Neurochemistry and epileptology


Address correspondence to Harry Meinardi, 49 Mauritskade 2514 HG, The Hague, The Netherlands. E-mail:


This paper gives an account of the global evolution of (neuro-)chemistry in epileptology with an emphasis on the role of the International League Against Epilepsy (ILAE), which declared in its constitution a mission “to make the epilepsy-problem the object of special study and to make practical use of the results of such study.” As Epilepsia is the scientific journal of the ILAE, the review emphasizes papers published in the journal.

It was six men of Indostan, to learning much inclined, who went to see the Elephant (though all of them were blind), that each by observation might satisfy his mind.
J. G. Saxe

Although many important contributions in the field of neurochemistry relevant to epilepsy will have appeared and still do appear in journals other than Epilepsia, the journal owned by the International League Against Epilepsy (ILAE), this journal will be used as primary source for this centenary review. We aim to give a global account of the role of (neuro-)chemistry in advancing epileptology, as facilitated by the League, which declared in its constitution a mission “to make the epilepsy-problem the object of special study and to make practical use of the results of such study.” Fair enough, then, to expect that articles published in its journal should reflect how (neuro-)chemistry supported the objectives of the ILAE.

Broadly speaking, contributions can be divided into two classes: those concerned with elucidating the pathogenesis of epilepsy and those geared to providing knowledge about the best way to treat the condition. Obviously each of these developments can cross-fertilize the other: a positive effect of a therapeutic agent can suggest the nature of the pathologic processes, and conversely, insight in a pathologic process may indicate the tools to restore normal function.

Because Epilepsia will serve as our window on progress, the overview will reflect the journal’s history. The current volume (50) belongs to the fourth series of Epilepsia, which started in 1959. The preceding half-century saw three interruptions of its publication. In the first half-century, too, the character of the contributions was inconsistent, and varied from simple compilations of references to papers published in other journals, critical overviews, and occasionally an original paper.

The First Half of the “ILAE Century”

Generally speaking, in the first 50 years of the ILAE (1909–1959), chemistry was mainly a tool to try and elucidate processes that could explain the pathophysiology of epilepsy. Also during this period, several successful new antiepileptic drugs were introduced.

The first series of Epilepsia

Endogenous compounds as pathogens

In the first article of the first issue of the new journal, Raymond and Sérieux (1909) noted that reports of contemporary authors were contradictory. First urea was believed to be a cause of epilepsy; later it was shown that urea is nontoxic. Next uric acid seemed to be present in higher concentrations in persons with epilepsy, but these authors had not heeded differences in the diets of the subjects studied. Ammonium carbonate was considered by Krainsky to be the culprit, but in reality the levels found in blood were too low to accept this hypothesis. A causative role of phosphates could also be rejected. Phosphates were not toxic, and their elevation was a consequence and not a cause. Donáth reported that persons with epilepsy showed choline in their cerebrospinal fluid (CSF) and blood. Claude and Blanchetière repeated these studies of persons with epilepsy as well as healthy controls and found the substance in both. That such was not known before was considered the result of the technique used by Donáth, which was subsequently also employed by Claude and Blanchetière.

The second article of the first issue was written by Otto Binswanger (1909, p. 37) and titled “Tasks and Objective of Epilepsy Research.” Interestingly, this scholarly treatise featured statements such as: “Functional neuroses and psychoses are ultimately due to (chemical/nutritional) disturbances of the neuronal substrate, in particular in the central nervous system, which will clear off (ausgleichbaren Störungen).” Yet, at the onset of the 20th century scientists had not the faintest idea what exactly these chemical/nutritional processes of the neurons were.

Discussing contemporary attempts to find toxic processes in persons with epilepsy, Binswanger cautioned that establishing that such processes exist does not solve the question of whether they are pathogenic or a consequence of epilepsy. He conceded that contemporary studies indicated that developmental or disease-driven changes in glands with internal secretion (e.g., thyroid and adrenal glands) could indeed cause autointoxication and certain nervous disorders. In addition, some studies had revealed how intestinal disturbances might be an etiologic factor in epilepsy, designated by Alt as “Magenepilepsie” and by American authors as intestinal epilepsy. By the same token, Binswanger posited that the fact that many people are exposed to the pathogens that are held responsible for “intestinal epilepsy” without developing epilepsy was much more supportive for the theory that genuine epilepsy is based on an innate developmental disorder.

Exogenous compounds cause epilepsy

The third paper in this first issue dealt with “alcohol epilepsy.” The author, Emil Redlich of Vienna, began his discourse by referring to a contemporary controversy over whether an entity such as epilepsy caused by excessive alcohol existed. According to Redlich (1909), most authors agreed that alcohol could cause epilepsy, although they differed about the incidence and pathogenesis. Others, however, maintained that epilepsy in dipsomaniacs was either fortuitous or at most based on alcohol abuse eliciting seizures in latent epilepsy. Redlich described how seizures can occur in some people when they are heavily intoxicated but in others 2 or 3 days after alcohol withdrawal. This phenomenon was in later years considered equivalent to phenobarbital withdrawal seizures, but it was not understood by Redlich and occasioned various hypotheses. One farfetched example held that alcohol abuse induces a natural antidote which, however, if not in balance with alcohol gives rise to seizures.

In 1910 Epilepsia appeared twice, in September and in December. William Aldren Turner (1910) of London addressed the role played by chemical substances in the etiology of epilepsy. Under the heading “Toxaemic Epilepsy” he wrote: “It has for a time been contended that an excitant of epileptic seizures might be found in toxic or auto-toxic causes arising in connection with the body metabolism or gastro-intestinal disorders. The discrepancies, however, in results of observations by different workers upon urinary and blood toxicity in epilepsy have not afforded satisfactory evidence upon which to found such a theory. It is now generally accepted that the altered condition of the urine and blood, which has been found in association with epileptic attacks by many observers, is the temporary effect and not the cause of seizures” (p. 31). Turner acknowledged, however, that Binswanger had called attention to a small group of epilepsies that might have a toxemic basis. Binswanger had referred to papers in which “a hypotoxic condition of the urine was found before and a hypertoxic after the fits” (p. 32), or where a relation was assumed with diurnal variations in uric acid. Alternatively, seizures were attributed to the formation of carbamic acid, a urea derivative. Donáth was the first to call attention to choline in the blood and CSF of persons with epilepsy. In 1904, Mott and Halliburton had attributed the presence of choline in the CSF to disintegration of the myelin sheaths of nerve fibers. Donáth proved experimentally that moderate doses of choline injected in the circulation do not produce convulsions, but that in high doses both convulsions and paralysis could occur. Donáth also tested other substances found in the blood and urine of people with epilepsy in dogs and guinea pigs. Uric acid, neutral sodium urate, creatine, and lactic acid did not produce seizures, although ammonia and organic ammonium bases, as well as trimethylamine, choline, creatinine, and guanidine did.

That same year Claude and Lejonne (1910) of Paris suggested that many young children have meningoencephalitis from which they appear to completely recover. Their brains, however, may “become susceptible to otherwise innocuous gastro-intestinal upsets or changes in the glands with internal secretion” (p. 11). This hypothesis was based on their experiments with dogs in which a meningoencephalitis was induced, which subsided seemingly without aftermath. But when, months later, the dogs were exposed to a toxin that was without effect in healthy counterparts, the experimental animals reacted with seizures and eventually died.

A few years later Bolten (1915) turned the tables and posited that in “genuine epilepsy there exists hypothyroidism and consequently metabolism is retarded and less complete, because all kinds of ferments (as well of the intestinal tract as of the intermediary metabolism) are insufficiently produced, and all kinds of toxic products of our own metabolism as well as breaking products of feeding materials come into the circulation and gradually they accumulate and become lodged in the cerebral cortex, which, when this accumulation of toxins has reached its maximum, react upon that with a so-called ‘discharge’, the epileptic fit, as a temporary remedy of the organism to deliver itself of the toxins for a short period” (p. 308). Bolten ended his thesis by saying: “Genuine epilepsy is a quite curable disease; by the administration rectally, of fresh press-extracts of the insufficient organs (thyroid and parathyroid glands), one succeeds to free the patients from all epileptic phenomena, as I have proved in a great series of cases” (p. 309).

The second series of Epilepsia

The century of international epileptology was seriously affected by the two world wars. The last issue of Epilepsia’s first period was in 1915. It was 21 years before the first issue of the second series appeared, in February 1937. This issue contains no information whatsoever pertaining to our topic. During this period, Epilepsia was published once a year. In the 1938 issue Stauder (1938) reported on progress in epileptology in Germany. He mentioned studies of Fritsch, who posited that there is a vicarious influence of albumin and globulin levels on epilepsy. Fritsch had noticed an increase in albumin levels shortly before seizures occurred. When increased globulin levels were induced in dogs, thrice normal stimulus strength had to be used to provoke seizures by electrical stimulation of the cortex. Kroll posited that extracts of brain areas that had been electrically stimulated to provoke seizures and extracts from a surgically removed human epileptic focus could provoke severe seizures if injected in other animals. Stauder cautioned that these experiments still needed independent confirmation. (In the 1939 issue of the journal, William Lennox reported that Keith and McEachern had punctured the reports of Kroll.) Lennox (1938) reviewed the literature of 1936; few articles are relevant for our purposes. Spiegel and Spiegel had done work on physicochemical mechanisms in convulsive reactivity. They concluded that epileptogenous agents probably act in two ways: “by production of a change in concentration of ions at cell surfaces,” which when supra threshold, results in a convulsion and “by diminution in the density of the cell surface film” (p. 150), resulting in increased permeability and, therefore, lowering of convulsive reactivity.

In the third issue of this second series, H. P. Stubbe Teglbjaerg (1939) from Denmark lamented. “If we go through the scientific results which in the course of the past 20 years—the period I have witnessed—have been reputed and ascertained as facts, we are disgraced on behalf of science by much uncritical acceptance of new allegations, which are even frequently joined on a defective basis and carried on, the result being that epilepsy researchers have repeatedly been led into a blind alley, I am especially thinking of certain ‘pathognomonic’ metabolic anomalies which are demonstrated over and over again. When conscientious researchers arrive at the true recognition of such mistakes they are often disillusioned as to the possibility of yielding a lasting positive contribution in this domain” (p. 189). In any event, the yield of information about scientific contributions in the field of neurochemistry in this issue is nil.

The fourth issue again is almost void of information relevant for this review. There is, however, mention in the report of P. C. Cloake (1940) on English studies of a paper by Tod and Stalker on neuropsychiatric aspects of bromide intoxication. In this paper, which can be considered a first example of “monitoring anti-epileptic drugs in body fluids,” the authors posit that a blood level of bromide under 100 mg produces no symptoms; 100–200 mg produces symptoms occasionally in elderly patients with cardiorenal inefficiency; and over 200 mg usually produces symptoms. In a report by Lennox (1940) from the United States, the only remarkable information concerns the “discovery and clinical use of sodium diphenyl hydantoinate as a new and effective anti-epileptic drug” (p. 285).

The Second World War made it necessary to move publication of Epilepsia to the United States, where volume II of the second series starts with issue 1 in July 1941. The number of full reports had declined, and there were delays in reviewing the literature. In the December 1942 issue Lennox presented papers published in various journals in 1940 (Lennox, 1942). Gibbs and Gibbs compared the oxygen and carbon dioxide content in arterial blood and that from the internal jugular vein in persons with epilepsy and healthy controls. While the respiratory quotient of healthy controls was 0.98, the investigators found that, especially in patients with “petit mal,” this was reduced to 0.90. Accordingly, they concluded that in these patients there is a deviation in glucose metabolism in the brain (p. 121). In a second paper they stated that, “whereas the oxygen, sodium and potassium were normal in epileptic persons, the value of carbon dioxide in both the arterial and internal jugular blood were abnormal in the following respects: in patients subject to petit mal seizures carbon dioxide values tended to be abnormally low, whereas in those subject to grand mal seizures they tended to be abnormally high; spontaneously occurring grand mal and petit mal seizures were preceded by abnormal fluctuations in the carbon dioxide content of arterial and internal jugular blood, time relations indicating a causal linkage between the carbon dioxide content of blood and seizures” (p. 123–124). Chick, El Sadr, and Worden were reported to have demonstrated that within 8–38 weeks of depriving rats of pyridoxine (vitamin B6) seizures would occur. Re-establishing a normal diet relieved the condition. There are many papers about the effects of antiepileptic drugs, in particular the recently introduced phenytoin. One paper by Cohen, Coombs, Cobb, and Talbott addressed the putative mechanism of action of azosulfamide. They found a decrease in carbon dioxide content and carbon dioxide-combining power and an elevation of chlorides in serum. The authors concluded that the anticonvulsant effect of both azosulfamide and phenobarbital coincided with a positive potassium balance. Ammonium chloride produced the same degree of acidosis as azosulfamide, without changing potassium chloride content, and did not have anticonvulsant action.

The first postwar issue was volume III, issue 1 of the second series. The acting editor, William Lennox, 1945, wrote: “Because of the enlarged subscription list publication of original communications is now possible” (p. 7). Yet a large portion of the issue still consisted of reviews of the literature.

Schütz posited that barbiturate withdrawal seizures could be due to the fact that phenobarbital therapy reduced serum cholinesterase and that it could take a while for normal values to be restored. From the précis (p. 35) presented it is not clear whether, before treatment, persons with epilepsy exhibited normal levels or low levels of serum cholinesterase.

This issue also has a paper by Houston Merritt and Tracy Putnam (Merritt & Putnam, 1945), who first reported their method to determine anticonvulsant properties of chemical compounds in 1937 (Putnam & Merritt, 1937) by applying electroshocks to cats, and here present a report on 700 compounds that, according to their structural formula, belong to barbiturates: benzoxazoles; hydantoins; ketones and phenyl ketones; oxazolidinediones; phenyl sulfides, sulphones, and sulfoxides; and miscellaneous. Of all compounds tested, 76 had a maximum rating for seizure suppression on a five-point scale.

The December 1946 issue contained a review of a paper by Boszormenyi claiming increased permeability of the blood–brain barrier during convulsions, and of one by Foster who posited that “the physiological cause of epilepsy may be defined as abnormalities of acetylcholine metabolism” (p. 141). A paper by Madsen published in 1943 was reviewed in which the study of ammonia output and electrolyte balances led to the statement: “Thus these studies confirm the theory of a reversible increase in cellular permeability, with altered colloid-osmotic conditions, resulting in increased irritability of the brain cells.” Lennox, the editor, added the following note to this review: “Kidneys are far from the brain. Can these urinary findings be backed by blood and brain wave change?” (pp. 142–143).

The September 1947 issue cited a paper by Pope, Morris, Jasper, Elliot and Penfield studying epileptogenic areas of cerebral cortex in man and the monkey. Interstitial fluid pH was measured with a glass electrode described by Nims. pH was not considered to be an important factor in determining heightened neuronal excitability in epileptogenic lesions as described, nor was an indicator of the cytochrome oxidase system significantly altered in the human foci studied. However, comparative studies on cholinesterase activity in normal and epileptogenic lesions both in humans and monkeys demonstrated increased activity of the enzyme in areas showing epileptiform electrical discharges (p. 234).

In a paper by Davies and Rémond (Anon, 1947) (appraised by the review editor as a brilliant demonstration of cortical physiology), following onset of electrocorticographic epileptiform changes, a decline in the oxygen tension of the cortical tissue and of the cortical venules was demonstrated without change in the oxygen tension of the arterioles. Direct measurement of the oxygen consumption of the cortical tissue demonstrated an increase coincident with the convulsion (p. 249).

A report by Alexander in the December 1948 issue titled “Neuropathology and Neurophysiology Including Electroencephalography in Wartime Germany” contained a remarkable statement: “Hallervorden, to whom went the brains of those dying in the ‘killing centers for the insane’, was offered but refused the brains of epileptics because he found that nothing of significance would be found in them” (p. 309).

Three abstracts concern agene (nitrogen trichloride), which at the time was used to whiten bread more. An editorial in the British Medical Journal of December 1947 signaled the finding of Mellanby that dogs fed such bread would develop running fits. Newell et al. reported in the Journal of the American Medical Society of November 1947 that the toxic factor was associated with the protein but not with the lipid or carbohydrate. Furthermore, they found that cats fed agene developed seizures, monkeys some electroencephalography (EEG) changes but no seizures, and rats, chicks, guinea-pigs, and people neither seizures nor EEG changes. (The editor wonders what would have happened if also people with epilepsy had participated in this last study.) In the same issue of JAMA, Silver et al. reported that following treatment with agene, mixtures of amino acids and the individual amino acids cysteine and cystine produced the same effect.

December 1949 heralded a new volume of Epilepsia, which eventually consisted only of two issues, 1949 and 1950. R. Eeg-Olofsson reviewed a book titled Total Protein, Globulin and Albumin in Lumbar Fluid in Cryptogenic Epilepsy. The book contained more than 200 references. Epilepsy hyperproteinorrhachia—an increase in spinal fluid total protein to 65 mg or above without an abnormal number of cells—was encountered in six cases. Abnormal low globulin-albumin quotients (under 0.14) were found in 14 patients. These abnormalities were not related to duration of the disease or to the near approach of a seizure. Refereed papers from various journals debated the role of acetylcholine and anticholinesterases. The section on pharmacology mentioned a paper by Torda and Wolff, who studied a long list of convulsant and anticonvulsant agents in terms of their influence on the synthesis and hydrolysis of acetylcholine and the sensitivity of effector organs to the substance. They state that their results indicate that most convulsants cause an accumulation of acetylcholine and most anticonvulsants a decrease.

Local oxygen consumption measured with an oxygen cathode was studied following strong stimulation of a cortical area. Another interest of the time appeared to be the role of glutamic acid in brain function; seven papers on the topic were cited. In 1949, Epilepsia lost the support of the organization that had carried it through the war (the American Epilepsy League) and ceased operation. It did not pick up again until 1952.

The third series of Epilepsia

The first volume of the third series of Epilepsia appeared in November 1952 and contained a statement that the annual abstract journal, Epilepsia had lost its “raison d’être.” The present editors (the publications committee of the American League Against Epilepsy, chaired by Jerome K. Merlis) wished to offer an opportunity to people desiring to critically examine what had been done in the past, to point out what new facts were required.

One such paper was by Toman and Taylor about the mechanism of action and metabolism of anticonvulsants. The following remarks can be found in this treatise: Action on cholinergic mechanisms had frequently been considered either as a factor in the production of seizures, or in the effects of protective drugs. It had been noted that acetylcholine-binding substances are altered in tissues from epileptogenic foci and that some anticonvulsants were capable of changing the biochemistry toward normal. In general, it could be said that there is as yet no consistent body of information relating cholinergic mechanisms either to epilepsy or antiepileptic drug action. In another part of the review the authors observed that, for some drugs like phenobarbital, there was a rapid rebound in “grand mal” after withdrawal, in others such as trimethadione remission of “petit mal” could persist for weeks to months following drug withdrawal. The paper ended with a remark which, although devoid of any biochemical explanation, is curious enough to be repeated: “Aureomycin can be used in the control of otherwise refractory ‘petit mal’ of viral etiology long after the acute infection has subsided, although aureomycin has no demonstrable general anticonvulsant properties.”

A paper by Pope updated epileptologists about the Krebs cycle. Pope noted that metabolic increments can be seen as the result of excessive neuronal discharges, but that convulsions may well be caused by oxygen or glucose deprivation or poisoning of critical enzymes such as the cytochrome system. He mentioned an important role of glutamic acid, which plays a part in buffering ammonia. Alone among amino acids, glutamic acid supported brain respiration in vitro and appeared to have an important role in regulating membrane permeability in the brain. Pope did not wish to take a stand in the discussion on the controversial role of acetylcholine in the production, propagation, and synaptic passage of nerve impulses; but he did emphasize that the enzyme systems for formation and hydrolysis of acetylcholine were ubiquitously present in excitable tissues and must be intact for normal function. Pope abstained from discussing what he nevertheless called “the revolutionary observations” of Caspersson and Hydén showing rapid intracellular protein and nucleoprotein turnover during neuronal discharge.

Tower, in contrast, took the acetylcholine system as point of attack to better understand the metabolic processes involved in seizure occurrence. He observed that in epileptogenic cortex there was an impairment of bound acetylcholine production without reduction in synthesis of free acetylcholine. Furthermore, he considered it likely that glutamic acid was involved in the mechanism that forms bound acetylcholine. With a defect in the mechanism responsible for retaining acetylcholine in its bound form, the presence of abnormal amounts of free (active) acetylcholine was certainly possible. And, according to Tower, “by now it is well recognized that experimentally acetylcholine can give rise to epileptiform activity in the electroencephalogram.” In reporting the deliberations of the Committee on Research on Experimental Epilepsy, Ward noted that Stavraky had presented evidence that led him to believe that the “law of denervation” applied to the central nervous system (CNS) as well as to the peripheral system. Therefore, deafferented neurons would become hypersensitive to acetylcholine, which explained the hyperactivity displayed by epileptogenic areas.

The opening paper of the third series, issue 2, November 1953, is a biochemistry-related paper, but it is the only one. Glaser (1953) discussed the relationship between adrenal cortical activity and the convulsive state. The author refers to many and often contradictory studies. In his opinion the mechanisms by which adrenal hormone imbalance causes brain hyperexcitability and seizures, and the nature of the antagonistic effect of desoxycorticosterone and cortisone had up to then not been adequately elucidated.

In the third volume of this series, a third-year student from New York, Jerome Fabricant (1954), was awarded a first prize for a paper on the role of adrenaline in epilepsy (According to PubMed, apart from this paper on epilepsy, Fabricant published on veterinary topics between 1950 and 1999.) This is the only paper in this volume to treat the biochemical aspects of epilepsy. From a study of the literature Fabricant posited that in patients with epilepsy, the electrolyte effects of adrenaline were not sufficiently checked by adrenocortical hormones, owing to adrenocortical hypofunction. In conditions of stress these ionic effects could be sufficient to excite discharges. Several mechanisms by which adrenaline might act were suggested: promotion of acetylcholine synthesis; anticholinesterase activity; and inhibition of carbohydrate metabolism leading to permeability defects with consequent leakage of potassium, increased free acetylcholine, and increased permeability to acetylcholine. Large amounts of adrenaline, however, depressed the action of acetylcholine. This was invoked as the mechanism by which the seizure ceased and the epileptic patient fell asleep.

The Second Half of the ILAE Century

Following an interval of 4 years, the present (fourth) series of Epilepsia began. The first volume of this series is labeled 1959/1960 and marks the second half of the ILAE century. This period would cover a span of time in which from the outset contributions were made by scientists and physicians who are still alive and are well aware of the achievements and disappointments in their field. They have seen the number of colleagues interested in epilepsy multiply almost exponentially. Constrained by space, but because we wish to provide a recognizable overview, our report alters somewhat in outlook and content. From here on we try to depict with broader strokes how chemistry has helped to understand the pathology of epilepsy. However, a major part of the information obtained with the help of chemistry published in Epilepsia concerns antiepileptic drugs, their clinical use, monitoring of their levels in body fluids, and their mechanisms of therapeutic action. The pages of the journal also included reports on the mechanisms of adverse effects of therapeutics, including drug interactions and novel insights regarding the role of “multiple drug resistance” in drug tolerance for adverse and beneficial effects. An attempt to succinctly summarize this information would far exceed the available space. Because a monograph on this topic also chronicles the history and is regularly updated, the reader is referred to that source (Levy et al., 2002). Furthermore, although molecular genetics can be argued to belong to the domain of chemistry, the review of progress in that field is left to the discussion of genetics.

Progress obviously depends on understanding processes; however, without the assistance of new technical developments such processes would never have been unraveled. The first part of this account, therefore, summarized which new techniques provided the tools that allow us to answer at least some of the questions that hopefully will bring us nearer to our goal. We also describe the substrates from which information was obtained, which we nicknamed windows. The next sections deal with cerebral energy and cerebral lipids, and end with cerebral peptides and amino acids.

Technical developments in solving chemical problems

Obviously the papers acknowledged here are those that first introduced new techniques to the readership of Epilepsia. Several technical developments antedated the moment when first mention of their use in epileptologic problems was made in the journal.

Clinical experience had made abundantly clear that medications offered to patients had variable outcomes. There was a logical desire to be informed whether the medication prescribed did reach its target in the amounts predicted by the amounts administered. The first mention in Epilepsia of the use of thin layer chromatography to show how much phenobarbital and phenytoin was actually present in plasma was written by Buchthal and Svensmark (1960). About a decade later (and two decades after Archer John Porter Martin developed the technique), a first report by Meijer (1971) appeared on how using gas chromatography would permit simultaneous determination in body fluids of currently used antiepileptic drugs. Gerber et al. (1979) first mentioned using gas chromatography coupled with mass spectroscopy, which they employed in identifying metabolites of mesantoin. Monaco et al. (1982) compared the results of enzyme multiplied immunoassay technique (EMIT) and high-pressure liquid chromatography (HPLC) to estimate phenobarbital, phenytoin, and carbamazepine. In the same year Li et al. (1982) reported the feasibility of a simple, sensitive, non-radioisotopic fluorescence immunoassay to measure carbamazepine. The proceedings of the 2nd Merritt-Putnam symposium (1984) from the 15th Epilepsy International Symposium in Washington, DC, presents discussions of technical advances such as computerized axial tomography scanning. Magnetic resonance imaging (MRI; also called nuclear magnetic resonance, or NMR, imaging) and positron emission tomography (so-called PET scans), which would later play a role in answering chemistry-related questions. Theodore et al. (1987) reported on the ability of [18F]-2-deoxyglucose (FDG-)PET to provide information regarding how antiepileptic drugs influence local cerebral metabolic rates of glucose utilization. Over the course of years, PET scans also used other markers. Using [11C]flumazenil PET, Savic et al. (1990) demonstrated a reduction of the cortical benzodiazepine (BZD) receptor density in the epileptic foci of patients with partial epileptic seizures. Werhahn et al. (2006) used a high-affinity dopamine (D2/D3) ligand [18F]fallypride-PET to visualize extrastriatal binding. Korja et al. (2007) used [11C]raclopride-PET to study dopaminergic abnormalities in Unverricht-Lundborg myoclonus epilepsy. Finally, very recently Natsume et al. (2008) reported how, with the help of α-[11C]methyl-L-tryptophan PET, he and his colleagues tried in vain to check for abnormal tryptophan metabolism to determine pathology in periventricular nodular heterotopia. Abnormal metabolism in the neocortex was indeed revealed but did not co-localize with the EEG focus.

In August 1990, Cochran et al. reported on the AccuLevel assay, a new whole-blood, noninstrumented immunochromatographic test for carbamazepine (CBZ), which permits determination of its blood level in the physician’s office. Duncan (1996) contributed an informative paper on the advantages of magnetic resonance spectroscopy (MRS), among them its noninvasiveness and its ability to readily combine with MRI. By this time, attention was focusing on proton (1H) and phosphorus (31P) MRS, and studies had been undertaken using single voxels or many voxels simultaneously (chemical-shift imaging, MRS imaging). The latter was more difficult and prone to artifact, but potentially yielded significantly more information. 1H MRS had principally yielded data on concentrations of N-acetyl aspartate (NAA), choline, creatine, and phosphocreatine. NAA is located primarily within neurons, and reduction of the ratio of NAA to choline, creatine, and phosphocreatine is a marker of neuronal loss and dysfunction. It was also possible to estimate cerebral concentrations of several neurotransmitters: glutamate, glutamine, and γ-aminobutyric acid (GABA). As with all functional imaging methods, co-registration with high-quality MRI was essential for interpreting data.

In 1999 Verhoeff et al. used [123I]iomazenil single-photon emission computed tomography (SPECT) to estimate central type A GABA (GABAA)/benzodiazepine receptors combined with MRS to assess tissue GABA levels under the influence of the antiepileptic drug vigabatrin (γ-vinyl-GABA). In the same issue Krakow et al. (1999) introduce studying epilepsy by using EEG-triggered functional MRI (fMRI), diffusion tensor imaging (DTI), and chemical shift imaging (CSI). The applied MRI methods describe functional, microstructural, and biochemical characteristics of the epileptogenic tissue that cannot be obtained with other noninvasive means and thus improve the understanding of the pathophysiology of epilepsy.

In November 1999 Adelson et al. (1999) reported the use of near-infrared spectroscopy (NIRS) to examine the changes in cerebral oxygenation in the periictal period in patients with seizures. de Vasconcelos et al. (2000) described a technique to study the role of nitric oxide and nitric oxide synthetase in inducing seizures. Kohane et al. (2002) reported the use of biodegradable lipid–protein–sugar particles of 4–5 μm diameter to deliver drugs at specific locations in brain. Tenney et al. (2003) were the first to make explicit mention of using blood-oxygenation-level–dependent (BOLD) fMRI (discovered by Seiji Ogawa in 1990).

Much later, Cucullo et al. (2007) claimed to have achieved a biotechnological breakthrough by developing a humanized in vitro blood–brain barrier model based on cocultures of human microvascular endothelial cells from “normal” and drug-resistant epileptic brain tissue with human brain astrocytes from epilepsy patients or controls. Yuto Ueda et al. (2007) then reported the use of X-band and L-band paramagnetic resonance spectroscopy to study antioxidant ability in biologic studies. They applied the method to a study of in vivo antioxidant properties in the dorsal hippocampi of rats that had seizures induced by FeCl3; antioxidant ability was shown to be significantly decreased bilaterally.

The last issue considered for this review was that of August 2008. In it, Akhtari et al. (2008) reported the development of nonradioactive and targeted magneto-nanoparticles, composed of iron oxide and dextran, capable of crossing the blood–brain barrier and of concentrating in the epileptogenic tissues of acute and chronic animal models of temporal lobe epilepsy to render these tissues visible on MRI.

Windows for observing pathologic and therapeutic processes

To understand the chemical processes involved in epilepsy and its treatment, an obvious distinction has to be made between observations in people with epilepsy and research that employs animal or even purely theoretical models.

For example, readily available blood allows biochemical analysis in humans, which has been particularly valuable in understanding the pharmacokinetic aspects of therapy. The CSF is one step nearer to the site where the epileptic process initiates and where therapeutic agents are expected to find their point of action. When pneumoencephalography was an accepted part of neuroimaging, CSF was also available for the laboratory. Thus Johannessen and Strandjord (1973) were able to discuss findings about carbamazepine concentrations in blood and CSF relative to the dose taken. In the same year a laboratory paper (Craig & Hartman, 1973) dealt with concentrations of amino acids in brains of rats made epileptic by inducing a focus with cobalt.

Measuring antiepileptic drugs in the CSF had the advantage that concentration in the CSF is an ultrafiltrate of the blood concentration and directly in touch with the cells where the substance is to exert its action. Apart from using laboratory methods to obtain an ultrafiltrate, a natural and accessible source of “filtrated” blood is available as tear fluid or as saliva. Both these fluids have been examined for their efficacy in guiding antiepileptic drug therapy. A first paper titled “Tears as the Best Practical Indicator of the Unbound Fraction of an Anticonvulsant Drug” was published in EpilepsiaMonaco et al. (1979). A second paper appeared on this subject by Paxton (1982) who focused on carbamazepine and also discussed whether the EMIT would be a suitable tool for measuring drug levels compared with HPLC.

In 2002 mention was first made in Epilepsia about using analysis of antiepileptic drug concentrations in subsequent sections of hair as a form of historical bookkeeping for compliance over a prolonged period of time. In such a way Williams et al. (2002) tried to assess whether women during pregnancy had reduced their antiepileptic medication without reporting it to their physician.

Cerebral energy

In the 1970s the assumption was examined and subsequently rejected that seizures end owing to oxygen depletion or hypercapnia (Caspers & Speckmann, 1972). After a brief early mention of the ketogenic diet (KD) (Millichap & Jones, 1964), a subsequent paper in 1974 examined its effects in rat brain. The only changes were elevations of d-β-hydroxybutyrate and sodium and lowering of acetoacetyl CoA transferase activity. It was posited that altered metabolism of glucose led to an elevated electroconvulsive threshold. Single papers appeared in 1978, 1986, and 1992, but from 1995 onward each year several papers on the KD topic were published, although up until now no satisfactory explanation of the mode of action has been found. The three most recent papers (Porta et al., 2008; Taha et al., 2008; Willis et al., 2008) are prepublications and reach opposite conclusions. They deal with polyunsaturated fatty acids (PUFAs), which have been suggested to modify brain phospholipids and explain the effect of the KD. The paper of Porta et al. examined docosahexaenoic acid and eicosapentaenoic acid, which significantly alter the fatty acid profile of the brain but do not alter seizure thresholds or behavior. The other two papers found an increase in the pentylenetetrazol threshold, but one concludes that this effect is unrelated to brain phospholipid composition and the other that linoleic and α-linolenic polyunsaturated fatty acids in a 4:1 ratio raises n-3 PUFA composition of unesterified fatty acids.

In the 1980s the cerebral energy topic centered on studies of ATP concentrations. There is uncertainty as to whether seizures increase expenditure or disrupt cellular and mitochondrial membrane function (Kogure & Schwartzman, 1980). McCandless and Schwartzenburg (1982) reported that, when seizures were evoked in DBA/2J mice, decreases in high-energy phosphates were more pronounced in the cerebellum than in the parietal cortex. It is posited that this is due to cerebellar activity trying to reduce the severity of the wild-run and tonic extension seizure. At the end of the 1980s Swartz et al. (1989) first reported that FDG-PET could demonstrate focal hypometabolism, which significantly correlated with the electroclinical ictal localization. Brines et al. (1995) later noted that although sodium pump protein in surviving neurons appears to be upregulated in epilepsy, sodium pump capacity may be limited by reduced levels of cytochrome C oxidase activity. Functional reduction in sodium pump capacity may be an important factor in hyperexcitability and neuronal death.

O’Brien et al. (1997) used PET and MRI to discover whether regional hypometabolism in hippocampal sclerosis is due to hippocampal cell loss. There was no correlation between the magnitudes of the FDG-PET asymmetry index and the MRI volume ratio for the mesial or lateral temporal regions. As in temporal lobectomy, specimens of patients with hippocampal sclerosis showed widespread microdysgenesis in the temporal neocortex. The authors posited that a decrease in synaptic activity in dysplastic neocortex could account in part for the diffuse regional hypometabolism seen on FDG-PET scans.

At the turn of the millennium, Zubal et al. (2000) addressed the diagnostic use of the fact that in epileptogenic tissue, perfusion and metabolism appear uncoupled. Interictal PET (18F-FDG) and interictal SPECT (99mTc-HMPAO) scans were acquired. The metabolism and perfusion images were three-dimensionally spatially registered, and a functional ratio image was computed. These functional maps were overlaid onto a three-dimensional rendering of the same patient’s MRI anatomy. Calculation of a functional ratio image demonstrated localized foci that in some cases could not be observed on the PET image alone. The ratio image also yielded a quantitative measure of the uncoupling phenomenon. Nickel et al. (2003) subsequently tried to correlate observed FDG-PET activity with the maximal glucose-oxidation rates in 400-μm-thick hippocampal subfields obtained after dissection of human hippocampal slices into the CA1 and CA3 pyramidal subfields and the dentate gyrus. A correlation was observed between FDG-PET activity and the maximal glucose oxidation rate for the CA3 pyramidal subfields but not for the CA1 regions and dentate gyrus. No correlation of the FDG-PET with the neuronal cell density of CA1, CA3, and the dentate gyrus was found. The investigators also asked whether gender-specific differences exist in the magnitude and distribution of hypometabolism in mesial temporal lobe epilepsy patients determined with FDG-PET and, if so, whether these findings are related to clinical and neuropsychological abnormalities. Extramesiotemporal hypometabolism prevailed in the male patients. Metabolic asymmetry in temporal and frontal regions was related to performance in the Trail-Making Test and WAIS-R subitems.

A recent investigation (Vielhaber et al., 2008) analyzed whether interictal decrease in NAA detected in mesial temporal lobe epilepsy by proton MRS is due to loss of pyramidal neurons or metabolic dysfunction. High-resolution 1H-MRS at 14.1 Tesla measured metabolite concentrations from hippocampal subfields (CA1, CA3, dentate gyrus) and the parahippocampal region. In contrast to four patients with medial temporal lobe epilepsy caused by lesions, a large variance of NAA concentrations was found in the individual hippocampal regions of patients with Ammon’s horn sclerosis (AHS). Specifically, in subfield CA3 of AHS patients, despite moderate preservation of neuronal cell densities, the concentration of NAA was significantly lowered, whereas the concentrations of lactate, glucose, and succinate were elevated. These subfield-specific alterations of metabolite concentrations in AHS are very likely caused by impairment of mitochondrial function and not related to neuronal cell loss.

Cerebral lipids

It is not until 1978 that a paper concerning the role of lipids appeared in the fourth series of Epilepsia (Bierkamper & Cenedella, 1978). The authors succinctly summarized the state of the art:

Several areas of investigation focused attention on an apparent relationship between lipids, especially sterols, and the etiology or expression of epilepsy. For instance, high fat diets protect epileptic children against intractable seizures and elevation of serum cholesterol levels decreases convulsive seizures in epileptic monkeys. Diets rich in cholesterol have also been shown to decrease the susceptibility of mice and rats to pentylenetetrazol-induced seizures…[L]ong-term subthreshold doses of pentylenetetrazol inhibit cholesterol synthesis in mice. Moreover reduction of brain cholesterol levels in the opossum with U18,66A, a cholesterol synthesis inhibitor, results in a chronic epileptic state. Similarly, treatment of neonatal rats with this drug leads to development of chronic spontaneous seizures. Induction of epilepsy in the rat by insertion of cobalt wire into the cerebral cortex is accompanied by a significant reduction in cortical free (nonesterified) cholesterol and an increase in cholesterol esters (CE)…. Cholesterol, the major sterol in adult brain is present almost exclusively in membranes, the highest concentrations being found in myelin and nerve endings (synaptosomes). The cerebral cortex contains less than one-third of the brain’s total myelin but is rich in nerve endings and synapses. Studies on artificial and biological membranes suggest that specific amounts of cholesterol are necessary to maintain membrane fluidity and stability. Thus variations in these critical amounts of cholesterol could theoretically lead to alterations in normal membrane function, e.g., selective permeability (pp. 155–156).

These authors report that in their experiments they found that the cortical concentration of free cholesterol decreased and the concentrations of CEs greatly increased in the adjacent area of rats implanted with epileptogenic metals (cobalt and nickel) but not in those implanted with nonepileptogenic metals (copper and stainless steel). They posited that these sterol changes reflected membrane alterations or disintegration. Four years later this same group published two papers in Epilepsia (Cenedella et al., 1982; Sarkar et al., 1982). Sarkar et al. noted that desmosterol might substitute for cholesterol in neuronal membranes without detriment, and, therefore, they searched for changes in other brain lipids. Changes in brain lipids were focused in the myelin fraction. Phospholipid levels and the sterol–phospholipid ratio of microsomes and synaptosomes, in contrast to myelin, were near normal; however, gangliosides were clearly elevated in all fractions. Cenedella et al. (1982) reported that the applied inhibitor of desmosterol reductase (U18666A) is not localized in a specific brain region or particular cell type.

Several years passed before another paper concerned with lipids appeared. Singh and Pathak (1990) investigated the relationship between lipid peroxidation, subsequent activation of antioxidative enzymes, and development of iron-induced epilepsy in the rat. Their data showed that, although levels of superoxide dismutase and glucose-6-phosphate dehydrogenase increased by ∼60% and glutathione reductase increased by ∼40%, the increases in the enzyme glutathione peroxidase (GP) and catalase (CA) were much lower, <20%. Therefore, comparatively less increase in CA and GP activities produced a deficiency of these two enzymes in the iron (ipsilateral) focus. It appeared reasonable to suggest that GP deficiency causes lipid peroxidation to increase tremendously during iron epileptogenesis.

That same year, the journal published the proceedings of the 9th Merritt-Putnam Symposium, in which Willmore (1990) reviewed cellular mechanisms of posttraumatic epilepsy. Extravasation of blood is followed by hemolysis and deposition of heme-containing compounds into the neuropil, initiating a sequence of univalent redox reactions and generating various free-radical species, including superoxides, hydroxyl radicals, peroxides, and perfenyl ions. Free radicals initiate peroxidation reactions by hydrogen abstraction from methylene groups adjacent to double bonds of fatty acids and lipids within cellular membranes. Intrinsic enzymatic mechanisms for control of free-radical reactions include activation of catalase, peroxidase, and superoxide dismutase. Following this full paper in 1990, however, up until 2008 only abstracts of papers presented at scientific meetings are found on this topic, and these are not included in this overview.

Cerebral amino acids

Just before the second half of the ILAE century, amino acids are recognized as major players in epilepsy. In 1956, factor I, a brain agent discovered earlier by McLennan (1955) , was chemically identified as the amino acid GABA (Basemore et al. 1956). GABA was first mentioned in Epilepsia in 1964 (Guerrero-Figueroa et al., 1964); when the amino acid was applied topically in animals with a chemical focus, a reduction of epileptiform discharges was found. However, the authors were unable to explain the suppressive effects of GABA by an action in the neuronal cortical organization. Although it was already accepted that GABA acts as an inhibitory transmitter, the mechanism by which it exerts its anticonvulsant action was still unclear. Since that time, more than 400 papers that touch on GABA have been published in Epilepsia concerning neurochemistry, but also pharmacology. The GABA hypothesis of epilepsy has had a very large impact on pharmaceutical developments.

From the early years of this period until today, changes—both increases and decreases—in free amino acids have been reported in the brain tissue of animals with induced seizures [e.g., in the cortex {Pintilie et al., 1970; Craig & Hartman, 1973)], but also in the amino acid content of serum and CSF of patients (e.g., Monaco et al., 1975) and animals with epilepsy. Diminution is the rule, especially that of GABA (e.g., Löscher et al., 1981; Griffith et al., 1991). Moreover, patients with epilepsy without medication have lower GABA concentrations than patients without epilepsy (Löscher & Siemes, 1985). These data were interpreted as evidence that indeed an impairment of GABA neurotransmission contributes to an increased seizure propensity. But later (van Gelder et al., 1983; Dolina et al., 1993), amino acid concentrations were viewed as being in delicate balance, and an increase in glutamic acid concentration in combination with failure of GABA inhibition was considered the most plausible explanation for hyperexcitability. Only the content of taurine was often shown to be increased; indeed, the role of taurine is still unclear (van Gelder et al., 1977;Airaksinen, 1979; Bonhaus et al., 1984; Goodman et al., 1989; Gupta et al., 2005).

In the mid-1980s it was recognized that not only the total content but also the compartment in which the amino acid is located might be of importance. Accordingly, subcellular distributions of amino acids were studied (Bonhaus et al., 1984). An increase in interstitial glutamic acid concentration was suggested to be the cause of penicillin-induced generalized seizures in cats (van Gelder et al., 1983).

Later it was reported that amino acid concentrations are highly dependent on the state of the epileptic brain. The amino acid content of the hippocampus in the pilocarpine model showed, during the acute period, a decrease in concentrations of excitatory amino acids (aspartate and glutamate) and an increased GABA level (Cavalheiro et al., 1994). The silent phase was, however, characterized by a decrease in GABA and glycine levels and an increase in glutamate concentration. The period of spontaneous recurrent seizures showed an increase in all amino acid concentrations. Indeed, the causal relation of changed amino acid concentrations with epilepsy was questioned; highly elevated amino acid levels were observed only in CSF from patients with symptomatic infantile spasms, whereas the idiopathic subgroup showed levels of free amino acids that were not statistically different from those of the nonneurologic control group (Spink et al. 1988).

The search for the underlying cause of the aberrations in amino acids started in the 1980s as well. Genetic factors in seizure disorders were sought. The results of a study of free amino acids in the CSF of multiply affected sibships do not support previously reported increases in plasma taurine, aspartic acid, or glutamic acid in seizure patients (Haines et al., 1985).

Just before the start of the ILAE centenary, in 1905, the concept of “receptor” was introduced independently by Paul Ehrlich and John Langley. At first little attention was paid to the topic in Epilepsia, but once progress had been made in understanding the mechanisms by which cells respond to extracellular signals, including the cascade of events triggered by second messengers, papers related to this topic began to show up in the journal. Thus Dubeau et al., 1992 reported that epileptic activity modulates the metabolism of the second messenger phosphoinositide and alters receptor/effector coupling. This finding was based on measurements in neocortex excised during epilepsy surgery. Knowledge about the subunit composition of the receptors warranted regular attention for altered subunit composition of neurotransmitter receptors in relation to epilepsy. Alterations in subunit messenger RNA (mRNA) expression levels of excitatory amino acids in cortex in kindled animals were reported (Kikuchi et al. 1996), and later in mRNA levels of both GABAA receptors and glutamate receptors in cortex of patients with hemimegalencephaly, a disorder associated with intractable epilepsy (Baybis et al., 2004).

Cerebral peptides

Following the identification of the classical neurotransmitter in the first half of the 20th century, in the early 1970s peptides were recognized as chemical messengers between neurons. Because communication between neurons was a main topic in relation to epilepsy, several papers in Epilepsia were dedicated to them. It became clear that opioid neuropeptides might modulate epileptic phenomena and vice versa. Ochi et al. (1988) considered that opioid peptides act as anticonvulsants: indeed, the leucine–enkephalin content of the brains of epileptic mice increased following a convulsion. The authors concluded that opioid peptides may play an important role in modulating the seizures. This conclusion is supported by several other reports on this group of neuropeptides (Caldecott-Hazard et al., 1983;Ortolani et al., 1990; Henry, 1996). Interestingly, endogenous antagonists of this seizure-inhibiting system were identified as well: Nociceptin, a neuropeptide that activates a receptor of the opioid family, was shown to play a facilitatory role in ictogenesis and in epileptogenesis (Bregola et al., 2002).

Another peptide that was a prominent topic in Epilepsia is adrenocorticotropic hormone (ACTH). It was reported that, in children with infantile spasms, CSF levels of ACTH are lower than in controls (Nalin et al., 1985), suggesting that ACTH had anticonvulsant properties. In addition, Tønder et al. (1994) showed a dense distribution of ACTH-immunoreactive neurons in the hippocampal formations of patients with mesial temporal lobe epilepsy syndrome. This finding was interpreted as an attempt by the damaged hippocampal circuit to restore the balance of excitatory/inhibitory neurotransmission in temporal lobe epilepsy.

Dozens of different peptides are known to be released by different populations of neurons in the mammalian brain. Papers in Epilepsia have focused in particular on substances P and K, neuropeptide Y, VIP, α-MSH, and brain-derived neurotrophic factors. Often several neuropeptides are mentioned in a single paper (Tønder et al., 1994) or in which the CSF levels of the peptides are correlated, for example, medication (Devinsky et al., 1993). Complex changes in neuropeptide expression in some principal hippocampal neurons and interneurons appear as a characteristic feature of spontaneous seizures in mice carrying a deletion of the galanin R1 receptor (Fetissov et al., 2003). Furthermore, the relation between classical neurotransmitters and peptides has been the subject of several papers.

Vezzani et al. (1991) considered the endogenous peptide diazepam-binding inhibitor (DBI) and were the only authors to do so. DBI-peptide fragments induce limbic seizures in rats. DBI and/or its natural processing products may play a role in the pathophysiology of epilepsy. Among other peptides, like DBI, substance P has been suggested to have endogenous proconvulsant properties (Gale, 1988). On the other hand, neuropeptide Y (NPY) is one of the brain peptides capable of inhibiting epileptiform discharges (Reibel et al., 2000; Baraban, 2002). At the end of the ILAE century, in the soup of endogenous neuropeptides, the search for markers for epileptic seizures is well under way (Rauchenzauner et al., 2007).


Conflict of interest: The contributing authors to this paper declare no conflict of interest.