Psychosomatic Medicine 61:599-617 (1999)
© 1999 American Psychosomatic Society
SPECIAL ISSUE: PSYCHOPHARMACOLOGY AND PSYCHOSOMATIC RESEARCH |
Regulation of Signal Transduction Pathways and Gene Expression by Mood Stabilizers and Antidepressants
Guang Chen, MD,
Khondakar A. Hasanat, MD,
Joseph M. Bebchuk, MD, FRCPC,
Gregory J. Moore, PhD,
Debra Glitz, MD and
Husseini K. Manji, MD, FRCPC
From the Department of Psychiatry and Behavioral Neurosciences (G.C., K.A.H., J.M.B., G.J.M., D.G., H.K.M.), Pharmacology (H.K.M.), Radiology (G.J.M.), and Cellular and Clinical Neurobiology Program (G.C., G.J.M., H.K.M.), Wayne State University School of Medicine, Detroit, MI.
Address reprint requests to: Husseini K. Manji, MD, FRCPC, Director, Laboratory of Molecular Pathophysiology, Department of Psychiatry and Behavioral Neurosciences, Wayne State University School of Medicine, UHC 9B, 4201 St. Antoine Blvd., Detroit, MI 48201. Email: hmanji{at}med.wayne.edu
 |
ABSTRACT
|
|---|
OBJECTIVE: To determine whether the currently available evidence supports the hypothesis that antidepressants and mood stabilizers may bring about some of their long-term therapeutic effects by regulating signal transduction pathways and gene expression in the central nervous system.
METHODS: To address this question, we reviewed the evidence showing that chronic administration of antidepressants and mood stabilizers involves alterations in signaling pathways and gene expression in the central nervous system.
RESULTS: A large body of data has shown that lithium and valproate exert effects on the protein kinase C signaling pathway and the activator protein 1 family of transcription factors; in contrast, antidepressants affect the cyclic adenosine monophospate pathway and may bring about their therapeutic effects by modulating cyclic adenosine monophosphate–regulated gene expression in the central nervous system.
CONCLUSIONS: Given the key roles of these signaling cascades in the amplification and integration of signals in the central nervous system, the findings have clear implications not only for research into the etiology and pathophysiology of the severe mood disorders but also for the development of novel and innovative treatment strategies.
Key Words: mood disorders • protein kinase C • gene expression • lithium • antidepressants • valproate.
Abbreviations: AC = adenylate cyclase; ADP = adenosine diphosphate; AP-1 = activator protein 1; cAMP = cyclic adenosinemonophosphate; CNS = central nervous system; CREB = cAMPresponse element–binding protein; DAG = diacylglycerol; DMI= desipramine; GAP-43 = growth-associated protein 43; GDP =guanosine diphosphate; GSK-3 = glycogen synthase kinase 3; GTP = guanosine triphosphate; IP3 = inositol1,4,5-trisphosphate; MARCKS = myristoylated alanine-rich C kinasesubstrate; PI = phosphatidylinositol; PIP2 =phosphotidylinositol-4,5-bisphosphate; PKC = protein kinaseC; PLC = phospholipase C; RGS = regulator of G proteinsignaling; TPA = 12-o-tetradecanoyl-phorbol13-acetate; ßAR = ß-adrenergic; 5HT = serotonin.
|
INTRODUCTION
|
|---|
Mood disorders are common, severe, chronic, and often life-threatening illnesses. Despite well-established genetic diatheses and extensive research, the biochemical abnormalities underlying the predisposition to, and the pathophysiology of, these disorders remain to be clearly established. Early biological theories regarding mechanisms of action of mood stabilizers and antidepressants focused on various neurotransmitters, in particular the biogenic amines. However, although a number of acute, in vitro effects of these agents are well established, their therapeutic effects are only seen after chronic administration, thereby precluding any simple mechanistic interpretations based on acute biochemical effects. The search for the mechanisms of action of mood stabilizers and antidepressants has been facilitated by a growing appreciation that rather than any single neurotransmitter system being responsible for depression or mania, multiple interacting and overlapping systems are likely involved in regulating mood and that most effective drugs likely do not work on any particular neurotransmitter system in isolation but rather affect the functional balance between interacting systems. Advances in our understanding of the cellular basis of neuronal communication have also led to a reconceptualization of the mechanisms by which neuronal function is regulated, and it has become increasingly appreciated that in addition to regulating the levels of neurotransmitters per se, the throughput of the extracellular signal can also be markedly regulated through alterations in intracellular signaling (1–5). In this context, signal transduction pathways are in a pivotal position in the CNS and thus represent attractive targets to explain the efficacy of pharmacological agents in treating multiple aspects of mood disorders (6–8). It is therefore not surprising that, in recent years, research aimed at elucidation of the cellular mechanisms underlying the therapeutic effects of mood stabilizers and antidepressants have focused on second messenger-generating systems. We now provide a brief overview of transmembrane signal transduction pathways and then discuss their potential involvement in mood disorders and as targets of mood stabilizers and antidepressants.
|
TRANSMEMBRANE CELLULAR SIGNAL TRANSDUCTION
|
|---|
There has been considerable recent progress in elucidation of the cellular mechanisms underlying neuronal communication, particularly with respect to defining the obligatory role of G proteins in the transduction of a vast array of extracellular, receptor-detected signals across cell membranes to intracellular effectors. Most of the molecules that initiate cellular signaling cannot penetrate the hydrophobic plasma membrane lipid bilayer but rather transmit their message via cell-surface receptors. These receptors facilitate the optimization and amplification of signal transmission by stimulating the production or modulation of one or, indeed, several intracellular second or third messengers. It has been estimated that about 80% of all known hormones, neurotransmitters, and neuromodulators elicit cellular responses through G proteins coupled to a variety of intracellular effectors (Table 1 ). Guanine nucleotide–binding proteins (G proteins) are a ubiquitous family of proteins that serve the critical role of transducers of information across the plasma membrane, coupling receptors to various effectors (1–5). The G proteins are heterotrimers localized to the inner surface of the plasma membrane and consist of an 
subunit, which binds (and hydrolyzes) GTP, and ß
subunits, which form a tightly but noncovalently linked dimer (Figure 1 ).
View larger version (26K):
[in this window]
[in a new window]
|
Fig. 1. G protein activation/deactivation cycle. This figure depicts activation of the stimulatory G protein (Gs) by the ßAR receptor, but a similar activation/deactivation cycle is thought to occur with most (if not all) G proteins. The G protein subunit cycles between an inactive GDP-bound heterotrimeric (ß) form and an active GTP-bound monomeric form. At rest, an equilibrium exists between the receptor in the high-affinity state (coupled to the G protein) and the low-affinity (uncoupled) state. Activation of receptors by an agonist induces a conformational change in the receptor, allowing it to interact with the G protein, leading to the release of GDP, and the formation of a high-affinity ternary complex (agonist-receptor-G protein). This high-affinity state is short lived, and binding of GTP to the empty nucleotide site on the subunit of the G protein leads to a destabilization of the high-affinity complex and a dissociation of the G protein into s-GTP and ß subunits. It is now well established that both the s-GTP and ß subunits are able to regulate the activity of various effectors. To date, the best characterized effects of the G protein ß subunits are potentiation of the activity of ACs II and IV and activation of certain PLC isozymes, ion channels, and receptor kinases. s-GTP is shown to activate ACs in this figure, but there is also evidence demonstrating the direct activation of L-type Ca2+ channels by s-GTP (at least in certain tissues). The continued activation of effectors by -GTP is terminated by the action of a GTPase enzyme intrinsic to the a subunit. The formation of s-GDP causes its dissociation from AC; the reassociation of s-GDP with ß is thermodynamically stable and completes the cycle with the formation of the inactive GDP-bound heterotrimeric (ß) G protein. ATP = adenosine triphosphate; PLC ß = phospholipase C ß isoenzyme; s = subunit of stimulatory G protein; ß = G protein ß subunits.
|
|
The subunits had previously been assumed to confer receptor and effector specificity to the G protein (and, indeed, the G proteins are named according to their subunits), but recent evidence has also demonstrated critical roles for ß subunits in signal transduction pathways. Information traverses catalytically across a G protein–coupled system in a GTPase cycle, resulting in a several thousand–fold amplification of the original signal. The G proteins that coordinate receptor-effector activity are derived from a large gene family, and recent cloning and sequencing of cDNAs encoding the subunits have defined four major classes, Gs, Gi, Gq, and G12, with 20 subtypes (9; Table 1). The G protein subunits have intrinsic GTPase activity that cleaves bound GTP to GDP, and interactions with certain effectors, and a novel class of proteins (RGSs) accelerates intrinsic subunit GTPase activity. GTP hydrolysis serves to "turn off" the G protein and allows reassociation of the subunits of the heterotrimer (Figure 1). Indirectly, this also contributes to the desensitization process. G protein–coupled signal transduction pathways play an especially important role in the CNS, where they serve the critical roles of first amplifying and "weighting" extracellularly generated neuronal signals and then transmitting these integrated signals to effectors, thereby forming the basis for a complex information-processing network (2, 4). Because a single receptor subtype can be coupled to multiple G proteins, and because multiple G proteins can converge to activate or inactivate a single effector (2, 4–6), G protein–coupled interactions form complex networks. Indeed, given their critical roles in cellular physiology, it is not surprising that abnormalities in G protein–coupled signaling pathways have now been identified in a number of clinical conditions (see Refs. 4 and 5 for excellent reviews; Table 2 ). The high degree of complexity generated by the interactions of G protein–coupled receptors may be one mechanism by which neurons acquire the flexibility for generating the wide range of responses observed in the nervous system. This has led to the proposal that G proteins may be involved in pathways regulating such diverse vegetative functions as mood, appetite, and wakefulness and may also play a key role in the mechanisms of action of molecular mechanisms of action of antidepressants and mood-stabilizing agents.
|
ABNORMALITIES IN SECOND MESSENGER-GENERATING SYSTEMS IN MOOD DISORDERS
|
|---|
cAMP-Generating System
The future development of selective receptor and second messenger ligands for positron emission tomography studies may permit the direct assessment of CNS receptor and postreceptor sensitivity in humans. To date, however, studies of receptor and postreceptor function in mood disorders have been limited to indirect research strategies or postmortem studies (6, 8). The most commonly used strategy has been to characterize receptor function in readily accessible blood elements. Much clinical research has focused on the activity of the cAMP-generating system in mood disorders. Overall, the preponderance of evidence suggests altered receptor and/or postreceptor sensitivity of the cAMP-generating system in the absence of consistent alterations in the number of receptors themselves (6, 8). Thus, using platelets, several (but not all) studies have demonstrated reduced prostaglandin E1 receptor stimulation and 2 inhibition of AC in unipolar depression (8, 10–15). Most studies measuring leukocyte ßAR-stimulated AC activity report decreased responsiveness in depressed patients compared with healthy volunteers (8, 16–21). One study suggested that this finding correlates with treatment response and that poor responders characteristically have lower pretreatment isoproterenol-stimulated AC activity (20). The consistently observed decrease in leukocyte ßAR function in depression could reflect an inherited abnormality of the ßAR/Gs/AC complex, as suggested by the findings of Wright et al. (22), who used Epstein-Barr virus–transformed lymphocytes from manic-depressive and control subjects. However, these findings need to be replicated, and a number of additional confounding factors need to be considered (8). In this context, twin studies suggest that variations in isoproterenol-stimulated cAMP production are most likely due to "environmental" effects on the number or sensitivity of ßAR receptors (23). Thus, additional studies are clearly needed to ascertain whether the consistent differences in leukocyte ßAR sensitivity in depression truly reflect comparable differences in the brain. Interestingly, patients with bipolar depression have not consistently demonstrated a similar attenuation of ßAR-stimulated AC activity (24, 25). These findings are of considerable interest and are compatible with results of recent studies suggesting the presence of elevated levels of Gs in bipolar affective disorder (discussed below).
G Proteins in Mood Disorders
In view of the indirect evidence of abnormalities at postreceptor sites described above, it is not surprising that several independent laboratories have examined G proteins in patients with mood disorders (see Ref. 8 for an excellent review). Young et al. were the first to report increased levels of Gs in patients with bipolar affective disorder in two separate studies (8, 26). Compared with control subjects matched for age, postmortem interval, and brain pH, patients with bipolar affective disorder had increased levels of Gs in frontal, temporal, and occipital cortex, but not in hippocampus, thalamus, or cerebellum, in postmortem brain tissue. The investigators also found increases in forskolin-stimulated AC activity in postmortem brain tissue compatible with a postreceptor abnormality in patients with bipolar affective disorder. The findings of elevated Gs levels and/or function are also supported by the recent findings of Wang and Friedman (27), who found increased agonist-activated [35S]GTPS binding to G protein subunits in frontal cortical membrane preparations of postmortem brain tissue from patients with bipolar affective disorder. Garcia-Sevilla et al. (28) reported increased levels of Gi1/2 in prefrontal cortical samples obtained postmortem from depressed patients who committed suicide, effects that were apparently attenuated by antemortem antidepressant treatment. Overall, the findings in postmortem brain tissue in patients with unipolar depression have been less consistent (see Ref. 8 for a summary of postmortem findings). In keeping with the G protein abnormalities in brain tissue obtained postmortem from bipolar affective disorder patients, Schreiber et al. (29) reported "hyperfunctional" G protein function in leukocytes of untreated manic patients by demonstrating that agonist-stimulated binding of [3H]Gpp(NH)p (a stable, nonhydrolyzable analog of GTP) was enhanced in leukocyte membranes of untreated manic patients compared with control subjects. These findings suggest the presence of increased levels of G proteins and/or enhanced receptor-mediated activation of G proteins in leukocytes from untreated manic subjects. More recently, investigators have reported significantly higher levels of Gs in mononuclear leukocytes from patients with bipolar, but not unipolar, depression (30); another study (31) quantitatively measured the levels of the major G protein subunits in leukocytes and platelets from both untreated (predominantly manic) and lithium-treated, euthymic bipolar affective disorder patients; in both platelet and leukocyte membranes, this study observed higher levels of the 45-kd form of Gs in the overall group of bipolar affective disorder patients (treated or untreated) compared with control subjects. A recent study found elevated levels of Gs mRNA in granulocytes obtained from patients with bipolar, but not unipolar, depression (32). This study also found nonsignificant elevations in the levels of Gi2 in unmedicated bipolar patients, which intriguingly were modulated by lithium in bipolar (but not unipolar) patients. Studies of G protein levels in peripheral cells from unipolar depressed patients have revealed somewhat conflicting results, with one study reporting decreased levels of leukocyte Gs and Gi in patients with unipolar depression (33) and another reporting elevated platelet Gi2 in these patients (34). Similar to what has been observed in the CNS (3), one recent study evaluated the role of platelet G proteins as "signal coincidence detectors" and found this function to be impaired in depressed patients (35). Recent studies have also suggested that peripheral cell G protein measures are altered in patients undergoing treatment with antidepressants or electroconvulsive shock and, furthermore, that these changes may be associated with treatment response (34, 36, 37). Because it is unlikely that electroconvulsive shock directly alters peripheral cell G protein levels, it is likely that the changes following treatments are secondary to alterations in circulating factors (eg, catecholamines or glucocorticoids). Overall, the most consistent finding to emerge is that in both peripheral cells and postmortem brain tissue from patients with bipolar depression, elevations are observed in the predominant subspecies of Gs present in the tissues examined (see Ref. 8). Because Gs is a ubiquitously expressed protein, it may seem counterintuitive that an abnormality in this protein may play a role in the pathophysiology of bipolar affective disorder. However, there is already precedence for clinical disorders arising from abnormalities in the levels of Gs, which present with limited clinical manifestations, despite the ubiquitous expression of the protein (see Refs. 4 and 5 for excellent reviews on clinical manifestations associated with signaling abnormalities). These heterogeneous clinical effects and tissue-specific manifestations have been postulated to arise from differences in receptor, G protein, and effector stoichiometries in different tissues and to tissue-specific differences in the ability of different cells to compensate for the abnormality (4, 5). It should be emphasized, however, that there is, at present, no evidence to suggest that alterations in the levels of Gs are due to a mutation in the Gs gene itself (38). Indeed, numerous transcriptional and posttranscriptional mechanisms regulate the levels of G protein subunits (see Ref. 8), and the elevated levels of Gs could potentially represent the sequelae of alterations in any one of these other biochemical pathways. Thus, at this point, considerable caution is clearly required in interpretation of the data because they are derived primarily from peripheral cell models and may not adequately reflect CNS pathology. The possibility of the presence of aberrant biochemical pathways that regulate Gs levels in bipolar disorder is currently being studied (8).
Phosphoinositide/PKC Signaling Pathway
Peripheral cell and postmortem brain studies have generally revealed modest abnormalities in the phosphoinositide/PKC signaling system in bipolar affective disorder. Thus, one study measured membrane phospholipids in platelets of seven medication-free patients in the manic phase of bipolar affective disorder and seven healthy control subjects. These investigators found that the relative percentage of platelet membrane PIP2 was significantly higher in manic patients than in control subjects (39). More recently, investigators from the same laboratory (40) studied PIP2 membrane values in a patient with bipolar disorder during different mood states in a single case study. They found that the relative percentage of PIP2 in platelet membranes increased with cycling from the euthymic into the manic state. After lithium treatment, the PIP2 level decreased and was similar to that during the euthymic state. Consistent with these findings, van Calker et al. (41) found increased sensitivity to agonist stimulation of the Ca2+ response in neutrophils of manic-depressive patients, effects that were normalized by lithium treatment. Interestingly, in addition to lithium, other psychotropic drugs used in the treatment of bipolar affective disorder (including antidepressants, haloperidol, and valproate) have also been demonstrated to have complex effects on the phosphoinositide signaling system both ex vivo and in vitro (42–44). Investigators have also attempted to determine whether the altered phosphoinositide signaling found in peripheral cells from bipolar affective disorder patients is also present in the CNS. In this context, Mathews et al. (45) found increased Gq/11 immunoreactivity in postmortem occipital cortex from patients with bipolar affective disorder. However, these elevated levels of Gq/11 were accompanied by reduced agonist-induced PI turnover (46), although the potential effects of long-term lithium treatment remained to be fully delineated. To date, only two studies have directly examined PKC in bipolar affective disorder. Friedman et al. (47) investigated PKC activity and translocation in response to serotonin in platelets obtained from patients with bipolar affective disorder before and during lithium treatment. They found that the ratios of platelet membrane bound to cytosolic PKC activities were elevated in the manic patients compared with patients with schizophrenia and healthy control subjects. In addition, serotonin-elicited platelet PKC translocation was found to be enhanced in those subjects. Similar to the results observed in rodent brain, lithium treatment for up to 2 weeks resulted in a reduction in cytosolic and membrane-associated PKC activities and in an attenuated PKC translocation in response to serotonin. These preliminary results suggest that alteration in platelet PKC is associated with the manic phase of bipolar illness. More recently, investigators from the same laboratory measured PKC isozyme levels, activity, and translocation in postmortem brain tissue from patients with bipolar affective disorder; they found increased PKC activity and translocation, as well as elevated levels of cytosolic - and membrane-associated - and 
PKC isozymes, in brain tissue obtained postmortem from patients with bipolar affective disorder compared with tissue from control subjects (27). These results may be specific to bipolar affective disorder, because investigators from another laboratory recently found significantly decreased [3H]phorbol 12, 13-dibutyrate (PDBU)-binding sites in both membranous and cytosolic postmortem brain samples (Brodmann’s areas 8 and 9) obtained from teenage suicide victims compared with control subjects (48).
|
ARE G PROTEIN–COUPLED SIGNAL TRANSDUCTION PATHWAYS TARGETS FOR ANTIDEPRESSANTS AND MOOD-STABILIZING AGENTS?
|
|---|
Antidepressants and G Proteins
Long-term administration of a variety of antidepressant treatments, including tricyclics, monoamine oxidase inhibitors, and electroconvulsive shock, seem to downregulate or desensitize the ßARs in rat forebrain (49) and enhance the synaptic efficacy of serotonergic neurotransmission via the 5HT1A receptor in rat hippocampus (50). These effects, however, do not explain the clinical efficacy of all antidepressants, and the dissociation between receptor number and their functional responsiveness has led to the investigation of possible postreceptor sites of action of these drugs. After long-term treatment, DMI is reported to cause a functional uncoupling of the ßAR from Gs in rat cortex (51, 52) and to interfere with the breakdown of the ßAR high-affinity ternary complex (53) and the subsequent activation of AC, perhaps by decreasing the affinity of Gs for guanine nucleotides (54). Okada et al. (55) reported that pertussis toxin treatment of rats overcomes DMI-induced ßAR desensitization (as assessed by measurements of isoproterenol-stimulated AC activity) without attenuating DMI-induced ßAR downregulation. These studies suggest that antidepressants modify the interactions between the ßAR and Gi/Go and are intriguing in view of the recently defined role of G protein ß subunits in regulating receptor desensitization (56).
Recent studies have investigated the effects of long-term administration of antidepressants on the levels of G protein subunits. Overall, the results are far from conclusive and have tended to reveal a complex, regional- and tissue-specific pattern of effects on various G protein subunits and their mRNA levels after long-term treatment with various antidepressants (8, 34, 37, 57–65). This complex pattern of effects is not altogether surprising in view of the differing pharmacological profiles of the drugs (and their metabolites) with respect to receptor binding, monoamine reuptake blockade, and enzymatic (monoamine oxidase) inhibition (66). Thus, one study found that antidepressants tended to decrease the immunolabeling of Gs in the hippocampus (57), consistent with the recent report of postreceptor desensitization of hippocampal AC activity after long-term DMI treatment (52) and suggesting a coordinate downregulation of the ß-adrenoceptor-Gs-AC complex by antidepressants (49, 66). In this study, long-term antidepressant treatment tended to decrease the immunolabeling of Gi1/2, in hippocampus, whereas tricyclics (but not the monoamine oxidase inhibitor) tended to increase levels of Go in the hippocampus (57). These findings are complementary to those of Okada et al. (55) described above and suggest the possible involvement of pertussis toxin substrates (Gi and Go) in mediating the effects of DMI. Reductions in the levels of Gi2 have also been observed in platelets of depressed patients treated with citalopram, imipramine, or clomipramine (34). Additionally, the opposite effects of the antidepressants on the levels of Gi and Go described above offer an explanation for the seemingly opposite effects of antidepressants on 5HT1A signaling that have been observed when using functional biochemical (cAMP) or electrophysiological measures (50, 66, 67). Thus, long-term administration of antidepressants may produce a relative shift in the hippocampal 5HT1A receptor response from those mediated by Gi toward those mediated by Go. Other studies have found that long-term administration of fluoxetine reduces the levels of Gi, Go, and Gz in the hypothalamus (58, 60) and reduces the levels of Go and Gi in the midbrain (61). Long-term administration of imipramine has been reported to decrease Go mRNA levels in CA1 and CA3 hippocampal subfields and dentate gyrus without affecting Gs or Gi mRNA levels in these regions (65). In contrast, repeated electroconvulsive shock has been reported to reduce Gs mRNA in CA1 and CA3, reduce the levels of Go mRNA in dentate gyrus, and reduce the levels of Gi2 in both dentate gyrus and CA3 (63). Several studies, however, have not found any alterations in the levels of G subunits or their mRNA levels after long-term antidepressant administration (59, 62, 64, 68).
Although the precise effects of antidepressants on the levels of G proteins remain inconclusive, several elegant studies have demonstrated an enhanced coupling between Gs and the catalytic unit of adenylate cyclase after long-term administration of antidepressants (68–71). Several studies have also demonstrated that the postreceptor components of the cAMP system are regulated by long-term antidepressant treatments, including cAMP-dependent protein kinase enzyme activity (72, 73).Taken together, these results suggest that antidepressants, through their complex effects on G proteins and the AC signaling system, may attenuate ßAR-mediated activation of Gs while enhancing the effects of agents operating by pathways independent of the ßAR receptor.
Consistent with these results, recent data from Duman et al. (74) and Nibuya et al. (75) have demonstrated that long-term treatment of rats with a variety of antidepressants increases the levels of CREB mRNA, CREB protein, and cAMP response element DNA–binding activity in hippocampus. These effects were selective for antidepressants and were not observed with a variety of nonantidepressants, including cocaine, morphine, and haloperidol. These results suggest that genes regulated by CREB may, in fact, also be long-term targets of antidepressants. In this context, the same investigators (74, 75) have demonstrated that long-term treatment of rats results in an increase in the expression of two genes known to be regulated by CREB, namely BDNF (brain-derived neurotrophic factor) and its receptor (trkB). These investigators have postulated that upregulation of the expression of a neurotrophic factor and its receptor may increase the survival of hippocampal neurons or promote the sprouting of neurons that innervate the hippocampus, such as those of the major monoaminergic pathways. These exciting findings have led to a molecular and cellular hypothesis of depression (74), which posits that stress-induced vulnerability and the therapeutic action of antidepressant treatments occur via intracellular mechanisms regulating neurotrophic factors necessary for the survival and functioning of critical neurons, with important implications for future drug development.
|
LITHIUM AND THE PHOSPHOINOSITIDE CYCLE
|
|---|
Over the last decade, research on the molecular mechanisms underlying the therapeutic effects of lithium has focused on intracellular second messenger-generating systems, in particular receptor-coupled hydrolysis of PIP2. Although inositol phospholipids are relatively minor components of cell membranes, they play a major role in receptor-mediated signal transduction pathways and are involved in a diverse range of responses in the CNS (reviewed in Ref. 76). Activation of a variety of neurotransmitter receptor subtypes (including muscarinic M1, M3, and M5, noradrenergic 1, serotonergic 5HT2, and several metabotropic glutamatergic receptors) induces hydrolysis of membrane phospholipids. In brief, agonists such as acetylcholine, norepinephrine, serotonin, and glutamate bind to specific cell-surface receptors to stimulate certain isoforms of the enzyme PLC. Activated PLC catalyzes the conversion of PIP2 to two second messengers, IP3 and DAG. IP3 stimulates mobilization of intracellular Ca2+, whereas DAG activates PKC. IP3 can be phosphorylated and dephosphorylated, leading to other inositol phosphate compounds or to unphosphorylated inositol. Inositol, in turn, is converted to phosphatidylinositol, which is phosphorylated to phosphatidylinositolphosphate (PIP) and PIP2 (Figure 2 ).
View larger version (36K):
[in this window]
[in a new window]
|
Fig. 2. Effects of lithium on the phosphoinositide cycle. In this scheme, occupancy of the receptor (R) by a specific agonist (A) initiates hydrolysis of PIP2 by PLC. Hydrolysis of PIP2 by PLC results in the formation of two major second messengers, IP3 and DAG. IP3 mobilizes calcium from intracellular stores, whereas DAG activates PKC (see text for details). IP3 is either dephosphorylated to form inositol 1,4-diphosphate (Ins 1,4,P2), inositol monophosphate (Ins P1), and, ultimately, free inositol, or phosphorylated to form inositol 1,3,4,5-tetraphosphate (Ins 1,3,4,5 P4), which is then dephosphorylated by sequential distinct pathways. Lithium, at therapeutically relevant concentrations, inhibits the dephosphorylation of inositol 1,3,4-triphosphate (Ins 1,3,4 P3), Ins 1,4,P2, and all three forms of inositol phosphatases (not shown in detail in the figure). Because the ability of a cell to maintain sufficient supplies of myoinositol is crucial to the resynthesis of the phosphoinositides, and because in most tissues inositol is derived primarily from recycling of inositol phosphates, one early consequence of lithium’s action is to reduce the levels of free inositol. As shown in the figure, inositol depletion can also perturb the DAG limb of the phosphoinositide turnover pathway. Thus, resynthesis of PI involves the transfer of the phosphatidic acidic moiety from cytidine diphosphate DAG to myoinositol. Presumably because of its effect of lowering levels of inositol, lithium treatment of a number of cells has been found to increase the levels of cytidine diphosphate DAG, and its interconvertible metabolite, DAG. Because DAG activates PKC, one consequence of lithium treatment is an activation of PKC. BBB = blood brain barrier.
|
|
The ability of a cell to maintain sufficient supplies of myoinositol is crucial to the resynthesis of the phosphoinositides and the maintenance and efficiency of signaling. Lithium, at therapeutically relevant concentrations, is an inhibitor of inositol monophosphatase (Ki, 0.8 mM) and results in an accumulation of inositol-1-monophosphate as well as a reduction in free inositol (77, 78). Lithium also inhibits inositol polyphosphate-1-phosphatase, which is involved in recycling inositol polyphosphates to inositol. Furthermore, because the mode of enzyme inhibition is uncompetitive, lithium’s effects have been postulated to be most pronounced in systems undergoing the highest rate of PIP2 hydrolysis. Thus, Berridge et al. (79) first proposed that the physiological consequence of lithium’s action is derived through a depletion of free inositol and that its selectivity could be attributed its preferential action (due to the uncompetitive nature of the inhibition) on the most overactive receptor-mediated neuronal. For these reasons, it has been hypothesized that a physiological consequence of lithium’s action is derived through a depletion of free myoinositol in the brain, the "inositol depletion hypothesis" (79). Because several subtypes of adrenergic, cholinergic, serotonergic, and metabotropic glutamatergic receptors are coupled to PIP2 hydrolysis in the brain, the inositol depletion hypothesis, as initially proposed, offered an attractive explanation for lithium’s therapeutic efficacy in treating multiple aspects of bipolar affective disorder.
However, numerous studies have examined the effects of lithium on receptor-mediated PI responses, and although some report a reduction in agonist-stimulated PIP2 hydrolysis in rat brain slices after short- or long-term lithium administration, these findings have often been small and inconsistent and subject to numerous methodological differences (see Refs. 7 and 80 for excellent recent reviews). Thus, despite its attractiveness, the inositol depletion hypothesis as originally articulated has recently been questioned (7, 80, 81). A body of preclinical data, however, suggests that some of the initial actions of lithium may occur with a relative depletion of myoinositol (82–85); this relative depletion of myoinositol may initiate a cascade of secondary changes at different levels of the signal transduction process and gene expression in the CNS, effects that are ultimately responsible for lithium’s therapeutic efficacy (80, 81). Our research group has recently undertaken a series of studies using magnetic resonance spectroscopy to determine whether lithium reduces the levels of myoinositol in critical brain regions of individuals with bipolar affective disorder (despite the attractiveness of the inositol depletion hypothesis, it has never been demonstrated to occur in human brain). Our results show that the therapeutic administration of lithium does, indeed, produce a significant brain region–specific reduction in myoinositol levels in patients with bipolar affective disorder (86). However, the major lithium-induced reductions in myoinositol levels occur after 5 days of lithium administration, at a time when the patients’ clinical state is largely unchanged. These results clearly indicate that lithium’s therapeutic effects are not due to a lowering of myoinositol per se. Studies are currently under way to determine whether the early lithium-induced reductions in myoinositol levels are associated with ultimate therapeutic response (likely mediated by changes in PKC-induced gene expression).
|
LITHIUM AND PKC
|
|---|
Considerable recent research has clearly shown that, in addition to its effects on PI turnover, the PKC signaling pathway is a target for the actions of long-term lithium (reviewed in Refs. 7, 81, and 87; Table 3 ). PKC is highly enriched in brain and plays a major role in regulating pre- and postsynaptic aspects of neurotransmission (88–90). PKC is one of the major intracellular mediators of signals generated on external stimulation of cells through a variety of neurotransmitter receptor subtypes that induce the hydrolysis of membrane phospholipids. PKC is now known to exist as a family of closely related subspecies, has a heterogeneous distribution in brain (with particularly high levels in presynaptic nerve terminals), and plays a major role in the regulation of neuronal excitability, neurotransmitter release, and long-term alterations in gene expression and plasticity (88–90).
Recent evidence accumulating from various laboratories has demonstrated that lithium exerts complex effects on PKC in a number of cell systems, including the brain (7, 81, 91). The preponderance of the currently available data suggest that short-term lithium treatment may activate PKC, whereas long-term lithium exposure results in an attenuation of phorbol ester–mediated responses, which may be accompanied by downregulation of PKC isozymes brain (7, 81, 91; Table 3). Activation of PKC is now known to facilitate the release of a number of neurotransmitters, and biochemical studies have revealed that chronic (3-week) lithium at "therapeutic" levels attenuates PKC-induced [3H]5-HT release in hippocampus (92). Although the precise mechanisms by which PKC activation facilitates neurotransmitter release remain to be fully elucidated, they have been postulated to involve the phosphorylation of key substrates in the nerve terminal, including MARCKS and GAP-43, by PKC and to play a key role in facilitating neurotransmitter release (93). In this context, it is noteworthy that Lenox et al. (94) have demonstrated that the level of MARCKS, a protein implicated in synaptic transmission, was significantly reduced after long-term lithium exposure. This laboratory has been very instrumental in identifying the MARCKS protein as a potentially therapeutically relevant target of the actions of lithium and valproate (94, 95). MARCKS is preferentially expressed in dendritic branches and axon terminals within limbic and limbic-associated regions of the brain. Additionally, MARCKS cross-links filamentous actin and has been implicated in cellular processes, including cytoskeletal restructuring, transmembrane signaling, and neurotransmitter release (discussed in Refs. 93–95). Thus, the possible involvement of MARCKS in the mechanism of action of lithium and valproate remains an exciting area for future research.
Using quantitative autoradiographic techniques, it has also been demonstrated that long-term (5-week) lithium administration results in a significant decrease in membrane-associated PKC in several hippocampal structures, most notably the subiculum and CA1 region, in the absence of any significant changes in the various other cortical and subcortical structures examined (96). Furthermore, immunoblotting using monoclonal anti-PKC antibodies has revealed isozyme-specific decreases in PKC and 
, which have been particularly implicated in facilitating neurotransmitter release, in the absence of significant alterations in PKC ß, PKC , PKC 
, or PKC . It is also noteworthy that exposure of neuroblastoma cells (97) or PC12 cells (98) to 1 mM lithium in vitro produces isozyme-selective decreases in PKC and, in the case of PC12 cells, PKC . In the absence of suitable animal models for the major psychiatric disorders, a major problem inherent in neuropharmacological research is the difficulty in precisely ascribing therapeutic relevance to any observed biochemical finding. One approach is to iden- tify common biochemical targets that are modified by drugs belonging to the same therapeutic class (eg, antimanic agents) but possessing distinct chemical structures (eg, lithium and valproate, the only two drugs approved by the US Food and Drug Administration for treatment of bipolar affective disorder) when administered in a "therapeutically relevant" paradigm. Although they likely do not work by precisely the same mechanisms, identifying the biochemical targets that are regulated in concert by these two chemically distinct agents may provide important clues about molecular mechanisms underlying mood stabilization in the brain. In view of the significant effects of lithium on PKC outlined above, the effects of valproate on various aspects of PKC functioning have also been investigated. It has been found that the structurally highly dissimilar agent, valproate, produces effects on the PKC signaling pathway that are strikingly similar to those of lithium (91, 99, 100; Table 4 ). Interestingly, long-term administration of lithium and valproate seems to regulate PKC isozymes by distinct mechanisms, with valproate’s effects seeming to be largely independent of myoinositol (100). This biochemical observation is consistent with clinical observations that some patients have a preferential response to one or the other agent and that additive therapeutic effects often occur when the two agents are coadministered.
PKC Inhibitors in the Treatment of Acute Mania
In view of the pivotal role of the PKC signaling pathway in regulation of neuronal excitability and neurotransmitter release (88–90, 101), we have postulated that the attenuation of PKC activity may play a major role in the therapeutic effects of the only two drugs approved by the US Food and Drug Administration for the treatment of mania (Table 4). There is thus a clear need to investigate the efficacy of PKC inhibitors in the treatment of mania; there is currently only one relatively selective PKC inhibitor available for human use, tamoxifen. This synthetic nonsteroidal antiestrogen has been widely used in the treatment of breast cancer (102, 103). A number of its effects are due to estrogen receptor antagonism (102), but it has become clear in recent years that it is also a potent PKC inhibitor at therapeutically relevant concentrations (104).We therefore initiated a pilot study investigating the efficacy of tamoxifen in the treatment of acute mania (105). Clearly, these results must be considered preliminary because of the small sample size thus far. Nevertheless, the significant (and, in some cases, rapid and striking) results we observed suggest that tamoxifen possesses antimanic properties (105). In view of the preliminary data suggesting the involvement of the PKC signaling system in the pathophysiology of bipolar affective disorder (vide supra), these results suggest that PKC inhibitors may be very useful agents in the treatment of bipolar affective disorder. Larger, double-blind, placebo-controlled studies of tamoxifen and of novel selective PKC inhibitors in the treatment of mania are clearly warranted.
|
LITHIUM AND AC
|
|---|
The other major receptor-coupled second messenger system on which lithium exerts significant effects is the cAMP-generating system (8, 81). Lithium has been demonstrated to exert complex effects on this system, and the preponderance of data demonstrates an elevation of basal AC activity with simultaneous attenuation of the receptor-mediated response (106–109). Lithium in vitro inhibits stimulation of AC by the poorly hydrolyzable analog of GTP, Gpp(NH)p, and also by Ca2+/calmodulin, suggesting that lithium in vitro is directly able to inhibit the catalytic unit of AC. Because these inhibitory effects of lithium in vitro can be overcome by Mg2+ (106, 110–114), they seem to be mediated (at least in part) by direct competition with magnesium (whose hydrated ionic radius is similar to that of lithium) for a binding site on the catalytic unit of AC (106, 110, 113, 114). However, the inhibitory effects of long-term lithium treatment on rat brain AC are not reversed by Mg2+ and still persist after the membranes are washed but are reversed by increasing concentrations of GTP (106, 113). These results suggest that the therapeutically relevant effects of lithium (ie, those seen with long-term administration that are not reversed immediately on discontinuation) may be exerted at the level of signal-transducing G proteins at a GTP-responsive step (81).
|
LITHIUM AND G PROTEINS
|
|---|
As noted above, abundant experimental evidence has shown that lithium attenuates receptor-mediated AC activity and PI turnover in rodents and humans in the absence of consistent changes in the density of the receptors themselves. There is considerable evidence that long-term (but not short-term) lithium administration affects G protein function (as assessed by second messenger function) (23, 107, 114–119; Table 5 ). At present, the possible effects of long-term lithium administration on the absolute levels of G protein subunits remain unclear; two independent laboratories have not observed any alterations (120, 121), whereas others have reported small but significant decreases in the levels of the s, i1, and i2 in rat brain (122, 123). However, long-term lithium administration has been shown to alter mRNA levels of a number of G proteins in rat brain, including s, i1, and i2 (120, 122, 123), suggesting that lithium produces complex transcriptional and posttranscriptional effects after long-term administration (vide infra).
Recent studies have also examined the effects of long-term lithium administration on G protein function in humans and have generally observed reduced receptor/G protein coupling (23, 67, 107, 116, 124–127). The effects of 2 weeks of lithium administration on G protein measures have also been examined in healthy volunteers, which prevents the potentially confounding, significant effects of alterations in mood state–dependent biochemical and neuroendocrine parameters (107, 124). Long-term lithium administration resulted in an increase in both basal and postreceptor-stimulated platelet AC activity in platelets, which is most compatible with attenuation of Gi function (107, 124). Similar to the findings in rat brain, lithium did not affect the levels of platelet G protein subunits but produced a significant 40% increase in pertussis toxin–catalyzed [32P]ADP-ribosylation, once again suggesting stabilization of the inactive undissociated ß heterotrimeric form of Gi (121, 128, 129). These results suggest that removal of the "inhibitory tone" by lithium might result in enhanced responses by agents activating the stimulatory pathway distal to the receptor. Consistent with this hypothesis, in an elegant series of recent studies, lithium has been shown to potentiate the hyperactivity induced by intraaccumbens cholera toxin administration, which activates the stimulatory G proteins Gs and Golf) (130).
Overall, the preponderance of data from cell culture, rodent, and human studies argues for an effect of long-term lithium administration on G protein function in both humans and rodents. Such allosteric modulation of G proteins may play a role in the long-term prophylactic efficacy of the cation in protecting susceptible individuals from spontaneous, stress-induced, and drug-induced (eg, by antidepressant and stimulants) cyclic affective episodes. These long-term effects of long-term lithium administration on G protein are likely attributable to an indirect posttranslational modification of the G proteins and a relative change in the dynamic equilibrium of the active and inactive states of protein conformation. In this context, it is noteworthy that investigators have demonstrated that lithium alters the levels of endogenous ADP-ribosylation in C6 glioma cells (131) and in rat brain (132), suggesting another mechanism by which long-term lithium administration may indirectly regulate the activity of these critical signaling proteins.
|
VALPROIC ACID AND G PROTEINS
|
|---|
Recent studies have examined the effects of valproate on components of the ßAR-coupled, cAMP-generating system (133). Long-term administration of valproate has been shown to produce a significant alteration of the ßAR-coupled, cAMP-generating system in cultured cells in vitro; these effects were observed at concentrations of valproate similar to those attained in the plasma in the clinical treatment of neuropsychiatric disorders. In contrast to what has been observed with chronic lithium treatment (discussed above), it was found that chronic valproate produced a significant reduction in the density of ßARs. Interestingly, the decrease in numbers of ßARs (approximately 30%) was accompanied by an even greater decrease in receptor- and postreceptor-mediated cAMP accumulation, suggesting that long-term valproate administration also exerts effects at the ßAR-Gs interaction or at postreceptor sites (eg, Gs or AC). Consistent with such a contention, it was indeed found that long-term, but not short-term, valproate incubation induced a marked decrease in the levels of Gs 45 but not any other G protein subunits examined (Gs 52, Gi1/2, Go, or Gq/11). In view of the suggested involvement of Gs in the pathophysiology of manic-depressive illness (vide supra), as well as the effects of lithium on the ßAR-Gs-AC system, these effects may play a role in the therapeutic effects of valproate and are worthy of further study.
|
GSK3ß: A THERAPEUTICALLY RELEVANT TARGET FOR THE ACTIONS OF LITHIUM?
|
|---|
In the last 2 years, a hitherto completely unexpected target for the action of lithium has been identified. Klein and Melton (134) were the first to demonstrate that lithium, at therapeutically relevant concentrations, is an inhibitor of GSK3ß. GSK3ß is an evolutionarily, highly conserved kinase that was originally identified as a regulator of glycogen synthesis. It is now known to play a critical role in the CNS, by regulating various cytoskeletal processes via its effects on tau and synapsin I, as well as in long-term nuclear events, through phosphorylation of c-jun and nuclear translocation of ß-catenin (134–136). Thus, lithium’s inhibition of GSK3ß may underlie some of its transcriptional and posttranscriptional actions in the brain and thereby many of its long-term therapeutic effects; this is clearly an exciting area for future research (80).
|
EFFECTS OF CARBAMAZEPINE ON THE cAMP-GENERATING SYSTEM
|
|---|
Considerable data suggest that carbamazepine (an atypical anticonvulsant) is an alternative or adjunctive treatment to lithium, both for acute manic episodes as well as for long-term prophylaxis in bipolar affective disorder (137). Despite the widespread clinical use of carbamazepine, the cellular mechanisms underlying both its anticonvulsant and mood-stabilizing effects have not been identified (138). In contrast to the effects observed with lithium and valproate described above, it has been demonstrated that carbamazepine has very modest effects on the PKC signaling pathway or G proteins (H. K. Manji, unpublished observations). Carbamazepine has, however, been demonstrated to have many effects on the cAMP signaling pathway. The cAMP-generating system plays a major role in the regulation of neuronal excitability and has been postulated to play a role in the pathophysiology of both seizure disorders (139, 140) and bipolar affective disorder (6, 8). It is thus noteworthy that carbamazepine decreases the basal concentrations of cAMP in mouse cerebral cortex and cerebellum (141) and reduces cAMP production induced by norepinephrine (141, 142), adenosine (141, 143, 144), and the epileptogenic compounds ouabain (141, 145) and veratridine (146) in brain slices. In manic patients, carbamazepine decreased elevated levels of cAMP in cerebrospinal fluid (147). Recent studies have also demonstrated that carbamazepine inhibits forskolin-induced c-fos gene expression in cultured pheochromocytoma (PC-12) cells (148). Thus, overall, considerable evidence indicates that carbamazepine inhibits cAMP formation.
Recent studies have investigated the possible mechanisms by which carbamazepine inhibits the cAMP-generating system. It has been found that carbamazepine, at therapeutically relevant concentrations, inhibited both basal AC and forskolin-stimulated cAMP accumulation in C6 glioma cells (149). Within the clinical therapeutic range (approximately 50 µM), carbamazepine inhibited basal cAMP levels by 10% to 20% and forskolin-stimulated cAMP production by 40% to 60%. Together, these data indicate that carbamazepine is more effective in inhibiting the activated AC system, although the possibility of "floor effects" (ie, an inability to lower basal cAMP levels beyond certain levels in this system) cannot be ruled out. To further characterize the site at which carbamazepine exerts its inhibitory effects, ACs were purified from rat cerebral cortex using a forskolin affinity purification column. It was found that similar to the situation observed in intact C6 cells and C6 cell membranes, carbamazepine inhibited both basal and forskolin-stimulated activity of purified AC (149). Together, the data suggest that carbamazepine inhibits cAMP production by acting directly on AC and/or through factors tightly associated with or copurified with AC. Consistent with these results, it has been demonstrated that carbamazepine attenuates forskolin-induced c-fos (an immediate-early gene) expression in PC-12 cells (148) and inhibits FSK-induced phosphorylation of CREB in C6 glioma cells (149). Because c-fos and CREB are known to involved in mediating a number of long-term neuronal responses (150), these effects might be postulated to play a role in the delayed therapeutic effect of carbamazepine. The effects of mood stabilizers on long-term genomic events are discussed next.
TOP
ABSTRACT
INTRODUCTION
