This detailed clinical reference dataset will greatly facilitate diagnostic confirmation and management of patients, and has provided important insights into disease pathogenesis. Ann Neurol 2018;83:1105-1124.

The purpose of this study was to describe unusual variants of Alexander's disease. We studied 10 patients who did not meet previously established magnetic resonance imaging (MRI) criteria for Alexander's disease, but for whom this diagnosis was considered because of Rosenthal fibers at histological examination or presence of some MRI features suggestive of Alexander's disease. Sequence analysis of the GFAP gene was performed. In eight patients, MRI results showed predominantly posterior fossa lesions, especially multiple tumor-like brainstem lesions. One patient had asymmetrical frontal white matter abnormalities and basal ganglia abnormalities. One patient (Patient 10) developed degeneration of the frontal white matter. In nine patients, a mutation was found that was concluded to be pathogenic, because the mutation was de novo (five patients), a known mutation was found (two patients), or assembly of the glial fibrillary acidic protein was abnormal in cultured cells (two patients). In Patient 10, a DNA variation was found that was also present in the patient's clinically unaffected father and was concluded to be a polymorphism. In conclusion, DNA diagnostics is warranted in patients who display MRI features suggestive of Alexander's disease, even if they do not meet the full set of previously established MRI criteria.

Induction of altered phenotypes during development in response to environmental input involves epigenetic changes. Phenotypic traits can be passed between generations by a variety of mechanisms, including direct transmission of epigenetic states or by induction of epigenetic marks de novo in each generation. To distinguish between these possibilities we measured epigenetic marks over four generations in rats exposed to a sustained environmental challenge. Dietary energy was increased by 25% at conception in F0 female rats and maintained at this level to generation F3. F0 dams showed higher pregnancy weight gain, but lower weight gain and food intake during lactation than F1 and F2 dams. On gestational day 8, fasting plasma glucose concentration was higher and β-hydroxybutyrate lower in F0 and F1 dams than F2 dams. This was accompanied by decreased phosphoenolpyruvate carboxykinase (PEPCK) and increased PPARα and carnitine palmitoyl transferase-1 mRNA expression. PEPCK mRNA expression was inversely related to the methylation of specific CpG dinucleotides in its promoter. DNA methyltransferase (Dnmt) 3a2, but not Dnmt1 or Dnmt3b, expression increased and methylation of its promoter decreased from F1 to F3 generations. These data suggest that the regulation of energy metabolism during pregnancy and lactation within a generation is influenced by the maternal phenotype in the preceding generation and the environment during the current pregnancy. The transgenerational effects on phenotype were associated with altered DNA methylation of specific genes in a manner consistent with induction de novo of epigenetic marks in each generation.

Childhood-onset autosomal recessive spinal muscular atrophy (SMA) is associated with absence of the telomeric survival motor neuron gene (SMNt) in most patients, and deletion of the neuronal apoptosis inhibitory protein (NAIP) gene in the majority of severely affected patients. Analysis of SMNt has been complicated by the existence of a centromeric copy, SMNc, which is almost identical to SMNt but which can be distinguished from it by restriction enzyme analysis. In this study 143 SMA patients have been genotyped for the presence or absence of the SMNt, SMNc and NAIP genes, and the data correlated with quantifiable clinical variables. Although a significant correlation was observed between the presence or absence of the NAIP gene and the severity of the clinical phenotype in SMA patients generally, there was no difference in age of onset or survival in type I patients with the NAIP + or NAIP-genotype. Fluorimetric PCR analysis of SMNc gene dosage in 57 patients homozygous for the absence of the SMNt gene but in whom the NAIP gene was present showed a highly significant correlation between SMNc copy number and SMA subtype, and between SMNc copy number and both age of onset and length of survival. The data provide strong statistical support for the emerging consensus that the clinical phenotype in SMA is directed primarily by the level of functional SMN protein. The lower SMNc copy number in type I patients in whom the NAIP gene is present suggests that the SMNt gene is removed by deletion in the majority of such patients, rather than by gene conversion as is the case in SMA types II and III.