A detailed study of loss of heterozygosity on chromosome 17 in tumours from Li ± Fraumeni patients carrying a mutation to the TP53 gene JM Varley 1 , M Thorncroft 1 , G McGown 1 , J Appleby 1 , AM Kelsey 2 , KJ Tricker 3 , DGR Evans 1 and JM Birch 3 1 CRC Department of Cancer Genetics, Paterson Institute for Cancer Research, Wilmslow Road, Manchester M20 9BX, UK; 2 Department of Histopathology and 3 CRC Paediatric and Familial Cancer Research Group, Royal Manchester Children's Hospital, Pendlebury, Manchester M27 4HA, UK We have studied a total of 36 tumours from 28 patients with germline mutations to the TP53 gene for loss of heterozygosity at TP53 using techniques of both direct sequencing and restriction fragment length polymorphism analysis. All patients were from families conforming to the de®nition of classical Li ± Fraumeni syndrome (LFS) or were Li ± Fraumeni-like (LFL). The data we have obtained show that loss of the wild-type TP53 gene is observed in under half (44%) of all tumours, and that the pattern of LOH at TP53 may be mutation speci®c. LOH has been observed in premalignant as well as invasive tumours. Two tumours (6%) show loss of the mutant allele and retention of the wild-type. To con®rm that TP53 is indeed the target for LOH events on chromosome 17, we have used additional microsatellite repeats to examine patterns of allelic imbalance along the length of chromosome 17. Data from this analysis indicate that TP53 is the target of loss, but reveal some other interesting patterns of allelic imbalance at other loci on chromosome 17. Keywords: chromosome 17; LOH; Li ± Fraumeni; tumours Introduction Li ± Fraumeni syndrome (LFS) is an inherited cancer predisposition syndrome characterised by sarcomas in children and young adults, early-onset breast cancer, brain tumours, adrenal cortical carcinoma and leukaemia (Li and Fraumeni, 1969; Li et al., 1988; Birch, 1990; Garber et al., 1991). Over half of all families conforming to the de®nition of classic LFS and approximately one quarter of those which are Li ± Fraumeni-like (LFL, Birch et al., 1994) carry germline mutations to the TP53 gene (Malkin et al., 1990; Birch et al., 1994; Frebourg et al., 1995; Varley et al., 1997). Furthermore, a signi®cant proportion of patients with tumours which are typical of those seen in Li ± Fraumeni syndrome have also been shown to carry TP53 germline mutations at an increased frequency (Malkin et al., 1992; Sameshima et al., 1992; Toguchida et al., 1992; Brugieres et al., 1993; Kyritsis et al., 1994; McIntyre et al., 1994; Wagner et al., 1994; Chen et al., 1995; Diller et al., 1995; Li et al., 1995). TP53 is considered to be a tumour suppressor gene (Hollstein et al., 1991; Levine et al., 1991; Lane, 1992), and as such some tumours from patients with germline TP53 mutations have been analysed for loss of heterozygos- ity (LOH). LOH at a particular locus/gene is considered to indicate the presence of a tumour suppressor at that locus (Ponder, 1988), with the LOH event unmasking the recessive mutation. A number of known tumour suppressor genes such as BRCA1 (Smith et al., 1992), RB1 (Cavenee et al., 1983), BRCA2 (Gudmundsson et al., 1995), APC (Ichii et al., 1992) and VHL (Tory et al., 1989) have been studied in families in which there is a germline mutation to those genes, and loss of the wild-type allele has been seen at high frequency in the tumours of patients carrying the mutation. All these genes therefore ful®l the criteria expected of tumour suppressor genes, namely obeying Knudson's `two hit' hypothesis (Knudson, 1971), whereby one defective allele is inherited, and the second is inactivated by loss of part or all of the wild-type chromosome (Ponder, 1988). To our knowledge 39 tumours from patients with germline mutations at TP53 have been studied for LOH, with LOH detected in 69% of those tumours (Malkin et al., 1990; Chung et al., 1991; Metzger et al., 1991; Bùrresen et al., 1992; Felix et al., 1992, 1993, 1995; Iavarone et al., 1992; Kovar et al., 1992; Prosser et al., 1992; Sameshima et al., 1992; Srivastava et al., 1992; Warneford et al., 1992; Eeles et al., 1993; Scott et al., 1993; Grayson et al., 1994; Gutierrez et al., 1994; Hamelin et al., 1994; Horio et al., 1994; Jolly et al., 1994; Plummer et al., 1994; Chen et al., 1995; Li et al., 1995; Lubbe et al., 1995). However, there has been no systematic study of LOH in a large series of patients with de®ned TP53 mutations, and in tumour material which has been characterised according to standard criteria by one pathologist. We have identi®ed 19 LFS and LFL families in which there are germline TP53 mutations (Birch et al., 1994; Varley et al., 1997), and from these we have obtained material from a total of 41 tumours from which we have been able to analyse LOH at TP53 in 36. In addition, we have used microsatellite repeat sequences at various positions along chromosome 17 to determine the extent of the loss, and the nature of the molecular events occurring in those tumours. Results and Discussion The results from the analysis of TP53 in tumour DNAs are shown in Table 1, with a representative example in Figure 1. The majority of tumour samples were from Correspondence: JM Varley Received 30 September 1996; revised 14 January 1997; accepted 14 January 1997 Oncogene (1997) 14, 865 ± 871  1997 Stockton Press All rights reserved 0950 ± 9232/97 $12.00