Distinct cerebral lesions in sporadic and 'D90A' SOD1 ALS: studies with [11C]flumazenil PET

Abstract

Five to ten percent of amyotrophic lateral sclerosis (ALS) cases are associated with mutations of the superoxide dismutase-1 (SOD1) gene, and the 'D90A' mutation is associated with a unique phenotype and markedly slower disease progression (mean survival time 14 years). Relative sparing of inhibitory cortical neuronal circuits might be one mechanism contributing to the slower progression in patients homozygous for the D90A mutation (homD90A). The GABA(A) receptor PET ligand [11C]flumazenil has demonstrated motor and extra-motor cortical changes in sporadic ALS. In this study, we used [11C]flumazenil PET to explore differences in the pattern of cortical involvement between sporadic and genetically homogeneous ALS groups. Twenty-four sporadic ALS (sALS) and 10 homD90A patients underwent [11C]flumazenil PET of the brain. In addition, two subjects homozygous for the D90A mutation, but without symptoms or signs ('pre-symptomatic', psD90A), also underwent imaging. Results for each group were compared with those for 24 healthy controls of similar age. Decreases in the binding of [11C]flumazenil in the sALS group were found within premotor regions, motor cortex and posterior motor association areas. In the homD90A group of ALS patients, however, decreases were concentrated in the left fronto-temporal junction and anterior cingulate gyrus. In the two psD90A subjects, a small focus of reduced [11C]flumazenil binding at the left fronto-temporal junction was seen, similar to the pattern seen in the clinically affected patients. Within the sALS group, there was no statistically significant association between decreases in cortical [11C]flumazenil binding and revised ALS functional rating scale (ALSFRS-R score), whereas the upper motor neuron (UMN) score correlated with widespread and marked cortical decreases over the dominant hemisphere. In the homD90A group, there was a stronger statistical association between reduced cortical [11C]flumazenil binding and the ALSFRS-R, rather than the UMN, score, and also with disease duration. This study provides evidence for differences in the distribution of reduced cortical [11C]flumazenil binding in homD90A compared with sALS patients. We hypothesize that this might reflect differences in cortical neuronal vulnerability.