- Email: [email protected]
Lithium–pilocarpine seizures as a model for lithium action in mania
ARTICLE IN PRESS
Neuroscience and Biobehavioral Reviews 31 (2007) 843–849 www.elsevier.com/locate/neubiorev
Review
Lithium–pilocarpine seizures as a model for lithium action in mania R.H. Belmaker, Yuly Bersudsky Ben Gurion University of the Negev, Beersheva Mental Health Center, P.O. Box 4600, Beersheva, Israel
Abstract Lithium (Li) pre-treatment of rats or mice given low dose pilocarpine induces a unique limbic seizure syndrome. This syndrome is stereospecifically reversed by myo-inositol, which suggests that it is a behavioral model for Li depletion of brain inositol. However, this syndrome has little face validity because seizures are not a component of bipolar disorder. Moreover, other animal species that maintain higher brain inositol levels than mice or rats do not show Li–pilocarpine seizures and a study in humans suggests that humans do not show this syndrome as well. It could be suggested that Li–pilocarpine seizures are an in vivo bioassay for inositol depletion. Recent studies of knockout mice lacking inositol monophosphatase-1 or the sodium myo-inositol transporter-1 found that both these knockout mice given pilocarpine develop limbic seizures as if they had been pre-treated with Li. These mice in addition to such pilocarpine sensitivity have other behaviors such as decreased immobility in the Porsolt forced swim test that suggests that their inositol depletion has Li-like effects. Thus, the Li–pilocarpine seizure model may, despite its lack of face validity, be a biochemical marker for a model of mania treatment in animals. r 2007 Elsevier Ltd. All rights reserved. Keywords: Li–pilocarpine; Seizures; Inositol; Mania
Contents 1. 2. 3. 4. 5.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The epi-inositol problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Species differences in susceptibility to Li–pilocarpine seizures . . . . . . . . . . Effect of Li on the physostigmine-induced behavioral syndrome in humans Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1. Introduction A widespread hypothesis explaining Li’s therapeutic and prophylactic effect in affective disorder is that inhibition of inositol monophosphatase (IMPase) impairs the operation of the phosphatidylinositol cycle (PI cycle). The membrane phospholipid, phosphatidylinositol (PI) is sequentially phosphorylated to form phosphatidylinositol bisphosphate (PIP2). Agonist-stimulated phospholipase C (PLC) cleaves Corresponding author. Tel.: +972 640 1602; fax: +972 8 640 1621.
E-mail address: [email protected] (R.H. Belmaker). 0149-7634/$ - see front matter r 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.neubiorev.2007.05.001
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
843 845 847 847 848 848
PIP2 to two second messengers: inositol 1,4,5-trisphosphate (IP3) and diacylglycerol (DAG). IP3 releases calcium sequestered in endoplasmic reticulum. IP3 is subsequently dephosphorylated to inositol monophosphate (IP), which is dephosphorylated by IMPase to free inositol. DAG, the second derivative of PIP2 activates protein kinase C. DAG is converted sequentially to cytidine disphosphate-diacyl glycerol (CDP-DG), which is combined with free inositol by PI synthase to re-form PI. Li was first shown to affect the system in 1971 (Allison and Stewart, 1971). Li inhibits the dephosphorylation of four of the inositol monophosphates as well as two inositol
ARTICLE IN PRESS 844
R.H. Belmaker, Y. Bersudsky / Neuroscience and Biobehavioral Reviews 31 (2007) 843–849
biphosphates, thereby increasing brain levels of IP and two biphosphates and reducing levels of myo-inositol (Allison et al., 1976, 1980; Sherman et al., 1981, 1985b; Honchar et al., 1989). With discovery of the importance of the cycle as a second messenger system, the Li-induced reduction of inositol immediately assumed potential importance as a key mechanism of Li action (Berridge, 1985). Hallcher and Sherman (1980) showed that Li at therapeutic doses inhibits rat brain IMPase, thereby explaining the reduction in inositol and accumulation of IP. Inhibition is uncompetitive with a Ki of 0.8 mM, thus within therapeutically effective serum concentrations in the range of 0.5–1.4 mM (Hallcher and Sherman, 1980). In rats, Li chloride (LiCl), 10 meq/kg, reduced brain inositol levels by 30% and increased IP levels 20-fold after 6 h (Allison et al., 1976, 1980) and 40-fold after 24 h (Sherman et al., 1985b). Chronic administration of Li has been reported to reduce brain inositol levels (Sherman et al., 1985b), but others have not replicated this critical finding (Jope and Williams, 1994). Berridge (1985) proposed that the Li-induced shortage of inositol in the brain, an organ to which plasma-borne inositol is essentially unavailable, leads to depletion of substrate for phosphatidylinositol resynthesis only in overactive neurons. This theory could explain Li’s specific therapeutic activity in psychopathological states with minimal effects on normal behavior (Belmaker et al., 1996). Various biological effects of Li can be reversed by addition of myo-inositol in vitro (Kofman and Belmaker, 1993). To ascertain if Li’s inhibition of IMPase is relevant to its therapeutic effects in patients with affective disorders, it is critical to demonstrate that behavioral effects of Li are also reversed by myo-inositol. Because myo-inositol does not easily penetrate the blood brain barrier when injected systemically (Spector and Lorenzo, 1975), it is often necessary to inject myo-inositol directly into the brain. One of the most robust behavioral effects of Li is that normally subconvulsant doses of muscarinic agonists will induce limbic seizures in rats pretreated with Li (Honchar et al., 1983). Induction of Li–pilocarpine seizures is concomitant with a reduction in cortical myo-inositol levels and an elevation of IP, which is about 10-fold greater than the effects elicited by either Li or pilocarpine alone (Sherman et al., 1985a, 1986). Tricklebank et al. (1991) reported that ICV injections of myo-inositol prolonged the latency to seizures elicited by ICV or systemic Li in mice. We found similar results in rats (Kofman et al., 1993). This effect is dramatic, and many inositol-treated animals do not seize at all (Kofman and Belmaker, 1993). For example, in our laboratory (Kofman and Belmaker, 1993; Belmaker et al., 1998) 28 male Sprague Dawley rats were implanted with guide cannulae in the lateral ventricle using standard stereotaxic procedures under pentobarbital anesthesia. Coordinates for the cannula placement were 0.8 mm posterior to bregma, 1.4 mm lateral to midline, and
5.0 mm below skull surface. Rats were randomly divided into three groups and injected ICV with myo-inositol or artificial cerebrospinal fluid (CSF) via an injection cannula attached with polyethelene tubing to a 100-ml Koehln microsyringe. As a control, the inositol stereoisomer, L-chiro-inositol was injected in a third group of rats. Myo- and L-chiro-inositol were injected in a dose of 10 mg/40 m1, and CSF was injected in a volume of 40 ml. Injections were made manually over a period of 2 min, and the injection cannula was left in place for 1 min before being replaced by the stylet. Then rats were injected with LiCl, 3 meq/kg in a volume of 15 cm3/kg intra peritoneal. Twenty-four hours later, they were reinjected ICV with the same drug they had received the previous day, and 30 min later, they were injected with pilocarpine, 30 mg/kg, sc, or 20 mg/kg sc. The animals were rated for signs of seizure according to a modified version of the scale used by Patel et al. (1988) once every 5 min for 75 min. The scoring was as follows: 0 ¼ no response; 1 ¼ gustatory movements and/or fictive scratching; 2 ¼ tremor; 3 ¼ head bobbing; 4 ¼ forelimb clonus; 5 ¼ rearing, clonus, and falling. In addition, the latency to attain forelimb clonus (score 4) was recorded for each rat. The observer was blind to the treatment condition. Brains were removed and frozen for biochemical analysis and confirmation of cannula site. The frozen brains were cut at the cannula site, and those subjects that did not have cannulae in the lateral ventricle were excluded from behavioral and biochemical analysis. Latency to clonus was significantly prolonged by myoinositol (41.2 min) as compared to artificial CSF (19.5 min) or L-chiro-inositol (16.9 min), (F ¼ 10.48, p ¼ 0.00074), in rats treated with Li and pilocarpine 30 mg/kg. Post-hoc Scheffe tests indicated that there was a significant difference between myo-inositol and L-chiro-inositol (po0.002) and between myo-inositol and artificial CSF (p ¼ 0.005). The seizure score was analyzed by Kruskal Wallis test at each time point. There was a significant difference between the three groups at 20, and 30–45 min (Fig. 1a) for 30 mg/kg pilocarpine. The latency to exhibit clonus in rats treated with Li and pilocarpine, 20 mg/kg (F ¼ 28.35, po0.00001) significantly increased with myo-inositol. Post-hoc Scheffe tests indicated that this difference was significant when myo-inositol was compared to vehicle (po0.00002) and to L-chiroinositol (po0.00002); using Kruskal Wallis, there was a significant effect of myo-inositol at 25–50 and 60–75 min (Fig. 1b). Eight of 16 rats treated with myo-inositol did not attain seizure scores of 4 (clonus) or higher during the 75 min observation period (and until sacrifice); whereas, only one of the 14 vehicle-treated rats did not exhibit clonus. All the rats treated with L-chiro-inositol had clonic seizures. A chisquare test for the number of animals that did not reach stage 4 indicated a significant difference between
ARTICLE IN PRESS R.H. Belmaker, Y. Bersudsky / Neuroscience and Biobehavioral Reviews 31 (2007) 843–849
845
Fig. 1. Score representing the intensity of pilocarpine-induced behaviors (ordinate) as described in the text. Abscissa represents the time in minutes following the injection of pilocarpine. Following Belmaker et al. (1998).
myo-inositol and vehicle groups (po0.01) and between myo-inositol and L-chiro-inositol groups (po0.004). There was no significant difference between the vehicle and L-chiro-inositol groups. There was no effect of myo-inositol in Li+-free rats injected with pilocarpine, 200 mg/kg. These results suggest that myo-inositol depletion is critical to the development of Li–pilocarpine seizures. The ICV injections of myo-inositol elevated cortical myoinositol levels and simultaneously attenuated and prevented Li–pilocarpine seizures. The fact that seizures induced by a high dose of pilocarpine alone were not attenuated by myo-inositol suggests that the myo-inositol reversal is specific to the Li effect. However, hyponatremia can lower brain inositol and hypernatremia can raise brain inositol. We found that induction of low brain inositol by hyponatremia followed by pilocarpine did not cause limbic seizures. Induction of high brain inositol using hypernatremia followed by Li–pilocarpine administration did not reverse limbic seizures. These data support the concept that inositol available for PI synthesis and inositol for osmotic function are sequestered in different cellular pools (Bersudsky et al., 1994a). 2. The epi-inositol problem However, Williams and Jope (1995) found that the stereoisomer epi-inositol, which was not incorporated into PI, also prevents Li–pilocarpine seizures. This unexpected effect of a stereoisomer, which, in contrast to myo-inositol, did not reverse either the teratogenic effect of Li in xenopus oocytes (Busa and Gimlich, 1989) or the suppression of
neuronal firing of suprachiasmatic nucleus cells by Li in vitro (Busa and Gimlich, 1989), suggested that the mechanism of the reversal of behavioral effects of Li by myo-inositol may be unrelated to the metabolism of PI-derived second messengers. Moreover, Williams and Jope (1995) reported that myo-inositol delayed, but did not prevent seizures induced by administration of pilocarpine to rats treated with chronic dietary Li, raising the possibility that myo-inositol may not be relevant to chronic behavioral effects of Li. Following the report by Williams and Jope (1995), we conducted a series of studies in an attempt to define the conditions under which the epi- and myo-inositol isomers reversed Li–pilocarpine seizures (Patishi et al., 1996). Three questions were raised: (1) is epi-inositol effective following acute and chronic Li treatment? (2) Can a low dose of myoinositol, combined with an inactive stereoisomer of inositol, effectively attenuate Li–pilocarpine seizures? Because epi-inositol could conceivably be converted to myo-inositol and may contain up to 10% of the active isomer (WR Sherman, personal communication), it was hypothesized that a low dose of myo-inositol, co-administered with a biologically inactive isomer, may be equally effective in reducing the effects of Li. Myo-inositol is sequestered in various pools, and it has been estimated that only about 15% of cellular myo-inositol is involved in second messenger synthesis (Fain and Berridge, 1979; Sherman, 1991). Conceivably, under conditions of Li-induced inositol depletion, the ‘‘false’’ inositol, might stimulate the release of myo-inositol from a hitherto ‘‘inactive’’ pool. (3) Is myo-inositol effective following chronic Li, as has been previously reported for acute Li treatment?
ARTICLE IN PRESS 846
R.H. Belmaker, Y. Bersudsky / Neuroscience and Biobehavioral Reviews 31 (2007) 843–849
We (Patishi et al., 1996) were able to replicate the results of Williams and Jope (1995), showing that epi-inositol is also effective in attenuating Li–pilocarpine seizures. Both myo-inositol and epi-inositol also effectively prevented Li–pilocarpine seizures following chronic Li (Table 1). Epi-inositol ICV attenuated seizures after both acute and chronic Li. After acute Li, 4 of 6 control rats, but only 1 of 7 rats treated with myo-inositol, and 0 of 7 rats treated with epi-inositol had seizures. Following chronic Li, epi-inositol also reduced the number of rats showing clonus, preventing seizures in three of eight rats (w2 ¼ 3.69, p ¼ 0.055). All of the control rats had seizures. The latency to onset of clonus was significantly longer in epi-inositol-treated rats (59.27 8.04 min, mean7SD) than in control rats (27711.67 min) (t ¼ 5.38, po0.0005). The serum Li levels were 1.247 0.5 mmol/l for the epi-inositol group and 1.2270.4 mmol/l for the control group. The inactive steroisomer L-chiro-inositol, adulterated with a low dose of myo-inositol, did not affect Li–pilocarpine seizures. Two of 12 rats treated with 9 mg l-chiroinositol and 1 mg myo-inositol and 2/10 rats treated with vehicle did not have behavioral seizures. This difference was not statistically significant (w2 ¼ 0.04). The latency to the onset of seizures was 33715.53 min (mean7SD) in the control rats and 42.1712.71 min in the inositol-treated rats. ICV myo-inositol prevented Li–pilocarpine seizures following chronic dietary Li. After chronic Li treatment, only two of seven rats pretreated with ICV myo-inositol had Li–pilocarpine seizures; whereas, all seven of the vehicle-treated rats showed seizures at a mean latency of 22.1478.49 (mean7SD) min. The reduction in the incidence of clonus by myo-inositol was significant (w2 ¼ 7.78, po0.01). The finding that epi-inositol effectively blocked the onset of seizures in rats treated with acute or chronic Li, followed by 20 mg/kg pilocarpine replicates the findings reported by
Table 1 Summary of the effect of ICV myo-inositol and epi-inositol following various regimens of Li administration and 20 mg/kg pilocarpine Li regimen
ICV treatment
No. animals with clonus
Mean+SD latency to onset of clonus (min)
Acute 3 meq/kg
Vehicle Epi-inositol Myo-inositol Vehicle Epi-inositol Vehicle 90% L-chiro-+ 10% myoinositol Vehicle Myo-inositol
4 of 6 0 of 7 1 of 7 8 of 8 5 of 8 8 of 10 10 of 12
NA
7 of 7 2 of 7
22.1478.49 NA
Chronic Dietary—21 days Acute 3 meq/kg
Chronic dietary—21 days
Following Belmaker et al. (1998).
27711.67 59.278.04 42.1712.71
Williams and Jope (1995). However, in contrast to their findings, we (Patishi et al., 1996) found that both myo- and epi-inositol isomers were effective when rats were treated with chronic Li using a lower dose of pilocarpine and a longer interval between the administration of myo-inositol and pilocarpine than they used. It is likely that the absence of blockade of chronic Li–pilocarpine seizures by myoinositol reported by Williams and Jope (1995) can be attributed to differences in the dose of pilocarpine (30 vs. 20 mg/kg) and the interval between administration of ICV inositol and pilocarpine (30 min vs. 1 h). Williams and Jope (1995) found an increased latency to onset of seizures in the myo-inositol-treated group following chronic Li, similar to that reported by our group using a higher dose (30 mg/kg) of pilocarpine and acute Li (Kofman et al., 1993). Although these data support the relevance of inositol depletion to chronic Li treatment, it was unclear why the epi-inositol stereoisomer would be as effective as myoinositol. Epi-inositol does not seem to be a substrate for phosphatidylinositol synthase (Benjamins and Agranoff, 1969) nor, unlike myo-inositol, does it reverse teratogenic effects of Li (Busa and Gimlich, 1989) or Li-induced suppression of suprachiasmatic nucleus cell firing in vitro (Mason and Biello, 1992). It was, therefore, critical to reexamine the claim of Williams and Jope (1995) that epi-inositol is biochemically inactive in the PI cycle. To avoid the confounding effects of exogenous inositols on the specific activity of radiolabeled myo-inositol, the turnover of the PI cycle was determined by measuring levels of tritiated cytidine monophosphorylphosphatidate ([3H]CMP-PA) in intact Chinese hamster ovary cells or cross-chopped slices of rat cerebral cortex (Richards and Belmaker, 1996). These tissues were incubated in physiological buffer containing [3H]cytidine in the absence or presence of increasing concentrations of myo-, epi-, L-chiro-, and scyllo-inositol for 60 min at 37 1C. Carbachol and Li (1 and 10 mM final concentrations, respectively) were then added, and the incubation continued for 30 min. Incorporation of radioactivity into cell membranes was quantified by scintillation spectrometry. In CHO cells, the accumulation of [3H]CMP-PA was 60710 times basal (n ¼ 13) in the presence of carbachol plus Li. Neither scyllo-inositol (up to 10 mM) nor L-chiro-inositol (up to 30 mM) had an effect on the stimulated formation of [3H]CMP-PA; whereas, 25 mM scyllo-inositol slightly inhibited the response (o20%). In contrast, both myo-inositol and epi-inositol concentration-dependently inhibited the accumulation of [3H]CMP-PA induced by carbachol plus Li. Full inhibition was attained with 10 mM myo-inositol; whereas, epi-inositol at 30 mM inhibited response by 8374% (n ¼ 6). Myo-inositol was slightly but significantly (po0.001, unpaired t-test) more potent than epi-inositol: EC50 values were 1.970.2, and 6.170.6 mM, respectively (n ¼ 6 for each). Although a smaller stimulation of [3H]CMP-PA by carbachol plus Li was found in rat cerebral cortex cross-chopped slices (9.871.2 times basal, n ¼ 3), the same order of activity was
ARTICLE IN PRESS R.H. Belmaker, Y. Bersudsky / Neuroscience and Biobehavioral Reviews 31 (2007) 843–849
observed for the four inositol isomers in the two tissues (Richards and Belmaker, 1996). Thus, there is a precise correlation between the effects of the four inositol isomers on PI turnover, as measured by their effects on the accumulation of [3H]CMP-PA in two tissues, with their effects on Li–pilocarpine-induced seizures. Both in vitro and in vivo, L-chiro- and scyllo-inositol had little or no effect; whereas, epi-inositol was less potent but almost as effective as myo-inositol. The perfect rank order of activity of these four inositol isomers in biochemical and behavioral tests strongly suggest a common basis of action. 3. Species differences in susceptibility to Li–pilocarpine seizures Li–pilocarpine seizures are a dramatic, easily quantifiable behavioral effect of Li in rats that can be clearly ascribed to Li inhibition of IMPase. Sherman (1991) raised the question of whether it occurred in other species such as mice as well as rats. Hokin’s group (Lee et al., 1992) studied the effects of cholinergic agonists on PI metabolism in cortical slices from several species, including monkeys, rats and guinea pigs. They concluded that Li–pilocarpine interactions occur only in species with low basal brain inositol levels, such as rats, and do not occur in monkeys or guinea pigs. In vitro the phenomenon can be elicited whenever cellular inositol levels are depleted in a low inositol medium. To test Hokin’s (Lee et al., 1992) hypothesis in vivo, we (Bersudsky et al., 1994b) attempted to elicit Li–pilocarpine seizures in a wide variety of species. Rats (Sprague Dawley, 300 g), mice (ICR25 g), guinea pigs (250 g), rabbits (500 g), chicks (70 g) and goldfish (30 g) were used. Animals were administered LiCl i.p. before the s.c. injection of pilocarpine and were observed for the 2 h following pilocarpine administration. Control groups were administered low dose pilocarpine alone, after preliminary experiments determined pilocarpine doses necessary to elicit seizures without Li. Rats, mice and goldfish show a standard limbic seizure syndrome after Li–pilocarpine and not with pilocarpine alone, with rhythmic contraction, loss of consciousness, fluctuating improvement and eventual death. One guinea pig out of nine with pilocarpine alone and one with Li–pilocarpine seized in a tonic–clonic manner and died immediately. One of two rabbits treated with Li–pilocarpine died without seizures, but neither had a Li–pilocarpine syndrome. Chicks show pilocarpine effects such as salivation, head bobbing and beak movements but there was no difference between pilocarpine vs. Li–pilocarpine; there was no loss of consciousness or rhythmic muscle contraction. Frogs showed no pilocarpine (1 g/kg) or carbachol (100 mg/kg) response up to a very high dose, and no response to Li–pilocarpine or carbachol. The present results suggest that Li–pilocarpine seizures are unique to rats, mice and goldfish, do not occur in all other mammals or birds, or even in other rodents such as
847
guinea pigs. This supports the concept of Hokin (Lee et al., 1992) that Li–pilocarpine interactions are dependent on low baseline inositol present in rat and mouse brain. The fact that goldfish show Li–pilocarpine interactions and guinea pigs do not suggests that the property is not restricted to a specific evolutionary branch of brain development, but is potentially present since at least teleost development (goldfish) but is dependent on specific aspects of brain physiology. It is interesting to speculate on the importance of the above finding for the use of Li–pilocarpine seizures as a model for the testing of Berridge’s inositol depletion theory of Li action. Contraction of guinea pig ileum was a key model for the study of serotonin receptors, even though guinea pig ileum is uniquely suited for these experiments. The underlying biochemical principles were generalizable to other species and other serotonin receptors, including brain. Li–pilocarpine seizures may be such a model with no face validity but useful for mechanistic study. 4. Effect of Li on the physostigmine-induced behavioral syndrome in humans Janowsky et al. (1972a, 1973) and El-Yousef et al. (1973) reported that physostigmine can reverse manic symptoms rapidly but temporarily and can induce depressive symptoms in marijuana-intoxicated subjects. Janowsky et al. (1972b) proposed an adrenergic–cholinergic balance theory of affective disorder, where adrenergic predominance is associated with mania and cholinergic predominance with depression. Physostigmine, an acetylcholinesterase inhibitor, blocks the breakdown of acetylcholine and thus increases cholinergic transmission (Koelle, 1970). We decided to attempt to replicate the work of Janowsky et al. (1972a, 1973) and El-Yousef et al. (1973) by extending his methods to a group of normal subjects. In addition, we decided to compare the behavioral responses of drug-free normal subjects with subjects receiving therapeutic doses of Li, with the aim of investigating whether Li can inhibit the possible ‘‘model depression’’ induced by physostigmine. This study was done (Oppenheim et al., 1979) before Li–pilocarpine seizures were discovered and thus was regarded as safe. It is important to note that in rats physostigmine can elicit seizures in animals pre-treated with Li just like pilocarpine (Honchar et al., 1983) Subjects consisted of five clinically asymptomatic bipolar manic depressive patients from the Li Clinic of Jerusalem Mental Health Center and five drug-free volunteer controls. All subjects gave informed consent and had no history or evidence of physical disease, especially asthma, ulcer or cardiac arrhythmia. Physostigmine was given intravenously through an indwelling catheter. In all but two cases, the dose was standardized at 1.25 mg in a single injection. Mood was rated every 10 min with a Hebrew translation of a rating scale modified from Janowsky et al. (1973) with
ARTICLE IN PRESS 848
R.H. Belmaker, Y. Bersudsky / Neuroscience and Biobehavioral Reviews 31 (2007) 843–849
a maximum retardation–behavioral depression score of 52. In addition, each subject was observed throughout the hour following physostigmine injection by two psychiatrists who made independent observations of the patient’s stream and content of thought, mood and behavior. Every subject exhibited a profound syndrome of psychomotor retardation within 15 min after physostigmine injection. Every subject reported slowed thoughts, apathy, absence of desire for interpersonal contact, and every subject was noted by the observers to exhibit a marked slowness of speech and movement despite efforts to stimulate conversation. There was no difference noted in the quality, depth or length of the physostigmine response between Li-treated and untreated subjects other than the appearance of depressed mood in the above mentioned three Li-treated subjects. There were no symptoms of exaggerated physostigmine response in the Li-treated subjects, and no seizures or loss of consciousness of any kind. Although it is impossible to draw major conclusions based only on one study with a small number of participants and a single dose of the drug, these data suggest that ‘‘Li–pilocarpine’’ seizures do not occur in humans, fortunately for the subjects and the investigators. 5. Conclusions Li–pilocarpine seizures clearly have no face validity as a model of mania or depression or bipolar disorder. Indeed much of the literature on these seizures uses them as a model of status epilepticus or temporal lobe epilepsy. Seizure models in psychiatry are of course influenced by the fact that anticonvulsants are one of the three major groups of mood stabilizers today (the others being Li or atypical antipsychotics). However, these anticonvulsants are if anything less effective in Li–pilocarpine seizures than in seizures-induced electrically or with GABA antagonists; and the fact that Li–pilocarpine seizures are induced by Li further reduces the face validity of this model. Moreover, Li–pilocarpine seizures appear not to occur in many species other than in rodents and almost certainly do not occur in humans. In what sense then can Li–pilocarpine seizures be useful to the study of bipolar disorder and the understanding of mood stabilizers? First of all, the strong stereospecific evidence that inositol depletion is the mechanism of Li–pilocarpine seizures allows these seizures to be a mammalian in vivo bioassay for inositol depletion. The inositol depletion hypothesis of Li action has recently received support from the study of Williams et al. (2002, 2004) who studied dorsal sensory root ganglia in culture. These cells present a behavior in culture of spreading of their growth cones. This behavior is altered by Li and other anticonvulsant mood stabilizers such as valproate and carbamazapine. The Li and valproate effects are reversed by myo-inositol but not by epi-inositol (Shaltiel et al., 2007). These data support the idea that this Li effect is dependent on IMPase inhibition and therefore depletion of
inositol; however the behavior deals with cells in vitro. The Li–pilocarpine seizure whole animal in vivo model adds to this concept and could be seen as a bridge between cellular models and true behavioral models with face validity. Most recently, studies of knockout mice have used Li–pilocarpine seizures in the sense that they have looked at pilocarpine sensitivity in mice whose inositol level was depressed by homozygous knockout of IMPase-1 (Cryns et al., 2007) or sodium-myo-inositol transporter-1 (Bersudsky et al., in press). In such mice a dose of pilocarpine that causes minimal cholinergic effects in normal mice gives the same picture of status epilepticus and limbic seizures as would be given by Li pre-treatment. These mice in addition to such pilocarpine sensitivity have other behaviors such as decreased immobility in the Porsolt forced swim test (Bersudsky et al., in press; Cryns et al., 2007) that suggests that their inositol depletion has Li-like effects. Thus, the Li–pilocarpine seizure model may, despite its lack of face validity, be a biochemical marker for a model of mania treatment in animals. References Allison, J.H., Stewart, M.A., 1971. Reduced brain inositol in lithiumtreated rats. Nature New Biology 233, 267–268. Allison, J.H., Blisner, M.E., Holland, W.H., Hipps, P.P., Sherman, W.R., 1976. Increased brain myo-inositol 1-phosphate in lithium-treated rats. Biochemical Biophysical Research Communications 71, 664–670. Allison, J.H., Boshans, R.L., Hallcher, L.M., Packman, P.M., Sherman, W.R., 1980. The effects of lithium on myo-inositol levels in layers of frontal cerebral cortex, in cerebellum, and in corpus callosum of the rat. Journal of Neurochemistry 34, 456–458. Belmaker, R.H., Bersudsky, Y., Agam, G., Levine, J., Kofman, O., 1996. How does lithium work on manic depression? Clinical and psychological correlates of the inositol theory. Annual Review of Medicine 47, 47–56. Belmaker, R.H., Agam, G., van Calker, D., Richards, M.H., Kofman, O., 1998. Behavioral reversal of lithium effects by four inositol isomers correlates perfectly with biochemical effects on the PI cycle: depletion by chronic lithium of brain inositol is specific to hypothalamus, and inositol levels may be abnormal in postmortem brain from bipolar patients. Neuropsychopharmacology 19, 220–232. Benjamins, J.A., Agranoff, B.W., 1969. Distribution and properties of CDP-diglyceride:inositol transferase from brain. Journal of Neurochemistry 16, 513–527. Berridge, M.J., 1985. Phosphoinositides and signal transductions. Review in Clinical and Basic Pharmacology 5 (Suppl.), 5S–13S. Bersudsky, Y., Kaplan, Z., Shapiro, Y., Agam, G., Kofman, O., Belmaker, R.H., 1994a. Behavioral evidence for the existence of two pools of cellular inositol. European Neuropsychopharmacology 4, 463–467. Bersudsky, Y., Mahler, O., Kofman, O., Belmaker, R.H., 1994b. Species differences in susceptibilty to Li-pilocarpine seizures. European Neuropsychopharmacology 4, 429–430. Bersudsky, Y., Shaldubina, A., Agam, G., Berry, G.T., Belmaker, R.H., Homozygote inositol transporter knockout mice show a lithium-like phenotype. Bipolar Disorders. Busa, W.B., Gimlich, R.L., 1989. Lithium-induced teratogenesis in frog embryos prevented by a polyphosphoinositide cycle intermediate or a diacylglycerol analog. Developmental Biology 132, 315–324. Cryns, K., Shamir, A., Levi, I., Daneels, G., Goris, I., Delille, P., Bouwknecht, A., Kass, S., Agam, G., Belmaker, R.H., Bersudsky, Y., Steckler, T., Moechars, D., 2007. IMPA1 is essential for embryonic
ARTICLE IN PRESS R.H. Belmaker, Y. Bersudsky / Neuroscience and Biobehavioral Reviews 31 (2007) 843–849 development and lithium-like pilocarpine sensitivity. Neuropsychopharmacology, in press. El-Yousef, M.K., Janowsky, D.S., Davis, J.M., Rosenblatt, J.E., 1973. Induction of severe depression by physostigmine in marihuana intoxicated individuals. British Journal on Addiction Alcoholism and Other Drugs 68, 321–325. Fain, J.N., Berridge, M.J., 1979. Relationship between phosphatidylinositol synthesis and recovery of 5-hydroxytryptamine-responsive Ca2+ flux in blowfly salivary glands. Biochemistry Journal 180, 655–661. Hallcher, L.M., Sherman, W.R., 1980. The effect of lithium ion and other agents on the activity of myo-inositol-1-phosphatase from bovine brain. Journal of Biological Chemistry 255, 10896–10901. Honchar, M.P., Olney, J.W., Sherman, W.R., 1983. Systemic cholinergic agents induce seizures and brain damage in lithium-treated rats. Science 220, 323–325. Honchar, M.P., Ackermann, K.E., Sherman, W.R., 1989. Chronically administered lithium alters neither myo-inositol monophosphatase activity nor phosphoinositide levels in rat brain. Journal of Neurochemistry 53, 590–594. Janowsky, D.S., el-Yousef, M.K., Davis, J.M., Hubbard, B., Sekerke, H.J., 1972a. Cholinergic reversal of manic symptoms. Lancet 1, 1236–1237. Janowsky, D.S., el-Yousef, M.K., Davis, J.M., Sekerke, H.J., 1972b. A cholinergic-adrenergic hypothesis of mania and depression. Lancet 2, 632–635. Janowsky, D.S., el-Yousef, K., Davis, J.M., Sekerke, H.J., 1973. Parasympathetic suppression of manic symptoms by physostigmine. Archives of General Psychiatry 28, 542–547. Koelle, G.B., 1970. Anticholinesterase agents. In: Goodman, L.S., Gillman, A. (Eds.), The Pharmacological Basis of Therapeutics. Macmillian, New York, pp. 442–465. Kofman, O., Belmaker, R.H., 1993. Ziskind-Somerfeld Research Award 1993. Biochemical, behavioral, and clinical studies of the role of inositol in lithium treatment and depression. Biological Psychiatry 34, 839–852. Kofman, O., Sherman, W.R., Katz, V., Belmaker, R.H., 1993. Restoration of brain myo-inositol levels in rats increases latency to lithiumpilocarpine seizures. Psychopharmacology (Berl.) 110, 229–234. Lee, C.H., Dixon, J.F., Reichman, M., Moummi, C., Los, G., Hokin, L.E., 1992. Li+ increases accumulation of inositol 1,4,5-trisphosphate and inositol 1,3,4,5-tetrakisphosphate in cholinergically stimulated brain cortex slices in guinea pig, mouse and rat. The increases require inositol supplementation in mouse and rat but not in guinea pig. Biochemistry Journal 282 (Pt 2), 377–385. Mason, R., Biello, S.M., 1992. A neurophysiological study of a lithium-sensitive phosphoinositide system in the hamster suprachiasmatic (SCN) biological clock in vitro. Neuroscience Letters 144, 135–138.
849
Oppenheim, G., Ebstein, R.P., Belmaker, R.H., 1979. Effect of lithium on the physostigmine-induced behavioral syndrome and plasma cyclic GMP. Journal of Psychiatric Research 15, 133–138. Patel, S., Meldrum, B.S., Fine, A., 1988. Susceptibility to pilocarpineinduced seizures in rats increases with age. Behavioural Brain Research 31, 165–167. Patishi, Y., Belmaker, R.H., Bersudsky, Y., Kofman, O., 1996. A comparison of the ability of myo-inositol and epi-inositol to attenuate lithium-pilocarpine seizures in rats. Biological Psychiatry 39, 829–832. Richards, M.H., Belmaker, R.H., 1996. Epi-inositol is biochemically active in reversing lithium effects on cytidine monophosphorylphosphatidate (CMP-PA). Short communication. Journal of Neural Transmission 103, 1281–1285. Shaltiel, G., Dalton, C.E., Belmaker, R.H., Harwood, J.A., Agam, G., 2007. Specificity of mood stabilizer action on neuronal growth cones. Bipolar Disorders 9, 281–289. Sherman, W.R., 1991. Lithium and the phosphoinositide signalling system. In: Birch, N.J. (Ed.), Lithium and the Cell: Pharmacology and Biochemistry. Academic Press, London, pp. 121–157. Sherman, W.R., Leavitt, A.L., Honchar, M.P., Hallcher, L.M., Phillips, B.E., 1981. Evidence that lithium alters phosphoinositide metabolism: chronic administration elevates primarily D-myo-inositol-1-phosphate in cerebral cortex of the rat. Journal of Neurochemistry 36, 1947–1951. Sherman, W.R., Honchar, M.P., Munsell, L.Y., 1985a. Dectection of receptor-linked phophoinositide metabolism in brain of lithiumtreated rats. In: Bleasdale, J.E., et al. (Eds.), Inositol and Phosphoinositides: Metabolism and Regulation. Humana Press, Clifton, NJ, pp. 49–65. Sherman, W.R., Munsell, L.Y., Gish, B.G., Honchar, M.P., 1985b. Effects of systemically administered lithium on phosphoinositide metabolism in rat brain, kidney, and testis. Journal of Neurochemistry 44, 798–807. Sherman, W.R., Gish, B.G., Honchar, M.P., Munsell, L.Y., 1986. Effects of lithium on phosphoinositide metabolism in vivo. Federation Proceedingss 45, 2639–2646. Spector, R., Lorenzo, A.V., 1975. Myo-inositol transport through the blood brain barrier. Neurochemical Research 13, 785–787. Tricklebank, M.D., Singh, L., Jackson, A., Oles, R.J., 1991. Evidence that a proconvulsant action of lithium is mediated by inhibition of myoinositol phosphatase in mouse brain. Brain Research 558, 145–148. Williams, M.B., Jope, R.S., 1995. Modulation by inositol of cholinergicand serotonergic-induced seizures in lithium-treated rats. Brain Research 685, 169–178. Williams, R.S., Cheng, L., Mudge, A.W., Harwood, A.J., 2002. A common mechanism of action for three mood-stabilizing drugs. Nature 417, 292–295. Williams, R., Ryves, W.J., Dalton, E.C., Eickholt, B., Shaltiel, G., Agam, G., Harwood, A.J., 2004. A molecular cell biology of lithium. Biochemical Society Transactions 32, 799–802.
Neuroscience and Biobehavioral Reviews 31 (2007) 843–849 www.elsevier.com/locate/neubiorev
Review
Lithium–pilocarpine seizures as a model for lithium action in mania R.H. Belmaker, Yuly Bersudsky Ben Gurion University of the Negev, Beersheva Mental Health Center, P.O. Box 4600, Beersheva, Israel
Abstract Lithium (Li) pre-treatment of rats or mice given low dose pilocarpine induces a unique limbic seizure syndrome. This syndrome is stereospecifically reversed by myo-inositol, which suggests that it is a behavioral model for Li depletion of brain inositol. However, this syndrome has little face validity because seizures are not a component of bipolar disorder. Moreover, other animal species that maintain higher brain inositol levels than mice or rats do not show Li–pilocarpine seizures and a study in humans suggests that humans do not show this syndrome as well. It could be suggested that Li–pilocarpine seizures are an in vivo bioassay for inositol depletion. Recent studies of knockout mice lacking inositol monophosphatase-1 or the sodium myo-inositol transporter-1 found that both these knockout mice given pilocarpine develop limbic seizures as if they had been pre-treated with Li. These mice in addition to such pilocarpine sensitivity have other behaviors such as decreased immobility in the Porsolt forced swim test that suggests that their inositol depletion has Li-like effects. Thus, the Li–pilocarpine seizure model may, despite its lack of face validity, be a biochemical marker for a model of mania treatment in animals. r 2007 Elsevier Ltd. All rights reserved. Keywords: Li–pilocarpine; Seizures; Inositol; Mania
Contents 1. 2. 3. 4. 5.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The epi-inositol problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Species differences in susceptibility to Li–pilocarpine seizures . . . . . . . . . . Effect of Li on the physostigmine-induced behavioral syndrome in humans Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1. Introduction A widespread hypothesis explaining Li’s therapeutic and prophylactic effect in affective disorder is that inhibition of inositol monophosphatase (IMPase) impairs the operation of the phosphatidylinositol cycle (PI cycle). The membrane phospholipid, phosphatidylinositol (PI) is sequentially phosphorylated to form phosphatidylinositol bisphosphate (PIP2). Agonist-stimulated phospholipase C (PLC) cleaves Corresponding author. Tel.: +972 640 1602; fax: +972 8 640 1621.
E-mail address: [email protected] (R.H. Belmaker). 0149-7634/$ - see front matter r 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.neubiorev.2007.05.001
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
843 845 847 847 848 848
PIP2 to two second messengers: inositol 1,4,5-trisphosphate (IP3) and diacylglycerol (DAG). IP3 releases calcium sequestered in endoplasmic reticulum. IP3 is subsequently dephosphorylated to inositol monophosphate (IP), which is dephosphorylated by IMPase to free inositol. DAG, the second derivative of PIP2 activates protein kinase C. DAG is converted sequentially to cytidine disphosphate-diacyl glycerol (CDP-DG), which is combined with free inositol by PI synthase to re-form PI. Li was first shown to affect the system in 1971 (Allison and Stewart, 1971). Li inhibits the dephosphorylation of four of the inositol monophosphates as well as two inositol
ARTICLE IN PRESS 844
R.H. Belmaker, Y. Bersudsky / Neuroscience and Biobehavioral Reviews 31 (2007) 843–849
biphosphates, thereby increasing brain levels of IP and two biphosphates and reducing levels of myo-inositol (Allison et al., 1976, 1980; Sherman et al., 1981, 1985b; Honchar et al., 1989). With discovery of the importance of the cycle as a second messenger system, the Li-induced reduction of inositol immediately assumed potential importance as a key mechanism of Li action (Berridge, 1985). Hallcher and Sherman (1980) showed that Li at therapeutic doses inhibits rat brain IMPase, thereby explaining the reduction in inositol and accumulation of IP. Inhibition is uncompetitive with a Ki of 0.8 mM, thus within therapeutically effective serum concentrations in the range of 0.5–1.4 mM (Hallcher and Sherman, 1980). In rats, Li chloride (LiCl), 10 meq/kg, reduced brain inositol levels by 30% and increased IP levels 20-fold after 6 h (Allison et al., 1976, 1980) and 40-fold after 24 h (Sherman et al., 1985b). Chronic administration of Li has been reported to reduce brain inositol levels (Sherman et al., 1985b), but others have not replicated this critical finding (Jope and Williams, 1994). Berridge (1985) proposed that the Li-induced shortage of inositol in the brain, an organ to which plasma-borne inositol is essentially unavailable, leads to depletion of substrate for phosphatidylinositol resynthesis only in overactive neurons. This theory could explain Li’s specific therapeutic activity in psychopathological states with minimal effects on normal behavior (Belmaker et al., 1996). Various biological effects of Li can be reversed by addition of myo-inositol in vitro (Kofman and Belmaker, 1993). To ascertain if Li’s inhibition of IMPase is relevant to its therapeutic effects in patients with affective disorders, it is critical to demonstrate that behavioral effects of Li are also reversed by myo-inositol. Because myo-inositol does not easily penetrate the blood brain barrier when injected systemically (Spector and Lorenzo, 1975), it is often necessary to inject myo-inositol directly into the brain. One of the most robust behavioral effects of Li is that normally subconvulsant doses of muscarinic agonists will induce limbic seizures in rats pretreated with Li (Honchar et al., 1983). Induction of Li–pilocarpine seizures is concomitant with a reduction in cortical myo-inositol levels and an elevation of IP, which is about 10-fold greater than the effects elicited by either Li or pilocarpine alone (Sherman et al., 1985a, 1986). Tricklebank et al. (1991) reported that ICV injections of myo-inositol prolonged the latency to seizures elicited by ICV or systemic Li in mice. We found similar results in rats (Kofman et al., 1993). This effect is dramatic, and many inositol-treated animals do not seize at all (Kofman and Belmaker, 1993). For example, in our laboratory (Kofman and Belmaker, 1993; Belmaker et al., 1998) 28 male Sprague Dawley rats were implanted with guide cannulae in the lateral ventricle using standard stereotaxic procedures under pentobarbital anesthesia. Coordinates for the cannula placement were 0.8 mm posterior to bregma, 1.4 mm lateral to midline, and
5.0 mm below skull surface. Rats were randomly divided into three groups and injected ICV with myo-inositol or artificial cerebrospinal fluid (CSF) via an injection cannula attached with polyethelene tubing to a 100-ml Koehln microsyringe. As a control, the inositol stereoisomer, L-chiro-inositol was injected in a third group of rats. Myo- and L-chiro-inositol were injected in a dose of 10 mg/40 m1, and CSF was injected in a volume of 40 ml. Injections were made manually over a period of 2 min, and the injection cannula was left in place for 1 min before being replaced by the stylet. Then rats were injected with LiCl, 3 meq/kg in a volume of 15 cm3/kg intra peritoneal. Twenty-four hours later, they were reinjected ICV with the same drug they had received the previous day, and 30 min later, they were injected with pilocarpine, 30 mg/kg, sc, or 20 mg/kg sc. The animals were rated for signs of seizure according to a modified version of the scale used by Patel et al. (1988) once every 5 min for 75 min. The scoring was as follows: 0 ¼ no response; 1 ¼ gustatory movements and/or fictive scratching; 2 ¼ tremor; 3 ¼ head bobbing; 4 ¼ forelimb clonus; 5 ¼ rearing, clonus, and falling. In addition, the latency to attain forelimb clonus (score 4) was recorded for each rat. The observer was blind to the treatment condition. Brains were removed and frozen for biochemical analysis and confirmation of cannula site. The frozen brains were cut at the cannula site, and those subjects that did not have cannulae in the lateral ventricle were excluded from behavioral and biochemical analysis. Latency to clonus was significantly prolonged by myoinositol (41.2 min) as compared to artificial CSF (19.5 min) or L-chiro-inositol (16.9 min), (F ¼ 10.48, p ¼ 0.00074), in rats treated with Li and pilocarpine 30 mg/kg. Post-hoc Scheffe tests indicated that there was a significant difference between myo-inositol and L-chiro-inositol (po0.002) and between myo-inositol and artificial CSF (p ¼ 0.005). The seizure score was analyzed by Kruskal Wallis test at each time point. There was a significant difference between the three groups at 20, and 30–45 min (Fig. 1a) for 30 mg/kg pilocarpine. The latency to exhibit clonus in rats treated with Li and pilocarpine, 20 mg/kg (F ¼ 28.35, po0.00001) significantly increased with myo-inositol. Post-hoc Scheffe tests indicated that this difference was significant when myo-inositol was compared to vehicle (po0.00002) and to L-chiroinositol (po0.00002); using Kruskal Wallis, there was a significant effect of myo-inositol at 25–50 and 60–75 min (Fig. 1b). Eight of 16 rats treated with myo-inositol did not attain seizure scores of 4 (clonus) or higher during the 75 min observation period (and until sacrifice); whereas, only one of the 14 vehicle-treated rats did not exhibit clonus. All the rats treated with L-chiro-inositol had clonic seizures. A chisquare test for the number of animals that did not reach stage 4 indicated a significant difference between
ARTICLE IN PRESS R.H. Belmaker, Y. Bersudsky / Neuroscience and Biobehavioral Reviews 31 (2007) 843–849
845
Fig. 1. Score representing the intensity of pilocarpine-induced behaviors (ordinate) as described in the text. Abscissa represents the time in minutes following the injection of pilocarpine. Following Belmaker et al. (1998).
myo-inositol and vehicle groups (po0.01) and between myo-inositol and L-chiro-inositol groups (po0.004). There was no significant difference between the vehicle and L-chiro-inositol groups. There was no effect of myo-inositol in Li+-free rats injected with pilocarpine, 200 mg/kg. These results suggest that myo-inositol depletion is critical to the development of Li–pilocarpine seizures. The ICV injections of myo-inositol elevated cortical myoinositol levels and simultaneously attenuated and prevented Li–pilocarpine seizures. The fact that seizures induced by a high dose of pilocarpine alone were not attenuated by myo-inositol suggests that the myo-inositol reversal is specific to the Li effect. However, hyponatremia can lower brain inositol and hypernatremia can raise brain inositol. We found that induction of low brain inositol by hyponatremia followed by pilocarpine did not cause limbic seizures. Induction of high brain inositol using hypernatremia followed by Li–pilocarpine administration did not reverse limbic seizures. These data support the concept that inositol available for PI synthesis and inositol for osmotic function are sequestered in different cellular pools (Bersudsky et al., 1994a). 2. The epi-inositol problem However, Williams and Jope (1995) found that the stereoisomer epi-inositol, which was not incorporated into PI, also prevents Li–pilocarpine seizures. This unexpected effect of a stereoisomer, which, in contrast to myo-inositol, did not reverse either the teratogenic effect of Li in xenopus oocytes (Busa and Gimlich, 1989) or the suppression of
neuronal firing of suprachiasmatic nucleus cells by Li in vitro (Busa and Gimlich, 1989), suggested that the mechanism of the reversal of behavioral effects of Li by myo-inositol may be unrelated to the metabolism of PI-derived second messengers. Moreover, Williams and Jope (1995) reported that myo-inositol delayed, but did not prevent seizures induced by administration of pilocarpine to rats treated with chronic dietary Li, raising the possibility that myo-inositol may not be relevant to chronic behavioral effects of Li. Following the report by Williams and Jope (1995), we conducted a series of studies in an attempt to define the conditions under which the epi- and myo-inositol isomers reversed Li–pilocarpine seizures (Patishi et al., 1996). Three questions were raised: (1) is epi-inositol effective following acute and chronic Li treatment? (2) Can a low dose of myoinositol, combined with an inactive stereoisomer of inositol, effectively attenuate Li–pilocarpine seizures? Because epi-inositol could conceivably be converted to myo-inositol and may contain up to 10% of the active isomer (WR Sherman, personal communication), it was hypothesized that a low dose of myo-inositol, co-administered with a biologically inactive isomer, may be equally effective in reducing the effects of Li. Myo-inositol is sequestered in various pools, and it has been estimated that only about 15% of cellular myo-inositol is involved in second messenger synthesis (Fain and Berridge, 1979; Sherman, 1991). Conceivably, under conditions of Li-induced inositol depletion, the ‘‘false’’ inositol, might stimulate the release of myo-inositol from a hitherto ‘‘inactive’’ pool. (3) Is myo-inositol effective following chronic Li, as has been previously reported for acute Li treatment?
ARTICLE IN PRESS 846
R.H. Belmaker, Y. Bersudsky / Neuroscience and Biobehavioral Reviews 31 (2007) 843–849
We (Patishi et al., 1996) were able to replicate the results of Williams and Jope (1995), showing that epi-inositol is also effective in attenuating Li–pilocarpine seizures. Both myo-inositol and epi-inositol also effectively prevented Li–pilocarpine seizures following chronic Li (Table 1). Epi-inositol ICV attenuated seizures after both acute and chronic Li. After acute Li, 4 of 6 control rats, but only 1 of 7 rats treated with myo-inositol, and 0 of 7 rats treated with epi-inositol had seizures. Following chronic Li, epi-inositol also reduced the number of rats showing clonus, preventing seizures in three of eight rats (w2 ¼ 3.69, p ¼ 0.055). All of the control rats had seizures. The latency to onset of clonus was significantly longer in epi-inositol-treated rats (59.27 8.04 min, mean7SD) than in control rats (27711.67 min) (t ¼ 5.38, po0.0005). The serum Li levels were 1.247 0.5 mmol/l for the epi-inositol group and 1.2270.4 mmol/l for the control group. The inactive steroisomer L-chiro-inositol, adulterated with a low dose of myo-inositol, did not affect Li–pilocarpine seizures. Two of 12 rats treated with 9 mg l-chiroinositol and 1 mg myo-inositol and 2/10 rats treated with vehicle did not have behavioral seizures. This difference was not statistically significant (w2 ¼ 0.04). The latency to the onset of seizures was 33715.53 min (mean7SD) in the control rats and 42.1712.71 min in the inositol-treated rats. ICV myo-inositol prevented Li–pilocarpine seizures following chronic dietary Li. After chronic Li treatment, only two of seven rats pretreated with ICV myo-inositol had Li–pilocarpine seizures; whereas, all seven of the vehicle-treated rats showed seizures at a mean latency of 22.1478.49 (mean7SD) min. The reduction in the incidence of clonus by myo-inositol was significant (w2 ¼ 7.78, po0.01). The finding that epi-inositol effectively blocked the onset of seizures in rats treated with acute or chronic Li, followed by 20 mg/kg pilocarpine replicates the findings reported by
Table 1 Summary of the effect of ICV myo-inositol and epi-inositol following various regimens of Li administration and 20 mg/kg pilocarpine Li regimen
ICV treatment
No. animals with clonus
Mean+SD latency to onset of clonus (min)
Acute 3 meq/kg
Vehicle Epi-inositol Myo-inositol Vehicle Epi-inositol Vehicle 90% L-chiro-+ 10% myoinositol Vehicle Myo-inositol
4 of 6 0 of 7 1 of 7 8 of 8 5 of 8 8 of 10 10 of 12
NA
7 of 7 2 of 7
22.1478.49 NA
Chronic Dietary—21 days Acute 3 meq/kg
Chronic dietary—21 days
Following Belmaker et al. (1998).
27711.67 59.278.04 42.1712.71
Williams and Jope (1995). However, in contrast to their findings, we (Patishi et al., 1996) found that both myo- and epi-inositol isomers were effective when rats were treated with chronic Li using a lower dose of pilocarpine and a longer interval between the administration of myo-inositol and pilocarpine than they used. It is likely that the absence of blockade of chronic Li–pilocarpine seizures by myoinositol reported by Williams and Jope (1995) can be attributed to differences in the dose of pilocarpine (30 vs. 20 mg/kg) and the interval between administration of ICV inositol and pilocarpine (30 min vs. 1 h). Williams and Jope (1995) found an increased latency to onset of seizures in the myo-inositol-treated group following chronic Li, similar to that reported by our group using a higher dose (30 mg/kg) of pilocarpine and acute Li (Kofman et al., 1993). Although these data support the relevance of inositol depletion to chronic Li treatment, it was unclear why the epi-inositol stereoisomer would be as effective as myoinositol. Epi-inositol does not seem to be a substrate for phosphatidylinositol synthase (Benjamins and Agranoff, 1969) nor, unlike myo-inositol, does it reverse teratogenic effects of Li (Busa and Gimlich, 1989) or Li-induced suppression of suprachiasmatic nucleus cell firing in vitro (Mason and Biello, 1992). It was, therefore, critical to reexamine the claim of Williams and Jope (1995) that epi-inositol is biochemically inactive in the PI cycle. To avoid the confounding effects of exogenous inositols on the specific activity of radiolabeled myo-inositol, the turnover of the PI cycle was determined by measuring levels of tritiated cytidine monophosphorylphosphatidate ([3H]CMP-PA) in intact Chinese hamster ovary cells or cross-chopped slices of rat cerebral cortex (Richards and Belmaker, 1996). These tissues were incubated in physiological buffer containing [3H]cytidine in the absence or presence of increasing concentrations of myo-, epi-, L-chiro-, and scyllo-inositol for 60 min at 37 1C. Carbachol and Li (1 and 10 mM final concentrations, respectively) were then added, and the incubation continued for 30 min. Incorporation of radioactivity into cell membranes was quantified by scintillation spectrometry. In CHO cells, the accumulation of [3H]CMP-PA was 60710 times basal (n ¼ 13) in the presence of carbachol plus Li. Neither scyllo-inositol (up to 10 mM) nor L-chiro-inositol (up to 30 mM) had an effect on the stimulated formation of [3H]CMP-PA; whereas, 25 mM scyllo-inositol slightly inhibited the response (o20%). In contrast, both myo-inositol and epi-inositol concentration-dependently inhibited the accumulation of [3H]CMP-PA induced by carbachol plus Li. Full inhibition was attained with 10 mM myo-inositol; whereas, epi-inositol at 30 mM inhibited response by 8374% (n ¼ 6). Myo-inositol was slightly but significantly (po0.001, unpaired t-test) more potent than epi-inositol: EC50 values were 1.970.2, and 6.170.6 mM, respectively (n ¼ 6 for each). Although a smaller stimulation of [3H]CMP-PA by carbachol plus Li was found in rat cerebral cortex cross-chopped slices (9.871.2 times basal, n ¼ 3), the same order of activity was
ARTICLE IN PRESS R.H. Belmaker, Y. Bersudsky / Neuroscience and Biobehavioral Reviews 31 (2007) 843–849
observed for the four inositol isomers in the two tissues (Richards and Belmaker, 1996). Thus, there is a precise correlation between the effects of the four inositol isomers on PI turnover, as measured by their effects on the accumulation of [3H]CMP-PA in two tissues, with their effects on Li–pilocarpine-induced seizures. Both in vitro and in vivo, L-chiro- and scyllo-inositol had little or no effect; whereas, epi-inositol was less potent but almost as effective as myo-inositol. The perfect rank order of activity of these four inositol isomers in biochemical and behavioral tests strongly suggest a common basis of action. 3. Species differences in susceptibility to Li–pilocarpine seizures Li–pilocarpine seizures are a dramatic, easily quantifiable behavioral effect of Li in rats that can be clearly ascribed to Li inhibition of IMPase. Sherman (1991) raised the question of whether it occurred in other species such as mice as well as rats. Hokin’s group (Lee et al., 1992) studied the effects of cholinergic agonists on PI metabolism in cortical slices from several species, including monkeys, rats and guinea pigs. They concluded that Li–pilocarpine interactions occur only in species with low basal brain inositol levels, such as rats, and do not occur in monkeys or guinea pigs. In vitro the phenomenon can be elicited whenever cellular inositol levels are depleted in a low inositol medium. To test Hokin’s (Lee et al., 1992) hypothesis in vivo, we (Bersudsky et al., 1994b) attempted to elicit Li–pilocarpine seizures in a wide variety of species. Rats (Sprague Dawley, 300 g), mice (ICR25 g), guinea pigs (250 g), rabbits (500 g), chicks (70 g) and goldfish (30 g) were used. Animals were administered LiCl i.p. before the s.c. injection of pilocarpine and were observed for the 2 h following pilocarpine administration. Control groups were administered low dose pilocarpine alone, after preliminary experiments determined pilocarpine doses necessary to elicit seizures without Li. Rats, mice and goldfish show a standard limbic seizure syndrome after Li–pilocarpine and not with pilocarpine alone, with rhythmic contraction, loss of consciousness, fluctuating improvement and eventual death. One guinea pig out of nine with pilocarpine alone and one with Li–pilocarpine seized in a tonic–clonic manner and died immediately. One of two rabbits treated with Li–pilocarpine died without seizures, but neither had a Li–pilocarpine syndrome. Chicks show pilocarpine effects such as salivation, head bobbing and beak movements but there was no difference between pilocarpine vs. Li–pilocarpine; there was no loss of consciousness or rhythmic muscle contraction. Frogs showed no pilocarpine (1 g/kg) or carbachol (100 mg/kg) response up to a very high dose, and no response to Li–pilocarpine or carbachol. The present results suggest that Li–pilocarpine seizures are unique to rats, mice and goldfish, do not occur in all other mammals or birds, or even in other rodents such as
847
guinea pigs. This supports the concept of Hokin (Lee et al., 1992) that Li–pilocarpine interactions are dependent on low baseline inositol present in rat and mouse brain. The fact that goldfish show Li–pilocarpine interactions and guinea pigs do not suggests that the property is not restricted to a specific evolutionary branch of brain development, but is potentially present since at least teleost development (goldfish) but is dependent on specific aspects of brain physiology. It is interesting to speculate on the importance of the above finding for the use of Li–pilocarpine seizures as a model for the testing of Berridge’s inositol depletion theory of Li action. Contraction of guinea pig ileum was a key model for the study of serotonin receptors, even though guinea pig ileum is uniquely suited for these experiments. The underlying biochemical principles were generalizable to other species and other serotonin receptors, including brain. Li–pilocarpine seizures may be such a model with no face validity but useful for mechanistic study. 4. Effect of Li on the physostigmine-induced behavioral syndrome in humans Janowsky et al. (1972a, 1973) and El-Yousef et al. (1973) reported that physostigmine can reverse manic symptoms rapidly but temporarily and can induce depressive symptoms in marijuana-intoxicated subjects. Janowsky et al. (1972b) proposed an adrenergic–cholinergic balance theory of affective disorder, where adrenergic predominance is associated with mania and cholinergic predominance with depression. Physostigmine, an acetylcholinesterase inhibitor, blocks the breakdown of acetylcholine and thus increases cholinergic transmission (Koelle, 1970). We decided to attempt to replicate the work of Janowsky et al. (1972a, 1973) and El-Yousef et al. (1973) by extending his methods to a group of normal subjects. In addition, we decided to compare the behavioral responses of drug-free normal subjects with subjects receiving therapeutic doses of Li, with the aim of investigating whether Li can inhibit the possible ‘‘model depression’’ induced by physostigmine. This study was done (Oppenheim et al., 1979) before Li–pilocarpine seizures were discovered and thus was regarded as safe. It is important to note that in rats physostigmine can elicit seizures in animals pre-treated with Li just like pilocarpine (Honchar et al., 1983) Subjects consisted of five clinically asymptomatic bipolar manic depressive patients from the Li Clinic of Jerusalem Mental Health Center and five drug-free volunteer controls. All subjects gave informed consent and had no history or evidence of physical disease, especially asthma, ulcer or cardiac arrhythmia. Physostigmine was given intravenously through an indwelling catheter. In all but two cases, the dose was standardized at 1.25 mg in a single injection. Mood was rated every 10 min with a Hebrew translation of a rating scale modified from Janowsky et al. (1973) with
ARTICLE IN PRESS 848
R.H. Belmaker, Y. Bersudsky / Neuroscience and Biobehavioral Reviews 31 (2007) 843–849
a maximum retardation–behavioral depression score of 52. In addition, each subject was observed throughout the hour following physostigmine injection by two psychiatrists who made independent observations of the patient’s stream and content of thought, mood and behavior. Every subject exhibited a profound syndrome of psychomotor retardation within 15 min after physostigmine injection. Every subject reported slowed thoughts, apathy, absence of desire for interpersonal contact, and every subject was noted by the observers to exhibit a marked slowness of speech and movement despite efforts to stimulate conversation. There was no difference noted in the quality, depth or length of the physostigmine response between Li-treated and untreated subjects other than the appearance of depressed mood in the above mentioned three Li-treated subjects. There were no symptoms of exaggerated physostigmine response in the Li-treated subjects, and no seizures or loss of consciousness of any kind. Although it is impossible to draw major conclusions based only on one study with a small number of participants and a single dose of the drug, these data suggest that ‘‘Li–pilocarpine’’ seizures do not occur in humans, fortunately for the subjects and the investigators. 5. Conclusions Li–pilocarpine seizures clearly have no face validity as a model of mania or depression or bipolar disorder. Indeed much of the literature on these seizures uses them as a model of status epilepticus or temporal lobe epilepsy. Seizure models in psychiatry are of course influenced by the fact that anticonvulsants are one of the three major groups of mood stabilizers today (the others being Li or atypical antipsychotics). However, these anticonvulsants are if anything less effective in Li–pilocarpine seizures than in seizures-induced electrically or with GABA antagonists; and the fact that Li–pilocarpine seizures are induced by Li further reduces the face validity of this model. Moreover, Li–pilocarpine seizures appear not to occur in many species other than in rodents and almost certainly do not occur in humans. In what sense then can Li–pilocarpine seizures be useful to the study of bipolar disorder and the understanding of mood stabilizers? First of all, the strong stereospecific evidence that inositol depletion is the mechanism of Li–pilocarpine seizures allows these seizures to be a mammalian in vivo bioassay for inositol depletion. The inositol depletion hypothesis of Li action has recently received support from the study of Williams et al. (2002, 2004) who studied dorsal sensory root ganglia in culture. These cells present a behavior in culture of spreading of their growth cones. This behavior is altered by Li and other anticonvulsant mood stabilizers such as valproate and carbamazapine. The Li and valproate effects are reversed by myo-inositol but not by epi-inositol (Shaltiel et al., 2007). These data support the idea that this Li effect is dependent on IMPase inhibition and therefore depletion of
inositol; however the behavior deals with cells in vitro. The Li–pilocarpine seizure whole animal in vivo model adds to this concept and could be seen as a bridge between cellular models and true behavioral models with face validity. Most recently, studies of knockout mice have used Li–pilocarpine seizures in the sense that they have looked at pilocarpine sensitivity in mice whose inositol level was depressed by homozygous knockout of IMPase-1 (Cryns et al., 2007) or sodium-myo-inositol transporter-1 (Bersudsky et al., in press). In such mice a dose of pilocarpine that causes minimal cholinergic effects in normal mice gives the same picture of status epilepticus and limbic seizures as would be given by Li pre-treatment. These mice in addition to such pilocarpine sensitivity have other behaviors such as decreased immobility in the Porsolt forced swim test (Bersudsky et al., in press; Cryns et al., 2007) that suggests that their inositol depletion has Li-like effects. Thus, the Li–pilocarpine seizure model may, despite its lack of face validity, be a biochemical marker for a model of mania treatment in animals. References Allison, J.H., Stewart, M.A., 1971. Reduced brain inositol in lithiumtreated rats. Nature New Biology 233, 267–268. Allison, J.H., Blisner, M.E., Holland, W.H., Hipps, P.P., Sherman, W.R., 1976. Increased brain myo-inositol 1-phosphate in lithium-treated rats. Biochemical Biophysical Research Communications 71, 664–670. Allison, J.H., Boshans, R.L., Hallcher, L.M., Packman, P.M., Sherman, W.R., 1980. The effects of lithium on myo-inositol levels in layers of frontal cerebral cortex, in cerebellum, and in corpus callosum of the rat. Journal of Neurochemistry 34, 456–458. Belmaker, R.H., Bersudsky, Y., Agam, G., Levine, J., Kofman, O., 1996. How does lithium work on manic depression? Clinical and psychological correlates of the inositol theory. Annual Review of Medicine 47, 47–56. Belmaker, R.H., Agam, G., van Calker, D., Richards, M.H., Kofman, O., 1998. Behavioral reversal of lithium effects by four inositol isomers correlates perfectly with biochemical effects on the PI cycle: depletion by chronic lithium of brain inositol is specific to hypothalamus, and inositol levels may be abnormal in postmortem brain from bipolar patients. Neuropsychopharmacology 19, 220–232. Benjamins, J.A., Agranoff, B.W., 1969. Distribution and properties of CDP-diglyceride:inositol transferase from brain. Journal of Neurochemistry 16, 513–527. Berridge, M.J., 1985. Phosphoinositides and signal transductions. Review in Clinical and Basic Pharmacology 5 (Suppl.), 5S–13S. Bersudsky, Y., Kaplan, Z., Shapiro, Y., Agam, G., Kofman, O., Belmaker, R.H., 1994a. Behavioral evidence for the existence of two pools of cellular inositol. European Neuropsychopharmacology 4, 463–467. Bersudsky, Y., Mahler, O., Kofman, O., Belmaker, R.H., 1994b. Species differences in susceptibilty to Li-pilocarpine seizures. European Neuropsychopharmacology 4, 429–430. Bersudsky, Y., Shaldubina, A., Agam, G., Berry, G.T., Belmaker, R.H., Homozygote inositol transporter knockout mice show a lithium-like phenotype. Bipolar Disorders. Busa, W.B., Gimlich, R.L., 1989. Lithium-induced teratogenesis in frog embryos prevented by a polyphosphoinositide cycle intermediate or a diacylglycerol analog. Developmental Biology 132, 315–324. Cryns, K., Shamir, A., Levi, I., Daneels, G., Goris, I., Delille, P., Bouwknecht, A., Kass, S., Agam, G., Belmaker, R.H., Bersudsky, Y., Steckler, T., Moechars, D., 2007. IMPA1 is essential for embryonic
ARTICLE IN PRESS R.H. Belmaker, Y. Bersudsky / Neuroscience and Biobehavioral Reviews 31 (2007) 843–849 development and lithium-like pilocarpine sensitivity. Neuropsychopharmacology, in press. El-Yousef, M.K., Janowsky, D.S., Davis, J.M., Rosenblatt, J.E., 1973. Induction of severe depression by physostigmine in marihuana intoxicated individuals. British Journal on Addiction Alcoholism and Other Drugs 68, 321–325. Fain, J.N., Berridge, M.J., 1979. Relationship between phosphatidylinositol synthesis and recovery of 5-hydroxytryptamine-responsive Ca2+ flux in blowfly salivary glands. Biochemistry Journal 180, 655–661. Hallcher, L.M., Sherman, W.R., 1980. The effect of lithium ion and other agents on the activity of myo-inositol-1-phosphatase from bovine brain. Journal of Biological Chemistry 255, 10896–10901. Honchar, M.P., Olney, J.W., Sherman, W.R., 1983. Systemic cholinergic agents induce seizures and brain damage in lithium-treated rats. Science 220, 323–325. Honchar, M.P., Ackermann, K.E., Sherman, W.R., 1989. Chronically administered lithium alters neither myo-inositol monophosphatase activity nor phosphoinositide levels in rat brain. Journal of Neurochemistry 53, 590–594. Janowsky, D.S., el-Yousef, M.K., Davis, J.M., Hubbard, B., Sekerke, H.J., 1972a. Cholinergic reversal of manic symptoms. Lancet 1, 1236–1237. Janowsky, D.S., el-Yousef, M.K., Davis, J.M., Sekerke, H.J., 1972b. A cholinergic-adrenergic hypothesis of mania and depression. Lancet 2, 632–635. Janowsky, D.S., el-Yousef, K., Davis, J.M., Sekerke, H.J., 1973. Parasympathetic suppression of manic symptoms by physostigmine. Archives of General Psychiatry 28, 542–547. Koelle, G.B., 1970. Anticholinesterase agents. In: Goodman, L.S., Gillman, A. (Eds.), The Pharmacological Basis of Therapeutics. Macmillian, New York, pp. 442–465. Kofman, O., Belmaker, R.H., 1993. Ziskind-Somerfeld Research Award 1993. Biochemical, behavioral, and clinical studies of the role of inositol in lithium treatment and depression. Biological Psychiatry 34, 839–852. Kofman, O., Sherman, W.R., Katz, V., Belmaker, R.H., 1993. Restoration of brain myo-inositol levels in rats increases latency to lithiumpilocarpine seizures. Psychopharmacology (Berl.) 110, 229–234. Lee, C.H., Dixon, J.F., Reichman, M., Moummi, C., Los, G., Hokin, L.E., 1992. Li+ increases accumulation of inositol 1,4,5-trisphosphate and inositol 1,3,4,5-tetrakisphosphate in cholinergically stimulated brain cortex slices in guinea pig, mouse and rat. The increases require inositol supplementation in mouse and rat but not in guinea pig. Biochemistry Journal 282 (Pt 2), 377–385. Mason, R., Biello, S.M., 1992. A neurophysiological study of a lithium-sensitive phosphoinositide system in the hamster suprachiasmatic (SCN) biological clock in vitro. Neuroscience Letters 144, 135–138.
849
Oppenheim, G., Ebstein, R.P., Belmaker, R.H., 1979. Effect of lithium on the physostigmine-induced behavioral syndrome and plasma cyclic GMP. Journal of Psychiatric Research 15, 133–138. Patel, S., Meldrum, B.S., Fine, A., 1988. Susceptibility to pilocarpineinduced seizures in rats increases with age. Behavioural Brain Research 31, 165–167. Patishi, Y., Belmaker, R.H., Bersudsky, Y., Kofman, O., 1996. A comparison of the ability of myo-inositol and epi-inositol to attenuate lithium-pilocarpine seizures in rats. Biological Psychiatry 39, 829–832. Richards, M.H., Belmaker, R.H., 1996. Epi-inositol is biochemically active in reversing lithium effects on cytidine monophosphorylphosphatidate (CMP-PA). Short communication. Journal of Neural Transmission 103, 1281–1285. Shaltiel, G., Dalton, C.E., Belmaker, R.H., Harwood, J.A., Agam, G., 2007. Specificity of mood stabilizer action on neuronal growth cones. Bipolar Disorders 9, 281–289. Sherman, W.R., 1991. Lithium and the phosphoinositide signalling system. In: Birch, N.J. (Ed.), Lithium and the Cell: Pharmacology and Biochemistry. Academic Press, London, pp. 121–157. Sherman, W.R., Leavitt, A.L., Honchar, M.P., Hallcher, L.M., Phillips, B.E., 1981. Evidence that lithium alters phosphoinositide metabolism: chronic administration elevates primarily D-myo-inositol-1-phosphate in cerebral cortex of the rat. Journal of Neurochemistry 36, 1947–1951. Sherman, W.R., Honchar, M.P., Munsell, L.Y., 1985a. Dectection of receptor-linked phophoinositide metabolism in brain of lithiumtreated rats. In: Bleasdale, J.E., et al. (Eds.), Inositol and Phosphoinositides: Metabolism and Regulation. Humana Press, Clifton, NJ, pp. 49–65. Sherman, W.R., Munsell, L.Y., Gish, B.G., Honchar, M.P., 1985b. Effects of systemically administered lithium on phosphoinositide metabolism in rat brain, kidney, and testis. Journal of Neurochemistry 44, 798–807. Sherman, W.R., Gish, B.G., Honchar, M.P., Munsell, L.Y., 1986. Effects of lithium on phosphoinositide metabolism in vivo. Federation Proceedingss 45, 2639–2646. Spector, R., Lorenzo, A.V., 1975. Myo-inositol transport through the blood brain barrier. Neurochemical Research 13, 785–787. Tricklebank, M.D., Singh, L., Jackson, A., Oles, R.J., 1991. Evidence that a proconvulsant action of lithium is mediated by inhibition of myoinositol phosphatase in mouse brain. Brain Research 558, 145–148. Williams, M.B., Jope, R.S., 1995. Modulation by inositol of cholinergicand serotonergic-induced seizures in lithium-treated rats. Brain Research 685, 169–178. Williams, R.S., Cheng, L., Mudge, A.W., Harwood, A.J., 2002. A common mechanism of action for three mood-stabilizing drugs. Nature 417, 292–295. Williams, R., Ryves, W.J., Dalton, E.C., Eickholt, B., Shaltiel, G., Agam, G., Harwood, A.J., 2004. A molecular cell biology of lithium. Biochemical Society Transactions 32, 799–802.