ARTICLE IN BRIEF
In an analysis of data from brain tissue removed at autopsy from 185 people free of Alzheimer's disease, investigators reported that the patterns of gene expression found in carriers of an apolipoprotein E4 allele were closely related to the changes found in late-onset Alzheimer's disease, even though the brain tissue showed no pathological changes.
Even healthy people who carry an allele apolipoprotein E4 (APOE4) show brain changes suggestive of Alzheimer's disease, according to an analysis of transcription data reported in the Aug. 1 issue of Nature.
Researchers at Columbia University led by Asa Abeliovich, MD, PhD, analyzed publicly available transcriptome-wide association data from brain tissue removed at autopsy from 185 people free of Alzheimer's disease. The patterns of gene expression found in carriers of an APOE4 allele were closely related to the changes found in late-onset Alzheimer's disease (LOAD), even though the brain tissue showed no pathological changes. Out of 8,449 gene transcripts, 215 were found to be up-regulated or down-regulated in both APOE4 and LOAD samples, suggesting the two conditions are related.
“We hypothesized that there would be consistent, underlying changes in APOE4 carriers who don't have Alzheimer's disease, and we found changes in gene expression that looked like early Alzheimer's even in carriers who didn't have the disease,” said Dr. Abeliovich, associate professor of pathology and neurology in the Taub Institute for Research on Alzheimer's Disease and the Aging Brain at Columbia University Medical Center.
The researchers subjected their transcription data to differential co-expression analysis (DCA), which identifies genes expressed in similar ways. Among the 20 most highly ranked candidate genes they found several that had been previously implicated in the processing of amyloid precursor protein (APP), long suspected of playing a pivotal role in causing Alzheimer's disease. The top-ranked DCA hit, ring-finger protein 219 (RNF219), had not been linked to LOAD, but common genetic variants have been linked to alternations in lipid metabolism, cognitive performance, and ventricle volume — all associated with LOAD.
The researchers also applied APOE protein variants to mouse neuroblastoma N2a cells that overexpress a human APP transgene. They determined that APOE4 protein, but not APOE2 or 3 protein, significantly increased levels of amyloid-beta (Abeta) 40 and 42 in the cells. “Thus, the DCA-identified node hits included many known or novel potential modifiers of APOE4-dependent Aβ accumulation,” the authors stated.
“The paper significantly advances our knowledge of molecular and genetic mechanisms that modify LOAD risk, further implicating defects in endocytosis and intracellular trafficking in Alzheimer's disease development,” according to Daniel H. Geschwind, MD, PhD, Gordon and Virginia MacDonald distinguished chair of human genetics, and co-director of the Center for Neurobehavioral Genetics at the David Geffen School of Medicine at University of California, Los Angeles, and post-doctoral student Vivek Swarup, PhD, who co-authored an editorial about the work in the same issue of Nature.
Marek-Marsel Mesulam, MD, who has often expressed reservations about the amyloid cascade hypothesis, which dominates Alzheimer's disease research, called the paper a “tour de force” that hints at an answer to the two questions that most puzzle him about Alzheimer's: Why is age such a strong risk factor? And why does the disease target the hippocampus and the entorhinal cortex?
Dr. Mesulam, Ruth Dunbar Davee professor in neuroscience and director of the Cognitive Neurology and Alzheimer's Disease Center at the Northwestern University Feinberg School of Medicine, has long suspected that Alzheimer's results from a decline in the capacity for neuroplasticity followed by a period of compensatory but ineffective hyperactivity that eventually leads to degeneration. “Gradually, as plasticity requires more and more biological work, there would be plasticity-related overactivity in specific areas whose business it is to maintain the life-long synaptic reorganization required for memory, such as the hippocampus,” he said. “In that scheme of things, to find an intermediate step associated with hippocampal hyperactivity would fit very nicely with this hypothesis.”
The authors of the Nature paper provide an impressive demonstration of what modern genetics can do, according to Dr. Mesulam. “If they can find out exactly what the proteins they identify do, they may provide more tools for dealing with Alzheimer's disease.”
Dennis Selkoe, MD, finds support in the Nature paper for the amyloid cascade hypothesis, which he champions as the most persuasive explanation of how Alzheimer's develops. “Conventional late-onset Alzheimer's always has Abeta buildup — it's the sine qua non needed for the diagnosis,” said Dr. Selkoe, Vincent and Stella Coates professor of neurologic diseases at Harvard Medical School. “Since their APOE4 cohort has some of the same transcriptional changes, they can say that there's something about APOE4 that drives genetic changes that lead to an abnormality in Abeta metabolism even before people develop Alzheimer's. That's supportive of what we've known.”
However, he believes the authors departed from the strictly unbiased approach they adopted for their transcriptome-wide genetic analysis and decided to “cherry-pick” certain genes that other scientists have implicated as possible causes of LOAD. “Their results suggest that some of these genes are involved in APP processing, and as an APP guy myself, I find that interesting,” said Dr. Selkoe. “But as much confidence as I have in the amyloid hypothesis, I think they're focusing their experiments to show that APP metabolism could be altered by these genes they've discovered, especially RNF219 and SV2A — their top hits.”
Dr. Selkoe also had reservations about bathing mouse neuroblastoma N2a cells that overexpress human APP with APOE protein variants. “They acknowledge that this is a simple system,” he said, “but it's not really physiologic because APOE4 needs to be properly lipidated. They also haven't yet gone into animal models, which might have been a more convincing confirmation of their idea that RNF219 and SV2A are important modulators of APP processing.”
Ellen M. Wijsman, PhD, lead author of a 2011 paper in PLoS Genetics that used data from 3,839 individuals involved in a genome-wide association study of LOAD families to analyze the relationship between APOE genotypes and age of onset of Alzheimer's, also had reservations about the methods used in the Nature paper.
“There are some problems I see with sample size,” said Dr. Wijsman, a professor in the Division of Medical Genetics and department of biostatistics at the University of Washington in Seattle. “They're looking at 8,000 genes, but in only a couple of hundred people. When you have 8,000 genes from only 200 people, there are going to be correlations among some genes just because genes are expressed in the same person.”
Overall, however, she found the design of the study interesting, especially the contrast between people who are not demented but who have an APOE4 allele, and those who have only APOE3 alleles.
“They're hypothesizing that certain genes might be involved in central pathways, and if any of these hubs turn out to be true, that will provide really good targets for pharmaceuticals,” Dr. Wijsman said. “This paper generates a lot of hypotheses, and there's nothing wrong with a paper that generates hypotheses. That's what science is all about — getting new hypotheses to test.”