NATURE|Vol 442|10 August 2006 NEWS & VIEWS

641

many of the features associated with Hunting-ton’s disease in humans: cognitive and move-ment deficits, degeneration of the striatalregion of the brain, increased sensitivity toexcitotoxicity (the toxic effects on neurons ofoveractivation) and — significantly — cleav-age of Htt with subsequent movement of theprotein fragments into the cell nucleus.

Building on evidence that Htt can be digestedby caspase-3 and caspase-6, and that thesecleavages might be linked to toxicity5, Grahamet al.2 generated YAC128 mice that expressvariations of the polyQ-mutant Htt in whicheach of the caspase target sites is disrupted byan amino-acid substitution. This yielded miceexpressing Htt with the disease-causing polyQexpansion but resistant to cleavage by cas-pase-6 (which has one possible cleavage site)or caspase-3 (two possible cleavage sites). Micethat entirely lack Htt die before birth, butexpression of the caspase-resistant versions ofmutant Htt allowed the animals to survive. So,a crucial biological function of Htt, and byinference the overall conformation of Htt, wasnot altered by the caspase-resistant mutations.

The major finding by Graham et al.2 wasthat, in contrast to mice expressing YAC128Htt or the two forms of caspase-3-resistantHtt, mice expressing caspase-6-resistant Httfailed to develop any Huntington’s-like symp-toms. In particular, these mice showed nosigns of abnormal motor function, neuro-degeneration, movement of Htt to the nucleusof striatal cells, or enhanced excitotoxicity.

Galvan et al.1 and Graham et al.2 demon-strate that, in APP and Htt, specific aspartateresidues that are the targets of caspases have central roles in the development ofAlzheimer’s and Huntington’s disease, respec-tively. Caspases are increasingly being linkedto a wide variety of biological processes, butthey are best known for being active duringprogrammed cell death. This activation islinked to the cell’s response to stress. If the

occulting bodies are about as distant as Pluto,the authors estimate that the total number ofbodies of between 10 and 100 metres across inthe outer Solar System is more than 1015 — aquadrillion (Box 1).

That number suggests that there is a 20-metre-sized planetesimal in every sphere of100,000 kilometres’ radius in the outer SolarSystem. Astronomically speaking, that isextreme overpopulation. The debris fieldresembles a crowded dance floor where,inevitably, frequent collisions occur. These col-lisions break up the bodies, grinding themdown to smaller objects and eventually to dustparticles. Through the interplay of solar radia-tion pressure and gravitational forces, theseparticles are slowly removed from the outerSolar System, with some drifting out intospace and others spiralling into the Sun. Com-puter models indicate that there should cur-rently be between a billion (109) and a trillion(1012) bodies of 10–100 metres in size6, far lessthan Chang and colleagues’ number1. This dis-crepancy implies that collisions are less fre-quent than thought. As faster-moving objectscollide more frequently, these small bodies arelikely to be moving slowly relative to the largerplanetesimals in the disk.

The occultation technique has an inherentproblem: a dip does not establish the exactgeometry of an event, as a small body nearbywould produce a similar dip to that producedby a large object farther away. Fortunately, onecan establish the distance to an occulting bodyindependently of its size. This is done byobserving the tiny modifications in the shapeof a dip as the light beam from the backgroundsource diffracts over the surface of the occult-ing body7.

But despite the high X-ray luminosity of theScorpius X-1 source, it was not bright enoughto allow a study of these expected smallchanges. A dedicated space-based monitoringmission, optimized for precision stellar occul-tation measurements at millisecond timeintervals, could search for diffraction patternsto establish the exact extent of the debris field8.Whipple, a mission currently under consider-ation by NASA, would be able to monitoroccultations due to small bodies as far away asthe hypothesized Oort cloud at the extremeedge of the Solar System.

The Taiwan–America Occultation Survey9

currently operates four robotic optical tele-scopes in central Taiwan that monitor opticallight from stars at 0.2-second intervals. It willsoon make a census of the number of kilo-metre-sized bodies on the outer edges of theSolar System. The cumulative thermal ‘black-body’ radiation emission from planetesimalscould also soon be revealed in exquisite far-infrared images from the European SpaceAgency’s Herschel and Planck missions, which are scheduled for launch in 2008. Meanwhile, theorists must scrutinize theirmodels to explain the dense debris field at thefringes. The reason for its denseness may hide

NEURODEGENERATIVE DISEASE

Cut to the chaseLisa M. Ellerby and Harry T. Orr

A family of enzymes called caspases — best known for their involvement in programmed cell death — now seems to be pivotal in the progression of two neurodegenerative diseases.

Caspases are protease enzymes that cut proteins at specific target amino-acid residues.By blocking caspase action on crucial proteinsassociated with Alzheimer’s or Huntington’sdisease, respectively, Galvan et al.1 and Grahamet al.2 show that the development of symptomsis also blocked in mouse models of the dis-eases. Although these studies are straightfor-ward in concept, they are each an experimentaltour de force, providing compelling in vivoevidence that caspases have a prominent role inthese devastating conditions.

One of the main characteristics ofAlzheimer’s disease is the build-up in the brainof ‘amyloid plaques’ made of clumps of amy-loid precursor protein (APP). The amyloidhypothesis posits that a peptide fragment ofAPP called A�42 initiates a cascade of eventsthat leads to full-blown Alzheimer’s disease. However, APP can be cleaved by caspase at a different site, aspartate amino acid 664(D664), releasing another toxic peptide, calledAPP-C31 (refs 3, 4). That suggests that there is more to the story.

Galvan et al.1 used a mouse model ofAlzheimer’s disease in which the mice expressa mutant form of human APP. The authorsthen mutated the APP further by changing theD664 site to an alanine (D664A). They showconvincingly that the mice with the D664Amutation no longer make the APP-C31 pep-tide. Strikingly, the mice still produce A�42and amyloid deposits, but fail to develop thepathological, physiological and behaviouraldeficits characteristic of Alzheimer’s disease.

By contrast, Graham et al.2 assess the role ofcaspases in Huntington’s disease. This is aninherited disorder associated with a mutatedform of huntingtin protein (Htt) that has anexpanded stretch of glutamate residues (thepolyQ tract). The group had previously genet-ically engineered a strain of mice expressingmutant Htt with an expanded polyQ tract(called YAC128 mice). These mice develop

2. Edgeworth, K. E. J. Br. Astron. Assoc. 53, 181–188 (1943).3. Luu, J. X. & Jewitt, D. C. Annu. Rev. Astron. Astrophys. 40,

63–101 (2002).4. Elliot, J. L. & Olkin, C. B. Annu. Rev. Earth Planet. Sci. 24,

89–124 (1996).5. Bailey, M. E. Nature 259, 290–291 (1976).6. Kenyon, S. J. & Bromely, B. C. Astron. J. 128, 1916–1926 (2004).7. Roques, F., Moncuquet, M. & Sicardy, B. Astron. J. 93,

1549–1558 (1987).8. Cooray, A. Astrophys. J. 589, L97–L101 (2003).9. http://taos.asiaa.sinica.edu.tw

important clues to the mass and extent of the primordial gas disk from which the SolarSystem formed. ■

Asantha Cooray is in the Department of Physicsand Astronomy, University of California, Irvine, California 92697, USA.e-mail: acooray@uci.edu

1. Chang, H.-K. et al. Nature 442, 660–663 (2006).

10.8 News & Views MH 4/8/06 5:19 PM Page 641

Nature Publishing Group ©2006

642

50 YEARS AGO“European Brewery Convention”— The idea of science infiltratinginto brewing still frequentlyproduces a reaction from thelayman on the grounds that, first,beer has been brewed for morethan six thousand years, so therecan be little still to learn; second,the less the scientist has to dowith it the better because it usedto be better than it is ( in fact, it is even hinted that nowadays“beer is made from chemicals”);third (triumphantly), brewing is an art and not a science. The first two points are born of ignorance of the real position, and the truth of thethird depends on the sense of the word ‘art’: in so far as it refers to experience and skill in ‘know how’ it may becorrect.From Nature 11 August 1956.

100 YEARS AGOPoverty and Hereditary Genius; aCriticism of Mr. Francis Galton’sTheory of Hereditary Genius —The criticism which Mr. Constable brings forward in this book is that reputation is not a test of ability, and asGalton’s theory of hereditarygenius is based on thisassumption, it has to bediscarded. The statisticalevidence given in “HereditaryGenius” has to be explainedaway, and Mr. Constableattempts to do this by what he calls the “swamping effect of poverty.” We quite agree with Mr. Constable that it isharder for a poor man withuninfluential parents to achievesuccess as a judge than for a rich one with influence, but thisdoes not seem to us to justify Mr. Constable in discarding theconclusions of “HereditaryGenius,” for if the socialconditions of both parents and offspring are relatively about the same, it seems as if the omission of the ability inpoverty-stricken parents andtheir children is rather likeleaving out of account theaddition of numbers to both the numerator and denominatorof a fraction.From Nature 9 August 1906.50 &

100

YEA

RS A

GO

NEWS & VIEWS NATURE|Vol 442|10 August 2006

aspartate mutations generated by Galvan et al.and Graham et al. are indeed having theireffect on the neurodegenerative diseases bypreventing caspase action, this implies thatcleavage of APP and Htt is downstream ofsome event that activates a cellular response tostress; that is, caspase cleavage and the subse-quent pathological events are components ofdisease progression.

It is notable that the D664A mutation inAPP did not affect the generation of A�42 oramyloid deposition. Perhaps the generation ofA�42 induces a stress signal that results in cas-pase activation and its pathological conse-quences. Similarly, caspase-6 cleavage of Httmight also be in response to some stress signal,presumably polyQ-mutant Htt. To understandHuntington’s disease fully, this initiating eventwill need to be identified. Alternatively, cas-pase-6 might not be the initiating protease forHuntington’s disease in vivo, in which caseidentification of this crucial protease would be of paramount concern. It will therefore beessential to examine whether manipulatingcaspase-6 activity in the brain alters diseaseprogression in YAC128 mice. Interestingly,lowering caspase-6 activity in cell-culturemodels of Huntington’s disease does protectneurons from degeneration6.

Caspase-6 has a unique set of substratespecificities that does not overlap with those ofother caspases. So selective inhibitors of cas-pase-6 might block the symptoms of Hunting-ton’s disease. Other caspases also interact withAPP and Htt, suggesting that blocking suchinteractions might be beneficial therapeuti-cally4,6. Specific caspase inhibitors are beingdeveloped by pharmaceutical companies, butmuch work needs to be done before we knowwhether they are suitable for clinical use. Thestudies of Graham et al.1 and Galvan et al.2 do,however, validate caspases in vivo as potentialtherapeutic targets for Alzheimer’s and Hunt-ington’s disease. ■

Lisa M. Ellerby is at the Buck Institute for AgeResearch, 8001 Redwood Boulevard, Novato,California 94947, USA. Harry T. Orr is at theInstitute of Human Genetics, University ofMinnesota, 516 Delaware Street SE, Minneapolis,Minnesota 55455-0374, USA. e-mail: orrxx002@.umn.edu

1. Galvan, V. et al. Proc. Natl Acad. Sci. USA 103, 7130–7135(2006).

2. Graham, R. K. et al. Cell 125, 1179–1191 (2006).3. Gervais, F. G. et al. Cell 97, 395–406 (2000).4. Lu, D. C. et al. Nature Med. 6, 397–404 (2000).5. Wellington, C. L. et al. J. Biol. Chem. 275, 19831–19838

(2000).6. Hermel, E. et al. Cell Death Differ. 11, 424–438 (2004).

MICROSCOPY

Nanotomography comes of ageDavid Attwood

The use of X-rays to construct three-dimensional tomographic images iswell established in medicine. The same principle is being extended to thenanoscale, bringing us startlingly accurate pictures of tiny objects.

Writing in Applied Physics Letters, Yin and col-leagues1 report an X-ray microscopy techniqueof broad potential for three-dimensional imag-ing in the physical and life sciences. By tuninghigh-energy X-rays, the authors manipulatethe contributions of specific chemical elementsto a series of two-dimensional images. Theythen use tomographic methods to combineimages taken at different incident X-ray angles,allowing internal structures and — given suffi-cient spectral resolution — chemical bondingsto be discerned with a spatial resolution ofaround 60 nanometres.

In essence, this technique is a nanometre-scale version of medical computed tomo-graphic (CT) imaging of humans. The highspatial resolution of the new system1 is largelydetermined by the 50-nm width of the outer-most transmitting zone of its zone plate lens.This lens is a circular diffraction grating con-sisting of alternate transparent and opaqueconcentric rings2,3 and is used to focus the X-ray photons. Higher doses of radiation arerequired for nanoscale imaging, so radiation-

sensitive samples such as biological tissue canrequire cryogenic or other ‘fixation’ techniquesthat limit structural damage to them.

For their experiments, the authors use pho-tons at wavelengths of between 0.11 and 0.15nanometres; as a photon’s wavelength isinversely proportional to its energy, this isequivalent to photon energies of between 11and 8 kiloelectronvolts. Such high-energy X-rays are known as hard X-rays. At the ‘absorp-tion edges’ of a chemical element, photonshave enough energy to lift an electron out of aparticular atomic orbital, and can therefore be absorbed. Thus, by tuning the X-ray energyto just above or below a prominent absorp-tion edge, the contributions of specific chemi-cal elements to the image can be enhanced or diminished.

Yin et al. tested their set-up on a thinnedportion of a silicon computer chip containingcopper–aluminium wires and an array oftungsten plugs. These plugs are used as electri-cal interconnects between the typically eightor nine layers of a computer chip. Such plugs

10.8 News & Views MH 4/8/06 5:19 PM Page 642

Nature Publishing Group ©2006