Brain metabolism and autoantibody titres predict functional impairment in systemic lupus erythematosus

Abstract

Objective: We investigated whether systemic lupus erythematosus (SLE) disease duration or serology associate with abnormal regional glucose metabolism as measured with [(18)F]2-fluoro-2-deoxy-D-glucose positron emission tomography (FDG-PET) and deficits on neuropsychological testing.

Methods: Subjects with SLE with stable disease activity, without brain damage or clinical symptoms of neuropsychiatric (NP) SLE, stratified by disease duration (short-term (ST)-SLE=disease ≤2 years, long-term (LT)-SLE=disease ≥10 years), underwent clinical assessments, neuropsychological testing, resting FDG-PET scan imaging and measurement of serum titres of antibody to N-methyl-d-aspartate receptor (DNRAb). FDG-PET scans were compared with age-matched and gender-matched healthy controls.

Results: Subjects with LT-SLE demonstrated hypometabolism in the prefrontal and premotor cortices that correlated with accrued SLE-related damage, but not with DNRAb titre or performance on NP testing. Independent of disease duration, subjects with SLE demonstrated hypermetabolism in the hippocampus and orbitofrontal cortex that correlated with impaired memory performance and mood alterations (depression, anxiety, fatigue). Serum DNRAb also correlated independently with impaired memory performance and increased anxiety. Together, serum DNRAb titre and regional hypermetabolism were more powerful predictors of performance than either alone.

Interpretation: The presence of serum DNRAbs can account for some aspects of brain dysfunction in patients with SLE, and the addition of regional measurements of resting brain metabolism improves the assessment and precise attribution of central nervous system manifestations related to SLE.

Keywords: Autoantibodies; Autoimmune Diseases; Systemic Lupus Erythematosus.

Figure 1

Voxel-based comparisons of FDG-PET images for SLE and healthy control subjects. SLE subjects (n=17) were age-matched and gender-matched with HCs (n=17). The SLE subjects demonstrated significantly increased metabolic activity (top) in the hippocampus (A), orbitofrontal cortex (BA 11, B) and putamen/globidus pallidus (GP, C) compared to HCs. In each region (bottom; p<0.001, one-way ANOVAs), metabolic activity was higher (p<0.05) in SLE subjects (open and filled triangles) than HC subjects (open circles), but did not differ (p>0.27) between the ST-SLE (open triangles) and LT-SLE (filled triangles) groups. Statistical parametric maps (SPM) were superimposed on a single-subject MRI brain template and thresholded at T=3.4, p<0.001 (uncorrected). Metabolic increases are color coded from red to yellow. The left side (L) is labeled on the display.

Figure 2

Voxel-based comparisons of FDG-PET images for SLE subjects with different disease duration and HC. Significantly decreased metabolic activity (top) was present in the prefrontal cortex (BA 9 and 10, A and B) and the premotor cortex (BA 6, C) of the LT-SLE subjects. The LT-SLE subjects (filled triangles) demonstrated significant hypometabolism compared to the ST-SLE subjects (open triangles) in each of these areas (bottom; p<0.05, one-way ANOVAs). However, compared to HCs (open circles), the LT-SLE subjects demonstrated significant hypometabolism only in the prefrontal cortex (p<0.05, A, B). Metabolism in the prefrontal and premotor regions did not differ significantly between the ST-SLE subjects and HCs. Statistical parametric maps (SPM) were superimposed on a single-subject MRI brain template and thresholded at T=3.4, p<0.001 (uncorrected). Metabolic decreases are color coded from blue to green. The left side (L) is labeled on the display.

Figure 3

Correlation between resting metabolism in SLE subjects and cognitive or behavioral performance. Increased metabolism in the left hippocampus correlated with poor performance on the running memory continuous performance test (RMCPT), a measure of sustained attention (A, p<0.01). Increased metabolism in the right hippocampus correlated with increased depression and anxiety scores (B and C, p<0.05); and increased metabolism in the left orbitofrontal cortex correlated with poor performance on the code substitution test, a measure of visual search, sustained attention and memory (D, p<0.005).

Figure 4

Predictive model of memory performance and mood alteration with serum DNRAb and metabolism in SLE. A. Multiple regression analysis demonstrates that the combined influence of elevated serum DNRAb titers (yellow) and increasing metabolism in the hippocampus (red) better predicts memory impairment in cognitive performance (blue) in the SLE patients (R²=0.67, p<0.0008) than either measure alone (hippocampal metabolism: R²=0.37, p<0.01; DNRAb: R²=0.51, p<0.002). B. Similarly, the combined effects of serum DNRAb (yellow) and increasing orbitofronal metabolism (red) are a more powerful predictor of increased anxiety (blue) (R²=0.44, p<0.03) than either measure independently (DNRAb: R²=0.31, p<0.03; orbitofrontal metabolism: R²=0.37, p=0.01).