Clinical and molecular phenotype of Aicardi-Goutieres syndrome

Abstract

Aicardi-Goutieres syndrome (AGS) is a genetic encephalopathy whose clinical features mimic those of acquired in utero viral infection. AGS exhibits locus heterogeneity, with mutations identified in genes encoding the 3'-->5' exonuclease TREX1 and the three subunits of the RNASEH2 endonuclease complex. To define the molecular spectrum of AGS, we performed mutation screening in patients, from 127 pedigrees, with a clinical diagnosis of the disease. Biallelic mutations in TREX1, RNASEH2A, RNASEH2B, and RNASEH2C were observed in 31, 3, 47, and 18 families, respectively. In five families, we identified an RNASEH2A or RNASEH2B mutation on one allele only. In one child, the disease occurred because of a de novo heterozygous TREX1 mutation. In 22 families, no mutations were found. Null mutations were common in TREX1, although a specific missense mutation was observed frequently in patients from northern Europe. Almost all mutations in RNASEH2A, RNASEH2B, and RNASEH2C were missense. We identified an RNASEH2C founder mutation in 13 Pakistani families. We also collected clinical data from 123 mutation-positive patients. Two clinical presentations could be delineated: an early-onset neonatal form, highly reminiscent of congenital infection seen particularly with TREX1 mutations, and a later-onset presentation, sometimes occurring after several months of normal development and occasionally associated with remarkably preserved neurological function, most frequently due to RNASEH2B mutations. Mortality was correlated with genotype; 34.3% of patients with TREX1, RNASEH2A, and RNASEH2C mutations versus 8.0% RNASEH2B mutation-positive patients were known to have died (P=.001). Our analysis defines the phenotypic spectrum of AGS and suggests a coherent mutation-screening strategy in this heterogeneous disorder. Additionally, our data indicate that at least one further AGS-causing gene remains to be identified.

Figure 1.

Numbers and percentages of AGS-affected families with biallelic mutations in TREX1, RNASEH2A, RNASEH2B, and RNASEH2C; single RNASEH2A, RNASEH2B, and TREX1 mutations; and those with no identifiable mutation(s).

Figure 2.

Schematic representation of the TREX1 protein, with the corresponding position of the mutations identified in TREX1. Regions 1, 3, and 4 represent the exonuclease regions (Exo1-3), which coordinate the binding of 2 Mg²⁺ ions required for catalysis. Region 2 represents the polyproline II motif. Region 5 represents the dileucine-repeat region. Numbers in parentheses after mutations represent the number of mutated alleles identified. An asterisk (*) denotes the de novo mutation identified in one family.

Figure 3.

Schematic representation of the RNASEH2B gene, with the position of identified mutations. Shaded areas with large numbers indicate the specified exons. Numbers in parentheses after mutations represent the number of mutated alleles identified. Splice-site variants and stop mutations always occur with a missense mutation.

Figure 4.

Schematic representation of the RNASEH2C gene, with the position of identified mutations. Shaded areas with large numbers indicate the specified exons. Numbers in parentheses after mutations represent the number of mutated alleles identified.

Figure 5.

Schematic representation of the RNASEH2A gene, with the position of identified mutations and polymorphisms. Shaded areas with large numbers indicate the specified exons. Numbers in parentheses after mutations represent the number of mutated alleles identified. An asterisk (*) denotes polymorphisms included in the SNP database; a double asterisk (**) denotes synonymous changes not found in controls or annotated as a SNP; ¶ denotes polymorphisms found in controls; † denotes polymorphisms found in combination with RNASEH2B mutations; ‡ denotes putative mutation found in patients with only a single identified RNASEH2A change; a number sign (#) denotes mutations found on one allele in a single patient.

Figure 6.

Age (mo) at presentation by gene for patients with TREX1, RNASEH2A, RNASEH2B, and RNASEH2C mutations. With the Mann-Whitney U test comparing age at presentation for patients with RNASEH2B mutations with that for patients with mutations in TREX1, RNASEH2C, and RNASEH2A, P<.0005. With the Kruskal-Wallis test comparing age at presentation with gene, P<.0005.

Figure 7.

Age at death (in years) of patients with TREX1, RNASEH2A, RNASEH2B, and RNASEH2C mutations; χ² test comparing number of deaths among children with TREX1, RNASEH2C, and RNASEH2A mutations against number of deaths among children with RNASEH2B mutations (P=.001).

Figure 8.

Examples of chilblain lesions seen in patients with AGS

Figure 9.

Examples of intracranial calcification on CT scan of patients with AGS. Calcification is seen in the basal ganglia (a and b), dentate nuclei of the cerebellum (c), a periventricular distribution (d), and within the deep white matter (e).

Figure 10.

Spectrum of brain changes seen on MRI of patients with AGS, showing hypointensity on T1-weighted imaging (a), hyperintensity on T2-weighted imaging (b and c) of white matter, extensive bitemporal cystic lesions (d), and significant thinning of the brain stem and cerebellar atrophy (e).

Figure 11.

CSF white-cell counts by age in patients with TREX1, RNASEH2A, RNASEH2B, and RNASEH2C mutations. With Kruskal-Wallis test comparing CSF WCC/mm³ with age group, P<.0005.

Figure 12.

CSF IFN-α titers by age in patients with TREX1, RNASEH2A, RNASEH2B, and RNASEH2C mutations. With Kruskal-Wallis test comparing CSF IFN-α IU/liter with age group, P<.0005.