A longitudinal investigation of GABA, glutamate, and glutamine across the insula during antipsychotic treatment of first-episode schizophrenia

Abstract

Individuals with first-episode schizophrenia (FES) typically present with acute psychotic symptoms. Though antipsychotic drugs are the mainstay for treatment, the neurobiology underlying successful treatment remains largely elusive. Recent evidence from functional connectivity studies highlights the insula as a key structure in the neural mechanism of response. However, molecular contributions to response across insular regions remain largely unknown. We used 7-Tesla magnetic resonance spectroscopic imaging (MRSI) to measure glutamate (Glu), Glutamine (Gln), and GABA from anterior and posterior regions of the insula across antipsychotic treatment. A total of 36 participants were examined, including 15 individuals with FES and moderate to severe psychosis who were scanned at two time points, while starting and after 6 weeks of antipsychotic treatment. Symptoms were carefully monitored across the study period to characterize treatment response. GABA, Glu, and Gln levels were calculated relative to creatine in anterior and posterior insular regions, bilaterally. In relation to psychotic symptom reduction, we observed a significant increase in Glu across all insular regions with (p < 0.001), but no corresponding changes in Gln or GABA. In group analyses, the FES cohort showed lower levels of Glu (p < 0.001) and GABA (p = 0.02) at baseline. Finally, in exploratory analyses, treatment remitters demonstrated a normalization of lower insular Glu levels across treatment, unlike non-remitters. Overall, these findings contribute to our understating of molecular changes associated with antipsychotic response and demonstrate abnormalities specific to the insula in FES.

Keywords: Antipsychotic treatment; First-episode schizophrenia; GABA; Glutamate; Glutamine; Insula; Magnetic resonance spectroscopic imaging; Psychosis; Treatment response.

Fig. 1.

MRSI Methods. In A), representative images in native space of voxels placed with anatomic precision within our MRSI slab in sagittal and coronal views are shown, and a selection of insular voxels in native space across our MRSI slice is displayed in the axial plane. Selected voxels were then transformed to standard MNI space. Attention was given to the central sulcus of the insula as an anatomical demarcation between anterior and posterior subregions. In B), the central coordinates for all voxels are displayed across insular regions in standard space anterior or posterior to the insular central sulcus. Panel C) shows a representative measured spectrum from one voxel and the decomposed metabolite or neurotransmitter spectra with corresponding CRLB values. The linewidth of the metabolite peaks displayed is 13 Hz.

Fig. 2.

Antipsychotic Efficacy vs. Change in Metabolites. We first examined percent change in metabolites (Glu, GABA, and Gln) relative to percent reductions in psychotic symptoms across antipsychotic treatment in our FES cohort. A significant increase in Glu/Cre was noted with treatment efficacy across all insular regions (p < 0.001). No corresponding changes were significant for either GABA or Gln.

Fig. 3.

Group Comparisons. In A) metabolites (Glu/Cre, GABA/Cre, and Gln/Cre) in FES vs HC at baseline are displayed. A significantly lower level of Glu/Cre (p < 0.001) and GABA/Cre (p = 0.02) were observed in the FES cohort. In B) we categorized FES participants as remitters or non-remitters to treatment for further exploratory analyses. Glu/Cre, which showed an increase in relation to reduction in psychotic symptoms across treatment, showed a normalizing effect. At baseline remitters show a lower level of Glu/Cre (p < 0.001), but not at follow-up (p = 0.31). Meanwhile, non-remitters show similar levels relative to the HC range across treatment (p = 0.02 at baseline, and p = 0.01 at follow-up). HC = healthy controls; FES = first-episode schizophrenia; NR = non-remitters; R = remitters.