NATURE MEDICINE • VOLUME 7 • NUMBER 2 • FEBRUARY 2001 153

NEWS

large sets of families to search for geneticlinkages. A number of putative loci havebeen identified (Fig. 1). These includeHPC1, PCAP, HPCX and CAPB. However,confirmatory linkage studies are generallylacking, and a strong candidate gene hasnot yet been cloned.

The hereditary form of prostate cancer isremarkably heterogeneous. This adds tothe other obstacles of linkage detection,which include a high rate of sporadic cases,late age of diagnosis, various modes of in-heritance (autosomal, X-linked/dominantor recessive), and strong influences frommodifying factors. In the February issue of

NEWS & VIEWS

Prostate cancer is the most common ma-lignancy in North American and

European men, and represents a majorpublic health challenge. Traditionally con-sidered a disease of elderly men, a consider-able proportion of cases occur in men intheir pre-retirement years. New means ofidentifying individuals at risk and strate-gies for early detection and preventive careare greatly needed.

Prostate cancer has, like other commoncancers, a recognized familial component1.Though major susceptibility genes for col-orectal and breast cancer have been identi-fied and characterized, the genetic factorsthat underlie inherited prostate cancer re-main elusive. Several research groups havedone extensive work to identify these fac-tors, performing genome-wide scans on

Nature Genetics, Tavtigian et al.2 identify acandidate prostate cancer susceptibilitygene; they report a significant linkage inhigh risk prostate cancer families to mark-ers on chromosome 17p (two point loga-rithm of odds (LOD) score of 4.5 forD17S1289, meaning that there is less thana 1 in 10,000 chance of random associa-tion). Moreover, the authors provide evi-dence of germline frameshift andnonconservative missense mutations, aswell as disease-associated common vari-ants, in a candidate gene named ELAC2(Fig. 1).

Homologues to ELAC2 have been described in yeast, bacteria and plants2 (www.ncbi.nml.nih.gov). Thesegenes are predicted to encode a highly conserved metal-dependent hy-

Hereditary prostate cancer: a new piece of the puzzle

ÅKE BORG

An eagerly awaited prostate cancer susceptibility gene has been announced. But does the candidate live up to expectations?

Although most parasites invade hostcells by forming an internalization vac-uole that carries them into the cell with-out disrupting the plasma membrane,Plasmodium spp. sporozoites have founda way to enter cells by a completelynovel mechanism. The sporozoite, theinfective stage of the malaria parasite, istransmitted by bite of infectedAnopheles mosquitoes. In humans,sporozoites migrate from the skin to livercell, where they enter hepatocytes andmature into the merozoites that enterthe bloodstream and infect red bloodcells. Little is known about the cell biol-ogy of the sporozoite, including themechanisms they use to enter and exitcells.

Videos made years ago by JeromeVanderberg at the New York UniversitySchool of Medicine showed sporozoitesthat appeared to enter and exit cellswithout the aide of an internalizationvesicle, but this behavior was originallybelieved to be an artifact. Upon re-exam-ining the old videos, NYU cell biologistsbecame intrigued by the phenomenonand decided to investigate further. “Atthe time people believed that the para-sites were going underneath the cells,but when we saw these movies, we reallyhad the impression that the sporozoiteswere moving through them,” said NYU

researcher Ana Rodriguez.Rodriguez and colleagues

began using time-lapsevideo analysis and fluores-cent antibody labeling ex-periments to track themovement of Plasmodiumsporozoites in cell cultures.In the 5 January issue ofScience, Mota et al. (Science291, 141–144) demon-strate that the parasites areactually capable of breach-ing the plasma membraneto enter and exit cells. Theconfocal image (picture)shows two sporozoites(blue) that have disruptedthe hepatocyte plasmamembrane (red) and are able to movethroughout the cytosol (green), inde-pendent of vacuolar membranes. Insome cases the cell is capable of rapidlyresealing the broken membrane, but inother cases the breakage causes cyto-plasmic leakage and cell death. Mota etal. determined that sporozoites trans-verse an average of four cells an hour,and suggest that they may need to tra-verse several cells in search of hepato-cytes that are suitable for infection.

This is the first demonstration that anon-viral pathogen can transverse the

cell membrane without forming an inter-nalization vacuole. The authors plan touncover the exact mechanism used bythe parasites to enter the cell. “Theprocess may be mediated by proteasesor lipases that the sporozoite releases todestroy the plasma membrane, or elseby the mechanical force of the parasiteitself,” said Rodriguez, the senior authoron the study. “Perhaps one day we maylearn how to inhibit this process, andthen malarial infection could be abro-gated at its first step.”

Kristine Novak

Picky Plasmodium search for the perfect host

©20

01 N

atu

re P

ub

lish

ing

Gro

up

h

ttp

://m

edic

ine.

nat

ure

.co

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

154 NATURE MEDICINE • VOLUME 7 • NUMBER 2 • FEBRUARY 2001

NEWS NEWS & VIEWS

drolase, and the amino acid sequence ofthis protein is similar to proteins involvedin DNA interstrand crosslink repair andmRNA editing. This discovery is intriguing,but it may still be too early to be declared amajor breakthrough, as it seems unclearhow to interpret the data from our currentperspective.

The prostate is an androgen-regulatedorgan. Androgens, such as testosterone, areknown to be strong tumor promoters, act-ing via the androgen receptor (AR) to stim-ulate cell division and enhance the effect ofendogenous and exogenous carcinogens.The risk of prostate cancer is more than50% higher in African-American men thanin European-American men, and two- tothree-fold lower in native Japanese andChinese men3. This may be partly be due tofactors such as the ethnic differences in cir-culating levels of free testosterone, genesassociated with androgen synthesis (suchas SRD5A2 and CYP17)3,4 or other differ-ences in AR-related signaling pathways.

Exon 1 of the X-linked gene encodingAR contains several trinucleotide repeats,and the length of a polymorphic polygluta-mine (CAG) tract has been shown to be in-versely related to receptor-coactivatorinteraction and transcriptional activity, aswell as to prostate cancer risk5. Likewise, agermline mutation in the hormone-bind-ing domain of AR, altering its transactiva-tional specificity, was found to besignificantly more common in both famil-ial and sporadic cases than in controls, andestimated to contribute to cancer develop-ment in up to 2% of Finnish prostate can-cer patients6. Moreover, homozygosity fora polymorphism in the androgen responseelement of the prostate-specific antigen(PSA) gene promotor was in itself associ-ated with a slightly elevated prostate can-cer risk, but was in combination with ashort (hyperactive) AR CAG allele found toresult in a five-fold increased risk7. Thesefindings indicate that critical combina-tions of common polymorphisms in rele-vant genes might promote prostate cancerdevelopment, both in a sporadic and famil-ial setting. Accordingly, the potential mod-ifying effect of these variants should beconsidered when assessing the risks attrib-uted by mutations in prostate cancer sus-ceptibility genes, and, whenever possible,should also be incorporated into models ofsegregation and linkage analysis.

The report by Tavtigian et al.2 reinforcesthe importance of considering a complexand possibly multigenic mode of prostatecancer inheritance. The authors analyzed atotal of 127 large Utah families, and found

42 families (or branches of them) having atleast 4 or 5 cases sharing identical alleles formarkers on chromosome 17p. Using thesame criteria, a similar number of HPC1- (alocus previously linked to prostate cancer)associated famlies were recorded, suggest-ing that this study had uncovered twomajor susceptibility loci. However, sequence analysis of ELAC2 genes from the people of the 42 17p-linked families re-sulted in only one detected mutation, aframeshift insertion leading to proteintruncation (1641insG; Fig. 1). To add to theconfusion, whereas this mutation wasidentified in the a 46-year-old patient froma family in which 6 of 8 cases on the pater-nal side share a 17p-haplotype, the1641insG mutation was found to originatefrom the maternal branch of the family,where prostate cancer is less prevalent.Furthermore, the youngest patient in thisfamily carries neither the mutation nor thehaplotype (Fig. 1).

How should these paradoxes be inter-preted? Although large rearrangements aswell as mutations in non-coding regions ofthe gene will escape detection, this proba-bly does not explain the low mutation rateobserved. Does the 17p region contain asecond susceptibility gene? Is the wild-type17p allele lost in tumors from linked fami-lies? Are we faced with a new mechanismof cancer risk inheritance?

Tavtigian et al.2 also identified two com-mon ELAC2 missense mutations, S217Land A541T (Fig. 1), and demonstrated thatthe strongly-linked Leu217/Thr541 variantsoccur more frequently in prostate cancercases than in age-matched controls, an as-sociation that has been confirmed in an in-dependent study8. Moreover, by screeninganother set of young patients, a secondharmful mutation was revealed. This mis-sense mutation, R781H, was shown topartly segregate with disease in this family.Moreover, R781H was found on an ELAC2

1

13

1720

X

CAPB

HPC1

PCAP

HPC2BRCA2

HPC20BRCA1

HPCX

AR

exon 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

200bp

Arg781HisAla541Thr 1641insGSer217Leu

42 51 46

58

90

67 91

59

87 75 63

46 65 60

71 8472

80

69

61

76

6974

43

Kindred 4102; 17p haplotype; ELAC2 1641insG

PSA 4.2 PSA 5.7

index

q12.3

q42q43

q24q25

p36

p12

q21 q13q12

q27q28

ELAC2

Fig. 1 Chromosomal localization of putative prostate cancer susceptibility genes. The putative prostatecancer susceptibility genes HPC1 and PCAP were found by linkage in early onset cancer families9,10,whereas CAPB was pinpointed by linkage in prostate/brain cancer families11. Segregation analyses haveindicated the existence of X-linked susceptibility loci, supporting the finding of linkage to HPCX (ref. 12).Variants of the androgen receptor (AR) gene have also been associated with prostate cancer risk5,6. Aminor proportion of prostate cancer can be ascribed to germline mutations in BRCA2, and possibly also toBRCA1, which are breast/ovarian cancer genes13. Recent evidence implicate linkage to a locus on chro-mosome 20, a region often amplified in human cancer and in which a novel androgen-regulated genehave been identified14,15. Additional, less well-characterized loci have been localized to chromosomes 2, 4,5, 7, 10, 11p, 12, 13, 14 and 16q (refs. 9,16,17). The ELAC2 gene has 24 coding exons spanning a ≈27 kbgenomic region on chromosome 17p11–p12, and is transcribed into an ≈3.0 kb transcript that encodesa protein of 826 amino acids. The locations of mutations and variants, identified by Tavtigian et al.2, areindicated (top). The pedigree of one prostate cancer family, kindred 4102 (bottom), in which 6 of 8 pa-tients in the paternal (right) branch of the index case share a 17p haplotype (blue), but where the ELAC21641insG mutation (red) originates from the maternal branch. Non-haplotype/mutation carriers are indi-cated in white. Prostate cancer cases are shown in red, two unaffected individuals with elevated PSA levelsare shown in green. Age at diagnosis or, for non-affected, at death/last follow-up is shown within symbols.

©20

01 N

atu

re P

ub

lish

ing

Gro

up

h

ttp

://m

edic

ine.

nat

ure

.co

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

NATURE MEDICINE • VOLUME 7 • NUMBER 2 • FEBRUARY 2001 155

allele that also carried the Leu217/Thr541variants, and the authors hypothesized thatthe combined effect of three missense alter-ations would give rise to a highly penetrantELAC2 mutant. Notably, the triple-changewas also present in an ovarian cancer pa-tient of this family, which also included anovarian cancer patient with theLeu217/Thr541 allele.

Identification of the ELAC2 gene has leftus with more questions than answers, andconfirmatory studies need to be performedbefore it can be called a major prostate can-cer susceptibility gene. However, the se-quence similarity between ELAC2 andDNA interstrand crosslink repair proteinsPSO2 (SMN1) is intriguing in light of theestablished connection between DNA re-pair deficiency and cancer susceptibility.Interestingly, from a clinical point of view,alkylating drugs such as mitomycin C—attimes used in treatment of advancedprostate cancer—are known to exert theircytotoxic effect by inducing DNA crosslinkformation. It would be useful to further in-vestigate whether the ELAC2 product and

its allelic variants have a role in repair of in-terstrand crosslinks or DNA damage causedby other carcinogens.

1. Carter, B.S., Beaty, T.H., Steinberg, G.D., Childs, B. &Walsh, P.C. Mendelian inheritance of familialprostate cancer. Proc. Natl. Acad. Sci. USA 89,3367–3371 (1992).

2. Tavtigian, S.V. et al. A strong candidate prostate can-cer susceptibility gene at chromosome 17p. NatureGenet. 27, 172-180 (2001).

3. Ross, R.K. et al. Androgen metabolism and prostatecancer: establishing a model of genetic susceptibil-ity. Cancer Res. 58, 4497–4504 (1998).

4. Wadelius, M., Andersson, A.O., Johansson, J.E.,Wadelius C. & Rane, E. Prostate cancer associatedwith CYP17 genotype. Pharmacogenet. 9, 635–639(1999).

5. Irvine, R.A., Yu, M.C., Ross, R.K. & Coetzee, G.A. TheCAG and GGC microsatellites of the androgen re-ceptor gene are in linkage disequilibrium in menwith prostate cancer. Cancer Res. 55, 1937–1940(1995).

6. Mononen, N. et al. Two percent of Finnish prostatecancer patients have a germ-line mutation in thehormone-binding domain of the androgen receptorgene. Cancer Res. 60, 6479–6481 (2000).

7. Xue, W. et al. Susceptibility to prostate cancer: inter-action between genotypes at the androgen receptorand prostate-specific antigen loci. Cancer Res. 60,839–841 (2000).

8. Rebbeck, T.R. et al. Association of HPC2/ELAC2 geno-types and prostate cancer. Am. J. Hum. Genet. 67,1014–1019 (2000).

9. Smith, J.R. et al. Major susceptibility locus for prostatecancer on chromosome 1 suggested by a genome-wide search. Science 274, 1371–1374 (1996).

10. Berthon, P. et al. Predisposing gene for early-onsetprostate cancer, localized on chromosome 1q42.2-43. Am. J. Hum. Genet. 62, 1416–1424 (1998).

11. Gibbs, M. et al. Evidence for a rare prostate cancer-susceptibility locus at chromosome 1p36. Am. J.Hum. Genet. 64, 776–787 (1999).

12. Xu, J. et al. Evidence for a prostate cancer suscepti-bility locus on the X chromosome. Nature Genet. 20,175–179 (1998).

13. Gayther, S.A. et al. The frequency of germ-line mu-tations in the breast cancer predisposition genesBRCA1 and BRCA2 in familial prostate cancer.Cancer Res. 60, 4513–4518 (2000).

14. Berry, R. et al. Evidence for a prostate cancer-suscep-tibility locus on chromosome 20. Am. J. Hum. Genet.67, 82–91 (2000).

15. Xu, L.L. et al. A novel androgen-regulated gene,PMEPA1, located on chromosome 20q13 exhibitshigh level expression in prostate. Genomics 66,257–263 (2000).

16. Gibbs, M. et al. A genomic scan of families withprostate cancer identifies multiple regions of inter-est. Am. J. Hum. Genet. 67, 100–109 (2000).

17. Suarez, B.K. et al. A genome screen of multiplex sib-ships with prostate cancer. Am. J. Hum. Genet. 66,933–944 (2000).

Department of OncologyUniversity Hospital, Lund, Swedenemail: ake.borg@onk.lu.se

NEWS & VIEWS

Spider venom helps hearts keep their rhythmAlthough the sight of a large hairy spider isenough to give some people a heart at-tack, spider venom may eventually beused to treat some heart condi-tions. In the 4 January issue ofNature, Bode et al. (Nature, 409,35–36) report that a peptide iso-lated from the venom of the spiderGrammostola spatulata (picture) in-hibits atrial fibrillation. Atrial fibrilla-tion, characterized by disorganizedatrial activity, occurs in patients suf-fering from valve disease, hyper-tension, and chronic lung disease.It is the most common form of car-diac arrhythmia, and can lead tofurther complications such as hy-potension, pulmonary congestion,strokes and angina pectoris.

The anti-arrhythmic effects ofthis peptide were discoveredserendipitously in a research pro-ject that initially had nothing to dowith hearts or spiders. Frederick Sachs, se-nior author on the study, began investi-gating mechanosensitive ion channelsover 15 years ago. In a search for agentsthat block or activate mechanical sensitiv-ity, his group began screening insect ven-oms, which are known to contain a varietyof neuroactive compounds. “We tried a

variety of spider and scorpion venoms inpatch clamp experiments, and finallyfound two that inhibited mechanosensi-

tive ion channel activity,” said Sachs.Fractionation of the venom led to the iso-lation of a peptide from tarantula venom,GxMtx-4, that the authors observed toblock the activity of cationic stretch-acti-vated ion channels (SACs) in adult astro-cytes and cardiocytes.

SACs convert gradients of stress into

gradients of electrical activity, and sus-tained mechanical gradients can developwhen systolic pressure and wall tension

are high, conditions that are presentin the diseased heart. “It was well-known that cardiac wall stretch in-creases heart rate, and we knew thatthese receptors were expressed inheart cells,” said Sachs. He teamedup with Frank Bode and Mike Franzat Georgetown University to showthat the peptide suppressed both theincidence and duration of atrial fibril-lation in rabbit hearts. They alsoshowed that GxMtx-4 acts only dur-ing stretch activation, as it has no ef-fect on the action potential of restingatrial cells. “It also seems to have noacute effects on normal hearts in theabsence of distension,” added Sachs,although further toxicity analysis isrequired.

This finding of Bode et al. consti-tutes the first demonstration that SAC acti-vation can generate atrial fibrillation. Theauthors therefore suggest that GsMtx-4initiates a new class of anti-arrhythmicagents that could act on the cause, ratherthan the effects of cardiac arrhythmias.

Kristine Novak

©20

01 N

atu

re P

ub

lish

ing

Gro

up

h

ttp

://m

edic

ine.

nat

ure

.co

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