The tauopathy progressive supranuclear palsy, or PSP, is a poorly understood movement disorder that has recently come to be of intense interest to age-related neurodegenerative disease researchers. This week, scientists got important hints to the cause of this neurodegenerative disease. On June 19 in Nature Genetics, researchers reported that a large, genomewide association study confirmed the importance of tau in this disease and identified three new genes that increase the risk for PSP. “The results point to a link between PSP and cellular stress responses, and to a role for non-neuronal oligodendrocytes in the disease,” said Gerard Schellenberg, University of Pennsylvania, who led the research, together with Ulrich Müller from Justus Liebig University, Giessen, Germany. The study is the largest genetic analysis of the disease to date. PSP is one of many tauopathies, which are characterized by the accumulation of tau deposits in the brain upon postmortem pathology. “This is a really important study that will likely lead to a flurry of new investigations,” suggested Mark Cookson from the National Institute on Aging, Bethesda, Maryland. Cookson is one of more than 100 coauthors on the paper, reflecting a recent trend of collaboration and sample sharing in the neurogenetics community that is transforming the field from one based on small, often irreproducible, single-center studies.

Though the prevalence of PSP is fairly low at about four per 100,000, it is the second most common cause of parkinsonism. Scientists still puzzle over the role of tau in PSP and other tauopathies. In some cases, such as frontotemporal dementia, mutations in tau are causative, while in others there is no genetic link to tau. Researchers linked PSP to MAPT polymorphisms in 1997 (see Conrad et al., 1997). A study two years later demonstrated that inversion of a piece of DNA on chromosome 17 containing MAPT increases risk for PSP (see Baker et al., 1999), while a second major inversion, H2, does not. Some other genes have been linked to PSP but have not been confirmed.

Schellenberg, Müller, and colleagues adopted a genomewide association strategy to search for PSP genes. Joint first authors Günter Höglinger, from Philipps-Universität, Marburg, Germany; Nadine Melhem at the University of Pennsylvania; Dennis Dickson from the Mayo Clinic, Jacksonville, Florida; and Patrick Sleiman from Children’s Hospital of Philadelphia, Pennsylvania, used a two-stage analysis. In stage 1 they compared over 500,000 single nucleotide polymorphisms (SNPs) among DNA samples from 1,114 people with PSP and DNA from roughly 3,200 controls. They tested only samples from autopsy-confirmed cases to avoid spurious results. Most cases and controls were of European ancestry. Nearly 5,000 SNPs that had modest statistical significance were retested in the second stage, where the authors looked for association with PSP in 1,051 clinically diagnosed cases, comparing against about 3,500 controls. Schellenberg told ARF that the number of samples they obtained was staggering given the rarity of the disease, and was a testament to the cooperation among many labs.

From the first-stage analysis, Höglinger and colleagues found SNPs in three regions that reached genomewide significance. One of the SNPs lies on chromosome 1 near the gene for syntaxin-6 (STX6). Another, on chromosome 3 is in the region of the MOBP gene, which codes for myelin-associated oligodendrocyte basic protein. The third region, containing 58 SNPs, overlaps the MAPT inversion polymorphism. Second-stage analysis confirmed these associations and also a new one, near the EIF2AK3 gene, which codes for eukaryotic translation initiation factor 2-α kinase. The tau SNPs were the most statistically significant, with one of them (rs8070723) having a p value to -51 power (the threshold for significance in this study was 1.8 x 10-8) and accounting for the effect of the H1/H2 inversion. According to the authors' calculations, this SNP increases the chances of having PSP by 5.5-fold, making it a stronger risk factor for PSP than is ApoE4 for Alzheimer’s disease (see ARF related news story). A second tau SNP (rs242557) was associated with PSP (p = 1.3 x 10-11), even after correcting for the H1/H2 haplotypes.

How do these genetic variations contribute to PSP? In fact, SNPs that show positive associations in genetic studies are not necessarily those that alter susceptibility to disease. Instead, disease-associated SNPS are often co-inherited with true causative genetic variations, which may even lie in adjacent genes. In the case of the STX6 and EIF2AK3, however, the authors found nearby variants in each gene that alter the coding region and are tightly linked to the risk SNPs, suggesting the coding changes may be pathogenic. In the case of MOBP and MAPT, the researchers did not find evidence for coding change, but did for hints of gene expression effects.

Cookson and coauthor Andrew Singleton, also from the NIA, have developed a system to measure how genetic variation affects expression of thousands of genes across the human brain (see Gibbs et al., 2010). The authors used this type of analysis to discover that the MOBP SNPs may alter expression of itself and also the nearby SLC25A38 gene, which codes for “mitochondrial protein solute carrier family 25, member 38.” SNPs in the MAPT locus not only affected tau expression, but also expression of three nearby genes; however, correcting for the H1/H2 inversion at the tau locus eliminated these effects. In short, with this particular analysis, the authors could not find evidence that the rs242557 SNP had any effect on gene expression. There are no coding variations in MAPT that can explain its impact on PSP risk, either. Höglinger and colleagues suggest that rs242557 modulates alternative splicing without affecting overall expression.

The identification of three new genes now gives researchers a molecular handle on PSP, which is poorly studied, said Cookson. Schellenberg is particularly interested in the EIF2AK3 association, because the kinase, also known as PERK, plays a role in the unfolded protein response (UPR) of the endoplasmic reticulum (ER). The UPR blocks protein synthesis to allow misfolded proteins to clear the ER. It is activated in neurodegenerative disease, including Alzheimer’s and Parkinson’s (see ARF related news story and ARF news story).

“It is intriguing that tau should not be in the ER,” said Schellenberg, who laid out one hypothesis for how the UPR might be linked to tauopathy. In his scenario, misfolded proteins, such as tau, block the proteasome, and that activates the UPR stress response. Going a step further, he suggested that seeds of toxic tau might find their way into the ER where they could trigger cell death. Syntaxin-6, which contributes to endosome trafficking, may even be part of this scenario. Myelin-associated oligodendrocyte basic protein is found in the myelin in the white matter of the medulla, pons, cerebellum, and midbrain, regions that are damaged in PSP. Schellenberg suggested that myelin or oligodendrocyte dysfunction might contribute to PSP, or that the MOBP could trigger an immune response, much like myelin basic protein does in animal models of multiple sclerosis.

One further link that came out of the study was between PSP and ApoE. This was based on a SNP in TOMM40, which lies adjacent to the ApoE gene and is used as a surrogate for ApoE in genomewide association studies because the SNP platforms contain no ApoE markers. Interestingly, the data suggest that the ApoE4 variant, which is the ApoE risk allele for Alzheimer’s, is protective for PSP. “This was a surprise,” said Schellenberg, but he cautioned that the data did not reach genomewide significance. “Until we have a bigger sample set, we cannot draw any firm conclusions,” he said. He told ARF that in another year he hopes to have 1,000 more samples and plans to reanalyzed the data.—Tom Fagan


  1. This tremendous collaborative effort has yielded very exciting results that will have great impact on tauopathy research. Here we want to highlight the association of risk for PSP with SNPs in the EIF2AK3 region, which encodes the endoplasmic reticulum (ER) stress transducer PERK. ER stress induces the unfolded protein response (UPR), of which PERK is a major component.

    Our lab previously reported activation of PERK in AD brain in close correlation with tau pathology (1). Because PERK activation is found in neurons with diffusely distributed phosphorylated tau and not in tangle-bearing neurons, we hypothesized that the activation of the UPR is a very early event in AD pathogenesis that may precede tau pathology.

    PSP is a primary tauopathy that has no common co-pathology like the Aβ pathology in AD. The association of EIF2AK3 and risk for this tauopathy points to a key role for the UPR in the development of tau pathology in general, and warrants further investigation of UPR activation in other tauopathies. The UPR is initiated to restore protein homeostasis caused by ER stress, for example, by activating autophagy (2). What the trigger for UPR activation is and how this affects tau pathology are important challenges for future research and may open the possibility of intervention in tau pathology via the UPR (3). 


    . The unfolded protein response is activated in pretangle neurons in Alzheimer's disease hippocampus. Am J Pathol. 2009 Apr;174(4):1241-51. PubMed.

    . Endoplasmic reticulum stress activates autophagy but not the proteasome in neuronal cells: implications for Alzheimer's disease. Cell Death Differ. 2011 Jun;18(6):1071-81. PubMed.

    . Endoplasmic reticulum protein quality control in neurodegenerative disease: the good, the bad and the therapy. Curr Med Chem. 2009;16(5):615-26. PubMed.

    View all comments by Wiep Scheper

Make a Comment

To make a comment you must login or register.


News Citations

  1. AD Genetics—Problems and Promise
  2. Aβ Assault on Neurons Targets ER, Calcium
  3. Presenilin-1 Interferes with Protein Folding

Paper Citations

  1. . Genetic evidence for the involvement of tau in progressive supranuclear palsy. Ann Neurol. 1997 Feb;41(2):277-81. PubMed.
  2. . Association of an extended haplotype in the tau gene with progressive supranuclear palsy. Hum Mol Genet. 1999 Apr;8(4):711-5. PubMed.
  3. . Abundant quantitative trait loci exist for DNA methylation and gene expression in human brain. PLoS Genet. 2010 May;6(5):e1000952. PubMed.

Further Reading


  1. . Identification of a novel risk locus for progressive supranuclear palsy by a pooled genomewide scan of 500,288 single-nucleotide polymorphisms. Am J Hum Genet. 2007 Apr;80(4):769-78. PubMed.

Primary Papers

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