. Association of MAPT Subhaplotypes With Risk of Progressive Supranuclear Palsy and Severity of Tau Pathology. JAMA Neurol. 2019 Jun 1;76(6):710-717. PubMed.


Please login to recommend the paper.


  1. This study involves mostly the same samples analysed in a previous study (Pittman et al., 2005). A general problem in genetic studies is that often the same samples end up in many different studies, and this makes it difficult to assess whether something is a genuine confirmation, a technical replication or, often, a mixture of both. This is a problem plaguing analyses, and anonymization makes this problem worse. One needs to employ an auditor to follow the provenance of samples between institutions and analyses.

    That said, analysis of additional samples in this study have strengthened the statistics of the association of the H1c sub-haplotype and the protective nature of H2 that was first described by Alan Pittman and colleagues (Pittman et al., 2005). Heckman et al. also uncovered other associated H1 sub-haplotypes. It is also notable that the outcomes of good old-fashioned, pre-NGS linkage-based approaches remain robust.

    By weight of these data, the rs242557 single-nucleotide polymorphism (SNP) that tags the H1c haplotype in particular, is a central player. The commonest of the risk sub-haplotypes (H1c, H1d, H1g), and H1o, all conferring odds ratios between 1.86 and 3.64, carry the risk allele (A) of rs242557.

    Most telling, though, is the conditional analysis in the PSP genome-wide association study (GWAS) (Höglinger et al., 2011) where, due to the monolithic nature of the MAPT inversion region, i.e., near complete linkage disequilibrium, all the H1/H2 SNPs in the large inversion region are associated with PSP (Fig 2A from Höglinger et al., 2011, below). For this reason the GWAS team sought out associations that are independent of the H1/H2 dichotomy, and only rs242557 and adjacent SNPs retained genome-wide significant association, clearly implicating the region containing this SNP (Fig 2B from Höglinger et al., 2011, below).

    (© Nature Genetics.)

    The rs242557 SNP is in a conserved regulatory domain in the large promoter region of MAPT. In a simple cellular assay, Myers and colleagues (Myers et al., 2007) showed that the risk allele (A) was associated with significantly higher transcription levels off the MAPT core promoter. However, this has not been replicated in vivo—single-cell quantitation may be necessary to resolve cell-type specific vulnerability.

    Suggestive correlations of sub-haplotype with pathological features are of interest, but would have to be confirmed in a larger study.


    . Linkage disequilibrium fine mapping and haplotype association analysis of the tau gene in progressive supranuclear palsy and corticobasal degeneration. J Med Genet. 2005 Nov;42(11):837-46. PubMed.

    . Identification of common variants influencing risk of the tauopathy progressive supranuclear palsy. Nat Genet. 2011 Jul;43(7):699-705. PubMed.

    . The MAPT H1c risk haplotype is associated with increased expression of tau and especially of 4 repeat containing transcripts. Neurobiol Dis. 2007 Mar;25(3):561-70. PubMed.

    View all comments by Rohan de Silva
  2. PSP (progressive supranuclear palsy) is a rare neurodegenerative disease and typical representative of primary tauopathies (Dickson et al., 2007). 

    The gold standard for making the diagnosis is the postmortem evidence of high density of neurofibrillary tangles (NFTs) and neuropil threads in the basal ganglia and brainstem (Hauw et al., 1994). Additionally, oligodendroglial coiled bodies can be verified in variable amounts and distributions (Kovacs, 2015). Further alterations in typical PSP comprise atrophy in the subthalamic nucleus, accumulation of pigments and spheroids in the globus pallidus and substantia nigra, and grumose degeneration in the dentate nucleus (Kovacs, 2015). 

    A genetic breakthrough was achieved by conducting a genome-wide association study that included only autopsy-confirmed PSP cases (Höglinger et al., 2011). In this study, most of the PSP-associated single-nucleotide polymorphisms (SNPs) in the MAPT region mapped directly or closely onto the major two haplotypes, H1 and H2.

    These two haplotype clades derive from an inversion polymorphism about 3 million years ago of approximately 900kb on chromosome 17q21 and including the MAPT gene (Stefansson et al., 2005). Since that time, these H1 and H2 haplotypes have been recombinationally suppressed and have accumulated sequence variations. As a result, a large number of SNPs within the 900kb can be genotyped to distinguish the two haplotypes.

    In the current study, Heckman et al. carefully examined the H1/H2 haplotypes and the risk of progressive PSP. Using TaqMan single-nucleotide polymorphism genotyping, MAPT variants rs1467967, rs242557, rs3785883, rs2471738, and rs7521 were genotyped. Associations between these variant MAPT haplotypes and risk of PSP were analyzed. Five different MAPT subhaplotypes were significantly associated with risk of PSP. The previously reported associations with PSP risk for the H2 (rs8070723) and H1c (rs242557) haplotypes and three other H1 subhaplotypes, including H1d, H1g, and H1o, were found to be significantly associated with risk of PSP, too. Counterintuitively, several H1 subhaplotypes (H1l and H1p) were associated with a nominally significant decreased risk of PSP. After correcting for multiple testing, no haplotypes were significantly associated with tau pathology scores (i.e. a composite measure describing the burden of coiled bodies, neurofibrillary tangles, tufted astrocytes, and neuropil threads overall tau pathology scores). Nonetheless, there were indications of associations that fell only just below the significant threshold.

    The results by Heckman et al. suggest that it is worth following up these associations in growing autopsy-confirmed PSP cohorts. Furthermore, this study encourages examining MAPT subhaplotypes in other cohorts with autopsy-confirmed related neurodegenerative diseases such as other tauopathies (corticobasal degeneration, frontotemporal dementia, and others) or synucleinopathies (Parkinson's disease, dementia with Lewy bodies, multiple system atrophy) and others.

    In ongoing and upcoming clinical trials, the combination of genetic markers such as subhaplotypes and their relation to tau pathology burden might help to stratify study participants to clinicopathological variants and monitor treatment response to tau specific therapy.

    In summary, the findings here are based on solid ground by examining a large autopsy-confirmed cohort. They further delineate the genetic basis of risk for PSP.


    . Progressive supranuclear palsy: pathology and genetics. Brain Pathol. 2007 Jan;17(1):74-82. PubMed.

    . Preliminary NINDS neuropathologic criteria for Steele-Richardson-Olszewski syndrome (progressive supranuclear palsy). Neurology. 1994 Nov;44(11):2015-9. PubMed.

    . Identification of common variants influencing risk of the tauopathy progressive supranuclear palsy. Nat Genet. 2011 Jul;43(7):699-705. PubMed.

    . Invited review: Neuropathology of tauopathies: principles and practice. Neuropathol Appl Neurobiol. 2015 Feb;41(1):3-23. PubMed.

    . A common inversion under selection in Europeans. Nat Genet. 2005 Feb;37(2):129-37. Epub 2005 Jan 16 PubMed.

    View all comments by Günter Höglinger
  3. It is important to note that the H1 haplotype, which is present in around 80 percent of Europeans, is highly associated with PSP. Put simply, the H1 haplotype is for PSP as APOE is for AD. An extremely strong association.

    However, and opposite to APOE, the variant driving the association in H1 is not known. There are literally thousands of variants in the H1 haplotype, and multiple genes. The gene that encodes tau, MAPT is also in this haplotype, so it is very likely that the association of this haplotype is mediated through MAPT, but there is not an experimental demonstration of this.

    Looking at different H1 haplotypes can help to narrow down the functional variant and the gene driving the association. This is the largest study using PSP neuropathologically-confirmed samples and also looking at PSP neuropath measurements. The study confirms that there are specific H1 subhaplotypes like H1C that are associated with PSP. The study also identifies novel haplotypes that will help us to understand the impact of the H1 haplotype in PSP, so it is a step in the right direction.

    At the same time, as the authors comment, a lot still needs to be done. It is unknown what the functional effect of this haplotype is, if it affects MAPT expression or splicing, or any other gene in the region, as this haplotype still covers a large "chunk" of Chr 17. The functional variant remains unknown and additional studies, likely using de novo long-reads sequencing, are needed to understand what is going on with the H1 haplotype.

    View all comments by Carlos Cruchaga

Make a Comment

To make a comment you must login or register.

This paper appears in the following:


  1. Unequal: Some Tau Haplotypes Carry More Risk Than Others