Bipolar disorder was traditionally considered an episodic illness showing complete remission of symptoms between bouts of elevated or depressed moods. However, many patients experience significant social impairment between episodes, and different lines of investigation, including brain imaging and neurocognitive assessment, provide evidence of structural or functional brain alterations that are independent of the illness episode. Neuroimaging studies have reported increased size and reduced glucose utilisation in the amygdala and basal ganglia, and parts of the prefrontal cortex appear to be smaller in
Mood Disorders: A Handbook of Science and Practice. Edited by M. Power. © 2004 John Wiley & Sons, Ltd. ISBN 0-470-84390-X.
bipolar patients than in controls. Phosphorous magnetic resonance spectroscopy (MRS) has revealed abnormalities of membrane phospholipid metabolism in frontal and striatal regions (Strakowski et al., 2000). The subgenual prefrontal cortex, which is part of the cin-gulate cortex, is of particular interest (Drevets et al., 1997). Abnormalities in this region were first identified in depression; subsequently positron-emission tomography (PET), magnetic resonance imaging (MRI) and post-mortem data have confirmed a significant reduction in grey matter volume in the subgenual prefrontal cortex in bipolar disorder (Ongur et al.,
1998). These changes may be a feature of all types of depression; imaging studies have not identified consistent changes specific to bipolar disorder. Overall, there is converging evidence from imaging studies that dysfunctions in the prefrontal cortex, amygdala and striatum play a part in bipolar mood swings.
Investigations of cognitive function in bipolar disorder suggest that patients experience impairment of memory and concentration during periods of illness, and some deficits persist after recovery. During an episode of acute mania, patients show deficits in sustained attention and verbal learning rather than in tests of executive function (Clark, et al., 2001). In some bipolar patients, cognitive deficits persist after the remission of acute symptoms, especially those with a chronic form of illness (Bearden et al., 2001). These studies are difficult to interpret because the states of depression and mania strongly influence the administration of tests and the subjects' motivation to take part. However, there is converging evidence from several studies that bipolar subjects, compared with controls, show impairments in verbal and visuospatial memory, and tasks requiring serial processing and higher-order cognitive functioning, such as abstraction. Results from some, but not all, studies distinguish bipolar from unipolar subjects. In one comparison of neuropsychological performance during an acute depressive episode, patients with bipolar disorder showed a higher degree of cognitive dysfunction connected with frontal lobe activity than patients with unipolar depression (Borkowska & Rybakowski, 2001). The difference between bipolar and unipolar patients could not be accounted for by differences in symptom severity or duration of the illness, and the level of cognitive dysfunction led the authors to suggest that there may be similarities between cognitive deficits observed in bipolar disorder and schizophrenia. Alterations in memory and executive function are not specific to bipolar disorder, and their neural correlates remain extremely speculative, but it has been suggested that these are consistent with impairment in the prefrontal cortex and striatal systems, as identified in imaging studies. A major question is whether the cognitive deficits found in patients during episodes of depression or of mania persist after full recovery when the patient is euthymic. Several studies on this have provided evidence for lasting deficits that are trait- rather than state-related variables. Both good- and poor-outcome bipolar patients performed worse than controls on a number of neuropsychological tests, and after controlling for age, premorbid IQ and depressive symptoms, it was found that executive function was consistently impaired (Ferrier et al.,
1999). Another study testing verbal learning, memory and executive function in euthymic patients and in controls found persisting impairment of verbal learning in bipolar subjects when fully recovered from a previous manic or depressive episode (Cavanagh et al., 2002).
Pharmacological treatments are central to the management of bipolar disorder in the acute phase and for the prevention of further episodes. Theories of the neurochemical basis of bipolar disorder have traditionally been based on knowledge of the targets of drugs known to be effective in the treatment of depression and mania, drugs known to cause mood changes and mood stabilisers effective in prophylaxis.
One of the earliest biological theories of mood disorder, the monoamine hypothesis, proposed that depression was due to a deficiency of the monoamine neurotransmitters noradrenalin (norepinephrenine), 5-hydroxytryptamine or serotonin (5-HT), and dopamine. This was based on the pharmacology of the first effective antidepressant drugs, the tricyclic antidepressants and monoamineoxidase inhibitors, which are known to increase the availability of monoamines at the synapse, in contrast to drugs, such as reserpine, that deplete monoamines and caused depression. Support came from biochemical and pharmacological studies of neurotransmitters, and their precursors and metabolites in serum, platelets and cerebrospinal fluid (CSF), and in post-mortem brain tissue, where receptor function has been directly and extensively studied (Stahl, 2000). Overall, evidence from these studies has been inconclusive. Direct measurement of brain monoamine receptors in post-mortem tissue has failed to reveal consistent changes linked to mood disorder, apart from the striking and consistent finding of increased 5-HT2 receptors in the frontal cortex of suicide victims. Noradrenalin metabolites are reduced in some depressed patients, and the main metabolite of 5-HT, 5-hydroxy indole acetic acid (5HIAA), is reduced in the CSF of depressed subjects.
However, the simple hypothesis that reduced monoamine availability at certain synapses is a cause of depression does not explain the delayed response to antidepressants, for, although antidepressants cause an immediate increase in monoamines, their therapeutic response is felt by the patient sometimes after a delay of several weeks. The focus of research has moved from neurotransmitters and their metabolites to their receptors, the control of gene expression regulating their synthesis and the post-synaptic signalling events of the downstream transmission of synaptic signals.
Long-term prophylaxis with mood stabilisers is an essential element of treatment for most patients. Lithium carbonate, first discovered as a treatment over 50 years ago, remains the first choice of mood stabiliser, firmly backed by clinical trials that prove its efficacy in the treatment and prophylaxis of mania and recurrent depression. The anticonvulsants sodium valproate (or valproate semisodium) and carbamazepine are alternatives to lithium and lamotrigine; moreover, an anticonvulsant is being increasingly used as a second-line treatment.
Post-synaptic signal transduction, the cascade of post-synaptic events set in train by the depolarisation of a monoamine receptor, involves a complex second-messenger system, part of which is the family of proteins called guanine nucleotide-binding proteins (G proteins). These bind to the post-synaptic receptor and are responsible for the further transmission of the signal initiated when the neurotransmitter binds to its receptor at the synapse. A number of enzymes, including inositol, modify the second-messenger system by binding to G proteins, and it is thought that lithium exerts its effects by depleting the level of inositol (Berridge et al., 1989).
The potential importance of the inositol system is enhanced by the recent discovery that three major mood stabilisers, lithium, carbamazepine and sodium valproate, have a common mode of action, causing inositol depletion via the cytoplasmic inositol-regulating protein prolyl oligopeptidase in a model system. Inositol depletion is likely to have an important effect in the regulation of signal transduction and indeed in neuronal growth (Williams et al., 2002). Much attention has been directed to these intracellular signalling pathways. G protein levels and function measured in peripheral blood mononuclear leucocytes are reported to be increased in mania, decreased in depression and altered in post-mortem tissue from bipolar patients. However, no DNA sequence variants associated with bipolar disorder have yet been detected in genes coding for proteins involved in these signal-transduction pathways (Avissar & Schreiber, 2002).
It has long been suspected and is now firmly established that bipolar disorder is familial, and there is a 10-fold increase in the risk of illness in a first-degree relative of someone with the disorder compared to the population risk. That this is partly due to genetic rather than purely environmental factors is confirmed by adoption studies and the well-replicated observation that concordance rates are significantly higher in identical than fraternal twins. However, the type of inheritance observed in families with bipolar disorder is not well understood. It is clear that the disorder is not usually caused by the dysfunction of any single gene or even two or three genes, and analysis of the segregation of the illness in families has not provided a clear explanation of how the illness is inherited. Diagnosis is essentially descriptive and is based on symptoms described by patients and observation of their behaviour, and there are no reliable biological markers to validate the descriptive definition. The disorder is most probably heterogeneous, encompassing several distinct disorders each with a different genetic basis. For example, an early age of onset or a maternal inheritance pattern may identify two subtypes. Some studies of the segregation of the disorder in families support a model in which single genes of large effect cause illness in families, and different genes are responsible for illness in different families (major genes with locus heterogeneity). This model is supported by some segregation analyses and linkage studies in large families (Blackwood et al., 2001; Blangero & Elston, 1989; Rice et al., 1987; Spence et al., 1995). However, there is evidence that subtypes of the disorder may be the result of the presence of several additive or interacting genes, each one alone being neither sufficient nor necessary for illness to develop (polygenic model). Further complexity arises when we consider other genetic effects that may be important in some forms of the illness. So-called epigenetic phenomena include anticipation, defined as an increase in illness severity and progressively earlier age at onset with each generation. The importance of anticipation, for which there is some evidence in familial bipolar disorder, is that, if present, it suggests a possible molecular mechanism to explain the clinical phenomenon. Other disorders showing anticipation, such as Huntington's disease, are caused by the expansion of unstable repeat DNA sequences in genes. Two other epigenetic phenomena of possible relevance to bipolar disorder are imprinting, describing a different expression of a disease when transmitted maternally as opposed to paternally, and mitochondrial inheritance, caused by a mutation in the mitochondrial genome, in which the disease is always transmitted from the maternal side. Each of these genetic hypotheses suggests candidate genes that have been investigated by linkage and association studies, as decribed below.
The success of linkage studies as the first step in mapping genes in other complex disorders, including Alzheimer's disease, diabetes and breast cancer, has given a strong stimulus to the search for genes in bipolar disorder (Baron, 2002; Craddock et al., 2001; Potash & DePaulo,
2000). Genetic linkage studies involve families where two or more members are affected by illness. The statistical analyses aim to detect the cosegregation of a genetic marker with the disease phenotype in a family. Studies can be based on extended, multiply affected families or collections of sibling pairs where both siblings are affected. Before the era of molecular genetics, one of the first linkages to be reported in bipolar disorder was with colour blindness and the glucose-6-phosphate dehydrogenase locus, highlighting the possibility of an X chromosome locus in bipolar disorder, since both of these markers were known to be on the X chromosome. Initial reports were not replicated, but subsequent linkage studies have maintained interest in a possible locus of bipolar disorder on the X chromosome (Baron, 2002). As a product of the Human Genome Project, many thousands of polymorphic markers are now available. These include microsatellites and single nucleotide polymorphisms (SNPs), each precisely mapped to a known location on a chromosome. Today, sequencing techniques can be used to detect SNPs directly, and a huge number have been generated by the SNP-mapping consortium and other sources, allowing the detailed mapping of large stretches of the genome (Taylor et al., 2001). In a typical linkage study, a series of polymorphic DNA markers, evenly spaced across the region of interest (that may include the whole genome), are typed with DNA obtained from family members. Typically, about 400 evenly spaced microsatellite markers may be used in a genome-wide scan for linkage. During the past decade, family linkage studies have identified several chromosome regions likely to harbour genes implicated in bipolar disorder. Recent results have been encouraging, and several chromosome regions have been identified in more than one linkage study. Chromosome regions where linkage has been confirmed or is suggestive include 1q, 4p, 6p, 10p, 10q, 12q, 13q, 18p, 18q, 21q, 22q and Xp. Further linkage studies may show that some of these are false-positive findings, but it is likely that some are true linkages. The task of finding genes in these regions by methods of association (linkage disequilibrium mapping) and direct sequencing of candidate genes is not trivial, because linkage typically has low resolution for locating genes and defines a broad chromosome region. For example, the candidate region identified by linkage on chromosome 4 may contain around 50 genes, several of which are good candidates for a role in mental disorders.
Was this article helpful?