Mitochondrial dysfunction and pathology in bipolar disorder and schizophrenia

Bipolar disorder (BPD) and schizophrenia (SZ) are severe psychiatric illnesses with a combined prevalence of 4%. A disturbance of energy metabolism is frequently observed in these disorders. Several pieces of evidence point to an underlying dysfunction of mitochondria: (i) decreased mitochondrial respiration; (ii) changes in mitochondrial morphology; (iii) increases in mitochondrial DNA (mtDNA) polymorphisms and in levels of mtDNA mutations; (iv) downregulation of nuclear mRNA molecules and proteins involved in mitochondrial respiration; (v) decreased high-energy phosphates and decreased pH in the brain; and (vi) psychotic and affective symptoms, and cognitive decline in mitochondrial disorders. Furthermore, transgenic mice with mutated mitochondrial DNA polymerase show mood disorder-like phenotypes. In this review, we will discuss the genetic and physiological components of mitochondria and the evidence for mitochondrial abnormalities in BPD and SZ. We will furthermore describe the role of mitochondria during brain development and the effect of current drugs for mental illness on mitochondrial function. Understanding the role of mitochondria, both developmentally as well as in the ailing brain, is of critical importance to elucidate pathophysiological mechanisms in psychiatric disorders.     

Mice with neuron-specific accumulation of mitochondrial DNA mutations show mood disorder-like phenotypes

There is no established genetic model of bipolar disorder or major depression, which hampers research of these mood disorders. Although mood disorders are multifactorial diseases, they are sometimes manifested by one of pleiotropic effects of a single major gene defect. We focused on chronic progressive external ophthalmoplegia (CPEO), patients with which sometimes have comorbid mood disorders. Chronic progressive external ophthalmoplegia is a mitochondrial disease, which is accompanied by accumulation of mitochondrial DNA (mtDNA) deletions caused by mutations in nuclear-encoded genes such as POLG (mtDNA polymerase). We generated transgenic mice, in which mutant POLG was expressed in a neuron-specific manner. The mice showed forebrain-specific defects of mtDNA and had altered monoaminergic functions in the brain. The mutant mice exhibited characteristic behavioral phenotypes, a distorted day-night rhythm and a robust periodic activity pattern associated with estrous cycle. These abnormal behaviors resembling mood disorder were worsened by tricyclic antidepressant treatment and improved by lithium, a mood stabilizer. We also observed antidepressant-induced mania-like behavior and long-lasting irregularity of activity in some mutant animals. Our data suggest that accumulation of mtDNA defects in brain caused mood disorder-like mental symptoms with similar treatment responses to bipolar disorder. These findings are compatible with mitochondrial dysfunction hypothesis of bipolar disorder.     

What Can Mitochondrial DNA Analysis Tell Us About Mood Disorders?

Variants in mitochondrial DNA (mtDNA) and nuclear genes encoding mitochondrial proteins in bipolar disorder, depression, or other psychiatric disorders have been studied for decades, since mitochondrial dysfunction was first suggested in the brains of patients with these diseases. Candidate gene association studies initially resulted in findings compatible with the mitochondrial dysfunction hypothesis. Many of those studies, however, were conducted with modest sample sizes (N < 1000), which could cause false positive findings. Furthermore, the DNA samples examined in these studies, including genome-wide association studies, were generally derived from peripheral tissues. One key unanswered question is whether there is an association between mood disorders and somatic mtDNA mutations (deletions and point mutations) in brain regions that accumulate a high amount of mtDNA mutations and/or are involved in the regulation of mood. Two lines of robust evidence supporting the importance of mtDNA mutations in brain tissues for mood disorders have come from clinical observation of mitochondrial disease patients who carry primary mtDNA mutations or accumulate secondary mtDNA mutations due to nuclear mutations and an animal model study. More than half of mitochondrial disease patients have comorbid mood disorders, and mice with neuron-specific accumulation of mtDNA mutations show spontaneous depression-like episodes. In this review, we first summarize the current knowledge of mtDNA and its genetics and discuss what mtDNA analysis tells us about neuropsychiatric disorders based on an example of Parkinson’s disease. We also discuss challenges and future directions beyond mtDNA analysis toward an understanding of the pathophysiology of “idiopathic” mood disorders.     

The other, forgotten genome: mitochondrial DNA and mental disorders.

This paper summarizes recent research on mitochondrial DNA (mtDNA)--which might be described as the "other, forgotten genome". Recent studies suggest the possible pathophysiological significance of mtDNA in schizophrenia and neurodegenerative and mood disorders. Decreased activity of the mitochondrial electron transport chain has been implicated in both Parkinson's and Alzheimer's disease and while age-related accumulation of mtDNA deletions has been suggested as a possible cause, there is no concrete evidence that particular mtDNA polymorphisms are responsible. In schizophrenia, the activity and/or mRNA expression of complex IV are involved, but the direction of the alteration is not the same and there is no evidence linking schizophrenia with mtDNA. In bipolar disorder, there is some evidence of parent-of-origin effects and association with mtDNA polymorphisms but further investigation is needed to elucidate the role of mtDNA in mental disorders.     

Reduced intracellular pH in the basal ganglia and whole brain measured by 31P-MRS in bipolar disorder

The authors have previously reported that intracellular pH measured by phosphorus-31 magnetic resonance spectroscopy (31P-MRS) was decreased in the frontal lobes of patients with bipolar disorder. In the present study, phosphorus metabolism in the basal ganglia was examined in 13 patients with bipolar disorder and 10 matched controls by localized 31P-MRS. While no significant alteration of peak area ratios was found for all phosphorus metabolites, intracellular pH was significantly reduced in the basal ganglia in patients with bipolar disorder (7.014 +/- 0.045) compared with control subjects (7.066 +/- 0.047, P < 0.05). Unexpectedly, non-localized 31P-MR spectra also showed significantly lower levels of intracellular pH (6.970 +/- 0.025) than controls (6.986 +/- 0.024, P < 0.05). These results suggest that decreased intracellular pH in the brain of patients with bipolar disorder is not caused by dysfunction of the frontal lobes but reflect altered metabolism at the cellular level.     

Inherited and somatic mitochondrial DNA mutations in Guam amyotrophic lateral sclerosis and parkinsonism-dementia

There is increasing evidence for mitochondrial dysfunction in neurodegenerative disorders, although the exact role of mitochondrial DNA (mtDNA) mutations in this process is unresolved. We investigated inherited and somatic mtDNA substitutions and deletions in Guam amyotrophic lateral sclerosis (ALS) and parkinsonism-dementia (PD). Hypervariable segment 1 sequences of Chamorro mtDNA revealed that the odds ratio of a PD or ALS diagnosis was increased for individuals in the E1 haplogroup while individuals in the E2 haplogroup had decreased odds of an ALS or PD diagnosis. Once the disorders were examined separately, it became evident that PD was responsible for these results. When the entire mitochondrial genome was sequenced for a subset of individuals, the nonsynonymous mutation at nucleotide position 9080, shared by all E2 individuals, resulted in a significantly low odds ratio for a diagnosis of ALS or PD. Private polymorphisms found in transfer and ribosomal RNA regions were found only in ALS and PD patients in the E1 haplogroup. Somatic mtDNA deletions in the entire mtDNA genome were not associated with either ALS or PD. We conclude that mtDNA haplogroup effects may result in mitochondrial dysfunction in Guam PD and reflect Guam population history. Thus it is reasonable to consider Guam ALS and PD as complex disorders with both environmental prerequisites and small genetic effects.     

Mitochondrial Etiology of Neuropsychiatric Disorders

The brain has the highest mitochondrial energy demand of any organ. Therefore, subtle changes in mitochondrial energy production will preferentially affect the brain. Considerable biochemical evidence has accumulated revealing mitochondrial defects associated with neuropsychiatric diseases. Moreover, the mitochondrial genome encompasses over a thousand nuclear DNA genes plus hundreds to thousands of copies of the maternally inherited mitochondrial DNA (mtDNA). Therefore, partial defects in either the nuclear DNA or mtDNA genes or combinations of the two can be sufficient to cause neuropsychiatric disorders. Inherited and acquired mtDNA mutations have recently been associated with autism spectrum disorder, which parallels previous evidence of mtDNA variation in other neurological diseases. Therefore, mitochondrial dysfunction may be central to the etiology of a wide spectrum of neurological diseases. The mitochondria and the nucleus communicate to coordinate energy production and utilization, providing the potential for therapeutics by manipulating nuclear regulation of mitochondrial gene expression.     

Frontal lobe bioenergetic metabolism in depressed adolescents with bipolar disorder: a phosphorus-31 magnetic resonance spectroscopy study

Shi X‐F, Kondo DG, Sung Y‐H, Hellem TL, Fiedler KK, Jeong E‐K, Huber RS, Renshaw PF. Frontal lobe bioenergetic metabolism in depressed adolescents with bipolar disorder: a phosphorus‐31 magnetic resonance spectroscopy study. Bipolar Disord 2012: 14: 607–617. © 2012 The Authors. Journal compilation © 2012 John Wiley & Sons A/S. Objectives: To compare the concentrations of high‐energy phosphorus metabolites associated with mitochondrial function in the frontal lobe of depressed adolescents with bipolar disorder (BD) and healthy controls (HC). Methods: We used in vivo phosphorus‐31 magnetic resonance spectroscopy (³¹P‐MRS) at 3 Tesla to measure phosphocreatine (PCr), beta‐nucleoside triphosphate (β‐NTP), inorganic phosphate (Pi), and other neurometabolites in the frontal lobe of eight unmedicated and six medicated adolescents with bipolar depression and 24 adolescent HCs. Results: Analysis of covariance, including age as a covariate, revealed differences in PCr (p = 0.037), Pi (p = 0.017), and PCr/Pi (p = 0.002) between participant groups. Percentage neurochemical differences were calculated with respect to mean metabolite concentrations in the HC group. Post‐hoc Tukey–Kramer analysis showed that unmedicated BD participants had decreased Pi compared with both HC (17%; p = 0.038) and medicated BD (24%; p = 0.022). The unmedicated BD group had increased PCr compared with medicated BD (11%; p = 0.032). The PCr/Pi ratio was increased in unmedicated BD compared with HC (24%; p = 0.013) and with medicated BD (39%; p = 0.002). No differences in β‐NTP or pH were observed. Conclusions: Our results support the view that frontal lobe mitochondrial function is altered in adolescent BD and may have implications for the use of Pi as a biomarker. These findings join volumetric studies of the amygdala, and proton MRS studies of n‐acetyl aspartate in pointing to potential differences in neurobiology between pediatric and adult BD.     

Quantification of total mitochondrial DNA and mitochondrial common deletion in the frontal cortex of patients with schizophrenia and bipolar disorder

Summary Data published during the last decade are suggestive of a role for mitochondrial dysfunction in the pathogenesis of schizophrenia, bipolar disorder and other psychiatric diseases. In order to determine if the mitochondrial deficits reported in the literature are caused by abnormalities in the mitochondrial DNA of psychiatric patients, we quantified mitochondrial DNA (mtDNA) levels and the 5 kb common mitochondrial deletion (CD) in postmortem frontal cortex tissue. The mitochondrial CD and mtDNA levels were measured in tissue obtained from the frontal cortex (Brodmann Area 46) of 144 individuals (45 patients with schizophrenia, 40 patients with bipolar disorder, 44 controls, and 15 patients with major depression). These variables were measured using newly developed SYBR green and TaqMan real time PCR assays. Both the TaqMan and the SYBR green assays gave similar results. There was no statistically significant difference for the quantity of the common mitochondrial deletion between controls and patients. We also did not detect a difference in the mtDNA levels amongst the diagnosis groups. There were statistically significant differences for the evaluated parameters for smokers, schizophrenic patients on antipsychotic drugs at time of death, and bipolar patients with antidepressant use and alcohol abuse. Based on this study and other reports, we conclude that neither the common mitochondrial deletion nor changes in mitochondrial DNA levels are likely to account for the mitochondrial changes associated with bipolar disorder or schizophrenia. The effect of premortem agonal factors and medication on mitochondrial dysfunction still needs further elucidation. First two authors contributed equally to this paper     

Mitochondrial-related gene expression changes are sensitive to agonal-pH state: implications for brain disorders

Mitochondrial defects in gene expression have been implicated in the pathophysiology of bipolar disorder and schizophrenia. We have now contrasted control brains with low pH versus high pH and showed that 28% of genes in mitochondrial-related pathways meet criteria for differential expression. A majority of genes in the mitochondrial, chaperone and proteasome pathways of nuclear DNA-encoded gene expression were decreased with decreased brain pH, whereas a majority of genes in the apoptotic and reactive oxygen stress pathways showed an increased gene expression with a decreased brain pH. There was a significant increase in mitochondrial DNA copy number and mitochondrial DNA gene expression with increased agonal duration. To minimize effects of agonal-pH state on mood disorder comparisons, two classic approaches were used, removing all subjects with low pH and agonal factors from analysis, or grouping low and high pH as a separate variable. Three groups of potential candidate genes emerged that may be mood disorder related: (a) genes that showed no sensitivity to pH but were differentially expressed in bipolar disorder or major depressive disorder; (b) genes that were altered by agonal-pH in one direction but altered in mood disorder in the opposite direction to agonal-pH and (c) genes with agonal-pH sensitivity that displayed the same direction of changes in mood disorder. Genes from these categories such as NR4A1 and HSPA2 were confirmed with Q-PCR. The interpretation of postmortem brain studies involving broad mitochondrial gene expression and related pathway alterations must be monitored against the strong effect of agonal-pH state. Genes with the least sensitivity to agonal-pH could present a starting point for candidate gene search in neuropsychiatric disorders.     

Differential brain activation during response inhibition in bipolar and attention-deficit hyperactivity disorders

Aim: To identify differential patterns of brain activation between adolescents with bipolar disorder and adolescents with attention‐deficit hyperactivity disorder (ADHD) to better understand the neurophysiology of both disorders. We hypothesized that subjects with ADHD would show altered activation in brain regions involved in executive and sustained attention. In contrast, we hypothesized that bipolar subjects would show altered brain activation in regions responsible for emotionally homeostasis, including the striatum and amygdala. Methods: Functional magnetic resonance imaging was performed during a continuous performance task with a response inhibition component in 11 adolescents with bipolar disorder during a manic episode, 10 adolescents with ADHD, and 13 healthy adolescents. Results: There were no differences in behavioural performance among the three groups. Compared with bipolar subjects, subjects with ADHD showed increased activation in the superior temporal lobe during successful response inhibition. Although bipolar subjects did not show activation differences in the striatum or amygdala compared with ADHD subjects, increased left parahippocampal activation in the bipolar group was associated with increased manic symptoms. Conclusions: The patterns of brain activation observed in the current study support divergent patterns of neurophysiological dysfunction in individuals with bipolar disorder as compared with those with ADHD. Therefore, the impulsive behaviour seen in both disorders may be the consequence of dysfunction in different brain regions, and further research may help identify neurobiological markers that are specific to each condition.     

Novel heteroplasmic mutation in the anticodon stem of mitochondrial tRNALys associated with dystonia and stroke‐like episodes

Gal A, Pentelenyi K, Remenyi V, Pal Z, Csanyi B, Tomory G, Rasko I, Molnar MJ. Novel heteroplasmic mutation in the anticodon stem of mitochondrial tRNA^Lys associated with dystonia and stroke‐like episodes.
Acta Neurol Scand: 2010: 122: 252–256.
© 2009 The Authors Journal compilation © 2009 Blackwell Munksgaard. Objectives – We report a novel heteroplasmic mitochondrial tRNA^Lys mutation associated with dystonia, stroke‐like episodes, sensorineural hearing loss and epilepsy in a Hungarian family. Material and methods – A 16‐year‐old boy, his brother and mother were investigated. Thorough clinical investigation as well as electrophysiological, neuroradiological and myopathological examinations were performed. Molecular studies included the analysis of the DYT1, DDP1/TIMM8A (deafness‐dystonia peptid‐1) genes and mitochondrial DNA (mtDNA). Results –  The mtDNA analysis of the proband revealed a heteroplasmic A8332G substitution in the anticodon stem of the tRNA^Lys gene. The mutation segregated in all affected family members. Besides this mutation 16 further mtDNA polymorphisms were detected. Complex I activity of the patient’s fibroblast cultures showed decreased activity confirming mitochondrial dysfunction. Conclusion –  The novel A8332G heteroplasmic mutation is most likely a new cause of dystonia and stroke‐like episodes due to mitochondrial encephalopathy. The synergistic effect of the G8697A, A11812G and T10463C single nucleotide polymorphisms may modify the phenotype.     

Early-onset cerebellar ataxia,myoclonus and hypogonadism in a case of mitochondrial complex III deficiency treated with vitamins K3 and C

A 16-year-old girl presented with early-onset cerebellar ataxia, myoclonus, elevated lactic acidosis and hypogonadotropic hypogonadism. Muscle biopsy specimens revealed fibres with a ragged appearance with increased mitochondria and lipid droplets. Biochemical investigation revealed a deficiency of complexbc ₁ (complex III) of the mitochondrial respiratory chain. Genetic analysis did not show either deletions or known mutations of mitochondrial DNA (mtDNA). Phosphorus magnetic resonance spectroscopy (³¹P-MRS) showed defective energy metabolism in brain and gastrocnemius muscle. A decreased phosphocreatine (PCr) content was found in the occipital lobes accompanied by normal inorganic phosphate (Pi) and cytosolic pH. These findings represented evidence of a high cytosolic adenosine diphosphate concentration and a relatively high rate of metabolism accompanied by a low phosphorylation potential. Muscle³¹P-MRS showed a high Pi content at rest, abnormal exercise transfer pattern and a low rate of PCr post-exercise recovery. These findings suggested a deficit of mitochondrial function. Therapy with vitamins K₃ and C normalized brain³¹P-MRS indices, whereas it did not affect muscle bioenergetic metabolism. In this patient, the endocrinological disorder is putatively due to a mitochondrial cytopathy. Although an unknown mtDNA mutation cannot be ruled out, the genetic defect may lie in the nuclear genome.     

Mitochondrial DNA sequence data reveals association of haplogroup U with psychosis in bipolar disorder

Converging genetic, postmortem gene-expression, cellular, and neuroimaging data implicate mitochondrial dysfunction in bipolar disorder. This study was conducted to investigate whether mitochondrial DNA (mtDNA) haplogroups and single nucleotide variants (SNVs) are associated with sub-phenotypes of bipolar disorder. MtDNA from 224 patients with Bipolar I disorder (BPI) was sequenced, and association of sequence variations with 3 sub-phenotypes (psychosis, rapid cycling, and adolescent illness onset) was evaluated. Gene-level tests were performed to evaluate overall burden of minor alleles for each phenotype. The haplogroup U was associated with a higher risk of psychosis. Secondary analyses of SNVs provided nominal evidence for association of psychosis with variants in the tRNA, ND4 and ND5 genes. The association of psychosis with ND4 (gene that encodes NADH dehydrogenase 4) was further supported by gene-level analysis. Preliminary analysis of mtDNA sequence data suggests a higher risk of psychosis with the U haplogroup and variation in the ND4 gene implicated in electron transport chain energy regulation. Further investigation of the functional consequences of this mtDNA variation is encouraged.     

A pathophysiological study of macular mutant mouse as a model of human Menkes kinky hair disease. II. Analysis of brain metabolism using 31P- and 1H-nuclear magnetic resonance spectroscopy

We studied the brain metabolism in macular mutant mice (Ml/y, +/y), an appropriate model of Menkes kinky hair disease, using 31P- and 1H-NMR spectroscopy to clarify the pathophysiological mechanisms of disturbed nervous function. An analysis of in vivo 31P-NMR spectra showed a decreased phosphocreatine (PCr)/inorganic phosphate (Pi) ratio and decreased ATP levels and decreased intracellular pH in Ml/y mice at 9 days, suggesting energy failure in the brain. This associated decline in ATP levels may reflect multiple causative factors including disturbed mitochondrial respiration and ischemia secondary to circulatory failure. Brain metabolites, including PCr, creatine, lactate and 7 amino acids were easily detectable quantitatively and qualitatively by in vitro 1H-NMR spectrum. An elevation in lactate levels and a decline in PCr/creatine ratio in Ml/y mice at 9 days were also noted with an in vitro study, supporting the in vivo data. NMR spectroscopy is a useful and promising tool to obtain the information on brain metabolism.     

Mitochondrial dysfunction in bipolar disorder: evidence from magnetic resonance spectroscopy research

Magnetic resonance spectroscopy (MRS) affords a noninvasive window on in vivo brain chemistry and, as such, provides a unique opportunity to gain insight into the biochemical pathology of bipolar disorder. Studies utilizing proton ((1)H) MRS have identified changes in cerebral concentrations of N-acetyl aspartate, glutamate/glutamine, choline-containing compounds, myo-inositol, and lactate in bipolar subjects compared to normal controls, while studies using phosphorus ((31)P) MRS have examined additional alterations in levels of phosphocreatine, phosphomonoesters, and intracellular pH. We hypothesize that the majority of MRS findings in bipolar subjects can be fit into a more cohesive bioenergetic and neurochemical model of bipolar illness that is both novel and yet in concordance with findings from complementary methodological approaches. In this review, we propose a hypothesis of mitochondrial dysfunction in bipolar disorder that involves impaired oxidative phosphorylation, a resultant shift toward glycolytic energy production, a decrease in total energy production and/or substrate availability, and altered phospholipid metabolism.     

Mitochondrial DNA deletions/rearrangements in parkinson disease and related neurodegenerative disorders

Inhibition of mitochondrial respiratory chain function may contribute to dopaminergic neurodegeneration in the substantia nigra (SN) of patients with Parkinson disease (PD). Since large-scale structural changes (e.g. deletions and rearrangements in mitochondrial DNA [mtDNA]) have been associated with mitochondrial dysfunction, we tested the hypothesis that increased total mtDNA deletions/rearrangements are associated with neurodegeneration in PD. This study employed a well-established technique, long-extension polymerase chain reaction (LX-PCR), to detect the multiple mtDNA deletions/rearrangements in the SN of patients with PD, multiple system atrophy (MSA), dementia with Lewy bodies (DLB), Alzheimer disease (AD), and age-matched controls. We also compared the total mtDNA deletions/rearrangements in different brain regions of PD patients. The results demonstrated that both the number and variety of mtDNA deletions/rearrangements were selectively increased in the SN of PD patients compared to patients with other movement disorders as well as patients with AD and age-matched controls. In addition, increased mtDNA deletions/rearrangements were observed in other brain regions in PD patients, indicating that mitochondrial dysfunction is not just limited to the SN of PD patients. These data suggest that accumulation of total mtDNA deletions/rearrangements is a relatively specific characteristic of PD and may be one of the contributing factors leading to mitochondrial dysfunction and neurodegeneration in PD.     

PDE4B polymorphisms and decreased PDE4B expression are associated with schizophrenia

Schizophrenia has a complex genetic underpinning and variations in a number of candidate genes have been identified that confer risk of developing the disorder. We report in the present studies that several single nucleotide polymorphisms (SNPs) and a two-SNP haplotype in PDE4B are associated with an increased incidence of schizophrenia in two large populations of Caucasian and African American patients. The SNPs in PDE4B associated with schizophrenia occur in intronic sequences in the vicinity of a critical splice junction that gives rise to the expression of PDE4B isoforms with distinct regulation and function. We also observed specific decreases in phosphodiesterase 4B (PDE4B) isoforms in brain tissue obtained postmortem from patients diagnosed with schizophrenia and bipolar disorder. PDE4B metabolically inactivates the second messenger cAMP to regulate intracellular signaling in neurons throughout the brain. Thus, the present observations suggest that dysregulation of intracellular signaling mediated by PDE4B is a significant factor in the cause and expression, respectively, of schizophrenia and bipolar disorder and that targeting PDE4B-regulated signaling pathways may yield new therapies to treat the totality of these disorders.     

Oral choline decreases brain purine levels in lithium-treated subjects with rapid-cycling bipolar disorder: a double-blind trial using proton and lithium magnetic resonance spectroscopy

Objectives: Oral choline administration has been reported to increase brain phosphatidylcholine levels. As phospholipid synthesis for maintaining membrane integrity in mammalian brain cells consumes approximately 10–15% of the total adenosine triphosphate (ATP) pool, an increased availability of brain choline may lead to an increase in ATP consumption. Given reports of genetic studies, which suggest mitochondrial dysfunction, and phosphorus (³¹P) magnetic resonance spectroscopy (MRS) studies, which report dysfunction in high‐energy phosphate metabolism in patients with bipolar disorder, the current study is designed to evaluate the role of oral choline supplementation in modifying high‐energy phosphate metabolism in subjects with bipolar disorder. Methods: Eight lithium‐treated patients with DSM‐IV bipolar disorder, rapid cycling type were randomly assigned to 50 mg/kg/day of choline bitartrate or placebo for 12 weeks. Brain purine, choline and lithium levels were assessed using ¹H‐ and ⁷Li‐MRS. Patients received four to six MRS scans, at baseline and weeks 2, 3, 5, 8, 10 and 12 of treatment (n = 40 scans). Patients were assessed using the Clinical Global Impression Scale (CGIS), the Young Mania Rating Scale (YRMS) and the Hamilton Depression Rating Scale (HDRS) at each MRS scan. Results: There were no significant differences in change‐from‐baseline measures of CGIS, YMRS, and HDRS, brain choline/creatine ratios, and brain lithium levels over a 12‐week assessment period between the choline and placebo groups or within each group. However, the choline treatment group showed a significant decrease in purine metabolite ratios from baseline (purine/n‐acetyl aspartate: coef = ?0.08, z = ?2.17, df = 22, p = 0.030; purine/choline: coef = ?0.12, z = ?1.97, df = 22, p = 0.049) compared to the placebo group, controlling for brain lithium level changes. Brain lithium level change was not a significant predictor of purine ratios. Conclusions: The current study reports that oral choline supplementation resulted in a significant decrease in brain purine levels over a 12‐week treatment period in lithium‐treated patients with DSM‐IV bipolar disorder, rapid‐cycling type, which may be related to the anti‐manic effects of adjuvant choline. This result is consistent with mitochondrial dysfunction in bipolar disorder inadequately meeting the demand for increased ATP production as exogenous oral choline administration increases membrane phospholipid synthesis.