Characterization of neocortical and hippocampal synaptosomes from temporal lobe epilepsy patients

Brain Research 837 Ž1999. 55–66 www.elsevier.comrlocaterbres

Research report

Characterization of neocortical and hippocampal synaptosomes from temporal lobe epilepsy patients Govert Hoogland a , Marion Blomenrohr ¨ a , Hilde Dijstelbloem a , Marina de Wit a , a Henk A. Spierenburg , Cees W.M. van Veelen b , Peter C. van Rijen b , Alexander C. van Huffelen c , Willem Hendrik Gispen a , Pierre N.E. de Graan a, ) a

c

Rudolf Magnus Institute for Neurosciences, Utrecht UniÕersity, UniÕersiteitsweg 100, 3584 CG Utrecht, Netherlands b Department of Neurosurgery, UniÕersity Hospital Utrecht, Heidelberglaan 100, 3584 CX Utrecht, Netherlands Department of Clinical Neurophysiology, UniÕersity Hospital Utrecht, Heidelberglaan 100, 3584 CX Utrecht, Netherlands Accepted 25 May 1999

Abstract To investigate epilepsy-associated changes in the presynaptic terminal, we isolated and characterized synaptosomes from biopsies resected during surgical treatment of drug-resistant temporal lobe epilepsy ŽTLE. patients. Our main findings are: Ž1. The yield of synaptosomal protein from biopsies of epilepsy patients was about 25% of that from rat brain. Synaptosomal preparations were essentially free of glial contaminations. Ž2. Synaptosomes from TLE patients and naive rat brain, quickly responded to Kq-depolarization with a 70% increase in intrasynaptosomal Ca2q ŽwCa2q x i ., and a 40% increase in B-50rGAP-43 phosphorylation. Ž3. Neocortical and hippocampal synaptosomes from TLE patients contained 20–50% of the glutamate and g-aminobutyric acid ŽGABA. contents of rat cortical synaptosomes. Ž4. Although the absolute amount of glutamate and GABA released under basal conditions from neocortical synaptosomes of TLE patients was lower than from rat synaptosomes, basal release expressed as percentage of total content was higher Ž16.4% and 17.3%, respectively. than in rat Ž11.5% and 9.9%, respectively.. Ž5. Depolarization-induced glutamate and GABA release from neocortical synaptosomes from TLE patients was smaller than from rat synaptosomes Ž3.9% and 13.0% vs. 21.9% and 25.0%, respectively.. Ž6. Analysis of breakdown of glial fibrillary acid protein ŽGFAP. indicates that resection time Žanoxic period during the operation. is a critical parameter for the quality of the synaptosomes. We conclude that highly pure and viable synaptosomes can be isolated even from highly sclerotic human epileptic tissue. Our data show that in studies on human synaptosomes it is of critical importance to distinguish methodological Ži.e., resection time. from pathology-related abnormalities. q 1999 Elsevier Science B.V. All rights reserved. Keywords: Temporal lobe epilepsy; Human brain synaptosomes; B-50; Protein phosphorylation; Calcium; Neurotransmitter release

1. Introduction Temporal lobe epilepsy ŽTLE. is a type of focal epilepsy in which the hippocampus often plays a central role in the generation of epileptic activity. Until the last two decades, the pathophysiology of human TLE has been investigated mainly by immunocytochemical and histological techniques w1x. Based on the pathological substrates found in these studies, two types of TLE patients can be distinguished. In the first type, the hippocampus is characterized by a typical pattern of neuronal loss and gliosis, referred to ) Corresponding author. Rudolf Magnus Institute for Neurosciences, P.O. Box 80040, 3508 TA Utrecht, The Netherlands. Fax: q31-30-2539032; E-mail: [email protected]

as mesial temporal sclerosis ŽMTS. w6,52x. The hippocampus of these MTS-associated TLE patients not only shows loss of synapses due to neuronal degeneration, but also reactive synaptogenesis as a result of sprouting of dentate granule cell mossy fibers w2x. In the second type of TLE patients, the pathological substrate is a lesion Žtumor or arteriovenous malformation.. In these so-called tumor-associated TLE patients, a neoplastic lesion Žsuch as a glioma or ganglioglioma. can be accompanied by neuronal loss and gliosis as well as perilesional neurochemical changes w31,51x. The functional implications of these anatomical changes in epileptic tissue are thus far not well understood. It has been proposed that the increased seizure susceptibility characterizing epilepsy may be due to instabilities in neu-

0006-8993r99r$ - see front matter q 1999 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 1 - 4 5 3 4 Ž 9 9 . 0 0 3 3 1 - 7

56

G. Hoogland et al.r Brain Research 837 (1999) 55–66

ronal networks, caused by excessive excitatory transmission andror impaired inhibitory transmission mediated by the amino acid transmitters glutamate and g-aminobutyric acid ŽGABA., respectively w4,19x. In fact, seizure-associated elevation of extracellular glutamate is thought to initiate excitotoxicity, that is a cascade of reactions involving NMDA receptor activation, massive calcium influx and eventually neuronal death w17,36x. Although probably all cells are excited during seizures, cell loss is selective and has been associated with a diminished calcium buffering capacity of calcium-binding proteins w28,29,34x. Thus far neurochemical studies on glutamatergic and GABAergic transmission in epileptic human brain tissue have focused primarily on analysis of total glutamate and GABA levels in brain homogenates w41,44,45x, brain slices w25,27x and after microdialysis w7,13,17,24,43,50x. From these studies, no clear pattern of changes in transmitter levels has emerged. This is probably due to the complexity of the amino acid transmitter systems. For instance, extracellular glutamate levels are regulated not only by presynaptic release, but also by neurotransmitter uptake into neurons and particularly glia cells. As the degree and extent of gliosis varies between different TLE patients, the contribution of glial cells in these studies may have varied likewise w52x. Moreover, since glutamate is a metabolic amino acid, large amounts of glutamate may be released from degenerating neurons, which cannot easily be distinguished from presynaptic exocytotic release w40x. To study changes in glutamate and GABA release in epileptic human tissue, it is therefore essential to study transmitter release from nerve terminals. Pinched-off and resealed nerve terminals Žsynaptosomes. can be obtained by applying shear forces to brain tissue followed by a subcellular fractionation to separate the synaptosomes from other subcellular compartments w20x. Of the various isolation techniques that have been described w3,23,38x, the procedure of Dunkley et al. w14–16x is particularly effective to obtain highly purified and viable synaptosomes. Viable synaptosomes are capable of generating ATP, essential to maintain a membrane potential and a low intrasynaptosomal calcium concentration, and show Ca2q-dependent neurotransmitter release upon depolarization w49x. The depolarization-induced increase in wCa2q x i is known to activate protein kinase C ŽPKC. and induce a transient phosphorylation of its substrate B-50 Ža.k.a. GAP-43. w12x. In the present study, we isolated synaptosomes from en bloc resected neocortical and hippocampal biopsies from TLE patients and neurochemically characterized them to assess their suitability to investigate epilepsy-associated changes in glutamatergic and GABAergic transmission. We determined yield, purity Žglial fibrillary acidic protein ŽGFAP. and B-50 content. and viability Ždepolarization-induced increases in wCa2q x i , B-50 phosphorylation and glutamaterGABA release.. To discriminate between epilepsy and non-epilepsy related phenomena, we compared

synaptosomes from human epileptic hippocampus with synaptosomes from neocortex of the same patient and from naive rat brain. Effects of anoxia during surgical resection were assessed by measuring GFAP proteolysis.

2. Materials and methods 2.1. Materials Fura-2-acetoxymethyl ester Žfura 2-AM., bovine serum albumin ŽBSA; fraction V., ionomycin, homoserine, phorbol 12-myristate 13-acetate ŽPMA. and ethylene glycolbis- Ž-aminoethyl ether .-N, N, N X , N X -tetra-acetic acid ŽEGTA. were purchased from Sigma ŽSt. Louis, MO, USA.. Percoll was purchased from Pharmacia ŽUppsala, Sweden., dithiothreitol ŽDTT. from Fluka ŽBuchs, Switzerland. and pansorbin from Calbiochem ŽLa Jolla, CA, USA.. Carrier-free 32 Pi Žspecific activity 40 mCirml. was purchased from Amersham Life Science ŽBuckinghamshire, UK.. All other chemicals were obtained from Merck ŽDarmstadt, Germany.. Monoclonal anti-rat B-50 antibody NM2 was generously donated by Dr. A. van der Voorde, Innogenetics, Gent, Belgium w37x. The polyclonal rabbit antiserum 9527, raised against recombinant rat hexa-His B-50, was made in collaboration with Dr. A.B. Oestreicher and Dr. L.H. Schrama in the Rudolf Magnus Institute for Neurosciences. Polyclonal rabbit anti-cow GFAP was purchased from Dako ŽGlostrup, Denmark. and w125 Ix-labeled donkey antirabbit and w125 Ix-labeled sheep anti-mouse Žspecific activity 0.1 mCirml. from Amersham Life Science ŽBuckinghamshire, UK.. 2.2. Human biopsies Neocortical and hippocampal biopsies were collected en bloc in the operating room, in the course of a selective temporal lobe respective procedure in patients suffering from medically intractable TLE w11x. Informed consent was given prior to surgical treatment and clinical criteria determined the extent of the surgical resection only; the tissue used in these experiments would otherwise have been discarded. In this study, we only included biopsies of patients who, based on the histopathological evaluation, were diagnosed with MTS- or tumor-associated TLE. One non-epileptic frontal cortex biopsy was obtained from a 65-year old patient who was operated for a metastatic tumor Žmamma carcinoma.. This biopsy was resected within the same time frame as neocortical biopsies from TLE patients. Biopsies were immediately cooled in ice-cold buffer A Ž120 mM NaCl, 5 mM KCl, 20 mM pipes, 1 mM CaCl 2 , 1 mM MgCl 2 , 25 mM DŽq.glucose; pH 7.0; equilibrated with 100% O 2 . w30x, and cut with a razor-blade into 0.5 cm sections that were used alternately for the isolation of

G. Hoogland et al.r Brain Research 837 (1999) 55–66

synaptosomes and histopathological examination. Sections used for the isolation of synaptosomes were directly placed in fresh ice-cold buffer A, and transported from the operating room to the laboratory within 15 min. Tissue wet weight was determined and a 10% Žwet weight to volume. homogenate was made in a Teflon to glass Elvejhem Potter tube in 0.32 M sucrose Žcontaining 1 mM EDTA and 0.25 mM DTT; pH 7.4. at 48C. The homogenates were used for the isolation of synaptosomes and for immunoblotting. To study the effect of the duration of the surgery Žhypoxia., the resection time was determined for each biopsy. The start of the neocortical resection was defined as the first coagulation in the neocortex, and the start of the hippocampal resection was defined as the first preparative coagulation in the removal of the amygdala-hippocampal complex. The resection time was then defined as the time between resection start and the placement of the biopsy in ice-cold buffer A. 2.3. Isolation of synaptosomes Synaptosomes were highly purified by Percoll-sucrose density gradient centrifugation w16x of homogenates from biopsy sections, rat cortex or rat hippocampus Žmale Wistar rats, 100–120 g.. Synaptosomes were collected from the 15–23% Percoll gradient interface and pelleted Ž15 000 = g, 15 min. in buffer B Ž124 mM NaCl, 5 mM KCl, 1.3 mM MgSO4 , 20 mM CaCl 2 , 26 mM NaHCO 3 , 10 mM DŽq.glucose and 1 mgrml BSA; pH 7.4; equilibrated with 95% O 2r5% CO 2 .. Synaptosomes were washed in BSAfree buffer B as described above and resuspended in BSA-free buffer B. Protein was determined according to the method of Bradford w5x using BSA as the standard. 2.4. B-50 phosphorylation Synaptosomes were resuspended at 3 mgrml in buffer C Žbuffer B containing 2 mM CaCl 2 . and incubated with 32 Pi Ž1 mCirmg. under 95% O 2r5% CO 2 flow for 50 min at 308C to label the endogenous ATP-pool w12x. Excess 32 Pi was removed by addition of 10 vol. buffer C containing 1.2 mM NaH 2 PO4 and subsequent pelleting Ž6000 = g, 3 min, room temperature.. Synaptosomes were resuspended at 2 mgrml in buffer C and incubated Ž20 mg protein in a final volume of 20 ml. for 4 min at 308C with either 50 mM 4-aminopyridine Ž4-AP. or 1 mM PMA, 30 mM KCl Žbuffer C with equimolar replacement of 30 mM Naq by Kq added at t s 3.45 min. or buffer C Žcontrol.. Incubations were stopped with 10 ml ice-cold denaturation buffer Žfinal concentration: 62.5 mM Tris–HCl ŽpH 6.8., 2% SDS Žwrv., 10% glycerol Žvrv., 5% b-mercaptoethanol Žvrv., 0.001% bromophenol blue. and stored on ice. 32 P-incorporation in B-50 was determined by quantitative immunoprecipitation as described previously w10x us-

57

ing the anti-B-50 antibody 9527 in a 1:175 dilution. Immunoprecipitated proteins were separated by 11% SDSpolyacrylamide gel electrophoresis ŽSDS-PAGE. and 32 Pincorporation was measured by phosphor imaging on a Fuji BAS 1000 ŽRaytest, Straubenhard, Germany. and quantified using TINA analysis software. Data were expressed as percentage of control. Total protein phosphorylation was assessed in 6.7 mg synaptosomal protein separated on 11% SDS-PAGE followed by phosphor imaging. 2.5. [Ca 2 q ]i measurements wCa2q x i was determined using fura 2-AM as described previously w47x, with slight modifications. Briefly, synaptosomal preparations were adjusted to 1 mg proteinrml in buffer B in the presence of 5 mM fura 2-AM and incubated at 308C for 30 min, followed by a dilution to 0.15 mg proteinrml in buffer B and another 10 min incubation. After fura 2-AM loading, synaptosomes were pelleted Ž6000 = g, 3 min, room temperature., resuspended to 0.1 mg proteinrml in buffer B containing 2 mM CaCl 2 and transferred to a stirred and thermostatted cuvette in a Perkin-Elmer LS-50B luminescence spectrometer. Fluorescence emission was measured at 510 nm at alternating excitation wavelengths of 340 and 380 nm. Calibration was performed by addition of 10 mM ionomycin to determine maximal fluorescence and by addition of 10 mM EGTAr100 mM Tris to determine minimal fluorescence w22x. Ratios of the emission from these two excitation wavelengths were used to calculate the basal and 30 mM KCl depolarized wCa2q x i . 2.6. Glutamate and GABA release Synaptosomal preparations were adjusted to 1–2 mg proteinrml in buffer B containing 6 mM homoserine and incubated Ž10–20 mg protein in a final volume of 60 ml. for 5 min at 258C under basal Žbuffer B containing 2 mM CaCl 2 . or depolarizing condition Žbuffer B with equimolar replacement of 30 mM Naq by Kq and containing 2 mM CaCl 2 .. Total synaptosomal glutamate and GABA contents were measured by lysing of separate synaptosomal samples in 0.05% Triton X-100 Žvrv. w40x. Incubations were stopped by centrifugation for 25 s at 10 000 = g and the supernatants were stored in 5% trichloroacetic acid Žvrv, final concentration. at y808C until further analysis. Amino acid analysis was performed by reversed-phase HPLC as described earlier w48x. Prior to analysis, samples were spun Ž5 min, 15 000 = g, room temperature. and the supernatants were neutralized ŽpH 7.5. with 10 M NaOH. Glutamate and GABA release were expressed as percentage of the total content or as nanomoles per milligram of synaptosomal protein, after correction for the internal standard. 2.7. B-50 and GFAP Western blotting Brain homogenate and synaptosomal proteins Ž20 mg for B-50 and 3 mg for GFAP immunoblotting. were dis-

G. Hoogland et al.r Brain Research 837 (1999) 55–66

58

solved in denaturation buffer, and then separated by 11% SDS-PAGE, transferred to nitrocellulose membranes ŽSchleicher and Schuell. for 1 h at 400 mA, and blocked in TBS-T Ž50 mM Tris–HCl ŽpH 7.5., 150 mM NaCl, 0.1% Tween-20 Žvrv.. containing 2.5% skim milk. Membranes were incubated with NM2 Ž1:1000 in TBS-T and 1% goat serum Žvrv., 1 h. or GFAP antibody Ž1:500 in TBS-T and 1% goat serum Žvrv., 1 h. as primary antibody and w125 Ix-labelled secondary antibody Žanti-mouse 1:500 or anti-rabbit 1:1000, both diluted in TBS-T and 1% goat serum Žvrv., 1 h.. B-50 and GFAP immunoreactivity ŽIR. was visualized using a Fuji BAS 1000 phosphor imager and quantified using TINA analysis software. Data were expressed as percentage of homogenate. GFAP degradation was defined as IR proteins below the 49 kDa GFAP band and expressed as percentage of total GFAP IR, i.e., degradation products plus the 49 kDa IR band. 2.8. Statistics Statistical analysis was performed by using the Student’s t-test for paired or unpaired samples as indicated. Data were presented as means " S.E.M., n being the number of independent experiments.

3. Results 3.1. Yield and purity of synaptosomes Table 1 shows the synaptosomal yields, expressed as micrograms of synaptosomal protein per gram of biopsy wet weight, from cortical and hippocampal biopsies. Control experiments in rat showed no significant difference in synaptosome yield from cortex Ž950 " 210 mgrg. and hippocampus Ž1390 " 400 mgrg.. The yield from one cortical biopsy of a non-epileptic control patient was similar Ž1320 mgrg.. The yields from neocortical and hippocampal biopsies of TLE patients were approximately four times lower than from naive rat and human control tissue. Comparing within patient groups, only the MTS-associated TLE patients

Table 2 Clinical characteristics of MTS- and tumor-associated TLE patients

n Age at onset Žyears. a Age at surgery Žyears. a Gender Žmalerfemale. Resection side Žleftrright. Seizuresrmontha Neuropathological diagnosisb

MTS

Tumor

14 14"3 37"2 8r6 7r7

9 16"3 28"3 7r2 6r3

8"1 Wyler 1 Ž ns 3. Wyler 3 Ž ns 2. Wyler 4 Ž ns8. Wyler? Ž ns1.

10"2 Glioma Ž ns 2. Ganglioglioma Ž ns 3. Oligodendroglioma Ž ns 2. Astrocytoma Ž ns1. DNET Ž ns1.

a

Data are means"S.E.M. of n patients. The degree of MTS was graded according to histopathological criteria as proposed by Wyler et al. w52x. In one patient, it was not possible to determine a Wyler grade due to lack of representative hippocampal sections ŽWyler?.. This patient was diagnosed as MTS-associated TLE based on radiological evaluation. Dysembryoplastic neuroepithelial tumor ŽDNET..

b

showed a significantly lower yield from hippocampus Ž p - 0.05 by paired Student’s t-test. than from the neocortex. Due to the operation procedure, the wet weight of neocortical biopsies was significantly higher than of hippocampal biopsies. However, rat control experiments showed that within the experimental range, the biopsy wet weight did not affect the yield. Table 2 shows key clinical characteristics of the patient groups described in Table 1. Synaptosomal purity was assessed by Western blot analysis of the neuronal marker B-50 and the glial marker GFAP. Fig. 1 shows a typical example of a Western blot loaded with homogenate and synaptosomes from a neocortical biopsy of a patient with severe MTS. B-50 immunostaining revealed a single IR band migrating at 51 kDa ŽFig. 1 left panel. w37x. GFAP staining, however, revealed multiple IR bands, the highest migrating at 49 kDa ŽFig. 1 right panel.. GFAP IR bands, which migrated below this 49 kDa band were considered to be GFAP degradation products w8x. Such breakdown products were not found in homogenates from acutely dissected rat brain. Total GFAP IR in synaptosomal fractions from neocortical and hip-

Table 1 Yield of rat and human synaptosomes from cortical and hippocampal tissue Species

Pathology

Tissue

n

Wet weight Žg. a

Synaptosomal protein Žmg. a

Yield Žmgrg. a

Rat Rat Human Human Human Human Human

control control tumor-controlb MTS-TLE MTS-TLE tumor-TLE tumor-TLE

cortex hippocampus cortex neocortex hippocampus neocortex hippocampus

11 11 1 14 14 9 9

0.40 " 0.05 0.20 " 0.02 0.98 3.01 " 0.38 0.90 " 0.10 1.79 " 0.36 0.73 " 0.20

380 " 86 279 " 80 1320 948 " 141 230 " 44 741 " 321 127 " 22

950 " 210 1390 " 400 1320 320 " 30 c 250 " 30 303 " 85 257 " 72

a

Data are means " S.E.M. of n experiments. The yield was calculated as micrograms of synaptosomal protein per gram of biopsy wet weight. Human cortical biopsy obtained from a non-epileptic tumor patient. c p - 0.05 comparing synaptosomal yield from neocortical and hippocampal biopsies of MTS-associated TLE patients by paired Student’s t-test. b

G. Hoogland et al.r Brain Research 837 (1999) 55–66

59

Fig. 1. Typical example of an immunoblot of brain homogenate and synaptosomes stained for B-50 Žleft panel. and GFAP Žright panel.. Homogenate ŽHom. and synaptosomes ŽSyn. were made from a neocortical biopsy of an epilepsy patient with severe MTS. For B-50 immunoblots 20 mg and for GFAP immunoblots 3 mg protein was separated by 11% SDS-PAGE, transferred to nitrocellulose and immunostained with monoclonal anti-rat B-50 NM2 or polyclonal anti-cow GFAP as primary antibody and w125 Ix-labelled secondary antibody. B-50 and GFAP immunostaining were visualized by using phosphor imaging. MW: position of molecular weight standards.

pocampal biopsies was - 10% compared to the homogenate ŽFig. 2.. B-50 IR in these fraction was about 40% compared to the homogenate ŽFig. 2..

To investigate a possible relationship between resection time and GFAP degradation, we plotted GFAP degradation, i.e., IR homogenate proteins below the 49 kDa band

Fig. 2. Relative amount of neuronal marker B-50 and glial marker GFAP in brain homogenates and synaptosomes from TLE patients. Synaptosomes were purified from homogenates of neocortical ŽNC; n s 12 for B-50 and n s 10 for GFAP. and hippocampal ŽH; n s 3 for B-50 and n s 4 for GFAP. biopsies. Proteins were separated by 11% SDS-PAGE, immunoblotted for B-50 or GFAP and quantified. GFAP was quantified as total GFAP, i.e., the 49 kDa immunostained band plus degradation products. Data are expressed as percentage of IR staining in the homogenate.

60

G. Hoogland et al.r Brain Research 837 (1999) 55–66

35 min. showed a significant correlation between GFAP degradation and resection time Ž p - 0.001, R 2 s 0.73., while no correlation was found in hippocampal biopsies Žresected within 48 to 81 min..

3.2. B-50 phosphorylation

Fig. 3. Correlation between resection time and degradation of GFAP in homogenates of neocortical Ž Ø : ns11. and hippocampal Žl: ns 4. biopsies from TLE patients. Resection time was defined as the time between the first coagulation and the placement of the biopsy in ice-cold buffer A. GFAP degradation was defined as immunostained proteins below the 49 kDa immunostained protein band and expressed as percentage of total GFAP immunostaining Ždegradation products plus the 49 kDa band..

expressed as percentage of total GFAP IR vs. resection time ŽFig. 3.. Neocortical biopsies Žresected within 10 to

To assess synaptosomal viability we measured depolarization-induced changes in B-50 phosphorylation, in wCa2q x i , and in the release of glutamate and GABA. Kq-induced, PKC-mediated B-50 phosphorylation is a sensitive measure for synaptosome viability w12x. Analysis of 32 P-incorporation into total synaptosomal proteins revealed several phosphobands in human synaptosomes ŽFig. 4, left panel., a pattern similar to rat synaptosomes, but with differences in labelling intensity of individual bands. Phosphorylation of these bands in human and rat synaptosomes was not affected by phorbol ester or Kq treatment. After B-50 immunoprecipitation and 11% SDS-PAGE a single phosphoband was detected in human and rat synaptosomes, with an apparent MW of 51 and 48 kDa, respectively ŽFig. 4, right panel.. Treatment of synaptosomes from human neocortex with 30 mM Kq, 50 mM 4-AP or 1 mM phorbol ester resulted in an increase in B-50 phosphorylation of 41%, 40% and 133% of control levels, respec-

Fig. 4. Phosphor image of 32 P incorporated in total synaptosomal protein Žleft panel. and B-50 after immunoprecipitation Žright panel.. Rat cortex synaptosomes and neocortical synaptosomes from epilepsy patients were labelled with 32 Pi and subsequently incubated in control buffer ŽC. or buffer containing 30 mM Kq ŽKq. . Total synaptosomal protein and immunoprecipitated B-50 were resolved by 11% SDS-PAGE and visualized by phosphor imaging.

G. Hoogland et al.r Brain Research 837 (1999) 55–66

61

Fig. 5. Depolarization and phorbol ester-induced B-50 phosphorylation in rat and TLE patient synaptosomes. Rat cortex synaptosomes and neocortical synaptosomes from epilepsy patients were labelled with 32 Pi and subsequently incubated with control buffer Ž4 min., 30 mM Kq Ž15 s., 50 mM 4-AP Ž4 min. or 1 mM PMA Ž4 min.. Immunoprecipitated B-50 was resolved by 11% SDS-PAGE, and 32 P-incorporation into B-50 was quantified by phosphor imaging. Data are expressed as percentage of control Žmean " S.E.M.; n s 7–26 observations from five independent experiments., UU significantly different from control, p - 0.01 by unpaired Student’s t-test.

tively ŽFig. 5.. Similar values were found for rat cortical synaptosomes Ž37%, 35% and 164%, respectively.. 3.3. [Ca 2 q ]i measurements

synaptosomes from TLE patients contained ; 50% less glutamate and ; 75% less GABA than synaptosomes from rat cortex. Hippocampal synaptosomes from TLE patients contained ; 75% less glutamate and ; 80% less GABA than synaptosomes from rat hippocampus.

The basal and depolarization-induced wCa2q x i is a sensitive measure for metabolic stress and deterioration of synaptosomes. Viable rat synaptosomes typically show a basal wCa2q x i of about 200 nM, which is elevated by 70% upon Kq-depolarization w49x. Fig. 6 shows the average neocortical and hippocampal wCa2q x i-traces from fura-2 loaded synaptosomes of three MTS-associated TLE patients. Hippocampal synaptosomes showed an average basal wCa2q x i level of 213 " 27 nM which increased to 363 " 58 nM after Kq-induced depolarization. In neocortical synaptosomes basal and depolarized wCa2q x i levels were almost twice as high, 387 " 27 and 664 " 83 nM, respectively. Expressed as percentage of basal levels, the Kq-induced elevation in wCa2q x i was identical for both types of synaptosomes Ž72 " 6% and 70 " 11%, respectively., and does not differ from rat synaptosomes w49x. 3.4. Total synaptosomal glutamate and GABA content The total synaptosomal glutamate and GABA content, including vesicular, cytosolic and mitochondrial pools, was measured after synaptosomal lysis ŽTable 3.. Neocortical

Fig. 6. Average wCa2q x i-traces of fura-2 loaded neocortical and hippocampal synaptosomes from the same three patients suffering from MTS-associated TLE. Basal neocortical and hippocampal wCa2q x i Žexpressed in nM. was monitored for 5 min prior to addition of 30 mM Kq Žarrow..

G. Hoogland et al.r Brain Research 837 (1999) 55–66

62

Table 3 Total glutamate and GABA content of rat and TLE patient synaptosomesa

Glutamate GABA

Rat cortex

TLE patient neocortex

Rat hippocampus

TLE patient hippocampus

47.2"7.6 18.8"2.9

22.9"2.2 4.6"0.4

47.2"6.8 16.6"2.1

10.9"1.9 3.1"0.6

a

Data are means"S.E.M. of n experiments; rat Ž ns11., TLE patients Ž ns14.. Total glutamate and GABA content of lysed synaptosomes is expressed as nanomoles of glutamaterGABA per milligram of synaptosomal protein.

3.5. Glutamate and GABA release The HPLC chromatograms of amino acids released from human synaptosomes were similar to those from rat

synaptosomes ŽFig. 7.. Table 4 presents the glutamate and GABA release from rat and human synaptosomes expressed in nanomoles per milligram and as percentage of total synaptosomal content. Expressed as nanomoles per milligram basal release from neocortical synaptosomes of TLE patients Žglutamate 3.9 and GABA 0.8, respectively. was lower than from synaptosomes from rat cortex Žglutamate 5.9 and GABA 1.4, respectively.. However, expressed as percentage of the total synaptosomal glutamate or GABA content, basal release from human neocortical synaptosomes Ž16.4% and 17.3%, respectively. was elevated compared to basal release from synaptosomes from rat cortex Ž11.5% and 9.9%, respectively.. Depolarization induced by 30 mM Kq significantly increased basal glutamate and GABA release from rat and

Fig. 7. Examples of reversed-phase HPLC chromatograms, representing amino acid analysis of the supernatants from rat ŽA, B. and TLE-patient ŽC, D. synaptosomes under basal ŽA, C. and depolarized ŽB, D. conditions. Amino acids that were analyzed were: aspartate Ž1., glutamate Ž2., serine Ž3., glutamine Ž4., homoserine Ž5., glycine Ž6., taurine Ž7. and GABA Ž8.. Compared to the analysis of supernatants from rat synaptosomes, the HPLC attenuator was set two times more sensitive when analyzing supernatants from TLE-patient synaptosomes.

G. Hoogland et al.r Brain Research 837 (1999) 55–66 Table 4 Glutamate and GABA release from rat cortical and TLE patient neocortical synaptosomes Condition

Rat a %

TLE patient a nmolrmg %

nmolrmg

Glutamate basal 11.5"1.3 5.9"1.6 16.4"1.4 3.9"0.7 30 mM Kq 33.4"1.6 16.3"3.9 20.3"1.2 4.8"0.7 GABA basal 9.9"0.9 1.4"0.3 17.3"1.7 0.8"0.1 30 mM Kq 34.9"3.2 4.9"0.9 30.3"1.7 1.5"0.2 a

Data are means"S.E.M. of n observations; rat Ž ns9 out of three individual experiments., TLE patients Ž ns9.. Released glutamate and GABA are expressed as percentage of total synaptosomal glutamaterGABA content Ž%. and as nanomoles of glutamaterGABA per milligram of synaptosomal protein.

human synaptosomes Ž p - 0.001 by paired Student’s ttest.. Expressed as percentage of total content, Kq induced similar amounts of glutamate and GABA release from synaptosomes from rat cortex Ž33.4% and 34.9%, respectively., whereas human neocortical synaptosomes released 20.3% and 30.3%, respectively.

4. Discussion The present study was undertaken to evaluate the yield, purity and viability of synaptosomes isolated from neocortical and hippocampal biopsies of patients operated for medically intractable TLE. Only limited information is available on the feasibility of preparing viable synaptosomes from biopsies of epilepsy patients w21,32,39,42x. In none of the previously published studies, the synaptosomal viability was analyzed by measuring depolarization-induced changes. 4.1. Yield and purity of synaptosomes Using the purification procedure as described by Dunkley et al. w16x, we isolated 250–320 mg synaptosomal protein per gram wet weight from biopsies of epilepsy patients ŽTable 1., which is about 25% of the yield obtained from rat brain Žcortex or hippocampus. and a cortical biopsy from a control patient. Thus, our data indicate that the severe reduction in synaptosomal yield from epileptic tissue is due to the pathological state, rather than to a species difference or the operation procedure. Indeed, neuropathological evaluation of the biopsies showed severe loss of neurons and nerve terminals in many hippocampal subareas of MTS-associated TLE patients, and to a lesser degree also in tumor-associated TLE patients ŽProper et al., submitted., confirming previously published observations w6,31,52x. Also in extra-hippocampal temporal lobe structures of tumor- and MTS-associated TLE patients various types of anatomical and neurochemical abnormalities have been reported, but the extent of neuronal cell loss is much less dramatic and not well documented

63

w6,33,35,51x. The low yield of synaptosomes from epileptic neocortical tissue may also be caused by an increased vulnerability of ‘‘epileptic’’ terminals to the conditions during the isolation procedure. In agreement with an earlier study w39x, we found lower synaptosomal yields from hippocampal than from neocortical biopsies of MTS-associated TLE patients. Since the biopsies used in this study exhibited various degrees of gliosis, it was important to assess the purity of synaptosomes. Synaptosomes isolated from neocortical and hippocampal biopsies of TLE patients contained less than 10% glial contaminants ŽFig. 2.. The reduction in neuronal marker B-50rGAP-43 in synaptosomes compared to homogenates is not surprising, since B-50rGAP-43 is not only localized in terminals, but also in axons w46x. The total synaptosomal glutamate and GABA content in epileptic hippocampus and neocortex is considerably lower than in rat, but amino acid analysis showed that synaptosomes from rat cortex and biopsies of TLE patients released the same spectrum of amino acids and that both types of synaptosomes release glutamate and GABA upon Kq-induced depolarization ŽFig. 7.. The glutamate and GABA content of neocortical TLE synaptosomes found here are similar to the levels found in another type of synaptosomal preparation from TLE neocortex w42x. In summary, our data show that, although the yield may vary, relatively pure synaptosomes can be isolated from biopsies of different types of TLE patients.

4.2. Synaptosome Õiability Three properties of human TLE synaptosomes were analyzed to assess their viability: basal- and depolarization-induced wCa2q x i levels, protein phosphorylation and glutamaterGABA release. Synaptosomes isolated from TLE biopsies maintain similar low basal wCa2q x i as those from rat cortex or hippocampus w49x, and the depolarization-induced rise in wCa2q x i Žq70%. is indistinguishable from their rat counterparts w49x. To understand the higher basal and stimulated wCa2q x i levels in neocortical synaptosomes from TLE patients, further experiments are required. Rat and human TLE synaptosomes incorporate similar amounts of 32 P-labelled phosphate into total phosphoproteins. Moreover, the phorbol ester-, Kq- and 4-AP-induced increases in B-50rGAP-43 phosphorylation are identical in both preparations and are in accordance with previously published data in rat synaptosomes w12,26,49x. These experiments show that TLE synaptosomes have an active energy metabolism and normal PKC activity. Due to the reduction in total glutamate and GABA content of neocortical synaptosomes of TLE patients, it is not surprising that the absolute amount of glutamate and GABA released under basal conditions was lower than

64

G. Hoogland et al.r Brain Research 837 (1999) 55–66

from rat synaptosomes. Interestingly, basal glutamate and GABA release expressed as percentage of total content was higher Ž16.4% and 17.3%, respectively. than in rat Ž11.5% and 9.9%, respectively.. Possibly this increased x found in basal release is due to the elevated basal wCa2q i these synaptosomes. TLE synaptosomes showed a significant Kq-induced release of glutamate and GABA, but it was smaller than in rat synaptosomes. Thus, our data show that synaptosomes from tumor- as well as MTS-associated neocortex biopsies are capable of amino acid release in a Kq-dependent manner. Together our data show that the human TLE synaptosomes are metabolically active and highly responsive to Kq-depolarization. Nevertheless, we found several differences between rat and human TLE synaptosomes, for instance in amino acid transmitter content and release. In interpreting these differences it is essential to know to what degree these can be attributed to the operation procedure Žanoxia. or to real epilepsy related pathology. 4.3. Hypoxia-associated changes An important factor which may affect human TLE synaptosome properties is the time required for biopsy excision during surgery. During this resection time hypoxia may result in time-dependent proteolysis w18x, for instance of GFAP w9x. Indeed, we observed a positive correlation between the relatively short resection times Ž- 35 min. of neocortical biopsies and GFAP degradation, but not for the longer, hippocampal resection times Ž48–81 min, Fig. 3.. These results closely correspond to the 48-h time course of GFAP degradation in human brain homogenate showing a rapidly initial breakdown phase and a much slower second phase in which more proteolysis resistant GFAP fragments are formed w8,9x. Resection time may also affect transmitter release from biopsy synaptosomes. Increased neurotransmitter release, eventually leading to depletion of the releasable transmitter pool, has been described as one of the intermediate Ž1–10 min. hypoxiaassociated events w18x. Indeed, we found synaptosomes from TLE patients to contain less glutamate and GABA, than rat cortical synaptosomes ŽTable 3.. Thus, our data indicate that at least part of the differences between TLE and rat synaptosomes may be caused by anoxia during biopsy resection. 4.4. Possible pathology-associated changes Several of the observed differences between TLE and rat synaptosomes are most likely due to the pathological state of the tissue. Assuming that anoxia-induced transmitter release is confined to the first 10 min w18x, the low total glutamate and GABA content of hippocampal synaptosomes compared to neocortical synaptosomes must be pathology related. These data confirm earlier reports on focal amino acid levels in whole brain tissue w45x and synaptosomes w32x. Previously, these reduced levels have been attributed to changes in energy metabolism due to

hyperactivity associated with a higher local energy demand w39x. Combined with the high frequency of neurotransmitter release in epileptic tissue, this may lead to depletion of neuronal glutamate and GABA. Possibly also the reduced glutamate and GABA content of neocortical synaptosomes Žcompared to rat. is in part pathology related, but here it is difficult to distinguish between anoxia and pathology related changes. Expressed as percentage of total transmitter content neocortical synaptosomes from TLE patients exhibited a higher basal release than rat cortical synaptosomes ŽTable 4.. Interestingly, also the basal wCa2q x i in these neocortical synaptosomes from TLE patients, was increased compared with rat synaptosomes. These data, together with perilesional neurochemical changes found in these patients w51x, indicate that also the neocortex is affected in this type of epilepsy patients. Based on the Kq-induced increase in wCa2q x i , which is comparable between synaptosomes from TLE patients and rat, we expected Kq-evoked glutamate and GABA release from these two types of synaptosomes to be similar. However, while Kq-evoked GABA release from neocortical synaptosomes of TLE patient and rat synaptosomes was indeed similar, Kq-evoked glutamate release from TLE patient synaptosomes was decreased compared to rat. In order to determine to what degree changes in glutamate and GABA release are pathology related we are presently studying basal and Kq-evoked glutamate and GABA release from hippocampal and neocortical synaptosomes from patients with tumor- and MTS-associated TLE. In these studies, we determine calcium-dependent and independent components of amino acid release to discriminate between vesicular and non-vesicular release ŽHoogland et al., in preparation..

Acknowledgements We thank Dr. L.C. Meiners for radiological evaluation of the patients and Drs. G.H. Jansen for neuropathological evaluation of the biopsies. GH was supported by grant A91 of the Epilepsy Fund of the Netherlands.

References w1x T.L. Babb, W.J. Brown, Pathological findings in epilepsy, in: J.J. Engel ŽEd.., Surgical Treatment of the Epilepsies, Raven Press, New York, 1987, pp. 511–540. w2x T.L. Babb, W.R. Kupfer, J.K. Pretorius, P.H. Crandall, M.F. Levesque, Synaptic reorganization by mossy fibers in human epileptic fascia dentata, Neuroscience 42 Ž1991. 351–363. w3x R.F.G. Booth, J.B. Clark, A rapid method for the preparation of relatively pure, metabolically competent synaptosomes from rat brain, Biochem. J. 176 Ž1978. 365–370. w4x H.F. Bradford, Glutamate, GABA and epilepsy, Prog. Neurobiol. 47 Ž1995. 477–511.

G. Hoogland et al.r Brain Research 837 (1999) 55–66 w5x M.M. Bradford, A rapid and sensitive method for the quantification of microgram quantities of protein utilizing the principle of proteindye binding, Anal. Biochem. 72 Ž1976. 248–254. w6x C.J. Bruton, The Neuropathology of Temporal Lobe Epilepsy, Oxford Univ. Press, New York, 1988. w7x H. Carlson, E. Ronne Engstrom, U. Ungerstedt, L. Hillered, Seizure related elevations of extracellular amino acids in human focal epilepsy, Neurosci. Lett. 140 Ž1992. 30–32. w8x D. Dahl, A. Bignami, Glial fibrillary acidic protein from normal and gliosed human brain. Demonstration of multiple related polypeptides, Biochem. Biophys. Acta 386 Ž1975. 41–51. w9x D. Dahl, Glial fibrillary acidic protein from bovine and rat brain. Degradation in tissues and homogenates, Biochim. Biophys. Acta 420 Ž1976. 142–154. w10x P.N.E. de Graan, L.V. Dekker, A.B. Oestreicher, L. van der Voorn, W.H. Gispen, Determination of changes in the phosphorylation state of the neuron-specific protein kinase C substrate B-50 ŽGAP-43. by quantitative immunoprecipitation, J. Neurochem. 52 Ž1989. 17–23. w11x R.M. Debets, C.W. van Veelen, A.C. van Huffelen, W. van Emde Boas, Presurgical evaluation of patients with intractable partial epilepsy: the Dutch epilepsy surgery program, Acta Neurol. Belg. 91 Ž1991. 125–140. w12x L.V. Dekker, P.N.E. de Graan, M. de Wit, J.J.H. Hens, W.H. Gispen, Depolarization-induced phosphorylation of protein kinase C substrate B-50 ŽGAP-43. in rat cortical synaptosomes, J. Neurochem. 54 Ž1990. 1645–1652. w13x K.Q. Do, J.M. Klancnik, B. Gahwiler, H. Perschak, H.G. Wieser, M. Cuenod, Release of excitatory amino acids: animal studies and epileptic foci studies in humans, in: B.S. Meldrum, F. Moroni, R.P. Simon, J.H. Woods ŽEds.., Excitatory Amino Acids, Raven Press, New York, NY, 1991, pp. 677–685. w14x P.R. Dunkley, P.E. Jarvie, J.W. Heath, G.J. Kidd, J.A. Rostas, A rapid method for isolation of synaptosomes on Percoll gradients, Brain Res. 372 Ž1986. 115–129. w15x P.R. Dunkley, J.A.P. Rostas, J.W. Heath, D.A. Powis, The preparation and use of synaptosomes for studying secretion of catecholamines, in: Poisner, Trifara ŽEds.., In Vitro Methods for Studying Secretion, Elsevier, 1987, pp. 315–334. w16x P.R. Dunkley, J.W. Heath, S.M. Harrison, P.E. Jarvie, P.J. Glenfield, J.A. Rostas, A rapid Percoll gradient procedure for isolation of synaptosomes directly from an S1 fraction: homogeneity and morphology of subcellular fractions, Brain Res. 441 Ž1988. 59–71. w17x M.J. During, D.D. Spencer, Extracellular hippocampal glutamate and spontaneous seizure in the conscious human brain, Lancet 341 Ž1993. 1607–1610. w18x A.A. Farooqui, S.E. Haun, L.A. Horrocks, Ischemia and hypoxia, in: G.J. Siegel, B.W. Agranoff, R.W. Albers, P.B. Molinoff ŽEds.., Basic Neurochemistry, Raven Press, New York, 1994, pp. 867–884. w19x M. Glass, M. Dragunow, Neurochemical and morphological changes associated with human epilepsy, Brain Res. Brain Res. Rev. 21 Ž1995. 29–41. w20x E.G. Gray, V.P. Whittaker, The isolation of nerve endings from brain. An electron microscopic study of cell fragments divided by homogenization and centrifugation, J. Anat. ŽLondon. 96 Ž1962. 79–88. w21x T. Grisar, A.V. Delgado Escueta, Astroglial contribution in human temporal lobe epilepsy: Kq activation of Naq,Kq-ATPase in bulk isolated glial cells and synaptosomes, Brain Res. 364 Ž1986. 1–11. w22x G. Grynkiewicz, M. Poenie, R.Y. Tsien, A new generation of Ca2q indicators with greatly improved fluorescence properties, J. Biol. Chem. 260 Ž1985. 3440–3450. w23x F. Hajos, An improved method for the preparation of synaptosomal fractions in high purity, Brain Res. 93 Ž1975. 485–489. w24x A. Hamberger, B. Nystrom, H. Silfvenius, Extra- and intracellular amino acids in the CNS of patients with epilepsy and other neurological disorders, Epilepsy Res., Suppl. 8 Ž1992. 375–381. w25x T.S. Haugstad, E. Hegstad, I.A. Langmoen, Calcium dependent

w26x

w27x

w28x

w29x

w30x

w31x

w32x

w33x

w34x

w35x

w36x w37x

w38x

w39x

w40x w41x w42x

w43x

w44x

w45x

65

release of g-aminobutyric acid ŽGABA. from human cerebral cortex, Neurosci. Lett. 141 Ž1992. 61–64. F.M. Heemskerk, L.H. Schrama, P.N.E. de Graan, W.H. Gispen, 4-Aminopyridine stimulates B-50 ŽGAP-43. phosphorylation in rat synaptosomes, J. Mol. Neurosci. 2 Ž1990. 11–17. E. Hegstad, T.S. Haugstad, E. Hauglie Hanssen, I.A. Langmoen, Calcium-dependent release of glutamate from human cerebral cortex, Brain Res. 585 Ž1992. 340–342. C.W. Heizmann, K. Braun, Changes in Ca2q-binding proteins in human neurodegenerative disorders, Trends Neurosci. 15 Ž1992. 259–264. W. Kamphuis, E. Huisman, W.J. Wadman, C.W. Heizmann, F.H. Lopes da Silva, Kindling induced changes in parvalbumin immunoreactivity in rat hippocampus and its relation to long-term decrease in GABA-immunoreactivity, Brain Res. 479 Ž1989. 23–34. A.R. Kay, R.K. Wong, Isolation of neurons suitable for patch-clamping from adult mammalian central nervous systems, J. Neurosci. Methods 16 Ž1986. 227–238. J.H. Kim, M.Y. Guimarais, M.Y. Shen, L.M. Masukawa, D.D. Spencer, Hippocampal neuronal density in temporal lobe epilepsy with and without gliomas, Acta Neuropathol. 80 Ž1990. 41–45. P.L. Lleu, D. Labiner, M. Weinand, R.J. Huxtable, Neuroactive amino acids in synaptosomes from focal and nonfocal temporal lobe tissue of patients with intractable complex partial seizures, Adv. Exp. Med. Biol. 359 Ž1994. 435–443. P. Marco, R.G. Sola, P. Pulido, M.T. Alijarde, A. Sanchez, C.-S. Ramon, J. DeFelipe, Inhibitory neurons in the human epileptogenic temporal neocortex. An immunocytochemical study, Brain 119 Ž1996. 1327–1347. M.P. Mattson, B. Rychlik, C. Chu, S. Christakos, Evidence for calcium-reducing and excito-protective roles for the calcium-binding protein calbindin-D28k in cultured hippocampal neurons, Neuron 6 Ž1991. 41–51. L.C. Meiners, A. van Gils, G.H. Jansen, G. de Kort, T.D. Witkamp, L.M.P. Ramos, J. Valk, R.M.C. Debets, A.C. van Huffelen, C.W.M. van Veelen, W.P.T.M. Mali, Temporal lobe epilepsy: the various MR appearances of histologically proven mesial temporal sclerosis, Am. J. Neuroradiol. 15 Ž1994. 1547–1555. B.S. Meldrum, Excitotoxicity and selective neuronal loss in epilepsy, Brain Pathol. 3 Ž1993. 405–412. M. Mercken, U. Lubke, M. Vandermeeren, J. Gheuens, A.B. Oestreicher, Immunocytochemical detection of the growth-associated protein B-50 by newly characterized monoclonal antibodies in human brain and muscle, J. Neurobiol. 23 Ž1992. 309–321. A. Nagy, A.V. Delgado Escueta, Rapid preparation of synaptosomes from mammalian brain using nontoxic isoosmotic gradient material ŽPercoll., J. Neurochem. 43 Ž1984. 1114–1123. A.K. Nagy, C.R. Houser, A.V. Delgado Escueta, Synaptosomal ATPase activities in temporal cortex and hippocampal formation of humans with focal epilepsy, Brain Res. 529 Ž1990. 192–201. D.G. Nicholls, T.S. Sihra, Synaptosomes possess an exocytotic pool of glutamate, Nature 321 Ž1986. 772–773. T.L. Perry, S. Hansen, Amino acid abnormalities in epileptogenic foci, Neurology 31 Ž1981. 872–876. O.A. Petroff, L.A. Pleban, D.D. Spencer, Symbiosis between in vivo and in vitro NMR spectroscopy: the creatine, N-acetylaspartate, glutamate, and GABA content of the epileptic human brain, Magn. Reson. Imaging 13 Ž1995. 1197–1211. E. Ronne Engstrom, L. Hillered, R. Flink, B. Spannare, U. Ungerstedt, H. Carlson, Intracerebral microdialysis of extracellular amino acids in the human epileptic focus, J. Cereb. Blood Flow Metab. 12 Ž1992. 873–876. A. Sherwin, Y. Robitaille, F. Quesney, A. Olivier, J. Villemure, R. Leblanc, W. Feindel, E. Andermann, J. Gotman, F. Andermann et al., Excitatory amino acids are elevated in human epileptic cerebral cortex, Neurology 38 Ž1988. 920–923. N.M. van Gelder, A.L. Sherwin, T. Rasmussen, Amino acid content

66

G. Hoogland et al.r Brain Research 837 (1999) 55–66

of epileptogenic human brain: focal versus surrounding regions, Brain Res. 40 Ž1972. 385–393. w46x M. van Lookeren Campagne, C.G. Dotti, E.R. Jap Tjoen San, A.J. Verkleij, W.H. Gispen, A.B. Oestreicher, B-50rGAP-43 localization in polarized hippocampal neurons in vitro: an ultrastructural quantitative study, Neuroscience 50 Ž1992. 35–52. w47x M. Verhage, E. Besselsen, F.H. Lopes da Silva, W.E. Ghijsen, Evaluation of the Ca2q concentration in purified nerve terminals: relationship between Ca2q homeostasis and synaptosomal preparation, J. Neurochem. 51 Ž1988. 1667–1674. w48x M. Verhage, E. Besselsen, F.H. Lopes da Silva, W.E. Ghijsen, Ca2q-dependent regulation of presynaptic stimulus-secretion coupling, J. Neurochem. 53 Ž1989. 1188–1194.

w49x M. Verhage, W.E. Ghijsen, F.H. Lopes da Silva, Presynaptic plasticity: the regulation of Ca2q-dependent transmitter release, Prog. Neurobiol. 42 Ž1994. 539–574. w50x C.L. Wilson, N.T. Maidment, M.H. Shomer, E.J. Behnke, L. Ackerson, I. Fried, J. Engel, Comparison of seizure related amino acid release in human epileptic hippocampus versus a chronic, kainate rat model of hippocampal epilepsy, Epilepsy Res. 26 Ž1996. 245–254. w51x H.K. Wolf, D. Roos, I. Blumcke, T. Pietsch, O.D. Wiestler, Perilesional neurochemical changes in focal epilepsies, Acta Neuropathol. Berl. 91 Ž1996. 376–384. w52x A.R. Wyler, F.C. Dohan, J.B. Schweitzer, A.D. Berry, A grading system for mesial temporal pathology Žhippocampal sclerosis. from anterior temporal lobectomy, J. Epilepsy 5 Ž1992. 220–225.