Reduced Brain Iron and Striatal Hyperdopaminergia in Schizophrenia: A Quantitative Susceptibility Mapping MRI and PET Study

Abstract

Objective: Neuroimaging studies have independently associated schizophrenia with low iron and elevated dopamine synthesis. While preclinical research demonstrates that midbrain iron deficiency leads to striatal hyperdopaminergia, this relationship has not been studied in schizophrenia. The authors conducted a case-control study to examine differences in tissue magnetic susceptibility, a marker of brain iron, and correlated these with striatal dopamine synthesis capacity.

Methods: Magnetic susceptibility in the substantia nigra and ventral tegmental area (SN-VTA) was measured using quantitative susceptibility mapping (QSM) MRI in 159 participants (control subjects, N=80; early-course schizophrenia, N=79, including patients who were antipsychotic-naïve or antipsychotic-free). Because magnetic susceptibility is increased by neuromelanin and reduced by myelin, neuromelanin-sensitive MRI (NM-MRI) and diffusion tensor imaging (DTI) were employed to investigate the influence of neuromelanin and myelin on the QSM findings in 99 participants (control subjects, N=38; schizophrenia patients, N=61). Dopamine synthesis capacity (K_i ^cer) was then assessed with [¹⁸F]-DOPA positron emission tomography in 40 people from the schizophrenia group to test whether low SN-VTA magnetic susceptibility was related to high striatal K_i ^cer.

Results: SN-VTA magnetic susceptibility was lower in patients with schizophrenia than in control subjects (d=-0.66, 95% CI=-0.98, -0.34). This difference remained significant in analyses controlling for mean diffusivity (a DTI measure inversely correlating with myelin concentration) and NM-MRI contrast-to-noise ratios. SN-VTA magnetic susceptibility was significantly inversely correlated with striatal K_i ^cer, independent of mean diffusivity and NM-MRI contrast-to-noise ratios (r=-0.44). In both analyses, the strongest effects were observed in the ventral SN-VTA.

Conclusions: These findings suggest that lower levels of non-neuromelanin-bound iron in the SN-VTA contribute to striatal hyperdopaminergia in schizophrenia. Further investigation is warranted to understand the role of low iron in schizophrenia and its potential as a treatment target.

Keywords: Biological Markers; Brain Imaging Techniques; Neuroanatomy; Neurocircuitry; Neuroimaging; Schizophrenia Spectrum and Other Psychotic Disorders.

Figure 1. Substantia nigra and ventral tegmental area (SN-VTA) quantitative susceptibility mapping (QSM) case-control analysis ^a

^a Panel A is a bar chart showing mean SN-VTA susceptibility, calculated by QSM, for the groups (95% confidence intervals) with values also plotted for each participant. Healthy control values are displayed in brown and data for patients with schizophrenia are in green. Susceptibility was significantly lower for participants with schizophrenia relative to the control group (d=-0.66; 95% CI=-0.98, -0.34). This remained significant when controlling for potential clinical confounders (p<0.001). Panel B shows the SN-VTA t-score map. The colorbar refers to the t-score, where brown indicates lower susceptibility in schizophrenia relative to controls and green lower susceptibility in controls. Panel C displays the 3 clusters (327 voxels out of the 1790 SN-VTA voxels) where schizophrenia was associated with significantly lower susceptibility (threshold-free cluster enhancement, Benjamini-Hochberg corrected p<0.05). The peak t-score of each cluster was in the right ventrolateral SN-VTA (x=7, y=-22, z=−20; t=-3.98; n voxels=187; yellow), left ventral SN-VTA (x=-7, y=-11, z=−10; t=-3.28; n voxels=114; purple), and left lateral SN-VTA (x=-11, y=-23, z=−13; t=-2.82; n voxels=26; red). ***=p<0.001. ppb=parts per billion.

Figure 2. Association between the voxelwise pattern of striatal dopamine synthesis capacity (K_i^cer) and substantia nigra and ventral tegmental area (SN-VTA) susceptibility ^a

^a Panel A shows the correlation between the average partial least squares regression (PLSR) generated out-of-sample scores assessing the voxelwise relationship between striatal K_i^cer and SN-VTA susceptibility. In panels B and C, z-score maps show the strength of relationship between SN-VTA susceptibility and each striatal voxel K_i^cer. Positive z-scores (red) indicated a direct relationship with SN-VTA susceptibility, while negative z-scores (blue) demonstrated an inverse relationship. The peak striatal z-score was in the right dorsal striatum (x=28, y=-4, z=−8; z-score=-4.46). Panel D displays the SN-VTA z-score map showing the strength of the voxel relationship with striatal K_i^cer. The peak z-score was in the left ventral SN-VTA (x=-5, y=-12, z=−11; z-score=-4.56).