Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Characterization of a 500 kb region on 17q25 and the exclusion of candidate genes as the familial Tylosis Oesophageal Cancer (TOC) locus


The locus for a syndrome of focal palmoplantar keratoderma (Tylosis) associated with squamous cell oesophageal cancer (TOC) has been mapped to chromosome 17q25, a region frequently deleted in sporadic squamous cell oesophageal tumours. Further haplotype analysis described here, based on revised maps of marker order, has reduced the TOC minimal region to a genetic interval of 2 cM limited by the microsatellite markers D17S785 and D17S751. Partial sequence data and complete physical maps estimate the actual size of this region to be only 0.5 Mb. This analysis allowed the exclusion of proposed candidate tumour suppressor genes including MLL septin-like fusion (MSF), survivin, and deleted in multiple human cancer (DMC1). Computer analysis of sequence data from the minimal region identified 13 candidate genes and the presence of 50–70 other ‘gene fragments’ as ESTs and/or predicted exons and genes. Ten of the characterized genes were assayed for mutations but no disease-specific alterations were identified in the coding and promoter sequences. This region of chromosome 17q25 is, therefore, relatively gene-rich, containing 13 known and possibly as many as 50 predicted genes. Further mutation analysis of these predicted genes, and others possibly residing in the region, is required in order to identify the elusive TOC locus.


Oesophageal cancer (predominantly squamous cell carcinoma; scc) is one of the 10 leading causes of cancer mortality worldwide, with marked regional variations in incidence and mortality (Munoz and Castellsague, 1994; Parkin et al., 1993). Mortality is high because of late presentation and diagnosis. In developed countries, the aetiology of SCC has been mainly attributed to tobacco smoking and heavy alcohol drinking (Brown et al., 2001), however, dietary deficiencies may also play a major role in both the high-risk populations of developing countries and in the western world (Brown et al., 2001; Franceschi et al., 2000).

Squamous cell oesophageal cancer has been associated with focal non-epidermolytic palmoplantar keratoderma (historically termed Tylosis) in three families (Howel-Evans et al., 1958; Ellis et al., 1994; Hennies et al., 1995; Stevens et al., 1996). The UK and US families are both large, each comprising more than six generations and containing at least 100 affected individuals, and a smaller German family has also been observed. The skin disorder and oesophageal cancer segregate together in all three pedigrees, thus implying that the same gene causes the two phenotypes in these individuals. The incidence of other common cancers in these families is not altered compared to the normal Caucasian population. The causative locus has been designated the tylosis oesophageal cancer (TOC) gene (Risk et al., 1994).

Tylosis is an autosomal, dominantly inherited disorder of the skin that manifests as focal thickening of the palmar and plantar surfaces together with oral lesions and shows complete penetrance by the age of 20 (Ellis et al., 1994). This mode of inheritance, together with the observation that squamous cell oesophageal cancer is intimately related to the skin condition in the UK and US families, allowed the localization of the oesophageal cancer gene by mapping the locus responsible for tylosis. Linkage analysis in all three families, and haplotype analysis in the UK and US pedigrees have located the TOC gene locus to a small region of chromosome 17q25 (Risk et al., 1994, 1999; Hennies et al., 1995; Stevens et al., 1996; Kelsell et al., 1996).

Although the incidence of tylosis with oesophageal cancer is low in the general population, studies of sporadically occurring squamous cell oesophageal carcinomas have demonstrated specific loss of heterozygosity (LOH) at microsatellite markers that are located close to the TOC locus (von Brevern et al., 1998; Iwaya et al., 1998). Thus, it is possible that a tumour suppressor gene resides in this region. In addition, a role for the TOC gene in Barrett's adenocarcinoma of the oesophagus and breast cancer has been suggested by recent data (Dunn et al., 1999; Kalikin et al., 1996; Fukino et al., 1999), so the TOC gene is implicated in the aetiology of a number of sporadic tumours that are an important cause of mortality worldwide.

The production of a partial physical map of the TOC region, coupled with haplotype analysis of intragenic polymorphisms, allowed the exclusion of a strong candidate gene, envoplakin, as the causative TOC locus (Risk et al., 1999). Recent publications of two complete physical maps in this region (Kuhlenbäumer et al., 1999; Kalikin et al., 1999), together with the first draft of the human genome sequence (International Human Genome Sequencing Consortium, 2001; Venter et al., 2001), have now allowed us to redefine the location of the TOC gene using haplotype analysis based on definitive microsatellite marker order. This redefined minimal region has excluded a number of plausible candidate genes for TOC. Bioinformatic analysis using a variety of computational databases and tools has allowed the identification of further candidate tumour suppressor genes located within the TOC minimal region, which were screened for disease-specific mutations.


Integration of three physical maps and assembly of contiguous sequence data

Several of the clones identified in the recently published physical maps were already present in our partial contig of the TOC region (Risk et al., 1999; Kuhlenbäumer et al., 1999; Kalikin et al., 1999). Collaborations with Dr G Kuhlenbäumer, Professor C Van Broeckhoven and Professor E Petty allowed us access to their mapping data before publication. Discrepancies between the maps were removed by analysis of clones from the RPCI-11 BAC library. The location of D17S785 was confirmed as proximal to D17S1817 whilst D17S722 was positioned distal to D17S801 but proximal to D17S937 (Figure 1a,b and HS Mills MPhil Thesis). D17S1603 was confirmed to map between D17S1839 and D17S785. Human genome sequence data from RPCI-11 BAC clone 666A8 (Accession number AC015802) also established that D17S1817 is located between D17S785 and D17S801. This clone, together with RPCI-11 BAC clones 318A15 (Accession number AC005837), 234P2 (Accession number AC024720) and 87G24 (Accession number AC016168) form a partial sequence contig covering approximately 99% of the TOC minimal region (Figure 1b,c). Sequence data from clones 666A8 and 87G24 confirm the location of D17S722 distal to D17S801 and D17S1817 and the location of D17S751 proximal to D17S722.

Figure 1

(a) Genetic, (b) physical and (c) sequence/transcript maps of the TOC region of chromosome 17. Drawn approximately to scale. (a) Genetic distances are taken from the Généthon and GDB combined maps. (b) All physical clones are RPCI-11 BACs except for 96G20 which is an RPCI-1 PAC. BAC clones with names in italics and underlined form the sequencing contig shown in (c). (c) Accession details and number of fragments are given for each BAC in the sequencing contig. Sizes of fragments are taken from sequencing data. Unmapped fragments of clone 234P2 are not shown. Gene symbols are shown above the line if the gene is coded for by the + strand and below the line if coded for by the − strand. 1: ribosomal protein R7L; 2: UBE2-230k promoter; 3: AANAT; 4: FLJ22341; 5: myoglobin-like; 6: SThM; 7: ST6GalNAcI; 8: MXRA7-like; 9: KIAA0585; 10: CG5013-like; 11: SFRS2; 12: ET; 13: MGAT5-like. EST symbols are shown above the line if coded for by the + strand and below the line if coded for by the − strand. EST symbol and gene symbol are overlaid if the EST is located within the gene sequence. (a) stSG4863; (b) stSG52568; (c) stSG52396; (d) stSG54053; (e) stSG50865; (f) stSG62536; (g) stSG51808; (h) Bda94c06; (i) stSG48272; (j) stSG161; (k) SHGC-12848; (m) stSG4039

Haplotype analysis with microsatellite markers

The boundaries of the TOC minimal region were redefined by re-evaluating the haplotype data in the light of the new marker order data. Sequence data demonstrates a definitive marker order of D17S785D17S1817D17S801D17S2192D17S751D17S722. The TOC gene is thus positioned between the markers D17S785 and D17S751 (Figures 1 and 2). The distal limit is set between D17S2192 and D17S751 by UK individuals VI:3 and V:107, while the proximal limit is set by a recombination event between D17S785 and D17S1817 in affected UK individual IV:57, or possibly between D17S1817 and D17S801 in the affected UK individual V:107. The proximal limit is unclear because of homozygosity in the mother of V:107 (IV:57) at D17S1817 but it is unlikely that a second recombination event has occurred in V:107 so close to the D17S2192D17S751 crossover that sets the distal limit.

Figure 2

Partial pedigrees of the UK (a) and US (b) families showing recombinant individuals and their immediate relatives. The affected haplotype for each family is given in column 2 and is enclosed within a bold box in affected individuals showing the extent of their inheritance of this haplotype and highlighting the sites of recombination. Shaded boxes across the whole family show the extent of the minimal region in each family

Identification and analysis of candidate genes

ESTs analysis

One hundred and eleven ESTs were identified as mapping in the genetic region D17S1352D17S836 on the Genebridge G4 and Stanford G3 radiation hybrid maps, using the GeneMap website (URL: This broad genetic region overlaps the TOC minimal region by 9.5 cM proximally and 8 cM distally, and allows the investigation of ESTs that may not have been accurately mapped using radiation hybrids. ESTs were mapped using DNA from clones constituting the minimal contig (RPCI-1 PAC 96G20 and RPCI-11 BACs 318A15 and 835H2) (Figure 1b), together with adjacent overlapping clones.

Twelve out of 111 ESTs mapped to at least one of the three clones forming the TOC minimal contig (Figure 1c), while the remaining 98 ESTs mapped outside the minimal contig. Of these 12 ESTs, three were associated with the known genes SThM (stSG50865), SFRS2 (SHGC-12848) and ET (stSG4039), and one with a partially characterized gene KIAA0585 (stSG48272).

Human genome sequence data analysis

The sequence gap in clone 666A8 on the public databases has been filled by sequence walking (Risk et al., 2002) and 11 fully or partially characterized genes were observed to be located on the two proximal, completely sequenced, clones, 666A8 and 318A15 using the NIX suite of genomic sequence annotation programs (Figure 1c). AANAT, SThM, KIAA0585, SFRS2 and ET had been previously identified by EST mapping and from other publications (Kuhlenbäumer et al., 1999; Kalikin et al., 1999) but six additional candidate genes were identified as mapping to the TOC minimal region for the first time: UBE2-230k, R7L, myoglobin-like, ST6GalNAcI, MXRA7-like and CG5013-like. Two of the GeneMap EST sequences, stSG161 and Bda94c06, were shown to be contained in the newly described CG5013-like and MXRA7-like genes, respectively (Figure 1c).

Many of the genes identified were considered as candidate TOC genes on the basis of known or inferred function. SFRS2 (or SC35) is a splicing factor required for spliceosome assembly and is expressed in a variety of tissues (Fu and Maniatis, 1992) while ET is transcribed from the opposite strand to SC35 in a tissue-specific manner, is alternatively spliced and may also be involved in splicing (Sureau et al., 1997). Human sialyl transferases like SThM and ST6GalNAcI are alternatively spliced in a tissue-specific manner and have been implicated in the inhibition of keratinocyte cell line differentiation and control of proliferation (Paller et al., 1993; Taniguchi and Matsumoto, 1998). MXRA7 is a matrix remodelling associated gene and thus may be implicated in the development of cancer (Walker MG, unpublished). UBE2-230k is a ubiquitin-conjugating enzyme (Yokota et al., 2001) and CG5013-like is homologous to a Drosophila gene showing similarity to zinc metalloproteases (Gadfly database: URL:http://hedgehog/ Little is known about the function of the KIAA0585 gene except that it shows similarity to a phosphatidylserine receptor (Nagase et al., 1998), whilst AANAT is an arylalkylamine N-acetyltransferase involved in melatonin formation and may not be a strong candidate (Zheng et al., 2001). However, since polymorphisms in the arylamine N-acetyltransferases 1 and 2 (NAT1, NAT2) have been implicated in oesophageal and breast cancer (Shibuta et al., 2001; Geylan et al., 2001), and RT–PCR data has indicated that AANAT is ubiquitously expressed in a variety of tissues including oesophagus and skin (data not shown), analysis of this gene was undertaken. The ribosomal and myoglobin-like genes were discounted as the TOC causative locus on the basis of their known or presumed functions.

In addition, NIX analysis of this 355 kb proximal sequence identified a previously undescribed gene of 16 exons (Figures 1c and 3b). The exon structure and boundaries were confirmed by alignment of genomic and EST sequences, many of which covered more than one exon. The more recent database submission of a cDNA sequence covering 13 out of 16 exons has confirmed our 3′ predicted structure (FLJ22341).

Figure 3

Exon–intron structure of nine of the 10 candidate genes located within the TOC minimal region. Boxes represent exons; lines represent introns. Accession numbers of reference sequences (r), mRNAs (m), Unigene clusters (u) and ESTs (e) and/or Trembl (t) alignments upon which this data is based are given in square brackets below. (a) AANAT [NM_001088 (r)]; (b) FLJ22341 [BI83911(e) (exons 1–4), Hs.25485(u) (exons 4–18), NM_024599 (r) (exons 6–18)]; (c) SThM [NM_006456 (r)]; (d) ST6GalNAcI [XM_036780 (r)]; (e) MXRA7-like [Hs.350199 (u) (exons 1–3, 5), AW888225 (e) (exons 1–5)]; (f) KIAA0585 [XM_036784 (r)]; (g) CG5013-like [Hs.74655 (u) (exons 1, 3–5), TR:Q9D1J5 (t) (exons 1–5)]; (h) SFRS2 [NM_003016 (r)]; (i) ET [BI084978 (e) (exons 1–2), AF015186 (m) (exons 2–17)]

Sequence data from the two distal clones, 234P2 and 87G24, is available as 13 and two pieces respectively. Many of the fragments contain clone-end and microsatellite markers that allow their provisional ordering and are of a size amenable to NIX analysis. However, only one additional gene showing homology to a characterized gene was identified as mapping to this approximately 200 kb region (Figure 1c). This gene showed homology to MGAT5 (Saito et al., 1994) at its 3′ end (exons 6–17) but appeared to possess an additional five 5′ exons. MGAT5-induced beta 1,6 branching on N-glycans may control T-cell receptor recruitment to the site of antigen presentation and dysresolution of MGAT5 may thus increase susceptibility to autoimmune diseases (Demetriou et al., 2001). This gene was thus excluded as a candidate.

After analysis of the NIX data from all four clones, an additional 20–25 putative genes were identified as overlapping ESTs with associated exon/gene predictions, but with no homology to known genes. BLAST analysis of the sequence data also identified 10–15 ESTs and/or Unigene clusters that did not appear to be associated with genes forecast by the exon prediction programs and showed no homology to known genes. In addition, 25–30 putative genes were predicted by at least three out of 12 exon and gene prediction programs that did not appear to be associated with BLAST matches to coding sequence. Studies investigating the expression patterns of these partial ‘genes’ are currently being undertaken in order to determine which are actually expressed and to obtain some data regarding size of transcript and tissue distribution.

Mutation analysis

The exon–intron structure of all candidate genes (except the ribosomal, myoglobin-like and MGAT5-like genes) was determined from NIX output and confirmed by alignment of cDNA/EST data with genomic sequence using the Sequencher program (GeneCodes Corp.) (Figure 3).

Mutation analysis covered all coding regions of these genes except for the UBE2-230k ubiquitin–conjugating enzyme gene, where only the promoter region is located within the TOC minimal region. Promoter regions of the four strongest candidate genes (SThM, ST6GalNacI, UBE2-230k, and FLJ22341) were also determined and analysed. Initial mutation screens comprised one affected member of the UK family, one affected US family member, and one unaffected individual from the UK family. This preliminary sample selection was deemed capable of identifying polymorphisms in either family that could then be further analysed in additional family members in order to ascertain the disease-specificity of the polymorphism. This second sample selection comprised at least five affected and two unaffected family members from both the UK and US families and included those individuals who demonstrated defining recombinations in the TOC region (UK individuals IV:57, VI:3 and V:107). The method of choice for mutation analysis was sequencing, but the sequence traces from exon 2 of FLJ22341, exon 1 of the MXRA7-like gene and exon 1 of KIAA0585 were of insufficient quality to determine polymorphisms accurately. SSCP analysis was used to assess these exons for mutations.

No tylosis-specific mutations were observed during these investigations. However, one SNP was detected in FLJ22341, two in SThM, one in ST6GalNAcI and two in the MXRA7-like gene (Table 1). None of these SNPs were informative in haplotype analysis of the UK and US families (Figure 2) and no coding changes are associated with the sequenced polymorphisms.

Table 1 Intragenic polymorphisms in candidate TOC genes

Expression analysis

The expression patterns of the four strongest candidate genes (SThM, ST6GalNacI, UBE2-230k, and FLJ22341) were analysed in a panel of tissues and in keratinocyte RNA from two UK family members. PCR primers were designed in exonic sequence to cover the whole of each gene or one of two internal fragments (details on request). No alternative spliced products were identified for any of these genes in any of the tissues or the cell lines examined (data not shown).


The recent emphasis on physical mapping to supplement genetic mapping of disease genes is reaching its climax as the Human Genome Project sequencing phase nears completion. However many breaks still remain in the sequencing data, thus ensuring that physical mapping using STS data from contigs of cloned human DNA remains necessary to bridge these gaps and to obtain cloned DNA for additional research. Two published physical maps of chromosome 17q25 also illustrate the disadvantages of relying predominantly on YAC data to order STS markers in the absence of sequence data. Thus, the map produced by Kalikin et al. (1999) contains many YAC clones with ‘missing’ STSs indicative of chimaeric clones, whilst the paper by Kuhlenbäumer et al. (1999) details the discarding of YAC clones in favour of more stable BAC clones. The approach is vindicated in this example by the partial sequence data that is currently available which corresponds with Kuhlenbäumer's marker order and places D17S722 distal to D17S1817 rather than proximal to D17S785 as in the Kalikin map. This new mapping data, coupled with the partial sequence of the region has been used in our study to refine the minimal region on chromosome 17q25 containing the tylosis with oesophageal cancer (TOC) gene and to identify and analyse candidate genes.

Haplotype data in this paper was based on the analysis of microsatellite markers and placed the TOC locus between D17S785 and D17S751, a genetic map distance of 2 cM. A physical distance cannot yet be assigned to this interval owing to the incompleteness of the Human Genome sequencing data in this region, but it is probably approximately 0.5 Mb, based upon the known sizes of sequenced fragments of clones 666A8 (206 kb), 318A15 (163 kb), 234P2 (157 kb) and 87G24 (102.5+90.5 kb), the known or estimated size of overlap between clones and the positions of the limiting microsatellite markers. Further haplotype analysis incorporating data from six intragenic SNPs in four of the candidate genes did not provide any additional refinement of the TOC area, although the genes are clearly located within the minimal region. Other SNPs can now be identified from sequence data, together with additional microsatellite markers, which will possibly enable us to further reduce the area under investigation. In addition, the availability of sequence data permitted the focussed large-scale analysis of this region by restriction endonuclease digestion and Southern blotting, but no gross alterations were observed in the TOC region in keratinocyte DNA isolated from two UK family members (data not shown).

One hundred and eleven ESTs identified from the GeneMap website (Deloukas et al., 1998) have also been investigated and 12 were located between our limiting markers. Four were shown to be associated with known genes and three with novel genes, whilst five have unknown function. Computer-based annotation of the available sequence data has identified 12 known genes and predicted a novel 16 exon gene. The NIX suite of programs was chosen in preference to other similar packages because of its ease of use and user friendly graphical output (Jones et al., 2002). More recently, annotation of the HGP sequence data on the Ensembl and GoldenPath websites (URL: and URL: has demonstrated that our sequence annotation did not overlook any genes within this region. Ten of these complete genes were investigated as the causative TOC gene using mutation analysis of family members. Many of these genes were good candidates based on their functional characteristics or homology with other proteins. However, no tylosis-specific mutations were observed in either exons or promoter regions of these genes. In addition, analysis of expression patterns and transcript sizes of these genes by RT–PCR failed to show any disease-associated or affected tissue-specific changes. The genes MLL septin-like fusion (MSF) or ovarian/breast septin, survivin (SVV), and deleted in multiple human cancer (DMC1), together with three other novel genes, have been recently described in the literature as candidates for the TOC locus (Osaka et al., 1999; Russell et al., 2000; Li and Altier, 1999; Islam et al., 2000; Harada et al., 2001). However, the refinement of the minimal region detailed in this paper excludes all of these on the basis of position, and they have also been excluded by BLAST analysis of the TOC region sequence. In addition, the candidate genes MAFG, GPR14, GPRC5C and MRPL12, located to the 17q25 cytogenetic region, have also been excluded as the TOC gene on the basis of sequence analysis (Blank et al., 1997; Protopopov et al., 2000; Robbins et al., 2000; Marty et al., 1997).

Other, weaker, gene predictions and individual ESTs without gene predictions will also require avenues of research that will enable us to determine whether they are suitable for further analysis. These data imply that this region of chromosome 17 is very gene-rich and will present research challenges for many months. On average, the gene density of the human genome has been estimated to be one per 20 kb for gene-rich regions and one per 200 kb for gene-poor regions (Fields et al., 1994). We have identified at least 13 genes and 50–75 putative genes in approximately 500 kb of sequence. Some of the predicted genes will almost certainly prove to be artefacts, but we estimate that at least 30 genes will eventually be shown to map to these four clones, two of which show evidence of being gene-rich segments.

These results demonstrate that the best opportunities for the identification of the TOC gene lie in the completion of the Human Genome Project sequencing of the TOC region and its further analysis to identify candidate genes. This approach should be coupled with the identification of additional polymorphisms (microsatellites and SNPs) within the minimal region that may be used in haplotype analysis in order to reduce the length of the area under investigation. Candidate genes identified from this data will undergo a number of analyses, including mutation analysis of coding and promoter regions and expression analysis in order to determine which is the causative TOC locus. A systematic approach is needed to identify the TOC gene and this may also include the investigation of the expression of these genes in sporadic squamous cell oesophageal carcinomas displaying LOH of 17q25. The importance of this gene in a larger population than those afflicted with the familial disease is indicated by the association of the genomic region containing the TOC gene with sporadic oesophageal tumours, Barrett's adenocarcinoma of the oesophagus, and primary breast cancers (von Brevern et al., 1998; Iwaya et al., 1998; Dunn et al., 1999; Kalikin et al., 1996; Fukino et al., 1999).

Materials and methods

Physical mapping

RPCI-11 BAC library clones were obtained from Research Genetics Inc. (Huntsville, USA). RPCI-1 PAC library clones were obtained from UK HGMP Resource Centre (Hinxton, UK). These clones were grown using standard conditions and DNA extracted using common methods. STS and EST marker content was determined by PCR (details on request). PCR products were then analysed on 2% agarose gels and visualized by ethidium bromide staining.

Haplotype analysis

Microsatellite markers

Forty-seven tylotic individuals (five with oesophageal cancer), 28 unaffected family members and six unaffected spouses from the UK family together with 45 tylotic individuals (four with oesophageal cancer), 14 unaffected family members and one unaffected spouse from the US family were genotyped using up to 20 polymorphic microsatellite repeat markers spanning 7 cM in the TOC region (D17S929D17S939). Four of these markers had not been previously investigated in these families (D17S602, D17S2192, D17S751 and D17S663). Five of the 47 UK affected individuals, two out of 28 UK unaffected family members and six out of 45 US affected family members had been previously shown to possess a recombination event in this region (Kelsell et al., 1996 and unpublished data). Genotyping was performed on 100 ng genomic DNA in a standard 25 μl PCR reaction at annealing temperatures of between 52–62°C, depending upon the primer pair under investigation (details on request). PCR products were analysed on 8–10% non-denaturing polyacrylamide gels and visualized by the silver staining method of Gottleib and Chavko (1987). This study was approved by the Regional Health Authority Ethical Committee.


Individuals from the UK and US families previously shown to possess a recombination event that may aid in the definition of the TOC region (Kelsell et al., 1996 and unpublished), together with distant family members who identified the allele associated with the affected haplotype, were genotyped for six SNPs by sequence analysis or SSCP.


Existing haplotypes were refined using the definitive order of microsatellites determined by sequence and physical mapping where possible (D17S1603D17S722), or the most statistically robust radiation hybrid and genetic mapping data for other areas. Genetic recombinations were identified. In meioses where phase could not be determined unequivocally, haplotypes showing the minimal number of recombination events were constructed.

ESTs identified on GeneMap

All ESTs mapping with the D17S1352D17S836 region on both the Genebridge G4 and the Stanford G3 panels were identified from the GeneMap web site (Deloukas et al., 1998) (URL: PCR primer details were obtained from this web site and synthesized either by Research Genetics Inc. or by MWG Biotech UK Ltd.

Analysis of genomic sequence data

The NIX suite of programs running at the UK HGMP Resource Centre integrates exon prediction, gene prediction and gene element prediction programs with BLAST results to assist in the identification of transcripts within genomic sequence (URL: These programs were used to analyse human genome sequence data from clones covering the TOC region. Genomic DNA was analysed in 20 kb segments that overlapped by 2 kb in order to allow gene prediction programs to identify genes spanning two segments. Exons predicted by at least three of the exon prediction and/or gene identification programs that also displayed BLAST matches to ESTs, Unigene clusters, mRNAs and/or proteins were considered to be candidate genes.

Mutation analysis

Gene structure

The exon–intron structure of candidate genes was determined from NIX output and confirmed by alignment of cDNA/EST data with genomic sequence using the Sequencher program (GeneCodes Corp.).

Sequencing template preparation

Primers were designed to cover all coding and promoter regions of candidate genes in 1 kb sequencing templates and nested sequencing primers were IRD-700 labelled (details on request). The primer 3 program at the Whitehead Institute (URL: was used for primer design, and primer synthesis was by MWG Biotech Ltd or Sigma-Genosys Ltd. Sequencing templates were prepared by PCR from 200 ng of genomic DNA using ELONGASE® Enzyme (GIBCO BRL) according to the manufacter's instructions. PCR products were assessed on a 1.5% agarose gel to ensure that a single product had been produced and templates for sequencing were cleaned using QIAquickTM Gel Extraction/PCR purification kits (Qiagen UK Ltd.).


PCR templates were sequenced using the Sequitherm EXCELTM II DNA sequencing kit-LC and loaded onto a 7% gel for analysis on a LI-COR DNA Sequencer 4200. Sequencing results were obtained for both directions from each of two separate templates and were analysed using the LI-COR base-calling software and mutation analysis was assessed using Sequencher. In addition, the gel images from each reaction were scanned by eye as heterozygous changes are visible as additional bands using this technology (separate lanes for each ddNTP).

SSCP analysis

SSCP analysis of three exons was also undertaken. PCR was performed as for microsatellite markers using primers located in the introns adjacent to exons that produced a product of no more than 300 bp (details on request). PCR products were then analysed under four gel conditions: 8 or 10% acrylamide at room temperature, and 8 or 10% acrylamide at 4°C, all with 5% glycerol. Gels were silver stained as before.


Expression patterns of candidate genes was determined by two-step RT–PCR in a bank of commerically available RNA comprising normal oesophagus, squamous cell oesophageal cancer, skin, brain, liver, breast, prostate, colon, ovary, stomach, foetal brain, and foetal liver, together with RNA from keratinocytes isolated from two UK family members. RNA samples that had not been DNAse treated by the manufacturer were subjected to DNase treatment using the DNA Free kit (Ambion). Reverse transcription was accomplished using the RETROscript kit (Ambion) and this step was checked using actin PCR primers before PCR of candidate genes was performed. PCR products were visualized by agarose gel electrophoresis.


  1. Blank V, Knoll JHM, Andrews NC . 1997 Genomics 44: 147–149

  2. Brown LM, Hoover R, Silverman D, Baris D, Hayes R, Swanson GM et al . 2001 Am. J. Epidemiol. 153: 114–122

  3. Deloukas P, Schuler GD, Gyapay G, Beasley EM, Soderlund C, Rodriguez-Tome P et al . 1998 Science 282: 744–746

  4. Demetriou M, Granovsky M, Quaggin S, Dennis JW . 2001 Nature 409: 733–739

  5. Dunn J, Garde J, Dolan K, Gosney JR, Sutton R, Meltzer SJ, Field JK . 1999 Oncogene 18: 987–993

  6. Ellis A, Field JK, Field AE, Friedmann PS, Fryer A, Howard P et al . 1994 Eur. J. Cancer Oral Oncol. 30B: 102–112

  7. Fields C, Adams MD, White O, Venter JC . 1994 Nature Genet. 7: 345–346

  8. Franceschi S, Bidoli E, Negri E, Zambon P, Talamini R, Ruol A et al . 2000 Int. J. Cancer 86: 626–631

  9. Fu XD, Maniatis T . 1992 Science 256: 535–538

  10. Fukino K, Iida A, Teramoto A, Sakamoto G, Kasumi F, Nakamura Y, Eni M . 1999 Genes Chromosomes Cancer 24: 345–350

  11. Geylan YS, Dizbay S, Guray T . 2001 Neoplasma 48: 108–111

  12. Gottleib M, Chavko M . 1987 Anal. Biochem. 165: 33–37

  13. Harada H, Nagai H, Tsuneizumi M, Mikami I, Sugano S, Emi M . 2001 J. Hum. Cancer 46: 90–95

  14. Hennies HC, Hagedorn M, Reis A . 1995 Genomics 29: 537–540

  15. Howel-Evans W, McConnell RB, Clarke CA, Sheppard PM . 1958 Quart. J. Med. 27: 413–429

  16. International Human Genome Sequencing Consortium. 2001 Nature 409: 860–921

  17. Islam A, Kageyama H, Takada N, Kawamato T, Takayasu H, Isogai E et al . 2000 Oncogene 19: 617–623

  18. Iwaya T, Maesawa C, Ogasawara S, Tamura G . 1998 Gastroenterology 114: 1206–1210

  19. Jones J, Field JK, Risk JM . 2002 Int. J. Oncol. 20: 697–705

  20. Kalikin LM, Qu X, Frank TS, Caduff RF, Svoboda SM, Law DJ, Petty EM . 1996 Genes Chromosomes Cancer 17: 64–68

  21. Kalikin LM, George RAV, Keller MP, Bort S, Bowler NS, Law DJ et al . 1999 Genomics 57: 36–42

  22. Kelsell DP, Risk JM, Leigh IM, Stevens HP, Ellis A, Hennies HC et al . 1996 Hum. Mol. Genet. 5: 857–860

  23. Kuhlenbäumer G, Schirmacher A, Mueleman J, Tissir F, Del-Favero J, Stogbauer F et al . 1999 Genomics 62: 242–250

  24. Li FZ, Altier DC . 1999 Cancer Res. 59: 3143–3151

  25. Marty L, Taviaux S, Fort P . 1997 Genomics 41: 453–457

  26. Munoz N, Castellsague X . 1994 Eur. J. Gastroenterol. Hepatol. 6: 649–655

  27. Nagase T, Ishikawa K, Miyajima N, Tanaka A, Kotani H, Nomura N, Ohara O . 1998 DNA Res. 5: 31–39

  28. Osaka M, Rowley JD, Zeleznikle NJ . 1999 Proc. Natl. Acad. Sci. USA 96: 6428–6433

  29. Paller AS, Arnsmeier SL, Alvarez-Franco M, Bremer EG . 1993 J. Invest. Dermatol. 100: 841–845

  30. Parkin DM, Pisani P, Ferlay F . 1993 Int. J. Cancer 54: 1–13

  31. Protopopov A, Kashuba V, Podowski R, Gizatullin R, Sonnhammer E, Wahlestedt C, Zabarovsky ER . 2000 Cytogenet. Cell Genet. 88: 312–313

  32. Risk JM, Whittaker J, Fryer A, Ellis A, Shaw JM, Field EA et al . 1994 Nature Genet. 8: 319–321

  33. Risk JM, Ruhrberg C, Hennies HC, Mills HM, Di Colandrea T, Evans KE et al . 1999 Genomics 59: 234–242

  34. Risk JM, Langan JE, Rowbottom L, Jones J, Liloglou T, Field JK . 2002 Biotechniques (in press)

  35. Robbins MJ, Michalovich D, Hill J, Calver AR, Medhurst AD, Gloger I et al . 2000 Genomics 67: 8–18

  36. Russell SHE, McIlhatton MA, Burrows JF, Donaghy PG, Chanduloy S, Petty EM et al . 2000 Cancer Res. 60: 4729–4734

  37. Saito H, Nishikawa A, Gu J, Ihara Y, Soejima H, Wada Y et al . 1994 Biochem. Biophys. Res. Commun. 198: 318–327

  38. Shibuta J, Eto T, Kataoka A, Inoue H, Ueo H, Suzuki T et al . 2001 Am. J. Gastroenterol. 96: 3419–3424

  39. Stevens HP, Kelsell DP, Bryant SP, Bishop DT, Spurr NK, Weissenbach J et al . 1996 Arch. Dermatol. 132: 640–651

  40. Sureau A, Soret J, Guyon C, Gaillard C, Dumon S, Keller M et al . 1997 Nucleic Acids Res. 25: 4513–4522

  41. Taniguchi A, Matsumoto K . 1998 Biochem. Biophys. Res. Commun. 243: 177–183

  42. Venter JC, Adams MD, Myers EW, Li PW, Mural RJ, Sutton GG et al . 2001 Science 291: 1304–1351

  43. von Brevern M, Hollstein MC, Risk JM, Garde J, Bennett WP, Harris CC et al . 1998 Oncogene 17: 2101–2105

  44. Yokota T, Nagai H, Harada H, Mine N, Terada Y, Fujiwara H et al . 2001 Gene 267: 95–100

  45. Zheng W, Scheignet KA, Ho AK, Cole PA . 2001 Chem. Biol. 8: 379–389

Download references


The authors wish to thank Dr G Kuhlenbäumer, Professor C Van Broeckhoven and Professor E Petty for their release to us of pre-publication data. This work was supported by grants from the Cancer Research Campaign (CRC) (grant numbers SP2384/0101 and SP2384/0201), the North West Cancer Research Fund, Isle of Man Anti Cancer Association, and a University of Liverpool PhD studentship.

Author information



Corresponding author

Correspondence to Janet M Risk.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Risk, J., Evans, K., Jones, J. et al. Characterization of a 500 kb region on 17q25 and the exclusion of candidate genes as the familial Tylosis Oesophageal Cancer (TOC) locus. Oncogene 21, 6395–6402 (2002).

Download citation


  • oesophageal cancer
  • transcript map
  • tylosis

Further reading


Quick links