Chromosome 1 abnormalities in human
neoplasia
There are multiple lines of evidence that implicate regions of chromosome 1 in the development and progression of cancer. These include regions of the chromosome that are duplicated or lost in tumours, translocations between chromosome 1 and other regions of the genome, genetic linkage studies in familial cancer syndromes, the analysis of candidate genes and functional rescue of tumourigenicity by chromosome transfer. Abnormalities have been identified in a wide range of solid tumours and haematopoietic malignancies however very few of the genes that are the targets of these alterations have been cloned. In many cases this is due to the relatively large critical interval sometimes covering many megabases of DNA. Presentations during the workshop included work towards refining these intervals and identifying the genes at the heart of the aberrations.
Previous chromosome 1 workshop reports provide an eloquent review of abnormalities in human neoplasia for this chromosome (see Vance et al. 1997 and references therein). This report will concentrate on developments that were presented during this workshop or published between this and the last workshop in 1997. Specific sections are present for frequently studied tumour types with a general review for tumour types that are either rare or less frequently associated with chromosome 1.
Chromosome 1 abnormalities in cancer
The distal portion of 1p continues to be the region of chromosome 1 thought to contain genes involved in cancer. Vortmeyer and colleagues (1998) found loss of heterozygosity (LOH) of 1p35-p36 in 7 out of 10 Merkel cell carcinomas while Williamson and co-workers (1997) detected loss of the same cytogenetic interval in 27% of 39 parathyroid adenomas and T. Martinsson (this report) observed LOH in this region in germ cell tumours. Proximally, 1p34-p36 was deleted in 5 of 11 informative pheochrmocytomas (Vargas et al., 1997). Amplification of 1q was reported in 43% (6 out of 14) and 73% (11 out of 15) of endometrial cancers analysed by comparative genomic hybridisation (Sonoda et al. 1997 and Suzuki et al. 1997 respectively) while amplification of 1q23 was observed in adenocarcinoma of the lung (Petersen et al., 1997). Almeida and colleagues (1998) identified a me Mer of the leucine-rich repeat superfamily, GAC1, that is amplified and over expressed in malignant gliomas. Loss of heterozygosity on 1q occurs less frequently than on 1p, for example Pietsch and colleagues (1997) found LOH between 1q31 and q32 in 36% of 30 medulloblastomas.
Breast and ovarian cancer
Multiple regions of chromosome 1 have been implicated in breast and ovarian cancer. Y. Hey (this report) reported considerable progress on the characterisation of a 200 Kb region on 1p31.1, an interval frequently lost in breast cancer. Sequencing of the critical interval is nearly complete and the analysis of the sequence is underway. Additional experiments to find genes in the region are in progress. Deletion of a proximal interval, 1p22-p31, has been implicated in local progression and metastasis (Tsukamoto et al. , 1998). Functional studies with the melanoma metastasis suppressor gene, KiSS 1, which maps to 1q32-q41, suggest it may suppress metastasis in some breast tumours (Lee et al. 1997) and co-inside with 1q deletions in late stage breast carcinomas. However the picture is clouded by the amplification of the MUC1 mucin gene at 1q21-q24 in breast tumours (as shown by Southern analysis, Bieche and Lidereau 1997). This demonstrates the heterogeneity of gains and losses of chromosome 1 in breast cancers. Thompson and colleagues (1997) confirmed the observation of 1p36 deletions in ovarian tumours and reported consistent translocations between 1p36 and chromosome 17 in 3 out of 11 cases. These suggestions suggested that this region is important in some sporadic ovarian cancers.
Colorectal cancer
A significant correlation has been reported between 1p deletions and aneuploidy in colorectal tumors. These observations suggest that loss of genes in this region may be implicated in chromosome instability (Di Vinci et al., 1998). More specifically, a comparison of disease free interval, survival and LOH at 5 microsatellite makers on 1p32 and 1p36 showed that allelic loss was an independent predictor of poor prognosis (Ogunbiyi et al., 1997).
Neuroblastoma
Many recent publications, and presentations at this workshop, indicate that investigation has continued into the involvement of chromosome 1 in neuroblastma. The number of different minimal regions that have been suggested may be due to experimental variation or could be biased by the collection method and phenotypes of the different tumour sets. Recent mapping data locates neuroblastoma genes in 1p36 (Martinsson et al., this report), 1p36 and 1p31-32 (Avigad et al. 1997) and 5 locations between 1p34-pter (Kageyama et al., this report). While these findings complement previous work, no specific genes have been identified by these methods. A limited study of familial neuroblastomas identified LOH of 1p in tumours from three patients (Tonini et al. 1997) however linkage studies in three large neuroblastoma kindreds excluded 1p36.2-36.3 as the location of a familial neuroblastoma gene (Maris et al., 1997). An alternative mapping approach identified two genes, Alx3 and p73, that were altered in neuroblastomas and map to 1p13-21 and 1p36 respectively (Zhu et al., this report). While a candidate gene approach has excluded CDC2L1 (Martinsson et al., 1997) and phospholipase A2 (Haluska et al., 1997) which are both located on 1p36. One possible route to reduce and clarify the minimal region thought to contain the neuroblastoma gene is to compare the minimal regions identified in different tumour types in the hope that only one gene is the target of the deletions. To this end Martinsson et al. (this report) have identified a common interval of 5 cM that is lost in both neuroblastomas and germ cell tumours.
Prostate cancer
The genetic analysis of prostate cancer families has continued with particular interest on chromosome 1. Cooney and co-workers (1997) confirmed that 1q24-q25 is likely to contain the prostate cancer gene called HPC1 (Smith et al. 1996) while Eeles and colleagues (1998) found no evidence of linkage in their prostate cancer families. A detailed analysis of the data from Smith and co-workers (1996) indicated that only 34% of familial prostate cancer may be linked to HPC1. Furthermore, linkage to this gene was found in large kindreds (with five or more members affected) with the disease diagnosed at an early age, suggesting that HPC1 is not the only familial prostate cancer gene (Gronberg et al., 1997). Indeed, Berthon and colleagues (1998) have identified a second familial prostate cancer locus on 1q42.2-q43 that is distinct from HPC1. Approximately 50% of the families in this European study were linked to this new locus. So far, neither of these familial prostate cancer genes have been cloned.
Sarcomas
Interest has continued on possible regions of amplification in sarcomas (see Forus et al. and Meza-Zepeda et al., this report). 1q21-q24 is amplified in liposarcomas but not in lipomas (Szymanska et al., 1997). A similar interval, 1q21-q22, was found to be amplified in the sarcomas studied by Meza-Zepeda and colleagues (this report). It would be easy to speculate that a single gene is the target of these aberrations in both studies. However Meza-Zepeda has refined the original amplicon to the extent that it may be two separate amplified regions indicating the presence of two genes. Gains of 1q were also observed in alveolar soft part sarcomas, although the size of the amplicon and significance of the event is unclear (Kiuru Kuhlefelt et al., 1998).