Abnormal glutamatergic neurotransmission and neuronal-glial interactions in acute mania

Abstract

Background: At excitatory synapses, glutamate released from neurons is taken up by glial cells and converted to glutamine, which is cycled back to neurons. Alterations in this system are believed to play a role in the pathophysiology of bipolar disorder, but they have not been characterized in vivo. We examined the glutamine/glutamate ratio and levels of other metabolites in acute mania and schizophrenia in this exploratory study.

Methods: Data were obtained from 2 x 2 x 2 cm voxels in the anterior cingulate cortex (ACC) and parieto-occipital cortex (POC) using two-dimensional J-resolved proton magnetic resonance spectroscopy at 4 Tesla and analyzed using LCModel. Fifteen bipolar disorder patients with acute mania and 17 schizophrenia patients with acute psychosis were recruited from an inpatient unit; 21 matched healthy control subjects were also studied. Glutamine/glutamate ratio and N-acetylaspartate, creatine, choline, and myo-inositol levels were evaluated in a repeated measures design. Medication effects and relationship to demographic and clinical variables were analyzed.

Results: Glutamine/glutamate ratio was significantly higher in ACC and POC in bipolar disorder, but not schizophrenia, compared with healthy control subjects. N-acetylaspartate was significantly lower in the ACC in schizophrenia. Patients on and off lithium, anticonvulsants, or benzodiazepines had similar glutamine/glutamate ratios.

Conclusions: The elevated glutamine/glutamate ratio is consistent with glutamatergic overactivity and/or defective neuronal-glial coupling in acute mania, although medication effects cannot be ruled out. Abnormalities in glutamatergic neurotransmission and glial cell function in bipolar disorder may represent targets for novel therapeutic interventions.

Figures

Figure 1

Parasagittal (left) and axial (right) views of the brain from T1-weighted images illustrate ACC (top) and POC (bottom) voxel placement in one control subject.

Figure 2

Contour plots of real 2D spectra from the ACC in a control (left) and BD subject (right). In each case, the X-axis is frequency (F2 in ppm) and the Y-axis is J (F1 in Hz). The spectral region from about −35 to +35Hz is shown. The main metabolite resonances recognizable in the plots are labeled. Although mI and Gln resonances are not well-resolved in these plots, the additional information available from 2D MRS allows improved fitting of these metabolites as discussed in the text. Note the variable nature of the macromolecule signal profile (highlighted in a box in the left panel) between the two spectra. Abbreviations as in the text; H2O – water, GSH – glutathione, MMs - macromolecules.

Figure 3

Sample 1D spectra extracted from 2D datasets at J=0.0Hz. These spectra are from the ACC in a control (left) and BD subject (right). Raw data are visible as a black line and the LCModel spectral fit is in red, with the residual shown in the top panel. Metabolite resonances identifiable in this spectral extraction are labeled; note that Gln resonances are not apparent at J=0.0Hz.

Figure 4

Scatter plot of Glu and Gln measures expressed as Arbitrary Units (AU) for all 10 subjects in the test-retest study. See text for details. Note that in general, within-subject variability across scans is less significant than between-subject variability, Glu levels are higher than Gln levels, and ACC levels of both metabolites are higher than those in the POC.

Figure 5

Gln/Glu ratios in the three study groups. The left panel shows a scatter plot of Gln/Glu in NC, BD, and SZ groups with ACC and POC data points presented on the same graph, and mean and standard errors depicted adjacent to the data for each group. Within the SZ group, data from patients with schizoaffective disorder are depicted with open squares. The right panel shows Gln/Glu in the ACC and POC voxels in NC, BD and SZ groups (black, grey and white bars, respectively). Mean±SE are shown.

Figure 6

Gln/Glu in BD patients grouped according to medication status. In each panel, data points on the left are from patients not taking the specified medication while those on the right are from patients taking the medication (Li – Lithium; AC – Anticonvulsants; BDZ – Benzodiazepines; left column – data from ACC; right column – data from POC).