By Turna Ray
The discovery of a new predictive gene for Alzheimer's disease by Duke University researchers may lead to the development of a new diagnostic test that can pinpoint whether people will develop the degenerative disease and approximately at what age the disease will manifest.
At the International Conference on Alzheimer's Disease in Vienna this weekend, a team of researchers led by Allen Roses, director of Duke's Deane Drug Discovery Institute, presented new data showing that long, repeated sequences of the TOMM40 gene linked to APOE3 and APOE4 polymorphisms is a predictor of whether a person over 60 years old will develop Alzheimer's disease within a five- to seven-year window.
Variations in apolipoprotein E (APOE) genotypes are associated with increased risk and age of onset for Alzheimer's disease. In particular, APOE4 accounts for 50 percent of late-onset Alzheimer's cases. Now, with this new discovery, APOE4 and TOMM40 together explain between 85 percent and 90 percent of the genetic effect on Alzheimer's disease, according to Roses.
The study presented at ICAD was funded by the Deane Drug Discovery Institute and Zinfandel Pharmaceuticals, a company started by Roses for the purpose of developing a predictive assay based on this discovery. Any IP developed from this study belongs to the Deane Drug Discovery Institute. Commercialization plans for the diagnostic — whether Duke or Zinfandel Pharma in collaboration with an industry partner — have not been solidified pending validation studies.
Although Roses' team has been working on this research project for over a year, it was "only in the last five weeks that we narrowed it down to the repeat" sequences of TOMM40, Roses said.
“Genome-wide screening detects big blocks of DNA inherited together, but it doesn’t tell us all the differences within that block,” Roses said in a statement. “We conducted a phylogenetic analysis to explore the evolution of the DNA and to see what changes take place on the backbone of other changes.”
Despite the fact that the finding still requires validation, Roses did confirm to Pharmacogenomics Reporter last week that his team intends to develop this into a diagnostic.
"We've already presented some of the data to the US Food and Drug Administration," Roses added.
Roses' team is currently sharing this initial research data with the FDA under the VXDS program, which allows researchers from academia and industry to get agency input on pharmacogenetic data for investigational products. The agency does use the data discussed under this program in its review process for future submissions.
Instead of using genome-wide association techniques, Roses and his team used phylogenetic analyses, previously used in HIV disease and flu virus research, to uncover the link between TOMM40 status and age of Alzheimer's onset.
"We're hearing a lot of noise about how complex diseases are going to be a whole bunch of small [gene associations] and we are going to need 50,000 patients to be able to do more genome scans," said Roses, who previously discovered the link between APOE and Alzheimer's. "I don't think that's true."
The abstract Roses originally intended to submit to ICAD discussed an algorithm using multiple SNPs to predict disease risk and age of onset. "We don't need to do that anymore," Roses said.
Rather than GWAS, Roses and his team used primary deep sequencing data from 340 individuals, which revealed that patients with variations in APOE3 were more likely to have either short or long repeated sequences of TOMM40, while patients with APOE4 had only the long repeat sequences of TOMM40. Polymorphic DNA Technologies, a sequencing services firm based in Alameda, Calif., conducted the sequencing reported in this study.
Based on this finding, Roses' team concluded that longer versions of TOMM40 that co-occur with both APOE3 and APOE4 are significantly associated with earlier disease onset, while the short repeat sequences were associated with a later onset of disease.
Specifically, researchers found that a single variable poly-T polymorphism, rs10524523 in the TOMM40 gene, accounts for differences in age of onset of Alzheimer's disease. The researchers also determined in this study that TOMM40 status associated with APOE3 strands is inherited independently from APOE4-TOMM40 strands.
[ pagebreak ]
According to Roses, patients with a short version of TOMM40 and APOE3 will not get Alzheimer's until after age 80. However, the long version of TOMM40 occurring with either APOE3 or APOE4 means the person will likely get Alzheimer's before the age of 80.
The Duke team is planning to validate how accurately APOE genotypes and TOMM40 status can predict age of disease onset within a five-year span for normal subjects recruited from epidemiologic populations between age 62 and 87. One part of this study will be a pharmacogenetic analysis using rs10524523 variants to help guide treatment with an Alzheimer's drug.
Roses did not name which drug would be used in this study, but said discussions are ongoing with several pharmaceutical companies regarding funding such a study and donating a therapeutic for the trial.
A genetic test developed out of this research, "can predict who is at high risk of getting Alzheimer's disease based on whether they are over 60 years, their TOMM40 poly-T repeat, and their APOE genotype," Roses said.
The predictive and prognostic assay that Roses intends to develop from this data will mostly likely be based on sequencing technology.
"One of the things we're testing is a whole bunch of next-generation sequencers," Roses said. "So, first you test the sample with PCR technology and then look at 100 or 1,000 cycles using a next-generation sequencer. These methods may be able to tell the difference between the long strand size and the small strand size" of TOMM40.
Roses' group has not yet decided which specific method to use in developing a commercial test, but once a test is designed, Roses said he intends to clear it through FDA.
In order to run the pharmacogenetic arm of the validation study, the FDA will need to simultaneously review the test as it looks at the design of the study, Roses noted. His team has already gotten some feedback from the FDA through its VXDS program.
"We want to submit to the FDA. We want to do it to show that if you have the right study design, [Rx/Dx co-development] is not impossible to do," Roses said. "This VXDS board that we're addressing has representatives from both the drug and diagnostic center. We're talking to both groups at the same time, in the same room, so they can discuss what we need to do."
Ultimately, Roses' team intends to use phylogenetic analyses to uncover gene associations with other diseases, such as autism, diabetes, and chronic obstructive pulmonary disease.