letter nature genetics • volume 24 • february 2000 197 Localization to Xq27 of a susceptibility gene for testicular germ-cell tumours Elizabeth A. Rapley 1 , Gillian P. Crockford 2 , Dawn Teare 3 , Patrick Biggs 1 , Sheila Seal 1 , Rita Barfoot 1 , Sandra Edwards 1 , Rifat Hamoudi 1 , Ketil Heimdal 4 , Sophie D. Fosså 5 , Kathy Tucker 6 , Jenny Donald 7 Felicity Collins 6 , Michael Friedlander 6 , David Hogg 8 , Paul Goss 8 , Axel Heidenreich 9 , Wilma Ormiston 10 , Peter A. Daly 10 , David Forman 2 , R. Timothy D. Oliver 2 , Michael Leahy 2 , Robert Huddart 1 , Colin S. Cooper 1 , Julia G. Bodmer 11 , Douglas F. Easton 3 , Michael R. Stratton 1 & D. Timothy Bishop 2 1 Sections of Cancer Genetics and Molecular Carcinogenesis, Institute of Cancer Research, Haddow Laboratories, Sutton, Surrey, UK. 2 Imperial Cancer Research Fund Genetic Epidemiology Lab, Ashley Wing, Leeds, UK. 3 CRC Genetic Epidemiology Unit, Strangeways Research Laboratories, Worts Causeway, Cambridge, UK. 4 Unit of Medical Genetics, Department of Oncology, The Norwegian Radium Hospital, Oslo, Norway. 5 Department of Oncology, The Norwegian Radium Hospital, Oslo, Norway. 6 Department of Oncology, Prince of Wales Hospital, Randwick, Australia. 7 Department of Biological Sciences, Macquarie University, N.S.W., Australia. 8 Department of Medicine, University of Toronto, Medical Sciences Building, Ontario, Canada. 9 Department of Urology, Philipps-University, Marburg, Germany. 9 Department of Medical Oncology, Hope Directorate, St James’s Hospital, Dublin, Ireland. 10 Imperial Cancer Research Fund, Cancer Genetics & Immunology Laboratory, John Radcliffe Hospital, Oxford, UK. Correspondence should be addressed to M.R.S. (e-mail: mikes@icr.ac.uk). Testicular germ-cell tumours (TGCT) affect 1 in 500 men and are the most common cancer in males aged 15–40 in Western Euro- pean populations 1 . The incidence of TGCT has risen dramatically over the last century 2-5 . Known risk factors for TGCT include a history of undescended testis (UDT), testicular dysgenesis, infer- tility 6 , previously diagnosed TGCT (ref. 7) and a family history of the disease 8-10 . Brothers of men with TGCT have an 8-10-fold risk of developing TGCT (refs 8,9), whereas the relative risk to fathers and sons is fourfold (ref. 9). This familial relative risk is much higher than that for most other types of cancer. We have collected samples from 134 families with two or more cases of TGCT, 87 of which are affected sibpairs. A genome-wide linkage search yielded a heterogeneity lod (hlod) score of 2.01 on chro- mosome Xq27 using all families compatible with X inheritance. We obtained a hlod score of 4.7 from families with at least one bilateral case, corresponding to a genome-wide significance level of P=0.034. The proportion of families with UDT linked to this locus was 73% compared with 26% of families without UDT (P=0.03). Our results provide evidence for a TGCT susceptibility gene on chromosome Xq27 that may also predispose to UDT. Localization of a TGCT predisposition gene has been hampered by the relative rarity of multigenerational pedigrees with several affected cases, which are most informative for genetic linkage analysis. Genetic linkage analysis of the International Testicular Cancer Linkage Consortium set of 134 families (Table 1) using polymorphic markers has excluded the possibility that suscepti- bility to TGCT is due to a single autosomal gene that accounts for all the familial risk (data not shown). We have previously reported regions with suggestive evidence in favour of linkage identified in our autosomal genome search 11-12 . The increased risk of TGCT to fathers or sons of cases has been reported to be less than the increased risk to brothers of cases 8,9 . As this could be interpreted as evidence of X inheritance, we extended the linkage search to include the X chromosome. In these analyses we excluded from genotyping the 35 (26% of all families) families which show male-to-male transmission and hence, a priori, are inconsistent with X linkage. Linkage analysis of the set of families compatible with X linkage, 80% of which are sibpairs, provided preliminary evidence for a TGCT predisposi- tion locus at Xq27-28 (maximum hlod score=2.01, α=0.32, maximum multipoint lod score under homogeneity of -19.92; Table 2, Figs 1 and 2). Evidence in favour of linkage was observed in families from all contributing groups (data not shown). We subsequently stratified families according to the presence of at least one bilateral case, the presence of UDT, histology and age (Table 2). Families with at least one case of bilateral disease, all of which are sibpairs except for one maternal cousins pedigree and one of three sibs, showed strong evidence of linkage to the locus on Xq27 (hlod score=4.76, α=1.00; Fig. 2) and were more likely to be linked to the X chromosome than families without a Table 1 • Breakdown of all families in consortium set Family type Australia Canada Germany Ireland Norway UK Total sibpair 6 3 12 13 53 87 sibtrio 4 a 4 father/son b 2 2 6 10 cousins 1 1 2 4 maternal cousins 1 1 c 2 4 paternal cousins 1 4 5 paternal uncle/nephew 3 1 4 maternal uncle/nephew 1 2 4 7 other 3 d 1 e 1 f 2 g 2 h 9 total 14 4 14 1 23 78 134 i a Pedigree 341; third affected sib deceased and no sample exists. b Father/son pedigrees were only used in the genome wide linkage search if another unaffected male sibling was available for genotyping. c Pedigree 155; two brothers married two sisters. d Four cousins, uncle/two nephews and second cousins. e Four sibs and second cousins once removed. f Father/son and cousin. g Father/two sons; three cousins. h Father/son and cousin; sibs and cousin. i There are 103 potentially X-linked families in this set. However, DNA from four families was unavailable for typing of X chromosome markers. © 2000 Nature America Inc. • http://genetics.nature.com © 2000 Nature America Inc. • http://genetics.nature.com