Introduction

Alexander disease (AxD) is a rare neurodegenerative disorder characterized by white matter degeneration and the formation of cytoplasmic inclusions known as Rosenthal fibers, which primarily accumulate in the astrocyte end-feet in the subpial and perivascular zones and consist of glial fibrillary acidic protein (GFAP), heat-shock protein 27 and αB-crystallin.1, 2, 3 Since 2001, studies have reported GFAP mutations in various clinical types of AxD.4 Although most of the mutations in the coding region of GFAP in patients with AxD are missense mutations, additions and/or deletions of one or a few amino acids and intronic mutations that result in frame shifts have also been identified.5 These mutations are expected to have dominant gain-of-function and dominant-negative effects.4, 6

AxD has been classified according to the age of onset into the following three subtypes: infantile AxD (<2 years of age), juvenile AxD (2–12 years of age) and adult AxD (>12 years of age). However, we recently proposed a novel classification of AxD into three distinct types based on neurological and magnetic resonance imaging findings: cerebral (type 1), bulbospinal (type 2) and intermediate (type 3).7, 8 Late-onset AxD, which includes types 2 and 3, exhibits a more variable onset and severity in comparison with early-onset AxD, which includes type 1, suggesting the existence of factors that modify the clinical phenotype, particularly of types 2 and 3.9, 10, 11

A single-nucleotide polymorphism (SNP) in the activator protein-1 binding site of the GFAP promoter, −250 bp upstream from the GFAP transcriptional start site (NCBI dbSNP: rs2070935), is a candidate factor for mediating the effects of GFAP mutations.12 The A allele at rs2070935, which is a minor allele with an allelic frequency of 0.43, would primarily modify the site sequence and create a novel activator protein-1 binding site. An investigation of healthy controls has revealed that, based on differences in the extent of binding between C and A alleles, the C allele has a greater association with transcriptional activity than the A allele.12 These results indicate higher levels of GFAP expression in the presence of the C allele than in that of the A allele.12

In this study, to evaluate whether rs2070935 affected the clinical course of late-onset AxD, we analyzed the SNP in the GFAP promoter and examined the relationship between the SNP, age of onset and ambulatory disability in patients with AxD that was classified as types 2 and 3.

Materials and methods

Gene analysis

Between 2005 and 2012, to analyze GFAP mutations, 44 patients suspected of having AxD were referred to Kyoto Prefectural University of Medicine from other hospitals all over Japan. Genomic DNA was extracted from the peripheral blood of all patients after obtaining their informed consent. To detect GFAP mutations, sequence analysis of genomic DNA was performed as described previously.13 Point mutations in GFAP were identified in 12 of these patients. Of these, six satisfied the criteria of our proposed novel classification for type 2 AxD, and four satisfied the criteria for type 3 AxD. These 10 patients were included in this study.

The genomic DNA containing rs2070935 was amplified using polymerase chain reaction (PCR) with the following primers: forward, 5′-GTCCCCAGTTCATAGCAGGA-3′; reverse, 5′-GTGATGCGTCTCCTCTCCAT-3′. Direct sequence analysis was performed using an ABI PRISM 3100 autosequencer (PE Applied Biosystems, Foster City, CA, USA) and Big Dye terminators according to the manufacturer’s instructions.

Subcloning

The genomic DNA containing the coding region of rs2070935 and codon 74 of the GFAP exon 1, with a PCR product size of 803 bp, was amplified using PCR with the following primers: forward, 5′-ACTCAGCCCTTTCCTTCCTT-3′; reverse, 5′-CAGATTGTCCCTCTCAACCTCC-3′. The PCR products subjected to the attachment of a poly-A tail (10 × A-attachment mix) were ligated into pGEM-T Easy Vector (Promega, Madison, WI, USA), and subsequently transformed into chemically competent Escherichia coli cells. After screening for successful insertion using Insert Check Ready (Toyobo, Osaka, Japan), direct sequence analysis of three clones for each PCR product was performed as described above.

Phenotype–genotype correlation

We evaluated the age of onset and the age of ambulatory disability among patients with the C/C, C/A and A/A genotypes. Furthermore, ambulatory disability was evaluated by constructing the Kaplan–Meier ambulatory curves of C/C versus C/A and A/A genotypic patients. The statistical analysis of the relationship between the SNP and age of onset was performed by Ekuseru–Toukei 2010. A P-value <0.05 was considered statistically significant.

Results

In 10 patients with type 2 or 3 AxD, we identified the following rs2070935 genotypes: four C/C homozygotes, five C/A heterozygotes and one A/A homozygote (Table 1). Of four patients with the C/C genotype, three had lost the ability to walk in their 30 s or 40 s; however, all six patients with the C/A or A/A genotypes retained the ability to walk throughout their 30 s (Figure 1). The mean age of onset in the patients with the C/C genotype was 9.5 years and that in patients with C/A and A/A genotypes was 40.2 years old (P<0.05). An R79H GFAP mutation was observed in three patients with the following rs2070935 genotypes: A/A homozygote (patient 7), C/C homozygote (patient 8) and C/A heterozygote (patient 10). Patient 8 with the C/C genotype exhibited earlier onset than patients 7 and 10 with the A/A and C/A genotypes, respectively. Patient 8 exhibited gait disturbance at the age of 5 years and lost the ability to walk at the age of 43 years (Table 2). In addition, in her mid-30 s, she developed bulbar dysfunction, which progressed gradually. She had difficulty in swallowing and underwent a percutaneous endoscopic gastrostomy at the age of 45 years. Patients 7 and 10 did not show gait disturbance until their late 30 s, and at the time of this study, they were able to walk and did not present with difficulties in swallowing. Although two patients (patient 2 and 3) in whom M74T (c.221−T>C) mutation was observed had the C/A genotype at rs2070935 and similar ages of onset, patient 2 showed spastic tetraparesis with muscle weakness and required the aid of a walker (Table 3).14 Gene analysis using the subcloning method revealed that the C allele of rs2070935 was in cis with the gene containing the mutant allele of codon 221 in patient 2 and in trans in patient 3.

Table 1 Clinical and genetic data in 10 patients of AxD
Full size table
Figure 1
figure 1

Ambulatory disability was evaluated by constructing the Kaplan–Meier ambulatory curves of C/C versus C/A and A/A genotypic patients.

Full size image
Table 2 Comparison of clinical data among three patients with R79H AxD
Full size table
Table 3 Comparison of clinical data between two patients with M74T
Full size table

Discussion

Investigations on clinical phenotype–genotype correlations and morphological and functional impairment of astrocytes due to GFAP mutations using cells or animal models have suggested that the clinical type of AxD is primarily determined by the GFAP mutation.4, 5, 13, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28 However, patients with late-onset AxD, which includes types 2 and 3, exhibit variable onset and severity, suggesting the existence of other factors that modify the clinical course.9, 10, 11

An investigation using a transgenic mouse model that overexpressed wild-type GFAP has revealed that the aggregates are similar to Rosenthal fibers and that a shorter lifespan is observed in patients with increased GFAP expression,29 suggesting that GFAP overexpression may cause AxD. Although 3% of the patients with AxD did not have detectable GFAP mutations,15 other causative factors such as GFAP multiplication8 have not been detected either.

The rs2070935 in the activator protein-1 binding site of the GFAP promoter, which may mediate the effects of GFAP mutations, is a candidate factor for the modification of the clinical course of AxD,12 and differences in the degree of binding between C and A alleles have demonstrated a greater association of the C allele with transcriptional activity compared with the A allele.12

Our results suggested that the presence of the C/C genotype at rs2070935 may contribute to the early onset and the severity of ambulatory function in late-onset AxD, supporting the findings of a previous study that demonstrated a greater association of the C allele with transcriptional activity compared with that of the A allele.12 Our investigation of two M74T mutations at C/A genotype at rs2070935 suggested that the patient in whom the C allele of rs2070935 was in cis with the GFAP mutation may show a more severe phenotype than the patient in whom the C allele of rs2070935 was in trans with the GFAP mutation.

Although the minor allelic frequency of the A allele in rs2070935 of the GFAP promoter is 0.43, the C/A SNP allelic frequency differs among races, as follows: the A allelic frequency is 0.35 in Asians, 0.44 in Europeans, 0.39 in Americans and 0.54 in Africans (1000 Genomes; http://www.1000genomes.org/). Thus, the severity of AxD may differ among races. Additional studies are needed to clarify the natural history of late-onset AxD among races.

A limitation of our investigation was the extremely small sample size. However, it may be difficult to examine a sufficient number of subjects because AxD is extremely rare, with an estimated prevalence of approximately one in 2.7 million people in Japan.7 Furthermore, identical GFAP mutations in late-onset AxD are much rarer.7

In conclusion, our investigation demonstrated the possibility that a C/C genotype at rs2070935 in patients with late-onset AxD may be associated with an earlier onset of motor impairment and more rapid progression of ambulatory disability than the C/A and A/A genotypes at rs2070935, although larger-scale investigation and long-term evaluation are required.