VIEWS: 36 PAGES: 9 POSTED ON: 9/6/2011
American Journal of Medical Genetics Part C (Semin. Med. Genet.) 123C:76 – 84 (2003) A R T I C L E The Neurobiology of Bipolar Disorder GREGORY S. BERNS* AND CHARLES B. NEMEROFF The neurobiology of bipolar disorder is reviewed. Bipolar disorder is associated with alterations in central nervous system (CNS) function from the level of large-scale brain circuits to intracellular signal transduction mechanisms. Because of the broad spectrum over which these abnormalities appear, the causative effects are most likely present in the lowest common denominator of all of these systems. Current evidence points to subtle alterations in signal transduction that reverberate downstream both intra- and extracellularly to produce the symptoms of bipolar disorder. ß 2003 Wiley-Liss, Inc. KEY WORDS: bipolar disorder; schizophrenia; magnetic resonance imaging INTRODUCTION In contrast, untreated schizophrenia is lesser severity and termed hypomania. generally characterized by a long pro- Bipolar II disorder has a lifetime pre- Bipolar disorder has been known to gressive decline without any signiﬁcant valence rate of 0.5% [Angst, 1998]. mankind since at least the time of the return to the premorbid state. Thus, bipolar I and II disorders have a ancient Greeks, but it was not until the Estimates of bipolar I disorder, or combined prevalence rate of about 2%, 20th century that it was truly recognized classic manic-depressive illness, have which is equal to or greater than that of as an illness distinct from other psychiat- ranged from 0.8–1.6% of the population schizophrenia. It is important to note ric disorders. The German psychiatrist [Kessler et al., 1994]. These patients that one-third of patients with bipolar Emil Kraepelin was likely the ﬁrst to experience both full-blown manic epi- disorder have their ﬁrst episode during draw attention to the distinction be- sodes and syndromal major depressions. their adolescent years, and 50% exhibit tween manic depression and schizophre- Various subtypes have been described comorbid drug or alcohol abuse. nia. While the acute presentations of that include rapid cycling bipolar dis- In trying to understand the biology both illnesses can appear identical with order and mixed states, the latter char- of bipolar disorder, we struggle to link delusions and hallucinations, manic de- acterized by patients who exhibit the obviously complex manifestations of pression is characterized by an abrupt concurrent symptoms of both mania the illness and its cyclic nature with its onset and a ﬂuctuating course in which and depression. Bipolar II disorder is biological substrate. How can one ferret the individual returns to a relatively characterized by syndromal major out the roots of a mental illness when normal state between episodes, particu- depressions and manic episodes that do we are only beginning to understand larly early in the course of the disorder. not fulﬁll criteria for mania but are of the complex machinery of the brain? Even so, extraordinary discoveries, using techniques of molecular biology, ge- Gregory S. Berns, M.D., Ph.D., is an Associate Professor in the Department of Psychiatry and netics, and neuroimaging, consistently Behavioral Sciences at Emory University School of Medicine and the Coulter School of Biomedical point to the fact that individuals with Engineering at the Georgia Institute of Technology. Dr. Berns received his Ph.D. in biomedical engineering from the University of California–Davis in 1990 and his M.D. from the University bipolar disorder have fundamental altera- of California–San Diego in 1994. He specializes in the use of brain imaging techniques to tions in brain function. We present here understand the functioning of human reward pathways and how this can be applied to some of the recent advances in neuro- understand human decision making, an area that has come to be known as neuroeconomics. Charles B. Nemeroff, M.D., Ph.D., is the Reunette W. Harris Professor and Chairman of the biology that ﬁrmly root the pathophy- Department of Psychiatry and Behavioral Sciences at Emory University School of Medicine. He siology of bipolar disorder in the brain. received both his Ph.D. and M.D. degrees from the University of North Carolina–Chapel Hill and completed residency training at Duke University in 1985. He was a Professor in the Departments of Psychiatry and Pharmacology at Duke University, relocating to Atlanta in 1991 to assume his current position. His research has concentrated on the biological basis of the major neuro- ALTERATIONS IN psychiatric disorders, including affective disorders, schizophrenia, and anxiety disorders. He serves as Editor-in-Chief of Neuropsychopharmacology and has published more than 700 research BRAIN STRUCTURE reports and reviews. He is currently the recipient of several research grants from the NIH, including a Conte Center for the Neurobiology of Major Mental Disorders. The idea that bipolar disorder may be Grant sponsor: NIH; Grant numbers: MH-61010, DA-00367, MH-42088, MH-39415, MH- related to an alteration in brain structure 58922. was derived from the astute clinical ob- *Correspondence to: Gregory S. Berns, Department of Psychiatry and Behavioral Sciences, Emory University School of Medicine, 1639 Pierce Drive, Suite 4000, Atlanta, GA 30322. servation that certain brain lesions pro- E-mail: email@example.com duced by brain tumors, stroke, or head DOI 10.1002/ajmg.c.20016 injury resulted in manic-like behavior ß 2003 Wiley-Liss, Inc. ARTICLE AMERICAN JOURNAL OF MEDICAL GENETICS (SEMIN. MED. GENET.) 77 enlargement is typically characteristic of white matter hyperintensities. Their The idea that bipolar disorder of cell loss, such as the neurodegenera- location suggests a potential role in dis- tions observed in Alzheimers disease rupting communicating ﬁbers between may be related to an or perhaps alterations in neural circuit fronto-temporal regions, which lends alteration in brain structure development, but issues of controlling support to earlier observations that was derived from the astute for other potentially important con- lesions in these regions cause mania. founding factors, such as alcohol and Follow-up postmortem studies of pa- clinical observation that certain drug abuse and head injury, preclude an tients with UBOs have demonstrated a brain lesions produced by easy interpretation of the results. Given number of histologic changes in these these limitations, volumetric imaging regions, including small vascular mal- brain tumors, stroke, or head studies have provided intriguing ﬁndings formations, dilated perivascular spaces, injury resulted in in bipolar disorder. As recently reviewed brain cysts, infarcts, and necrosis. These manic-like behavior. by Strakowski et al. [2002a], both bipolar are surprisingly nonspeciﬁc lesions that and unipolar depression are reportedly can occur from multiple causes. It is associated with smaller prefrontal lobe possible that these lesions represent [Cummings and Mendez, 1984; Cum- volumes, but in contrast both the basal damage from a comorbid disease process mings, 1993]. In general, any brain ganglia and thalamus are larger in bi- unrelated to bipolar disorder; however, lesion is far more likely to cause depres- polar patients [Aylward et al., 1994; recent studies in children and adoles- sion than mania, but lesions that induce Strakowski et al., 2002a]. Moreover, cents with mania continue to reveal an mania occur more commonly in the both hippocampal and amygdala enlar- abundance of these UBOs [Lyoo et al., frontal and temporal lobes and sub- gement in bipolar disorder has also been 2002; Pillai et al., 2002]. More re- cortically in the head of the caudate and occasionally reported [Swayze et al., cently, the MRI changes in bipolar the thalamus [Cummings and Mendez, 1992; Strakowski et al., 2002a], but not disorder have been noted to bear a 1984; Starkstein et al., 1991], so-called consistently. striking similarity to an autosomal secondary manias. It has been repeatedly Volumetric measurements of var- dominant disorder called cerebral auto- suggested that lesions of the left frontal ious brain regions are of interest, espe- somal dominant arteriopathy with lobe result in depression, whereas right cially to identify structures that deserve subcortical infarcts and leukoencephalo- fronto-temporal lesions produce mania. further scrutiny, but the interpretation pathy (CADASIL), and there is some However, these generalizations about of volumetric assessments remains pro- evidence for a higher incidence of laterality are far too simplistic, and blematic. Speciﬁc changes in regional bipolar disorder in CADASIL patients many exceptions to this rule have been volume may occur in response to a [Ahearn et al., 1998, 2002]. observed. variety of factors and may not be per- Before the advent of noninvasive manent. MRI has the capability of look- ALTERATIONS IN brain imaging techniques, the only ing beyond simple structure by yielding BRAIN FUNCTION methods available to examine patients’ information about both neurochemical brains were autopsy, brain biopsy, and alterations and the neural activity of spe- Both CT and MRI yield static informa- pneumoencephalography. There have ciﬁc regions. When brain MR images tion about brain structure—a kind of been few postmortem anatomic studies were obtained in bipolar patients, it was snapshot—but the brain is a dynamic of patients with conﬁrmed bipolar dis- quickly noted that such patients had an organ, and to understand function we order; however, neuroimaging using inordinate number of hyperintense re- must choose different techniques. Func- both computed tomography (CT) and gions. These unidentiﬁed bright objects tional neuroimaging can measure subtle magnetic resonance imaging (MRI) has (UBOs) are typically associated with changes in receptor density, blood ﬂow, revealed multiple structural alterations. vascular diseases, including systemic hy- and glucose metabolism. Although we CT scans were the ﬁrst noninvasive pertension, Binswangers disease, and cannot yet image neuronal activity modality to systematically scrutinize carotid arteriosclerosis. Why are they directly, recent technological advances brain structure, but the relatively poor present in patients with a mental dis- in MRI have led to the development of demarcation between different brain order? Further studies revealed that they functional MRI (fMRI), which can regions allowed for only the grossest of tend to localize in deep white matter detect changes in blood ﬂow on the observations. Several investigators have structures. The percentage of bipolar scale of seconds with millimeter resolu- suggested that patients with bipolar dis- patients exhibiting these ﬁndings has tion, and such changes are clearly coupl- order have larger ventricles than normal ranged from 5–50%, compared with ed to neuronal activity [Ogawa et al., controls, a ﬁnding much more clearly about 3% for controls [Aylward et al., 1990; Kwong et al., 1992]. established in patients with schizophre- 1994; Altshuler et al., 1995; Dupont Most functional neuroimaging nia [Schlegel and Krtezschmar, 1987; et al., 1995; Norris et al., 1997; studies take advantage of a critical ob- Dewan et al., 1988; Swayze et al., 1990; McDonald et al., 1999]. Elderly bipolar servation about neuronal activity and Strakowski et al., 1993]. Ventricular patients have larger and a higher number brain blood ﬂow. When synaptic activity 78 AMERICAN JOURNAL OF MEDICAL GENETICS (SEMIN. MED. GENET.) ARTICLE increases in a particular brain region, the patients recovered from their depression, system. Is there some aspect of neuro- blood ﬂow to that region transiently these abnormalities disappeared. These nal function that renders patients with increases [Logothetis et al., 2001]. The ﬁndings have not been completely rep- manic-depressive illness more prone to blood ﬂow apparently increases beyond licated. Some studies have reported these shifts? The new science of chaos the metabolic requirements of the tissue, relatively normal cortical metabolism, theory characterizes these states as at- so a surfeit of oxygenated blood tem- but more localized abnormalities in tractors. Consider the simplistic case of porarily bathes the region. Both positron subcortical regions such as the caudate three mood states: euthymia (normal emission tomography (PET) and fMRI or subgenual prefrontal cortex [Drevets mood), depression, and mania. For most can be used to measure this blood ﬂow et al., 1995]. individuals, euthymia is the usual state. increase, thereby indirectly measuring Frontal regions, especially the dor- Unpleasant events cause transient dys- neuronal activity. PET can also be used solateral prefrontal cortex (DLPFC), phoria, but most people quickly return to measure directly local glucose meta- have been identiﬁed as having both to their usual mood state. Similarly, bolism. By using a glucose analog, 2- decreased metabolism and blood ﬂow winning a lottery makes most people deoxy-glucose (2-DG), this compound in depression. Decreases in left DLPFC very happy, but does not shift them to a can be labeled with a positron emitter, metabolism have been correlated with permanent state of elation. Euthymia is ﬂuorine-18. Like glucose, 2-DG is severity of depression, but this is not therefore a stable state for most people— transported into cells and metabolized. likely to be speciﬁc to bipolar disorder perturbations are small and the return to Unlike glucose, 2-DG is metabolized [Strakowski et al., 2002b]. Frontal hypo- euthymia is invariant. Patients with bi- through only one step of glycolysis and metabolism has been reported repeat- polar disorder often switch into extreme subsequently becomes trapped in the edly in schizophrenia. Presumably, any mania or depression without returning cell. Thus it serves as a marker of both alteration in these regions is associated to a euthymic state for a considerable glucose uptake and metabolism. Because with profound effects on cognition and period of time. the process of uptake and metabolism emotion. These brain regions are well takes some time, ﬂuoro-deoxyglucose known to be integral to many functions ALTERATIONS IN (FDG) studies are more appropriate for that are altered in psychiatric disorders BRAIN CHEMISTRY the measurement of stable state changes such as attention and working memory. in the brain. Other frontal regions, especially those MR spectroscopy (MRS) has been used The earliest functional imaging on the innermost folds of the brain, extensively to measure changes in rela- studies focused on large-scale changes are poorly understood but seem to tive concentrations of several important in both cerebral blood ﬂow and meta- be involved in conﬂict monitoring neuroregulators in the brains of bipolar bolism [Strakowski et al., 2002b]. These [Carter et al., 1998], reward valuation patients. The most common method, studies showed that bipolar-depressed [Montague and Berns, 2002], and proton-MRS, is used both routinely and patients had signiﬁcantly lower cortical response inhibition. This suggests that now clinically. Proton-MRS measures metabolism than either controls or pa- both state- and trait-dependent inter- the relative concentrations of N-acetyl tients with unipolar depression [Baxter actions with performance on cognitive aspartate (NAA), creatine (Cr), phos- tasks may serve as a ﬁner probe of dys- phocreatine (PCr), and various choline function with brain imaging [Berns (Cho)-containing compounds. Because The earliest functional imaging et al., 2002]. lithium increases Cho concentrations in Functional neuroimaging in bi- human red blood cells [Jope et al., 1978], studies focused on large-scale polar disorder dispelled a common myth it was reasonable to look for similar changes in both cerebral about the organization of the brain, changes in Cho concentration in the blood ﬂow and metabolism. namely, that speciﬁc cognitive processes brain. The MRS data on Cho concen- can be completely localized to isolated trations are not entirely consistent, but These studies showed brain regions. Virtually every imaging there does seem to be a consensus that that bipolar-depressed patients study has identiﬁed networks of activity. there is at least an elevated Cho/Cr ratio In this context, it becomes clear why the in the basal ganglia of bipolar patients had signiﬁcantly lower search for regional abnormalities has not [Stoll et al., 2000; Strakowski et al., cortical metabolism than either yielded consistent results. If a cognitive 2002b]. It is likely that this ﬁnding is state process requires the coordinated func- dependent because similar elevations controls or patients tion of several brain regions, then a small have been observed in depression. with unipolar depression. alteration in one region may cause dra- Because the Cho peak in proton- matic effects on the whole circuit. MRS represents several compounds, pho- Both mania and depression are char- sphorous-MRS has been used to further et al., 1985; Buchsbaum et al., 1986]. acterized by profound global changes in delineate the nature of these alterations. Furthermore, these changes were state brain function. These state changes are Phosphorous-MRS can distinguish ATP, dependent, meaning that when the manifest at multiple levels in the nervous PCr, and phosphomonoesters (PMEs) ARTICLE AMERICAN JOURNAL OF MEDICAL GENETICS (SEMIN. MED. GENET.) 79 like phosphocholine, phosphoinositol, cult to precisely document. NE, like its metabolite, 3-methoxy-4-hydroxy- and phosphoethanolamine; it can also many neurotransmitters, appears extra- phenylethyleneglycol (MHPG), to NE, measure indirectly intracellular pH and cellularly in small amounts. Further- ranged from 64–107% greater in several free magnesium. However, because of the more, it is metabolized to several other cortical regions. Signiﬁcant decreases relatively low concentrations of these compounds that appear in CSF, plasma, in both serotonin and dopamine meta- compounds, phosphorous-MRS is tech- and urine. Thus, alterations in NE cir- bolism were found in the same brain nically demanding and suffers from limited cuits may appear as a change in either the regions. sensitivity. Like the proton-MRS data, neurotransmitter or any of its metabo- In addition to the monoamine there are discordant ﬁndings. Most studies lites. Most evidence points to a deﬁ- neurotransmitters, others have also been have found changes in PMEs in the frontal ciency in depression and an excess in implicated in the pathophysiology of lobes of symptomatic patients, but whether mania, but this may simply reﬂect the bipolar disorder. Because of its prepon- it is increased or decreased, or whether global neural activity of these states, as derance in the brain, glutamate has there are left/right asymmetries, is not well as contributions from the sympa- received growing attention. Glutamate agreed upon [Stoll et al., 2000; Strakowski thetic nervous system, which utilizes exerts its effects through four major et al., 2002b]. At a minimum, these studies NE as the neurotransmitter of post- receptor families. Three are ionotropic: suggest that alterations in phospholipid ganglionic neurons. Interestingly, NE N-methyl-D-aspartate (NMDA), a- metabolism occur in bipolar disorder elevations purportedly precede the amino-3-hydroxy-5-methyl-isoxazole- [Yildiz et al., 2001]. 4-propionic acid (AMPA), and kainate. The ionotropic receptors are coupled to NEUROCHEMICAL Changes are evident at different ion channels, and when gluta- mate binds to them, the ionic conduc- CHANGES virtually all levels of the central tances are altered. The fourth family is Changes are evident at virtually all levels nervous system in bipolar metabotropic, and these receptors are of the central nervous system (CNS) in patients. If the illness is coupled to intracellular G-proteins. A bipolar patients. If the illness is manifest recent postmortem of the striatum by changes in brain attractor states, as manifest by changes in brain found increased expression of mRNA opposed to lesions of a speciﬁc region, attractor states, as opposed transcripts for both the NR2D subtype then we have merely shifted the search of the NMDA receptor and the AMPA for the cause to more fundamental levels. to lesions of a speciﬁc receptor in bipolar patients [Meador- Numerous biochemical abnormalities region, then we have merely Woodruff et al., 2001]. have been detected by measuring one shifted the search for the or another neurotransmitter metabolites cause to more fundamental NEUROENDOCRINE or hormones in plasma, cerebrospinal CHANGES ﬂuid (CSF), and postmortem tissue levels. Numerous studies. Although depression has often For many years it has been recognized been conceptualized as due to a relative biochemical abnormalities that certain endocrine disorders are asso- deﬁciency in the activity of certain have been detected by ciated with a greater than expected monoamine-containing systems, e.g., occurrence in bipolar disorder. The measuring one or another serotonin, dopamine, and norepinephr- hypothalamic-pituitary-adrenal (HPA) ine (NE), these have not yet been clearly neurotransmitter metabolites axis has received the most attention in implicated in the pathophysiology of or hormones in plasma, mood disorders. Corticotropin-releasing bipolar disorder. Many antidepressants, factor (CRF) is released from neurons which increase the activity of one or cerebrospinal ﬂuid, and in the paraventricular nucleus of the more of these neurotransmitter circuits, postmortem tissue studies. hypothalamus, and CRF is transported can precipitate the development of to the anterior pituitary, causing ACTH mania. to be released systemically. ACTH acts Concentrations of NE, or its major switch into mania. Although NE may upon the adrenal cortex, where it re- metabolite, are consistently altered in the not itself be the causative mediating leases cortisol [Wang and Nemeroff, CSF of patients with bipolar disorder. factor, it is further evidence for the 2003]. Mixed mania has been associated NE was originally proposed by Schildk- idea of unstable cortical states. In one with both an elevated CSF and urinary raut  as the major culprit in both comprehensive postmortem study, free cortisol concentration [Swann et al., depression and mania. The catechola- there were no differences in the con- 1992], but this has also been observed mine hypothesis stated that depression centration of NE, serotonin, or dopa- in unipolar major depression. The as- resulted from low levels of NE and mania mine in any brain region of bipolar sessment of HPA function has typically resulted from high levels [Schildkraut, patients [Young et al., 1994], but NE been done with either of two tests: the 1965]. This has been remarkably difﬁ- turnover, as measured by the ratio of dexamethasone suppression test (DST) 80 AMERICAN JOURNAL OF MEDICAL GENETICS (SEMIN. MED. GENET.) ARTICLE and the CRF-stimulation test. The a-subunit that typically binds to the ion events, including calcium mobilization, practical limitations of the DST have channel. In addition to regulation of cytoskeletal organization, and phos- been covered elsewhere [Shapiro et al., ion channels, G-proteins interact with phoinositol metabolism. Rap1 has also 1983], but the CRF-stimulation test several intracellular second messengers, been found to be involved in the re- retains utility. In the latter, CRF is ad- including cyclic AMP (cAMP), cyclic gulation of signal cascades coupled to ministered intravenously (usually 1 mg/ neurotrophic factors [Bos et al., 2001]. kg or 100-mg dose), and blood samples This is very intriguing because recent are obtained for ACTH and cortisol data also suggest a role for both anti- at 30-min intervals for 2–3 hr. When The heterogeneity of depressants and mood stabilizers as compared to normal control subjects, both imaging ﬁndings and neuroprotective agents [Duman et al., the ACTH response to exogenous CRF 2001; Manji and Duman, 2001]. Inter- neurotransmitter alterations is blunted in depression but not in mania. estingly, another downstream substrate, When dexamethasone was combined has not yielded a single cAMP response element-binding pro- with CRF-stimulation, depressed bipo- underlying hypothesis for the tein (CREB) has not be found to be lar patients were reported to have a increased in bipolar patients and may be signiﬁcantly greater elevation of cortisol pathophysiology of bipolar decreased in the temporal lobes [Stewart than either normal controls or unipolar disorder. There does, however, et al., 2001]. depressed patients [Schmider et al., Protein kinase C (PKC) is yet 1995; Rybakowski and Twardowska, appear to be consistent another second messenger-dependent 1999]. The mechanisms for these altera- evidence pointing to signal kinase, dependent on calcium, not tions in the HPA axis of patients with transduction as one major locus cAMP. At rest, PKC isozymes exist as affective disorders are unknown, but both cystolic and membrane-bound glucocorticoid resistance, which is ana- of pathophysiology. forms, but mostly cystolic. Activation logous to the insulin resistance of dia- of receptors coupled to phospholipase C betes mellitus, has been one mechanism facilitates the translocation of cystolic proposed [Pariante and Miller, 2001; GMP (cGMP), calcium, metabolites of PKC to the membrane [Manji et al., Watson and Young, 2002]. the phosphatidyl-inositol (PI) pathway, 1995]. Like PKA, PKC is elevated arachidonic acid, and nitric oxide. Upon in the platelets of bipolar patients neurotransmitter binding, the bg-sub- [Friedman et al., 1993]. Acute lithium SIGNAL TRANSDUCTION unit separates from the receptor and exposures apparently facilitate many The heterogeneity of both imaging ﬁnd- modulates the activity of adenylate cy- PKC-mediated effects, but longer expo- ings and neurotransmitter alterations has clase, in effect changing intracellular sure results in downregulation of some not yielded a single underlying hypoth- levels of cAMP. cAMP then phosphor- PKC isozymes. The effects of lithium- esis for the pathophysiology of bipolar ylates a number of cAMP-dependent induced changes of the PKC signaling disorder. There does, however, appear protein kinases, activating their respec- pathway can be measured on down- to be consistent evidence pointing to tive functions. Protein kinase A (PKA) is stream products, just as in the PKA signal transduction as one major locus of the most prominent of these kinases and system. Chronic lithium exposure has pathophysiology. For historical reasons is referred to as a third messenger. been demonstrated to reduce the ex- alluded to above, the signal transduction Particular attention has been focus- pression of myristolated alanine-rich pathway for catecholamine receptors has ed on the cAMP/PKA transduction C kinase substrate (MARKS), espe- been the most extensively characterized pathway for several reasons. First, cially in the hippocampus. MARKS in mood disorder. What follows is largely lithium exerts complex effects on ade- has been implicated in the regulation of related to the NE system [Duman and nylate cyclase [Risby et al., 1991], and neuroplastic events [Manji and Lenox, Nestler, 1995]. this is believed to be manifest as down- 2000]. The neurotransmitter itself, in this stream changes on the CAMP/PKA case NE, is referred to as the ﬁrst mes- pathway [Manji and Lenox, 2000]. LITHIUM senger, and it binds to one or more Consistent with this notion, an increased adrenergic receptor subtypes (see above). concentration of Ga-subunits was re- Because lithium revolutionized the Depending on the receptor subtype, a ported in the CNS in a postmortem treatment of bipolar disorder, it also number of different intracellular events study of bipolar patients [Young et al., provided a potential window into un- may occur. G-proteins on the intracel- 1991, 1993]. Looking farther down- derstanding the disease-related altera- lular side of the receptor can bind to ion stream, Rap1, a PKA substrate, has been tions that occur at the cellular level. channels, thereby inﬂuencing the mem- reported to exhibit increased levels of Lithium was identiﬁed as an element brane potential of the cell. The G- phosphorylation in the platelets of bi- more than 150 years ago, and it wasn’t proteins are generally composed of three polar patients [Perez et al., 2000]. Rap1 long after its discovery that it was used subunits, labeled a, b, and g, and it is the may be involved in several intracellular as a therapeutic agent for a variety of ARTICLE AMERICAN JOURNAL OF MEDICAL GENETICS (SEMIN. MED. GENET.) 81 ailments. Lithium’s mood-stabilizing has been reported to increase neurogen- lethal doses as little as two times the effects were demonstrated in the 1950s. esis in the hippocampus [Chen et al., therapeutic dose. Lithium treatment is Unlike other medications used to treat 2000]. often associated with a number of un- psychiatric patients, lithium is a salt, toward effects, ranging from tremor and consequently, it does not have a and gastrointestinal side effects (nausea, CLINICAL CORRELATES receptor to which it binds in the brain. diarrhea, and cognitive slowing) to Rather, it is actively transported into The term bipolar disorder is somewhat hypothyroidism and diabetes insipidus. the cell through the sodium channel. misleading because it implies that indi- Fortunately, results from studies utilizing When a neuron depolarizes, the sodium viduals exist in either a depressed or biological models of bipolar disorder channel opens and both sodium and manic state, and that these states are at served as an impetus for research on lithium rush into the cell. The sodium opposite ends of a spectrum. The reality other pharmacologic treatments, and is then actively pumped out, using is more complex, and this has important one line of research led to the anti- the sodium-potassium-ATP pump, but implications for treatment. While it is convulsants carbamazepine and valproic lithium remains in the intracellular true that the depressed and manic states acid; they are now accepted as effective compartment. are far beyond the realm of normal treatments. Valproic acid is FDA- As described above, lithium appears emotions, they are not at opposite ends approved for the treatment of mania. to modulate several second messenger of a continuum. In fact, these states may Olanzapine, a recently FDA-approved systems, including cAMP and phosphoi- represent two dimensions of emotion atypical antipsychotic, is also effective nositol pathways. Lithium may blunt that, to a certain degree, are independent in the treatment of mania and perhaps receptor activation of adenylate cyclase of each other. As many as 40% of bipolar in the depression of bipolar disorder as activity, although separating the effects patients enter a mixed state, a condition well. of lithium from the alterations that occur with either the coexistence or rapid Manic-depressive cycles are neither naturally in bipolar disorder is not always alternation of symptoms of both depres- random nor predictable. Many, if not easy. Rather than causing large changes sion and mania, sometimes called dys- most, patients show a pattern of in- in baseline cellular activity, lithium phoric mania. Similarly, the depression creasing frequency over time. This seems to attenuate responsivity to other of bipolar disorder is generally not the phenomenon occurs in other areas of neurotransmitters. One might say that it same symptomatically as the depression neuroscience and has suggested a model ‘‘turns down the gain.’’ This may explain of unipolar major depression. Bipolar based on kindling and sensitization. its efﬁcacy in bipolar disorder—decreas- depression tends to be atypical with Kindling refers to increased responsivity ing sensitivity to both internal and prominent fatigue, hypersomnia, and to repeated low-level electrical stimula- external stimuli. Lithium also affects reverse diurnal mood variability, as tion. This is analogous to a seizure dis- other neurotransmitter systems, includ- opposed to insomnia in unipolar depres- order, in which a seizure focus becomes ing serotonin, dopamine, and g-amino- sion. increasingly sensitive to other electrical butyric acid (GABA) circuits, and its For these and other reasons, bipolar events (i.e., the more seizures one has, efﬁcacy may possibly be related to its disorder is generally more difﬁcult to the more likely the occurrence of ad- wide-ranging neurobiological effects treat than simple major depression. ditional seizures). The kindling hypoth- rather than to a single mechanism. Antidepressants do not typically work esis also explains the observation that One of the intriguing properties of as well for bipolar disorder, and they can early manic episodes tend to be triggered lithium treatment in mania is that a time destabilize patients by switching them by external events, like crossing time lag of several days is required before into manic or mixed states. Of the zones or drug abuse, whereas after seve- lithium exerts its clinical effect. More- available antidepressants, the best ral episodes they tend to occur without over, lithium’s beneﬁcial effects on choices based on the current limited any precipitants. Certain anticonvul- mood stabilization do not disappear database would support the use of sants, especially carbamazepine and val- immediately upon its discontinuation selective serotonin reuptake inhibitors proic acid, are effective treatments for [Goodwin and Jamison, 1990]. One (SSRIs) (ﬂuoxetine, paroxetine, sertra- certain patients with bipolar disorder, possibility is that lithium exerts its effects line, etc.) and bupropion. There is a lending further support to the kindling by resetting the ionic homeostasis in growing consensus that to obtain an hypothesis. It should be noted, however, neurons either directly or through its optimal response in most bipolar dis- that not all anticonvulsants are effective interaction with second messenger sys- order patients, multiple-drug therapy is in the treatment of bipolar disorder (e.g., tems. Lithium also protects cells from required. Lithium is one of the few Food phenytoin, phenobarbital). Moreover, other chemical insults [Nonaka et al., and Drug Administration (FDA)- the clinical trial data supporting the 1998]. The neuroprotective properties approved drugs for acute treatment of efﬁcacy of the anticonvulsants ranges of lithium may explain lithium-induced mania, and it is the only FDA-approved from valproic acid, which is FDA ap- inhibition of NMDA receptor-mediated maintenance treatment for bipolar dis- proved, to gabapentin and topiramate, calcium inﬂux. Beyond neuropro- order. Lithium, however, has a disturb- which have no published efﬁcacy data. tection, lithium, like antidepressants, ingly narrow therapeutic index, with Moreover, in spite of the attractive 82 AMERICAN JOURNAL OF MEDICAL GENETICS (SEMIN. MED. GENET.) ARTICLE nature of the kindling hypothesis, no or depression are usually discrete, rapid of speciﬁc clock genes offers an excit- convincing neurobiological data have eye movement (REM) sleep and body ing opportunity to bring to bear in- provided any support that this phenom- temperature cycles change more slowly sights from nonlinear dynamical systems. enon actually occurs in patients with and have been documented to precede bipolar disorder. the switch in mood [Goodwin and Manic-depressives suffer profound Jamison, 1990]. Although these cycles The recent elucidation alterations in sleep-wake cycles during change slowly, it appears that once some of both afferent and efferent both the manic and depressive phases of threshold is reached, the bipolar patient their illness, but subtle disturbances in is catapulted into either mania or de- pathways from the circadian rhythms often precede the pression. The outward manifestations of central clock in the SCN and full-scale shift in mood state. Mania is mood may therefore appear to change characterized by a markedly decreased quite suddenly, even though the under- the roles of speciﬁc clock need for sleep. It is well known that sleep lying dynamics are more subtle. genes offers an exciting deprivation [Wehr, 1989] or even travel- ing across time zones may trigger a opportunity to bring PUTTING IT to bear insights from nonlinear manic episode in vulnerable individuals ALL TOGETHER [Jauhar and Weller, 1982; Young, 1995]. dynamical systems. Here again is evidence for some basic In reviewing the data ranging from circuit instability that is subject to behavior to brain state to intracellular transient changes in sleep patterns. The events, one is struck both by the diversity Although largely qualitative at this normal sleep-wake cycle is determined and discordance of the extant ﬁndings in point, much is known about what by a combination of internal circadian bipolar disorder. Although there is no happens when collections of oscillators, rhythms and external cues—day and smoking gun, there is a biological crime like the SCN and its slaves, interact with night. The basic internal rhythm can scene. Our job is to sift through the each other. It will now be possible to be observed across many biologic mea- evidence and determine what happened. model how even subtle alterations in sures: body temperature, heart and re- Continuing the analogy, we deal with a clock synchronization might lead to spiration rate, and secretion of various contaminated crime scene. The diag- chaotic behavior, both biologically and hormones (e.g., growth hormone, cor- nosis of bipolar disorder is never straight- behaviorally. At this point, we do not tisol). The discovery of a ‘‘master clock’’ forward, sometimes being confused with know which is cause and effect, but in the suprachiasmatic nucleus (SCN) schizophrenia, and frequently it is over- one lesson from dynamical systems has revolutionized our understand- laid against a background of substance theory is that any alteration in the func- ing of the coordination of circadian use. The panoply of medications used to tion of a complex system will be rhythms [Reppert and Weaver, 2002]. treat the illness wreak further havoc on manifest throughout the system. This It is now known that the SCN entrains the CNS changes, making it increasingly occurs precisely because all the parts are a multitude of pacemakers both in the difﬁcult to sort out nascent brain changes interconnected. brain and out (e.g., liver). The basic from pharmacologically induced ones. If we continue looking into smaller oscillatory function depends on two Nevertheless, it is worthwhile to attempt scales in the brain, then we run into transcriptional factors, termed CLOCK at least some generalizations about the the most incontrovertible evidence of and BMAL1. Emerging data in both neurobiology. systemic dysfunction at the level of depression and bipolar disorder are The fact that there are not grossly signal transduction. Although the cate- suggestive of mutations in these genes consistent alterations in regional brain cholamine hypothesis may be correct [Bunney and Bunney, 2000; Mitterauer, function is the clearest evidence that roughly in the extreme, it too should 2000], but further research is neces- bipolar disorder is not localized to a be considered as part of the endophe- sary to investigate this exciting link. speciﬁc part of the brain. Although notype and not causative. Alterations Interestingly, lithium has been reported strokes can induce manic behavior, these in signal transduction appear to offer to lengthen the circadian period of are more likely syndromic expressions of the most explanatory power for the individual SCN neurons [Abe et al., a ﬁnal behavioral phenotype that coin- range of symptomatology in bipolar 2000]. cidentally resembles the manic state of disorder. The symptoms manifest them- Phase instability, that is, sensitivity bipolar disorder. What functional altera- selves as ampliﬁcations of the range of to perturbations in the circadian rhythm, tions do exist in the brains of bipolar both human emotion and behavior. appears to be one characteristic of bi- patients seemingly represent an exten- Unlike schizophrenia, the fact that polar disorder. This is concordant with sion of the phenotype, the so-called most bipolar patients return to a state the idea of a chaotic system that is more endophenotype. of relative normality, even in the ab- sensitive to slight changes—changes that The recent elucidation of both sence of treatment, is strongly sugges- throw the entire system from one state to afferent and efferent pathways from the tive for alterations in a modulatory another. Although transitions to mania central clock in the SCN and the roles mechanism. ARTICLE AMERICAN JOURNAL OF MEDICAL GENETICS (SEMIN. MED. GENET.) 83 ACKNOWLEDGMENTS hippocampal neurogenesis by lithium. J Manji HK, Lenox RH. 2000. Signaling: cellular Neurochem 75:1729–1734. insights into the pathophysiology of bipolar The authors are supported by grants Cummings JL. 1993. The neuroanatomy of disorder. Biol Psychiatry 48:518–530. from the NIH: MH-61010 and DA- depression. J Clin Psychiatry 54:14–20. Manji HK, Duman RS. 2001. Impairments of Cummings JL, Mendez MF. 1984. Secondary neuroplasticity and cellular resilience in 00367 (G.S.B.); MH-42088, MH- mania with focal cerebrovascular lesions. severe mood disorders: implications for the 39415, and MH-58922 (C.B.N.). Am J Psychiatry 141:1084–1087. development of novel therapeutics. Psycho- Dewan MJ, Haldipur CV, Lane EE, Ispahani A, pharmacol Bull 35:5–49. Boucher MF, Major LF. 1988. Bipolar Manji HK, Potter WZ, Lenox RH. 1995. Signal REFERENCES affective disorder. I. Comprehensive quan- transduction pathways: molecular targets for titative computed tomography. Acta Psy- lithium’s actions. Arch Gen Psychiatry 52: Abe M, Herzog ED, Block GD. 2000. Lithium chiatr Scand 77:670–676. 531–543. lengthens the circadian period of individual Drevets WC, Price JL, Videen TO, Todd RD, McDonald WM, Tupler LA, Marsteller FA, Figiel suprachiasmatic nucleus neurons. Neurore- Raichle ME. 1995. Metabolic abnormalities GS, DiSouza S, Nemeroff CB, Krishnan port 11:3261–3264. in the subgenual prefrontal cortex and KR. 1999. Hyperintense lesions on mag- Ahearn EP, Steffens DC, Cassidy F, Van Meter SA, ventral striatum in mood disorders. Soc netic resonance images in bipolar disorder. Provenzale JM, Seldin MF, Weisler RH, Neurosci Abs 21:260. Biol Psychiatry 45:965–971. Krishnan KR. 1998. Familial leukoence- Duman RS, Nestler EJ. 1995. Signal tran- Meador-Woodruff JH, Hogg AJ, Smith RE. phalopathy in bipolar disorder. Am J sduction pathways for catecholamine recep- 2001. Striatal ionotropic glutamate receptor Psychiatry 155:1605–1607. tors. In: Bloom FE, Kupfer DJ, editors. expression in schizophrenia, bipolar disor- Ahearn EP, Speer MC, Chen YT, Steffens DC, Psychopharmacology: the fourth genera- der, and major depressive disorder. Brain Cassidy F, Van Meter S, Provensale JM, tion of progress. New York: Raven Press. Res Bull 55:631–640. Weisler RH, Krishnan KR. 2002. Investiga- p 303–320. Mitterauer B. 2000. Clock genes, feedback loops tion of Notch3 as a candidate gene for Duman RS, Nakagawa S, Mahlberg J. 2001. and their possible role in the etiology of bipolar disorder using brain hyperintensities Regulation of adult neurogenesis by anti- bipolar disorders: an integrative model. Med as an endophenotype. Am J Med Genet depressant treatment. Neuropsychopharma- Hypotheses 55:155–159. 114:652–658. cology 25:836–844. Montague PR, Berns GS. 2002. Neural econom- Altshuler LL, Curran JG, Hauser P, Mintz J, Dupont RM, Jernigan TL, Heindel W, ics and the biological substrates of valuation. Denicoff K, Post R. 1995. T2 hyperinten- Butters N, Shafer K, Wilson T, Hesselink Neuron 36:265–284. sities in bipolar disorder: magnetic reso- J, Gillin JC. 1995. Magnetic resonance Nonaka S, Hough CJ, Chuan DM. 1998. Chro- nance imaging comparison and literature imaging and mood disorders. Localization nic lithium treatment robustly protects meta-analysis. Am J Psychiatry 152:1139– of white matter and other subcortical ab- neurons in the central nervous system 1144. normalities. Arch Gen Psychiatry 52:747– against excitotoxicity by inhibiting N- Angst J. 1998. The emerging epidemiology of 755. methyl-D-aspartate receptor-mediated cal- hypomania and bipolar II disorder. J Affect Friedman E, Hoau YW, Levinson D, Connell TA, cium inﬂux. Proc Natl Acad Sci USA 95: Disord 50:143–151. Singh H. 1993. Altered platelet protein 2642–2647. Aylward EH, Roberts-Twillie JV, Barta PE, kinase C activity in bipolar affective dis- Norris SD, Krishnan KR, Ahearn E. 1997. Kumar AJ, Harris GJ, Geer M, Peyser CE, order, manic episode. Biol Psychiatry 33: Structural changes in the brain of patients Pearlson GD. 1994. Basal ganglia volumes 520–525. with bipolar affective disorder by MRI: and white matter hyperintensities in patients Goodwin FK, Jamison KR. 1990. Manic-depres- a review of the literature. Prog Neuro- with bipolar disorder. Am J Psychiatry 151: sive illness. New York: Oxford University psychopharmacol Biol Psychiatry 21:1323– 687–693. Press. 1337. Baxter LR, Phelps ME, Mazziotta JC, Schwartz Jauhar P, Weller MP. 1982. Psychiatric morbidity Ogawa S, Lee TM, Kay AR, Tank DW. 1990. JM, Gerner RH, Selin CE, Sumida RM. and time zone changes: a study of patients Brain magnetic resonance imaging with 1985. Cerebral metabolic rates for glucose in from Heathrow airport. Br J Psychiatry contrast dependent on blood oxygena- mood disorders studied with positron emis- 140:231–235. tion. Proc Natl Acad Sci USA 87:9868– sion tomography (PET) and (F-18)-ﬂuoro- Jope RS, Jenden DJ, Ehrlich BE, Diamond JM. 9872. 2-deoxyglucose (FDG). Arch Gen Psychia- 1978. Choline accumulates in erythrocytes Pariante CM, Miller AH. 2001. Glucocorticoid try 42:441–447. during lithium therapy. N Engl J Med 299: receptors in major depression: relevance to Berns GS, Martin M, Proper SM. 2002. Limbic 833–834. pathophysiology and treatment. Biol Psy- hyperreactivity in bipolar II disorder. Am J Kessler RC, McGonagle KA, Zhao S, Nelson CB, chiatry 49:391–404. Psychiatry 159:304–306. Hughes M, Esleman S, Wittchen HU, Perez J, Tardito D, Mori S, Racagni G, Smeraldi Bos JL, de Rooij J, Reedquist KA. 2001. Rap1 Kendler KS. 1994. Lifetime and 12-month E, Zanardi R. 2000. Altered Rap1 endo- signalling: adhering to new models. Nat prevalence of DSM-II-R psychiatric dis- genous phosphorylation and levels in plate- Rev Mol Cell Biol 2:369–377. orders in the United States. Results from the lets from patients with bipolar disorder. Buchsbaum MS, Wu J, DeLisi LE, Holcomb H, National Comorbidity Survey. Arch Gen J Psychiatr Res 34:99–104. Kessler R, Johnson J, King AC, Hazlett E, Psychiatry 51:8–19. Pillai JJ, Friedman L, Stuve TA, Trinidad S, Langston K, Post RM. 1986. Frontal cortex Kwong KK, Belliveau JW, Chesler DA, Goldberg Jesberger JA, Lewin JS, Findling RL, Swales and basal ganglia metabolic rates assessed by IE, Weisskoff RM, Poncelet BP, Kennedy TP, Schulz SC. 2002. Increased presence of positron emission tomography with [18F]2- DN, Hoppel BE, Cohen MS, Turner R, white matter hyperintensities in adolescent deoxyglucose in affective illness. J Affect Cheng HM, Brady TJ, Rosen BR. 1992. patients with bipolar disorder. Psychiatry Disord 10:137–152. Dynamic magnetic resonance imaging of Res 114:51–56. Bunney WE, Bunney BG. 2000. Molecular clock human brain activity during primary sen- Reppert SM, Weaver DR. 2002. Coordination of genes in man and lower animals: possible sory stimulation. Proc Natl Acad Sci USA circadian timing in mammals. Nature 418: implications for circadian abnormalities in 89:5675–5679. 935–941. depression. Neuropsychopharmacology 22: Logothetis NK, Pauls J, Augath M, Trinath T, Risby ED, Hsiao JK, Manji HK, Bitran J, Moses F, 335–345. Oeltermann A. 2001. Neurophysiological Zhou DF, Potter WZ. 1991. The mechan- Carter CS, Braver TS, Barch DM, Botvinick investigation of the basis of the fMRI signal. isms of action of lithium. Arch Gen MM, Noll D, Cohen JD. 1998. Anterior Nature 412:150–157. Psychiatry 48:513–524. cingulate cortex, error detection and the on- Lyoo IK, Lee HK, Jung JH, Noam GG, Renshaw Rybakowski JK, Twardowska K. 1999. The dexa- line monitoring of performance. Science PF. 2002. White matter hyperintensities on methasone/corticotropin-releasing hormone 280:747–749. magnetic resonance imaging of the brain in test in depression in bipolar and unipolar Chen G, Rajkowska G, Du F, Seraji-Bozorgzad children with psychiatric disorders. Comp affective illness. J Psychiatr Res 33:363– N, Manji HK. 2000. Enhancement of Psychiatry 43:361–368. 370. 84 AMERICAN JOURNAL OF MEDICAL GENETICS (SEMIN. MED. GENET.) ARTICLE Schildkraut JJ. 1965. The catecholamine hypo- Strakowski SM, Wilson DR, Tohen M, Woods Handbook of medical psychiatry. New thesis of affective disorders: a review of BT, Douglass AW, Stoll AL. 1993. Structural York: Marcel Dekker. supporting evidence. Am J Psychiatry 122: brain abnormalities in ﬁrst-episode mania. Watson S, Young AH. 2002. Hypothalamic- 509–522. Biol Psychiatry 33:602–609. pituitary-adrenal axis function in bipolar Schlegel S, Krtezschmar K. 1987. Computed Strakowski SM, Adler CA, DelBello MP. 2002a. disorder. Clin Approaches Bipolar Disord tomography in affective disorders. Part I. Volumetric MRI studies of mood disorders: 1:57–64. Ventricular and sulcal measurements. Biol do they distinguish unipolar and bipolar Wehr TA. 1989. Sleep loss: a preventable cause Psychiatry 22:4–14. disorder? Bipolar Disord 4:80–88. of mania and other excited states. J Clin Schmider J, Lammers CH, Gotthardt U, Dettling Strakowski SM, DelBello MP, Adler C, Cecil Psychiatry 50:45–47. M, Holsboer F, Heuser IJ. 1995. Combin- KM, Sax KW. 2002b. Neuroimaging in Yildiz A, Sachs GS, Dorer DJ, Renshaw PF. 2001. ed dexamethasone/corticotropin-releasing bipolar disorder. Bipolar Disord 2:148– 31P nuclear magnetic resonance spectro- hormone test in acute and remitted manic 164. scopy ﬁndings in bipolar illness: a meta- patients, in acute depression, and in normal Swann AC, Stokes PE, Casper R, Secunda SK, analysis. Psychiatry Res 106:181–191. controls: I. Biol Psychiatry 38:797–802. Bowden CL, Berman N, Katz MM, Robins Young DM. 1995. Psychiatric morbidity in Shapiro MF, Lehman AF, Greeﬁeld S. 1983. Biases E. 1992. Hypothalamic-pituitary-adreno- travelers to Honolulu, Hawaii. Comp Psy- in the laboratory diagnosis of depression in cortical function in mixed and pure mania. chiatry 36:224–228. medical practice. Arch Intern Med 143: Acta Psychiatr Scand 85:270–274. Young LT, Li PP, Kish SJ, Siu KP, Warsh JJ. 1991. 2085–2088. Swayze VW, Andreasen NC, Alliger RJ, Ehrhardt Postmortem cerebral cortex Gs alpha-sub- Starkstein SE, Fedoroff P, Berthier ML, Robinson JC, Yuh WT. 1990. Structural brain unit levels are elevated in bipolar disorder. RG. 1991. Manic-depressive and pure abnormalities in bipolar affective disorder. Brain Res 553:323–326. manic states after brain lesions. Biol Psy- Ventricular enlargement and focal signal Young LT, Li PP, Kish SJ, Siu KP, Kamble A, chiatry 29:149–158. hyperintensities. Arch Gen Psychiatry 47: Hornykewiwicz O, Warsh JJ. 1993. Cerebral Stewart RJ, Chen B, Dowlatshahi D, MacQueen 1054–1059. cortex Gs alpha protein levels and forskolin- GM, Young LT. 2001. Abnormalities in the Swayze VW, Andreasen NC, Alliger RJ, Yuh WT, stimulated cyclic AMP formation are cAMP signaling pathway in post-mortem Ehrhardt JC. 1992. Subcortical and tem- increased in bipolar affective disorder. J brain tissue from the Stanley Neuropathology poral structures in affective disorder and Neurochem 61:890–898. Consortium. Brain Res Bull 55:625–629. schizophrenia: a magnetic resonance ima- Young LT, Warsh JJ, Kish SJ, Shannak K, Stoll AL, Renshaw PF, Yurgelun-Todd DA, ging study. Biol Psychiatry 31:221–240. Hornykewiwicz O. 1994. Reduced brain Cohen BM. 2000. Neuroimaging in bipolar Wang X, Nemeroff CB. 2003. Biological distinc- 5-HT and elevated NE turnover and disorder: what have we learned? Biol tion between unipolar and bipolar dis- metabolites in bipolar affective disorder. Psychiatry 48:505–517. orders. In: Soares JC, Gershon S, editors. Biol Psychiatry 35:121–127.
"The Neurobiology of Bipolar Disorder"