- Email: [email protected]
Neurotoxic effects ofl -α-aminoadipic acid on the carp retina: A long term observation
/v’euro.wrence Vol. 36. No. I. pp. 155-163, Printed ,n Great Britain
NEUROTOXIC THE
0306.4522:90
1990
EFFECTS
CARP
$3.00 + 0.00
Pergamon Press plc i 1990 IBRO
RETINA:
OF A
L-a-AMINOADIPIC
LONG
K. SUC;AWARA,* K. TORIC;OE,* S. OtcoVAMA,P
TERM
ACID
ON
OBSERVATION
K. NEGISHI* and S. KAT~*$
*Molecular Neurobiology Group, Departments of Neurophysiology and Biophysics. Neuroinformation Research Institute (NIRI) and tDepartment of Anatomy. University of Kanazawa School of Medicine, Kanazawa, Ishikawa 920. Japan Abstract-The hypothesis has been tested that the enantiomers of r-aminoadipic acid have different target effects; the L-isomer has both glio- and neurotoxic actions, while the in_-isomer has a gliospecific action in the CNS. Electrophysiological and morphological studies were carried out on the retina of the carp (Cyprinus carpio) for one to two months after intraocular injection with a-aminoadipic acids at various doses, Intracellular recording from horizontal cells and extracellular recording of spike discharges from ganglion cells in isolated retinal preparations were made from control and pretreated retinas at various intervals after intraocular injection with the enantiomers. In control retinas, application of IS mM L-a-aminoadipic acid in the superfusate resulted in hyperpolarization of all horizontal cells and in a decrease in amplitude of their light responses (S-potentials). In the retinas pretreated with L-r-aminoadipic acid (8 pmol), low amplitude S-potentials were seen during an early phase 24 h after ocular injection. but the normal appearance of S-potentials was restored one day after injection. In control retinas. a brief period of iontophoretic application of L-cc-aminoadipic acid resulted in a slight activation of the spontaneous spike firing of ganglion cells but a slight decrease in the rate of light-induced firing. In retinas pretreated with intraocular L-a-aminoadipic acid (4 pmol) 4 h prior to eye removal, however, light-induced spike discharges were abolished from nearly all spontaneously firing ganglion cells (>90%). Their unresponsiveness to light stimuli lasted for more than two months after injection, and was accompanied by insensitivity to iontophoretically applied putative neurotransmitters. Use of the DL-a-aminoadipic acid. even at a higher dose (16 pmol), did not cause this semi-permanent loss of ganglion cell responses. Under light-microscopic examination, both DL-r-aminoadipic acid and L-a-aminoadipic acid were found to produce a marked swelling of glial Miiller cells within one day after ocular injection, while the L-isomer additionally caused neuronal damage, particularly in the inner retina. The glial swelling completely disappeared in two to three weeks, and the neuronal damage disappeared gradually over one to two months. The present data show that the neurotoxicity caused by L-a-aminoadipic acid (>4 pmol) in the carp retina is toward neurons of the inner retina particularly to ganglion cells, and that the electrophysiologic effect is irreversible. Gliotoxicity of both L- and oL-r-aminoadipic acid. however, is reversible. the latter racemic mixture being gliospecific.
The glutamate analogue, alpha-aminoadipic acid (a-AAA) is well known as a CNS ghotoxin which affects the retina.” Among the enantiomers of a-AAA, the L-isomer has both gliotoxic and neurotoxic effects on different regions of the CNS, whereas the pr,-isomer specifically acts as a gliotoxin.‘~‘4 The concept
of neurotoxicity
caused
by L-t(-AAA
studied the long term effects of a-AAA isomers on the fish retina. In the preceding paper,’ we reported a gliotoxic action of r-AAA isomers measured by biochemical and electroretinographic parameters. In the present study, we made electrophysiological recordings from horizontal and ganglion cells in intact carp retinas and in retinas prepared at various intervals after intraocular injection of cc-AAA isomers. Both types of cell were chosen as representative neurons of the outer and inner retina, respectively, and because their large cell soma rendered them suitable for electrophysiology. Histopathological examination after injection of a-AAA isomers was also followed for a one to two month long period.
derives
from morphological observations of degenerative and necrotic neuronal cells. With respect to the retina several questions about the actions of enantiomers of r-AAA need to be resolved: (1) which types of neuron are sensitive or insensitive to L-a-AAA?; (2) is the neurotoxic effect reversible or irreversible?; and (3) is the neurotoxic action truly stereospecific, or does DL-c(-AAA at higher concentrations still act gliospecifically? To address these questions we have
EXPERIMENTAL PROCEDURES Applications of a-aminoadipic acid isomers $To whom correspondence should be addressed. Ahhr~iarion.~: r-AAA, alpha-aminoadipic acid; electroretinogram; GS, glutamine synthetase; phosphate-buffered saline.
Inrraocular injection of a-AAAs in viva. Carp (C~pprinus carpio, 600-800 g) were purchased from a local dealer. Their left eyes were intraocularly injected with 50 ~1 of sterilized phosphate-buffer saline (PBS. pH 7.4) containing various
ERG, PBS,
I55
IS6
K. $1 (;~\\vAK.\
doses (from 1.0 to 16 pmol:eye) of’ L-. I>- or tti.-r-AAA (Sigma). Their right eyes were injected with the same volume of PBS. At various time intervals (from 2 h to two months) after intraocular injection of cr-AAAs or P3S, the retinas were isolated for electrophysiological and morphological examinations. Statistical significances were assessed with Student’s t-test. ?-AAA upplicution in vitro. To test the acute effect of r-AAAs on retinal horizontal cells, the isolated retina was superfused with normal fish Ringer solution (pH 7.8) aerated with 97% O2 and 3% COz. A superfusate containing one of the three r-AAA isomers was introduced into the chamber via a liquid switch system r(~nti)~~(~retieupplicutif~~ of r-A.4 As. in order to apply r-AAAs locally into the inner plexiform layer, a multibarrel glass nlicroelectrode was tilled with 0.25 M I.-, u- or DLa-AAA (pH 6.0). and an iontophor~tjc application was performed through a constant current device.”
Ekrtrophysiologiwl
proreduws
Treated and control carp were kept in darkness for I h. The eye was excized under MS 222 (Sankyo) anaesthesia and the retina was isolated. The retina was mounted photoreceptor side-up in a small (0.5 ml) chamber, and superfused with a fish Ringer solution comprised of (in mM): NaCl 119.5, KC1 3.0, CaClz 1.15. MgSO, 1.04, NaHCO, 22.6, NaH2P0, 0.1, Na,HPO, 0.4 and D-f$UCOSe 10.0. The perfusate was aerated with 97% O2 plus 3% COz to obtain a pH of 7.8.’ The flow rate of perfusate was 1.2mljmin. A single glass microelectrode filled with 3 M potassium acetate was used for intracellular recording of horizontal cells. The electrode resistance ranged from 50 to IO0 MR. The signal output was fed into a conventional highimpedance d.c. preamplifier. The type of horizontal cells recorded was identified accord&I to the classification described previously.’ The recording microelectrodes for spike discharges of ganglion cells were In&X-coated tungsten wires with a 223 pm diameter tip. The output was fed into an ac. amplifier with a time constant of 0.003 s, and the firing rates were analysed with a rate meter at I-s intervals and displayed on a pen recorder.
Opticaf .s,wtem for light stimzrlr The two-channel optical system used in the present study was the same as described previousIy.ih A 500 W Xenon lamp was used as the source of light. The unattenuated intensity measured at the retinal surface was 0.7~Wicm~. which was controlled by interposing neutral density filters in 0.5 log unit steps. A small spot (0.4 mm diameter) was used as a stimulus to evoke the centre responses of horizontal and ganglion cells. An annular light stimulus (1.7 mm in outer diameter and 1.0 mm in inner diameter) was used to elicit surround responses.
Histological procedures Light microscopy. Retinal pieces for morphological examination were fixed immediately after dissection by immersion in 3% glutaraIdehyde/l% paraformaidehyde (pH 7.4) for 2 h at 4°C followed by post-fixation with 2% osmium tetroxide in 0.1 M phosphate buffer for 1h at 4°C and embedding in Epon. Blocks were sectioned at 2 pm on an LKB 8800 ultramicrotome and stained with Totuidine Blue. Photomicrographs were taken on Plus-X pan film (Kodak). i4utoradiograph~. A dose of l.OfiCi/SO~l saline/eye of [methyl-l’,2’-3H]thymidine (New England Nuclear; specific activity, 100 Ci/mmol) was injected intravitreally one day after injection with PBS or r-AAAs. Five to seven days later, the retinas were isolated. fixed, embedded and sectioned at 2 urn for autoradiographical examination. Sections were prepared according- to conventional procedures” and examined under bright-field illumination after Cresyl Violet counter-staining. -
(‘i
‘II RESC’LTS
Intracellular recordings from horizontal cells in isolated carp retinas superfused with aerated fish Ringer solution revealed that the addition of 15mM of L-a-AAA to the superfusate evoked a rapid hyperpolarization (12.5 + 2.5 mV) of a luminosity (L)-type horizontal cell within I5 s (Fig. IA). The hyperpolarization was accompanied by a diminution of the S-potential amplitude, produced in response to both central and annular light stimuli. A recovery from the hyperpolarization began I -2 min after washout of L-a-AAA, reached 70% at 5min but required more than 10 min to be completed. When ut-r-AAA (I5 mM) was introduced into the surface superfusate, the same horizontal cell again hyperpolarized (6.8 jI 24mV) and had a reduced S-potential (Fig. 1B). Addition of 15 mM of D-%-AAA to the superfusate similarly decreased the S-potentials with a hyperpolarization (7.4 _t 3.2 mV) in another L-type horizontal cell (Fig. IC). The same effects were obtained with both L-type cells (n = 20) and chromaticity (C)-type cells (n = 10). In contrast, L-glutamate (8 mM), applied in the same way, drastically and transiently depolarized (18 mV) horizontal cells and completely abolished the S-potentials (Fig. lD), as shown previously.” Horizontal intraocular
cell responses in the retina treated L-a:-aminoadipic acid in vivo
with
The recorded resting potential of horizontal cells in control retinas was -21.6 + 3.4 mV. The amplitude of the S-potential elicited by a small field (0.4 mm) of moderate intensity (0.007 FWcm’) was 14.2 Lt:4.1 mV (tr = 256 cells). Two to four hours after intraocular injection of 8gmol L-a-AAA, the horizontal cells recorded (n = 224) under identical light stimulus conditions had reduced S-potentials (~8 mV) and relatively low resting potentials (- 15.7 _t 2.5 mV, P < 0.01, compared to control). One to two days after intraocular injection of Lx-AAA, the horizontal cells (n = 70) appeared to have normal voltage properties: the resting potential was -20.8 + 3.6 mV with a large S-potential (> 10 mV). The normal membrane potential values of horizontal ceils (n = 68) were seen at any time after two days post-injection. Ganglion celi respowes CI-aminoadipic acids
in the retina to i~~to~h~~etic
Since it is difficult to determine whether the effects of ganglion cells of a drug applied in a superfusate are direct or indirect, we employed an iontophoretic ejection of cx-AAAs onto ganglion cells to more closely localize the drug effect (Fig. 2). A multibarrel glass-microelectrode was prepared and attached alongside a recording tungsten wire. The tip of the
Neurotoxic
..... .._._._.
effect of L-a-AAA
,...“.*,.....*......
..“.“....““.““.....“.....
r._._.L._._.L._._._._
._._._.-.
_
.-._._.
n
I
_,_.r.~,~.~._.~,_l-.~~~.~,~.~.~.~
.-.-.-.
‘......_,a
.....*.“.“,......l. -\-V.-.-I*
~.-._,rl_._l_._._.C\-.~.~.~.~.~._
/
~.ii,rii.ii.il*,l,,r.ii,ilil,i,iiiliii,li.
157
in the retina
141
/ /
.
,
J
,
i
,
*
,
1
“C
.-.-.
L
._._.
+
._._.___.__
_.__,
dJill...,>>,._./,,.
-10
A -
L-d-AAA,
15mM
-
20
-10 -20 Lmin
30
-4OmV
L
J 1 min
-
D-d-AAA,
15mM
Fig. I. Intracellular recording of horizontal cells demonstrating the effect of z-AAAs on the isolated carp retina superfused with Ringer solution. L-s(-AAA (I 5 mM) in the superfusate induced a hyperpolarization with diminishing of S-potentials in an L-type cell (A). DL-c(-AAA induced a similar but weaker hyperpolarization in the same cell (B). I,-r-AAA also evoked a small hyperpolarization in another L-type cell (C). L-Glutamate (8 mM) markedly depolarized thus cell, abolishing S-potentials (D). Voltage and time scales are indicated at the right. Central (c) and annular (a) light stimuli are indicated as square pulses in the upper part of the trace.
recording electrode was positioned 40-60 Itrn lateral to, and 2040 ILrn vertical to the iontophoretic glass electrode.“’ Each form of the iontophoreticallyapplied S-AAA evoked a weak facilitation of spontaneous spike discharges in both on-centre (Fig. 2A) and off-centre (Fig. 2B) ganglion cells. Their lightwere slightly diminished. The induced firings potency of spike facilitation evoked was in the order of L- > DL- > D-a-AAA. In contrast, L-glutamate evoked a strong facilitation and even a depolarization block of spike discharges after an intense iontophoretic current (Fig. 2B).
L-Glu I
Ganglion
I
-
acid.~ in vivo
Spike discharges of numerous ganglion cells in both control and pretreated retinas were recorded under constant stimulus conditions (0.4 mm in Lightspot diameter at - 2.0 log units intensity). responsive units were readily found in control retinas; of the 122 units recorded 51 were centre-on cells (41%) 65 were centre-off cells (53%) and six. on-off type cells (5%). At one day after injection of 4 pmol L-a-AAA. almost all units (64/70 cells) were unresponsive to light, although their
, ------I10
cell responses to z-aminoadipir~
,
D-d-AAA
L-
80
spikoslsec
-
1 min
Fig. 2. The effect of iontophoretically applied r-AAAs on the unit recording of ganglion cells. All types of r-AAA isomer evoked a weak facilitation of spontaneous spike discharges of different classes of ganglion cells (A). centre on-type; (B). centre off-type cells. L-Glutamate evoked a stronger facilitation of spike discharges than r-AAAs. Number of spikes evoked per I s and time scale are indicated at the right. Central light stimuli (c) are indicated in the lower part of each trace.
Control
1 L-d-AAA 4pmol
Id
14d
30d
M&&&U&U-
T DL-d-AAA 16j_tmol Id
spikeslsec
21 d
20
7d 3
0
Fig. 3. Ganglion cell activities in the retina pretreated with intravitreal injection of z-AAAs. In the control, units responded well to light stimuli. In the retina treated with 4 prnol L-a-AAA, light-induced firing was absent from almost all units; spontaneous firing could be seen at one day after injection (A). The loss of light-induced spikes continued for at least one month. In contrast to L-a-AAA, light responses in the retinas treated with DL-X-AAA (16 pmol) began to be seen one week after injection, and nearly all cells responded fully to light stimuli two to three weeks later (B).
spontaneous firing rates were relatively high (Fig. 3A, one day). One to two weeks later, units (59/65 cells) continued to be unresponsive, and they showed a reduced rate of spontaneous firing (Fig. 3A, two weeks). At one month after injection, nearly all units (55/60 cells) recorded were still light-insensitive, with a very low rate of spontaneous firing (Fig. 3A, one month). The results obtained after intraocular injection of a higher dose (16 pmol) of DL-a-AAA were greatly different from those with 4pmol L-a-AAA. At one day the lack of light responsivity was similar to the L-a-AAA treated retinas (54/61 cells) and had a high spontaneous firing rate (Fig. 3B, one day). However, one week later most units (43/71 cells. 61%) became weakly responsive to light stimuli (Fig. 3B, one week). and normal ganglion cell responses were stored by three weeks (Fig. 3B. three weeks). The restoration from the loss of light-induced spikes after injection of DL-a-AAA was even more rapid when lower doses were injected. The use of the D-isomer also resulted in a similar transient loss of light responses. These various results are quantified in Fig. 4. A transient loss by D- and DL-a-AAA was confirmed and a permanent loss of light-responsiveness on ganglion cells was observed after L-a-AAA. The minimum effective dose of Lx-AAA resulting in permanent loss of light response was about 4pmol.
Figure 5A shows an example of an off-centre type cell in the control retina, which was sensitive to iontophoretic L-glutamate, GABA, glycine and acetylcholine. L-Glutamate applied to control retinas facilitated the spontaneous spike discharge of offcentre cells but GABA, glycine and acetylcholine inhibited them. In contrast, ganglion cells recorded from the retina pretreated with 8 pmol L-c(-AAA were neither responsive to light stimuli nor to these neurotransmitters (Fig. 5B and C).
Ganglion cell in vitro responses to retinal neurotransmitter candidates
Fig. 4. Occurrence (percentage of the total number of cells examined) of light-responsive ganglion cells at various intervals after intraocular injection of various doses of r-AAAs. Responses (n = 70 cells) were recorded at the indicated periods (4 h to two months). The recovery from the loss of light-induced responses after injection of D-a-AAA (8 hmol) occurred earliest (five to seven days) and with DL-t(-AAA somewhat later (two to three weeks). Recovery of light responses was never seen after treatment with more than 4pmol of L-a-AAA.
We have demonstrated that spike discharges of ganglion cells in the carp retina were sensitive to various neurotransmitters.“.” In order to investigate the responsiveness of ganglion cells in the retinas pretreated with L-a-AAA to neurotransmitters, both control and pretreated retinas were compared.
Morphological acids
effects of intraocular (x-aminoadipic
Semi-thin sections of the retina were produced from eyes injected with 8-16pmol r-AAAs at
0Jlmo1
16JmOl
* -C-
L-U-AAA
+
l-I-AAA
.c” 0 _I
0 4h 1
3
5
Days after intraocular
7
14 injection
lpmol
l-CI_AAALpmol
21 30 50 of U-AAAs
8Mmol
-
Neurotoxic
J
effect of L-x-AAA
in the retina
i
c
L-Glu
,
Ach
GABA
200
200
200
300
In the control retina (A), iontophoretic Fig. 5. Responses of ganglion cells to neurotransmitters. r-glutamate facilitates spontaneous spike discharges of ganglion cells, while GABA. glycine and acetylcholine are inhibitory. In the retinas treated with 8 prnol of L-a-AAA. these excitatory and inhibitory substances evoked negligible or no changes in spontaneous firing of ganglion cells at one day (B) or one month (C) after injection.
various times prior to eye removal in order to assess the morphologic consequences of toxin administration (Figs 6 and 7). The Dt_-mixture (S pmol) caused a swelling of the inner retina; Miiller cells had extensive pcrikaryal and cytoplasmic edema (Fig. 6B) as compared with those in control retina (Fig. 6A). Neuronal elements also appeared slightly edematous. Injury to the glia by the o-isomer was appreciably less than that caused by DL-n-AAA (not shown). Four days after injection of DLr-AAA (i6f1mol). cells of the inner nuclear layer were swollen and an edematous change of inner plexiform layer was conspicuous (Fig. 6C). The edematous changes of the inner retina were fairly well repaired seven to IO days later and the normal appearance of retina was restored two to three weeks after injection (Fig. 6D). The injection of 8 ~~mol L-r-AAA caused much more severe damage (Fig. 7A) than did the r,t.-mixture (8 iimol. cf. Fig. 6B). Miiller cells showed pycnotic changes. Furthermore, neuronal damage was extensive with vacuoliration in the inner nuclear, inner plexiform and nerve fibre layers one day after injection (Fig. 7A). Cilia1 and neuronal changes produced by L-‘Y-AAA were less evident by three to four weeks and the retinal neurons appeared normal by two months after injection (Fig. 7B).
Our morphological data with semi-thin sections showed a reversible character of gliotoxicity produced by either t.- or DL-a-AAA. To investigate whether some glial cell proliferation (regeneration)
had occurred, we examined the incorporation of [7H]thynl~dine in retinas pretreated with intraocular DL-Z-AAA. [‘HjThymidine (I .OiiCi,‘eyc) was injected intraocularly one to two days after injection of DL-s(-AAA (&I6 pmol) and five to seven days later the eyes were enucleated for autoradiographical examination. In control retinas. cells incorporating ]3H]thymidinc were sparse and were found only in the outer nuclear layer (Fig. 8A, marked with arrows). In retinas pretreated with intraocular DL-TV-AAA, besides those cells located in the outer nuclear layer, other cells with dense accumulation of grains of [‘Hlthymidine were closely associated with the swollen Miiller cells in the inner nuclear layer (Fig. EiB, marked with arrow heads). DISCUSSION
The effects of CX-AAAs, a hyperpolarization 01 horizontal cells and reduction of S-potential in amplitudes, were transient when the toxins were applied in the perfusate. However, at 224 h after intravitreal injection all the horizontal cells penetrated showed a relatively low resting potential in addition to the reduced S-potential. This difference may have been due to differences in time of exposure of retinas receiving r-AAAs by the two routes. Nevertheless their effects were found to be transient in both in z*itro and in z.iro preparations. However, the recovery from the slight hyperpolarization produced by either one of x-AAAs was very slow (IO min), as compared to the immediate recovery from a large depolarization induced by L-glutamate (Fig. I). Although the reason
Fig, 6. Retinal morphology, at various times after intraocular injection of UL-a-AAA. Semi-thin sections of retinas from the eyes mjected with 8 nmoi of IJL-a-AAA one day prior to eye removal were stained with Toiuidine Blue. As compared with control retina (A), the gliai compartment of E,L-a-AAA injected eyes (B) was more densely stained (arrows). At four days after injection of 16.~mol IJL-a-AAA, changes in the inner nuclear and inner plexiform layers are evident (C). The glial changes caused by UL-z-AA/\ (16pmol) gradually recovered, and the normal appearance of retina could be seen at three weeks after injection (D). Scaie bar = 40 pm.
for the slow recovery process with r-AAAs was not explored, it could have been due to the absence of an uptake system for I-AAAs in the fish retina. The experiments reported in the preceding’ and present papers aimed to determine the extent of glio- and neurotoxicity of r-AAAs in the carp retina and to test the specificity of its isomers. The electrophysiological data showed that horizontal cells were hyperpoiari~d with decreasing S-potentials by superfusion of any form of r-AAA isomers (15 mM). It may be possible that the drug exerts a toxic effect on the photoreceptors which, in
turn, reduces the dark release of glutamate (see our preceding paper’). Iiowever, the current study revealed a restoration of the normal responses of horizontal cell (IO-20 mV) by one to two days after injection of any form of ix-AAAs. Acute changes in horizontal cell responses after DL-a-AAA injection were recorded by Wu and Dowling’9 and by Szamier rt ~71.‘~ Therefore horizontal cell changes appear to be of much shorter duration than the biochemical glutamine synthetase (GS) and electroretinogram (ERG) changes (b-wave) noted by Kato
L’t ld7
Neurotoxic
effect of L-c(-AAA in the retina
Ihl
Fig. 7. Morphological changes after intraocular injection of L-r-AAA. Severe changes of glial Miiller cells with pycnosis (arrow) were produced one day after injection of8 pmol L-Y-AAA (A). Additional neuronal injury occurred tn the mner retina with extensive vacuoliration. At two weeks after Injection of 8 Lrmol I-z-AAA. ghal and neuronal changes were similar to those seen four days after injection of lh/tmol I)L-X-AAA (cf. Fig. 6C). A normal architecture was gradually restored over 50~60 days after injection (B, two months). Scale bar = 40/rm.
Although ganglion cell spike activity was weakly activated by iontophoretic application of CC-AAAs, pretreatment with intraocular L-s(-AAA (>4/~mol) produced a long term suppression, and the lightinduced spike discharges of almost all units (>90%) failed to be restored even two months after injection. In contrast, the suppression of ganglion cell activities
caused by a four-fold higher dose of tn.-x-AAA (16 pmol) was reversible, being fully restored two to three weeks later. Bonaventure ct ~1. reported a block of spike discharges in some frog ganglion cells (25%) by intraocular administration of DL-a-AAA. However, they did not indicate whether or not the ganglion cell blockade was reversible. Furthermore. the unresponsiveness of ganglion cell activtties to iontophoretic application of various neurotransmitters indicates that the permanent suppression
Fig. 8. Autoradiograph of [?H]thymidine incorporation in the retina after intraocular injection of IIL-r-AAA. One mtcrocurie of [‘Hlthymidine was injected mto eyes pretreated with an mtravitrcal mjection of DL-a-AAA (I6 pmol) one to t&o days earlier, and the fish were allowed to survive for three to five days. In the control retina (A) [‘Hlthymidme incorporating cells were present in the outer nuclear layer (arrows). In m-u-AAA-pretreated retinas (B) additional lahelled cells (arrowheads) were present adjacent to abnormal appearing Miiller cells, as well as in the outer nuclear layer (arrows). Scale bar = 50 I’m.
caused by L-z-,4AA was not due to a biockadc ot’ presynaptic iransm~ssion but due to an injury to ganghon cells themselves. In their entirety, the present results strongly suggest that the neurotoxic action of CX-AAAisomers in the retina was limited to the L-form of the toxin and had long term effects only on inner neurons, particularly ganglion cells. The apparent recovery of n~orph~logical appearance caused one to two months after injection of L-cc-AAA was at odds with the suppression of ganglion cell activities produced by the same isomer two months after injection. Further structural analyses of the retinas treated with a-AAAs must be conducted to resolve this discrepancy in the fttture.
In the preceding paper,7 we quantified the extent of gliotoxicity produced by t.- and IX-c(-AAA in the fish retina. The gliotoxicity produced by L-r-AAA was comparable to that produced by two-fold higher dose of DL-a-AAA in the reduction of giutamine synthetase and ERG b-wave activities. The ghotoxicity of L-z-AAA even with a high dose (8 Lcrnoi) was transient. The present morphological data further confirmed the reversible gliotoxicity of L- and DLa-AAA. We found that the recovery of the ERG b-wave after intravitreal injection with II prno# I.a-AAA or with 16 pmoi IX-X-AAA required a longer time (two months for 50% recovery in amplitude) than that of the biochemical change (one month for 100% recovery in the GS activity). Since the suppressing effect of the t)L-isomer ( 16{irnoi) on ganglion cells was much weaker than that of the
t.-isomer (4 ~cmof). the functional states of ncuronaI ceils appear to be not invuived in their recovery processes of glial cells observed. The cellular mechanism for the transient nature of the action of sr-AAAs on Miiller cells was suggested by the results of the autoradiogrdphicaf study of ~3~]thyn~dine incorporation. Newly pro~iferat~ilg neurons of ail classes of retinal cell are known to be normaily added to the peripheral retina from the ora serrata, and new rod cells are added to the photoreceptor layer in the outer nuclear laycr.‘~‘~“~” !’ fn the DL-x-AAA (16iimol) treated animals in this study, cells which had incorporated [~~]t~yrnjd~ne were associated with swollen Mtiller cells as well as in germinal cells in the outer nuclear layer. The results suggest that the reversibility from partial destruction of the cells of inner nuclear layer by LXor L-r-AAA may be accomplished by a proliferation of [3H]thymidine incorporating cells. A similar localization of lH]thymidine-incorporating cells in the inner nuclear layer has been observed in juvenile carp retina treated with a high dose of k-hydroxy-
dopamine.” This may indicate a common response of the fish retina to high doses of neurotoxins and:or gliotoxins. The fate of these [‘Hjthymidine-ijthymidine-itlcorpctrating cells, whether differentiating into new Miiltcr c&s or into other classes oi‘ ceils. is unknown.
AIknuwledRemcrrts--We thank Dr A. W. Spira for his correction of our English and also Mrs Tami Urano for her secretarial as&tan& This work was supported in part by research grants (Nos 59570055, 4077~S6, 614801~5 and 01659506 to SK.) from the Ministry of Education, Science and Culture, Japan.
REFERENCES Bonaventure N., Roussel D. and Wioland N. (1981) Effects of ~,L-~pha-amjnoad~pic acid on Miiller ceils in frog and chicken retina in viva: relation to ERG b-wave, ganglion ceti discharge and tectal evoked potentials. Neumsri. Left. 27, 81-87. 2. Casper D. S. and Reif-Lehrer L. (19833 Effects of alpha-aminoadipate isomers on the mo~ho~ogy of the isolated chick embryo retina. invest. Dphthal. his. Sci. 24, 1480-1488. -3 Johns P. R. (1981) Growth of fish retinas, Am. Zoo/. 21, 441-453. 4. Johns P. R. and Fernald R. D. (1981) Genesis of rods in teleost fish retina. Nuture 293, 141-142. 5. Kato S. (1979) C-type horizontal cell .responses to annular stimuli. Expl Eye Res. 28, 627-639. 6. Kato S. and Neeishi (19781 Effects of variations in the oerfusate on the ERG and discharge ganglion cells in carp _ of -_ retina. Expl Eyp Res.‘26, j63-376. 7. Kate S., Sugawara K., ~ats~kawa T. and Negishi K. (1990) Ghotoxic effects of ~aminoadipic acid isomers on the carp retina: a long term observation. Neuroscience 36, 145-153. 8. Kato S., Teranishi T. and Negishi K. (1985) ~-Glutamate depolarizes ON-OFF transient type of amacrine cells in the carp retina: an ionophoretic study. Brain Res. 329, 390-394. 3. Kock J. H. (1982) Neuronal addition and retina! expansion during growth of the crucian carp eye. .J. c’u~prp.Neur
Neurotoxic
effect of L-r-AAA
in the retina
163
IS. Olney J. W.. Ho 0. L. and Rhee V. (1971) Cytotoxic effects of acidic and sulphur ~(~nt~ining amino acids on the infant mouse central nervous system. E.$ Brain Rev. 14, 61 76. 16. Sugawara K. (1985) Lateral actions at the inner plexiform layer of the carp retina: effects of turning windmill pattern stimulus. Visiorr Res. 25, I 179 1186. 17. Sugawara K. and Negishi K. (1973) Effects of some amino acids on light-induced responses in the isolated carp retina C’ision Rcs. 13, 2479--24X9. IX. Szamier R. B.. Ripps H. and Chappell R. L. (1981) Changes in ERG b-wave and Miiller cell structure induced by aipha-aminoadipic acid. Na~~~sci. L~rt. 21, 307 312. 19. Wu S. M. and Dowling J. E. (1978) L-Aspartate: evidence for ;i role in cone photoreceptor synaptic tr~~nsrniss~oll in the curp retina. PFOC.. N&z. Ami. Sci. U.S.A. 75, 5205 5209.
NEUROTOXIC THE
0306.4522:90
1990
EFFECTS
CARP
$3.00 + 0.00
Pergamon Press plc i 1990 IBRO
RETINA:
OF A
L-a-AMINOADIPIC
LONG
K. SUC;AWARA,* K. TORIC;OE,* S. OtcoVAMA,P
TERM
ACID
ON
OBSERVATION
K. NEGISHI* and S. KAT~*$
*Molecular Neurobiology Group, Departments of Neurophysiology and Biophysics. Neuroinformation Research Institute (NIRI) and tDepartment of Anatomy. University of Kanazawa School of Medicine, Kanazawa, Ishikawa 920. Japan Abstract-The hypothesis has been tested that the enantiomers of r-aminoadipic acid have different target effects; the L-isomer has both glio- and neurotoxic actions, while the in_-isomer has a gliospecific action in the CNS. Electrophysiological and morphological studies were carried out on the retina of the carp (Cyprinus carpio) for one to two months after intraocular injection with a-aminoadipic acids at various doses, Intracellular recording from horizontal cells and extracellular recording of spike discharges from ganglion cells in isolated retinal preparations were made from control and pretreated retinas at various intervals after intraocular injection with the enantiomers. In control retinas, application of IS mM L-a-aminoadipic acid in the superfusate resulted in hyperpolarization of all horizontal cells and in a decrease in amplitude of their light responses (S-potentials). In the retinas pretreated with L-r-aminoadipic acid (8 pmol), low amplitude S-potentials were seen during an early phase 24 h after ocular injection. but the normal appearance of S-potentials was restored one day after injection. In control retinas. a brief period of iontophoretic application of L-cc-aminoadipic acid resulted in a slight activation of the spontaneous spike firing of ganglion cells but a slight decrease in the rate of light-induced firing. In retinas pretreated with intraocular L-a-aminoadipic acid (4 pmol) 4 h prior to eye removal, however, light-induced spike discharges were abolished from nearly all spontaneously firing ganglion cells (>90%). Their unresponsiveness to light stimuli lasted for more than two months after injection, and was accompanied by insensitivity to iontophoretically applied putative neurotransmitters. Use of the DL-a-aminoadipic acid. even at a higher dose (16 pmol), did not cause this semi-permanent loss of ganglion cell responses. Under light-microscopic examination, both DL-r-aminoadipic acid and L-a-aminoadipic acid were found to produce a marked swelling of glial Miiller cells within one day after ocular injection, while the L-isomer additionally caused neuronal damage, particularly in the inner retina. The glial swelling completely disappeared in two to three weeks, and the neuronal damage disappeared gradually over one to two months. The present data show that the neurotoxicity caused by L-a-aminoadipic acid (>4 pmol) in the carp retina is toward neurons of the inner retina particularly to ganglion cells, and that the electrophysiologic effect is irreversible. Gliotoxicity of both L- and oL-r-aminoadipic acid. however, is reversible. the latter racemic mixture being gliospecific.
The glutamate analogue, alpha-aminoadipic acid (a-AAA) is well known as a CNS ghotoxin which affects the retina.” Among the enantiomers of a-AAA, the L-isomer has both gliotoxic and neurotoxic effects on different regions of the CNS, whereas the pr,-isomer specifically acts as a gliotoxin.‘~‘4 The concept
of neurotoxicity
caused
by L-t(-AAA
studied the long term effects of a-AAA isomers on the fish retina. In the preceding paper,’ we reported a gliotoxic action of r-AAA isomers measured by biochemical and electroretinographic parameters. In the present study, we made electrophysiological recordings from horizontal and ganglion cells in intact carp retinas and in retinas prepared at various intervals after intraocular injection of cc-AAA isomers. Both types of cell were chosen as representative neurons of the outer and inner retina, respectively, and because their large cell soma rendered them suitable for electrophysiology. Histopathological examination after injection of a-AAA isomers was also followed for a one to two month long period.
derives
from morphological observations of degenerative and necrotic neuronal cells. With respect to the retina several questions about the actions of enantiomers of r-AAA need to be resolved: (1) which types of neuron are sensitive or insensitive to L-a-AAA?; (2) is the neurotoxic effect reversible or irreversible?; and (3) is the neurotoxic action truly stereospecific, or does DL-c(-AAA at higher concentrations still act gliospecifically? To address these questions we have
EXPERIMENTAL PROCEDURES Applications of a-aminoadipic acid isomers $To whom correspondence should be addressed. Ahhr~iarion.~: r-AAA, alpha-aminoadipic acid; electroretinogram; GS, glutamine synthetase; phosphate-buffered saline.
Inrraocular injection of a-AAAs in viva. Carp (C~pprinus carpio, 600-800 g) were purchased from a local dealer. Their left eyes were intraocularly injected with 50 ~1 of sterilized phosphate-buffer saline (PBS. pH 7.4) containing various
ERG, PBS,
I55
IS6
K. $1 (;~\\vAK.\
doses (from 1.0 to 16 pmol:eye) of’ L-. I>- or tti.-r-AAA (Sigma). Their right eyes were injected with the same volume of PBS. At various time intervals (from 2 h to two months) after intraocular injection of cr-AAAs or P3S, the retinas were isolated for electrophysiological and morphological examinations. Statistical significances were assessed with Student’s t-test. ?-AAA upplicution in vitro. To test the acute effect of r-AAAs on retinal horizontal cells, the isolated retina was superfused with normal fish Ringer solution (pH 7.8) aerated with 97% O2 and 3% COz. A superfusate containing one of the three r-AAA isomers was introduced into the chamber via a liquid switch system r(~nti)~~(~retieupplicutif~~ of r-A.4 As. in order to apply r-AAAs locally into the inner plexiform layer, a multibarrel glass nlicroelectrode was tilled with 0.25 M I.-, u- or DLa-AAA (pH 6.0). and an iontophor~tjc application was performed through a constant current device.”
Ekrtrophysiologiwl
proreduws
Treated and control carp were kept in darkness for I h. The eye was excized under MS 222 (Sankyo) anaesthesia and the retina was isolated. The retina was mounted photoreceptor side-up in a small (0.5 ml) chamber, and superfused with a fish Ringer solution comprised of (in mM): NaCl 119.5, KC1 3.0, CaClz 1.15. MgSO, 1.04, NaHCO, 22.6, NaH2P0, 0.1, Na,HPO, 0.4 and D-f$UCOSe 10.0. The perfusate was aerated with 97% O2 plus 3% COz to obtain a pH of 7.8.’ The flow rate of perfusate was 1.2mljmin. A single glass microelectrode filled with 3 M potassium acetate was used for intracellular recording of horizontal cells. The electrode resistance ranged from 50 to IO0 MR. The signal output was fed into a conventional highimpedance d.c. preamplifier. The type of horizontal cells recorded was identified accord&I to the classification described previously.’ The recording microelectrodes for spike discharges of ganglion cells were In&X-coated tungsten wires with a 223 pm diameter tip. The output was fed into an ac. amplifier with a time constant of 0.003 s, and the firing rates were analysed with a rate meter at I-s intervals and displayed on a pen recorder.
Opticaf .s,wtem for light stimzrlr The two-channel optical system used in the present study was the same as described previousIy.ih A 500 W Xenon lamp was used as the source of light. The unattenuated intensity measured at the retinal surface was 0.7~Wicm~. which was controlled by interposing neutral density filters in 0.5 log unit steps. A small spot (0.4 mm diameter) was used as a stimulus to evoke the centre responses of horizontal and ganglion cells. An annular light stimulus (1.7 mm in outer diameter and 1.0 mm in inner diameter) was used to elicit surround responses.
Histological procedures Light microscopy. Retinal pieces for morphological examination were fixed immediately after dissection by immersion in 3% glutaraIdehyde/l% paraformaidehyde (pH 7.4) for 2 h at 4°C followed by post-fixation with 2% osmium tetroxide in 0.1 M phosphate buffer for 1h at 4°C and embedding in Epon. Blocks were sectioned at 2 pm on an LKB 8800 ultramicrotome and stained with Totuidine Blue. Photomicrographs were taken on Plus-X pan film (Kodak). i4utoradiograph~. A dose of l.OfiCi/SO~l saline/eye of [methyl-l’,2’-3H]thymidine (New England Nuclear; specific activity, 100 Ci/mmol) was injected intravitreally one day after injection with PBS or r-AAAs. Five to seven days later, the retinas were isolated. fixed, embedded and sectioned at 2 urn for autoradiographical examination. Sections were prepared according- to conventional procedures” and examined under bright-field illumination after Cresyl Violet counter-staining. -
(‘i
‘II RESC’LTS
Intracellular recordings from horizontal cells in isolated carp retinas superfused with aerated fish Ringer solution revealed that the addition of 15mM of L-a-AAA to the superfusate evoked a rapid hyperpolarization (12.5 + 2.5 mV) of a luminosity (L)-type horizontal cell within I5 s (Fig. IA). The hyperpolarization was accompanied by a diminution of the S-potential amplitude, produced in response to both central and annular light stimuli. A recovery from the hyperpolarization began I -2 min after washout of L-a-AAA, reached 70% at 5min but required more than 10 min to be completed. When ut-r-AAA (I5 mM) was introduced into the surface superfusate, the same horizontal cell again hyperpolarized (6.8 jI 24mV) and had a reduced S-potential (Fig. 1B). Addition of 15 mM of D-%-AAA to the superfusate similarly decreased the S-potentials with a hyperpolarization (7.4 _t 3.2 mV) in another L-type horizontal cell (Fig. IC). The same effects were obtained with both L-type cells (n = 20) and chromaticity (C)-type cells (n = 10). In contrast, L-glutamate (8 mM), applied in the same way, drastically and transiently depolarized (18 mV) horizontal cells and completely abolished the S-potentials (Fig. lD), as shown previously.” Horizontal intraocular
cell responses in the retina treated L-a:-aminoadipic acid in vivo
with
The recorded resting potential of horizontal cells in control retinas was -21.6 + 3.4 mV. The amplitude of the S-potential elicited by a small field (0.4 mm) of moderate intensity (0.007 FWcm’) was 14.2 Lt:4.1 mV (tr = 256 cells). Two to four hours after intraocular injection of 8gmol L-a-AAA, the horizontal cells recorded (n = 224) under identical light stimulus conditions had reduced S-potentials (~8 mV) and relatively low resting potentials (- 15.7 _t 2.5 mV, P < 0.01, compared to control). One to two days after intraocular injection of Lx-AAA, the horizontal cells (n = 70) appeared to have normal voltage properties: the resting potential was -20.8 + 3.6 mV with a large S-potential (> 10 mV). The normal membrane potential values of horizontal ceils (n = 68) were seen at any time after two days post-injection. Ganglion celi respowes CI-aminoadipic acids
in the retina to i~~to~h~~etic
Since it is difficult to determine whether the effects of ganglion cells of a drug applied in a superfusate are direct or indirect, we employed an iontophoretic ejection of cx-AAAs onto ganglion cells to more closely localize the drug effect (Fig. 2). A multibarrel glass-microelectrode was prepared and attached alongside a recording tungsten wire. The tip of the
Neurotoxic
..... .._._._.
effect of L-a-AAA
,...“.*,.....*......
..“.“....““.““.....“.....
r._._.L._._.L._._._._
._._._.-.
_
.-._._.
n
I
_,_.r.~,~.~._.~,_l-.~~~.~,~.~.~.~
.-.-.-.
‘......_,a
.....*.“.“,......l. -\-V.-.-I*
~.-._,rl_._l_._._.C\-.~.~.~.~.~._
/
~.ii,rii.ii.il*,l,,r.ii,ilil,i,iiiliii,li.
157
in the retina
141
/ /
.
,
J
,
i
,
*
,
1
“C
.-.-.
L
._._.
+
._._.___.__
_.__,
dJill...,>>,._./,,.
-10
A -
L-d-AAA,
15mM
-
20
-10 -20 Lmin
30
-4OmV
L
J 1 min
-
D-d-AAA,
15mM
Fig. I. Intracellular recording of horizontal cells demonstrating the effect of z-AAAs on the isolated carp retina superfused with Ringer solution. L-s(-AAA (I 5 mM) in the superfusate induced a hyperpolarization with diminishing of S-potentials in an L-type cell (A). DL-c(-AAA induced a similar but weaker hyperpolarization in the same cell (B). I,-r-AAA also evoked a small hyperpolarization in another L-type cell (C). L-Glutamate (8 mM) markedly depolarized thus cell, abolishing S-potentials (D). Voltage and time scales are indicated at the right. Central (c) and annular (a) light stimuli are indicated as square pulses in the upper part of the trace.
recording electrode was positioned 40-60 Itrn lateral to, and 2040 ILrn vertical to the iontophoretic glass electrode.“’ Each form of the iontophoreticallyapplied S-AAA evoked a weak facilitation of spontaneous spike discharges in both on-centre (Fig. 2A) and off-centre (Fig. 2B) ganglion cells. Their lightwere slightly diminished. The induced firings potency of spike facilitation evoked was in the order of L- > DL- > D-a-AAA. In contrast, L-glutamate evoked a strong facilitation and even a depolarization block of spike discharges after an intense iontophoretic current (Fig. 2B).
L-Glu I
Ganglion
I
-
acid.~ in vivo
Spike discharges of numerous ganglion cells in both control and pretreated retinas were recorded under constant stimulus conditions (0.4 mm in Lightspot diameter at - 2.0 log units intensity). responsive units were readily found in control retinas; of the 122 units recorded 51 were centre-on cells (41%) 65 were centre-off cells (53%) and six. on-off type cells (5%). At one day after injection of 4 pmol L-a-AAA. almost all units (64/70 cells) were unresponsive to light, although their
, ------I10
cell responses to z-aminoadipir~
,
D-d-AAA
L-
80
spikoslsec
-
1 min
Fig. 2. The effect of iontophoretically applied r-AAAs on the unit recording of ganglion cells. All types of r-AAA isomer evoked a weak facilitation of spontaneous spike discharges of different classes of ganglion cells (A). centre on-type; (B). centre off-type cells. L-Glutamate evoked a stronger facilitation of spike discharges than r-AAAs. Number of spikes evoked per I s and time scale are indicated at the right. Central light stimuli (c) are indicated in the lower part of each trace.
Control
1 L-d-AAA 4pmol
Id
14d
30d
M&&&U&U-
T DL-d-AAA 16j_tmol Id
spikeslsec
21 d
20
7d 3
0
Fig. 3. Ganglion cell activities in the retina pretreated with intravitreal injection of z-AAAs. In the control, units responded well to light stimuli. In the retina treated with 4 prnol L-a-AAA, light-induced firing was absent from almost all units; spontaneous firing could be seen at one day after injection (A). The loss of light-induced spikes continued for at least one month. In contrast to L-a-AAA, light responses in the retinas treated with DL-X-AAA (16 pmol) began to be seen one week after injection, and nearly all cells responded fully to light stimuli two to three weeks later (B).
spontaneous firing rates were relatively high (Fig. 3A, one day). One to two weeks later, units (59/65 cells) continued to be unresponsive, and they showed a reduced rate of spontaneous firing (Fig. 3A, two weeks). At one month after injection, nearly all units (55/60 cells) recorded were still light-insensitive, with a very low rate of spontaneous firing (Fig. 3A, one month). The results obtained after intraocular injection of a higher dose (16 pmol) of DL-a-AAA were greatly different from those with 4pmol L-a-AAA. At one day the lack of light responsivity was similar to the L-a-AAA treated retinas (54/61 cells) and had a high spontaneous firing rate (Fig. 3B, one day). However, one week later most units (43/71 cells. 61%) became weakly responsive to light stimuli (Fig. 3B, one week). and normal ganglion cell responses were stored by three weeks (Fig. 3B. three weeks). The restoration from the loss of light-induced spikes after injection of DL-a-AAA was even more rapid when lower doses were injected. The use of the D-isomer also resulted in a similar transient loss of light responses. These various results are quantified in Fig. 4. A transient loss by D- and DL-a-AAA was confirmed and a permanent loss of light-responsiveness on ganglion cells was observed after L-a-AAA. The minimum effective dose of Lx-AAA resulting in permanent loss of light response was about 4pmol.
Figure 5A shows an example of an off-centre type cell in the control retina, which was sensitive to iontophoretic L-glutamate, GABA, glycine and acetylcholine. L-Glutamate applied to control retinas facilitated the spontaneous spike discharge of offcentre cells but GABA, glycine and acetylcholine inhibited them. In contrast, ganglion cells recorded from the retina pretreated with 8 pmol L-c(-AAA were neither responsive to light stimuli nor to these neurotransmitters (Fig. 5B and C).
Ganglion cell in vitro responses to retinal neurotransmitter candidates
Fig. 4. Occurrence (percentage of the total number of cells examined) of light-responsive ganglion cells at various intervals after intraocular injection of various doses of r-AAAs. Responses (n = 70 cells) were recorded at the indicated periods (4 h to two months). The recovery from the loss of light-induced responses after injection of D-a-AAA (8 hmol) occurred earliest (five to seven days) and with DL-t(-AAA somewhat later (two to three weeks). Recovery of light responses was never seen after treatment with more than 4pmol of L-a-AAA.
We have demonstrated that spike discharges of ganglion cells in the carp retina were sensitive to various neurotransmitters.“.” In order to investigate the responsiveness of ganglion cells in the retinas pretreated with L-a-AAA to neurotransmitters, both control and pretreated retinas were compared.
Morphological acids
effects of intraocular (x-aminoadipic
Semi-thin sections of the retina were produced from eyes injected with 8-16pmol r-AAAs at
0Jlmo1
16JmOl
* -C-
L-U-AAA
+
l-I-AAA
.c” 0 _I
0 4h 1
3
5
Days after intraocular
7
14 injection
lpmol
l-CI_AAALpmol
21 30 50 of U-AAAs
8Mmol
-
Neurotoxic
J
effect of L-x-AAA
in the retina
i
c
L-Glu
,
Ach
GABA
200
200
200
300
In the control retina (A), iontophoretic Fig. 5. Responses of ganglion cells to neurotransmitters. r-glutamate facilitates spontaneous spike discharges of ganglion cells, while GABA. glycine and acetylcholine are inhibitory. In the retinas treated with 8 prnol of L-a-AAA. these excitatory and inhibitory substances evoked negligible or no changes in spontaneous firing of ganglion cells at one day (B) or one month (C) after injection.
various times prior to eye removal in order to assess the morphologic consequences of toxin administration (Figs 6 and 7). The Dt_-mixture (S pmol) caused a swelling of the inner retina; Miiller cells had extensive pcrikaryal and cytoplasmic edema (Fig. 6B) as compared with those in control retina (Fig. 6A). Neuronal elements also appeared slightly edematous. Injury to the glia by the o-isomer was appreciably less than that caused by DL-n-AAA (not shown). Four days after injection of DLr-AAA (i6f1mol). cells of the inner nuclear layer were swollen and an edematous change of inner plexiform layer was conspicuous (Fig. 6C). The edematous changes of the inner retina were fairly well repaired seven to IO days later and the normal appearance of retina was restored two to three weeks after injection (Fig. 6D). The injection of 8 ~~mol L-r-AAA caused much more severe damage (Fig. 7A) than did the r,t.-mixture (8 iimol. cf. Fig. 6B). Miiller cells showed pycnotic changes. Furthermore, neuronal damage was extensive with vacuoliration in the inner nuclear, inner plexiform and nerve fibre layers one day after injection (Fig. 7A). Cilia1 and neuronal changes produced by L-‘Y-AAA were less evident by three to four weeks and the retinal neurons appeared normal by two months after injection (Fig. 7B).
Our morphological data with semi-thin sections showed a reversible character of gliotoxicity produced by either t.- or DL-a-AAA. To investigate whether some glial cell proliferation (regeneration)
had occurred, we examined the incorporation of [7H]thynl~dine in retinas pretreated with intraocular DL-Z-AAA. [‘HjThymidine (I .OiiCi,‘eyc) was injected intraocularly one to two days after injection of DL-s(-AAA (&I6 pmol) and five to seven days later the eyes were enucleated for autoradiographical examination. In control retinas. cells incorporating ]3H]thymidinc were sparse and were found only in the outer nuclear layer (Fig. 8A, marked with arrows). In retinas pretreated with intraocular DL-TV-AAA, besides those cells located in the outer nuclear layer, other cells with dense accumulation of grains of [‘Hlthymidine were closely associated with the swollen Miiller cells in the inner nuclear layer (Fig. EiB, marked with arrow heads). DISCUSSION
The effects of CX-AAAs, a hyperpolarization 01 horizontal cells and reduction of S-potential in amplitudes, were transient when the toxins were applied in the perfusate. However, at 224 h after intravitreal injection all the horizontal cells penetrated showed a relatively low resting potential in addition to the reduced S-potential. This difference may have been due to differences in time of exposure of retinas receiving r-AAAs by the two routes. Nevertheless their effects were found to be transient in both in z*itro and in z.iro preparations. However, the recovery from the slight hyperpolarization produced by either one of x-AAAs was very slow (IO min), as compared to the immediate recovery from a large depolarization induced by L-glutamate (Fig. I). Although the reason
Fig, 6. Retinal morphology, at various times after intraocular injection of UL-a-AAA. Semi-thin sections of retinas from the eyes mjected with 8 nmoi of IJL-a-AAA one day prior to eye removal were stained with Toiuidine Blue. As compared with control retina (A), the gliai compartment of E,L-a-AAA injected eyes (B) was more densely stained (arrows). At four days after injection of 16.~mol IJL-a-AAA, changes in the inner nuclear and inner plexiform layers are evident (C). The glial changes caused by UL-z-AA/\ (16pmol) gradually recovered, and the normal appearance of retina could be seen at three weeks after injection (D). Scaie bar = 40 pm.
for the slow recovery process with r-AAAs was not explored, it could have been due to the absence of an uptake system for I-AAAs in the fish retina. The experiments reported in the preceding’ and present papers aimed to determine the extent of glio- and neurotoxicity of r-AAAs in the carp retina and to test the specificity of its isomers. The electrophysiological data showed that horizontal cells were hyperpoiari~d with decreasing S-potentials by superfusion of any form of r-AAA isomers (15 mM). It may be possible that the drug exerts a toxic effect on the photoreceptors which, in
turn, reduces the dark release of glutamate (see our preceding paper’). Iiowever, the current study revealed a restoration of the normal responses of horizontal cell (IO-20 mV) by one to two days after injection of any form of ix-AAAs. Acute changes in horizontal cell responses after DL-a-AAA injection were recorded by Wu and Dowling’9 and by Szamier rt ~71.‘~ Therefore horizontal cell changes appear to be of much shorter duration than the biochemical glutamine synthetase (GS) and electroretinogram (ERG) changes (b-wave) noted by Kato
L’t ld7
Neurotoxic
effect of L-c(-AAA in the retina
Ihl
Fig. 7. Morphological changes after intraocular injection of L-r-AAA. Severe changes of glial Miiller cells with pycnosis (arrow) were produced one day after injection of8 pmol L-Y-AAA (A). Additional neuronal injury occurred tn the mner retina with extensive vacuoliration. At two weeks after Injection of 8 Lrmol I-z-AAA. ghal and neuronal changes were similar to those seen four days after injection of lh/tmol I)L-X-AAA (cf. Fig. 6C). A normal architecture was gradually restored over 50~60 days after injection (B, two months). Scale bar = 40/rm.
Although ganglion cell spike activity was weakly activated by iontophoretic application of CC-AAAs, pretreatment with intraocular L-s(-AAA (>4/~mol) produced a long term suppression, and the lightinduced spike discharges of almost all units (>90%) failed to be restored even two months after injection. In contrast, the suppression of ganglion cell activities
caused by a four-fold higher dose of tn.-x-AAA (16 pmol) was reversible, being fully restored two to three weeks later. Bonaventure ct ~1. reported a block of spike discharges in some frog ganglion cells (25%) by intraocular administration of DL-a-AAA. However, they did not indicate whether or not the ganglion cell blockade was reversible. Furthermore. the unresponsiveness of ganglion cell activtties to iontophoretic application of various neurotransmitters indicates that the permanent suppression
Fig. 8. Autoradiograph of [?H]thymidine incorporation in the retina after intraocular injection of IIL-r-AAA. One mtcrocurie of [‘Hlthymidine was injected mto eyes pretreated with an mtravitrcal mjection of DL-a-AAA (I6 pmol) one to t&o days earlier, and the fish were allowed to survive for three to five days. In the control retina (A) [‘Hlthymidme incorporating cells were present in the outer nuclear layer (arrows). In m-u-AAA-pretreated retinas (B) additional lahelled cells (arrowheads) were present adjacent to abnormal appearing Miiller cells, as well as in the outer nuclear layer (arrows). Scale bar = 50 I’m.
caused by L-z-,4AA was not due to a biockadc ot’ presynaptic iransm~ssion but due to an injury to ganghon cells themselves. In their entirety, the present results strongly suggest that the neurotoxic action of CX-AAAisomers in the retina was limited to the L-form of the toxin and had long term effects only on inner neurons, particularly ganglion cells. The apparent recovery of n~orph~logical appearance caused one to two months after injection of L-cc-AAA was at odds with the suppression of ganglion cell activities produced by the same isomer two months after injection. Further structural analyses of the retinas treated with a-AAAs must be conducted to resolve this discrepancy in the fttture.
In the preceding paper,7 we quantified the extent of gliotoxicity produced by t.- and IX-c(-AAA in the fish retina. The gliotoxicity produced by L-r-AAA was comparable to that produced by two-fold higher dose of DL-a-AAA in the reduction of giutamine synthetase and ERG b-wave activities. The ghotoxicity of L-z-AAA even with a high dose (8 Lcrnoi) was transient. The present morphological data further confirmed the reversible gliotoxicity of L- and DLa-AAA. We found that the recovery of the ERG b-wave after intravitreal injection with II prno# I.a-AAA or with 16 pmoi IX-X-AAA required a longer time (two months for 50% recovery in amplitude) than that of the biochemical change (one month for 100% recovery in the GS activity). Since the suppressing effect of the t)L-isomer ( 16{irnoi) on ganglion cells was much weaker than that of the
t.-isomer (4 ~cmof). the functional states of ncuronaI ceils appear to be not invuived in their recovery processes of glial cells observed. The cellular mechanism for the transient nature of the action of sr-AAAs on Miiller cells was suggested by the results of the autoradiogrdphicaf study of ~3~]thyn~dine incorporation. Newly pro~iferat~ilg neurons of ail classes of retinal cell are known to be normaily added to the peripheral retina from the ora serrata, and new rod cells are added to the photoreceptor layer in the outer nuclear laycr.‘~‘~“~” !’ fn the DL-x-AAA (16iimol) treated animals in this study, cells which had incorporated [~~]t~yrnjd~ne were associated with swollen Mtiller cells as well as in germinal cells in the outer nuclear layer. The results suggest that the reversibility from partial destruction of the cells of inner nuclear layer by LXor L-r-AAA may be accomplished by a proliferation of [3H]thymidine incorporating cells. A similar localization of lH]thymidine-incorporating cells in the inner nuclear layer has been observed in juvenile carp retina treated with a high dose of k-hydroxy-
dopamine.” This may indicate a common response of the fish retina to high doses of neurotoxins and:or gliotoxins. The fate of these [‘Hjthymidine-ijthymidine-itlcorpctrating cells, whether differentiating into new Miiltcr c&s or into other classes oi‘ ceils. is unknown.
AIknuwledRemcrrts--We thank Dr A. W. Spira for his correction of our English and also Mrs Tami Urano for her secretarial as&tan& This work was supported in part by research grants (Nos 59570055, 4077~S6, 614801~5 and 01659506 to SK.) from the Ministry of Education, Science and Culture, Japan.
REFERENCES Bonaventure N., Roussel D. and Wioland N. (1981) Effects of ~,L-~pha-amjnoad~pic acid on Miiller ceils in frog and chicken retina in viva: relation to ERG b-wave, ganglion ceti discharge and tectal evoked potentials. Neumsri. Left. 27, 81-87. 2. Casper D. S. and Reif-Lehrer L. (19833 Effects of alpha-aminoadipate isomers on the mo~ho~ogy of the isolated chick embryo retina. invest. Dphthal. his. Sci. 24, 1480-1488. -3 Johns P. R. (1981) Growth of fish retinas, Am. Zoo/. 21, 441-453. 4. Johns P. R. and Fernald R. D. (1981) Genesis of rods in teleost fish retina. Nuture 293, 141-142. 5. Kato S. (1979) C-type horizontal cell .responses to annular stimuli. Expl Eye Res. 28, 627-639. 6. Kato S. and Neeishi (19781 Effects of variations in the oerfusate on the ERG and discharge ganglion cells in carp _ of -_ retina. Expl Eyp Res.‘26, j63-376. 7. Kate S., Sugawara K., ~ats~kawa T. and Negishi K. (1990) Ghotoxic effects of ~aminoadipic acid isomers on the carp retina: a long term observation. Neuroscience 36, 145-153. 8. Kato S., Teranishi T. and Negishi K. (1985) ~-Glutamate depolarizes ON-OFF transient type of amacrine cells in the carp retina: an ionophoretic study. Brain Res. 329, 390-394. 3. Kock J. H. (1982) Neuronal addition and retina! expansion during growth of the crucian carp eye. .J. c’u~prp.Neur
Neurotoxic
effect of L-r-AAA
in the retina
163
IS. Olney J. W.. Ho 0. L. and Rhee V. (1971) Cytotoxic effects of acidic and sulphur ~(~nt~ining amino acids on the infant mouse central nervous system. E.$ Brain Rev. 14, 61 76. 16. Sugawara K. (1985) Lateral actions at the inner plexiform layer of the carp retina: effects of turning windmill pattern stimulus. Visiorr Res. 25, I 179 1186. 17. Sugawara K. and Negishi K. (1973) Effects of some amino acids on light-induced responses in the isolated carp retina C’ision Rcs. 13, 2479--24X9. IX. Szamier R. B.. Ripps H. and Chappell R. L. (1981) Changes in ERG b-wave and Miiller cell structure induced by aipha-aminoadipic acid. Na~~~sci. L~rt. 21, 307 312. 19. Wu S. M. and Dowling J. E. (1978) L-Aspartate: evidence for ;i role in cone photoreceptor synaptic tr~~nsrniss~oll in the curp retina. PFOC.. N&z. Ami. Sci. U.S.A. 75, 5205 5209.