The inferior colliculus of GEPRs contains greater numbers of cells that express glutamate decarboxylase (GAD67) mRNA.

The inferior colliculus of GEPRs contains greater numbers of cells that express glutamate decarboxylase (GAD67) mRNA.


Introduction
Previous studies have shown that 7-aminobutyric acid (GABA) is a major inhibitory neurotransmitter in the inferior colliculus (IC) and plays a role in seizure initiation following auditory stimulation in both normal Sprague-Dawley (SD) rats and genetically epilepsy-prone rats (GEPR-9s) 1'7'1°. For example, the GABA antagonist, bicuculline, has been shown to induce audiogenic seizure-like activity in rats when injected into the IC I°. In addition, morphological studies using immunocytochemistry with an antiserum to glutamate decarboxylase (GAD) have shown greater numbers of GABAergic and total neurons in the IC of GEPR-9s as compared to non-epileptic SD rats 1~'tt'. Furthermore, biochemical analysis of the central nucleus of the IC has revealed elevated GABA content in the GEPR-gs as compared to SD rats 13. Thus, there appears to be an abnormality of the GA-BAergic system in the IC of GEPR-9s, and it may play a vital role in their seizure susceptibility. The objective of this study was to use in situ hybridiza-tion to determine whether abnormalities in the numbers and distributions of neurons containing GAD67 mRNA exist in the inferior coliiculus of GEPR-9s.
In the present study, we examined the autoradiographic localization of hybridization of a 35S-labeled RNA probe, complementary to the 67-kD form of GAD mRNA, in tissue sections through the three major nuclei of the IC (the central nucleus, external nucleus and dorsal cortex) of both GEPR-9s and SD rats. Light microscopy was used to count the labeled neuronal cell bodies. Also, counts were made in the frontal cortex, an area where the number of labeled neuronal cell bodies in the GEPR-9s and SD rats was not anticipated to differ because this region is not required for the initiation or propagation of seizures in GEPR-9s 4 and previous biochemical studies showed no differences in GABA content ~3. This analysis has shown that the number of cells that express GAD67 mRNA is greater in the central and external nuclei of the IC of GEPR-9s.

Methods
Genetically epilepsy-prone rats that exhibited maximal seizures (GEPR-9s) (n=3) and age matched Sprague-Dawley (SD) rats (n ~ 3) were used in the present study. Paired GEPR-9s (without evidence of seizure behavior h~amediately prior to sacrifice) and SD rats were killed by sodium pentobarbital overdose and perfused intracardially with 0.9% NaC! followed by 4% paraformaldehyde in 0.1 M phosphate buffer, pH 7,2 (PB). The brains were postfixed in the perfusate for 24 h at 4°C and then transferred to 20% sucrose in 4% paraformaldehyde/PB tbr 48 h at 4"C, Sections of 30 lma thick were cut in the sagittal plane using a freezing microtome and collected into cold 4% paraformaldehyde/PB and stored at 4"C until processed.
Tissue sections were processed tbr the in situ hybridization localization of glutamate decarboxylase (GAD67) mRNA as previously described 9. Briefly, sections were pretreated with proteinase K (1 #g/ mi in 0,1 M Tris buffer, pH 8.0, with 50 mM EDTA) at 37°C for 30 min, and then were incubated with a 35S-labeled cRNA prepared from cDNA pBSGAD 2"3 (provided by D.L. Benson and E.G. Jones). The GAD cDNA corresponds to bases 1324-1683 of cat GAD cDNA reported to encode the 67-kD form of GAD 2'6. The antisense cRNA was transcribed from the BamHl linearized cDNA using T7 RNA polymerase in the presence of aSS-UTP. The incubation with the labeled cRNA was done in a hybridization buffer containing 50% formamide, 10% dextran sulfate, 0.7% Ficoll, 0.7% polyvinyl pyrrolidone, 0.7% bovine serum albumin, 0.15 mg yeast tRNA, 0.33 mg/ml denatured herring sperm DNA, 40 #M dithiothreitol, 24 mM sodium chloride and 2.4 mM sodium citrate. The hybridization incubation was done at 60°C for 30-40 h with the cRNA at 1 x 106cpm/100#l.The sections were then treated with ribonuclease A (20 #g/ml in 10 mM Tris-saline, pH 8.0, with 1 mM EDTA) tbr 30 min at 45°C and washed through descending concentrations of saline sodium citrate (SSC) buffer (pH 7.0) to a final stringency of 0.1 x SSC. The sections were mounted onto gelatincoated-slides and air-dried.
The distribution of cellular cRNA hybridization was evaluated using emulsion autoradiography (Kodak NTB2 nuclear track emulsion mixed 1:1 with H20) and exposures of 2-3 weeks at 4°C. Following the development of the tissue autoradiograms, the sections were stained with cresyl violet and coverslipped with Permount.
Controls for the cRNA probe hybridization specificity included hybridization of tissue with the 3sS-labeled sense RNA sequence and treatment of tissue sections with ribonuclease A prior to normal incubation with the antisense cRNA probe. The sense-strand RNA was prepared by transcription from Pvull linearized pBSGAD using T3 RNA polymerase in the presence of 35S-UTP. In both cases, hybridization was not detected. In addition, the distribution of the hybridization-labeled neurons in the present material agreed with previous immunocytochemical reports on the localization of GAD-containing cells in the rat brain, including the inferior colliculus I 1, 16 The labeled neuronal cell bodies were counted using a light microscope at 20 x objective magnification. Dark-field microscopy was used to visualize the neuronal cell bodies that were labeled by the autoradiographic method. If a cluster of silver grains was observed over a cell body and the cluster displayed a significantly greater number of grains than the background (by two standard deviations), this neuron was determined to be autoradiographically labeled by the in situ hybridization method. Labeled cells in both GEPR-9s and SD rats were counted on coded slides by the same individual. Counts were made for the central nucleus ot" the IC and the frontal cortex using a grid with an area of 32,250 /tm 2. A similar method was used for counting the external nucleus and dorsal cortex of .. the IC except that only half of the grid ~,as used. In this way, it was possible to determine specifically the number of labeled cells in the latter two regions and to eliminate the possible error of counting adjacent regions of the IC (e.g., the central nucleus). Each area was identified using previously described criteria 8"~2.
The Macintosh StatView program was used to ,vbtain the means and standard errors for the GEPR-9s and SD rats. Statistical significance was determined with an unpaired t-test.

Results
The in situ hybridization autoradiograms of parasagittal sections through the brains of GEPR-9s and SD rats showed specific hybridization of the GAD67 cRNA to neurons throughout many brain regions (Fig. 1). The distribution of labeled neu-,'ons was similar to that previously described for this cRNA 2. For example, Purkinje cells were densely labeled in the cerebellar cortex. Other autoradiographically labeled neurons were found in the inferior colliculus, cerebral cortex, superior colliculus, zona incerta, thalamic reticular nucleus, substantia nigra and throughout the brainstem. The appearance and numb~er of labeled cells were most evident using darkfield illumination (Fig. 1). How- ever, it is irnportant to note that in brightfield optics many large, Nissl stained cell bodies were overlaid with clusters of black silver grains (Fig. 2). Therefore, the clusters of grains observed in darkfield optics corresponded to neuronal cell bodies and were not found to be overlying capillaries or glial cells. The inferior colliculus (IC) of the GEPR-9s (Fig,  3B) contained more autoradiographically labeled cells than that of the SD rats (Fig. 3A). This increase in number was observed to be greatest in the central nucleus of the IC. Within this region there was a broad range in the apparent size of the cRNA labeled cells in the SD rat (Fig. 4). A similar range of sizes for cRNA labeled cells was found in GEPR-9s but the number of labeled small cells (with nuclei less than 15 lma in diameter) appeared to be greater irt these rats (Fig. 4B). The quantitative results confirmed these qualitative observations (Fig. 5). In the central nucleus of the IC the mean number of labeled cells per unit area was 67.7 in GEPR-9s as compared to 37.5 in SD rats. Quantification of GAD67 cRNA labeled cells in the external nucleus of the IC indicated a difference in this region as well, Specifically, GEPR-9s had 64 labeled cells per unit area compared to 38.5 for SD rats. The means for both of these nuclear regions were significantly different (P < 0,01, unpaired t-test). Cell counts from the dorsal cortex of the IC and the frontal cortex revealed no significant differences in GAD67 cRNA labeling between GEPR-9s and SD rats (Fig. 5), although the dorsal cortex of the IC showed a trend for greater numbers of cells in GEPR-9s.
The nurnber of silver grains overlying similarsized, labeled neuronal somata in the IC of GEPR-9s appeared to be equivalent to that of SD rats. Ill contrast, in the superficial layers of the adjacent superior collicu!us, individual neurons appeared to be labeled with a greater density of autoradiographic grains in the GEPR-9s as compared to the SD rats. However, the elevated labeling density in the superior colliculus did not appear to be associated with an increase in the number of labeled neurons (Fig. 3A,B).  . These data were obtained Ii"om three SD rats tha~' did not display audiogenic seizures and from three GEPR-gs that displayed maximal audiogenie seizures. The frontal cortex (FC) was also counted. Asterisks indicate significant differences in GEPR-9s as compared to SD rats. Standard errors are indicated by small brackets.

Discussion
The major finding of this study is that the number of neurons that express GAD67 mRNA is significantly greater in the central and external nuclei of the inferior colliculus (IC) of GEPR-9s as compared to SD rats. These data confirm the previous immunocytochemical and biochemical data that showed greater numbers of GAD-immunoreactive neurons and GABA content, respectively, in GEPR-9s 13"16. The results of these studies support the hypothesis that an abnormality of the GA-BAergic system exists in the inferior colliculus of this genetic model of epilepsy.
The greater number of GAD67 cRNA labeled cells in the IC of GEPR-9s observed in the present study probably reflects an elevated number of GAD-containing cells as previously reported by Roberts et al. 16 rather than aberrant GAD67 mRNA expression by populations of normally non-GABAergic neurons of the IC. Previous quantitative studies of 40 #m thick and 2/am semithin sections showed that there were more neurons in the central nucleus of the IC of GEPR-9s as compared to SD rats 15'~6, and that this increase was due to a greater number of small neurons. The category of small cells was also fotmd to be greater in the quantitative study of GAD-positive neurons in GEPR-9s ~6. Similarly, in the present study we observed elevated numbers of small GAD67 cRNA labeled cells in the IC of GEPR-9s.
There was no apparent difference in the density of autoradiographic grains overlying GAD67 cRNA labeled cells in the IC from SD rats and GEPR-9s (cf. Fig. 4A,B). Thus, GAD67 mRNA content does not appear to be expressed at supranormal levels by IC neurons in GEPR-9s. However, we cannot exclude the possibility that a subpopulation of neurons that express GAD67 mRNA at very low levels and are not detectably labeled with the cRNA in SO rats increase expression to detectable levels in GEPR-9s, and are therefore only counted as GAD67 mRNA-positive in this latter group. It should be noted that this possibility is also relevant to immunocytochemical studies where greater numbers of GAD-immunoreactive neurons were counted. However, the greater total number of neurons in the IC of GEPR-9s, as compared to SD rats 16, suggests that differences between numbers of GAD-synthesizing neurons contribute to the increased counts found in both in situ hybridization and immunocytochemical preparations. Moreover, the immunocytochemical study showed that the proportion of GAD-positive neurons to total neurons was greater ill GEPR-9s as compared to SD rats 16. Since the ribonuclease treatment included in tile present ill situ hybridization method removes much of the Nissl staining capacity of neurons, it is not possible to count with confidence the numbers of total neurons in such preparations and determine whether the proportion of GAD67 mRNA labeled cells is also greater in GEPR-9s.
In contrast to this finding in the IC, the superficial layers of the superior colliculus of GEPR-9s sho~ved a higher density of silver grains overlying individual neurons. This finding indicates that GABA is synthesized in greater amounts in GEPR-9s in this region of the mesencephalon. The superior colliculus is considered an important structure for the propagation of seizures in this model of epilepsy 14. Thus, the superior colliculus provides further evidence for a link between aberrations in GAD67 mRNA expression and seizure propagation. It will be of interest to determine if GABA mediated inhibition, as assessed electrophysiologically, is abnormal in the superior colliculus as has been reported for the inferior colliculus of GEPR-9s 7.
Although the data support the existence of a defect in GAD expression in the IC of GEPR-9s, it is probably not the only abnormality that is underlying epileptogenesis in the IC of this model. Previous biochemical studies have shown not only elevated GABA content in the IC of GEPR-9s but also greater than normal levels of the excitatory neurotransmitter, glutamate ~3. Therefore, both inhibitory and excitatory mechanisms appear to be abnormal in GEPR-9s.
Finally, the data fi'om the present study confirm that a change in the nurnber of GABAergic neurons occurs in the IC of GEPR-gs. The elevated number of GABAergic neurons is indicative of a greater potential for GABA release and action. However, there are two t'orms of GAD mRNA, 65 kD and 67 kD (also known as GAD65 and GADt,7, respectively) that are encoded by different genes s'6 and are differentially regulated in neurons of adult brains under some experimental conditions. For example, levels of GAD67 mRNA, the form evaluated in the present study, are altered within the striatum following deafferentation 17. It will be of interest to determine if GAD65 mRNA is also abnormal and elevated in IC of GEPR-9s and if the expression of these mRNAs is abnormal prior to the developmental appearance of seizure activity. Abnormalities in the expression of both isoforms in neonatal GEPR-9s would argue for congenital disturbances in the regulation of GAD gene expression in this model of epilepsy as opposed to a~terations in GAD expression that are secondary to changes in physiological activity (i.e., seizures) or connectivity.