For decades, scientists thought that Down’s syndrome resulted from genes carried on the extra copy of chromosome 21 that causes the disorder. More recently, researchers have recognized that gene expression becomes altered throughout the entire genome. Now, researchers led by Stylianos Antonarakis at University of Geneva Medical School, Switzerland, examine these genetic changes in minute detail. By examining a set of monozygotic twins, one of whom had Down’s syndrome, they were able to see that cells with an added chromosome 21 exhibit a distinct pattern of gene dysregulation across the genome. Chromosomal areas that are usually highly expressed quiet down, while those that are normally silent perk up. As the researchers report in the April 17 Nature, the findings suggest that epigenetic changes could be at play. 

“The authors may have hit upon a genetic mechanism that could explain some of the specific clinical and neurobiological phenotypes you see in Down’s syndrome,” said Elliott Mufson, Rush University, Chicago, who was not involved in the study. “Whether it’s going to lead to a game-changing effect in understanding the disorder is hard to know.” 

Researchers have known for some time that an extra copy of chromosome 21 causes genome-wide changes in gene expression (see Vilardell et al., 2011). Just how the genes are disrupted remained a mystery, however, because high-background genetic variation between people masks the effects of an extra chromosome. Antonarakis and colleagues had a rare opportunity to circumvent this limitation when they became aware of a unique pair of monozygotic twin fetuses that had been aborted at 16 weeks, one with trisomy 21 and the other without. These two were genetically identical, save for that one extra chromosome. By examining gene expression differences between them, the researchers were able to avoid the background variation that occurs between unrelated people.

First author Audrey Letourneau and colleagues measured levels of messenger RNA from four separate samples of skin cells from each of the twins. They found that an alternating pattern of expression changed throughout the genome of the twin with Down’s. Compared to the normal twin, some groups of genes were activated while others were suppressed. These up- and downregulated blocks lay next to each other on all chromosomes. The authors called these regions gene expression dysregulation domains (GEDDs). Remarkably stable, these GEDDs persisted when the researchers derived induced pluripotent stem cells from skin fibroblasts. To the authors’ surprise, the Ts65Dn mouse model of Down’s syndrome showed GEDDs that affected the same genes, even though they lie on different chromosomes in mice and humans.  

The GEDDs overlapped with specific chromosomal regions called lamina-associated domains (see Guelen et al., 2008) and early replication domains (see Hansen et al., 2010). LADs are areas of low gene expression that are marked by suppressive patterns in the local chromatin; conversely, early replication domains are highly active regions of the genome. In the trisomic twin, as well as the Ts65Dn mouse, LADs were unusually active, whereas early replication domains were curiously subdued (see image below). “The result is a genome-wide flattening of gene expression,” wrote Benjamin Pope and David Gilbert of Florida State University, Tallahassee, in an accompanying News and Views article.

Genomic domains that are normally associated with low or high levels of gene expression are respectively up- or downregulated in Down’s syndrome. [Image courtesy of Nature.]

The authors are unsure what drives this. They found more methylation of histone H3 (H3K4me3) in the areas of upregulated genes, suggesting that the DNA there was more accessible. 

Antonarakis proposes that an extra copy of one or more particular genes on chromosome 21 leads to widespread chromatin changes that disrupt transcriptional regulation and explain some of the phenotypes in Down’s syndrome. Alternatively, the mere presence of extra DNA in the nucleus could be the reason, he said. That would suggest other trisomies could lead to similar genome modifications. Antonarakis’ team is testing that possibility by comparing trisomic and normal cells from people who have a mix of these cells in their bodies. People with cancer often accumulate trisomic cells.

“Clearly trisomy has an impact on the entire genome,” said Roger Reeves, Johns Hopkins University School of Medicine, Baltimore. “Though it is sometimes subtle in magnitude, it has a profound effect.”—Gwyneth Dickey Zakaib 



  1. In general, I think the finding that there are large chromosomal regions that show either increased or decreased gene expression is fascinating. With regard to Down’s syndrome, the observation that there is a “flattening” of these domains—that is, domains of high gene expression have lower expression levels in Down’s syndrome, while domains of lower gene expression are more modestly decreased—will need to be taken into account when considering gene-expression studies on Down’s syndrome and other aneuploidies as well. The finding that the phenomenon of flattening of regulation of gene expression was masked by genetic heterogeneity in unrelated humans is quite important. It is consistent with, but greatly extends, other work that hinted that this might be the case.

    A very important finding for future work is that the gene expression domain pattern is conserved in a widely used mouse model of Down’s syndrome, the Ts65Dn mouse.  This finding is reassuring to investigators using mouse models of Down’s syndrome or other human disorders.

    It now will be quite important to determine the generality (or not) of this observation in other human aneuploidies and in various mouse models of Down syndrome and other conditions. 

    View all comments by David Patterson
  2. The authors analyzed gene expression differences (transcriptome differences) in a pair of monozygotic twins in which only one (T1DS) had Down’s syndrome. Gene expression was analyzed in fetal fibroblasts and in pluripotent stem cells derived from fibroblasts from both T1DS and the normal twin (T2N).

    Transcriptome analyses in humans have revealed that there are discrete domains in the genome that carry genes that are highly expressed (often housekeeping genes essential for all cellular processes) and other discrete domains where the located genes have low expression. Comparing the expression levels of genes within specific domains in the twins and in controls, the general effect observed was a flattening of expression levels in the Down’s syndrome twin.  For genes in domains where high levels of expression generally occur, the T1DS showed lower levels of expression than occurred in  T2N. For genes in domains where gene expression is generally low, T1DS had levels that were not as low as those in the T2N. The key point was that this flattening of gene expression occurred at many regions throughout the genome. Interestingly, in mouse models of Down’s syndrome the flattening of expression occurred in genes in regions that matched in gene content (syntenic regions). The authors determined that the altered gene expression levels were due to changes in chromatin, i.e., epigenetic changes. The chromatin alterations particularly involved the histones and nucleosomes in the trisomic cells.

    This study indicated that increased dosage of genes on chromosome 21 in Down’s syndrome, and increased levels of expression of specific chromosome 21-located genes, likely alter chromatin modifications and nucleosome structure and function as downstream effects. Of particular interest, therefore, are chromosome 21 genes that impact histones and nucleosomes. These include holocarboxylase synthetase (HCS), which adds biotin to histones and reduces gene expression, and HMGN1, which interacts with chromatin and nucleosomes and enhances gene expression. Other genes located on chromosome 21 that have epigenetic effects include DYRK1A, BRWD1, and RUNX1. The authors also related the domains of altered expression to structural features of the nucleus and nuclear membrane, e.g., laminin, that impact genomic architecture and transcription.

    The importance of this paper is that it demonstrates in Down’s syndrome differences in expression of genes located in specific domains throughout the genome. These differences are distinct from population differences in gene expression. In addition, the authors demonstrate that changes in gene expression in Down’s syndrome are epigenetic, i.e., due to alteration in histones and nucleosomes. Furthermore, the epigenetic alterations most likely reflect the downstream effects of increased dosage of specific chromosome 21-located genes.

    The study raises several important questions that are especially relevant to Alzheimer disease.

    1. Do any of the regions of altered gene expression in T1DS include Alzheimer’s-related genes?

    2. Would it be possible to alter expression levels or counteract excess gene product of the chromosome 21 genes, e.g., HCS, HMGS, or DYRK1, and determine if this “normalizes” patterns of gene expression in Down’s fibroblasts?

    3. And would it be possible to correlate studies reported by Letourneau et al. with the recently reported (Lu et al., 2014) studies indicating that altered levels of expression of REST promote Alzheimer’s disease? There is also data that altered levels of expression of the transcription factor ZNF335, which impacts histone methyltransferase expression, impact the size of the neural stem cell population and lead to microcephaly (Yang  et al., 2012).


    . REST and stress resistance in ageing and Alzheimer's disease. Nature. 2014 Mar 27;507(7493):448-54. Epub 2014 Mar 19 PubMed.

    . Microcephaly gene links trithorax and REST/NRSF to control neural stem cell proliferation and differentiation. Cell. 2012 Nov 21;151(5):1097-112. PubMed.

    View all comments by Moyra Smith
  3. Fascinating!

  4. The study suggested that nuclear compartments of trisomic cells undergo modifications of the chromatin environment influencing the overall transcriptome, and that gene expression dysregulation domains (GEDDs) may therefore contribute to phenotypic changes observed in individuals with Down’s syndrome (DS). To further substantiate this hypothesis, similar GEDDs were also found in the fibroblasts from a model of DS, the Ts65Dn mouse, suggesting that it may serve as a good model for genome-related alterations in DS and how they relate to phenotypes. Studies suggest that both the genetic and epigenetic background of each individual may affect the severity of phenotypes during both development and aging. Understanding genetic and epigenetic determinants of phenotypic development is therefore a major focus in DS research, and this manuscript by Letourneau and colleagues provides an elegant platform for future genome-wide studies in individuals with DS. Because people with DS now live longer than before (the mean life span is now >50 years of age), and almost all develop Alzheimer's neuropathology and are at high risk to develop dementia, studies into phenotype/genetic alterations in DS are timely and important.

    The study investigated gene expression differences between a pair of monozygotic twins—one born with DS and one without. The triplication of chromosome 21 (HSA21) occurred due to segregation errors following the twinning event. Using the twins for the genetic studies may eliminate bias of genome variability and also environmental differences that can give rise to epigenetic alterations and thus have effects on the transcriptome. Interestingly, the investigators found that differences in gene expression between the twins were organized in domains along all chromosomes, not just on HSA21. The same differential gene expression across the genome was observed in Ts65Dn mice versus normosomic littermates.  The investigators utilized mRNA sequencing to study the transcriptome of fetal skin cells from the twins, and found reductions in gene expression of secreted proteins involved in cytokine-cytokine receptor interactions, as well as inflammatory responses. These findings are highly relevant in terms of both the age of the mother at conception and the accelerated aging process that is observed in some organ systems in individuals with DS, including the brain. For example, genetic alterations in cytokine-cytokine receptor interactions have been observed in older versus younger women (Diez-Fraile et al., 2014), and these findings could be relevant in relation to DS since HSA21 trisomy occurs with higher frequency in women who are older when they conceive.  In addition, alterations in genes related to inflammation are known to be dysregulated in aged brains in general (Buga et al., 2014) as well as in the brains of those with DS (Rosety-Rodriguez et al., 2013), providing further potential clues to the phenotypic meaning of the genome-wide alterations observed herein. 

    When Letourneau et al. focused on general changes in all chromosomes, they found well-defined chromosomal domains that were upregulated, followed by large downregulated domains, suggesting a well-organized discordant expression profile, rather than random up- or downregulated genes, such that a “flattening” of the gene expression profile was observed. These segments could contain more than 500 genes, but have a median number of 20 genes in each segment.  These types of up- or downregulated chromosomal domains were absent in a pair of healthy twins, suggesting that this segmental gene-expression profile was specific to the HSA21 trisomy segment. In Ts65Dn mice, the investigators found that the GEDDs were also organized in segments along all chromosomes, just as in the human twin cases. The GEDDs were well conserved between the DS twin's fibroblasts and the mouse fibroblasts and the direction of dysregulation was maintained between the two species. This is highly important, since the Ts65Dn mouse model only copies about 130 of the nearly 300 genes coded on HSA21. The findings suggested that the GEDDs were independent of karyotypic context, since the syntenic regions were located in other chromosomes in the murine genome versus the human, and suggests that phenotypic development may be at least partially similar between the human condition and the mouse model.   The natural gene variability between unrelated individuals in general is extensive, and only stably expressed genes can unmask the effects observed with the discordant monozygotic twins. 

    The investigators propose two possible explanations for these genome-wide effects: 1) the over-expression of one or more HSA21 genes modifies the overall chromatin environment of the nuclear compartments of trisomic cells, which will then lead to perturbation of the transcriptome in all future cells, and 2) the GEDDs could simply be the result of the additional chromosomal material caused by the triplication itself. Future studies will have to determine whether the first or second alternative proposed here holds true for the expression changes observed. The second scenario could easily be tested using another form of trisomy to see if it gives rise to similar expression changes. However, monozygotic twins in whom one exhibits a chromosomal whole or partial trisomy are rare, but can provide valuable information, as Letourneau et al., demonstrated. Here, they had an important and elegant control, examining the transcriptome of two monozygotic twins without chromosomal trisomy; they found no genome-wide GEDDs. These findings are interesting and important and may explain some of the very general phenotypic alterations observed in individuals with DS—not the least the alterations that occur with aging.  It is possible that at least some of these GEDDs are related to basic metabolic function and energy homeostasis of cells; hence the altered metabolic imbalance observed both in normal aging (Hu and Liu, 2014) and in individuals with DS (Perluigi et al., 2014).  A recent study (Perluigi et al., 2014) reports alterations in the PI3K/Akt/mTOR axis in individuals with DS, which were present already early in life. This pathway regulates the basic metabolic rate of cells and if dysregulation of this pathway occurs in fetal development, it could explain the “early aging” that occurs in the brain and other organ systems in DS. Perluigi et al. pointed out a close connection between the dysregulated mTOR pathway and hyperphosphorylation of Tau protein, a neuropathological marker for Alzheimer’s disease (AD). Early alterations in basic cell metabolism, such as might occur due to some of the GEDDs observed herein, may therefore “set the stage” for the AD neuropathology that develops unusually early in individuals with DS (see, e.g., Zigman, 2013).  The “flattening” of gene expression may alter cellular transcription output. This may be consistent with altered basic cellular metabolic rate, and may therefore affect the mTOR pathway described above. 

    In conclusion, the manuscript by Letourneau et al. points to several important issues and provides a platform for future genome studies on DS and other genetic diseases, and for their impact on life span, health span, and pathology. The study raises several important questions regarding the interaction between all chromosomes and genes, and whether the observed GEDDs are always present in trisomy or other genetic abnormalities, and if that is the case; what functional effects these GEDDs might have. 


    . Transcriptomics of post-stroke angiogenesis in the aged brain. Front Aging Neurosci. 2014;6:44. Epub 2014 Mar 18 PubMed.

    . Age-associated differential microRNA levels in human follicular fluid reveal pathways potentially determining fertility and success of in vitro fertilization. Hum Fertil (Camb). 2014 Jun;17(2):90-8. Epub 2014 Mar 31 PubMed.

    . Targeting tissue-specific metabolic signaling pathways in aging: the promise and limitations. Protein Cell. 2014 Jan;5(1):21-35. Epub 2014 Jan 29 PubMed.

    . Neuropathological role of PI3K/Akt/mTOR axis in Down syndrome brain. Biochim Biophys Acta. 2014 Jul;1842(7):1144-53. Epub 2014 Apr 13 PubMed.

    . Resistance circuit training reduced inflammatory cytokines in a cohort of male adults with Down syndrome. Med Sci Monit. 2013 Nov 7;19:949-53. PubMed.

    . Atypical aging in Down syndrome. Dev Disabil Res Rev. 2013 Aug;18(1):51-67. PubMed.

    View all comments by Lotta Granholm

Make a Comment

To make a comment you must login or register.


Research Models Citations

  1. Ts65Dn

Paper Citations

  1. . Meta-analysis of heterogeneous Down Syndrome data reveals consistent genome-wide dosage effects related to neurological processes. BMC Genomics. 2011;12:229. PubMed.
  2. . Domain organization of human chromosomes revealed by mapping of nuclear lamina interactions. Nature. 2008 Jun 12;453(7197):948-51. Epub 2008 May 7 PubMed.
  3. . Sequencing newly replicated DNA reveals widespread plasticity in human replication timing. Proc Natl Acad Sci U S A. 2010 Jan 5;107(1):139-44. Epub 2009 Dec 4 PubMed.

Further Reading


  1. . APP-dependent alteration of GSK3β activity impairs neurogenesis in the Ts65Dn mouse model of Down syndrome. Neurobiol Dis. 2014 Jul;67:24-36. Epub 2014 Mar 15 PubMed.
  2. . Modelling and rescuing neurodevelopmental defect of Down syndrome using induced pluripotent stem cells from monozygotic twins discordant for trisomy 21. EMBO Mol Med. 2014 Feb;6(2):259-77. Epub 2013 Dec 27 PubMed.

Primary Papers

  1. . Domains of genome-wide gene expression dysregulation in Down's syndrome. Nature. 2014 Apr 17;508(7496):345-50. PubMed.