Sci. Aging Knowl. Environ., 27 October 2004
Vol. 2004, Issue 43, p. ns8
[DOI: 10.1126/sageke.2004.43.ns8]


The Goldilocks Genes

Our chromosomes can get sloppy and delete or pick up genes. Researchers are starting to probe how these DNA miscues affect health and longevity

Mitch Leslie

Abstract: From cancer biologists to geneticists, scientists are tackling the question of how raising or lowering the number of gene copies from the standard two affects how we live and how soon we die. New genomic studies show that large-scale DNA duplications and deletions are more common than scientists imagined, even in healthy people. Other work demonstrates that our supply of gene copies matters in Parkinson's disease and cancer. The results raise the possibility that how fast we age depends in part on how many duplicates of certain genes we carry.

Football fans don't usually load up on seven-packs of beer. Krispy Kreme customers rarely opt for only 11 doughnuts. And genes are supposed to come in pairs. However, researchers are learning that missing and extra gene copies are more prevalent than they thought. When it comes to the number of genes they carry, some cells are as fussy as Goldilocks: Raising or lowering the count can elicit trouble. These gene excesses and deficiencies can impair health--some forms of Parkinson's disease (PD) stem from an oversupply of a particular gene, for example, and Alzheimer's disease (AD) might too. The notion that the number of gene copies influences disease and life span gained support last summer when two studies uncovered a profusion of unexpected duplications and deletions in the genome. Those findings indicate that most people are not equal, at least when it comes to how many gene copies they have.

Here a Copy, There a Copy Back to Top

The physical disparities between nebbish filmmaker Woody Allen and 340-pound basketball star Shaquille O'Neal are immense. But the genetic differences between any two people probably boil down to minute genetic alterations--or so researchers thought until last summer. Two broad surveys of the genome revealed that our chromosomes are littered with large duplicated segments and display telltale signs of multiple deletions. Shaq and Woody--and any other pair of humans--probably carry unequal numbers of these DNA segments. Researchers had long known that repeated regions of noncoding DNA abound in the genome, but these studies showed that our chromosomes also sport multiple copies of protein-encoding stretches.

Cytogeneticist Charles Lee of Harvard Medical School in Boston led one team, and molecular geneticist Michael Wigler of Cold Spring Harbor Laboratory in New York headed the other. The groups used different types of microarrays to assess genetic diversity across the genome but reached a similar conclusion: On average, any two people carry different numbers of about 11 or 12 particular DNA chunks. Because each stretch can be large, these variations could translate into unequal numbers of many genes. For example, Lee and colleagues tested for the gene that encodes amylase, a starch-splitting enzyme pumped out by the pancreas. The 55 subjects in their study harbored from six to 14 copies of the gene.

In Wigler's analysis, the 20 subjects carried different numbers of 70 genes. Researchers had ascribed most genetic differences among humans to single-nucleotide polymorphisms, one-letter changes in the DNA code, Lee says. If you think of the genome as a book, "we've been looking for spelling mistakes, but we've missed whole chapters," he says.

How chromosomes add or shed so many segments isn't clear, but one possibility is that DNA repair mechanisms blunder. What is clear from comparisons of human and rodent DNA, says Wigler, is that we sport more repeated stretches than do our beady-eyed distant cousins. By boosting genetic adaptability, the profusion of duplications probably fostered rapid evolution in our species and explains "why we have ascended to the top of the food chain," says Wigler. But the price of this flexibility could be genetic diseases. Circumstantial evidence supports that suggestion. Wigler's group found that some of the duplicated and deleted sections encompass genes that are linked to a rare developmental disorder called Cohen syndrome, to cancer, and to weight control--although all of the subjects were hale.

Built to Break Back to Top

However, some duplications clearly do provoke diseases, as geneticist James Lupski of Baylor College of Medicine in Houston, Texas, and colleagues have discerned. Their work reveals that many patients with Charcot-Marie-Tooth (CMT) disease--in which nerves in the legs, hands, and feet deteriorate--show a duplication of one region on chromosome 17. The doubled segment includes not only the CMT-causing gene but 24 others. A particular feature of the genome's structure seems to promote duplications along with deletions, Lupski says. The culprit area is rich in so-called low-copy repeats, identical or nearly identical DNA snippets large enough to hold several genes. These repeats can cause confusion when cells try to mend broken DNA.

That confusion arises when cells fix sundered strands through homologous recombination. Cells deploying homologous recombination for repair seek to restore the broken strand. So they usually swap DNA from a sister chromatid, an exact copy of the chromosome that the cell makes as it prepares to divide, and patch the gap using the intact strand as a blueprint.

Normally, the enzymatic repair crew scans the sister chromatid for a sequence that matches the broken section, using the low-copy repeats as landmarks. For example, to fix a break at point A (see figure), the enzymes would seek a match near the low-copy repeat shown in pink. However, the presence of numerous similar sequences can confuse the enzymes. Instead of weaving in DNA from a matching gene, they might select a sequence from a different part of the sister chromatid. The blunder can extend one chromatid and truncate the other. This aberrant homologous recombination could potentially produce gene duplications and deletions such as those that Lee and Wigler detected.

View larger version (24K):
[in this window]
[in a new window]
Unequal exchange. Homologous recombination can fix damaged DNA by swapping a matching section of an identical chromosome (upper panel). The letters designate genes, and the colored boxes indicate low-copy repeats. The exchange of uppercase "A" for lowercase "a" indicates that the strands have traded DNA. But sometimes enzymes look for a matching sequence on the wrong part of the chromosome, leading to shortened or lengthened strands (lower panel). [Source: James Lupski]

The genetic error that prompts CMT disease occurs during sperm and egg formation, Lupski says, and the mistake passes to the next generation. But mouse studies suggest that similar bungled DNA swaps might also occur when nonreproductive cells copy their DNA and divide, he adds. If that's true in humans, cells in the brain or muscles could accumulate or eject genes during a person's lifetime. So far, nobody has identified a patient who became ill because of such a change, Lupski notes. But clinical investigators rarely compare different tissues to determine whether only some show chromosome alterations, he says.

Quadruple Trouble Back to Top

The first study to implicate extra gene copies in a disease of aging came from neuroscientist John Hardy of the National Institute on Aging in Bethesda, Maryland, and colleagues. They were puzzling over a family whose members began to suffer PD symptoms in their mid-30s. Defective versions of the protein {alpha}-synuclein can cause rare types of inherited PD that begin early in life, but the affected individuals all bore normal versions of the gene. In a study published last fall in Science, Hardy's team showed that these unlucky folks sported four copies of the {alpha}-synuclein gene instead of the usual two. The extras rode on a slab of chromosome 17 that had triplicated sometime in the family's history; the researchers aren't sure how this molecular Xeroxing occurred.

The connection between gene number and brain deterioration suggests an explanation for how AD and other conditions provoke illness, Hardy says. Take AD, for example. Enzymes in the brain snip apart {beta}-amyloid precursor protein, APP, to yield {beta} amyloid, the glutinous protein that piles up in the brain (see "Detangling Alzheimer's Disease"). AD patients occasionally display errors in genes for APP or for the presenilins, proteins that form part of the molecular shears that trim APP (see Wolfe Perspective). These mutations hike production of a sticky form of {beta} amyloid that is more likely to clump. However, most AD patients--those who typically develop symptoms after age 65--show neither genetic abnormality. But they might be pumping out excess amounts of APP, Hardy speculates--either because they're endowed with extra copies of the gene or because their copies are overactive.

For decades, Hardy says, researchers have been staring at powerful evidence that backs the proposal: the high rate and early onset of AD among Down syndrome patients. Their cells tote an extra copy of chromosome 21, which harbors the gene for APP, and their AD symptoms typically begin before they reach age 50, about 20 years earlier than usual. But a few Down syndrome patients carry a pruned version of the chromosome that's missing the APP gene, and they don't get AD. Testing the hypothesis by measuring protein output is challenging, Hardy says, because researchers need ample brain tissue from patients.

However, researchers have tackled the question using mice engineered to mimic Down syndrome. The rodents carry an artificial chromosome with 104 genes that reside on the human chromosome 21, so they have extra copies of all these genes, including APP. Geneticist Roger Reeves of John Hopkins University School of Medicine in Baltimore, Maryland, and colleagues scrutinized the rodents' brains but found no evidence of plaques, even in 3-year-old animals. "That's quite old for a mouse," Reeves says. He adds that the lack of protein buildup in these rodents doesn't disprove the hypothesis that extra copies of APP spur AD. The animals might not live long enough for plaques to condense, he says.

Lose a Gene, Gain a Tumor? Back to Top

One could be more than a lonely number for some genes. Having only a single working copy of a gene, a situation known as haploinsufficiency, can provoke problems, particularly if the gene encodes a tumor suppressor, a protein that reins in cell growth and checks cancer. For about 30 years, cancer experts didn't fret about this possibility because they adhered to the "two-hit model," in which both copies of the tumor suppressor must fail in order to unleash rampant cell division. But researchers are amassing evidence that disabling even one copy can promote tumors, says cancer biologist Frederick Alt of Harvard Medical School.

The idea that losing one gene copy might drive a cell closer to cancer upsets even audiences of scientists, Alt says: "It sounds very scary, but it's not just a matter of flipping a switch." Other genes must also malfunction before cell division goes awry, he says (see "Dangerous Liaisons").

His team's work on the H2AX histone, one of the protein spools around which DNA wraps to form chromosomes, illustrates how subtle changes in gene number might precipitate problems. Mice bearing one copy of the H2AX gene seem healthy. But tumors proliferate if the animals also lack the protein p53, which normally goads cells with broken DNA to either make repairs or kill themselves. These haploinsufficient animals live only about half as long as rodents fully endowed with H2AX and also missing p53. The haploinsufficient mice's susceptibility to abnormal growths suggests that deleting one copy of the H2AX gene knocks down one of the cell's usual barriers to cancer. The mice don't have as many of the repair-promoting histone molecules per length of DNA as do normal animals, and Alt and colleagues hypothesize that this deficiency underlies their vulnerability to cancer. H2AX is vital for refurbishing broken DNA strands (see figure). The researchers conjecture that the proteins hold out the ragged ends and serve as a landing site and assembly area for a squad of fix-it proteins, which grab the frayed DNA and knit it together (see figure, right). Without enough of the DNA-holding histones, some fractures don't get mended. "If you don't repair all your damage, you are putting yourself at risk" for cancer, Alt says.

View larger version (25K):
[in this window]
[in a new window]
The enzyme has landed. Undamaged DNA coils around histones (pale blue cylinders, left). After the strands break (indicated by DSB), repair enzymes hustle to the scene (right). The reddish barrels are H2AX histones that are holding out the severed DNA strands. The enzyme conglomeration, denoted by the green and dark blue cylinders and the orange "tongs," has grabbed the DNA and is ready to stitch it back together. [Reprinted with permission from Landes Bioscience. C. H. Bassing, F. W. Alt, Cell Cycle 3, 149-153 (2004), figure 1, p. 151]

The H2AX situation is just one instance in which knocking out a gene promotes cancer in rodents. However, finding evidence to implicate the mechanism in humans is tougher, Alt notes. By the time a tumor has grown large enough for diagnosis, the cells often show chromosomal chaos: Some DNA segments have duplicated, some have shifted position, and some have disappeared. These changes could obliterate signs that gene loss helped launch the tumor, Alt says. In rodents, by contrast, researchers know which gene they disabled, so the chromosomal scrambling isn't an obstacle to understanding what's going on. Even if researchers can identify a lost gene in a human tumor, the effects on cell growth are likely to be subtle and hard to discern, says Wigler. However, for a few genes, such as BRCA1, which is associated with breast cancer, a single missing copy dramatically accelerates human tumor growth.

Measuring the Dose Back to Top

Nobody has systematically tallied genes for which the number of copies matters--called dosage dependent--so nobody knows how many we carry. "That's the million-dollar question," says Lupski. But some likely candidates have turned out not to be dosage-dependent for the characteristics that matter most. For example, molecular biologist Arlan Richardson of the University of Texas Health Sciences Center in San Antonio and colleagues gauged the effect on mice of deleting one copy of the gene for MnSOD, an enzyme that quenches reactive oxygen species spewed by metabolism (see "Is Less Enough?"). The mice accumulated excessive oxidative damage, as expected, and endured more tumors than do normal rodents. But they lived just as long.

However, recent work on Down syndrome mice suggests that, for chromosome 21 at least, the count of dosage-dependent genes is higher than many scientists thought. Most patients with Down syndrome have a third copy of chromosome 21, but a few patients carry only an extra fragment. By comparing these pieces, some researchers identified a shared region that, they argued, held the dosage-dependent genes responsible for many of the symptoms of Down syndrome. But in a study published last week in Science, Reeves and colleagues showed that mice with surplus copies of the 33 genes from this portion of chromosome 21 don't show the telltale facial deformities of Down syndrome. According to previous studies, these abnormalities do occur in animals with 104 genes from the chromosome. The results suggest that "there aren't just one or two important genes" for Down syndrome, Reeves says.

The work doesn't reveal exactly how many genes contribute to Down syndrome, never mind the total number of dosage-dependent genes in the genome. But Lupski has made a rough estimate from his results. Based on the occurrence of uncommon genetic diseases that produce discernible DNA gains or losses, the value could be as low as 1 in 100 genes, he says. On the other hand, the duplicated stretch he fingered on chromosome 17 harbors 25 genes, but only the one for CMT disease displays an obvious dosage effect, which gives a figure of 1 in 25. If the lower value is correct, it translates to about 300 dosage-dependent genes strewn around the genome. But if the higher number is closer to the actual figure, we could have some 1200 of these genes, which might provide more targets for diseases.

Lost Genes Equal Lost Life? Back to Top

Whether those 300 to 1200 genes influence longevity is a question researchers are just beginning to probe. One rare inherited disease known as dyskeratosis congenita suggests how gene dose might alter symptoms associated with aging. In one form of the disease, patients carry a single working copy of the gene that encodes part of telomerase, the enzyme that prevents the tips of chromosomes from eroding due to cell division. Their telomeres wear down in fast-dividing tissues such as the bone marrow, and many patients suffer from cancer or become anemic because they can't create enough blood cells. Researchers long ago proposed that the gradual whittling away of telomeres might contribute to aging (see "More Than a Sum of Our Cells"), a hypothesis that remains controversial. Although dyskeratosis congenita is not normally age-related, it shows how telomere shrinkage can accelerate physical deterioration. Researchers are also intrigued by the possibility that the disease could be age-related under rare circumstances. A parent with the disease might bequeath to his or her children a functional telomerase gene and marginally short telomeres that could erode unusually quickly. No one has yet documented this phenomenon, but the children thus might develop dyskeratosis congenita although they make normal amounts of telomerase.

Dyskeratosis congenita is rare, but more common variations in gene copy number could also help dictate life span, scientists hypothesize. A person who has only one working copy of a gene might survive decades without any sign of sickness, says pathologist George Martin of the University of Washington, Seattle (editor-in-chief of SAGE KE). But protein production often drops off with age, and cells don't recycle faulty proteins promptly (see Cuervo Perspective). Proteins that sit around can accumulate damage from reactive oxygen species or advanced glycation end products, which result when sugars glom onto amino acids (see "Aging Research Grows Up"). They can also undergo other chemical changes that impede their function. That development could leave haploinsufficient people short on key proteins or force them to rely on faulty ones. People with two copies of the gene could compensate to some degree, and they might not age as fast, says Martin. "There must be a reason why you have two copies [of most genes]," he says.

Haploinsufficiency could also leave stem cells vulnerable to mutations, Wigler notes. A person with two copies of a gene might be able to sacrifice one to mutation without suffering ill effects. But somebody born haploinsufficient can't afford to give up even one version. The body needs stem cells to replace cells that wear out and die, but mutations in haploinsufficient people might gradually deplete the stockpiles of these regenerating cells, thus speeding aging, says Wigler. He predicts that long-lived folks will carry more duplicates of key dosage-dependent genes than do short-lived people. Researchers have yet to test this hypothesis.

Aging could pummel more than genes in the nucleus. Time's passing might spur changes in numbers of mitochondrial genes, says endocrinologist K. Sreekumaran Nair of the Mayo Clinic in Rochester, Minnesota. In a 2000 study, Nair and colleagues showed that as rats age, mitochondria in the muscles and liver shed up to 50% of their copies of certain genes. However, the heart held tight to its mitochondrial genes. The explanation for the difference might lie in how hard the tissues work, Nair says. The heart labors continuously throughout life. Its unceasing activity might spur the mitochondria to duplicate their DNA and replace damaged strands, he hypothesizes. Whether humans show similar gene losses in their cellular powerhouses isn't clear. Researchers in Australia who scrutinized mitochondria from the muscles and heart found no changes in mitochondrial gene copy number with age.

Regardless of whether scientists substantiate this possible connection to aging, knowledge about how altered numbers of gene copies can disrupt cells could soon be helping patients with inherited DNA duplications. For example, it might be possible to treat CMT patients with small interfering RNA, snippets of nucleic acid that impede synthesis of the troublemaking protein. Similar advances might combat other dosage-related diseases and help cells keep their number of genes just right.

October 27, 2004

Mitchell Leslie writes about science from Portland, Oregon, where he hopes that the amount of rainfall this winter won't be too much or too little.

  1. R. Barazzoni, K. R. Short, K. S. Nair, Effects of aging on mitochondrial DNA copy number and cytochrome c oxidase gene expression in rat skeletal muscle, liver, and heart. J. Biol. Chem. 275, 3343-3347 (2000). [Abstract/Free Full Text]
  2. C. H. Bassing and F. W. Alt, H2AX may function as an anchor to hold broken chromosomal DNA ends in close proximity. Cell Cycle 3, 149-153 (2004). [Medline]
  3. C. H. Bassing and F. W. Alt, The cellular response to general and programmed DNA double strand breaks. DNA Repair 3, 781-796 (2004). [Medline]
  4. C. H. Bassing et al., Histone H2AX: a dosage-dependent tumor suppressor of oncogenic translocations and tumors. Cell 114, 359-370 (2003). [CrossRef][Medline]
  5. A. J. Iafrate et al., Detection of large-scale variation in the human genome. Nat. Genet. 9, 949-951 (2004).
  6. P. J. Mason and M. Bessler, Heterozygous telomerase deficiency in mouse and man. Cell Cycle 3, e39-e41 (2004).
  7. P. J. Miller et al., Precise determination of mitochondrial DNA copy number in human skeletal and cardiac muscle by a PCR-based assay: lack of change of copy number with age. Nucleic Acids Res. 31, e61 (2003). [Abstract/Free Full Text]
  8. J. R. Mitchell, E. Wood, K. Collins, A telomerase component is defective in the human disease dyskeratosis congenital. Nature 402, 551-555 (1999). [CrossRef][Medline]
  9. M. Morey, F. Serras, M. Corominas, Halving the selenophosphate synthetase gene dose confers hypersensitivity to oxidative stress in Drosophila melanogaster. FEBS Lett. 534, 111-114 (2003). [CrossRef][Medline]
  10. L. E. Olson, J. T. Richtsmeier, J. Leszl, R. H. Reeves, A chromosome 21 critical region does not cause specific Down syndrome phenotypes. Science 306, 687-690 (2004). [Abstract/Free Full Text]
  11. M. Santarosa and A. Ashworth, Haploinsufficiency for tumor suppressor genes: when you don't need to go all the way. Biochim. Biophys. Acta 1654, 105-122 (2004). [Medline]
  12. J. Sebat et al., Large-scale copy number polymorphism in the human genome. Science 305, 525-528 (2004). [Abstract/Free Full Text]
  13. A. Singleton, A. Myers, J. Hardy, The law of mass action applied to neurodegenerative disease: a hypothesis concerning the etiology and pathogenesis of complex diseases. Hum. Molec. Genet. 13, R123-R126 (2004). [Abstract/Free Full Text]
  14. A. Singleton et al., Association between cardiac denervation and parkinsonism caused by {alpha}-synuclein gene triplication. Brain 127, 768-772 (2003).
  15. A. B. Singleton et al., {alpha}-Synuclein locus triplication causes Parkinson's disease. Science 302, 841 (2003). [Free Full Text]
  16. P. Stankiewicz and J. R. Lupski, Molecular-evolutionary mechanisms for genomic disorders. Curr. Opin. Gen. Dev. 12, 312-319 (2002). [CrossRef][Medline]
  17. P. Stankiewicz and J. R. Lupski, Genome architecture, rearrangements and genomic disorders. Trends Genet. 18, 74-82 (2002). [CrossRef][Medline]
  18. P. Stankiewicz et al., Genome architecture catalyzes nonrecurrent chromosomal rearrangements. Am. J. Hum. Genet. 72, 1101-1116 (2003). [CrossRef][Medline]
  19. E. Thomas et al., Distribution of short paired duplications in mammalian genomes. Proc. Natl. Acad. Sci. U.S.A. 101, 10349-10354 (2004). [Abstract/Free Full Text]
  20. H. Van Remmen et al., Life-long reduction in MnSOD activity results in increased DNA damage and higher incidence of cancer but does not accelerate aging. Physiol. Genomics 16, 29-37 (2003). [Abstract/Free Full Text]
  21. J. M. Y. Wong, M. J. Kyasa, L. Hutchins, K. Collins, Telomerase RNA deficiency in peripheral blood mononuclear cells in X-linked dyskeratosis congenita. Hum. Genet. 115, 448-455 (2004). [CrossRef][Medline]
Citation: M. Leslie, The Goldilocks Genes. Sci. Aging Knowl. Environ. 2004 (43), ns8 (2004).

Science of Aging Knowledge Environment. ISSN 1539-6150