Bipolar disorder continues to present many problems for psychiatrists, starting with the diagnosis itself. Since Kraepelin first distinguished manic depressive insanity from dementia praecox late in the nineteenth century, our concept of “pure” bipolar disorder, now called bipolar I, in which episodes of mania and depression alternate with periods of functioning has changed little. But extensive research and clinical experience have taught us that the clinical picture is more complicated: so-called mixed states and rapid cycling are common, most patients spend more time depressed than manic, and many patients have problematic mood elevation or instability but do not meet criteria for any of the three DSM diagnoses of bipolar I, bipolar II, and cyclothymic disorders. Hagop Akiskal has proposed a classification of several more “bipolar spectrum” syndromes with more subtle mood shifts and possible comorbidity with anxiety disorders.
In practical terms, this diagnostic fuzziness can make it difficult to know which patients need mood stabilizing medication to prevent recurrent mood episodes or suicide and who is at risk for exacerbation of mood problems by an antidepressant. Other clinical dilemmas include the frequent anxiety and attentional symptoms experienced by patients with bipolar disorder, and the difficulties treating bipolar depression.
With hopes that a better understanding of bipolar disorder's biology would point to some answers to these important clinical questions, I tackled a comprehensive review by Vladimir Maletic of the University of South Carolina and Charles Raison of the University of Arizona. (The authors had editorial support funded by a pharmaceutical company, but I do not see any evidence of bias in the article, and their findings are similar to the other reviews at the end of this post.) I was astounded by how much is known about the biology of bipolar disorder, but also by the number of different biological systems involved and the limited evidencemany findings. Here I will briefly outline the main areas of research, note particularly surprising or interesting findings, and conclude with some thoughts about what this all means for clinicians.
It is important to recognize the difficulties inherent in researching bipolar disorder. In addition to the symptom heterogeneity that plagues its clinical diagnosis, researchers must consider the current mood states of their subjects, since biological findings can be expected to vary with whether they are manic, hypomanic, euthymic, depressed, or in mixed states. Bipolar disorder generally begins in adolescence or young adulthood and frequently worsens with age, and some have proposed staging this disorder, which would include premorbid and prodromal stages, since biological abnormalities are found in those at risk for the disorder. Gender differences are also important, even though overall rates are similar in men and women. Medication, and even a history of medication, is likely to affect biological measurements. On a larger scale, the historical era, culture, ethnicity, and socioeconomic are likely to influence what behaviors are exhibited, considered abnormal, diagnosed, and treated, as well as measurable biological variables. And there are many practical difficulties in conducting research, from the ethics of studying untreated patients suffering from a disorder with significant morbidity and mortality to getting agitated manic patients to remain still in claustrophobia-inducing brain scanners.
Maletic and Raison note that genetic abnormalities in bipolar disorder overlap extensively with other psychiatric disorders, particularly schizophrenia and major depression. Many genes are involved, each with small to moderate effects, which interact with each other, with epigenetic phenomena, and with environmental influences. Many of these genes are involved with basic “housekeeping” functions of cellular metabolism, ion exchange, and synaptic development. Others regulate myelination, neurotransmission, neuronal plasticity, and apoptosis.
In the last decade a great deal of work has gone into identifying risk factors to predict the emergence and course of bipolar illness. These have included genes as well as behavioral phenomena such as sleep problems, aggression, anxiety, self esteem, and suicidality. While neither genetic or clinical findings have much predictive power, brain imaging is beginning to produce more robust predictors.
Structural neuroimaging fairly consistently finds larger lateral cerebral ventricles, which means less brain tissue, in patients with bipolar disorder, especially after multiple episodes. While functional imaging has found some abnormalities in specific areas of the brain, stronger findings associate the function of brain networks, rather than regions, with emotional, cognitive, behavioral, autonomic, neuroendocrine, immune, and circadian abnormalities. Two prefrontal-limbic networks may be important. The first is an “Automatic/Internal emotional regulatory network” connecting the ventromedial prefrontal cortex, subgenual anterior cingulate cortex, nucleus accumbens, globus pallidus, and thalamus. The authors note its overlap with the “salience network.” It modulates “endogenously generated feeling states, such as melancholic feelings induced by memories of past losses,” and appears to operate mainly unconsciously. The second is a Volitional/External regulatory network, involving the ventrolateral prefrontal cortex, mid and dorsal cingulate cortex, ventromedial striatum, globus pallidus, and thalamus, which overlaps with the first network. It is closely related to the executive control network. It “modulates externally induced emotional states, assists with voluntary (cognitive) emotional regulation, and suppresses maladaptive affect." These networks “collaboratively regulate amygdala responses in complex emotional circumstances.” In addition, abnormalities are found in the Default Mode network, involving midline structures: the subgenual anterior cingulate cortex, ventromedial prefrontal cortex, dorsolateral prefrontal cortex, precuneus, and medial temporal structures. Its functions include “self-reflection, processing social information, creative work, future planning, reminiscing, and conjuring autobiographical memories.”
Maletic and Raison next turn to changes in brain activation with mood state. Overall, elevated moods are associated with impaired prefrontal functioning resulting in “compromised regulation of limbic and paralimbic areas,” which many account for a number of manic symptoms. Bipolar depression is also associated with dysfunction in prefrontal areas and “inadequate modulation of limbic and subcortical areas;” this is most apparent in response to negative emotional stimuli. These neuroimaging findings clearly distinguish depression from other mood states in bipolar disorder. Abnormalities in euthymic bipolar patients are prominent in older patients with longer duration of illness, which supports the idea of bipolar disorder as a progressive disease.
Comparing findings in bipolar and unipolar depression, the evidence suggests greater impairment in volitional than automatic prefrontal-limbic emotional control circuitry in bipolar disorder.
In a very interesting section on pathohistological findings, the authors note that in contrast to typical neurodegenerative diseases, in which neurons are lost and glial cells multiply, bipolar disorder is associated with loss of all three types of glial cells. Most prominent is the loss of oligodendroglia, which produce the myelin which ensheathes nerve axons and increases their conducting speed; on the basis of neuroimaging and histologic evidence, the authors believe “oligodendroglial deficits may be the key CNS cellular abnormality in bipolar disorder.” Microglia, the main immune cells in the central nervous system, are substantially increased in the dorsolateral prefrontal cortex, anterior cingulate cortex, and mediodorsal thalamus in bipolar patients who commit suicide compared with bipolar patients who died of other causes, suggesting a role for immune dysfunction and inflammation in suicide which might be independent of the bipolar diagnosis.
Neuroendocrine and autonomic dysregulation in bipolar disorder is well established. The hypothalamus releases high levels of corticotropin releasing factor (CRF), causing the pituitary gland to release adrenocortical hormone (ACTH), which in turn stimulates the adrenal gland to release glucocorticoids, which among many other effects stimulate glucocorticoid receptors in the amygdala. Individuals with bipolar disorder, even when euthymic, have less responsive glucocorticoid receptors, and this increases with multiple episodes.
In addition, bipolar disorder is associated with excessive sympathetic and reduced parasympathetic nervous system activity. The combination of sympathetic overactivity, parasympathetic underactivity, glucocorticoid receptor insensitivity, and increased inflammatory signaling may contribute to increased risk of metabolic syndrome and vascular disease. The authors cite evidence that vascular disease rather than suicide is the leading cause of excess death in bipolar disorder. Elevated glucocorticoids also suppress thyroid stimulating hormone (TSH) secretion and interfere with the conversion of thyroxine (T4) to the more biologically active triiodothyronine (T3).
Many biological phenomena that vary with the 24-hour circadian cycle are disrupted in bipolar disorder. These include sleep; mood; body temperature; and corticosteroid, catecholamine, serotonin, and melatonin levels. Much of this is coordinated by the suprachiasmatic nucleus of the hypothalamus, which sends projections to several other areas of the brain. Both melatonin and corticosteroids participate prominently in this entrainment of diurnal rhythms. Interpersonal and social rhythm therapy, a behavioral treatment based on resolving stressful interpersonal problems and maintaining regular daily rhythms, has been found helpful for patients with bipolar disorder. Light therapy might also be useful, but it has been reported to precipitate mood instability and suicidality.
Immune dysfunction is another important aspect of bipolar disorder. Limbic areas such as the amygdala, insula, and anterior cingulate cortex participate in the regulation of immune functions. Peripheral inflammatory cytokines are elevated in both bipolar depressed and manic patients. Although the brain is generally isolated from the rest of the body by capillary endothelial cells (the blood-brain barrier), several pathways allow peripheral inflammatory signaling to enter the brain. Imaging research shows that artificial stimulation of peripheral cytokines changes activity of limbic and paralimbic areas. Inflammatory cytokines activate microglia, which release other inflammatory molecules. In Maletic and Raison’s words, this “chemical cocktail of oxidative stress and inflammatory signals” produces changes in astrocyte function, which reduces production of neurotrophins and alters glutamate neurotransmission. In addition, cytokines increase the production of serotonin and dopamine transporters, disrupting monoamine signaling. One study found that increased expression of inflammatory genes correlated with greater hemodynamic response to emotional stimuli in limbic areas in patients with mood disorder. These inflammatory disturbances are likely related to the high rates of cardiovascular, respiratory, and gastrointestinal disorders in patients with bipolar disorder.
Alterations in monoamine neurotransmitters—norepinephrine, dopamine, and serotonin—have long been viewed as central to the pathophysiology of bipolar disorder. Monoamine excess was seen as corresponding to mania and monoamine deficit to depression, but more recent neuroimaging studies have been inconclusive. There is stronger evidence for altered glutamate signaling, which involves glial cells and interacts with the endocrine dysfunctions already discussed.
It has for some time been clear that actions at neurotransmitters receptors trigger cascades of chemical signaling within nerve cells which reach into the nucleus and alter gene expression. Most drugs used to treat bipolar disorder act on these intracellular signaling pathways. There is evidence that some of these stress-activated kinase pathways regulate oligodendroglia and the myelination of nerve axons.
Finally, many of the genes that have been linked to bipolar disorder code for mitochondrial proteins, indicating dysfunction at the most basic levels of cellular energy utilization.
Maletic and Raison conclude that “from a neurobiological perspective, there is no such things as bipolar disorder,” since many different pathologies appear to underlie similar clinical syndromes. They note that attempts to localize dysfunction in bipolar disorder with fMRI have produced remarkably inconsistent results. More consistent findings involve dysregulation of glial-neuronal interactions, especially overactivity of microglia, the brain’s main immune cells, along with inflammation in the rest of the body in both the depressed and manic states. Many of these findings overlap with other disorders, including schizophrenia and major depression. Many of the genetic abnormalites associated with bipolar disorder are not primarily associated with neuronal functioning but rather with basic “housekeeping” functions necessary for most cells of the body. This, as well as the disturbances in immune function, are likely related to bipolar disorder’s high cormorbidity with diseases of other body systems.
There is a great deal of information to metabolize here. The neuroimaging research points to limbic-emotional and frontal-control neurocircuits. Within those circuits, the pathophysiology appears to include impairment of oligodendroglial-modulated myelination; inflammatory/immune dysfunctions of microglia; and neurotransmitter abnormalities involving astrocytes; all of which may include disordered intracellular signaling and mitochondrial dysfunction. These interact with the circadian clock, HPA axis, and other bodily systems to produce the derangements of mood, energy, cognition, and behavior we recognize as bipolar disorder. Maletic and Raison argue for dynamic models of these dysfunctions, rather than static categorical diagnoses or unidimensional traits.
In a similar review, Ather Muneer proposes a more integrated cyclical model in which HPA axis dysfunction produces pro-inflammatory cytokines, which in turn produce reactive oxygen species, leading to increases in neurotoxic kynurenine metabolites, which disrupt the circadian molecular clock, leading back to HPA axis dysfunction. Andreazza and Young have a more complex multidimensional model. It may be that stress-induced dysfunction in any of several biological systems that are connected by multiple self-reinforcing feedback loops tends to be amplified, and if homeostatic checks are weak, will expand to produce overt signs and symptoms.
If this is true, any overall concept of bipolar disorder will have to be a systems-based model. Bipolar disorder will not turn out to be several distinct disorders. Each patient will have a unique combination of dysfunctions in many biological systems.
While this means no single treatment or combination of treatments is likely to work for all patients with bipolar disorder, its large number of biological dysfunctions means there are many targets for treatment. We can hope for biomarkers to guide treatment selection—patients with prominent inflammatory abnormalities might benefit from targeted anti-inflammatory treatment, while those with neurotransmitter or endocrine dysfunctions might do better with other approaches. We can watch the literature for convincing trials of such biomarker-guided treatments.
Bipolar disorder is a serious illness that reaches deep into the basic biological functions of the brain and body. It is associated with endocrine, cardiovascular, and gastrointestinal disorders, among others. Since it is progressive, aggressive treatment is likely to reduce medical illness and mortality as well as prevent psychiatric disability and suicide.