- rophysiol. 52:999-1013.

HOLLANDER, H., BISTI, S., MAFFEI, L., and HEBEL, R. ( 1984). Electroretinographic responses and retrograde changes of retinal morphology after intracranial optic nerve section. A quantitative analysis in the cat. Exp. Brain Res. 55483-493.

HOLLANDER, H., BISTI, S., and MAFFEI, L. (1985). Long term survival of cat retinal ganglion cells after intracranial optic nerve transection. Exp. Bruin Res. 59: 633-635.

HONIG, M. G. and HUME, R. I. ( 1986). Fluorescent carbocyanine dyes allow living neurons of identified ori- gin to be studied in long-term cultures. J. Cell Biol.

HUGHES, A. (1975). A quantitative analysis of the cat retinal ganglion cell topography. J. Cornp. Neiuol. 163: 107- 128.

HUGHES, A. (198 I ) . Population magnitude and distribution of the major modal classes of cat retinal ganglion cell as estimated from HRP filling and a systematic survey of the soma diameter spectra for classical neurons. J. Comp. Neurol. 197:303-339.

HUGHES, A. and WASSLE, H. (1976). The cat optic nerve: fibre total count and diameter spectrum. J. Comp. Neurol. 169: 17 1 - 184.

MAFFEI, L. and FIORENTINI, A. (198 I ) . Electroretino- graphic responses to alternating gratings before and after section of the optic nerve. Science 21 1:953-955.

MAFFEI, L. and FIORENTINI, A. (1982). Electroretino-

103: 17 1 - I 87.

graphic responses to alternating gratings in the cat. Exp. Brain Rex 48:327-334.

MAFFEI, L., CARMINGNOTO, G., PERRY, V. H., CAN- DEO, P., and FERRARI, G. (1 990). Schwann cells pro- mote the survival of rat retinal ganglion cells after optic nerve section. Proc. Natl. Acad. Sci. USA 87:1855- 1859.

MANSOUR-ROBAEY, S., CLARKE, D. B., WANG, Y.-C., BRAY, G. M., and AGUAYO, A. J. (1994). Effects of ocular injury and administration ofbrain-derived neurotrophic factor on survival and regrowth of axotomized retinal ganglion cells. €‘roc. Nut]. Acad. Sci. USA

MEY, J. andT~ANos, S. (1993). Intravitreal injections of neurotrophic factors support the survival of axotomized retinal ganglion cells in adult rats in vivo. Bruin Rex 602:304-3 17.

MORSE, J. K., WIEGAND, S. J., ANDERSON, K., You, Y., CAI, N., CARNAHAN, J., MILLER, J., DISTEFANO, P. S., ALTAR, C. A., LINDSAY, R. M., and ALDERSON, R. F. ( 1993). Brain-derived neurotrophic factor (BDNF) prevents the degeneration of medial septa1 cholinergic neurons following fimbria transection. J. Neurosci. 13:

MURRAY, M. and GRAFSTEIN, B. (1969). Changes in the morphology and amino acid incorporation of regener- ating goldfish optic neurons. Exp. Neurol. 23544- 560.

PEICHL, L., OTT, H., and BOYCOTT, B. B. (1987). Alpha ganglion cells in mammalian retinae. Proc. R. SOC. Lond. B231: 169- 197.

RODIECK, R. W. and WATANABE, M. (1986). Morpho- logic diversity in the ganglion cell projection to different zones within the cat lateral geniculate nucleus. Soc. Neurosci. Abstr. 12:1038.

RODIECK, R. W. and WATANABE, M. (1993). Survey of the morphology of macaque retinal ganglion cells that project to the pretectum, superior colliculus, and par- vicellular laminae of the lateral geniculate nucleus. J. Comp. Ntwrol. 338:289-303.

SCHERER, J. and SCHNITZER, J. (1991). Intraorbital transection ofthe rabbit optic nerve: consequences for ganglion cells and neuroglia in the retina. J . Comp. Neu- rol. 31 2: 1 75- 1 92.

SILVEIRA, L. C. L., RUSSELAKIS-CARNEIRO, M., and PERRY, V. H. (1994). The ganglion cell response to optic nerve injury in the cat: differential responses revealed by neurofibrillar staining. J . Neurocytol. 23:75- 86.

STONE, J. (1978). The number and division of ganglion cells in the cat’s retina. J. Conzp. Neurol. 180:753-772.

STONE, J. and FUKUDA, Y. (1974). Properties of cat retinal ganglion cells: a comparison of W-cells with X- and Y-cells. J. Ncurophysiol. 37~722-748.

nal diameter-based rate of retrograde degeneration in the rat optic nerve. Neurosci. Rex 18:S139.

91:1632-1636.

4 146-4 156.

SUGIOKA, M., SAWAI, H., and FUKUDA, Y. (1993). Axo-

Retinal Ganglion Cells Surviving Axolomy 203

TABATA, T. and FUKUDA, Y. (1992). Dendntic regrowth of retinal ganglion cells in adult rats. Neuroreport 3:

TAKANO, M. and HORIE, H. (1994). Critical period for degradation of adult rat retinal ganglion cells and their regeneration capability after axotomy. Neurosci. Lett. 175: 129-1 32.

THANOS, S. (1988). Alterations in the morphology of ganglion cell dendrites in the adult rat retina after optic nerve transection and grafting of peripheral nerve seg- ments. CelI Tissue Res. 254599-609.

TAUCHI, M. and MASLAND, R. H. ( I 984). The shape and arrangement of cholinergic neurons in the rabbit retina. Proc. R. Soc. Lond. B223: I0 1 - 1 19.

VILLEGAS-P~REZ, M.-P., VIDAL-SANZ, M., RASMINSKY, M., BRAY, G. M., and AGUAYO, A. J. (1993). Rapid and protracted phases of retinal ganglion cell loss fol- low axotomy in the optic nerve of adult rats. J. Neuro- bid. 24:23-36.

709-7 12.

WATANABE, M. and FUKUDA, Y. (1993). Number and morphology of cat’s retinal ganglion cells that survived long after axotomy. Neurosci. Rex 18:s 184.

WATANABE, M. and RODIECK, R. W. (1989). Parasol and midget ganglion cells of the primate retina. J. Comp. Neurol. 289:434-454.

WATANABE, M., SAWAI, H., and FUKUDA, Y. (1991). Axonal regeneration of retinal ganglion cells in the cat geniculocortical pathway. Bruin Res. 560:330-333.

WATANABE, M., SAWAI, H., SUGIOKA, M., RASMINSKY, M., and FUKUDA, Y . (1993a). Physiological properties of retinal ganglion cells that regenerated their axons into peripheral nerve grafts in cats. Invest. Ophthal- mol. Vis. Sci. 34: 1 1 5 5 .

WATANABE, M., SAWAI, H., and FUKUDA, Y . (1993b). Number, distribution and morphology of retinal ganglion cells with axons regenerated into peripheral nervegraft in adult cats. J. Neurosci. 13:2105-2117.

Documents

Number and dendritic morphology of retinal ganglion cells that survived after axotomy in adult cats