Brain Research 964 (2003) 144–152 www.elsevier.com / locate / brainres

Research report

The retinal anatomy and function of the myelin mutant taiep rat ´ E. Chavez ´ a , Manuel Roncagliolo a , Heidrun Kuhrt b , Andreas Reichenbach b , Andres ´ G. Palacios a , * Adrian a

´ , Molecular Cellular Center for Neuroscience of Valparaiso and Department of Physiology, Faculty of Science. University of Valparaıso ´ , Chile P.O. Box 5030, Valparaıso b Paul Flechsig Institute for Brain Research, University of Leipzig, Leipzig, Germany Accepted 22 November 2002

Abstract Purpose: To study the histology and the physiological function of the retina in the neurological myelin mutant, taiep rats during the postnatal developmental period (P20–P360). Methods: Electroretinography (ERG) was applied to evaluate intensity dependence and spectral sensitivity of the responses to light. Retinal histology, morphometry, and immunocytochemistry were used to characterize the ¨ structure of the retina, with particular emphasis on the Muller (glial) cells. Results: In the taiep rats of all ages studied, the scotopic ERG showed normal a- and b-wave amplitudes and latencies; likewise, the scotopic spectral sensitivity function was the same for control and taiep animals, with a maximal sensitivity ( lmax ) at 500 nm. However, in adult taiep rats (P90 to P360) a secondary cornea-positive wave (‘b 2 ’) was observed in response to high stimulus intensities, which never occurred in controls. This correlated with the observation that in the photopic ERG responses of the taiep rats, the b-wave was reduced in amplitude, and was followed by a rapid cornea-negative after-potential. After 1 year of life, in taiep rats the outer plexiform layer (OPL) became slightly thinner and the inner plexiform / ganglion cell layers (IPL / GCL) appeared to be swollen, and increased in thickness; in addition, the number of retinal neurons (particularly, of ¨ photoreceptor cells) slightly decreased. Increased GFAP immunoreactivity revealed a hypertrophy and reactivity of the Muller cells in 1-year-old taiep rats. Conclusions: The present results suggest the occurrence of a relatively mild and slowly progressing neural retinal alteration in taiep rats, which becomes histologically and functionally evident at the end of the first year of life, and mainly affects the circuit(s) of the photopic ON-response. It is speculated that this alteration is due to missing / altered signals from demyelinated optic nerve.  2002 Elsevier Science B.V. All rights reserved. ¨ (glial) cells; Retina; Visual system Keywords: Electroretinography (ERG); Neurological myelin mutant (taiep); Muller

1. Introduction The autosomal recessive myelin mutant taiep rats show a progressive deterioration of the CNS myelination as a consequence of altered microtubules in the oligodendrocytes [5,9,21] but not in the Schwann cells [14]. Homozygote recessive rats display a progressive neurological syndrome characterized by tremor, ataxia, episodes of immobility, audiogenic epilepsy and hindlimb paralysis. Furthermore in taiep rats, sensory alterations were found in the auditory pathway, optic nerve, and visual cortex [17,28–30]. Unlike in the brain and optic nerve, oligodendrocytes

*Corresponding author. Tel.: 156-32-508-048; fax: 156-32-283-320. E-mail address: [email protected] (A.G. Palacios).

are absent in the retina of most mammalian species, with the exception of the rabbit [31]. The vertebrate retina contains only one predominant type of macroglia cells, the ¨ Muller cells, which span all layers of the retina and are involved in the homeostasis of extracellular ions such as K 1 and H 1 , as well as of neuronally released signal molecules such as glutamate, GABA, and others [3,22]. It is interesting to note in this context that there is evidence ¨ for structural and functional similarities between Muller ¨ cells and oligodendrocytes. For example, Muller cells react with antibodies raised against myelin / oligodendrocyte-specific protein (MOSP) [24] and contain microtubules that are characteristic of oligodendrocytes but not of astrocytes ¨ [25]. Furthermore, Muller cells may be used as ‘early markers’ of various forms of retinal pathology, as they express (increased amounts of) the intermediate filament protein, GFAP, even before functional deficiencies of

0006-8993 / 02 / $ – see front matter  2002 Elsevier Science B.V. All rights reserved. doi:10.1016 / S0006-8993(02)04114-8

´ et al. / Brain Research 964 (2003) 144–152 A.E. Chavez

retinal neurons can be measured (for review, see Ref. [26]). Our study was a first study of retinal structure and function in taiep rats. In particular, we wondered whether ¨ the Muller cells might be involved in retinal pathology, either causally or, at least, as ‘indicators’. Some of the data have already been published in abstract form [6].

2. Material and methods

145

The ERG signals were recorded with a pair of Ag /AgCl electrodes, amplified and band-pass filtered (0.1 kHz to 300 Hz) with a high-gain amplifier (DP-301 Warner Instruments, or Grass Instrument). An oscilloscope was used to visualize the light responses during the experiments. Before each experiment, the photon flux emission from the lamp, between 300 and 800 nm, was measured with a calibrated photocell positioned at the level of the cornea (Optometry S370, UDT Instruments). The software controlling the monochromator, optic wedge and shutter was designed in our laboratory.

2.1. Animals 2.3. Recording and analysis Both taiep and control rats were bred locally at the ´ Animal Facility in the University of Valparaıso, and maintained at between 18 and 20 8C, in a 12-h light / 12-h dark schedule (light on at 08.00 h), with water supply and feeding ad libitum. A total of 54 rats of both sexes (25 controls and 29 taiep rats) were used at various postnatal ages (P) from P20, P30, P60, P90, P120, and P180 to P360. For each of these developmental stages, n54 animals were used from both taiep and control rats, except if indicated otherwise. The early identification of young taiep rats was possible at the end of the first month, by a noticeable tremor of the hind limbs (a symptom that is later associated with other neurological defects).

2.2. Electroretinogram ( ERG) The ERG, a light-induced trans-retinal voltage response, was recorded under both scotopic and photopic conditions. Before each experiment the animals were dark or light adapted (either with a white lamp filament bulb (75 W), or an illuminator given a background illumination of 240 mW/ cm 2 -sr at the cornea, placed at 1 m above the animal), respectively, for 45 min. Then, they were anesthetized with an intraperitoneal injection of a ketamine hydrochloride (120 mg / kg) and xylazine (4 mg / kg) mixture. A local corneal anesthetic (1% lidocaine), and some drops of 1% atropine sulfate for dilation of the pupil, were applied to the eye before the contact electrode was placed. The body temperature was maintained by means of a regulated thermal bed. The optical system consisted of a quartz lamp illumination (250 W), a motorized monochromator (1200 lines, Oriel), an electronic shutter (Vincent Associate) delivering flashes of 10–250 ms duration, and a quartz optical neutral density filter wedge to attenuate the incident photon flux. The maximum light intensity (given as log I50), depending on the experimental day, ranged from 5500 to 9000 photons mm 22 for a 10-ms flash duration and l5500 nm, delivered at the cornea; lower intensities were generated by using the neutral density filter, and are given as log I52n. A quartz fiber optics (Oriel, 77577) was used to direct the light into the eye. In later experiments quartz condenser lenses replaced the optic fiber.

The ERG response was evoked by an increasing number of photons per flash (with 1–2-s intervals between the flashes) at fixed wavelength(s). After a complete series of responses was obtained, the amplitude, latency, time to peak and integration time of the responses were measured for the a-wave (negative peak), b-wave (positive peak) and c-wave (positive peak). The response amplitudes were measured between baseline and peak. The latency was measured from the onset of the stimulus to the onset of the response. The amplitude response was normalized by r / r max 5i /i1s. Where i is the incident photon number at the cornea, r /r max is the normalized response (b-wave) and s is the half-saturating response. For the spectral sensitivity experiments, the amplitude of the b-wave was measured in response to an average of n550 dim flashes, in the range of 330–700 nm. Individual action spectra S( l) were normalized in the vertical axis, using a theoretical template—for values between 450 and 540 nm—that describes the spectra of vertebrate visual pigments [18]. Statistical differences were evaluated by analysis of variance (ANOVA). Excel and Origin software were used for analysis and graphic data presentation.

2.4. Histology, immunocytochemistry, and morphometry After the ERG experiments the animals were profoundly anesthetized, and the eyes were carefully excised. The eyes were incised at the corneal limb, and immediately fixed in 4% paraformaldehyde in phosphate-buffered solution (PBS; 0.1 M, pH 7.4) overnight. Then, the tissue was either embedded in paraffin for conventional processing, or cryostat sections were made (10 mm). The paraffin sections were stained with haemalaun to label the cell nuclei. For immunocytochemistry on the cryostat sections, the following primary antibodies were employed: monoclonal mouse anti-vimentin (V9 clone, Immunotech) and monoclonal mouse-anti-glutamine synthetase (Chemicon). The preparations were pre-incubated in normal goat serum (5% in PBS) for 20 min and then incubated in the primary antibody overnight at 4 8C. Immunoreactivity was visualized with the fluorescent secondary antibody cy 2 -goat-antimouse. In one series of experiments we utilized a poly-

146

´ et al. / Brain Research 964 (2003) 144–152 A.E. Chavez

clonal rabbit anti-GFAP serum (Dako), visualized by the ABC technique (Vectastain Elite ABC kit, Vector Lab) after pronase (1 mg / ml, Sigma) treatment of de-waxed paraffin sections for 15 min at 37 8C, and after blocking of the endogenous peroxidase by incubating the sections in a fresh solution of 0.3% hydrogen peroxide in PBS for 30 min at room temperature. For morphometrical analysis, the retinae were fixed in 2.5% glutaraldehyde (in PBS; 0.1 M, pH 7.4) overnight, post-fixed in 1% OsO 4 (Fluka) in 0.1 M phosphate buffer (pH 7.4), and embedded in Durcupan  . Semi-thin sections were cut (0.5 mm thickness) and stained with toluidine blue. Morphometry was performed by using the Soft¨ Imaging-Systems software (SIS, Munster, Germany). Cell nuclei were counted in all three nuclear retinal layers, within grids corresponding to a length of 20 mm along the retinal sections. The rough data were corrected for the different incidence of cutting cell nuclei in the three layers (due to the different mean diameter of the nuclei, correlated to the thickness of the sections) according to Floderus [11]. The densities of neuronal cell nuclei were compared ¨ to those of Muller cell processes, as counted in sections processed for glutamine synthetase immunocytochemistry ¨ (for this purpose, one cell nucleus per Muller cell process was to be subtracted from the number of nuclei counted in the inner nuclear layer, INL). As a result, the neuron-to¨ Muller cell ratio was estimated for each cellular layer, and for the retina as a whole (cf. Ref. [27]). Statistical differences were evaluated by analysis of variance (ANOVA). Excel and Origin software were used for analysis and graphic data presentation.

3. Results

3.1. Scotopic ERG Fig. 1A shows a characteristic series of scotopic ERG responses ( l 5500 nm) recorded from a P60 taiep rat (top) and a P60 control rat (bottom) elicited by short light flashes (10 ms) of increasing intensity. The responses were similar for the two groups, in both absolute amplitudes and intensity dependence. The inset shows a response of a control rat eye to a high intensity flash. Fig. 1B shows a normalized ERG response-intensity diagram obtained from control and taiep rats at P20 (n57) and P360 (n57); the continuous line represents the Michaelis–Menten function (best fit). Based on these experiments the half-saturating response (s ) intensity was determined for each rat (Table 1). No apparent statistical differences between taiep and control rat scotopic ERGs were found throughout the range of ages studied (P.0.05). However, the data of the taiep group showed a high variability in all parameters measured (cf. Table 1). Fig. 2 shows the averaged spectral sensitivity function measured between 300 and 700 nm; it corresponds well to

Fig. 1. Electroretinographic (ERG) recordings. (A) Characteristic ERG scotopic responses recorded from taiep (top) and control rats (down) at postnatal day (P) 60. Each trace is the average response of 50 trials with l 5500 nm and 10-ms flashes. In taiep the dimmest flash was 0.8 photons mm 22 incident at the cornea, and successive flashes were 1.5, 4, 8, 15, 39, 78, 155, 389, 775, 1804, 4530 and 9039 photons mm 22 and 0.8, 1.5, 4, 8, 15, 39, 78, 155, 389, 775, 1547, 3886 and 7753 photons mm 22 in control (Inset, single scotopic ERG response to a high-intensity flash (log I50), showing the initial negativity (the a-wave), and the subsequent positivity (the b-wave)). (B) Normalized ERG (b-wave) responses (r /r max ) of young (P20; filled circles, n57) and adult (P360; open circles, n57) to 10-ms light flashes of increasing intensities. Data from control and taiep rats of the same age were pooled together since no differences were found between the groups. The curve represents the intensity function, as the best fit of the data by the Michaelis–Menten equation. The vertical dashed line indicates the threshold of the ‘b 2 -wave’ observed in taiep rats at P360 (see Figs. 3 and 4).

the scotopic spectral sensitivity functions described for rats and other mammals [16]. The maximum sensitivity ( lmax ) was 50061.1 nm in control (n528; P20–P360; range, 498–503 nm) and 50061.0 nm in taiep (n528; P20– P360; range, 497–502 nm) rats. Other parameters, such as the amplitudes and latencies of a- and b-wave, showed no significant differences between taiep and control rats at all

´ et al. / Brain Research 964 (2003) 144–152 A.E. Chavez

147

Table 1 Comparison between ERG parameters (means6S.D.) of control (C) and taiep (T) rats Postnatal day (number)

s (photons mm 22 )

a-wave (mv)

Latency a-wave (ms)

b-wave (mv)

Latency b-wave (ms)

C-P20 (n54) C-P120 (n54) C-P360 (n54) T-P20 (n54)

71641 191696 125689 111616

24.660.7 25.160.3 23.060.9 26.960.7

3163.3 3860.4 3260.8 3561.8

1860.2 1663.2 2163.0 1863.4

5160.3 5661.2 5864.2 58610.6

T-P120 (n54)

37626

24.260.1*

3760.6

2662.0*

5262.3

T-P360 (n54)

116637

22.360.5*

3363.3

2461.3*

5869.4

b2 (mv)

Latency b2 (10 ms)

Latency b2 (50 ms)

0.94 (n51) 160.3 (n52) 160.1

230.7 (n51) 24568.0 (n52) 254612.2

28669.3

Four animals for each group were study during the ontogeny (P20–P360). s, flash intensity at the cornea for half-saturating responses; P, postnatal age in days. Asterisks show statistically significant differences within the taiep group (adult / aged versus P20) (P.0.05). For simplicity data for ages of P30, P60, P90, 180 are not shown in the table.

ages studied (P.0.05) (Table 1). Again, we observed a high variability in the amplitudes of a- and b-waves in the taiep group. Nevertheless, if the data were compared among the three ages of the taiep group, slight but statistically significant differences were found; in aged taiep rats (P120, P360), the amplitudes of the a-waves were reduced, and the amplitudes of the b-waves were increased as compared to those of young animals (P20) (P,0.05, see Table 1). The most intriguing observation was a cornea-positive potential that followed the normal b-wave in all adult taiep rats (P180–P360), and also in one P20 taiep rat, but never in one of the controls. Fig. 3 illustrates the ERG responses in one control (a) and two taiep rats (b,c), at five different intensities. A secondary positive wave (‘b 2 ’) occurred only

Fig. 2. Normalized spectral sensitivity function of the ERG b-wave, recorded from control (filled circles; n528; four animals at P20–P360) and taiep rats (open circles; n528; four animals at P20–P360); means and standard deviation (bars) are given. The lmax was found to be at 500 nm in both cases.

in response to high-intensity flashes (the dotted line in Fig. 1B shows the threshold value above which the ‘b 2 -wave’ started to be recordable). This wave reached about 5% of the amplitude of the b-wave; its latency was about 220– 250 ms (Table 1). To discard the possibility that this ‘b 2 -wave’ was an OFF response, we increased the stimulus duration. Fig. 4 shows that increasing the flash duration from 10 to 250 ms failed to enhance the latency of the ‘b 2 -wave’, as would have been the case for an OFF response or d-wave. Rather, the ‘b 2 -wave’ was reduced or even disappeared in the responses to 100 and 250 ms (Fig. 4).

3.2. Photopic ERG In order to obtain more information about the mechanism(s) generating the ‘b 2 -wave’ we performed experiments under photopic conditions (light adaptation 240 mW cm 2 -sr corneal radiance). Fig. 5 shows representative photopic ERGs elicited by short-flashes (10 ms) from two animals at P360 (top, control rat; bottom, taiep rat). In both cases, the three characteristic waves (a-, b- and c-wave) were observed; no significant differences were found in amplitude and latency of these waves between the control and the taiep rats (P.0.05). Similar results were obtained from other animals aged between P20 and P180 (data not shown); a ‘b 2 -wave’ was never observed under photopic conditions (Fig. 5). However, in aged taiep rats the cornea-negative after-potential (after the peak of the b-wave) showed a rapid and strong peak (Fig. 5, asterisk), which was not observed in the photopic ERGs of control rats, or of young taiep rats. Obviously, the ‘b 2 -wave’ was induced in aged taiep rats only under mesopic conditions; i.e., with high intensities in dark-adapted animals (Figs. 3 and 4) when both rod and cone responses contribute to the generation of the ERG [23]. At corresponding high intensities under photopic conditions (i.e., when the rod response is suppressed by light adaptation, and only the cone-driven responses contribute to the ERG), the ‘b 2 -wave’ was missing but a rapid

148

´ et al. / Brain Research 964 (2003) 144–152 A.E. Chavez

Fig. 3. ERG recordings (original traces) from a control (a) and two taiep rats (b,c) at P360; recorded after dark adaptation in response to l 5500-nm light flashes of 10 ms (each trace is the average of 50 trials). Each trace represents a different optical density attenuation (from top to down, 0, 20.3, 21, 22 and 23 OD). The arrows point to the ‘b 2 -wave’ which occurs in the ERG of taiep rats at high light intensities.

large after-negativity occurred in the ERG of aged taiep rats (Fig. 5). Probably, under mesopic conditions a normal saturated rod-driven ERG response will add to the altered photopic response (Fig. 5, bottom, asterisk) in the recorded summated ERG. This summation may cause the appearance of an additional cornea-positive ‘b 2 -wave’ after the cornea-negative ‘incision’ in the course of the cone-driven ERG components (Fig. 3).

3.3. Histology /immunocytochemistry There were no apparent alterations in the histology of young (P20) taiep rats (data not shown). In 1-year-old animals, however, moderate but obvious alterations were found as compared to controls of the same age (Fig. 6A,B). In particular, the innermost retinal layers (inner plexiform and ganglion cell layers, IPL / GCL) appeared to be ‘swollen’ or ‘dispersed’ (Fig. 6B). Furthermore, the thickness of the outer plexiform layer (OPL) was slightly reduced, whereas the thickness of the IPL / GCL was increased; these changes were statistically significant (P, 0.001; cf. Fig. 7C). Furthermore, the number of retinal cells per unit length of retinal section was lower than in the controls (42.05 vs. 44.30 cells / 20 mm); this was mainly

due to a significant (P,0.01) reduction of the cellular density in the outer nuclear layer (ONL) (29.79 vs. 31.93 cells / 20 mm) (Fig. 7A). Assuming a volume shrinkage factor of about 50% due to the fixation and staining procedures, this would roughly correspond to a reduction of photoreceptor cell densities from 316 000 to 295 000 cells mm 22 . ¨ When Muller cells were immunolabeled for the cell type-specific proteins, vimentin and glutamine synthetase, no differences in cell densities or labeling intensities were apparent between taiep and control rats, at any age studied (data not shown). In both controls and taiep rats, the ¨ Muller cells became weakly immunoreactive for GFAP at P60 (taiep) to P90 (controls) (not shown). At P360, GFAP¨ immunolabeling was significantly stronger in Muller cells of taiep rats than in cells from controls (Fig. 6C–D). ¨ Moreover, the GFAP-immunolabeled Muller cell processes were thicker and coarser in the taiep retinae than in the controls (Fig. 6D vs. C). ¨ When the Muller cell nuclei were identified by their irregular shape and dark staining (arrows in Fig. 6B) [13], ¨ and counted in the semi-thin sections, neuron-to-Muller cell ratios could be estimated. In accordance with earlier counts [4], about 29.7 neuronal cells (0.58 cells in the

´ et al. / Brain Research 964 (2003) 144–152 A.E. Chavez

Fig. 4. ERG recordings (each trace is the average of 50 trials) from a control (left column of traces) and a taiep rat (right column of traces) both recorded at P360 in response to 500-nm light stimuli (log I50). The stimulus duration was varied between 10, 50, 100 and 250 ms but neither the latency nor the peak times of the ‘b 2 -wave’ (arrows) were shifted. This indicates that the ‘b 2 -wave’ is not an OFF response.

GCL, 3.85 cells in the INL, and 24.27 cells in the ONL) ¨ were counted per Muller cell in the control rats (Fig. 7B). ¨ In the taiep rat retina, the number of neurons per Muller cells was reduced to 26.96; this reduction was mainly due to a decrease of the cell number in the ONL (21.81). It is concluded that a slight but specific and significant (P, 0.01) loss of (mainly, photoreceptor) neurons occurs in the retina of aged taiep rats.

4. Discussion Summarizing the present results, there were rather subtle but distinct histological and functional alterations in the retinae of taiep rats at the end of the first year of life. ¨ According to these observations (Figs. 3–5 and 7), Muller cells became reactive (‘gliotic’) as indicated by the enhanced expression of GFAP and the hypertrophy of their processes (Fig. 6C–D). As in many other cases of rat ¨ retinal pathology [2,7,10,12,13] this Muller cell reactivity may be indicative of injuries of retinal neurons. In

149

Fig. 5. Typical averaged for scotopic (dashed traces) and normalized photopic (continuous traces: light adaptation background 240 mW/ cm 2 -sr corneal radiance). The 10-ms flash stimulus intensity, for scotopic and photopic condition, was 2020 photon 21 mm 2 , l 5500 nm. ERG responses from a control (top) and a taiep rat (bottom) recorded at P360. The asterisk labels the rapid large after-negativity following the peak of the b-wave in taiep rats.

astrocytes of taiep rats, an enhanced immunoreactivity for GFAP was also observed; however, this astroglial reactivity was already found at birth in the spinal cord ventral columns, and within the first months of life in most brain regions [19]. The comparatively late onset of glial reactivi¨ ty in the retina (present data) may indicate that Muller cells are not causally involved in the retinal alterations of taiep rats, and may be explained by the fact that oligodendrocytes (which are primarily affected in the mutants) are present in spinal cord and brain but not in the retina. There remain the questions of (i) which type(s) of retinal neurons are primarily concerned, and (ii) how the taiep mutation may affect the retina, despite of its lack of oligodendrocytes. The ERG technique, widely used in the diagnosis of retinal diseases, may provide some initial cues. There were no significant differences in the a-wave amplitudes and latencies under both scotopic and photopic conditions between control and taiep rats. The significant but slight decrease in the a-wave amplitudes of taiep rats at P120 and P360 (as compared to taiep rats at P20; Table 1) may correspond to the histologically assessed loss of about 7% of the photoreceptor cells (Fig. 7C). Certainly this does not

150

´ et al. / Brain Research 964 (2003) 144–152 A.E. Chavez

Fig. 6. Retinal morphology. Cresyl toluidine blue semi-thin sections (A,B) and immunocytochemically processed cryostat sections (C,D) of control (A,C) ¨ and taiep rat retinae (B,D) at P360. In the taiep rat retinae, the inner retinal layers (IPL and GCL) showed a ‘swollen’ appearance (B versus A); the Muller cells of these retinae showed a dramatically enhanced immunoreactivity for glial fibrillary acidic protein (GFAP) and a hypertrophy of their processes (D ¨ ¨ versus C). The arrows in (B) point to some of the irregularly shaped Muller cell nuclei; in one case, a connection between such a nucleus and a Muller cell process in the IPL (arrowhead) is visible. NFL / GCL, nerve fiber / ganglion cell layer; IPL, inner plexiform layer; INL, inner nuclear layer; OPL, outer plexiform layer; ONL, outer nuclear layer; PRL, photoreceptor layer. The calibration bar (valid for A–D) represents 25 mm.

represent a selective loss of cones, as cones constitute less than 3–5% of all rat photoreceptors [32,33] and as the photopic ERG should be dramatically reduced if the majority of cones were lost. However, the a- and b-wave amplitudes of the photopic ERGs were not significantly decreased even in aged taiep rats (Table 1). Furthermore, the scotopic spectral sensitivity function (measured from the amplitude of the b-wave) was normal, and the latencies and amplitudes of the b-wave in scotopic ERGs were also not decreased in taiep rats at all intensities between threshold and rod saturation (about 2 log units above threshold). Thus, as a first summary it may be stated that although a minority of photoreceptors (predominantly, rods?) are lost after 1 year of life, the gross function of cone and rod photoreceptors is virtually unchanged, and the circuits of the scotopic system seem to be well maintained. However, our data suggest that the function of secondand / or third-order neurons of the photopic system might be impaired. At P360, we observed the appearance of the

novel ‘b 2 -wave’ in the mesopic ERG of taiep rats. Although future experiments (pharmacology, intra-retinal recordings) will be necessary to verify the source of the ‘b 2 -wave’, there is good reason to assume that it is generated by the rapid after-negativity following the peak of the photopic b-wave (Fig. 5) which would then constitute the principal change in the ERG of aged taiep rats. At this time-point of its course, the photopic sum ERG is thought to result from the balance between the corneanegative component PIII (generated by the cone photo¨ receptors and the Muller cells [20]) and the cornea-positive component PII . As the amplitude of the a-wave (which represents the commencing part of the PIII and is generated by the photoreceptors) is unchanged or even slightly decreased (Table 1), and as the number of photoreceptors is even reduced in taiep rats (Fig. 7), the rapid afternegativity is probably caused by an accelerated decay of the PII rather than by an enlarged PIII . The PII , as the component underlying the b-wave, is thought to be generated by interactions between ON-bipolar cells and amac-

´ et al. / Brain Research 964 (2003) 144–152 A.E. Chavez

151

the inner plexiform layer (swelling?) and the reduced thickness of the outer plexiform layer (shrinkage?) might reflect functional and / or structural alterations of the (ONbipolar cell) synapses involved in the generation of PII and b-wave. Finally, it remains difficult to speculate about the pathological mechanism(s) that cause the observed alterations in the taiep retina. First, we found no evidence for ¨ any primary involvement of Muller cells, as neither the ¨ density of Muller cells nor the expression of the important ¨ Muller cell-specific enzyme, glutamine synthetase, were changed. Thus, despite some structural and functional ¨ similarities between oligodendrocytes and Muller cells ¨ [24,25], the taiep mutation seems not to affect the Muller cells directly. Most probably, there occur slow (hitherto unknown) retrograde effects from the demyelinated optic nerve where the taiep mutation is known to affect the oligodendrocytes [30]. For instance, it is conceivable that the retrograde transport of neurotrophic factors along the optic nerve (for a recent review, see Ref. [1]) may be disturbed; however, there was no apparent loss of retinal ganglion cells (Figs. 6 and 7) and thus, retrograde degeneration (if any) must be very slow, and must cause transsynaptic functional impairments before detectable cell death occurs in the ganglion cell layer. Future quantitative electron microscopic studies are required to elucidate the pathological mechanism(s) underlying the retinal alterations in the taiep mutant rat. At the present state, it can be summarized that we were able to demonstrate, for the first time, functional and structural changes in the retina of the taiep mutant rat. This means that even such CNS tissues which fail to contain oligodendrocytes, and which receive only weak centrifugal innervations by myelinated fibers [15], may finally be affected by the taiep mutation.

Fig. 7. Mean (6S.D.) thickness of retinal layers (A) and cell densities (B) in control (grey columns) and taiep rats (black columns). (A) Thickness of retinal layers, as measured in semi-thin sections (cf. Fig. 6A,B; IPL1GCL, thickness of the inner retinal layers, i.e., between inner plexiform layer, IPL, and inner retinal surface). (B) Numbers of cell nuclei counted in the three nuclear layers (GCL, ganglion cell layer; INL, ¨ inner nuclear layer; MC, Muller cell nuclei; INL-MC, difference between ¨ all nuclei in the INL and Muller cell nuclei, i.e., number of neuronal cell nuclei in the INL; ONL, outer nuclear layer; neurons, all neuronal cell nuclei) along a distance of 20 mm of the semi-thin sections (cf. Fig. 6A,B). Significant differences between control and taiep data are ¨ indicated by **(P,0.01) and *** (P,0.001). (C) Neuron-to-Muller cell ratios calculated for retinal neurons in the three nuclear layers. No standard deviation or significant differences are indicated for these indirectly derived data.

¨ rine cells (and, possibly, Muller cells) [8]. Thus, the most reasonable assumption is that in aged taiep rats, the coneON-bipolar cell- and / or amacrine cell-generated responses are diminished in amplitude and duration. Although we found no indications of reduced neuronal cell numbers in the inner nuclear layer (Fig. 7), the increased thickness of

Acknowledgements We would like to thank Guillermo Rodriguez for his invaluable technical assistance and animal care. This work was supported by Fondecyt [1000396 (A.P.), [1991004 (M.R.), Iniciativa Milenio Mideplan [ P99-037-F, and by the Deutsche Forschungsgemeinschaft, Graduiertenkolleg ‘Intercell’ (GRK 250 / 2-99) and by the Bundesministerium ¨ Bildung, Forschung und Technologie (BMB1F), fur interdisciplinary Center for Clinical Research (IZKF) at the University of Leipzig (01KS9504, Project C5). The authors are indebted to Dr. Peter Laming for his critical reading of the manuscript.

References [1] M. Bahr, Live or let die—retinal ganglion cell death and survival during development and in the lesioned adult CNS, Trends Neurosci. 23 (2001) 483–490.

152

´ et al. / Brain Research 964 (2003) 144–152 A.E. Chavez

¨ [2] A. Bignami, D. Dahl, The radial glia of Muller in the rat retina and their response to injury. An immunofluorescence study with antibodies to the glial fibrillary acidic (GFA) protein, Exp. Eye Res. 28 (1979) 63–69. [3] A. Bringmann, S. Skatchkov, T. Pannicke, B. Biedermann, H. ¨ Wolburg, R. Orkand, A. Reichenbach, Muller glial cells in Anuran retina, Microsc. Res. Tech. 50 (2000) 384–393. [4] T.I. Chao, J. Grosche, K.J. Friedrich, B. Biedermann, M. Francke, T. ¨ Pannicke, W. Reichelt, M. Wulst, C. Muhle, S. Pritz-Hohmeier, H. Kuhrt, F. Faude, W. Drommer, M. Kasper, E. Buse, A. Reichenbach, ¨ Comparative studies on mammalian Muller (retinal glial) cells, J. Neurocytol. 26 (1997) 439–454. [5] E. Couve, J. Cabello, J. Krsulovic, M. Roncagliolo, Binding of microtubules to transitional elements in oligodendrocytes of the myelin mutant taiep rat, J. Neurosci. Res. 47 (1997) 573–581. ´ [6] A. Chavez, M. Roncagliolo, H. Kuhrt, A. Reichenbach, A.G. Palacios, Functional and morphological study of the retina in the myelin mutant taiep, Biol. Res. 34 (3–4) (2001) R42. [7] D.A. DiLoreto, J.R. Ison, G.P. Bowen, C. Cox, C. Del Cerro, A functional analysis of the age-related degeneration in the Fischer 344 rat, Curr. Eye Res. 14 (1995) 303–310. [8] C. Dong, W. Hare, Contribution to the kinetics and amplitude of the electroretinogram b-wave by third-order retinal neurons in the rabbit retina, Vision Res. 40 (2000) 579–589. ´ [9] I. Duncan, K.F. Lunn, B. Holmgren, R. Urba-Holmgren, L. Brignolo-Holme, The taiep rat: a myelin mutant with an associated oligodendrocyte microtubular defect, J. Neurocytol. 21 (1992) 870– 874. ¨ [10] A.J. Eisenfeld, A.H. Bunt-Milam, P.V. Sarthy, Muller cell expression of glial fibrillary acidic protein after genetic and experimental photoreceptor degeneration in the rat retina, Invest. Ophthalmol. Vis. Sci. 25 (1984) 201–214. ¨ [11] S. Floderus, Untersuchungen uber den Bau der menschlichen ¨ Hypophyse mit besonderer Berucksichtigung der mikromorphologis¨ chen Verhaltnisse. Acta Pathol. Mikrobiol. Scand. Suppl. 53 (1944) 1–23. ¨ [12] J. Grosche, W. Hartig, A. Reichenbach, Expression of glial fibrillary acidic protein (GFAP), glutamine synthetase (GS), and Bcl-2 ¨ protooncogene protein by Muller cells in retinal light damage of rats, Neurosci. Lett. 185 (1995) 119–122. [13] J. Grosche, D. Grimm, N. Clemens, A. Reichenbach, Retinal light damage vs. normal aging of rats: altered morphology, intermediate ¨ filament expression, and nuclear organization of Muller (glial) cells, J. Brain Res. 38 (1997) 459–470. ´ [14] B. Holmgren, R. Urba-Holmgren, L. Riboni, E.C. Vega-Saenz de Miera, Sprague–Dawley rat mutant whit tremor, ataxia, tonic immobility episodes, epilepsy and paralysis, Lab. Animal Sci. 39 (1989) 226–228. [15] S.K. Itaya, P.W. Itaya, Centrifugal fibers to the rat retina from the medial pretectal area and the periaqueductal grey matter, Brain Res. 326 (1985) 362–365. [16] G. Jacobs, J. Neitz, J. Deegan, Retinal receptors in rodents maximally sensitive to ultraviolet light, Nature 53 (1991) 655–656.

[17] J. Krsulovic, E. Couve, M. Roncagliolo Dysmyelination, demyelination and reactive astrogliosis in the optic nerve of the taiep rat, Biol. Res. 32 (1999) 253–262. [18] T.D. Lamb, Photoreceptor spectral sensitivities: common shape in the long-wavelength region, Vision Res. 22 (35) (1995) 3083–3091. [19] B.A. Leon Chavez, J. Guevara, S. Galindo, J. Luna, A. Ugarte, O. Villegas, R. Mena, J. Eguibar, D. Martinez-Fong, Regional and temporal progression of reactive astrocytosis in the brain of the myelin mutant taiep rat, Brain Res. 900 (2001) 152–155. [20] T. Matsuura, W.H. Miller, T. Tomita, Cone-specific c-wave in the turtle retina, Vision Res. 18 (1978) 767–776. ¨ [21] J.R. Moller, P.G. Durr, R.H. Quarles, I.D. Duncan, Biochemical analysis of myelin proteins in a novel neurological mutant: the taiep rat, J. Neurochem. 69 (1997) 773–779. ¨ [22] E. Newman, A. Reichenbach, The Muller cell: a functional element of the retina, Trends Neurosci. 19 (1996) 307–312. [23] P. Nixon, B. Bui, J. Armitage, J. Vingrys, The contribution of cone response to rat electroretinograms, Clin. Exp. Ophthalmol. 29 (2001) 193–196. [24] F. Prada, A. Quesada, Y. Aguilera, C. Santano, C. Prada, The Muller cell expresses myelin oligodendrocyte specific protein (MOSP), Ophtalmic Res. 33 (2001) 237–250. [25] A. Reichenbach, Attempt to classify glial cells by means of their ¨ process specialization using the rabbit retinal Muller cell as an example of cytotopographic specialization of glial cells, Glia 2 (1989) 250–259. ¨ [26] A. Reichenbach, S.R. Robinson, The involvement of Muller cells in the outer retina, in: M. Djamgoz, S.N. Archer, S. Vallerga (Eds.), Neurobiology and Clinical Aspects of the Outer Retina, Chapman & Hall, London, 1995, pp. 395–416. [27] A. Reichenbach, M. Ziegert, J. Schnitzer, S. Pritz-Hohmeier, P. Schaaf, W. Schober, H. Schneider, V. Development of the rabbit retina, The question of columnar units, Dev. Brain Res. 79 (1994) 72–84. [28] M. Roncagliolo, J. Contreras, M. Valenzuela, Visual evoked potentials and desmyelination in optic nerves of neurological mutant taiep rats, Not. Biol. 1 (2) (1993) 106. ´ [29] M. Roncagliolo, J. Benıtez, J. Eguibar, Progressive deterioration of central components of auditory brainstem responses during postnatal development of the myelin mutant taiep rat, Audiol. Neuro-Otol. 5 (2000) 267–275. ´ J. Silva, J.R. Eguibar, The recovery cycle [30] M. Roncagliolo, C. Leon, of excitability in isolated optic nerve of myelin mutant taiep rats during postnatal development, J. Physiol. 523P (2000) 54P. [31] J. Schnitzer, Distribution and immunoreactivity of glia in the retina of the rabbit, J. Comp. Neurol. 240 (1985) 128–142. [32] S. Sokol, Cortical and retinal spectral sensitivity of the hooded rat, Vis. Res. 10 (1970) 253–262. ´ P. Rohlich, ¨ [33] A. Szel, Two cone types of rat retina detected by antivisual pigment antibodies, Exp. Eye Res. 55 (1992) 47–52.