Brain mosaicism and altered gene copy numbers could explain Alzheimer's

© EEG

I've tackled the problem of the missing heritability in the past, i.e. the fact that despite all the research on genetic studies and disease associations, we can explain only a small fraction of cancers and disorders. Today we know a lot more than what we knew back when the human genome project was completed, and in particular we know how much we don't know. I think we are only beginning to understand the complexity of human diseases and genetics. Back when I started working on disease associations, in 2004, we roughly thought that we could find a few "buttons" that would trigger cancer. Today we know that it's not about finding a few buttons. We thought DNA was more or less a keyboard, when in fact we have a whole orchestra: DNA, mRNA, proteome, epigenome, etc. Mutations can occur at any level, and besides genetic alterations there can be epigenetic alterations, proteins that don't fold correctly, an abnormal accumulation of proteins, and so many other ways that things can go wrong. To this add exercise, body mass, diet, and all kinds of other environmental exposures.

Bottom line: we set out looking for a few "keys" to play on the DNA keyboard when in fact we should be looking for a whole symphony, and the symphony may very well change from person to person.

Alzheimer's disease is among the many disorders that have baffled researchers. Recent studies have found that rather than genetic mutations in the DNA we should be looking at abnormal accumulation of misfolded proteins called amyloids. They can accumulate inside cells to a level where they become toxic and cause cell death. Amyloids have been associated with many diseases, not just Alzhemeir's, but in the case of Alzheimer's in particular they seem to be responsible for the progressive loss of neurons and brain connections. A gene called APP codes for a protein that is an amyloid precursor, and mutations in this gene have been associated with familial Alzheimer's (when the disease occurs before age 60). However, the vast majority of Alzheimer's cases occur much later in life and are not associated with those mutations. Do all these cases fall into yet another missing heritability mystery?

As it turns out, genetic alterations come in many forms, not just mutations. The key, in the case of APP, could be not in what kind of gene allele one carries, but rather on how many copies we carry. Yes, we may have more than two copies in different parts of the brain.

Remember when they taught us in school that we are born with one DNA and that DNA is identical throughout our cells and tissues? Forget that. Of all organs, the brain is one of the most plastic regions of our body, with numerous retrotransposons and mobile genetic elements that attest for its plasticity. Errors in chromosome segregation when cells divide can also produce cells with a gain or loss in chromosome numbers. This variability has been shown to be a common feature of the normal brain: our brains are genetic mosaics made of genetically distinct cell lines that have, over the years, accumulated somatic mutations [1].

I've discussed retrotransposons a while back: these are genetic elements that can make copies of themselves and then reinsert in different parts of the DNA. They are particularly active in the brain and thanks to their activity the brain is in fact a somatic mosaic of genetically distinct neurons. Retrotransposons were first discovered in maize and explain why, for example, a single cob can have kernels of many different colors: the cob is in fact a mosaic and the repositioning of the retrotransposons causes the kernels to display different colors. So now you can think of the human brain as a cob and the neurons are kernels of different colors. ;-)

There are multiple lines of evidence that seem to point at a correlation between number of copies of the APP gene in the brain and an increased risk of developing Alzheimer's. People with Down Syndrome, for example, have three copies of the APP gene and by age 65 they have a 75% chance of developing Alzheimer's. In a recent paper [2], a group of researchers from the Scripps Research Institute analyzed the nuclei of neurons harvested from the prefrontal cortex and cerebellum of postmortem brains, for a total of 134 samples (of which 47 from subjects with Alzheimer's) and concluded that the gene APP is mosaically amplified in brains affected by Alzheimer's. Some neurons had up to 12 copies of the APP gene. While their findings do not exclude the fact that the high APP gene dosage could be an effect of the disease, rather than the cause, even if it is an aftermath effect, it certainly plays a role in the progression of the disease.

I find it fascinating that more and more evidence seems to indicate that the etiology of diseases goes beyond single mutations. Protein folding anomalies like the accumulation of amyloids and gene dosage add new pieces to an incredibly complicated puzzle.

[1]Bushman DM, & Chun J (2013). The genomically mosaic brain: aneuploidy and more in neural diversity and disease. Seminars in cell & developmental biology, 24 (4), 357-69 PMID: 23466288

[2]Bushman, D., Kaeser, G., Siddoway, B., Westra, J., Rivera, R., Rehen, S., Yung, Y., & Chun, J. (2015). Genomic mosaicism with increased amyloid precursor protein (APP) gene copy number in single neurons from sporadic Alzheimer's disease brains
eLife, 4 DOI: 10.7554/eLife.05116