As is true with getting into college or receiving a job offer, it’s not just pedigree but also life experiences that may determine whether a person will develop Alzheimer disease. So suggests an analysis of identical twins—one who died of AD, one without AD—reported this month in the publicly accessible journal PLoS ONE. Researchers led by Paul Coleman, Sun Health Research Institute, Sun City, Arizona, examined postmortem brain tissue and found that cortical neurons from the AD twin had reduced DNA methylation, a biochemical process that can disrupt genes’ accessibility for transcription by attaching methyl groups to individual nucleotides.

In an earlier study (Mastroeni et al., 2008), first author Diego Mastroeni and colleagues found lower levels of DNA methylation, as well as reduced expression of DNA methyltransferase and other methylation regulators, in affected brain areas of sporadic AD patients. “This led to the question of whether these epigenetic effects we saw in AD were related to the [people’s] genes or to their life experience,” said Coleman, who is also a professor emeritus at the University of Rochester Medical Center, New York. In the sporadic AD study, genetic backgrounds were all over the map—which is why the scientists leaped at the opportunity to analyze epigenetic markers in identical twins discordant for AD. “This was a situation in which the genetic background would be quite similar, if not identical, and anything we saw could be attributed to life experience,” Coleman said. Other research has shown that identical twins who are genetically prone to AD can differ markedly in their age of onset and degree of pathology (Brickell et al., 2007).

The twins in the current study were white males who attended the same schools and worked as chemical engineers. One encountered extensive pesticides in his work and died at age 76 after a 16-year battle with Alzheimer disease. The other worked in a different environment and was cognitively normal when he died of prostate cancer at age 79. Pathologically, their brains could not have looked more different. At the time of his death, the twin with AD had an anterior temporal neocortex riddled with amyloid plaques and neurofibrillary tangles, the two key pathological hallmarks of AD. In his non-demented brother, however, “we had to hunt through the brain sections in order to find even one neurofibrillary tangle,” Coleman told ARF. The cognitively intact man also had comparatively higher expression of 5-methylcytosine, a marker of methylated cytosine-guanine (CpG) sites on DNA, in neurons, reactive astrocytes, and microglia of brain areas typically vulnerable to AD.

Apart from disease status, DNA methylation appears to vary with age and environmental factors. In a recent analysis of 217 non-pathological human tissues, published this month in PLoS Genetics (Christensen et al., 2009), researchers report that genes in CpG islands become increasingly methylated as people get older, whereas genes outside of these methylation hotspots lose methylation with age. Methylation status also correlated with environmental exposures such as tobacco smoking in that analysis, led by Karl Kelsey at Brown University. In an earlier study, Manel Estreller and colleagues at the Spanish National Cancer Center, Madrid, analyzed identical twins and found that DNA methylation status was very similar when the siblings were young but diverged more and more as they got older (Fraga et al., 2005). Those papers “make the case for environmental and aging effects on methylation,” Coleman said of the Estreller and Kelsey studies. “Our research shows that the concept of life events affecting DNA methylation may apply to development of the AD phenotype. It also stresses the potential importance of epigenetic phenomena in molecular mechanisms of AD.”

The new data may have ramifications for interpreting studies of AD genetics. “One study will find that, yes, this gene is a risk factor for AD, and others say, no, it’s not, and the statistics have some uncertainty in them,” Coleman said. “We raise the question of whether the probabilistic nature of the relationship between some genes and AD may be due to the fact that the genetic effects can be modulated by life experience.”

A recent study in Iceland may offer a case in point. Researchers at the University of Iceland and at deCODE Genetics, Reykjavik, reported a drastically shortened lifespan over the last 20 years in people with a hereditary amyloid angiopathy, and attribute this to diet changes that may have exacerbated the effects of a genetic mutation tied to the disease (Palsdottir et al., 2008 and ARF related news story). Studies in AD mouse models that overexpress mutant amyloid precursor protein (TgCRND8 and 129Sv) offer another example of a diet-gene interaction. When put on a diet deficient in folate, B1, and B6, the AD mice had reduced brain methylation activity in conjunction with amyloid-β overproduction and cognitive impairment (Fuso et al., 2008).

A link between epigenetics and AD also came up in a recent investigation led by Axel Schumacher at the Klinikum Rechts der Isar, Munich, Germany. However, unlike the current study, which reveals global demethylation in affected brain areas of the AD twin, Schumacher’s showed that most DNA methylation changes in AD brains are subtle and restricted to specific genes, including several involved in amyloid-β processing (PSEN1, ApoE) and methylation homeostasis (MTHFR, DNMT1) (Wang et al., 2008 and ARF related news story). In an e-mail to ARF, Schumacher noted that analyzing late-stage disease tissue makes it hard to determine whether the observed epigenetic phenotypes are the cause or the result of the disease. In the new study, “the global demethylation in the affected brain areas may indicate that specific components of the epigenetic machinery, such as DNA maintenance methylation, were inactivated, which in turn could indicate that the observed epigenetic patterns result from the course of the disease,” he wrote (see full comment below).

Coleman hopes to address this possibility in a genomewide study to identify specific genes affected by DNA methylation in AD, he told ARF. Future work in this area may benefit from a new approach that uses flow cytometry and state-of-the-art sequencing techniques to quantify the number of methylated molecules in a sample. Its developers show the method is sensitive enough to detect one methylated molecule in about approximately 5,000 unmethylated molecules in DNA from plasma or fecal samples. In a report published online 16 August in Nature Biotechnology (Li et al., 2009), researchers led by Sanford Markowitz, Case Western Reserve University, Cleveland, Ohio, and Bert Vogelstein at Johns Hopkins University School of Medicine, Baltimore, Maryland, have used the technology to detect early-stage colorectal cancer.—Esther Landhuis


Make a Comment

To make a comment you must login or register.

Comments on News and Primary Papers

  1. There are many observations, including from our own laboratory, that indicate that epigenetic drift is likely to be a substantial mechanism predisposing individuals to LOAD and contributing to the course of disease. In this context, the study by Mastroeni et al. is a very interesting report, as we may gain more insight into epigenetic events in AD. However, in my opinion, the study presents a potentially unusual epigenetic phenotype in the affected co-twin. In a previous study from our group (Wang et al., 2008), we were able to show that most DNA methylation changes in AD brains are restricted to specific genes and are rather subtle. In this new study of discordant twins, the authors found significant global demethylation in the affected brain areas of the AD twin. In general, such rare monozygotic twins discordant for a disease offer a great opportunity to study molecular events that may contribute to a predisposition or the development of a complex disease such as AD. And indeed, this observation is highly interesting as it demonstrates clearly that epigenetic mechanisms are affected in AD. However, the intricacy is that, similar to epigenetic studies in cancer, we look at the endpoint of the disease, where it is difficult to establish if the observed epigenetic phenotype is the cause or the result of the disease. In this case, the global demethylation in the affected brain areas may indicate that specific components of the epigenetic machinery (such as DNA maintenance methylation) were inactivated, which in turn could indicate that the observed epigenetic patterns are rather the result of the course of the disease.

    In addition, we also see that merely measuring DNA methylation levels in postmortem brain samples of AD patients with a long AD history may not be enough in the long run, as we primarily observe the endpoint of the disease. In this case, the affected twin had lived already more than 16 years with the disease. Hence, it is important to identify epigenetic events that happen during very early stages of AD, or even before AD symptoms occur in the first place! It may also be necessary to study the age effects that are evident in AD. For example, in our study, we identified a notably age-specific epigenetic drift in AD patients, supporting a potential role of epigenetic effects in the development of the disease. The occurrence of early epigenetic changes in a significant subset of younger AD patients may be indicative of AD-specific epigenetic abnormalities predisposing to AD. It seems that specific genes in the human brain have a higher likelihood of developing abnormal epigenetic patterns, meaning they are epigenetically unstable. Such metastability could be due to vulnerable chromosomal regions, to environmentally induced changes affecting specific pathways in the brain, but also simply to stochastic fluctuations. For example, we found that some genes that participate in amyloid-β processing (PSEN1, ApoE) and methylation homeostasis (MTHFR, DNMT1) show a significant interindividual epigenetic variability, which may contribute to LOAD predisposition. For the present study on the twins, this could potentially indicate that the affected twin may have had an unfavorable epigenetic event affecting the DNMT1 or MTHFR gene, thereby interrupting methylation homeostasis in certain areas of the brain.

    It is noteworthy, though, that the non-affected twin also shows weak signs of AD pathology, indicating that he also may have been predisposed to AD. It seems likely that, had he lived longer, he might have developed AD symptoms as well. Such observations in AD twins are not new; from older twin studies we know that the onset of AD in identical twins can differ by more than 20 years. Rather than genetic causes, epigenetic factors are probably much better suited to explain the observed anomalies in AD, as individual people may acquire aberrant epigenetic patterns during many developmental stages. It is important to note that it is unlikely that age-dependent epigenetic drift will manifest itself by switching AD susceptibility genes completely on or off, as observed in the affected twins in this study. That is true especially if the majority of changes are due to stochastic fluctuations, which could be more common than is generally assumed.

    One important finding of this study is that epigenetic abnormalities were restricted to certain brain tissues. This finding could indicate, again, that the observed methylation patterns are the result of the disease and not the cause. On the other hand, it is also plausible that epigenetic events happened during early tissue differentiation stages, predisposing the twins to AD, because later environmental factors (such as work-related chemical exposures) are unlikely to affect only specific brain areas, but rather the whole brain. Small epimutations in the critical genes may be tolerated to a certain degree and merely reflect the range of interindividual variance. Environmental factors could be the triggers that push an epigenome across the disease threshold with the result that the brain starts to malfunction (see also “Epigenetic theory of late onset AD” in Wang et al., 2008). In future twin studies, it may be helpful to study additional tissues outside the affected brain, to learn when during the development of a human being critical epimutations occur and how the environment affects these events.

    View all comments by Axel Schumacher
  2. After reading with great interest the comment by Dr. Schumacher and the response by Dr. Coleman, I'd like to point out that the demonstration that B vitamin deficiency led to decreased DNA methylation (missing in our 2008 paper) was actually given in our recent paper on PS1 promoter demethylation (Fuso et al., 2009).

    I completely agree with the conclusion that there is much more to understand in the area of epigenetic changes in LOAD. It seems to me of great importance that different approaches are applied by different groups to investigate this topic.


    . Changes in Presenilin 1 gene methylation pattern in diet-induced B vitamin deficiency. Neurobiol Aging. 2011 Feb;32(2):187-99. PubMed.

  3. Dr. Schumacher’s commentary about our paper makes a number of valid points that, in their totality, emphasize that there is much still to be learned about epigenetics with regard to the normally aging and Alzheimer brain. For example, he refers to “epigenetic drift” and “stochastic fluctuations,” phrases that imply a random process. We, on the other hand, prefer to use the term “life events,” which implies a causal connection between specific events and epigenetic consequences. Such causal connection is consistent with the work of Fuso et al. (2008), which shows that “PS1 and BACE genes can be upregulated even in vivo by B vitamin deficiency, a condition that limits methylation activity.” Of course, what is missing here is the demonstration that the experimental B vitamin deficiency led to decreased DNA methylation (or other epigenetic regulator) of the specific genes affected in their animals.

    The hypothesis that life events, rather than a stochastic process, influence epigenetic phenomena is also consistent with the comment in Fraga et al. (2005) that similarities in the epigenome of the identical twins they studied was related to the amount of time they spent together. In a stochastic process one would expect that time only would determine similarity/dissimilarity, rather than time spent together.

    Of course, an influence of life events and a stochastic process are not mutually exclusive (e.g., Poulsen et al., 2007). Further research is needed to determine the role that each may play in the epigenetics of aging and Alzheimer disease.

    Dr. Schumacher also raises the appropriate issue of whether epigenetic changes in AD are a result or a cause of the disease. It appears to us that this is more complicated than an either/or proposition. For example, data indicate that the APP gene can be methylated (West et al., 1995) and also that Aβ induces epigenetic effects (Chen et al., 2009). Again, further research is needed to resolve this issue.

    Dr. Schumacher raises other important areas needing further research, including the elucidation of epigenetic events during the very early stages of AD and their relationship to age effects “that are evident in AD.”

    View all comments by Paul Coleman


News Citations

  1. Party of Three: Genes, Environment, and Epigenetics
  2. Drifting Toward AD—Epigenetic Changes Linked to Disease

Paper Citations

  1. . Epigenetic changes in Alzheimer's disease: decrements in DNA methylation. Neurobiol Aging. 2010 Dec;31(12):2025-37. PubMed.
  2. . Clinicopathological concordance and discordance in three monozygotic twin pairs with familial Alzheimer's disease. J Neurol Neurosurg Psychiatry. 2007 Oct;78(10):1050-5. PubMed.
  3. . Aging and environmental exposures alter tissue-specific DNA methylation dependent upon CpG island context. PLoS Genet. 2009 Aug;5(8):e1000602. PubMed.
  4. . Epigenetic differences arise during the lifetime of monozygotic twins. Proc Natl Acad Sci U S A. 2005 Jul 26;102(30):10604-9. PubMed.
  5. . A drastic reduction in the life span of cystatin C L68Q carriers due to life-style changes during the last two centuries. PLoS Genet. 2008 Jun;4(6):e1000099. PubMed.
  6. . B-vitamin deprivation induces hyperhomocysteinemia and brain S-adenosylhomocysteine, depletes brain S-adenosylmethionine, and enhances PS1 and BACE expression and amyloid-beta deposition in mice. Mol Cell Neurosci. 2008 Apr;37(4):731-46. PubMed.
  7. . Age-specific epigenetic drift in late-onset Alzheimer's disease. PLoS One. 2008;3(7):e2698. PubMed.
  8. . Sensitive digital quantification of DNA methylation in clinical samples. Nat Biotechnol. 2009 Sep;27(9):858-63. PubMed.

Further Reading


  1. . Age-specific epigenetic drift in late-onset Alzheimer's disease. PLoS One. 2008;3(7):e2698. PubMed.
  2. . Epigenetic changes in Alzheimer's disease: decrements in DNA methylation. Neurobiol Aging. 2010 Dec;31(12):2025-37. PubMed.
  3. . A drastic reduction in the life span of cystatin C L68Q carriers due to life-style changes during the last two centuries. PLoS Genet. 2008 Jun;4(6):e1000099. PubMed.

Primary Papers

  1. . Epigenetic differences in cortical neurons from a pair of monozygotic twins discordant for Alzheimer's disease. PLoS One. 2009;4(8):e6617. PubMed.