Characterization of patients with major psychiatric disorders with AMPA receptor positron emission tomography

Abstract

Synaptic phenotypes in living patients with psychiatric disorders are poorly characterized. Excitatory glutamate α-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid receptor (AMPAR) is a fundamental component for neurotransmission. We recently developed a positron emission tomography (PET) tracer for AMPAR, [¹¹C]K-2, the first technology to visualize and quantify AMPARs density in living human brain. In this study, we characterized patients with major psychiatric disorders with [¹¹C]K-2. One hundred forty-nine patients with psychiatric disorders (schizophrenia, n = 42; bipolar disorder, n = 37; depression, n = 35; and autism spectrum disorder, n = 35) and 70 healthy participants underwent a PET scan with [¹¹C]K-2 for measurement of AMPAR density. We detected brain regions that showed correlation between AMPAR density and symptomatology scores in each of four disorders. We also found brain areas with significant differences in AMPAR density between patients with each psychiatric disorder and healthy participants. Some of these areas were observed across diseases, indicating that these are commonly affected areas throughout psychiatric disorders. Schizophrenia, bipolar disorder, depression, and autism spectrum disorder are uniquely characterized by AMPAR distribution patterns. Our approach to psychiatric disorders using [¹¹C]K-2 can elucidate the biological mechanisms across diseases and pave the way to develop novel diagnostics and therapeutics based on the synapse physiology.

Fig. 1. State regions and altered AMPAR distribution in patients with schizophrenia compared with healthy participants.

A Brain regions showing a significant negative correlation between SUVR_{30–50 min}WB and PANSS score for positive symptoms in patients with schizophrenia (P < 0.05, T < −1.68, one-tailed, FDRc). Significant clusters displayed on an axial, coronal and sagittal slices (Left), and scatter plot between averaged SUVR_{30–50 min}WB in significant clusters and PANSS scores for positive symptoms (Right, two-tailed Pearson correlation analysis: correlation coefficient = −0.7514, P < 0.0001). B Brain regions showing a significant negative correlation between SUVR_{30–50 min}WB and PANSS score for negative symptoms in patients with schizophrenia (P < 0.05, T < −1.68, one-tailed, FDRc). Significant clusters displayed on an axial, coronal and sagittal slices (Left), and scatter plot between averaged SUVR_{30–50 min}WB in significant clusters and PANSS score for negative symptoms (Right, two-tailed Pearson correlation analysis: correlation coefficient = −0.6371, P < 0.0001). C Relative reduction (blue) and increase (red) of [¹¹C]K-2 retention in patients with schizophrenia compared to healthy participants (P < 0.05, increase of [¹¹C]K-2 retention: T > 1.66, reduction of [¹¹C]K-2 retention: T < −1.66, one-tailed, FDRc). Significant clusters displayed on a 3D-rendered brain (binary image, Left), and axial, coronal and sagittal slices (T-map, Right). A, B and C (right panel) were adjusted for covariates (age, sex). C (left panel) was not adjusted for covariates.

Fig. 2. State regions and altered AMPAR distribution in patients with bipolar disorder compared with healthy participants.

A Brain regions showing a significant negative (blue) and positive (red) correlation between SUVR_30-50minWB and the 17-item HAM-D score in patients with bipolar disorder (P < 0.05, positive correlation: T > 1.69, negative correlation: T < −1.69, one-tailed, FDRc). Significant clusters displayed on an axial, coronal and sagittal slices (Left), and scatter plot between averaged SUVR_{30–50 min}WB in significant clusters and HAM-D score (Right, two-tailed Pearson correlation analysis: positive correlation; correlation coefficient = 0.6841, P < 0.0001, negative correlation; correlation coefficient = −0.5208, P = 0.001). B Brain regions showing a significant negative (blue) and positive (red) correlation between SUVR_{30–50 min}WB and YMRS score in patients with bipolar disorder (P < 0.05, positive correlation: T > 1.69, negative correlation: T < −1.69, one-tailed, FDRc). Significant clusters displayed on an axial, coronal, and sagittal slices (Left), and scatter plot between averaged SUVR_{30–50 min}WB in significant clusters and YMRS score (Right, two-tailed Pearson correlation analysis: positive correlation; correlation coefficient = 0.5842, P = 0.0001, negative correlation; correlation coefficient = −0.7012, P < 0.0001). C Relative reduction (blue) and increase (red) of [¹¹C]K-2 retention in patients with bipolar disorder compared to healthy participants (P < 0.05, increase of [¹¹C]K-2 retention: T > 1.66, reduction of [¹¹C]K-2 retention: T < −1.66, one-tailed, FDRc). Significant clusters displayed on a 3D-rendered brain (binary image, Left), and axial, coronal, and sagittal slices (T-map, Right). A, B and C (right panel) were adjusted for covariates (age, sex). C (left panel) was not adjusted for covariates.

Fig. 3. State regions in depression.

A Brain regions showing a significant negative correlation between SUVR_{30–50 min}WB and the 17-item HAM-D score in patients with depression (P < 0.05, T < −1.69, one-tailed, FDRc). Significant clusters displayed on an axial, coronal and sagittal slices (Left), and scatter plot between averaged SUVR_{30–50 min}WB in significant clusters and HAM-D score (Right, two-tailed Pearson correlation analysis: correlation coefficient = −0.5479, P = 0.0007). B Brain regions showing a significant negative correlation between SUVR_{30–50 min}WB and the 17-item HAM-D score in patients with depression as determined above (blue) and bipolar disorder (green) (P < 0.05, T < −1.69, one-tailed, FDRc). Red regions show where the two regions overlap. MNI coordinates for the z-axis (axial slices), y-axis (coronal slices) and x-axis (sagittal slices) were shown bottom of each slice. C No significant difference of SUVR_{30–50 min}WB between the patients with depression and healthy participants. All panels were adjusted for covariates (age, sex).

Fig. 4. State regions and altered AMPAR distribution in patients with ASD compared with healthy participants.

A Brain regions showing a significant positive correlation between SUVR_{30–50 min}WB and ADOS-2 Module 4 CSS in patients with ASD (P < 0.05, T > 1.69, one-tailed, FDRc). Significant clusters displayed on an axial, coronal, and sagittal slices (Left), and scatter plot between averaged SUVR_{30–50 min}WB in significant clusters and ADOS-2 Module 4 CSS (Right, two-tailed Pearson correlation analysis: correlation coefficient = 0.6027, P = 0.0001). B Relative reduction (blue) and increase (red) of [¹¹C]K-2 retention in patients with ASD compared to healthy participants (P < 0.05, increase of [¹¹C]K-2 retention: T > 1.66, reduction of [¹¹C]K-2 retention: T < −1.66, one-tailed, FDRc). Significant clusters displayed on a 3D-rendered brain (binary image, Left), and axial, coronal and sagittal slices (T-map, Right). A and B (right panel) were adjusted for covariates (age, sex). B (left panel) was not adjusted for covariates.

Fig. 5. Commonly affected areas across psychiatric disorders.

A Common brain regions showing reduced (blue) and increased (red) SUVR_{30–50 min}WB, respectively, in patients with schizophrenia, bipolar disorder and ASD compared with healthy participants. B　Common brain regions showing reduced (blue) SUVR_{30–50 min}WM, in patients with schizophrenia, bipolar disorder and ASD compared with healthy participants. MNI coordinates for the z-axis (axial slices), y-axis (coronal slices) and x-axis (sagittal slices) were shown bottom of each slice.