- Email: [email protected]
Up-regulation of GABAA receptor binding on neurons of the prefrontal cortex in schizophrenic subjects
Pergamon
PII:
Neuroscience Vol. 75, No. 4, pp. 1021–1031, 1996 Copyright ? 1996 IBRO. Published by Elsevier Science Ltd Printed in Great Britain 0306–4522/96 $15.00+0.00 S0306-4522(96)00328-4
UP-REGULATION OF GABAA RECEPTOR BINDING ON NEURONS OF THE PREFRONTAL CORTEX IN SCHIZOPHRENIC SUBJECTS F. M. BENES,*†‡ S. L. VINCENT,* A. MARIE* and Y. KHAN* *Laboratory for Structural Neuroscience, McLean Hospital, 115 Mill Street, Belmont, MA 02178, U.S.A. †Department of Psychiatry and the Program in Neuroscience, Harvard Medical School, Boston, Massachusetts, U.S.A. Abstract––Recent investigations have reported a reduced density of interneurons and an increase of GABAA receptor binding occurring preferentially in layer II of the anterior cingulate cortex of schizophrenic subjects [Benes F. M. et al. (1992) J. Neurosci. 12, 924–929]. Since a reduction in the density of interneurons has also been found in layer II of the prefrontal cortex, this study has sought to determine whether an up-regulation of the GABAA receptor binding activity might also be found in this region of schizophrenics. A high-resolution autoradiographic analysis of bicuculline-sensitive [3H]muscimol (GABAA) receptor binding on individual neuron cell bodies in layers II, III, IV and VI has been applied to Brodmann area 10 from normal controls (n = 16) and schizophrenic (n = 7) subjects. A computerassisted technique has been used under strictly blind conditions to determine whether differences in binding occur in the schizophrenic group. A significant increase of GABAA receptor binding activity has been observed in layers II, III, V and VI in the schizophrenic group. When the binding is expressed as a density with respect to neuronal cell size, there is a gradient of binding across layers II, III, V and VI, with neuronal cell bodies in layer II having the greatest density of grains. When different subpopulations of neurons distinguished according to size criteria are examined separately, large (pyramidal) neurons show significantly higher binding, particularly in layer II, where it was increased by 90% in schizophrenics. Small (non-pyramidal) cells do not show significant differences in binding in schizophrenics, except in layer VI, where there was a 135% increase. Potential confounding effects from age and post mortem interval do not explain the differences between the two groups, because both young and old schizophrenics, as well as schizophrenics with long and short post mortem intervals, showed increased GABAA receptor binding activity when compared to control cases distinguished in a corresponding fashion. These data suggest that there may be a preferential reduction of inhibitory GABAergic inputs to pyramidal neurons, particularly in layer II of the prefrontal cortex, in schizophrenia. This change could potentially result in an increased excitatory outflow from the prefrontal area to other cortical regions of the schizophrenic brain. Overall, these results are consistent with the idea that reduced amounts of GABAergic activity in the prefrontal cortex could be related to a perinatal disturbance and could be a potentially important component of the pathophysiology of psychosis. Copyright ? 1996 IBRO. Published by Elsevier Science Ltd. Key words: Pyramidal cells, non-pyramidal neurons, layer II, Brodmann area 10.
In recent years, a renewed interest in the question of whether there are histopathological changes in the brains of schizophrenic patients has been generated by reports of volume loss,13,14 decreased numbers of neurons4,6,12,18,24 and other subtle cytoarchitectural variations3,5,7,23,27 in corticolimbic brain regions of individuals with schizophrenia. One study reported a reduction in the density of non-pyramidal neurons, particularly in layer II, of the anterior cingulate cortex and suggested that the missing cells might be GABAergic interneurons.6 To test the hypothesis that this latter change might result in a compensatory ‡To whom correspondence should be addressed. Abbreviations: ANOVA, analysis of variance; PMI, post mortem interval.
up-regulation of the postsynaptic receptor, a highresolution receptor binding technique for GABAA receptor binding activity48 was modified for human post mortem cortex.10 Using this approach, a marked increase of receptor binding was found in the neuropil of layer I and on neuronal cell bodies of layers II and III in the anterior cingulate cortex of schizophrenics;10 there was no increase of binding in the deeper laminae of this region. These results were interpreted as being consistent with a loss of inhibitory interneurons having occurred during the perinatal period, a time when basket cells in superficial laminae are actively differentiating.30,31 Although cortical areas differ considerably with regard to their anatomical and functional specializations, one study has suggested that diverse regions of
1021
1022
F. M. Benes et al. Table 1. Demographic analysis of post mortem cases
Brain no.
Diagnosis
Gender
Age (years)
Cause of death
PMI (h)
1317 1524 1542 1545 1547 1597 1603 1638 1696 1700 1705 1706
Control Control Control Control Control Control Control Control Control Control Control Control
Male Male Female Female Female Male Male Male Female Female Male Female
72 71 78 58 75 82 75 35 63 91 53 77
16.5 8 20 18 5 14 5.5 7.5 19 13 6 6
1733 1734 1737 1880
Control Control Control Control
Female Male Male Male
60 40 72 64
acute respiratory distress bronchopneumonia cardiac arrest bronchopneumonia coronary artery disease lung carcinoma lung carcinoma congestive heart failure thyroid carcinoma acute peritonitis lung carcinoma chronic obstructive pulmonary disease bronchopneumonia cardiac arrest coronary artery disease pancreatic carcinoma
18 48 24 15.9
1313 1318 1348 1350 1420 1576 1702
Schizophrenic Schizophrenic Schizophrenic Schizophrenic Schizophrenic Schizophrenic Schizophrenic
Male Female Female Male Male Male Male
34 37 74 36 35 67 62
bronchopneumonia suicide ischemic necrosis of colon coronary atherosclerosis suicide myocardial infarction carcinoma of large bowel
12 23 22.5 24 6.2 18 20.7
primate cortex develop in a relatively synchronous fashion.37 In this setting, a prenatal disturbance could potentially interfere with normal cell migration during development and induce preferential changes in superficial laminae of other cortical regions, such as the prefrontal cortex, where a reduced density of interneurons was also observed in layer II.6 To explore this, a high-resolution analysis of GABA receptor binding on cell bodies of two basic categories of cortical neurons, the so-called pyramidal and non-pyramidal neurons,35 has been performed on a cohort of post mortem normal and schizophrenic subjects. EXPERIMENTAL PROCEDURES
Post mortem specimens Post mortem specimens of prefrontal cortex (Brodmann area 10) from normal control (n = 16) and schizophrenic (n = 7) subjects were obtained through the Human Brain Tissue Resource Center at McLean Hospital (Table 1). All psychiatric cases for which a diagnosis of schizophrenia could be established were included in the study. The diagnosis of schizophrenia was made by performing a retrospective review of patient records and applying the criteria of Feighner et al.19 All brain specimens, including the prefrontal area, were evaluated by a neuropathologist to detect evidence of intracranial diseases, such as Alzheimer’s dementia, cerebrovascular disease or other more diffuse changes arising from alcohol abuse. Demographic data, including age (years), post mortem interval (PMI; h) and neuroleptic exposure (expressed as the chlorpromazineequivalent dose), were also collected. Tissue handling All brain specimens were removed at the time of autopsy and transported on ice prior to removal of tissue samples. Blocks of prefrontal cortex (Brodmann area 10) were placed
Neuroleptic exposure
0 — 400 — 200 0 100
in ice-cold 0.1% formaldehyde in 0.1 M phosphate buffer (pH 7.4) for 1.5 h prior to cryoprotection in ice-cold 30% sucrose in 0.1 M phosphate buffer (pH 7.4). The tissue blocks were left overnight in the sucrose solution, before being frozen on dry ice, imbedded in Tissue Tek OTC Compound (Miles, Elkhart, IN, U.S.A.) and stored at "70)C. Specimens from both the control and schizophrenic groups were stored for one to four years at "70)C. Only one control case was stored for a period of less than one year. The tissue blocks were sectioned (10 µm in thickness) on a cryostat at "20)C, thaw-mounted on gelatin-coated, acid-cleaned glass slides and maintained at "20)C until all blocks had been sectioned. Tissues from normal and schizophrenic cases were sectioned alternately to avoid a sequencing bias that might potentially be incurred during cutting and storage of the slide-mounted sections. All control and schizophrenic tissue sections were processed simultaneously to minimize experimental variance. Six sections per case (three for total binding and three for inhibited binding) were incubated in 0.31 M Tris–acetate (pH 7.4) at room temperature for 20 min to wash out endogenous ligand. The slides were then incubated for 40 min at room temperature in a solution containing 5 nM [3H]muscimol (specific activity 25.8 Ci/mmol; DuPont/New England Nuclear) in Tris–acetate buffer. Non-specific binding was assessed by incubating a parallel series of sections in tritiated ligand plus unlabelled 100 µM (+)-bicuculline methiodide. All of the sections were then rapidly washed in buffer, immersed in cold distilled water, rapidly dried with a stream of air and stored at 4)C. Using a modification10 of a previously described method,48 the slides were prepared for autoradiography by apposing acid-cleaned coverslips (no. 00, Corning Glass Works, Pittsburgh, PA, U.S.A.) coated with Kodak NTB-3 nuclear emulsion (diluted 1:1 with distilled water). To ensure that the grains representing binding activity were kept in proper registration with the tissue section, the coverslips were glued with cyanoacrylate to one end of the slide-mounted sections under darkroom conditions, secured with small binder clips and exposed for approximately six weeks. After exposure, the binder clips were removed under safelight conditions and the coverslips were gently bent
Up-regulation of GABAA receptor binding in schizophrenia
1023
Table 2. The average size of large and small neurons in the prefrontal cortex of normal control and schizophrenic subjects Large neurons Layer II III V VI
Small neurons
Control
Schizophrenic
Control
Schizophrenic
108.4 & 2.0 163.3 & 3.6 172.0 & 3.5 134.3 & 2.6
119.6 & 5.1 179.9 & 7.5 187.3 & 7.5 144.7 & 4.3
65.1 & 0.9 83.4 & 1.5 84.1 & 1.4 76.8 & 1.3
69.3 & 1.2 81.6 & 2.4 82.4 & 2.3 80.1 & 2.1
The data shown are the average size (µm2) & S.E.M. of large and small neurons in layers II, III, V and VI of normal controls and schizophrenic subjects. The values shown represent the average of the total number of cells sampled.
away from the tissue sections with a small Teflon spacer that was placed between the coverslips and slid about halfway down their length. The latent images were developed in a solution of Dektol for 4 min at 19)C, then washed in distilled water for 15 s, fixed with Kodak rapid fixer (without hardener) and rinsed with water for 30 min prior to staining with Thionin. Microscopic analyses All control and schizophrenic cases that were processed were included in the analyses. The autoradiographic preparations were visualized with a #100 oil immersion objective lens on a Leitz Laborlux microscope interfaced with a Bioquant MEG IV Image Analysis System (R and M Biometrics, Nashville, TN, U.S.A.) and a Wyse 386 PC computer via a Dage CCD71 video camera and a Summagraphics digitizing board. The final image magnification on the video monitor was #3800 and was of sufficiently high resolution to permit the visualization of both neuronal cell bodies and autoradiographic grains. The morphometric sampling technique employed was derived from the principles described by Weibel,46 with all procedures being conducted under strictly ‘‘blind’’ conditions using a technique described in detail elsewhere.10 Briefly, the slides were analysed by moving the field laterally through each layer, choosing the first 18 cells that could be identified as neurons, and tracing around their neuronal cell body to determine their size (µm2). This was followed by a semiautomated determination of the corresponding number of grains at a level of focus above the selected neuron.10 Non-pyramidal neurons were distinguished from glial cells using standard criteria: (i) the presence of euchromatin in neurons, but heterochromatin in glia; (ii) the presence of a nucleolus in neurons, but not glia; and (iii) the presence of Thionin-stained cytoplasm and dendritic processes in neurons, but not glia.4,6 To distinguish between pyramidal and non-pyramidal neurons, the size distributions for these two cell types were determined independently in tissue sections that were not subjected to the mechanical trauma arising from the apposition of emulsion-coated coverslips. A neuron was designated as ‘‘pyramidal’’ if it showed a characteristic triangular shape and an apical dendritic shaft, while multipolar neurons were considered to be ‘‘non-pyramidal’’ in nature. The average sizes of pyramidal neurons in layers II, III, V and VI were 96 & 12, 148 & 25, 158 & 31 and 117 & 18 µm2, respectively. For non-pyramidal neurons in these same laminae, the average sizes were 73 & 7, 90 & 7, 86 & 11 and 84 & 14 µm2, respectively. The size of pyramidal neurons in the Cresyl Violet-stained sections was approximately 14% larger than that for ‘‘large’’ neurons in the Thionin-stained autoradiograms, while that for nonpyramidal cells was 7% larger than ‘‘small’’ neurons. The greater difference for the pyramidal neurons can probably be explained by the somewhat better visualization of the proximal portions of apical dendritic shafts in the Cresyl Violet-stained sections when compared to the Thioninstained autoradiograms. This factor would have less of an
impact on the size of non-pyramidal neurons where the proximal dendritic branches are quite small relative to the overall size of the somata. Because the distribution profiles for both cell types were relatively narrow, one standard deviation was subtracted from the mean for both pyramidal and non-pyramidal neuron sizes in each layer, and the midway point between each was established. This latter number was used as a cut-off between ‘‘small’’ and ‘‘large’’ neurons in layers II (82 µm2), III (110 µm2), V (112 µm2) and VI (98 µm2) of the actual cell samples that had been collected for the autoradiographic analyses of GABA receptor binding activity. The average sizes of the resulting ‘‘large’’ and ‘‘small’’ neurons are shown in Table 2. Although it is likely that some small pyramidal neurons were included in the non-pyramidal cell population, the actual number was probably relatively low. The average diameter of each grain was approximately 0.3–0.5 µm. The computer-generated counts were manually corrected for underestimates arising from overlapping grains. This procedure was performed in each cortical layer of both the inhibited and uninhibited autoradiograms for each case. Neurons that were overlapping other cells were not included in the sample. Neuronal cell bodies in layer I were too sparse to be sampled, while those in layer IV were not analysed because of difficulties in identifying this lamina in the autoradiographic preparations. Statistical analyses Specific GABAA receptor binding was determined by subtracting the number of grains in sections with [3H]muscimol + bicuculline from those with [3H]muscimol alone for each normal and schizophrenic case. The individual case averages were used to generate a ‘‘mean of means’’ & S.E.M. for the respective groups. Student’s t-test was used to assess whether the differences between the groups were significant. The Bonferroni correction is required for multiple comparisons when selected comparisons, simultaneous examination of several different treatment contrasts or data dredging occurs.15 The cell types examined have been implicated in the pathophysiology of schizophrenia,6,8 and the six layers each have a unique pattern of connections, specific functional implications and precisely timed appearances during development. Thus, it was necessary to examine each cell type and layer separately and perform multiple comparisons in order to properly test the central hypothesis of this study that layer II plays a central role in schizophrenia. A two-way repeated measures analysis of variance (ANOVA) was used to evaluate whether the composite data for layers II, III, V and VI were different for the control vs schizophrenic groups. RESULTS
Visual inspection of the Thionin-stained autoradiogram assemblies indicated that there was both good grain development and adequate preservation
1024
F. M. Benes et al.
Fig. 1. Composite light photomicrographs of pyramidal neurons in the prefrontal cortex (Brodmann area 10) of a normal control (A) and a schizophrenic (B) subject, showing autoradiographic grains representing bicuculline-sensitive [3H]muscimol binding in layer II. The pyramidal cells were photographed in register with their corresponding grains. Fields of this type were analysed with a userinteractive automated image analysis system. Scale bar = 10 µm.
of morphological detail in both the normal and schizophrenic cases (Fig. 1A, B). Computer-assisted analyses of the autoradiograms revealed that the non-specific [3H]muscimol binding remaining after inhibition with bicuculline (range 2.5–5.0 grains/ neuron) was similar across the various layers of both the normal and schizophrenic groups. As shown in Fig. 2, the average size of neuronal cell bodies was similar for the normal and schizophrenic groups in each of the layers examined, but for both groups, those in layer II were considerably smaller than those in layers III, V and VI. Figure 3 indicates that the number of grains per individual neuronal cell body was significantly higher for the schizophrenic group, with the repeated measures ANOVA yielding F1,21 = 7.1 for the main effect of group (P = 0.015) in layers II (70%), III (44%), V (48%) and VI (66%). When the data were expressed as the density of receptor binding activity (number of grains/µm2 of cell size), the schizophrenic group showed a gradient across the layers, with layer II having the highest and layer VI having the lowest amount of binding activity (Fig. 4). Using the repeated measures ANOVA, the main effect of group had an F1,21 = 4.23 (P = 0.05). Only layers II and VI had differences that were statistically significant. As Table 2 shows, the sizes of
Fig. 2. A bar graph showing the size of neuronal cell bodies in layers II, III, V and VI of the prefrontal cortex (Brodmann area 10) from the control (solid bar) and schizophrenic (open bar) groups. The data shown are the average size & S.E.M. expressed in µm2 for the respective control (n = 16) and schizophrenic (n = 7) group. The mean size of neurons for each individual case was averaged for the respective groups.
Up-regulation of GABAA receptor binding in schizophrenia
Fig. 3. A bar graph showing specific GABAA receptor binding activity on neuronal cell bodies in layers II, III, V and VI of the prefrontal cortex (Brodmann area 10) from normal control (solid bar) and schizophrenic (open bar) subjects. The data shown were obtained by using the average number of grains per neuron for each subject in the respective control (n = 16) and schizophrenic (n = 7) groups to generate a mean and S.E.M. for each group. Using a repeated measures ANOVA, there is significantly more binding activity on neurons of the schizophrenic group (P = 0.015) in each of the four layers examined.
both large (pyramidal) and small (non-pyramidal) neurons were similar for both groups in layers II, III, V and VI. When the specific receptor binding activity was evaluated separately for the two populations of neuronal cell bodies (Fig. 5, upper and lower panels), the density of binding activity on large neurons was higher in layers II, III and V, with the main effect of group having a repeated measures F1,21 = 6.9 (P = 0.016). The difference in GABAA receptor binding was greatest in layer II, where large neurons showed a 90% increase relative to the normal controls, but the magnitude of the differences progressively decreased in layers III, V and VI. The binding activity for small cells was significantly increased (repeated measures F1,21 = 5.3, P = 0.03) only in layer VI, where it was 135% higher in the schizophrenic group. Table 3 indicates the average age and PMI for the normal and schizophrenic cases. The age of the control group (66.4 years) was considerably higher than that for the schizophrenic group (49.1 years), while the PMI was lower in the controls (12.6 h) than in the schizophrenics (18.2 h). In both cases, the differences between the two groups were not significant. To assess whether these two potentially confounding variables may have influenced the pattern of results described above, the normal and schizophrenic cases were pooled and both age and PMI were correlated with GABAA receptor binding. There was no correlation between either potential confound and the receptor binding activity found in any of the layers (Table 4). When the data were
1025
Fig. 4. A bar graph showing the density of specific GABAA receptor binding activity on neuronal cell bodies in layers II, III, V and VI of the prefrontal cortex (Brodmann area 10) from the normal control (solid bar) and schizophrenic (open bar) groups. The density data shown were obtained by dividing the number of grains per neuron by the size of each neuron. The mean number of grains per µm2 of neuronal cell body for each case was then averaged for control (n = 16) and schizophrenic (n = 7) groups. Using a repeated measures ANOVA, there is a significantly higher density of grains on neuronal cell bodies in layers II and VI of the schizophrenic group (P = 0.05). Overall, the grain density in the schizophrenics shows a gradient across the various laminae, with layer II showing the highest receptor binding activity and layer VI showing the lowest activity.
further broken down according to those above and below 45 years of age, young (mean = 37 & 4 years) and old (mean = 70 & 10 years) controls both showed binding activity on neuronal cell bodies that was much lower than that for young (mean = 36 & 1 years) and old (mean = 67 & 6 years) schizophrenics (Table 5). Similarly, when the data were distinguished according to short (<15 h) and long (>15 h) PMIs, the GABAA receptor binding on neuronal cell bodies of normal controls was considerably lower than that in corresponding subgroups of schizophrenic subjects (Table 6). To assess the potential effect of neuroleptic exposure on the increase of GABAA receptor binding, two schizophrenics with no neuroleptic exposure were evaluated separately from the rest of the group. The average binding in layer II for these two cases (7.2 & 1.5 grains/neuron) was lower than that for the subjects who had received neuroleptic medication (12.9 & 2.2 grains/neuron), and was only slightly higher than the average binding in this lamina for the control group (6.6 & 0.7 grains/neuron). However, in a previous study of the anterior cingulate cortex,8 these same two cases were observed to have GABAA receptor binding activity in layer II (46.6 grains/ neuron) that was much higher than that found for the schizophrenic group as a whole (38.0 grains/neuron). While it is conceivable that there are differential neuroleptic effects on prefrontal vs anterior cingulate
1026
F. M. Benes et al. Table 4. Correlation of GABAA receptor binding activity with age and post mortem interval Layer II III V VI
Age (years)
PMI (h)
0.082 0.047 0.027 0.046
0.10 0.19 0.08 0.10
The numbers shown are the correlation coefficient, r, obtained for the GABAA receptor binding activity (grains/ cell) in layer II of both the control and schizophrenic cases. Similar values were obtained when the receptor binding activity was expressed as a density (grains/µm2).
Fig. 5. A set of bar graphs comparing the total [3H]muscimol binding activity on large versus small neurons in layers II, III, V and VI of the prefrontal cortex from the control (solid bar) and schizophrenic (open bar) groups. The data are expressed as the number of autoradiographic grains per neuron & S.E.M. for the total number of large and small neurons in the various laminae of each group. The numbers above the horizontal brackets indicate the level of significance for the differences between the normal control and schizophrenic groups. There was a total of 480 large and 668 small neurons analysed in the various laminae of the control group, and 292 large and 212 small neurons in the various laminae of the schizophrenic group included in the data represented in graphs A–D. Using a repeated measures ANOVA, there is a significantly higher density of grains on large neurons of the schizophrenic group (P = 0.016) in layers II, III, V and VI. The density of autoradiographic grains is significantly higher on small neurons in layers II and VI, but not III and V. Table 3. Average age and post mortem interval for the control and schizophrenic groups
Controls Schizophrenics
n
Age (years)
PMI (h)
16 7
66.4 & 14.9 49.1 & 17.4
12.6 & 6.4 18.2 & 7.1
The age and PMI for the control and schizophrenic groups are expressed as mean & S.D.
cortices, it seems more likely that factors other than drug exposure could have influenced the GABAA receptor binding activity in the prefrontal area of the two schizophrenic subjects who received no neuroleptic medication (see Discussion section). DISCUSSION
The results of this study demonstrate that there is an increase in GABAA receptor binding activity on
individual neuronal cell bodies in the prefrontal cortex (Brodmann area 10) of schizophrenic subjects. In a pattern similar to that found previously in the anterior cingulate region,8 the density of autoradiographic grains showed a gradient of increase across layers II–VI of schizophrenics, with the greatest percentage increase in binding occurring on large neurons in layer II. In the anterior cingulate cortex,8 when the binding was expressed either as the number of grains per cell or as a density, an inside-out gradient in the overall binding across layers II–IV was observed. In the prefrontal cortex, where the size of neurons is much smaller in layer II than in layers III and V, a gradient of this type was found when the data were expressed as the percentage increase in the density (number of grains per unit area of cell size) for individual neurons, but not when expressed as the number of grains per large neuron. It was important to consider whether the differences in age or PMI between the controls and schizophrenic groups could possibly account for the increased GABAA receptor binding found in the patients. This concern was diminished when a breakdown of cases according to young vs old and short vs long PMIs showed a persistence of increased receptor binding activity in the schizophrenic subjects. This latter observation is consistent with reports from other laboratories that receptor binding activity remains remarkably stable for approximately 24 h after death.33,47 Moreover, the fact that there were differences in the degree to which GABAA receptor binding activity was increased on large vs small neurons, as well as on large neurons of superficial laminae vs those in deep layers, further argues against the possibility that differences in PMI are responsible for the findings reported here. A confounding effect of this type would be expected to non-specifically affect all neuronal types irrespective of the layer in which they are found. A further area of concern emerged when it was found that the two neuroleptic-naive cases did not show increased GABAA activity in the prefrontal area. This was puzzling since the same two cases showed a marked up-regulation in the anterior cingulate region,8 making it seem less likely that the findings reported here could be explained by a drug
Up-regulation of GABAA receptor binding in schizophrenia
1027
Table 5. Differences in GABAA receptor binding in the prefrontal cortex of young and old subjects Controls Age range (years)
Schizophrenics
<45
>45
<45
>45
Average age (years)
37 & 4 n=2
70 & 10 n = 14
36 & 1 n=4
67 & 6 n=3
Layer II III V VI
6.6 & 2.0 7.1 & 5.0 5.1 & 2.6 2.2 & 1.7
6.6 & 0.8 9.8 & 1.1 7.1 & 1.0 3.6 & 0.5
11.3 & 1.9 14.6 & 2.7 10.6 & 1.4 5.3 & 0.7
11.3 & 3.4 12.6 & 3.5 9.6 & 2.7 6.0 & 2.3
The data shown are the mean & S.E.M. grains/cell for GABAA receptor binding activity on pyramidal neurons in layers II, III, V and VI of young (<45 years) and old (>45 years) subjects. The average age & S.D. is shown for each age range of the respective control and schizophrenic groups. Table 6. Differences in GABAA receptor binding in the prefrontal cortex of subjects with short and long post mortem intervals Controls PMI range (h) Average PMI (h) Layer II III V VI
Schizophrenics
<15 6.9 & 2.3 n=8
>15 18.3 & 2.8 n=8
<15 8.6 & 3.4 n=2
>15 22.0 & 2.8 n=5
10.1 & 1.0 13.4 & 2.1 10.3 & 1.5 5.4 & 1.0
6.2 & 1.3 9.4 & 4.1 7.7 & 1.9 3.4 & 0.9
14.2 & 3.2 18.2 & 6.6 12.2 & 0.7 5.5 & 1.2
14.4 & 2.7 16.2 & 2.8 13.6 & 1.8 6.7 & 1.6
The data shown are the mean & S.E.M. grains/cell for GABAA receptor binding activity on pyramidal neurons in layers II, III, V and VI for subjects with short and long PMIs.
effect. It is noteworthy that one of the neuroleptictreated cases (B1350) also showed GABAA receptor binding in the prefrontal cortex that overlapped with the control group, but not in the anterior cingulate region, where it was increased. This observation suggests that there may be region-to-region differences in GABAA receptor binding activity that occur independently of neuroleptic exposure, and it may well be that such cases may eventually be distinguished according to subtle variations in the schizophrenic syndrome. Nevertheless, the possibility that neuroleptic effects might account for the differences in GABAA receptor binding in the prefrontal cortex of schizophrenics, though unlikely, cannot as yet be definitively excluded. Abnormalities of GABAergic neurotransmission have long been suspected as playing a role in the pathophysiology of schizophrenia.38 Earlier investigations have reported an increase of [3H]muscimol binding,22 decreased glutamate decarboxylase activity,11 decreased GABA uptake40 and a reduction in the number of neurons expressing mRNA for glutamate decarboxylase1 in the frontal area of the schizophrenic brain. While the data reported in these studies first suggested that a decrease of GABAergic activity might be present in the cortex of schizophrenics, small effects and large variances discouraged interest in these findings. With the high-resolution technique employed here, it was
possible to analyse binding on neuronal cell bodies where GABAA receptor binding activity is high10 and eliminate areas where it is entirely absent (nonneuronal compartments). Consequently, the emulsion coverslip technique has made it possible to obtain data with a large effect size and small variance that has intrinsically greater statistical power. An additional advantage of the high-resolution method employed here is the fact that receptor binding activity can also be differentially assessed on neuronal subpopulations, such as large and small cells. Although these latter two neuronal populations are probably not composed exclusively of one cell type, it seems likely that the majority of large neurons are pyramidal in nature, while the small neurons may be predominantly composed of non-pyramidal cells. While pyramidal neurons are the principle source of efferent flow from the cortex, non-pyramidal neurons modulate the activity of intrinsic circuits locally within each cortical layer.18,26 As a result, changes in GABAA receptor binding on one or the other cell type would likely result in different effects on the activity of intrinsic circuits within the prefrontal cortex. In the current study, an increased amount of GABAA receptor binding was found most consistently on pyramidal neurons in layers II, III and V, suggesting that these latter neurons would be more affected by a decrease of inhibitory activity in schizophrenic subjects (see
1028
F. M. Benes et al.
Fig. 6. A schematic diagram presenting a hypothetical model to account for the differences in specific GABAA receptor binding activity on neurons of the prefrontal cortex in schizophrenic subjects as compared to normal controls. (A) A low-power Nissl-stained section of human prefrontal cortex (Brodmann area 10), showing layers I–VI. Scale bar = 250 µm. (B) Normal circuit. In the middle panel, a normal circuit is shown with one pyramidal neuron present in layers II and V. Each pyramidal neuron receives an inhibitory input from two GABAergic interneurons. (C) Schizophrenic circuit. In the right-hand panel, a similar circuit in a schizophrenic subject also has one pyramidal neuron in layer II. In layer V, however, there are two pyramidal neurons, rather than one as in the normal circuit, as suggested by the increased density of these cells in this lamina of schizophrenics.6 In layer II, there is only one GABAergic interneuron providing an inhibitory input to the pyramidal neuron in layer II to account for the up-regulation of GABAA on these cells. In layer V, each pyramidal neuron also receives inhibitory input from only one GABAergic interneuron, rather than two, because there was no increase of non-pyramidal neurons noted previously in this layer.6 In the setting of increased pyramidal neurons, there would be a relative decrease of inhibitory activity from the normal number of GABA cells and this would give rise to compensatory up-regulation of GABAA receptor binding activity on pyramidal neurons of layer V in the schizophrenic group.
below) than similar cells in the deeper laminae. It is important to emphasize that the up-regulation of GABAA binding has also been found in the neuropil of layers I–III of the anterior cingulate region of schizophrenics.8 Although this compartment was not included in the current study, it is likely that this receptor binding activity was also increased in the neuropil of the prefrontal cortex. Accordingly, the up-regulation of the GABAA receptor in schizophrenics may also be associated with distal dendritic branches and axon terminations that are abundant components of the neuropil. Recent studies of the anterior cingulate cortex have emphasized the importance of alterations in layer II of schizophrenics. The changes noted in this region have included smaller clusters of neurons,3 a decreased density of interneurons,6 increased GABAA receptor binding activity8 and an increased
density of vertical axons5 that are probably glutamatergic afferents to the cingulate region.7 Layer II abnormalities have also been described in the entorhinal23 and prefrontal areas6 of schizophrenics. The fact that the cerebral cortex develops in an ‘‘inside-out’’ fashion, with layer II being the last cell-rich lamina to appear and differentiate,30,31,36,39 makes it plausible that a prenatal disturbance in cell migration could account for changes in the frontal area of schizophrenics.1 At the time of birth in humans, layer II basket cells are still quite immature,31 and theoretically could be more vulnerable to a pre- and/or perinatal injury than cells that are fully differentiated.2,8 Functional considerations An important observation reported in this study is that there appears to be a greater up-regulation of GABAA receptor binding on pyramidal neurons
Up-regulation of GABAA receptor binding in schizophrenia
than on small neurons, which are presumed to be primarily non-pyramidal in nature. This implies that the putative decrease of GABAergic inhibition in schizophrenics6,8 would give rise to an increased firing of these large projection neurons. This, in turn, could result in a greater outflow of activity from the prefrontal area to other cortical and subcortical sites with which it is connected, although it is possible that this increase of receptor binding activity on postsynaptic pyramidal neurons could compensate for a reduction of released GABA. In general, there was less change of GABAA receptor binding on non-pyramidal neurons in the prefrontal area of schizophrenics, except for layer VI interneurons, which did show a very striking increase. In this regard, it is important to point out that there are many different types of interneurons, each with unique laminar distributions and functional roles.16 For example, small and large basket cells are GABAergic inhibitory neurons distributed throughout layers II–VI;26 chandelier cells are also GABAergic cells, but ones found principally in upper cortical laminae;34 double bouquet cells are believed to be GABAergic inhibitory neurons that play a disinhibitory role;41 granule cells are excitatory cells found predominantly in layer IV;29 Martinotti cells are excitatory elements found primarily in layers V and VI.42 The heterogeneity of the nonpyramidal cell population16 and the inability of the autoradiographic technique employed in this study to distinguish the various subtypes noted above make it difficult to predict how the up-regulation of GABAA receptor binding activity on these latter cells of layer VI would impact on the overall output of these laminae. To understand the general outcome of a decreased inhibitory input to pyramidal neurons, it is pertinent to note that the termination sites of these cells in layers V and VI are in the thalamus, caudate nucleus and other subcortical loci, while those in layers II and III send an efferent outflow primarily to other cortical regions.20,25 Many pyramidal neurons in superficial cortical laminae also give off collateral branches to deeper laminae, particularly layer V.32 It is noteworthy
1029
that the density of pyramidal cells in layer V of the prefrontal cortex has been found to be significantly higher in schizophrenic subjects, while nonpyramidal cells in this layer showed no difference.6 With an increase of pyramidal neurons, the normal number of interneurons might provide inadequate inhibitory GABAergic input to each of these projection cells in layer V and result in increased efferent outflow (Fig. 6). Another pertinent question is whether the alterations of GABAergic activity in the prefrontal cortex of schizophrenics reported here may be related in some way to the mesocortical dopamine projections. Direct interactions between dopamine-containing axon varicosities and GABAergic interneurons have recently been proposed,21,28 and empirical evidence for this relationship has been reported,9,43 although such contacts are non-synaptic in nature43 and may only involve modulatory influences.9 The interaction of dopamine varicosities with GABAergic neurons could well be functional in nature, because dopamine receptor binding activity is also localized on non-pyramidal neurons of the prefrontal cortex.44,45 Thus, the ability of neuroleptic drugs to relieve the symptoms of schizophrenia could involve, at least in part, the blockade of dopamine receptors on an impaired population of inhibitory interneurons.2 Future studies will seek to determine whether other corticolimbic brain regions considered key regions of interest in schizophrenia research, such as the hippocampal formation and entorhinal region, also show increased GABAA receptor binding activity in subjects with this disorder. If so, there will be compelling evidence for the idea that alterations of GABAergic transmission may be an important component of the pathophysiology of schizophrenia.
Acknowledgements—This work has been supported by NIMH grants MH00423, MH42261 and MH31154, and an award from the Stanley Foundation. The authors wish to thank Drs Steven Matthysse and Alan M. Zaslavsky for their consultation concerning the statistical analyses.
REFERENCES
1. 2. 3. 4. 5. 6. 7.
Akbarian S., Bunney W. E., Potkin S. G., Wigal S. B., Hagman J. D., Sandman C. A. and Jones E. G. (1993) Altered distribution of nicotinamide-adenine dinucleotide phosphate-diaphorase cells in frontal lobe of schizophrenics implies disturbances of cortical development. Archs gen. Psychiat. 50, 227–230. Benes F. M. (1995) Is there a neuroanatomic basis for schizophrenia? An old question revisited. The Neuroscientist 1, 104–115. Benes F. M. and Bird E. D. (1987) An analysis of the arrangement of neurons in the cingulate cortex of schizophrenic patients. Archs gen. Psychiat. 44, 608–616. Benes F. M., Davidson J. and Bird E. D. (1986) Quantitative cytoarchitectural studies of cerebral cortex of schizophrenics. Archs gen. Psychiat. 43, 31–35. Benes F. M., Majocha R., Bird E. D. and Marrotta C. A. (1987) Increased vertical axon numbers in cingulate cortex of schizophrenics. Archs gen. Psychiat. 44, 1017–1021. Benes F. M., McSparren J., Bird E. D., Vincent S. L. and SanGiovanni J. P. (1991) Deficits in small interneurons in prefrontal and anterior cingulate cortex of schizophrenic and schizoaffective patients. Archs gen. Psychiat. 48, 996–1001. Benes F. M., Sorensen I., Vincent S. L., Bird E. D. and Sathi M. (1992) Increased density of glutamateimmunoreactive vertical processes in superficial laminae in cingulate cortex of schizophrenic brain. Cerebral Cortex 2, 502–512.
1030 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43.
F. M. Benes et al. Benes F. M., Vincent S. L., Alsterberg G., Bird E. D. and SanGiovanni J. P. (1992) Increased GABA-A receptor binding in superficial layers of cingulate cortex in schizophrenics. J. Neurosci. 12, 924–929. Benes F. M., Vincent S. L. and Molloy R. (1993) Dopamine-immunoreactive axon varicosities form nonrandom contacts with GABA-immunoreactive neurons in rat medial prefrontal cortex. Synapse 15, 285–295. Benes F. M., Vincent S. L. and SanGiovanni J. P. (1989) High resolution imaging of receptor binding in analysing neuropsychiatric diseases. Biotechniques 7, 970–979. Bird E. D., Spokes E. G. S. and Iversen L. L. (1979) Increased dopamine concentration in limbic areas of brain from patients dying with schizophrenia. Brain 102, 347–360. Bogerts B. and Falkai P. and Tutsch J. (1986) Cell numbers in the pallidum and hippocampus of schizophrenics. In Biological Psychiatry (eds Shagass C. et al.), Vol. 27, pp. 1178–1180. Elsevier, Amsterdam. Bogerts B., Meertz E. and Schonfieldt-Bausch R. (1985) Basal ganglia and limbic system pathology in schizophrenia: a morphometric study of brain volume and shrinkage. Archs gen. Psychiat. 42, 784–791. Brown R., Colter N., Corsellis J. A. N., Crow T. J., Frith C. D., Jagoe R., Johnstone E. C. and Marsh L. (1986) Postmortem evidence for structural brain changes in schizophrenia. Differences in brain weight, temporal horn area and parahippocampal gyrus width as compared with affective disorder. Archs gen. Psychiat. 43, 36–42. Dunnett C. and Goldsmith C. (1981) When and how to do multiple comparisons. In Statistics in the Pharmaceutical Industry (eds Buncher R. G. and Tsya I.-Y.), pp. 397–435. Marcel Dekker, New York. Fairen A. and DeFelipe J. and Regidor J. (1982) Nonpyramidal neurons. In Cerebral Cortex, Vol. 1, Cellular Components of Cerebral Cortex (eds Peters A. and Jones E. G.) , pp. 201–253. Plenum, New York. Falkai P. and Bogerts B. (1986) Cell loss in the hippocampus of schizophrenics. Eur. Arch. psychiat. neurol. Sci. 236, 154–161. Falkai P., Bogerts B. and Rozumek M. (1988) Cell loss and volume reduction in the entorhinal cortex of schizophrenics. Biol. Psychiat. 24, 515–521. Feighner J. P., Robins E. and Guze S. B. (1972) Diagnostic criteria for use in psychiatric research. Archs gen. Psychiat. 26, 57–63. Galaburda A. M. and Pandya D. N. (1983) The intrinsic architectonic and connectional organization of the superior temporal region of the rhesus monkey. J. comp. Neurol. 221, 169–184. Goldman-Rakic P. S., Leranth C., William S. M., Mons N. and Geffard M. (1989) Dopamine synaptic complex with pyramidal neurons in primate cerebral cortex. Proc. natn. Acad. Sci. U.S.A. 86, 9015–9019. Hanada S., Mita T., Nishinok N. and Tankaka C. (1987) 3H-Muscimol binding sites increased in autopsied brains of chronic schizophrenic. Life Sci. 40, 259–266. Jakob H. and Beckmann H. (1986) Prenatal developmental disturbances in the limbic allocortex in schizophrenics. J. neural Transm. 65, 303–326. Jeste D. and Lohr J. B. (1989) Hippocampal pathologic findings in schizophrenia. Archs gen. Psychiat. 46, 1019–1024. Jones E. (1984) Laminar distribution of cortical efferent cells. In Cerebral Cortex, Vol. 1, Cellular Components of the Cerebral Cortex (eds Peters A. and Jones E. G.), pp. 521–548. Plenum, New York. Jones E. G. and Hendry S. H. C. (1984) Basket cells. In Cerebral Cortex, Vol. 1, Cellular Components of Cerebral Cortex (eds Peters A. and Jones E. G.), pp. 309–336. Plenum, New York. Kovelman J. A. and Scheibel A. B. (1984) A neurohistological correlate of schizophrenia. Biol. Psychiat. 19, 1601–1621. Lewis D. A., Hayes T. L., Lund J. S. and Oeth K. M. (1992) Dopamine and the neural circuitry of primate prefrontal cortex: implications for schizophrenia research. Neuropsychopharmacology 425, 263–274. Lund J. S. (1984) Spiny stellate neurons. In Cerebral Cortex, Vol. 1, Cellular Components of the Cerebral Cortex (eds Peters A. and Jones E. G.), pp. 255–308. Plenum, New York. Marin-Padilla M. (1970) Prenatal and early postnatal ontogenesis of the human motor cortex: a Golgi study. I. The sequential development of the cortical layers. Brain Res. 23, 167–183. Marin-Padilla M. (1970) Prenatal and early postnatal ontogenesis of the human motor cortex: a Golgi study. II. The basket–pyramidal system. Brain Res. 23, 185–191. Martin K. A. C. (1984) Neuronal circuits in cat striate cortex. In Cerebral Cortex, Vol. 2, Functional Properties of Cortical Cells (eds Jones E. G. and Peters A.), pp. 241–284. Plenum, New York. Palacios J. M., Probst A. and Cortes R. (1986) Mapping receptors in human brain. Trends Neurosci. 9, 284–289. Peters A. (1984) Chandelier cells. In Cerebral Cortex, Vol. 1, Cellular Components of the Cerebral Cortex (eds Peters A. and Jones E. G.), pp. 361–380. Plenum, New York. Peters A. and Jones E. G. (1984) Classification of cortical neurons. In Cerebral Cortex, Vol. 1, Cellular Components of Cerebral Cortex (eds Peters A. and Jones E. G.), pp. 107–122. Plenum, New York. Rakic P. (1974) Neurons in rhesus monkey visual cortex. Systematic relation between time of origin and eventual disposition. Science 183, 425–427. Rakic P., Bourgeois J. P., Eckenhoff M. F., Zecevic N. and Goldman-Rakic P. S. (1986) Concurrent overproduction of synapses in diverse regions of the primate cerebral cortex. Science 232, 232–235. Roberts E. (1972) An hypothesis suggesting that there is a defect in the GABA system in schizophrenia. Neurosci. Res. Bull. 10, 469–482. Sidman R. and Rakic P. (1973) Neuronal migration with special reference to developing human brain. Brain Res. 62, 1–35. Simpson M. D., Slater P., Deakin J. F., Royston M. C. and Skan W. J. (1989) Reduced GABA uptake sites in the temporal lobe in schizophrenia. Neurosci. Lett. 107, 211–215. Somogyi P. and Cowey A. (1984) Double bouquet cells. In Cerebral Cortex, Vol. 1, Cellular Components of Cerebral Cortex (eds Peters A. and Jones E. G.), pp. 337–360. Plenum, New York. Tombol T. (1984) Layer VI cells. In Cerebral Cortex, Vol. 1, Cellular Components of Cerebral Cortex (eds Peters A. and Jones E. G.), pp. 479–519. Plenum, New York. Verney C., Alvarez C., Geffard M. and Berger B. (1990) Ultrastructural double-labelling study of dopamine terminals and GABA-containing neurons in rat anteromedial cerebral cortex. Eur. J. Neurosci. 2, 960–972.
Up-regulation of GABAA receptor binding in schizophrenia 44. 45. 46. 47. 48.
1031
Vincent S. L., Khan Y. and Benes F. M. (1993) Cellular distribution of dopamine D1 and D2 receptors in rat medial prefrontal cortex. J. Neurosci. 13, 2551–2564. Vincent S. L., Khan Y. and Benes F. M. (1995) Cellular colocalization of dopamine D1 and D2 receptors in rat medial prefrontal cortex. Synapse 19, 112–120. Weibel E. R. (1979) Stereological methods. In Practical Methods for Biological Morphometry, Vol. 1, pp. 63–100. Academic, New York. Whitehouse P. J., Lynch D. and Kuhar M. J. (1984) Effects of postmortem delay and temperature on neurotransmitter receptor binding in a rat model of the human autopsy process. J. Neurochem. 432, 553–559. Young W. S. and Kuhar M. J. (1979) A new method for receptor autoradiography: [3H]opioid receptors in rat brain. Brain Res. 179, 225–270. (Accepted 28 May 1996)
PII:
Neuroscience Vol. 75, No. 4, pp. 1021–1031, 1996 Copyright ? 1996 IBRO. Published by Elsevier Science Ltd Printed in Great Britain 0306–4522/96 $15.00+0.00 S0306-4522(96)00328-4
UP-REGULATION OF GABAA RECEPTOR BINDING ON NEURONS OF THE PREFRONTAL CORTEX IN SCHIZOPHRENIC SUBJECTS F. M. BENES,*†‡ S. L. VINCENT,* A. MARIE* and Y. KHAN* *Laboratory for Structural Neuroscience, McLean Hospital, 115 Mill Street, Belmont, MA 02178, U.S.A. †Department of Psychiatry and the Program in Neuroscience, Harvard Medical School, Boston, Massachusetts, U.S.A. Abstract––Recent investigations have reported a reduced density of interneurons and an increase of GABAA receptor binding occurring preferentially in layer II of the anterior cingulate cortex of schizophrenic subjects [Benes F. M. et al. (1992) J. Neurosci. 12, 924–929]. Since a reduction in the density of interneurons has also been found in layer II of the prefrontal cortex, this study has sought to determine whether an up-regulation of the GABAA receptor binding activity might also be found in this region of schizophrenics. A high-resolution autoradiographic analysis of bicuculline-sensitive [3H]muscimol (GABAA) receptor binding on individual neuron cell bodies in layers II, III, IV and VI has been applied to Brodmann area 10 from normal controls (n = 16) and schizophrenic (n = 7) subjects. A computerassisted technique has been used under strictly blind conditions to determine whether differences in binding occur in the schizophrenic group. A significant increase of GABAA receptor binding activity has been observed in layers II, III, V and VI in the schizophrenic group. When the binding is expressed as a density with respect to neuronal cell size, there is a gradient of binding across layers II, III, V and VI, with neuronal cell bodies in layer II having the greatest density of grains. When different subpopulations of neurons distinguished according to size criteria are examined separately, large (pyramidal) neurons show significantly higher binding, particularly in layer II, where it was increased by 90% in schizophrenics. Small (non-pyramidal) cells do not show significant differences in binding in schizophrenics, except in layer VI, where there was a 135% increase. Potential confounding effects from age and post mortem interval do not explain the differences between the two groups, because both young and old schizophrenics, as well as schizophrenics with long and short post mortem intervals, showed increased GABAA receptor binding activity when compared to control cases distinguished in a corresponding fashion. These data suggest that there may be a preferential reduction of inhibitory GABAergic inputs to pyramidal neurons, particularly in layer II of the prefrontal cortex, in schizophrenia. This change could potentially result in an increased excitatory outflow from the prefrontal area to other cortical regions of the schizophrenic brain. Overall, these results are consistent with the idea that reduced amounts of GABAergic activity in the prefrontal cortex could be related to a perinatal disturbance and could be a potentially important component of the pathophysiology of psychosis. Copyright ? 1996 IBRO. Published by Elsevier Science Ltd. Key words: Pyramidal cells, non-pyramidal neurons, layer II, Brodmann area 10.
In recent years, a renewed interest in the question of whether there are histopathological changes in the brains of schizophrenic patients has been generated by reports of volume loss,13,14 decreased numbers of neurons4,6,12,18,24 and other subtle cytoarchitectural variations3,5,7,23,27 in corticolimbic brain regions of individuals with schizophrenia. One study reported a reduction in the density of non-pyramidal neurons, particularly in layer II, of the anterior cingulate cortex and suggested that the missing cells might be GABAergic interneurons.6 To test the hypothesis that this latter change might result in a compensatory ‡To whom correspondence should be addressed. Abbreviations: ANOVA, analysis of variance; PMI, post mortem interval.
up-regulation of the postsynaptic receptor, a highresolution receptor binding technique for GABAA receptor binding activity48 was modified for human post mortem cortex.10 Using this approach, a marked increase of receptor binding was found in the neuropil of layer I and on neuronal cell bodies of layers II and III in the anterior cingulate cortex of schizophrenics;10 there was no increase of binding in the deeper laminae of this region. These results were interpreted as being consistent with a loss of inhibitory interneurons having occurred during the perinatal period, a time when basket cells in superficial laminae are actively differentiating.30,31 Although cortical areas differ considerably with regard to their anatomical and functional specializations, one study has suggested that diverse regions of
1021
1022
F. M. Benes et al. Table 1. Demographic analysis of post mortem cases
Brain no.
Diagnosis
Gender
Age (years)
Cause of death
PMI (h)
1317 1524 1542 1545 1547 1597 1603 1638 1696 1700 1705 1706
Control Control Control Control Control Control Control Control Control Control Control Control
Male Male Female Female Female Male Male Male Female Female Male Female
72 71 78 58 75 82 75 35 63 91 53 77
16.5 8 20 18 5 14 5.5 7.5 19 13 6 6
1733 1734 1737 1880
Control Control Control Control
Female Male Male Male
60 40 72 64
acute respiratory distress bronchopneumonia cardiac arrest bronchopneumonia coronary artery disease lung carcinoma lung carcinoma congestive heart failure thyroid carcinoma acute peritonitis lung carcinoma chronic obstructive pulmonary disease bronchopneumonia cardiac arrest coronary artery disease pancreatic carcinoma
18 48 24 15.9
1313 1318 1348 1350 1420 1576 1702
Schizophrenic Schizophrenic Schizophrenic Schizophrenic Schizophrenic Schizophrenic Schizophrenic
Male Female Female Male Male Male Male
34 37 74 36 35 67 62
bronchopneumonia suicide ischemic necrosis of colon coronary atherosclerosis suicide myocardial infarction carcinoma of large bowel
12 23 22.5 24 6.2 18 20.7
primate cortex develop in a relatively synchronous fashion.37 In this setting, a prenatal disturbance could potentially interfere with normal cell migration during development and induce preferential changes in superficial laminae of other cortical regions, such as the prefrontal cortex, where a reduced density of interneurons was also observed in layer II.6 To explore this, a high-resolution analysis of GABA receptor binding on cell bodies of two basic categories of cortical neurons, the so-called pyramidal and non-pyramidal neurons,35 has been performed on a cohort of post mortem normal and schizophrenic subjects. EXPERIMENTAL PROCEDURES
Post mortem specimens Post mortem specimens of prefrontal cortex (Brodmann area 10) from normal control (n = 16) and schizophrenic (n = 7) subjects were obtained through the Human Brain Tissue Resource Center at McLean Hospital (Table 1). All psychiatric cases for which a diagnosis of schizophrenia could be established were included in the study. The diagnosis of schizophrenia was made by performing a retrospective review of patient records and applying the criteria of Feighner et al.19 All brain specimens, including the prefrontal area, were evaluated by a neuropathologist to detect evidence of intracranial diseases, such as Alzheimer’s dementia, cerebrovascular disease or other more diffuse changes arising from alcohol abuse. Demographic data, including age (years), post mortem interval (PMI; h) and neuroleptic exposure (expressed as the chlorpromazineequivalent dose), were also collected. Tissue handling All brain specimens were removed at the time of autopsy and transported on ice prior to removal of tissue samples. Blocks of prefrontal cortex (Brodmann area 10) were placed
Neuroleptic exposure
0 — 400 — 200 0 100
in ice-cold 0.1% formaldehyde in 0.1 M phosphate buffer (pH 7.4) for 1.5 h prior to cryoprotection in ice-cold 30% sucrose in 0.1 M phosphate buffer (pH 7.4). The tissue blocks were left overnight in the sucrose solution, before being frozen on dry ice, imbedded in Tissue Tek OTC Compound (Miles, Elkhart, IN, U.S.A.) and stored at "70)C. Specimens from both the control and schizophrenic groups were stored for one to four years at "70)C. Only one control case was stored for a period of less than one year. The tissue blocks were sectioned (10 µm in thickness) on a cryostat at "20)C, thaw-mounted on gelatin-coated, acid-cleaned glass slides and maintained at "20)C until all blocks had been sectioned. Tissues from normal and schizophrenic cases were sectioned alternately to avoid a sequencing bias that might potentially be incurred during cutting and storage of the slide-mounted sections. All control and schizophrenic tissue sections were processed simultaneously to minimize experimental variance. Six sections per case (three for total binding and three for inhibited binding) were incubated in 0.31 M Tris–acetate (pH 7.4) at room temperature for 20 min to wash out endogenous ligand. The slides were then incubated for 40 min at room temperature in a solution containing 5 nM [3H]muscimol (specific activity 25.8 Ci/mmol; DuPont/New England Nuclear) in Tris–acetate buffer. Non-specific binding was assessed by incubating a parallel series of sections in tritiated ligand plus unlabelled 100 µM (+)-bicuculline methiodide. All of the sections were then rapidly washed in buffer, immersed in cold distilled water, rapidly dried with a stream of air and stored at 4)C. Using a modification10 of a previously described method,48 the slides were prepared for autoradiography by apposing acid-cleaned coverslips (no. 00, Corning Glass Works, Pittsburgh, PA, U.S.A.) coated with Kodak NTB-3 nuclear emulsion (diluted 1:1 with distilled water). To ensure that the grains representing binding activity were kept in proper registration with the tissue section, the coverslips were glued with cyanoacrylate to one end of the slide-mounted sections under darkroom conditions, secured with small binder clips and exposed for approximately six weeks. After exposure, the binder clips were removed under safelight conditions and the coverslips were gently bent
Up-regulation of GABAA receptor binding in schizophrenia
1023
Table 2. The average size of large and small neurons in the prefrontal cortex of normal control and schizophrenic subjects Large neurons Layer II III V VI
Small neurons
Control
Schizophrenic
Control
Schizophrenic
108.4 & 2.0 163.3 & 3.6 172.0 & 3.5 134.3 & 2.6
119.6 & 5.1 179.9 & 7.5 187.3 & 7.5 144.7 & 4.3
65.1 & 0.9 83.4 & 1.5 84.1 & 1.4 76.8 & 1.3
69.3 & 1.2 81.6 & 2.4 82.4 & 2.3 80.1 & 2.1
The data shown are the average size (µm2) & S.E.M. of large and small neurons in layers II, III, V and VI of normal controls and schizophrenic subjects. The values shown represent the average of the total number of cells sampled.
away from the tissue sections with a small Teflon spacer that was placed between the coverslips and slid about halfway down their length. The latent images were developed in a solution of Dektol for 4 min at 19)C, then washed in distilled water for 15 s, fixed with Kodak rapid fixer (without hardener) and rinsed with water for 30 min prior to staining with Thionin. Microscopic analyses All control and schizophrenic cases that were processed were included in the analyses. The autoradiographic preparations were visualized with a #100 oil immersion objective lens on a Leitz Laborlux microscope interfaced with a Bioquant MEG IV Image Analysis System (R and M Biometrics, Nashville, TN, U.S.A.) and a Wyse 386 PC computer via a Dage CCD71 video camera and a Summagraphics digitizing board. The final image magnification on the video monitor was #3800 and was of sufficiently high resolution to permit the visualization of both neuronal cell bodies and autoradiographic grains. The morphometric sampling technique employed was derived from the principles described by Weibel,46 with all procedures being conducted under strictly ‘‘blind’’ conditions using a technique described in detail elsewhere.10 Briefly, the slides were analysed by moving the field laterally through each layer, choosing the first 18 cells that could be identified as neurons, and tracing around their neuronal cell body to determine their size (µm2). This was followed by a semiautomated determination of the corresponding number of grains at a level of focus above the selected neuron.10 Non-pyramidal neurons were distinguished from glial cells using standard criteria: (i) the presence of euchromatin in neurons, but heterochromatin in glia; (ii) the presence of a nucleolus in neurons, but not glia; and (iii) the presence of Thionin-stained cytoplasm and dendritic processes in neurons, but not glia.4,6 To distinguish between pyramidal and non-pyramidal neurons, the size distributions for these two cell types were determined independently in tissue sections that were not subjected to the mechanical trauma arising from the apposition of emulsion-coated coverslips. A neuron was designated as ‘‘pyramidal’’ if it showed a characteristic triangular shape and an apical dendritic shaft, while multipolar neurons were considered to be ‘‘non-pyramidal’’ in nature. The average sizes of pyramidal neurons in layers II, III, V and VI were 96 & 12, 148 & 25, 158 & 31 and 117 & 18 µm2, respectively. For non-pyramidal neurons in these same laminae, the average sizes were 73 & 7, 90 & 7, 86 & 11 and 84 & 14 µm2, respectively. The size of pyramidal neurons in the Cresyl Violet-stained sections was approximately 14% larger than that for ‘‘large’’ neurons in the Thionin-stained autoradiograms, while that for nonpyramidal cells was 7% larger than ‘‘small’’ neurons. The greater difference for the pyramidal neurons can probably be explained by the somewhat better visualization of the proximal portions of apical dendritic shafts in the Cresyl Violet-stained sections when compared to the Thioninstained autoradiograms. This factor would have less of an
impact on the size of non-pyramidal neurons where the proximal dendritic branches are quite small relative to the overall size of the somata. Because the distribution profiles for both cell types were relatively narrow, one standard deviation was subtracted from the mean for both pyramidal and non-pyramidal neuron sizes in each layer, and the midway point between each was established. This latter number was used as a cut-off between ‘‘small’’ and ‘‘large’’ neurons in layers II (82 µm2), III (110 µm2), V (112 µm2) and VI (98 µm2) of the actual cell samples that had been collected for the autoradiographic analyses of GABA receptor binding activity. The average sizes of the resulting ‘‘large’’ and ‘‘small’’ neurons are shown in Table 2. Although it is likely that some small pyramidal neurons were included in the non-pyramidal cell population, the actual number was probably relatively low. The average diameter of each grain was approximately 0.3–0.5 µm. The computer-generated counts were manually corrected for underestimates arising from overlapping grains. This procedure was performed in each cortical layer of both the inhibited and uninhibited autoradiograms for each case. Neurons that were overlapping other cells were not included in the sample. Neuronal cell bodies in layer I were too sparse to be sampled, while those in layer IV were not analysed because of difficulties in identifying this lamina in the autoradiographic preparations. Statistical analyses Specific GABAA receptor binding was determined by subtracting the number of grains in sections with [3H]muscimol + bicuculline from those with [3H]muscimol alone for each normal and schizophrenic case. The individual case averages were used to generate a ‘‘mean of means’’ & S.E.M. for the respective groups. Student’s t-test was used to assess whether the differences between the groups were significant. The Bonferroni correction is required for multiple comparisons when selected comparisons, simultaneous examination of several different treatment contrasts or data dredging occurs.15 The cell types examined have been implicated in the pathophysiology of schizophrenia,6,8 and the six layers each have a unique pattern of connections, specific functional implications and precisely timed appearances during development. Thus, it was necessary to examine each cell type and layer separately and perform multiple comparisons in order to properly test the central hypothesis of this study that layer II plays a central role in schizophrenia. A two-way repeated measures analysis of variance (ANOVA) was used to evaluate whether the composite data for layers II, III, V and VI were different for the control vs schizophrenic groups. RESULTS
Visual inspection of the Thionin-stained autoradiogram assemblies indicated that there was both good grain development and adequate preservation
1024
F. M. Benes et al.
Fig. 1. Composite light photomicrographs of pyramidal neurons in the prefrontal cortex (Brodmann area 10) of a normal control (A) and a schizophrenic (B) subject, showing autoradiographic grains representing bicuculline-sensitive [3H]muscimol binding in layer II. The pyramidal cells were photographed in register with their corresponding grains. Fields of this type were analysed with a userinteractive automated image analysis system. Scale bar = 10 µm.
of morphological detail in both the normal and schizophrenic cases (Fig. 1A, B). Computer-assisted analyses of the autoradiograms revealed that the non-specific [3H]muscimol binding remaining after inhibition with bicuculline (range 2.5–5.0 grains/ neuron) was similar across the various layers of both the normal and schizophrenic groups. As shown in Fig. 2, the average size of neuronal cell bodies was similar for the normal and schizophrenic groups in each of the layers examined, but for both groups, those in layer II were considerably smaller than those in layers III, V and VI. Figure 3 indicates that the number of grains per individual neuronal cell body was significantly higher for the schizophrenic group, with the repeated measures ANOVA yielding F1,21 = 7.1 for the main effect of group (P = 0.015) in layers II (70%), III (44%), V (48%) and VI (66%). When the data were expressed as the density of receptor binding activity (number of grains/µm2 of cell size), the schizophrenic group showed a gradient across the layers, with layer II having the highest and layer VI having the lowest amount of binding activity (Fig. 4). Using the repeated measures ANOVA, the main effect of group had an F1,21 = 4.23 (P = 0.05). Only layers II and VI had differences that were statistically significant. As Table 2 shows, the sizes of
Fig. 2. A bar graph showing the size of neuronal cell bodies in layers II, III, V and VI of the prefrontal cortex (Brodmann area 10) from the control (solid bar) and schizophrenic (open bar) groups. The data shown are the average size & S.E.M. expressed in µm2 for the respective control (n = 16) and schizophrenic (n = 7) group. The mean size of neurons for each individual case was averaged for the respective groups.
Up-regulation of GABAA receptor binding in schizophrenia
Fig. 3. A bar graph showing specific GABAA receptor binding activity on neuronal cell bodies in layers II, III, V and VI of the prefrontal cortex (Brodmann area 10) from normal control (solid bar) and schizophrenic (open bar) subjects. The data shown were obtained by using the average number of grains per neuron for each subject in the respective control (n = 16) and schizophrenic (n = 7) groups to generate a mean and S.E.M. for each group. Using a repeated measures ANOVA, there is significantly more binding activity on neurons of the schizophrenic group (P = 0.015) in each of the four layers examined.
both large (pyramidal) and small (non-pyramidal) neurons were similar for both groups in layers II, III, V and VI. When the specific receptor binding activity was evaluated separately for the two populations of neuronal cell bodies (Fig. 5, upper and lower panels), the density of binding activity on large neurons was higher in layers II, III and V, with the main effect of group having a repeated measures F1,21 = 6.9 (P = 0.016). The difference in GABAA receptor binding was greatest in layer II, where large neurons showed a 90% increase relative to the normal controls, but the magnitude of the differences progressively decreased in layers III, V and VI. The binding activity for small cells was significantly increased (repeated measures F1,21 = 5.3, P = 0.03) only in layer VI, where it was 135% higher in the schizophrenic group. Table 3 indicates the average age and PMI for the normal and schizophrenic cases. The age of the control group (66.4 years) was considerably higher than that for the schizophrenic group (49.1 years), while the PMI was lower in the controls (12.6 h) than in the schizophrenics (18.2 h). In both cases, the differences between the two groups were not significant. To assess whether these two potentially confounding variables may have influenced the pattern of results described above, the normal and schizophrenic cases were pooled and both age and PMI were correlated with GABAA receptor binding. There was no correlation between either potential confound and the receptor binding activity found in any of the layers (Table 4). When the data were
1025
Fig. 4. A bar graph showing the density of specific GABAA receptor binding activity on neuronal cell bodies in layers II, III, V and VI of the prefrontal cortex (Brodmann area 10) from the normal control (solid bar) and schizophrenic (open bar) groups. The density data shown were obtained by dividing the number of grains per neuron by the size of each neuron. The mean number of grains per µm2 of neuronal cell body for each case was then averaged for control (n = 16) and schizophrenic (n = 7) groups. Using a repeated measures ANOVA, there is a significantly higher density of grains on neuronal cell bodies in layers II and VI of the schizophrenic group (P = 0.05). Overall, the grain density in the schizophrenics shows a gradient across the various laminae, with layer II showing the highest receptor binding activity and layer VI showing the lowest activity.
further broken down according to those above and below 45 years of age, young (mean = 37 & 4 years) and old (mean = 70 & 10 years) controls both showed binding activity on neuronal cell bodies that was much lower than that for young (mean = 36 & 1 years) and old (mean = 67 & 6 years) schizophrenics (Table 5). Similarly, when the data were distinguished according to short (<15 h) and long (>15 h) PMIs, the GABAA receptor binding on neuronal cell bodies of normal controls was considerably lower than that in corresponding subgroups of schizophrenic subjects (Table 6). To assess the potential effect of neuroleptic exposure on the increase of GABAA receptor binding, two schizophrenics with no neuroleptic exposure were evaluated separately from the rest of the group. The average binding in layer II for these two cases (7.2 & 1.5 grains/neuron) was lower than that for the subjects who had received neuroleptic medication (12.9 & 2.2 grains/neuron), and was only slightly higher than the average binding in this lamina for the control group (6.6 & 0.7 grains/neuron). However, in a previous study of the anterior cingulate cortex,8 these same two cases were observed to have GABAA receptor binding activity in layer II (46.6 grains/ neuron) that was much higher than that found for the schizophrenic group as a whole (38.0 grains/neuron). While it is conceivable that there are differential neuroleptic effects on prefrontal vs anterior cingulate
1026
F. M. Benes et al. Table 4. Correlation of GABAA receptor binding activity with age and post mortem interval Layer II III V VI
Age (years)
PMI (h)
0.082 0.047 0.027 0.046
0.10 0.19 0.08 0.10
The numbers shown are the correlation coefficient, r, obtained for the GABAA receptor binding activity (grains/ cell) in layer II of both the control and schizophrenic cases. Similar values were obtained when the receptor binding activity was expressed as a density (grains/µm2).
Fig. 5. A set of bar graphs comparing the total [3H]muscimol binding activity on large versus small neurons in layers II, III, V and VI of the prefrontal cortex from the control (solid bar) and schizophrenic (open bar) groups. The data are expressed as the number of autoradiographic grains per neuron & S.E.M. for the total number of large and small neurons in the various laminae of each group. The numbers above the horizontal brackets indicate the level of significance for the differences between the normal control and schizophrenic groups. There was a total of 480 large and 668 small neurons analysed in the various laminae of the control group, and 292 large and 212 small neurons in the various laminae of the schizophrenic group included in the data represented in graphs A–D. Using a repeated measures ANOVA, there is a significantly higher density of grains on large neurons of the schizophrenic group (P = 0.016) in layers II, III, V and VI. The density of autoradiographic grains is significantly higher on small neurons in layers II and VI, but not III and V. Table 3. Average age and post mortem interval for the control and schizophrenic groups
Controls Schizophrenics
n
Age (years)
PMI (h)
16 7
66.4 & 14.9 49.1 & 17.4
12.6 & 6.4 18.2 & 7.1
The age and PMI for the control and schizophrenic groups are expressed as mean & S.D.
cortices, it seems more likely that factors other than drug exposure could have influenced the GABAA receptor binding activity in the prefrontal area of the two schizophrenic subjects who received no neuroleptic medication (see Discussion section). DISCUSSION
The results of this study demonstrate that there is an increase in GABAA receptor binding activity on
individual neuronal cell bodies in the prefrontal cortex (Brodmann area 10) of schizophrenic subjects. In a pattern similar to that found previously in the anterior cingulate region,8 the density of autoradiographic grains showed a gradient of increase across layers II–VI of schizophrenics, with the greatest percentage increase in binding occurring on large neurons in layer II. In the anterior cingulate cortex,8 when the binding was expressed either as the number of grains per cell or as a density, an inside-out gradient in the overall binding across layers II–IV was observed. In the prefrontal cortex, where the size of neurons is much smaller in layer II than in layers III and V, a gradient of this type was found when the data were expressed as the percentage increase in the density (number of grains per unit area of cell size) for individual neurons, but not when expressed as the number of grains per large neuron. It was important to consider whether the differences in age or PMI between the controls and schizophrenic groups could possibly account for the increased GABAA receptor binding found in the patients. This concern was diminished when a breakdown of cases according to young vs old and short vs long PMIs showed a persistence of increased receptor binding activity in the schizophrenic subjects. This latter observation is consistent with reports from other laboratories that receptor binding activity remains remarkably stable for approximately 24 h after death.33,47 Moreover, the fact that there were differences in the degree to which GABAA receptor binding activity was increased on large vs small neurons, as well as on large neurons of superficial laminae vs those in deep layers, further argues against the possibility that differences in PMI are responsible for the findings reported here. A confounding effect of this type would be expected to non-specifically affect all neuronal types irrespective of the layer in which they are found. A further area of concern emerged when it was found that the two neuroleptic-naive cases did not show increased GABAA activity in the prefrontal area. This was puzzling since the same two cases showed a marked up-regulation in the anterior cingulate region,8 making it seem less likely that the findings reported here could be explained by a drug
Up-regulation of GABAA receptor binding in schizophrenia
1027
Table 5. Differences in GABAA receptor binding in the prefrontal cortex of young and old subjects Controls Age range (years)
Schizophrenics
<45
>45
<45
>45
Average age (years)
37 & 4 n=2
70 & 10 n = 14
36 & 1 n=4
67 & 6 n=3
Layer II III V VI
6.6 & 2.0 7.1 & 5.0 5.1 & 2.6 2.2 & 1.7
6.6 & 0.8 9.8 & 1.1 7.1 & 1.0 3.6 & 0.5
11.3 & 1.9 14.6 & 2.7 10.6 & 1.4 5.3 & 0.7
11.3 & 3.4 12.6 & 3.5 9.6 & 2.7 6.0 & 2.3
The data shown are the mean & S.E.M. grains/cell for GABAA receptor binding activity on pyramidal neurons in layers II, III, V and VI of young (<45 years) and old (>45 years) subjects. The average age & S.D. is shown for each age range of the respective control and schizophrenic groups. Table 6. Differences in GABAA receptor binding in the prefrontal cortex of subjects with short and long post mortem intervals Controls PMI range (h) Average PMI (h) Layer II III V VI
Schizophrenics
<15 6.9 & 2.3 n=8
>15 18.3 & 2.8 n=8
<15 8.6 & 3.4 n=2
>15 22.0 & 2.8 n=5
10.1 & 1.0 13.4 & 2.1 10.3 & 1.5 5.4 & 1.0
6.2 & 1.3 9.4 & 4.1 7.7 & 1.9 3.4 & 0.9
14.2 & 3.2 18.2 & 6.6 12.2 & 0.7 5.5 & 1.2
14.4 & 2.7 16.2 & 2.8 13.6 & 1.8 6.7 & 1.6
The data shown are the mean & S.E.M. grains/cell for GABAA receptor binding activity on pyramidal neurons in layers II, III, V and VI for subjects with short and long PMIs.
effect. It is noteworthy that one of the neuroleptictreated cases (B1350) also showed GABAA receptor binding in the prefrontal cortex that overlapped with the control group, but not in the anterior cingulate region, where it was increased. This observation suggests that there may be region-to-region differences in GABAA receptor binding activity that occur independently of neuroleptic exposure, and it may well be that such cases may eventually be distinguished according to subtle variations in the schizophrenic syndrome. Nevertheless, the possibility that neuroleptic effects might account for the differences in GABAA receptor binding in the prefrontal cortex of schizophrenics, though unlikely, cannot as yet be definitively excluded. Abnormalities of GABAergic neurotransmission have long been suspected as playing a role in the pathophysiology of schizophrenia.38 Earlier investigations have reported an increase of [3H]muscimol binding,22 decreased glutamate decarboxylase activity,11 decreased GABA uptake40 and a reduction in the number of neurons expressing mRNA for glutamate decarboxylase1 in the frontal area of the schizophrenic brain. While the data reported in these studies first suggested that a decrease of GABAergic activity might be present in the cortex of schizophrenics, small effects and large variances discouraged interest in these findings. With the high-resolution technique employed here, it was
possible to analyse binding on neuronal cell bodies where GABAA receptor binding activity is high10 and eliminate areas where it is entirely absent (nonneuronal compartments). Consequently, the emulsion coverslip technique has made it possible to obtain data with a large effect size and small variance that has intrinsically greater statistical power. An additional advantage of the high-resolution method employed here is the fact that receptor binding activity can also be differentially assessed on neuronal subpopulations, such as large and small cells. Although these latter two neuronal populations are probably not composed exclusively of one cell type, it seems likely that the majority of large neurons are pyramidal in nature, while the small neurons may be predominantly composed of non-pyramidal cells. While pyramidal neurons are the principle source of efferent flow from the cortex, non-pyramidal neurons modulate the activity of intrinsic circuits locally within each cortical layer.18,26 As a result, changes in GABAA receptor binding on one or the other cell type would likely result in different effects on the activity of intrinsic circuits within the prefrontal cortex. In the current study, an increased amount of GABAA receptor binding was found most consistently on pyramidal neurons in layers II, III and V, suggesting that these latter neurons would be more affected by a decrease of inhibitory activity in schizophrenic subjects (see
1028
F. M. Benes et al.
Fig. 6. A schematic diagram presenting a hypothetical model to account for the differences in specific GABAA receptor binding activity on neurons of the prefrontal cortex in schizophrenic subjects as compared to normal controls. (A) A low-power Nissl-stained section of human prefrontal cortex (Brodmann area 10), showing layers I–VI. Scale bar = 250 µm. (B) Normal circuit. In the middle panel, a normal circuit is shown with one pyramidal neuron present in layers II and V. Each pyramidal neuron receives an inhibitory input from two GABAergic interneurons. (C) Schizophrenic circuit. In the right-hand panel, a similar circuit in a schizophrenic subject also has one pyramidal neuron in layer II. In layer V, however, there are two pyramidal neurons, rather than one as in the normal circuit, as suggested by the increased density of these cells in this lamina of schizophrenics.6 In layer II, there is only one GABAergic interneuron providing an inhibitory input to the pyramidal neuron in layer II to account for the up-regulation of GABAA on these cells. In layer V, each pyramidal neuron also receives inhibitory input from only one GABAergic interneuron, rather than two, because there was no increase of non-pyramidal neurons noted previously in this layer.6 In the setting of increased pyramidal neurons, there would be a relative decrease of inhibitory activity from the normal number of GABA cells and this would give rise to compensatory up-regulation of GABAA receptor binding activity on pyramidal neurons of layer V in the schizophrenic group.
below) than similar cells in the deeper laminae. It is important to emphasize that the up-regulation of GABAA binding has also been found in the neuropil of layers I–III of the anterior cingulate region of schizophrenics.8 Although this compartment was not included in the current study, it is likely that this receptor binding activity was also increased in the neuropil of the prefrontal cortex. Accordingly, the up-regulation of the GABAA receptor in schizophrenics may also be associated with distal dendritic branches and axon terminations that are abundant components of the neuropil. Recent studies of the anterior cingulate cortex have emphasized the importance of alterations in layer II of schizophrenics. The changes noted in this region have included smaller clusters of neurons,3 a decreased density of interneurons,6 increased GABAA receptor binding activity8 and an increased
density of vertical axons5 that are probably glutamatergic afferents to the cingulate region.7 Layer II abnormalities have also been described in the entorhinal23 and prefrontal areas6 of schizophrenics. The fact that the cerebral cortex develops in an ‘‘inside-out’’ fashion, with layer II being the last cell-rich lamina to appear and differentiate,30,31,36,39 makes it plausible that a prenatal disturbance in cell migration could account for changes in the frontal area of schizophrenics.1 At the time of birth in humans, layer II basket cells are still quite immature,31 and theoretically could be more vulnerable to a pre- and/or perinatal injury than cells that are fully differentiated.2,8 Functional considerations An important observation reported in this study is that there appears to be a greater up-regulation of GABAA receptor binding on pyramidal neurons
Up-regulation of GABAA receptor binding in schizophrenia
than on small neurons, which are presumed to be primarily non-pyramidal in nature. This implies that the putative decrease of GABAergic inhibition in schizophrenics6,8 would give rise to an increased firing of these large projection neurons. This, in turn, could result in a greater outflow of activity from the prefrontal area to other cortical and subcortical sites with which it is connected, although it is possible that this increase of receptor binding activity on postsynaptic pyramidal neurons could compensate for a reduction of released GABA. In general, there was less change of GABAA receptor binding on non-pyramidal neurons in the prefrontal area of schizophrenics, except for layer VI interneurons, which did show a very striking increase. In this regard, it is important to point out that there are many different types of interneurons, each with unique laminar distributions and functional roles.16 For example, small and large basket cells are GABAergic inhibitory neurons distributed throughout layers II–VI;26 chandelier cells are also GABAergic cells, but ones found principally in upper cortical laminae;34 double bouquet cells are believed to be GABAergic inhibitory neurons that play a disinhibitory role;41 granule cells are excitatory cells found predominantly in layer IV;29 Martinotti cells are excitatory elements found primarily in layers V and VI.42 The heterogeneity of the nonpyramidal cell population16 and the inability of the autoradiographic technique employed in this study to distinguish the various subtypes noted above make it difficult to predict how the up-regulation of GABAA receptor binding activity on these latter cells of layer VI would impact on the overall output of these laminae. To understand the general outcome of a decreased inhibitory input to pyramidal neurons, it is pertinent to note that the termination sites of these cells in layers V and VI are in the thalamus, caudate nucleus and other subcortical loci, while those in layers II and III send an efferent outflow primarily to other cortical regions.20,25 Many pyramidal neurons in superficial cortical laminae also give off collateral branches to deeper laminae, particularly layer V.32 It is noteworthy
1029
that the density of pyramidal cells in layer V of the prefrontal cortex has been found to be significantly higher in schizophrenic subjects, while nonpyramidal cells in this layer showed no difference.6 With an increase of pyramidal neurons, the normal number of interneurons might provide inadequate inhibitory GABAergic input to each of these projection cells in layer V and result in increased efferent outflow (Fig. 6). Another pertinent question is whether the alterations of GABAergic activity in the prefrontal cortex of schizophrenics reported here may be related in some way to the mesocortical dopamine projections. Direct interactions between dopamine-containing axon varicosities and GABAergic interneurons have recently been proposed,21,28 and empirical evidence for this relationship has been reported,9,43 although such contacts are non-synaptic in nature43 and may only involve modulatory influences.9 The interaction of dopamine varicosities with GABAergic neurons could well be functional in nature, because dopamine receptor binding activity is also localized on non-pyramidal neurons of the prefrontal cortex.44,45 Thus, the ability of neuroleptic drugs to relieve the symptoms of schizophrenia could involve, at least in part, the blockade of dopamine receptors on an impaired population of inhibitory interneurons.2 Future studies will seek to determine whether other corticolimbic brain regions considered key regions of interest in schizophrenia research, such as the hippocampal formation and entorhinal region, also show increased GABAA receptor binding activity in subjects with this disorder. If so, there will be compelling evidence for the idea that alterations of GABAergic transmission may be an important component of the pathophysiology of schizophrenia.
Acknowledgements—This work has been supported by NIMH grants MH00423, MH42261 and MH31154, and an award from the Stanley Foundation. The authors wish to thank Drs Steven Matthysse and Alan M. Zaslavsky for their consultation concerning the statistical analyses.
REFERENCES
1. 2. 3. 4. 5. 6. 7.
Akbarian S., Bunney W. E., Potkin S. G., Wigal S. B., Hagman J. D., Sandman C. A. and Jones E. G. (1993) Altered distribution of nicotinamide-adenine dinucleotide phosphate-diaphorase cells in frontal lobe of schizophrenics implies disturbances of cortical development. Archs gen. Psychiat. 50, 227–230. Benes F. M. (1995) Is there a neuroanatomic basis for schizophrenia? An old question revisited. The Neuroscientist 1, 104–115. Benes F. M. and Bird E. D. (1987) An analysis of the arrangement of neurons in the cingulate cortex of schizophrenic patients. Archs gen. Psychiat. 44, 608–616. Benes F. M., Davidson J. and Bird E. D. (1986) Quantitative cytoarchitectural studies of cerebral cortex of schizophrenics. Archs gen. Psychiat. 43, 31–35. Benes F. M., Majocha R., Bird E. D. and Marrotta C. A. (1987) Increased vertical axon numbers in cingulate cortex of schizophrenics. Archs gen. Psychiat. 44, 1017–1021. Benes F. M., McSparren J., Bird E. D., Vincent S. L. and SanGiovanni J. P. (1991) Deficits in small interneurons in prefrontal and anterior cingulate cortex of schizophrenic and schizoaffective patients. Archs gen. Psychiat. 48, 996–1001. Benes F. M., Sorensen I., Vincent S. L., Bird E. D. and Sathi M. (1992) Increased density of glutamateimmunoreactive vertical processes in superficial laminae in cingulate cortex of schizophrenic brain. Cerebral Cortex 2, 502–512.
1030 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43.
F. M. Benes et al. Benes F. M., Vincent S. L., Alsterberg G., Bird E. D. and SanGiovanni J. P. (1992) Increased GABA-A receptor binding in superficial layers of cingulate cortex in schizophrenics. J. Neurosci. 12, 924–929. Benes F. M., Vincent S. L. and Molloy R. (1993) Dopamine-immunoreactive axon varicosities form nonrandom contacts with GABA-immunoreactive neurons in rat medial prefrontal cortex. Synapse 15, 285–295. Benes F. M., Vincent S. L. and SanGiovanni J. P. (1989) High resolution imaging of receptor binding in analysing neuropsychiatric diseases. Biotechniques 7, 970–979. Bird E. D., Spokes E. G. S. and Iversen L. L. (1979) Increased dopamine concentration in limbic areas of brain from patients dying with schizophrenia. Brain 102, 347–360. Bogerts B. and Falkai P. and Tutsch J. (1986) Cell numbers in the pallidum and hippocampus of schizophrenics. In Biological Psychiatry (eds Shagass C. et al.), Vol. 27, pp. 1178–1180. Elsevier, Amsterdam. Bogerts B., Meertz E. and Schonfieldt-Bausch R. (1985) Basal ganglia and limbic system pathology in schizophrenia: a morphometric study of brain volume and shrinkage. Archs gen. Psychiat. 42, 784–791. Brown R., Colter N., Corsellis J. A. N., Crow T. J., Frith C. D., Jagoe R., Johnstone E. C. and Marsh L. (1986) Postmortem evidence for structural brain changes in schizophrenia. Differences in brain weight, temporal horn area and parahippocampal gyrus width as compared with affective disorder. Archs gen. Psychiat. 43, 36–42. Dunnett C. and Goldsmith C. (1981) When and how to do multiple comparisons. In Statistics in the Pharmaceutical Industry (eds Buncher R. G. and Tsya I.-Y.), pp. 397–435. Marcel Dekker, New York. Fairen A. and DeFelipe J. and Regidor J. (1982) Nonpyramidal neurons. In Cerebral Cortex, Vol. 1, Cellular Components of Cerebral Cortex (eds Peters A. and Jones E. G.) , pp. 201–253. Plenum, New York. Falkai P. and Bogerts B. (1986) Cell loss in the hippocampus of schizophrenics. Eur. Arch. psychiat. neurol. Sci. 236, 154–161. Falkai P., Bogerts B. and Rozumek M. (1988) Cell loss and volume reduction in the entorhinal cortex of schizophrenics. Biol. Psychiat. 24, 515–521. Feighner J. P., Robins E. and Guze S. B. (1972) Diagnostic criteria for use in psychiatric research. Archs gen. Psychiat. 26, 57–63. Galaburda A. M. and Pandya D. N. (1983) The intrinsic architectonic and connectional organization of the superior temporal region of the rhesus monkey. J. comp. Neurol. 221, 169–184. Goldman-Rakic P. S., Leranth C., William S. M., Mons N. and Geffard M. (1989) Dopamine synaptic complex with pyramidal neurons in primate cerebral cortex. Proc. natn. Acad. Sci. U.S.A. 86, 9015–9019. Hanada S., Mita T., Nishinok N. and Tankaka C. (1987) 3H-Muscimol binding sites increased in autopsied brains of chronic schizophrenic. Life Sci. 40, 259–266. Jakob H. and Beckmann H. (1986) Prenatal developmental disturbances in the limbic allocortex in schizophrenics. J. neural Transm. 65, 303–326. Jeste D. and Lohr J. B. (1989) Hippocampal pathologic findings in schizophrenia. Archs gen. Psychiat. 46, 1019–1024. Jones E. (1984) Laminar distribution of cortical efferent cells. In Cerebral Cortex, Vol. 1, Cellular Components of the Cerebral Cortex (eds Peters A. and Jones E. G.), pp. 521–548. Plenum, New York. Jones E. G. and Hendry S. H. C. (1984) Basket cells. In Cerebral Cortex, Vol. 1, Cellular Components of Cerebral Cortex (eds Peters A. and Jones E. G.), pp. 309–336. Plenum, New York. Kovelman J. A. and Scheibel A. B. (1984) A neurohistological correlate of schizophrenia. Biol. Psychiat. 19, 1601–1621. Lewis D. A., Hayes T. L., Lund J. S. and Oeth K. M. (1992) Dopamine and the neural circuitry of primate prefrontal cortex: implications for schizophrenia research. Neuropsychopharmacology 425, 263–274. Lund J. S. (1984) Spiny stellate neurons. In Cerebral Cortex, Vol. 1, Cellular Components of the Cerebral Cortex (eds Peters A. and Jones E. G.), pp. 255–308. Plenum, New York. Marin-Padilla M. (1970) Prenatal and early postnatal ontogenesis of the human motor cortex: a Golgi study. I. The sequential development of the cortical layers. Brain Res. 23, 167–183. Marin-Padilla M. (1970) Prenatal and early postnatal ontogenesis of the human motor cortex: a Golgi study. II. The basket–pyramidal system. Brain Res. 23, 185–191. Martin K. A. C. (1984) Neuronal circuits in cat striate cortex. In Cerebral Cortex, Vol. 2, Functional Properties of Cortical Cells (eds Jones E. G. and Peters A.), pp. 241–284. Plenum, New York. Palacios J. M., Probst A. and Cortes R. (1986) Mapping receptors in human brain. Trends Neurosci. 9, 284–289. Peters A. (1984) Chandelier cells. In Cerebral Cortex, Vol. 1, Cellular Components of the Cerebral Cortex (eds Peters A. and Jones E. G.), pp. 361–380. Plenum, New York. Peters A. and Jones E. G. (1984) Classification of cortical neurons. In Cerebral Cortex, Vol. 1, Cellular Components of Cerebral Cortex (eds Peters A. and Jones E. G.), pp. 107–122. Plenum, New York. Rakic P. (1974) Neurons in rhesus monkey visual cortex. Systematic relation between time of origin and eventual disposition. Science 183, 425–427. Rakic P., Bourgeois J. P., Eckenhoff M. F., Zecevic N. and Goldman-Rakic P. S. (1986) Concurrent overproduction of synapses in diverse regions of the primate cerebral cortex. Science 232, 232–235. Roberts E. (1972) An hypothesis suggesting that there is a defect in the GABA system in schizophrenia. Neurosci. Res. Bull. 10, 469–482. Sidman R. and Rakic P. (1973) Neuronal migration with special reference to developing human brain. Brain Res. 62, 1–35. Simpson M. D., Slater P., Deakin J. F., Royston M. C. and Skan W. J. (1989) Reduced GABA uptake sites in the temporal lobe in schizophrenia. Neurosci. Lett. 107, 211–215. Somogyi P. and Cowey A. (1984) Double bouquet cells. In Cerebral Cortex, Vol. 1, Cellular Components of Cerebral Cortex (eds Peters A. and Jones E. G.), pp. 337–360. Plenum, New York. Tombol T. (1984) Layer VI cells. In Cerebral Cortex, Vol. 1, Cellular Components of Cerebral Cortex (eds Peters A. and Jones E. G.), pp. 479–519. Plenum, New York. Verney C., Alvarez C., Geffard M. and Berger B. (1990) Ultrastructural double-labelling study of dopamine terminals and GABA-containing neurons in rat anteromedial cerebral cortex. Eur. J. Neurosci. 2, 960–972.
Up-regulation of GABAA receptor binding in schizophrenia 44. 45. 46. 47. 48.
1031
Vincent S. L., Khan Y. and Benes F. M. (1993) Cellular distribution of dopamine D1 and D2 receptors in rat medial prefrontal cortex. J. Neurosci. 13, 2551–2564. Vincent S. L., Khan Y. and Benes F. M. (1995) Cellular colocalization of dopamine D1 and D2 receptors in rat medial prefrontal cortex. Synapse 19, 112–120. Weibel E. R. (1979) Stereological methods. In Practical Methods for Biological Morphometry, Vol. 1, pp. 63–100. Academic, New York. Whitehouse P. J., Lynch D. and Kuhar M. J. (1984) Effects of postmortem delay and temperature on neurotransmitter receptor binding in a rat model of the human autopsy process. J. Neurochem. 432, 553–559. Young W. S. and Kuhar M. J. (1979) A new method for receptor autoradiography: [3H]opioid receptors in rat brain. Brain Res. 179, 225–270. (Accepted 28 May 1996)