Reduced NAA Levels in the Dorsolateral Prefrontal Cortex of Young ...

Article

Reduced NAA Levels in the Dorsolateral Prefrontal Cortex of Young Bipolar Patients Roberto B. Sassi, M.D., Ph.D. Jeffrey A. Stanley, Ph.D. David Axelson, M.D. Paolo Brambilla, M.D. Mark A. Nicoletti, M.Sc. Matcheri S. Keshavan, M.D. Renato T. Ramos, M.D., Ph.D. Neal Ryan, M.D. Boris Birmaher, M.D. Jair C. Soares, M.D.

Objective: Converging evidence implicates prefrontal circuits in the pathophysiology of bipolar disorder. Proton spectroscopy studies performed in adult bipolar patients assessing prefrontal regions have suggested decreased levels of N -acetylaspartate (NAA), a putative marker of neuronal integrity. In order to examine whether such abnormalities would also be found in younger patients, a 1H spectroscopy study was conducted that focused on the dorsolateral prefrontal cortex of children and adolescents with bipolar disorder. Method: The authors examined the levels of NAA, creatine plus phosphocreatine, and choline-containing molecules in the left dorsolateral prefrontal cortex of 14 bipolar disorder patients (mean age= 15.5 years, SD=3, eight female) and 18 healthy comparison subjects (mean age= 17.3, SD=3.7, seven female) using short echo time, single-voxel in vivo 1 H spec-

troscopy. Absolute metabolite levels were determined using the water signal as an internal reference. Results: Bipolar patients presented significantly lower NAA levels and a significant inverse correlation between cholinecontaining molecules and number of previous affective episodes. No differences were found for other metabolites. Conclusions: These findings suggest that young bipolar patients have decreased NAA levels in the dorsolateral prefrontal cortex, similar to what was previously reported in adult patients. Such changes may reflect an underdevelopment of dendritic arborizations and synaptic connections. These neuronal abnormalities in the dorsolateral prefrontal cortex of bipolar disorder youth are unlikely to represent long-term degenerative processes, at least in the subgroup of patients where the illness had relatively early onset. (Am J Psychiatry 2005; 162:2109–2115)

B

ipolar disorder is a serious and chronic psychiatric illness whose neuropathology is still largely unknown. The involvement of prefrontal brain regions in bipolar disorder has been replicated by various research groups and is supported by several lines of evidence. In vivo structural neuroimaging studies have reported decreased prefrontal cortical volumes among bipolar patients (1–3). Postmortem analysis of this region has revealed specific abnormalities such as decreased neuronal and glial density (4). Investigations employing functional magnetic resonance imaging, positron emission tomography, or single photon emission computed tomography all have consistently reported decreased frontal glucose metabolism (5, 6) and lower frontal blood flow in bipolar patients, mostly during depressive episodes (3, 7). Discrete prefrontal cortical regions have been investigated in bipolar disorder patients, with interesting although sometimes conflicting results. Drevets et al. (8) repor ted decreased volume and functioning of the subgenual prefrontal cortex among bipolar patients, familial subtype. A postmortem study (9) found subgenual abnormalities, mainly glial density reduction, in this same population. Our group, however, has failed to identify volumetric abnormalities in this region among mood disorAm J Psychiatry 162:11, November 2005

der patients (10). Hypoactivation of orbital and rostral prefrontal cortices has been reported in manic patients (11), although other investigators have not found any rCBF abnormalities in bipolar disorder (12). The dorsolateral region of the prefrontal cortex, encompassing Brodmann’s areas 9 and 46 (13), is believed to play a major role in decision making (14) and working memory (15, 16), acting as an interface between cognition and emotion (17–19). Utilizing proton magnetic resonance spectroscopy (1H-MRS), a methodology that provides information on the brain content of several metabolites in vivo, Winsberg et al. (20) observed significantly decreased N-acetylaspartate (NAA) in the dorsolateral prefrontal cortex of unmedicated adult bipolar patients. NAA is found primarily in neurons (21) and is a nonspecific marker of neuronal integrity (22). On the other hand, Moore et al. (23) reported a significant increase in NAA levels after 4 weeks of lithium treatment in bipolar patients, possibly as a consequence of lithium’s neurotrophic actions (24). However, most neurobiological studies of affective disorder were conducted with adult patient groups. For instance, the average age of subjects in the Winsberg et al. (20) and Moore et al. (23) studies was 37.9 and 36.3 years, respectively. Therefore, it is not clear if the reported http://ajp.psychiatryonline.org

2109

NAA REDUCTIONS IN BIPOLAR YOUTH TABLE 1. Demographic and Clinical Characteristics of Bipolar Disorder Youth Age Race Subject (years)a Gender 1 10 M White

a

2

12

M

White

3

13

F

White

4

14

M

White

5

14

F

White

6

14

M

White

7

15

F

White

8

15

F

White

9

17

F

White

10

17

M

White

11

18

F

White

12

18

F

13

19

F

14

21

M

Diagnosis and Current Mood Episode Bipolar disorder not otherwise specified, depressed Bipolar II disorder, euthymic

Age at Onset (years) 7

Previous Antipsychotic Use Comorbidity Medications (daily oral dose) Yes Conduct disorder Drug-free for at least 4 weeks

6

Yes

ADHD

Bipolar I disorder, euthymic Bipolar I disorder, euthymic Bipolar II disorder, euthymic Bipolar I disorder, euthymic

10

Yes

10

Yes

11

No

Lithium (1050 mg); methylphenidate (36 mg); levothyroxine (50 µg) Lithium (900 mg); valproate (750 mg) Oppositional Lithium (1350 mg); valproate defiant disorder (750 mg) Valproate (1000 mg)

10

Yes

ADHD

Bipolar I disorder, euthymic Bipolar I disorder, euthymic

14

Yes

Valproate (500 mg); clomipramine (150 mg); dextroamphetamine (5 mg) Valproate (1000 mg)

14

Yes

ADHD

Bipolar II disorder, euthymic Bipolar I disorder, euthymic

13

No

Lithium (1350 mg); valproate (1000 mg); fluoxetine (10 mg); lorazepam (1 mg); levothyroxine (50 µg) Drug-free for at least 5 months

12

Yes

ADHD, substance abuse (cannabis) in remission

Lithium (1350 mg); valproate (2000 mg); trazodone (100 mg)

Bipolar I disorder, euthymic White Bipolar I disorder, euthymic White Bipolar I disorder, euthymic African Bipolar I disorder, American euthymic

15

Yes

8

Yes

17

No

Valproate (1875 mg); levothyroxine (50 µg) Lithium (1200 mg); citalopram (60 mg) Lithium (900 mg)

15

No

Lithium (1050 mg)

ADHD

Mean=15.5 years (SD=3).

abnormalities were present since the early stages of the illness or whether they might be related to illness progression or medication use. The purpose of the present study was to investigate metabolite abnormalities in the dorsolateral prefrontal cortex of bipolar disorder youth. A younger patient population was chosen in order to minimize the effects of confounding variables such as long-term medication usage. We hypothesized that lower NAA levels would be found among patients with bipolar disorder, representing abnormal prefrontal processes of possible neurodevelopmental origin that are already present at the early stages of the disease. Moreover, levels of choline-containing compounds in this same region would allow us to indirectly evaluate membrane phospholipid metabolism and signaling of the phosphatidylcholine system, providing information on neuronal membrane functioning.

21]; female: N=7, male: N=11; African American: N=3, Caucasian: N=15). After having understood all issues involved in participation in the study protocol, all subjects and their parents or legal representatives provided signed informed study consent. This research study was approved by the University of Pittsburgh Biomedical Institutional Review Board. The patients were recruited at the outpatient facilities of the University of Pittsburgh Medical Center or through advertisements in the local media. The inclusion criteria were age between 10 and 21 years and a diagnosis of bipolar disorder, any subtype, in any mood state. All patients met DSM-IV criteria for bipolar disorder. For patients 10–17 years old, diagnosis was determined with the Schedule for Affective Disorders and Schizophrenia for School-Age Children—Present and Lifetime Version (K-SADS-PL) (25). Patients 18–21 years old were assessed with the Structured Clinical Interview for DSM-IV (SCID), patient edition (26). Information about family history of psychiatric disorders, age at onset of illness, length of illness, number of previous DSM-IV affective episodes, number of weeks receiving lithium treatment, current lithium dose, and medication history was retrieved from psychiatric interviews with the patients and medical charts.

Method

All subjects had normal physical examination results and no history of neurological problems. The patients did not have any comorbid psychiatric diagnoses with the exception of ADHD (five of 14), conduct disorder (one of 14) and oppositional defiant disorder (one of 14). At the time of the study, one patient was in a depressive episode and 13 were euthymic. Only two patients were drug-free, and 12 patients were receiving medication treatment at

Subjects Thirty-two subjects were studied, of whom 14 were DSM-IV bipolar disorder patients (Table 1) and 18 were healthy comparison subjects (mean age=17.3 years [SD=3.7, median=18, range=11–

2110

http://ajp.psychiatryonline.org

Am J Psychiatry 162:11, November 2005

SASSI, STANLEY, AXELSON, ET AL. study entry. Ten patients had a history of previous antipsychotic usage, whereas four patients had no such history. Only one patient did not have a positive family history of mood disorders in a first-degree relative. First-degree relatives were considered positive for mood disorders if they ever received a diagnosis of unipolar or bipolar disorder by a psychiatrist, as ascertained by patient and relative reports or available medical records. Furthermore, the patients did not have any current medical problems or alcohol/substance abuse in the 6 months preceding the study. One patient had a previous history of substance abuse (cannabis) that had been in remission for more than 6 months before the study, and no patient had a lifetime history of substance dependence (Table 1). We could not ascertain the number of previous affective episodes for one patient. Our patient group had the following clinical characteristics: length of illness: mean=3.79 years (SD=2.39, median=3, range=1–10); age at first affective episode: mean=11.71 years (SD=3.24, median=11.5, range=6–17); number of previous affective episodes: mean=4.85 (SD=2.34, median=3, range=2–9). Healthy subjects had no DSM-IV axis I disorders, as determined either with the SCID or K-SADS-PL depending on subject age. Comparison subjects also had no current medical problems, no lifetime history of substance dependence or substance abuse in the 6 months preceding the study, and no history of psychiatric disorders among first-degree relatives.

FIGURE 1. Sagittal View of the 8-cm 3 Voxel Placement in the Dorsolateral Prefrontal Cortex

MRS Method In vivo 1H MRS was conducted on a GE Signa Imaging System (General Electric Medical Systems, Milwaukee), at field strength of 1.5 T. A three-dimensional spoiled gradient-recall acquisition was performed in the coronal plane (TR=25 msec, TE=5 msec, flip angle=40°, field of view=24 cm, slice thickness=1.5 mm, number of excitations=1, matrix size=256×192) to obtain 124 images covering the entire brain. A double spin-echo sequence was also used to obtain T 2 and proton density images in the axial plane to screen for neuroradiological abnormalities. The single-voxel short TE MRS data were collected with a STEAM sequence (TE=20 msec, TM=13.6 msec, TR=1.5 seconds, bandwidth=2 kHz, 2,048 complex data points, 300 acquisitions, voxel dimension=2.0×2.0×2.0 cm3). This 8-cm 3 voxel was placed in the left dorsolateral prefrontal cortex (Figure 1), which was identified on the set of sagittal and coronal MR images using standard anatomical atlases (28, 29). A second STEAM spectrum was collected without water suppression (16 acquisitions). On the basis of a semiautomated histogram method (30, 31), the percent volume of gray matter, white matter, and CSF within the MRS voxels was estimated from the three-dimensional spoiled gradientrecall acquisition data by using the NIH Image software package, version 1.62 (National Institutes of Health, Bethesda, Md.). The intraclass correlation coefficients for the histogram measurements obtained by two independent raters (P.B., M.A.N.) in a group of 10 subjects were 0.94 for gray matter, 0.94 for white matter, and 0.91 for CSF. The MRS postprocessing and quantification steps were 100% automated. The unsuppressed water spectrum was used to correct for any eddy current effects. No apodization was applied, and any residual water signal was removed by using the operatorindependent, singular-value-decomposition-based method (32). We investigated three major metabolites: NAA, creatine plus phosphocreatine, and choline-containing molecules (mostly glycerophosphocholine plus phosphocholine [33]). Five Gaussian damped sinusoids (NAA at 2.01 ppm, creatine plus phosphocreatine at 3.02 and 3.93 ppm, glycerophosphocholine plus phosphocholine at 3.21 ppm, and myo-inositol at 3.54 ppm) were used to model the in vivo data in the time domain using the Marquardt algorithm. To ensure that the signals of overlapping and of lesser amplitudes (i.e., metabolites with multiplet structures and macAm J Psychiatry 162:11, November 2005

romolecules) had negligible influence on the fitting of the singlets, the first 37 msec of the free-induction decay signal were omitted in the fitting, which has been shown to reliably and accurately quantify NAA, creatine plus phosphocreatine, and glycerophosphocholine plus phosphocholine (34). The unsuppressed water signal along with the appropriate correction factors was applied to obtain absolute quantification values with units of mmol/kg wet weight.

Statistical Analyses All analyses were conducted by using the SPSS for Windows software, version 10.0.5 (SPSS Inc., Chicago), and two-tailed statistical significance level was set at p