Strong association of the Saitohin gene Q7 variant with progressive supranuclear palsy

Patients and methods.

The sample population consisted of 49 patients with PSP and 62 control subjects. As previously described, the PSP cases were either pathologically confirmed, clinically typical PSP cases (n = 20), or met modified clinical diagnostic criteria for National Institute of Neurological Disorders and Stroke probable PSP (n = 29).6 Although prominent early postural instability was used as a positive inclusion criterion, falls in the first symptomatic year were not considered mandatory and diagnostic exclusion criteria were applied as available from case notes. Pathologic confirmation of the diagnosis of PSP was made following standardized criteria.6 Normal controls from a European Brain Bank series had neither clinical evidence of neurodegenerative disease nor abnormal histopathology (average age at death of 72 years). The study population was primarily of white origin from Britain and western Europe. This work was approved by the Joint Medical Ethics Committee of the National Hospital of Neurology and Neurosurgery.

Tau haplotype was unambiguously determined as previously described3 by analysis of the presence of the 238-bp intron 9 deletion of the H2 haplotype (figure 1).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. Representation of intron 9 of the tau gene (MAPT), showing relative distances (kb = kilobase pairs) of Saitohin gene (STH) coding sequence (CDS) from MAPT exons 9 and 10. Arrowhead over STH gene indicates direction of transcription. The relative position of the intron 9 deletion found in the H2 haplotype is indicated.

The STH gene was amplified with PCR primers FF-Cel 2: 5′-CCAAGTTCAGTTGCCATCTCC-3′ and OR Cel 2: 5′-CTCT-TGTGCATGGACCCTGTA-3′ flanking the STH coding sequence and the STH a → g polymorphism was analyzed after HinFI digestion (figure 2).1 Four random samples were sequenced in order to ensure accuracy of interpretation. We also sequenced the STH coding sequence of three pathologically confirmed, clinically typical PSP cases and three control samples that are H1/H1 homozygous in order to establish whether PSP cases carried any unique causative polymorphisms on the H1 backbone that were absent in unaffected controls. Statistical tests for Hardy-Weinberg equilibrium (HWE) were carried out by standard Pearson χ² tests. Tests for genetic association were carried out by Fisher exact test (QR and RR genotypes were combined in the genotype test). The test for LD was carried out by likelihood ratio testing of haplotype frequencies inferred assuming LD (using EM estimation) against haplotype frequencies inferred assuming no LD.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. Restriction enzyme analysis of Saitohin Q7R polymorphism. PCR products of the Q allele have four HinFI restriction sites resulting in five fragments (261, 243, 194, 54, and 38 bp). The R allele creates a new HinFI site, cutting the 194 bp fragment into two 97 bp fragments. The 54 and 38 bp fragments are not shown.

Results.

Automated sequence analysis of three pathologically confirmed clinically typical PSP cases and three unaffected controls homozygous for the H1 haplotype showed the absence of any pathogenic sequence variants of STH that belonged to the H1 haplotype, that exclusively segregates with the disorder. All six samples were homozygous for the Q variant, indicating that it predicts the H1 haplotype, as was further confirmed in the case-control study. There was no departure from HWE in the PSP (χ² = 0.0213, df = 1, p = 0.8841) and control (χ² = 1.7494, df = 1, p = 0.186) groups. The H2/H2, H2/H1, and H1/H1 genotypes associated perfectly with the RR, QR, and QQ genotypes. The inferred LD between the two loci appeared to be complete (χ² = 94.08, df = 1, p = 3 × 10⁻²²). Haplotype and genotype counts in PSP cases and controls are given in the table: comparison of allele frequencies shows that the Q variant and QQ genotype are considerably enriched in PSP (test for allelic association in all cases: p = 9 × 10⁻⁶; test for genotypic association: p = 4 × 10⁻⁶).

Discussion.

We have shown that the STH Q7R polymorphism appears to be in complete LD with the extended MAPT H1/H2 haplotype, with the geno-type state for STH completely predicting that of the MAPT haplotype. By definition, this extends the well-established association of the MAPT haplotype with PSP and CBD to STH and implicates the Q allele of this non-silent STH polymorphism as a potentially important candidate pathogenic variant in four-repeat tauopathies. However, the Q allele (H1 haplotype) is also the most common haplotype in the normal controls. This implies that there should be unfavorable interactions with other genetic or environmental risk factors or that the actual culprit variation that is unique to the H1 haplotype remains undiscovered. It would therefore be important to continue the search for H1-specific variations between epidemiologically matched case-control groups.

The strong LD and association are not surprising because the STH gene is nested between exons 9 and 10 of MAPT1 (see figure 1). Noting the complete LD between the STH Q7R polymorphism and the extended MAPT H1/H2 haplotypes, the observation of an association of the R allele and RR genotype with AD1 should apply to the entire H2 haplotype of the MAPT locus. However, several studies have failed to consistently confirm an association of the MAPT haplotype with AD (see references in reference 8), at best only implicating the MAPT locus as a weak risk factor in AD.7 More recently, three separate groups failed to show significant associations of the STH Q7R polymorphism with AD in much larger case-control groups.7-9⇓⇓ It is interesting that these studies with AD are not consistent. Because the STH Q7R polymorphism is part of the MAPT H1/H2 haplotypes, these differences in association that are observed in AD are not likely to be due to differences in LD.

The significance of the Q7R polymorphism remains to be elucidated. The STH gene codes for a protein of unknown homologies and function but has an expression profile similar to that of MAPT. A tantalizing possibility is the hypothesis proposing a functional relationship between tau and Saitohin based on parallels from the choline acetyltransferase (ChAT)/vesicular acetylcholine transporter (VAChT) locus.1 The VAChT gene is nested in the first intron of the ChAT gene and both genes are required for expression of the cholinergic phenotype and could be coregulated.10 Saitohin may play a similarly important role in tau regulation, and the Q7R polymorphism may cause a defect in this role, causing a predisposition to pathogenesis of PSP and CBD.