Decoding the Biology of Bipolar Disorder: An Update on Recent Findings in Genetics, Imaging, and Immunology

Family history has long been recognized as an important clinical feature of bipolar disorder. Family, twin, and adoption studies all support a significant genetic burden in bipolar disorder (1). Tremendous effort has been made to study the genetic risk factors for bipolar disorder. Early investigations into chromosomal aberrations identified candidate genes through linkage studies (2). Although linkage studies have been very successful in identifying a single or a small number of genes as causative factors in diseases such as Huntington disorder, linkage studies are less useful in studying the genetics in psychiatric illnesses including bipolar disorder. More recently, genome-wide association studies, which are more powerful than linkage studies, have been used to assess for multiple causative disease genes without a priori assumptions about which genes might be involved. It is now widely accepted that many genes, each with small effects, cumulatively contribute to the pathogenesis of bipolar disorder. These possible genetic loci (among others) include CACNA1C on chromosome 12p13, which encodes an α-1 subunit of a voltage-dependent calcium channel; ANK3 on chromosome 10q21, which encodes ankyrins, a family of proteins that are believed to connect integral membrane proteins to the spectrin-actin cytoskeleton; and DGKH on chromosome 13q14, which encodes diacylglycerol kinase that may be implicated in the mood-stabilizing effects of lithium (3–7). A recent analysis of combined genome-wide association studies samples from three large effectiveness studies of schizophrenia (CATIE), bipolar disorder (STEP-BD), and major depressive disorder (STAR*D) demonstrated that genetic variants near the adrenomedullin (ADM) gene on chromosome 11p15 may be specific to bipolar II disorder (8). However, no single causative gene for bipolar disorder has been identified to date.

The major challenge for genetic studies in bipolar disorder is that bipolar disorder is phenotypically heterogeneous; many bipolar symptoms are also seen in other psychiatric conditions such as schizophrenia, schizoaffective disorder, and major depressive disorder. It is therefore not surprising that considerable genetic overlap, such as common polymorphisms, have been found between bipolar disorders and schizophrenia, more so than previously suspected (9–11). Although further studies are needed to elucidate the fundamental genetic pathway for bipolar disorder and other psychiatric disorders, the possible genetic association between different psychiatric diseases suggests that a common neurodevelopmental process might exist in these disorders, supporting the merits of early intervention. The diagnostic heterogeneity in bipolar disorder also suggests that a better classification system for the disease may be required to make further progress in defining the genetic etiologies (3).

Finally, epigenetics (gene expression) investigates gene and environment (G×E) interplay. There has been some interesting G×E data that suggest mechanisms whereby genetically vulnerable individuals might or might not develop bipolar disorder in a stressful environment. Twin studies have shown that decreases in white matter were more likely related to the genetic risk of developing bipolar disorder, whereas a strong environmental influence was found for cortical gray matter. Lithium use was found to accentuate the environment effects (12). Moreover, family studies of lithium responses across generations by Grof et al. (13) suggest that lithium responders may represent a distinct subgroup of bipolar disorder with characteristic clinical profiles, which may be observed in their affected offspring as well. Lithium responders may have a unique genetic makeup that is distinct from that of individuals with other bipolar subtypes. In fact, clinicians have long used family history to inform their diagnostic and treatment formulation. However, the biological markers of lithium response remain elusive despite intensive research (13). Padmos et al. (14) demonstrated that except for a small set of genes the majority of monocyte proinflammatory gene expression in bipolar disorder (structural equation modeling data: 94% [95% confidence interval, 53%–99%]) was related to shared environment. Although we still do not fully understand the precise mechanism of the G×E interaction, it is encouraging to see these recent findings begin to reveal the biological foundation of psychosocial and behavioral intervention in bipolar disorder.

Neuroimaging studies of bipolar disorder have been rapidly increasing in recent decades. However, reports have been inconsistent largely because of heterogeneous sample selections, imaging methods, and small sample sizes. Despite the fact that no imaging markers have been established yet to aid diagnosis and treatment, structural and functional imaging studies suggest the presence of subtle brain abnormalities in bipolar disorder.

Structurally, for example, bipolar disorder has been associated with lateral ventricular enlargement, white matter changes, and a decrease in total cortical volume (12, 15–17). A recent international collaboration compared structural magnetic resonance imaging (MRI) differences among 321 individuals with bipolar I disorder and 442 healthy control subjects and found that patients with bipolar disorder had increased right lateral ventricular, left temporal lobe, and right putamen volumes. In addition, cerebral volume reduction may be associated with illness duration in bipolar disorder (15). The authors also suggest that lithium may exert mood-stabilizing effects through its neurotropic effects. Hippocampal and amygdala enlargement were found in patients with bipolar disorder taking lithium versus patients not treated with lithium and healthy control subjects (15). A smaller (N=28) study by Moore et al. (18) observed significant increases in total brain gray matter volume in subjects with bipolar disorder after 4 weeks of lithium treatment (plasma level at ∼0.8 mEq/liter×3 weeks); interestingly, only lithium responders (>50% decrease in Hamilton Depression Rating Scale total score) showed a significant increase in gray matter volume in the prefrontal cortex. For more information regarding the neurotrophic effects of lithium, please see a review by Quiroz et al. (19).

Functional imaging technologies, such as positron emission tomography, single-photon emission computed tomography, and functional MRI detect changes in brain metabolism or blood flow in vivo when subjects are either at rest or during cognitive tasks designed to elicit activities of specific neural circuitries. Although there are inconsistencies across studies, data generally suggest that there is frontal hypoactivity and limbic hyperactivity in individuals with bipolar disorder relative to that in healthy control subjects (20). More specifically, a recent meta-analysis of 65 functional MRI studies by Chen et al. (21) showed deactivation in the inferior frontal cortex or ventrolateral prefrontal cortex, particularly in manic but not in euthymic and depressed states. With the exception of the amygdala, hyperactivation was seen in all limbic regions, including medial temporal structures (parahippocampal gyrus, hippocampus, and amygdala) and basal ganglia across all mood states. In the amygdala, hyperactivation was observed only during euthymia when region-of-interest studies were included. Finally, hypoactivation in the inferior frontal cortex was elicited by both cognitive and emotional tasks, whereas increased limbic activation was mainly related to emotional processing.

Further, a recent meta-analysis of 10 whole-brain diffusion tenor imaging studies by Vederine et al. (17) has demonstrated that compared with healthy control subjects, patients with bipolar disorder have two significant clusters of decreased fractional anisotropy within the right side of the brain: one was located in the right white matter, close to the parahippocampal gyrus; the other was located close to the right anterior and subgenual cingulate cortex (17).

More evidence is emerging to support the use of imaging markers for the diagnosis of bipolar disorder and differentiation of bipolar disorder from unipolar depression. Researchers in Pittsburgh, Pennsylvania, found that when compared with patients with unipolar depression, adult patients with bipolar disorder has decreased white matter fractional anisotropy in the left occipitotemporal and primary sensory regions, suggesting that visuospatial and sensory processing in bipolar disorder may be different from that in unipolar depression (22). Other interesting findings from the same group suggested that the abnormal hyperactivation in the left amygdala in response to mild sad and neutral faces might be a depression-specific marker in bipolar disorder but not unipolar depression (23). In addition, increased activity in response to sad stimuli in the right amygdala-orbitofrontal cortical (OFC) functional connectivity (FC) might represent a trait marker for depression, whereas abnormally elevated left amygdala-OFC FC to sad stimuli and abnormally reduced amygdala-OFC FC to intense happy stimuli might represent a depression state marker in bipolar disorder (24).

Other researchers have found that, compared with patients with major depression, patients with bipolar disorder had higher rates of deep white matter hyperintensities, a smaller corpus callosum cross-sectional area, and bigger hippocampus and basal ganglia (25). Delvecchio et al. (26) suggested that decreased ventrolateral prefrontal cortical activity in bipolar disorder along with relative hypoactivation of the sensorimotor cortices in unipolar depression and increased responsiveness in the thalamus and basal ganglia were associated with bipolar disorder. Meanwhile, compared with healthy control subjects, both patients with bipolar disorder and those with unipolar depression shared some common features, such as increased lateral ventricle volume and increased rates of subcortical gray matter hyperintensities as well as limbic hyperactivation.

In recent decades, evidence demonstrating altered inflammatory response in major mental illnesses, such as major depressive disorder and schizophrenia, has been accumulating. Evidence from unipolar depression suggests that the interaction between proinflammatory cytokines, the hypothalamic-pituitary-adrenal axis, and central monoamines may be implicated in the pathogenesis of depression through neuronal damage and degeneration (27, 28). In contrast, the role of inflammation in bipolar disorder has not been investigated until recently.

A recent systematic review of publications between 1950 up to April 2008 by Goldstein et al. (29) summarized findings from 27 articles on cytokines. Data generally support the fact that there are increased proinflammatory markers, such as C-reactive protein, soluble interleukin (IL)-2 receptor, IL-6, and tumor necrosis factor-α, during mania (29). However, the change in anti-inflammatory markers (e.g., IL-4 and IL-10) or imbalance between pro- and anti-inflammatory markers is not consistent among studies. Furthermore, compared with unipolar depression and mania, inflammatory alteration has been less studied during bipolar depression and euthymia.

The interplay between the immune system and bipolar disorder is very complex at multiple levels; for a detailed review, see Drexhage et al. (30). At the gene expression level, about two-thirds of patients with bipolar disorder were found to have a higher expression of a coherent set of 34 inflammatory genes—an inflammatory gene expression “signature”—in their circulating monocytes (30, 31). This high inflammatory set point of circulating monocytes at the transcriptome level was found in both schizophrenia and bipolar disorder, with differences in inflammatory gene expression (30, 32). Aside from monocyte lineage, the authors also reported altered functions in T-cell networks among patients with bipolar disorder (33).

Finally, some have suggested that the mechanism of mood-stabilizing agents may involve the inhibition of cyclooxygenase-2 (COX-2) and a reduction in proinflammatory cytokines; lithium, as well as antipsychotic drugs, was found to decrease expression of some inflammatory genes (31). Although the specific mechanism underlying inflammation and pathogenesis of bipolar disorder has not been identified, researchers have speculated that anti-inflammatory medications may be useful as a novel treatment for bipolar disorder. In fact, celecoxib (a selective COX-2 inhibitor) was tested as an adjunct to treat depression or mixed episodes of bipolar disorder; unfortunately, the results have not been very encouraging (34).

The inflammation story of bipolar disorder does not end here; the increasing medical burden of bipolar disorder has been gaining recognition in the medical community (35). Moreover, inflammation is well known to relate to many medical conditions such as arthritis, obesity, diabetes, cardiovascular diseases, and pain, as well as behaviors such as smoking, alcohol use, physical activities, diet, and sleep, which are commonly problematic among patients with bipolar disorder. Together, medical comorbidity and behaviors may further worsen the added inflammatory burden to bipolar disorder.

Brain-derived neurotrophic factor (BDNF) appears to play a critical role in neurogenesis and neuroplasticity. Some proposed that BDNF may mediate the progression of bipolar disorder via neuroplasticity (36). Although the findings have not been consistent and the precise mechanism of BDNF has not been defined, there are trends suggesting that BDNF levels are decreased during manic and depressive episodes and then recover after treatment for acute mania. Interestingly, BDNF has been observed to decrease with age and length of illness in euthymia as well (37, 38). Because BDNF crosses the blood-brain barrier, its levels in serum are highly correlated with its levels in CSF. The authors suggest that peripheral BDNF may serve as a biomarker of mood states and disease progression for bipolar disorder (38).

Bipolar disorder is one of the top 10 most disabling illnesses worldwide. It is biologically complex with abnormalities at the molecular level as well as within neural networks and within the brain structure. Because the biological basis of bipolar disorder is not well understood, patients with this illness are often misdiagnosed, mistreated, or undertreated, leading to less-than-optimal outcome. It is therefore crucial to identify biological markers and to decode the pathophysiological process of bipolar disorder. Although much still remains unknown, it is encouraging to see that the field of biological research in bipolar disorder is rapidly evolving with a lot of informative findings on the horizon, which ultimately will translate into a better diagnosing system and novel and personalized treatment for this devastating condition.

Writing of this manuscript was supported by the Office of Academic Affiliations, Advanced Fellowship Program in Mental Illness Research and Treatment, Department of Veterans Affairs.