A comparison of identical human twins, only one of whom has Down's syndrome, reveals a genome-wide flattening of gene-expression levels in the affected individual. See Article p.345
Down's syndrome occurs when humans have an extra copy of chromosome 21 (ref. 1), a situation referred to as trisomy 21. Because each chromosome contains a distinct set of genes that serve as blueprints for the expression of cellular components, it has been presumed for decades that the condition is mainly caused by an overabundance of the products of chromosome 21 genes. But on page 345 of this issue, Letourneau et al.2 report a case of Down's syndrome that is associated with altered gene expression across every chromosome, not just chromosome 21. This observation implies that the expression of any number of genes on any chromosome may contribute to Down's syndrome, and raises the possibility that an extra copy of any chromosome can disrupt general gene regulation.
The discovery was made possible by an elegantly controlled experiment that compared a set of twins derived from the same fertilized egg (monozygotic, or 'identical', twins) in which one twin had an extra chromosome 21 and the other did not, owing to chromosome-segregation errors that occurred before the twinning event3. This unusual circumstance allowed the effects of the extra chromosome 21 to be studied in isolation. Although gene expression has been extensively studied in individuals with Down's syndrome, the genome-wide effect discovered by Letourneau et al. had gone undetected because, as they show, natural variation among individuals is strong enough to mask the effect.
Importantly, the authors found that the altered gene expression followed a consistent pattern, with increased and decreased gene-expression levels alternating across large chromosomal segments. The discovery of these up- and downregulated segments, which Letourneau et al. call gene expression dysregulation domains (GEDDs), supports mounting evidence that chromosomes contain functional domains that may help to provide cells with access to the genetic information at the appropriate place and time. The positions of the GEDDs align with chromosome domains defined by other structural and functional properties, such as domains that associate with nuclear lamina proteins (lamina-associated domains; LADs4) or that are replicated at different times during the DNA-synthesis phase of the cell-division cycle5. These findings strengthen the idea that chromosome functions reflect underlying structural domains.
The authors also report the presence of GEDDs in mice that carry an extra piece of chromosome 16 (the mouse counterpart to most of human chromosome 21) and that show several features of Down's syndrome6. The GEDDs were observed throughout the mouse genome at positions corresponding to their locations on human chromosomes. Furthermore, the authors demonstrate that the domains were largely preserved after the human twins' cells were artificially reprogrammed to a developmental state resembling that of embryonic stem cells (induced pluripotent stem cells)7. The authors understandably focus on the similarities between GEDDs before and after this reprogramming. However, the differences that they observed may correspond to the changes in replication timing, or perhaps lamina association, that occur during reprogramming5.
Intriguingly, Letourneau and colleagues show that GEDDs with increased expression corresponded to otherwise repressed genomic domains, whereas GEDDs with decreased expression corresponded to domains normally characterized by active transcription (Fig. 1). This means that there is a diminished difference between expressed and repressed genes in people with Down's syndrome, suggesting that the extra chromosome 21 interferes with the cell's ability to regulate transcriptional output.
The authors made several attempts to understand the mechanism behind GEDDs, but they found no significant changes in LADs or in patterns of DNA methylation — a modification that affects gene-transcription rates. They did find that levels of trimethylation at amino-acid residue lysine 4 on histone H3 correlated well with the transcriptional changes seen in GEDDs (histones are proteins around which DNA is wound in the nucleus, forming a complex called chromatin), but this is to be expected because such post-translational histone modification tracks with expressed genes8. The results of the authors' investigation of chromatin accessibility within GEDDs (the accessibility of chromatin to gene-transcription machinery also regulates expression levels) were difficult to interpret.
So how could the addition of a single, relatively small chromosome — chromosome 21 is the smallest human chromosome and accounts for less than 2% of the genome — dampen transcriptional differences across the genome? Two kinds of mechanism seem most plausible. First, and perhaps most simply, it is possible that the increased dosage of one or more genes on chromosome 21 is responsible. For example, human chromosome 21 and mouse chromosome 16 carry the HMGN1 gene, the product of which competes9 with histone H1 for access to the linker DNA between nucleosomes, the repeating units of chromatin. Because H1 is associated with less-accessible chromatin, an increase in dosage of HMGN1 would be consistent with an increase in global chromatin accessibility. Increased access to normally inaccessible chromatin would be expected to dilute the activity of factors that switch on genes in other parts of the genome, or release factors that repress genes in active regions, or both, with the net effect of flattening gene-expression levels genome-wide. An obvious experiment would be to examine the effect of controlled overexpression of HMGN1 on global transcription levels.
A second, much less defined possibility is that the phenomenon described by Letourneau et al. results from the extra DNA content, for example by sequestering factors that regulate expression10. This hypothetical mechanism need not be specific to chromosome 21 and could be explored further by comparing monozygotic twins that differ in other trisomies. Although less common, most other trisomies do cause some of the clinical features of Down's syndrome, and extra copies of larger chromosomes are associated with more-extreme effects11. Sex chromosomes, which are much more benign in a trisomic context, are the exception11.
Letourneau and colleagues have used a set of well-controlled, carefully performed and reproducible experiments to report a provocative new phenomenon. Their findings raise more questions than they answer, and open the door to exciting further research.
LeJeune, J., Gautier, M. & Turpin, R. C.R. Hebd. Séanc. Acad. Sci. 248, 602–603 (1959).
Letourneau, A. et al. Nature 508, 345–350 (2014).
Dahoun, S. et al. Am. J. Med. Genet. A 146A, 2086–2093 (2008).
Guelen, L. et al. Nature 453, 948–951 (2008).
Hiratani, I. et al. PLoS Biol. 6, e245 (2008).
Davisson, M. T. et al. Prog. Clin. Biol. Res. 384, 117–133 (1993).
Takahashi, K. & Yamanaka, S. Cell 126, 663–676 (2006).
Li, B., Carey, M. & Workman, J. L. Cell 128, 707–719 (2007).
Catez, F., Brown, D. T., Misteli, T. & Bustin, M. EMBO Rep. 3, 760–766 (2002).
Liu, X., Wu, B., Szary J., Kofoed, E. M. & Schaufele, F. J. Biol. Chem. 282, 20868–20876 (2007).
Hassold, T. & Hunt, P. Nature Rev. Genet. 2, 280–291 (2001).
About this article
Transcriptional Aneuploidy Responses of Brassica rapa-oleracea Monosomic Alien Addition Lines (MAALs) Derived From Natural Allopolyploid B. napus
Frontiers in Genetics (2019)
Physiological Reviews (2018)
Trisomy of a Down Syndrome Critical Region Globally Amplifies Transcription via HMGN1 Overexpression
Cell Reports (2018)
Nature Reviews Genetics (2017)
Role of DNA Methylation on the Expression of the Anthracycline Metabolizing Enzyme AKR7A2 in Human Heart
Cardiovascular Toxicology (2016)