Alzheimer’s and Parkinson’s diseases both include tau pathology. Yet while the tau gene, MAPT, is a risk factor for Parkinson’s, its status for AD remains hazier, with some studies finding an association and others not. Recently, neuroscientists led by Rahul Desikan at the University of California, San Diego, analyzed data from several large genome-wide association studies (GWAS). They used a relatively new “gatekeeper hypothesis” method, which increased the statistical power to definitively link the MAPT region to Alzheimer’s. In the February 17 online edition of Molecular Psychiatry, they report that an AD risk variant near MAPT correlated with higher tau expression and greater atrophy of the entorhinal cortex and hippocampus. Intriguingly, these effects were most pronounced among people who lacked an ApoE4 allele, a major genetic risk factor for sporadic AD. Overall, the findings underscore the role of tau in Alzheimer’s, and suggest it may play a particularly significant role in non-ApoE4 carriers, Desikan said.

Commentators praised the methodology and said it holds promise for finding other genes that link different neurodegenerative diseases. “This sets the basis for future studies in which large GWAS data and a pleiotropic framework can be applied for the identification of novel AD-associated risk variants,” Mikko Hiltunen at the University of Eastern Finland, Kuopio, wrote to Alzforum (see full comment below). Carlos Cruchaga at Washington University in St. Louis noted, “The paper is technically very good, and as far as I know, the largest study comparing Alzheimer’s and Parkinson’s genetics.”

Previous GWAS tied variation at the MAPT locus to Parkinson’s disease in European populations, while additional genetic studies confirmed a role for MAPT in other tauopathies (see Tobin et al., 2008Simón-Sánchez et al., 2009; and Pittman et al., 2005). Alzheimer’s GWAS gave more equivocal results, however, with some finding that MAPT variants affected age of onset, but not overall risk, and others reporting only a weak effect (see Jun 2008 newsSep 2005 newsGerrish et al., 2012; Allen et al., 2014). 

Desikan and colleagues approached the problem from a different angle. Because Parkinson’s and Alzheimer’s can share some pathological features, such as tau and α-synuclein accumulation, they wondered if the same genetic variants might be risk factors for both. The authors selected eight SNPs that had genome-wide significance for Parkinson’s disease, and tested them for an association in five large Alzheimer’s GWAS. A typical GWAS study compares millions of SNPs, and to correct for the fact that some will appear significant just by chance, these studies have to set a high threshold for significance. By contrast, because the authors were analyzing only a handful of loci, they could use a lower significance threshold. This provided much greater statistical power to find an association, Desikan told Alzforum.

Their analysis turned up just one significant hit among the eight SNPs. In the MAPT region, SNP rs393152 reached significance in four of the cohorts. A meta-analysis of the five cohorts, which together comprised more than 21,000 cases and 51,000 controls, again turned up rs393152. However, standard GWAS statistical methods of the same groups would have lacked sufficient power to find the association, the authors noted. The more common adenosine allele of rs393152 conferred higher AD risk. This is the same allele that has been linked to Parkinson’s as well as to tauopathies such as progressive supranuclear palsy and corticobasal degeneration.

This SNP may not be the functional allele, as it tags a large haplotype, or block of DNA that is inherited en masse. Known as H1, this haplotype includes several other genes besides MAPT. To find out if MAPT was the actual risk gene, the authors performed additional genetic analyses on SNPs in this region that suggested most of the association with AD came from polymorphisms within or near the MAPT gene. The best association came from rs1981997, which sits within MAPT and was inherited along with rs393152 in every case.

The authors also examined expression of all genes in this region, and found that only MAPT transcripts varied with rs393152 genotype. They saw a dose effect, with two risk alleles leading to higher tau expression than one. Moreover, people carrying the risk allele lost brain volume in the hippocampus and entorhinal cortex more quickly than non-carriers did. When the authors divided people into ApoE4 carriers and non-carriers, they found that accelerated atrophy occurred mostly in the non-carriers. While it is not clear exactly why this is, it may be that in ApoE4 carriers amyloid pathways predominate, while non-carriers succumb more to tau-related neurodegeneration, Desikan speculated.

While the data support a link between MAPT and AD, the evidence to date has not proven that the tau gene drives the association, because many of the variants in the H1 region are inherited together, Cruchaga noted. Desikan said that in ongoing work, he is looking for the functional polymorphism in the region, and will try to further delineate whether the entire association with AD comes from the MAPT gene. He will also examine whether the risk allele associates with tangle pathology in postmortem samples.

What does the finding mean for Alzheimer’s? The A allele has a frequency of 77 percent in European populations, meaning that about 95 percent of people carry at least one copy. However, the risk it confers is quite small, increasing the odds of Alzheimer’s by only about 10 percent. This is similar to the risk conferred by GWAS hits such as BIN1 and clusterin. The importance of the finding lies instead in what it tells researchers about the pathobiology of Alzheimer’s disease, Desikan said. “It opens a window into what happens in non-ApoE4 carriers, and suggests other risk genes that may be important in that population. Overall, our findings suggest a need to develop a better understanding of tau pathobiology, as well as therapeutic agents targeting the tau protein.”—Madolyn Bowman Rogers


  1. Based on the stepwise pleiotropic approach, Desikan et al. identified an SNP (rs393152) in the MAPT gene region that affected the risk of both PD and AD. Further investigations revealed that the observed pleiotropy between AD and PD was non-polygenic and most likely linked to the MAPT locus. Importantly, this pleiotropic locus associated with increased MAPT transcript expression levels and augmented brain atrophy at the hippocampus and entorhinal cortex, particularly among the ApoE epsilon 4 non-carriers. This is a very important study, which emphasizes the potential genetic overlap between AD and PD at the MAPT locus via altered expression of tau.

    Although this study suggests shared pathology between AD and PD at the MAPT locus, it still possible that some other gene at the chromosome 17 locus could be responsible for the observed genetic association with AD. However, given the well-established role of tau in neurodegenerative disease, it is plausible that the observed results are actually mediated through MAPT-related mechanisms. As further support for this, the pleiotropic variant in MAPT tags the H1 haplotype, which has been previously associated with several tauopathies, such as corticobasal degeneration or progressive supranuclear palsy. It is also important to note that the present study successfully applied the use of hippocampal and entorhinal cortex atrophy rates as endophenotypes in the identification of AD-associated risk variant. Finally, this study sets the basis for future studies in which large GWAS data and a pleiotropic framework can be applied to the identification of novel AD-associated risk variants and their relationships between various neurodegenerative diseases.  

    View all comments by Mikko Hiltunen
  2. The stepwise gatekeeper approach to uncover genetic overlap (pleiotropy) between AD and PD implemented in this study by Desikan et al. detected a single, study-wide, statistically significant association at the MAPT locus. Meta-analysis of five independent AD case-control cohorts achieved highly significant association, providing convincing evidence that genetic variation in MAPT plays a role in risk not only for PD, but also for AD. In this article, the authors also demonstrate that a MAPT variant in tight linkage disequilibrium with the top PD GWAS hit associates with increased risk of AD, increased levels of MAPT expression in brain, and increased atrophy in two medial temporal lobe regions that are selectively affected by neurofibrillary tangle pathology in the earliest stages of AD. Although these results are not entirely surprising given previous reports that show association of MAPT variants with AD risk and with MAPT gene expression (Allen et al., 2014), the study by Desikan et al. provides validation of the latter findings and implicates these associations in AD neurodegeneration.


    . Association of MAPT haplotypes with Alzheimer's disease risk and MAPT brain gene expression levels. Alzheimers Res Ther. 2014;6(4):39. Epub 2014 Jul 1 PubMed.

    View all comments by Minerva Carrasquillo
  3. The major caveat of this paper is that the AD cases in all of the datasets used are not pathologically confirmed (i.e., the AD case status is assigned based solely on clinical diagnosis). This, combined with the fact that the association of the “Tau” SNP was dependent on the absence of the APOE4 allele, makes me wonder whether the “Tau” SNP association could be driven by misdiagnosis of other dementias. Indeed, the vast majority of APOE4 carriers with clinical diagnosis of AD dementia also have demonstrable Aβ deposition by PiB or autopsy. On the other hand, a sizable fraction of APOE non-carriers with clinical diagnosis of AD dementia are misdiagnosed, since they don’t show beta amyloidosis by PiB or autopsy. Therefore, validation of this finding with an independent PiB or autopsy-confirmed AD cohort is warranted.

  4. This is a helpful comment. We did actually use a small cohort of autopsy-confirmed AD cases/controls to confirm our findings (see Discussion and Supplemental information from our paper). In this small cohort of autopsy-confirmed AD cases and controls, we replicated the directionality and magnitude of the A allele of rs393152 (Supplementary Figure 5) indicating that our AD-associated findings are not due to misdiagnosis from other dementias.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Variations in Tau Gene Linked to Age of Onset for AD
  2. Tau Shows Subtle Hints of Genetic Association

Paper Citations

  1. . Haplotypes and gene expression implicate the MAPT region for Parkinson disease: the GenePD Study. Neurology. 2008 Jul 1;71(1):28-34. PubMed.
  2. . Genome-wide association study reveals genetic risk underlying Parkinson's disease. Nat Genet. 2009 Dec;41(12):1308-12. PubMed.
  3. . 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.
  4. . The Role of Variation at AβPP, PSEN1, PSEN2, and MAPT in Late Onset Alzheimer's Disease. J Alzheimers Dis. 2011 Oct 25; PubMed.
  5. . Association of MAPT haplotypes with Alzheimer's disease risk and MAPT brain gene expression levels. Alzheimers Res Ther. 2014;6(4):39. Epub 2014 Jul 1 PubMed.

External Citations

  1. MAPT
  2. BIN1 
  3. clusterin

Further Reading

Primary Papers

  1. . Genetic overlap between Alzheimer's disease and Parkinson's disease at the MAPT locus. Mol Psychiatry. 2015 Feb 17; PubMed.