news & views

nature genetics • volume 29 • october 2001 103

On pages 160–165 and 166–173of this issue, Yi Yang and col-leagues1 and Shinji Hadano andcolleagues2 reveal the identity ofa gene that, when mutated, isresponsible for a rare autosomalrecessive variant of amyotrophiclateral sclerosis (ALS). This dis-order, called ‘juvenile ALS,’ isprevalent in populations ofNorth Africa and the MiddleEast, and was previously linkedto chromosome 2q33 (ref. 3).This is an important advance forthe field of ALS research—onethat may have a broad impact onstudies of neurodegeneration.

The clinical pictureTo understand the implicationsof this work, juvenile ALS needsto be placed in context. The‘usual’ form of ALS, whichcaused the deaths of Lou Gehrigand David Niven, is a lethal para-lytic disorder causing dysfunc-tion and death of the majorneuronal components of themotor system—the upper motorneurons (UMN) in the motorcortex and lower motor neurons(LMN) of the brainstem andspinal cord (Fig. 1). This is usually a diseaseof middle and later life, with an averagesurvival of less than 3 years, and is respon-sible for the death of approximately 1 in1,000 adults, most often from respiratoryfailure. Apparently sporadic in 90% ofcases, it may also present as a familial, usu-ally autosomal dominant, disease. JuvenileALS (ALS2) is a rarely encountered variantthat starts before the age of 25 and tends toprogress more slowly than classical ALS4.

The clinical picture of ALS variesdepending on the anatomical ‘center ofgravity’ of the disease process. ClassicalALS produces a combination of UMN(predominantly spasticity) and LMN(muscle weakness and atrophy) features.

The site of onset is usually focal, but there isinexorable spread of the disease to involvemotor neurons innervating muscles of thelimbs and bulbar (speech and swallowing)regions. Some affected individuals mayhave features of an LMN disorder exclu-sively. They may then develop UMN signsand progress to ALS, but some do not, andtheir disease is termed ‘progressive muscu-lar atrophy’. A rare minority present withpure UMN signs and never develop LMNfeatures, and this group is said to have the‘primary lateral sclerosis’ (PLS) variant. It isnow widely accepted that these subgroupsrepresent part of a clinicopathologicalspectrum based on a common neurode-generative pathway.

The gene huntAlthough neurodegeneration inALS was first described by Char-cot in 1869 (ref. 5), its mecha-nism has remained an enigma.The first major breakthroughcame in 1993 when a consor-tium of researchers discoveredthe genetic basis of approxi-mately 20% of cases of autoso-mal dominant familial ALS6.This turned out to be a gene onchromosome 21 encodingCu/Zn superoxide dismutase(SOD1), a very well character-ized enzyme that protects cellsfrom damage by free radicals.Familial ALS due to SOD1 alter-ations includes at least 60 differ-ent point mutations throughoutall 5 exons, with a smattering ofother types of genetic lesion.The gene is ubiquitouslyexpressed, and yet the mutatedform causes only motor neurondegeneration. The mechanismby which mutations in SOD1bring about motor neuronalinjury is not yet fully clarified,but is considered to be a toxicgain of function rather than aloss of normal SOD1 activity.

Despite the explosion of ALS-relatedresearch engendered by this breakthroughand supported by the availability of a goodtransgenic mouse model of the diseaseand several cellular models, there are threemajor competing theories of SOD1 toxic-ity, including oxidative stress resultingfrom aberrant enzymatic specificity, cop-per toxicity, and abnormal protein aggre-gation7,8. Reassuringly for its relevance tosporadic ALS, pathological studies ofSOD1-related familial ALS show evidenceof shared molecular pathological features,although the extent of central nervous sys-tem involvement is often more extensivethan it is in sporadic cases9. There are afew other genetic factors that may predis-

Genetic inroads in familial ALSPamela J. Shaw

Academic Neurology Unit, Medical School, University of Sheffield, UK. e-mail: pamela.shaw@sheffield.ac.uk

Amyotrophic lateral sclerosis (ALS) is a common neurodegenerative disease causing cell death of motor neurons and progressivemuscle weakness. The disease is familial in ten percent of cases, of which one-fifth are due to mutations in the gene encodingCu/Zn superoxide dismutase (SOD1). Two papers in this issue of Nature Genetics describe homozygous mutations in a new geneon chromosome 2q33 in 4 families of Arabian origin with a rare form of juvenile onset ALS (ALS2). The predicted protein structurehas domains homologous to GTPase regulatory proteins, and both the types of mutation and the pattern of inheritance suggestthat motor neuron degeneration is the result of a loss of function. Further work will determine the relevance of this breakthroughto other, more common forms of ALS.

BOB CRIMI

The human motor system

brain

brainstem

spinalcord

upper motor neurons (damaged in PLSand ALS)

corticospinal tract fromupper motor neurons

lower motor neurons (damaged in ALS)

limb muscles andbreathing muscles

motor neurons controllingeye movements, facial muscles,speech and swallowing (damaged in ALS)

Fig. 1 Organization of the human motor system. In ALS, both upperand lower motor neurons are damaged. In PLS, only the upper motor neu-rons are damaged.

©20

01 N

atu

re P

ub

lish

ing

Gro

up

h

ttp

://g

enet

ics.

nat

ure

.co

m© 2001 Nature Publishing Group http://genetics.nature.com

news & views

104 nature genetics • volume 29 • october 2001

pose to sporadic ALS, including abnor-malities of the neurofilament heavy sub-unit, NFH10,11, apolipoprotein e4genotype12 and the genes linked to spinalmuscular atrophy (Fig. 2)13,14. Until now,however, no gene mutation has beenfound in the remaining 80% of familialALS cases, and the field has been keenlyawaiting a ‘new’ ALS gene, with all theattendant opportunities for elucidatingkey pathways for motor neuron injury.

We now have another gene, but, unlikeSOD1, it encodes a protein with functionsthat can only be predicted from its struc-ture. Previous work had localized thegene to a critical region of chromosme 2qin which there were no obvious candidategenes. Both groups therefore embarkedon a laborious screen of all the exons pre-sent in the many transcripts from theregion. Each identified mutations linkedto affected individuals with familial juve-nile ALS in a gene comprising 34 exons ina genomic region of 83 kb. The gene,ALS2, encodes a protein termed ‘alsin’ byone of the groups1. There is a full-lengthform of 184 kD comprising 1,657 aminoacids, as well as a smaller, alternativelyspliced transcript that is only 396 aminoacids long. Yang and colleagues1 studiedtwo unrelated families from Saudi Arabia

and Tunisia. The Tunisian family has asingle–base pair deletion from exon 3 inthose with an ALS phenotype. The Saudifamily has a PLS phenotype, with a 2-bpdeletion in exon 9. Hadano and col-leagues2 found similar results in two fam-ilies from Kuwait and Tunisia. Affectedmembers of the Tunisian family have aclinical phenotype of ALS associated witha 1-bp deletion in exon 3. In contrast, theKuwaiti family has a PLS phenotype and ahomozygous 2-bp deletion in exon 5.Both groups assume that the most likelymechanism for disease is loss of function.The genotype–phenotype correlation isinteresting and is postulated to reflectmore severe disease (ALS) in affectedindividuals with a deletion mutationaffecting both long and short transcripts.The milder phenotype (PLS) in affectedindividuals with deletions in exons 5 and9 may arise because of an intact short pro-tein or some preservation of function ofthe mutated longer protein.

Form and functionThe authors of both papers use sequencehomology data to predict functionaldomains. Overall, the protein may begrouped with GTPase regulatory pro-teins based on the presence of several

guanine-nucleotide exchange factordomains. Other proteins with similardomains activate the Ras superfamily ofGTPases. Such proteins may play impor-tant roles in signaling pathways, intracel-lular trafficking and the organization ofthe cytoskeleton. Additional functionaldomains identified indicate potentialmembrane attachment (pleckstrindomain) and phosphatidylinositol sig-naling function (MORN repeats).Hadano et al.2 show that ALS2 mRNA iswidely expressed in several regions of thebrain and spinal cord and is also foundin many other tissues outside the centralnervous system.

Selective vulnerability of motor neu-rons in ALS is an important and tantaliz-ing issue. We now have two genes that areexpressed in many cell types and yet causea highly selective clinical syndrome ofmotor system failure. For SOD1-relatedALS, however, we know from clinical,imaging and autopsy studies that diseasephenotypes can vary and that degenera-tion in the brain is often present outsidethe motor system. This is also the case fora significant number of people with spo-radic ALS. The data in this issue of NatureGenetics show that different mutations ina single gene can cause variable clinicalmotor system phenotypes through arecessive mechanism. This is an importantopportunity, because loss-of-functionmechanisms are potentially much moreamenable to elucidation through geneticknockout strategies than is an unknowngain of function. The field of ALS researchis therefore at the start of a new endeavorthat will reveal much basic information onmechanisms of neuronal survival andinjury, as well as potential strategies forneuroprotection. As this new gene is asso-ciated with a very rare variant of ALS, itsrelevance to the more commonly encoun-tered forms of the disease and to the widerspectrum of neurodegenerative disordersremains to be seen. �1. Yang, Y. et al. Nature Genet. 29,160–165 (2001).2. Hadano, S. et al. Nature Genet. 29,166–173 (2001).3. Hentati, A. et al. Nature Genet. 7, 425–428 (1994).4. Ben Hamida, M. et al. Brain 113, 347–363 (1990).5. Charcot, J.M. & Joffroy, A. Arch. Physiol. Norm.

Pathol. 2, 354–367 (1869).6. Rosen, D.R. et al. Nature 362, 59–62 (1993).7. Cookson, M.R. & Shaw, P.J. Brain Pathol. 9, 165–186

(1999).8. Cleveland, D.W. & Liu, J. Nature Med. 6, 1320–1321

(2000).9. Ince, P.G. et al. J. Neuropathol. Exp. Neurol. 57,

895–904 (1998).10. Figlewicz, D.A. et al. Hum. Mol. Genet. 3,

1757–1761 (1994).11. Tomkins, J. et al. NeuroReport 9, 3967–3970 (1998).12. Al-Chalabi, A. et al. Lancet 347, 159–160 (1996).13. Jackson, M. et al. Ann. Neurol. 39, 796–800 (1996).14. Veldink, J.H. et al. Neurology 56, 753–757 (2001).

familial ALS

sporadic ALS

chromosomallinkage sites

genes

other forms of motorneuron degeration • 9q34

• 9q21-q22• 15q15-q21• X centromere

• deletions/insertions in KSP repeat region of neurofilament heavy subunit• SMN (survival motor neuron)• NAIP (neuronal apoptosis inhibitory protein)• apolipoprotein E ε 4• excitatory amino acid transporter 2 (EAAT2)• cytochrome c oxidase subunit 1• AP endonuclease• ? Mn SOD

• Kennedy's disease— androgen receptor gene• spinal muscular atrophy –SMN –NAIP

• SOD1 (Cu/Zn superoxide dismutase)• ALS2 (alsin)

BOB CRIMI

Fig. 2 Maintaining motor neurons. Genetic factors currently identified that may be associated with, orpredisposed to, motor neuron degeneration.

©20

01 N

atu

re P

ub

lish

ing

Gro

up

h

ttp

://g

enet

ics.

nat

ure

.co

m© 2001 Nature Publishing Group http://genetics.nature.com