Allelic imbalance at the DNA mismatch repair loci, hMSH2, hMLH1, hPMS1, hPMS2 and hMSH3, in squamous cell carcinoma of the head and neck

Oral Oncology 39 (2003) 115–129 www.elsevier.com/locate/oraloncology

Allelic imbalance at the DNA mismatch repair loci, hMSH2, hMLH1, hPMS1, hPMS2 and hMSH3, in squamous cell carcinoma of the head and neck J. Nunna, S. Naginib, J.M. Riska, W. Primec, P. Maloneya,d, T. Lilogloua,d, A.S. Jonese, S.R. Rogersf, J.R. Gosneyc, J. Woolgarg, J.K. Fielda,d,* a

Molecular Genetics and Oncology Group, Department of Clinical Dental Science, The University of Liverpool, Liverpool L69 3BX, UK b Department of Biochemistry, Faculty of Science, Annamalai University, Tamilnadu, India c Department of Pathology, The University of Liverpool, Liverpool L69 3BX, UK d The Roy Castle International Centre for Lung Cancer Research, Liverpool L3 9TA, UK e Department of Otorhinolaryngology, The University of Liverpool, Liverpool L69 3BX, UK f Maxillofacial Unit, Fazakerley Hospital, Liverpool, UK g Department of Oral Pathology, The Dental School, The University of Liverpool, Liverpool L69 3BX, UK Received 8 April 2002; accepted 30 April 2002

Abstract Background: Squamous cell carcinoma of the head and neck (SCCHN) is one of the 10 most frequently occurring cancers in the world. Defective mismatch repair, as exhibited by the phenomenon of microsatellite instability, has been observed in SCCHN although no reports of mismatch repair gene mutations or altered protein expression have been published. In a variety of microsatellite instability (MSI) positive cancers where mutations in the mismatch repair (MMR) genes were not observed, allelic imbalance at the loci of the MMR genes was prevalent. Objective: To investigate whether allelic imbalance at the MMR genetic loci contributes to the development of SCCHN. Materials and Methods: 35 matched normal/tumour SCCHN pairs were studied using 29 microsatellite markers located within and adjacent to six known DNA mismatch repair genes. In addition, mutational analysis and protein expression of hMSH2 and hMLH1 were investigated. Results and conclusions: We demonstrated that 36 and 17% of the analysed SCCHN specimens exhibited allele imbalance at the hMLH1 and hMSH3 genetic loci, respectively. Allelic instability at these two loci was found to be correlated with the MSI status of the SCCHN tumours. Allelic instability was found to be uncommon at the other MMR gene loci analysed. One mutation was found in hMSH2 and none in hMLH1 in this series of tumours. 23 of 24 (96%) of the examined SCCHN tumours showed reduced expression of either hMSH2 or hMCH1 genes. Allelic instability in the MMR genes, hMLH1 and hMSH3, is proposed to be involved in the aetiology of SCCHN tumours. # 2002 Elsevier Science Ltd. All rights reserved. Keywords: Carcinoma; Squamous cell; Microsatellite markers; Head and neck cancer; Allelic imbalance; Mutation analysis; DNA; DNA repair; Loss of heterozygosity; Immunohistochemistry; hMLH1; hMSH3

1. Introduction Oral cancer is one of the 10 most frequently occurring cancers in the world [1,2] but this figure disguises the global geographical differences that exist its occurrence [3–6]. In all parts of the world, the incidence of oral cancer is higher in males compared with females [2,5–7], and becomes more frequent with advancing age, usually demonstrated by a sharp rise after 40 years of age [3,5]. * Corresponding author. Tel.: +44-151-794-8900; fax: +44-151794-8989. E-mail address: j.k.fi[email protected] (J.K. Field).

Human cancers develop by accumulating a range of somatic genetic changes throughout their progression [8], however the molecular basis for these changes remains unclear in the majority of cancers. It has been postulated that genetic damage affecting the majority of autosomal chromosome arms, including activation of oncogenes and/or inactivation of tumour suppressor genes (TSGs), possibly working in conjunction with an impaired capacity of DNA repair mechanisms, may be a factor in the development of squamous cell carcinoma of the head and neck (SCCHN) [9–16]. To date, there are a large number of reports demonstrating that a significant proportion of the genome of

1368-8375/03/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved. PII: S1368-8375(02)00028-3

116

J. Nunn et al. / Oral Oncology 39 (2003) 115–129

human tumours contains losses or duplication of certain loci. The repeated observation of allelic imbalance of a specific chromosome marker in cells from a particular tumour types suggests the presence of a closely mapping TSG, the loss of which is involved in tumour pathogenesis [17]. Allelic imbalance resulting in loss of one allele has also been termed ‘loss of heterozygosity’ (LOH). Extensive allelotype and loss of heterozygosity studies have been used to identify key chromosomal regions in the human genome that encode TSGs [18–21]. The largest of these studies, in SCCHN, analysed 80 specimens using 145 microsatellite markers on 39 chromosomal arms [20]. This study identified a number of key chromosomal regions including chromosome 1, 2p, 3p, 5q, 6, 8, 9, 11q, 13q, 17, 18q, and 19q involved in the development of SCCHN. In addition, the genetic material in every living cell is constantly damaged by environmental mutagens and replication errors and if this damage is not repaired it may lead to the eventual death of the cell or its malignant transformation. The efficient and faithful transmission of genetic information to the next generation is of paramount importance for the survival of the daughter cells. Loss of any of the DNA repair mechanisms increases the risk of the development of cancers in humans. Defective human DNA mismatch repair mechanism manifests itself as the phenomenon termed ‘microsatellite instability’ (MSI), and is deemed a separate entity to allelic imbalance or loss of heterozygosity. This event was first reported in Hereditary Non-Polyposis Colorectal Cancer (HNPCC) tumours [22–24], later observed in sporadic colorectal cancers and a variety of other cancers [23–33]. It has been suggested that, like TSGs, inactivation of human DNA mismatch repair genes requires two steps [34–36]. The first event is assumed to be a deleterious mutation and the second mutational event was initially reported as allelic imbalance at the hMLH1 locus in HNPCC patients [37], and later in sporadic colorectal cancer (CRC), sporadic breast cancer, non-small cell lung cancer (NSCLC) and cervical squamous cell carcinoma [38–44]. It has been considered that, in microsatellite instability positive (MSI+) tumours, LOH might occur frequently at the DNA mismatch repair (MMR) gene loci, thereby inactivating one MMR allele in a way similar to allele loss at tumour suppressor gene loci. Seven human mismatch repair genes have been located, all of which are homologous to known DNA MMR genes found in the bacteria, Escherichia coli, and the yeast, Saccharomyces cerevisiae. The human homologue of the bacterial MutS gene, hMSH2, was the first MMR gene to be mapped and isolated on chromosome 2p16 [45, 46]. Three human homologues of the MutL gene; hMLH1, hPMS1 and hPMS2 are located at chromosome 3p21, 2q31–33 and 7p22, respectively [47–49]. Since then two new MutS homologues have been

reported, hMSH3 and hMSH6, located at 5q11–13 and, within 1 mega base of hMSH2, at 2p16, respectively [50–53] and a new MutL homologue, hMLH3, is located at 14q24.3 [54]. To date no study has investigated whether allelic imbalance of human DNA mismatch repair genes is involved in the development of SCCHN. Previous SCCHN allelotype studies have only analysed a small number of microsatellite markers on each chromosome, therefore no in-depth knowledge of allelic imbalance at the loci for DNA mismatch repair genes exists. This study details the investigation of the involvement of allelic imbalance at six mismatch repair gene loci in SCCHN, together with mutational analysis and protein expression of hMSH2 and hMLH1.

2. Materials and methods 2.1. Patients Thirty-five squamous cell carcinoma of the head and neck specimens were obtained from the Department of Otorhinolarynology, Royal Liverpool University Hospital, and the Maxillofacial Unit, Walton Hospital, Liverpool, UK. Specimens were obtained at the time of surgery; part of the tissue was snap frozen in liquid nitrogen and stored long term at 80  C, while the remaining tissue was fixed routinely in 10% buffered Formalin for histological examination. Clinical data was available from the case notes for the majority of the patients investigated and included the sex of the patient, the site of the tumour, the patient’s date of birth, the age of the patient at diagnosis, the TNM clinical staging of the tumour, the histopatholgy of the tumour, the patient’s smoking and alcohol intake history, whether the tumour was a primary or recurrent tumour, follow up, and the fate of the patient. Ethical approval had been obtained for this study. Subsequently frozen sections were prepared from the stored tissue and micro-dissected to provide tissue samples with over 60% tumour material and matched normal tissue; in most cases the DNA was prepared from samples with over 70% tumour cells. Genomic DNA was extracted, both normal and tumour tissue, using the Nucleon II DNA extraction kit (Nucleon Biosciences, Scotlab, Scotland), following the manufacturers instructions. Genomic DNA samples were stored at 4  C. 2.2. Microsatellite analysis Matched normal and tumour DNA pairs were analysed by PCR at 29 polymorphic microsatellite loci: chromosome 2p15–16, D2S177, D2S119, D2S391, D2S288, D2S123, D2S378; 3p21.3, D3S1609, D3S1619, D3S1561, D3S1612, D3S1611, D3S1260; 2q31–33,

J. Nunn et al. / Oral Oncology 39 (2003) 115–129

D2S300, D2S152, D2S118, D2S116, D2S155, D2S325; 5q11–13, D5S118, D5S637; D5S39, HMGCR, D5S112, D5S107; 7p22, D7S531, D7S472, D7S511, D7S481, and D7S513 (Fig. 1). The primers were obtained from Research Genetics (Huntsville, USA) and Isogen (Amsterdam). Amplification conditions were 25 cycles of 94  C for 30 s, 45–60  C for 30s and 72  C for 30 s in a 25 ml reaction volume containing 100 ng of genomic DNA, 200 mM of each dNTP, 10 pmol of each forward and reverse primer, 2.5 ml of 10PCR buffer [670mM Tris–HCl pH 8.5, 166 mM ammonium sulphate, 67 mM magnesium chloride, 1.7 mg/ml bovine serum albumin (BSA), 100 mM b-mercaptoethanol, 1% (w/v) Triton X100] and 0.2 units of Taq polymerase (Bioline) A 10 ml

117

aliquot of the PCR product was analysed on a 10% polyacrylamide gel (Accugel 19:1) and visualised by silver staining. Two observers (JN and JR) independently read the silver stained electrophoresis gels by eye to generate the results. To negate experimental variation, differences in the number and mobility of bands between tumour and comparable normal samples were reproduced by three separate, independent amplifications of different DNA preparations. Loss of heterozygosity, or allelic imbalance, was scored by direct visual comparison of the allelic ratios of the normal and tumour specimens. A reduction of at least 50% in the intensity of one allele in the tumour was considered as LOH. For any ambiguous

Fig. 1. Location of microsatellite markers used in the analysis of the human DNA mismatch repair genes. Integration of published data, genetic and BAC data from several sources provided the microsatellite marker chromosomal locations [46,60–63].

118

J. Nunn et al. / Oral Oncology 39 (2003) 115–129

results densitometry using the UVP Laboratory Imaging and Analysis system (UVP Ltd., Cambridge, UK) was undertaken to provide a more accurate result, with a cut off allelic imbalance value of 50%. A tumour was designated as having no LOH if it did not show allelic imbalance at any of the markers and was informative for at least two of the markers at each locus. Microsatellite instability was scored as the presence of extra band(s) and/or band shift in the tumour DNA specimen that was not present in the corresponding normal DNA specimen. If microsatellite instability occurred at any marker, then the result at that locus was excluded from the LOH analysis. 2.3. Mutation analysis of hMSH2 and hMLH1 Mutational analysis was undertaken on two of the DNA mismatch repair genes, hMSH2 and hMLH1. Single stranded confirmation polymorphism and heteroduplex analysis was undertaken in the 35 SCCHN specimens that were investigated by microsatellite analysis. Three cell lines were selected as controls for mutational analysis. HCT116, a human colon carcinoma cell line, contains a hemizygous C–– > A mutation at codon 252 in exon 9 of the hMLH1 gene, which results in a truncated protein [48,55]. Lo Vo, a human colon adenocarcinoma cell line, contains a deletion of exons 3–8 of the hMSH2 gene [56–58]. HCT-15, a human colon adenocarcinoma cell line, contains no mutations in either of hMSH2 or hMLH1 genes. HCT-15 has been reported to contain three mutations, one in the exonuclease domain of the polymerase delta gene, and two in the hMSH6 gene [52,53,59]. The three-monolayer cell lines were established and propagated. The cell lines were obtained as frozen cultures from the European Collection of Animal Cell Cultures (ECACC, Salisbury, UK). All cell lines were maintained in their optimal media supplemented with 10% fetal bovine serum, 1% penicillin and 1% streptomycin (Sigma). DNA from the cell lines was extracted using TRIzolTM reagent (GibcoBRL Life Technologies) as per the manufacturer’s instructions. 2.4. Automated sequencing Samples that showed altered mobility by SSCP and/or HA analysis were reamplified and removed from unincorporated primers and dNTPs using the Promega Wizard PCR clean up kits (Promega, Madison, WI). These samples were cycle sequenced [47] using appropriate sequencing primers and BIG DYETM terminator kits from PE Biosystems (The Perkin Elmer Corporation, Foster City, CA). The resultant reaction was precipitated with 75% isopropan-2-ol to remove excess fluorescent dyes and salts. The pellets were re-suspended in Template Suppression ReagentTM from PE Biosystems,

denatured for 5 min and chilled on ice. The samples were loaded into an ABI 310 Genetic Analyser (PE Biosystems) and electrophoresed in the analyser’s capillary, through Performance Optimised Polymer 6 (PE Biosystems) at 15 kV for 25 min. The resultant data was analysed using ABI Sequence Analysis software. To reduce the possibility of missing mutations during SSCP/HA analysis, five tumour DNA samples that were observed to be negative by these methods, were also randomly chosen for each exon and sequenced. 2.5. Immunohistochemistry of hMSH2 and hMLH1 Twenty-four of the 35 SCCHN specimens investigated for microsatellite analysis were examined for expression of two DNA MMR proteins, hMSH2 and hMLH1. Formalin-fixed paraffin embedded sections (2.5 mm) were mounted on APES treated glass slides and dried overnight at 37  C. Prior to immunohistochemistry the slides were de-paraffinised and rehydrated. The sections were microwave pre-treated using citrate buffer before incubation with the primary antibody. For hMSH2 experiments, the monoclonal anti-hMSH2-N terminal antibody (Cat No. MCA1672, Serotec, UK) was used at a dilution of 1:20. For hMLH1 experiments, the monoclonal anti-hMLH1 antibody (Cat No.13271A, Pharmingen, A Becton Dickinson Company, UK) was used at a dilution of 1:10. In the negative control reactions, the primary antibody was replaced with a non-immune serum for the same species (Biogenex, USA). Sections of tonsil tissue were used as the positive controls for both antibodies. After primary antibody incubation and washing of the slides, a Biotinylated anti-mouse immunoglobulin (Biogenex Stravigen Multilink kit, Biogenex, USA) was used as the second antibody. After incubation and washing, the slides were then incubated with streptoavidin alkaline phosphatase immunoglobulin (Biogenex Stravigen Multilink kit, Biogenex, USA) and stained using the chromagen Vector Red using Vectastain ABC AP kit (Vector Laboratories, Burlingame, USA). Two consecutive sections were stained for each specimen. A number of sections also contained normal epithelium, providing an internal control. JN and WP independently assessed the staining intensity of hMSH2 and hMLH1 samples. The amount of hMSH2 and hMLH1 staining was tabulated using the following criteria: > 75%, more than 75% of tumour nuclei stained; 50– 75%, between 50 and 75% of tumour nuclei stained; 10–50%, between 10–50% of tumour nuclei stained; < 10%, less that 10% of tumour nuclei stained. The intensity of staining was tabulated on a scale from ( ) to (+++) [as absent ( ), weak (+), moderate (++) and strong (+++)] without the knowledge of clinical, mutational and LOH status data.

119

J. Nunn et al. / Oral Oncology 39 (2003) 115–129

2.6. Statistical analysis Specimens were included for analysis, in Table 1, if at least two of the microsatellite markers, used at each of the genetic loci, demonstrated informative results. Statistical analysis was undertaken, on quantitative data, by Chi square analysis or Fisher’s exact test where appropriate, using SPSS software for Windows. A P value of < 0.05 was considered statistically significant.

3. Results 3.1. Allelic imbalance The exact localisation for the DNA mismatch repair genes and for each microsatellite marker is now known [60–63] (Fig. 1) and has been of value in the interpretation of the allelic imbalance data. A detailed account of the results with microsatellite markers located close to the MMR genes, hMSH2, hMSH6, hMLH1, hPMS1, hPMS2 and hMSH3 are shown in Table 1 and Fig. 2. Analysis of the data revealed allelic imbalance in 28 out of the 35 (80%) SCCHN specimens for at least one of the microsatellite markers mapping to the six chromosomal locations corresponding to the DNA MMR genes investigated. The majority of these specimens (20/28) showed AI at only one or two of the six chromosomal regions investigated. However, eight samples showed a higher rate of AI that may be indicative of a more widespread genomic instability. Allelic imbalance at two or more microsatellite markers was observed most frequently at the hMLH1 region (12/33; 36%; Table 1). Allelic imbalance at two or more markers in the regions of the other mismatch repair gene loci was observed in less than 17% of SCCHN specimens. Six microsatellite markers were found close to the hMLH1 gene sequence position (40.77 Mb) on chromosome 3p21-p24 (Fig. 1). The sequence span of these microsatellite markers was 14.77 Mb, and included D3S1611 at 40.49 Mb and D3S1561 at 40.90 Mb, on either side of the hMLH1 gene location. Allelic imbalance was observed in 18/33 (54.5%) SCCHN specimens at one or more of the markers examined (Figs. 2A and 3,

Table 1). Allelic imbalance at either D3S1611, which lies within an intron of the hMLH1 gene, or D3S1561 was found in 10/32 (31%) of informative SCCHN specimens. The highest percentage allelic imbalances were exhibited at D3S1612 (48%) and by D3S1609 (47%). From these results one minimal region of allelic imbalance can be identified between the D3S1612 and D3S1611 markers (39.90–40.49 Mb) as defined proximally by tumours 1232, 351, 1062, 1197, 427, 327, 346, 438 and distally be tumours 335, 341, 438. This observation indicates that one or two genes residing between 27.40–40.49 Mb maybe concerned with SCCHN development, and that this gene may be the hMLH1 gene. Six microsatellite markers were found to be close to the hMSH3 gene (67.28 Mb; Fig. 1). The sequence span was 28.90 Mb and the hMSH3 gene resides between D5S637 (62.55 Mb) and D5S39 (68.78 Mb). A minimal region of loss between markers D5S637 and D5S39 (62.55–68.78 Mb) is defined distally by tumours 324, 351, 318, 432 and 1219, and proximally by tumours 352 and 1219 (Fig. 2B). The marker showing the greatest AI was D5S637 (6/21; 29%). Six microsatellite markers were located close to the hMSH2/hMSH6 genes (61.64 and 58.56 Mb, respectively; Fig. 1). The sequence span of these microsatellite markers was 11.96 Mb and included D2S119 at 57.23 Mb and D2S391 at 61.89 Mb on either side of the location of the two mismatch repair genes. The maximum AI was observed at marker D2S378 (4/21; 19%) and one minimal region of loss was identified (Fig. 2C). Located between D2S119 and D2S288 (57.23–62.05 Mb) this region is defined proximally by tumours 438 and 342 and distally by tumour 438. Five microsatellite markers were located close to the hPMS2 gene (6.90 Mb; Fig. 1). The sequence span of these microsatellite markers was 12.93 Mb and included D7S511 and D7S481 (6.12 and 7.08 Mb, respectively) located either side of the hPMS2 gene. The maximum AI was observed at the D7S472 marker (4/23; 17%; Fig. 2D). A minimal region of allelic imbalance between D7S531 and D7S511 (3.77–6.12 Mb) was defined proximally by tumour 341 and distally by tumour 1228 (Fig. 2D). The hPMS2 gene is not located within this region. Six microsatellite markers were located close the hPMS1 gene 195.55 Mb (Fig. 1). The sequence span of

Table 1 Summary of allelic imbalance observed at DNA mismatch repair (MMR) gene loci Mismatch repair gene

No. of microsatellite markers

Chromosomal location

% LOH at 1 loci

% LOH at >2 loci

hMLH1 hMSH3 hMSH2 and hMSH6 hPMS2 hPMS1

6 6 6 5 6

3p21.3 5q11–13 2p15–16 7p22 2q31–33

18/33 10/34 6/33 8/32 14/33

12/33 6/34 4/33 3/32 4/33

(54.5%) (29%) (18%) (25%) (42%)

(36%) (17%) (12%) (9%) (12%)

Specimens were only included in analysis if they exhibited informative results (heterozygous or LOH) with two or more markers per locus.

120 J. Nunn et al. / Oral Oncology 39 (2003) 115–129

Fig. 2. Results obtained from microsatellite analysis surrounding the DNA mismatch repair gene loci.

J. Nunn et al. / Oral Oncology 39 (2003) 115–129

the microsatellite markers was 32.71Mb and included D2S152 and D2S118 located at 194.26 and 200.13 Mb, respectively on either side of the hPMS1 gene. The D2S152 and D2S325 markers showed the maximum AI in this region (6/22 and 7/26, respectively; 27%; Fig. 2E). One region of allelic imbalance was observed between D2S300 and D2S118 (186.70–200.13 Mb) and contained the hPMS1 gene. This region was defined distally by tumours 1062 and 341, and proximally by tumours 1197, 335 and 313 (Fig. 2E). 3.2. MSI status The 35 SCCHN specimens have been previously investigated for microsatellite instability with a range of microsatellite markers (range 21–186 markers), on 8 to 38 chromosomal arms [21]. Tumours were classified as MSI–High (MSI–H) when> 30% of microsatellite markers exhibit microsatellite instability MSI–Low (MSI–L) when 54.5 < 30% exhibit microsatellite instability and MSI–Stable (MSS) when 50 < 4.5% microsatellite instability is observed. Four out of 35 were identified as either MSI–H or MSI–L tumours, while the remaining 31 SCCHN were MSS tumours. Of the five minimal regions of loss encompassing MMR genes that are identified within this investigation (Fig. 2), allelic imbalance at D3S1611 and D5S39 (Fig. 2A and B, respectively) was found to be associated with the microsatellite instability status of the SCCHN specimens. Thus 3/3 informative MSI+ tumours showed AI at D3S1611, while only 6/20 MSS tumours showed AI at this marker. Allelic imbalance at the remaining three MMR gene loci was not associated with the microsatellite instability status of the SCCHN specimens analysed (Table 2). However, as the numbers of specimens with MSI are small, this association has to be treated with caution.

Table 2 Allelic imbalance in relation to microsatellite instability (MSI) status MMR gene

Marker

MSI+tumoursa MSS tumours P

hMLH1 hMSH3 hMSH2/hMSH6 hPMS1

D3S1611b D5S39c D2S391 D2S152d

3/3 (100%) 2/2 (100%) 0/2 1/3 (33%)

6/20 (30%) 3/22 (14%) 2/20 (10%) 5/19 (26%)

0.047* 0.036* 0.82 0.64

MMR, mismatch repair. a MSI+tumours=MSI–H or MSI–L tumours. b Three tumours show MSI at this marker so could not be assessed for AI. c One tumour shows MSI at this marker so could not be assessed for AI. d Two tumours show MSI at this marker so could not be assessed for AI. * Statistically significant.

121

3.3. Mutation analysis of hMSH2 and hMLH1 in SCCHN Mutational analysis of the hMSH2 and hMLH1 genes was undertaken in these 35 SCCHN specimens. Only one specimen (sample 1215) was found to have a genetic alteration in the hMSH2 gene (Fig. 4). Specimen 1215 was found to contain a 1 base pair insertion within exon 5 of the hMSH2 gene. This caused a missense mutation (F-> C; phenylalanine –> cysteine) followed by a frameshift mutation, which produced a premature termination codon. However, this specimen was a MSS tumour, and did not demonstrate allelic imbalance at any of the microsatellite markers close to the six known DNA mismatch repair genes. No mutations were observed in the hMLH1 gene. 3.4. Protein expression of hMSH2 and hMLH1 in SCCHN Positive control sections from tonsil tissue confirmed the specificity of the two antibodies. In all cases normal squamous epithelium demonstrated strong staining (+++) for both proteins. Therefore tumours demonstrating strong (+++) immunoreactivity were classified as ‘normal expression’ while tumours demonstrating absent, weak and moderate immunoreactivity were classified as ‘reduced expression’ (Figs. 5–7). Both hMSH2 and hMLH1 expression were examined in 24/35 SCCHN tissues (Table 3). Ten out of 24 (42%) was found to have intense hMLH1 staining (+++) and 14 (58%) showed reduced (absent, weak or moderate) staining ( , +, ++) with this antibody. Of the specimens with reduced expression, one showed absence ( ) of hMLH1 expression whereas six showed weak (+) and seven showed moderate (++) expression. In contrast, only 3 / 24 (12.5%) SCCHN specimens were found to have strong expression (+++) of hMSH2 while 21 (87.5%) showed reduced (absent, weak or moderate) expression ( , +, ++). These results confirm the localisation of both hMSH2 and hMLH1 as exclusively nuclear and very prominent in the most proliferating parts of the epithelium. Expression of both hMSH2 and hMLH1 was observed most prominently just above the basal cell layer of the epithelium. The results agree with the co-localisation of the hMLH1 and hMSH2 proteins consistent with the current understanding of their proposed biochemical function in the DNA mismatch repair system [64]. It is of note that 23/24 (96%) of all the examined SCCHN tumours showed reduced expression of at least one of the two investigated genes. Additionally, simultaneous reduced expression for both hMLH1 and hMSH2 was found in 12/24 (50%) SCCHN specimens examined. Samples with reduced expression of both mismatch repair proteins did not show any associations

122 J. Nunn et al. / Oral Oncology 39 (2003) 115–129 Fig. 3. Sequence traces of (A) Exon 5 of the hMSH2 gene in squamous cell carcinoma of the head and neck (SCCHN) specimen 1215, (B) Exon 5 of the hMSH2 gene in the human colorectal cancer cell line HCT116.

J. Nunn et al. / Oral Oncology 39 (2003) 115–129

123

Fig. 4. Sequence traces of (A) Exon 5 of the hMSH2 gene in squamous cell carcinoma of the head and neck (SCCHN) specimen 1215, (B) Exon 5 of the hMSH2 gene in the human colorectal cancer cell line HCT116.

with any of the clinicopathological parameters examined, compared with samples with reduced expression of only one of the examined proteins. All of the 24 specimens examined had been investigated for allelic imbalance at the chromosomal locations of the DNA mismatch repair genes. A comparative analysis of hMLH1 and hMSH2 protein expression with allelic imbalance data on the corresponding chromosome gene location showed that reduced hMLH1 or hMSH2 expression did not correlate with allelic imbalance at these chromosomal loci.

Table 3 Summary of protein expression for hMLH1 and hMSH2 in squamous cell carcinoma of the head and neck (SCCHN) hMLH1 expression Reduced – 1

+ 6

+ 7

hMSH2 expression Normal

Reduced

+++ 10

0

+ 4

Normal + 17

+++ 3

, absent in expression; +, weak expression; ++, moderate expression; +++, strong expression.

4. Discussion The activation of oncogenes, loss of tumour suppressor genes and impaired mismatch repair function are known to be involved in the development of cancer. Defects in DNA mismatch repair genes are known to lead to microsatellite instability. The suggestion that defective DNA repair is a contributing factor in the development of SCCHN has been supported by reports of microsatellite instability in SCCHN and oral cancer [31,32,65–68]. This investigation studied 35 SCCHN for allelic imbalance at microsatellite markers surrounding six of the known DNA mismatch repair genes. Based on the location of the mismatch repair genes and the microsatellite markers from the recent integration of genetic data from several sources [60–63], the incidence of observed allelic imbalance at the six DNA mismatch repair genes was reported. We have demonstrated that 12/33 (36%) and 6/34 (17%) of the SCCHN analysed specimens analysed in this study were affected by allelic imbalance at two or more microsatellite markers at the hMLH1 and hMSH3 genes, respectively. Allelic imbalance was found to be less common at the hMSH2/hMSH6 (12%), hPMS1 (12%) and hPMS2 (9%) loci in these specimens.

124

J. Nunn et al. / Oral Oncology 39 (2003) 115–129

Fig. 5. Immunohistochemistry staining of the hMLH1 protein in human oral epithelial tissue. (A) Normal epithelium. The nuclei in the epithelial cells of the basal layer show +3 intensity (Arrow indicates normal epithelium). (B) SCCHN showing +1 intensity nuclear staining. (C) SCCHN showing +2 intensity nuclear staining. (D) SCCHN showing +3 nuclear staining. Magnification 20 (A), 20 (B), 20 (C), 20 (D).

Fig. 6. Immunohistochemistry staining of the hMLH1 protein in human oral epithelial tissue. (A) Normal epithelium. The nuclei in the epithelial cells of the basal layer show +3 intensity (Arrow indicates normal epithelium). (B) SCCHN showing +1 intensity nuclear staining. (C) SCCHN showing +2 intensity nuclear staining. (D) SCCHN showing +3 nuclear staining. Magnification 20 (A), 20 (B), 20 (C), 20 (D).

J. Nunn et al. / Oral Oncology 39 (2003) 115–129

Six previous studies have analysed the occurrence of LOH at the mismatch repair gene loci in sporadic colorectal (CRC), HNPCC tumours, sporadic breast cancer, and non-small cell lung cancer [37–41,43] and it has been concluded that LOH of markers within or adjacent to the hMLH1 gene occurs in tumours of patients in which the disease phenotype co segregates with hMLH1 [37]. A slight excess of LOH was additionally observed in this tumour series in the region of hMSH2 locus in tumours with no evidence of hMLH1 involvement. Thus it was proposed that hMLH1 resembles a tumour suppressor gene, in that it follows Knudson’s ‘two- hit’ model of inactivation [34–36], and that all the DNA mismatch repair genes may follow the same mechanism, though it may only be the MSI+ tumours [37–40]. Previous studies have investigated the involvement of chromosome 3 in the development of SCCHN and implicated chromosome 3p loss as being an early event, with three regions of minimal loss being described; 3p24–26, 3p21.3–22.1, and 3p12.1–14.2 [69–71]. Allele loss has been demonstrated at chromosome 3p in 44– 71% of cases [18–20,69–72], and at the 3p21 region in 7– 53% of cases [70,73–75]. In this study, allele loss was observed at one or more microsatellite markers in 18/33 (54.5%) of SCCHN at the chromosome 3p21 region. This is slightly higher than previously reported for this region in SCCHN. Approximately a third (10/32; 31%)

Fig. 7. Immunohistochemistry staining of the hMSH2 protein in SCCHN specimen 1215. Magnification 10 (A), 20 (B).

125

of the SCCHN investigated exhibited allelic imbalance at both of the adjacent microsatellite markers to the hMLH1 gene (D3S1611 and D3S1561). A minimal region of loss was observed to be between the D3S1612 and D3S1611 markers, with D3S1611 located within an intron of the hMLH1 gene. Allelic imbalance at D3S1611 was more frequently observed in MSI+ tumours (3/3; 100%) compared to MSS tumours (6/20; 30%) (P=0.047). This indicates that allelic imbalance at this microsatellite marker maybe associated with the phenomenon of microsatellite instability. As the numbers of specimens with MSI are small, this association should to be treated with caution. A small number of SCCHN allelotype studies have analysed microsatellite markers on the majority of chromosomal arms [18–20], and give an indication of the amount of allele loss on chromosomes 2p, 2q, 5q and 7p. Only Field et al. [20] investigated all of these regions with more than one microsatellite marker. Table 4 provides a summary of the SCCHN allelotype studies in relation to the chromosomal arms bearing DNA mismatch repair genes and also provides information on the number of microsatellite markers investigated and their genetic location. Allele loss on 5q, mainly in the APC gene region of 5q21–22, has previously been demonstrated in 25–43% of SCCHN cases [18–20] (Table 4). In this study, the 5q11–13 region was analysed with a total of six microsatellite markers and 10/34 (29%) of SCCHN cases exhibited allele loss at one or more markers. Allelic imbalance at D5S39 was found to be significantly more frequently observed in MSI+ tumours (2/2, 100%) compared with MSS tumours (3/22, 14%; P=0.036). This indicates that allelic imbalance at this microsatellite marker maybe associated with the phenomenon of microsatellite instability. Allelic imbalance at the other three chromosomal locations containing MMR genes has rarely been studied in SCCHN (see Table 4). In this study, allelic imbalance at more than two microsatellite markers at these locations was found in less than 20% of tumours. No association between allelic imbalance and microsatellite instability was exhibited in the SCCHN specimens investigated. It has been argued that rather than detecting a complete loss of an allele, in many cases we are actually observing a partial shift in the ratio of both parental alleles in the tumour DNA relative to that in normal DNA. Various chromosomal- and tumour-specific mechanisms have been suggested which may explain the loss of an allele in a tumour DNA specimen [76–79]. These include: (1) deletion of the wild type chromosome resulting in hemizygosity at all loci near the TSG; (2) loss followed by amplification resulting in two copies of one allele and loss of the other; (3) mitotic recombination between homologues, resulting in heterozygosity at loci in the proximal region and homozygosity

126

J. Nunn et al. / Oral Oncology 39 (2003) 115–129

Table 4 Table summarising the percentage allelic imbalance observed at the DNA mismatch repair gene locations in the previous squamous cell carcinoma of the head and neck (SCCHN) allelotype studies Percentage LOH observed at specific chromosomal regions 2p

2q

3p

5q

7p

Ah-See et al. [18] No. of markers used/arm Chromosomal location investigated

0% 1 2p12

7% 1 2q36–37

44% 3 3p13-p22

43% 5 5q11.2-q34

8% 1 7p11.2-p12

Nawroz et al. [19] No. of markers used/arm Chromosomal location investigated

19% 1 2p24

15% 1 2q23–2q24

67% 2 3p24.1–3p22 & 3q21.3–3q25.2

25% 1 5q22-q23

23% 1 7p21–7p15

Field et al. [20] No. of markers used/arm Chromosomal location investigated

23% 6 2p12-p25

13% 3 2q13-q37

52% 18 3pter-p12 (5 markers at 3p21 region)

29% 8 5cen-qter (2 markers at 5q11-q13 region)

8% 1 7pter-7p15

throughout the rest of the chromosome, including the TSG locus; (4) localized events, such as point mutations, small deletions and gene conversions; (5) oxidative damage which is widespread in human cancer; and (6) chromosomal aneuploidy [76–79]. Therefore it may be impossible to casually link an observed LOH event in a single tumour, which usually extends over a large chromosomal region, to the inactivation of any gene in that region, but it can highlight regions of potential interest. Mutational analysis of the hMSH2 and hMLH1 genes found only one mutation amongst the 35 SCCHN specimens analysed. A 1 base pair insertion within exon 5 of the hMSH2 gene was found in SCCHN specimen 1215. This base pair insertion produced a missense mutation (F– > C), causing a frameshift and a premature truncated protein. The alteration was proposed to be a mutation, and not a polymorphism, due to its detrimental affect on the protein. This SCCHN specimen did not exhibit any sign of MSI, LOH at MMR gene loci or reduced expression of the hMLH1 or hMSH2 proteins. Thus, it may be suggested that the two MMR genes that were studied may not contain genetic mutations relating to MSI observed in this SCCHN. It is possible that mutations may lie in the remaining MMR genes and that different MSI+ cancers demonstrate different mutation spectrums or pathways leading to defective mismatch repair. Alternatively mutations may reside in the promoter elements of these genes [80]. The immunohistochemical analysis demonstrated that 58% of the examined SCCHN tumours had reduced expression of hMLH1 and 87.5% had reduced expression of hMSH2, while 50% demonstrated reduction of expression in both these genes and 23/24 (96%) of all the examined SCCHN tumours showed reduced

expression of at least one of the two investigated genes. The results of this study agree with those of Lo Muzio et al. [81] where an absence of expression for both hMSH2 and hMLH1 in oral SCC was reported and it was suggested that this might constitute a hallmark of potential mutator phenotype for this type of neoplasm [81]. Previous studies, mainly in CRC, have reported a close relationship between mismatch repair protein expression and MSI status. Loss of hMSH2 and hMLH1 immunostaining has been observed to be restricted entirely to the high MSI cases (MSI–H), while tumours with low MSI frequency (MSI–L or MSS) do not show loss of hMSH2 or hMLH1 expression [82–85]. However, the results in this study on SCCHN tumours do not support this relationship, indicating that the association between MSI status and mismatch repair protein expression may be tissue specific. DNA mismatch repair genes are considered to be inactivated in a two-hit model. Thus the relationship between protein expression levels and mutational analysis of two MMR genes together with allelic imbalance located near six known MMR genes was investigated. No relationship was found between the protein expression levels, mutation analysis and allelic imbalance at any microsatellite marker analysed. However, it can be suggested from this study, that chromosome 3p and 5q allelic imbalance is involved in the development of SCCHN and may be associated with the occurrence of microsatellite instability in these tumours.

Acknowledgements The University of Liverpool and the North West Cancer Research Fund supported this research.

J. Nunn et al. / Oral Oncology 39 (2003) 115–129

References [1] Parkin DM, Pisani P, Ferlay J. Estimates of the worldwide incidence of eighteen major cancers in 1985. Int J Cancer 1993;54: 594–606. [2] Coleman MP, Esteve J, Damiecki P, Arslan A, Renard H. World Health Organisation. International Agency for Research on Cancer. Trends in cancer incidence and mortality. IACR Scientific Publications, No. 121, Lyon; 1993. [3] Johnson NW. A global view of the epidemiology of oral cancer: Johnson NW, editor. Oral cancer detection of patients and lesions at risk. (vol 2). Cambridge & New York: Cambridge University Press, 1991. p. 3–26. [4] Dafarty DK, Murti PR, Bhonsle RB, Gupta PC, Mehta FS, Pindborg JJ. Risk factors and risk markers for oral cancers in high incidence areas of the world. In: Johnson NW, editor. Oral cancer detection of patients and lesions at risk. (vol. 2). Cambridge & New York: Cambridge University Press, 1991. p. 29–61. [5] Smith CJ. Epidemiology & Aetiology. In: Langdon JD, Henk JM, editors. Malignant Tumours of the Mouth, Jaws & Salivary Glands, 2nd ed. London & Boston: Edward Arnold, 1995. p. 1– 13. [6] Pisani P, Parkin DM, Bray F, Ferlay J. Estimates of the worldwide mortality from 25 cancers in 1990. Int J Cancer 1990;83:18– 29. [7] Alvi A, Myers EN, Johnson JT. Cancer of the oral cavity In: Myers EN, Suen JY, editors. Cancer of the head and neck, 3rd ed. Philadelphia: WB Saunders Company, 1996. p. 321–60. [8] Fearon ER, Vogelstein B. A genetic model for colorectal tumorigenesis. Cell 1990;61(5):757–67. [9] Field JK. The role of oncogenes and tumour-suppressor genes in the aetiology of oral, head and neck squamous cell carcinoma. J R Soc Med 1995;88(1):35P–9P. [10] Scholes AGM, Field JK. Genomic instability in head and neck cancer. In: Seifert G, editor. Current topics in pathology, oral pathology—actual diagnostic and prognostic aspects, 1st ed. Heidelberg: Springer-Verlag, 1996. p. 201–22. [11] Scholes AGM, Woolgar JA, Boyle MA, Brown JS, Vaughan ED, Hart CA, et al. Synchronous oral carcinomas: independent or common clonal origin? Cancer Res 1998;58(9):2003–6. [12] Liloglou T, Scholes AGM, Spandidos DA, Vaughan ED, Jones AS, Field JK. p53 mutations in squamous cell carcinoma of the head and neck predominate in a subgroup of former and present smokers with a low frequency of genetic instability. Cancer Res 1997;57(18):4070–4. [13] Field JK. Genomic instability in squamous cell carcinoma of the head and neck. Anticancer Research 1996;16(4C):2421–31. [14] Scully C, Field JK. Genetic aberrations in squamous cell carcinomas of the head and neck (SCCHN), with reference to oral carcinoma (review). Int J Oncol 1997;10(1):5–21. [15] Scully C, Field JK, Tanzawa H. Genetic aberrations in oral or head and neck squamous cell carcinoma (SCCHN): 1. Carcinogen metabolism, DNA repair and cell cycle control Oral Oncol 2000;36(3):256–63. [16] Scully C, Field JK, Tanzawa H. Genetic aberrations in oral or head and neck squamous cell carcinoma (SCCHN): 2 Chromosomal aberrations. Oral Oncol 2000;36(4):311–27. [17] Hansen MF, Cavenee WK. Genetics of cancer predisposition. Cancer Res 1987;47(21):5518–27. [18] Ah-See KW, Cooke TG, Pickford IR, Soutar D, Balmain A. An allelotype of squamous carcinoma of the head and neck using microsatellite markers. Cancer Research 1994;54:1617– 21. [19] Nawroz H, Van der Riet P, Hruban RH, Koch W, Ruppert JM, Sidransky D. Allelotype of head and neck squamous cell carcinoma. Cancer Res 1994;54(5):1152–5.

127

[20] Field JK, Kiaris H, Risk JM, Tsiriyotis C, Adamson R, Zoumpourlis V, et al. Allelotype of squamous cell carcinoma of the head and neck: fractional allele loss correlates with survival. British Journal of Cancer 1995;72:1180–8. [21] Nunn J, Scholes AGM, Liloglou T, Nagini S, Jones AS, Vaughan ED, et al. Fractional allele loss indicates distinct genetic populations in the development of squamous cell carcinoma of the head and neck (SCCHN). Carcinogenesis 2000;20(12):2219–28. [22] Peltomaki P, Aaltonen LA, Sistonen P, Pylkkanen L, Mecklin JP, Jarvinen H, et al. Genetic mapping of a locus predisposing to human colorectal cancer. Science 5109;1993a:260 810–812. [23] Aaltonen LA, Peltomaki P, Leach FS, Sistonen P, Pylkkanen L, Mecklin JP, et al. Clues to the pathogenesis of familial colorectal cancer. Science 5109;1993:260 812–816. [24] Thibodeau SN, Bren G, Schaid D. Microsatellite instability in cancer of the proximal colon. Science 5109;1993:260 816–819. [25] Lothe RA, Peltomaki P, Meling GI, Aaltonen LA, NystromLahti M, Pylkkanen L, et al. Genomic instability in colorectal cancer—relationship to clinicopathological variables and family history. Cancer Res 1993;53(24):5849–52. [26] Aaltonen LA, Peltomaki P, Mecklin JP, Jarvinen H, Jass JR, Green JS, et al. Replication errors in benign and malignant tumours from hereditary non-polyposis colorectal cancer patients. Cancer Res 1994;54(7):1645–8. [27] Peltomaki P, Lothe RA, Aaltonen LA, Pylkkanen L, NystromLahti M, Seruca R, et al. Microsatellite instability is associated with tumours that characterize the hereditary non-polyposis colorectal carcinoma syndrome. Cancer Res 1993;53(24):5853–5. [28] Keller G, Rotter M, Vogelsang H, Bischoff P, Becker KF, Mueller J, et al. Microsatellite instability in adenocarcinomas of the upper gastrointestinal tract. Relation to clinicopathological data and family history. Am J Pathol 1995;147(3):593–600. [29] Chung YJ, Park SW, Song JM, Lee KY, Seo EJ, Choi SW, et al. Evidence of genetic progression in human gastric carcinomas with microsatellite instability. Oncogene 1997;15(14):1719–26. [30] Pifarre A, Rosell R, Monzo M, De Anta JM, Moreno I, Sanchez JJ, et al. Prognostic value of replication errors on chromosomes 2p and 3p in non-small-cell lung cancer. Br J Cancer 1997;75(2): 184–9. [31] Ogawara K, Uzawa K, Nakanishi H, Yokoe H, Wang X, Tanzawa H, et al. Frequent microsatellite instability in oral cancer. Oncology Reports 1997;4:161–5. [32] Piccinin S, Gasparotto D, Vukosavljevic T, Barzan L, Sulfaro S, Maestro R, et al. Microsatellite instability in squamous cell carcinomas of the head and neck related to field cancerisation phenomena. Br J Cancer 1998;78(9):1147–51. [33] Han HJ, Yanagisawa A, Kato Y, Park JG, Nakamura Y. Genetic instability in pancreatic cancer and poorly differentiated type of gastric cancer. Cancer Res 1993;53(21):5087–9. [34] Knudson AG. Mutation and cancer: Statistical study of retinoblastoma. Proceedings of the National Academy of Sciences of the USA 1971;68:820–3. [35] Knudson AG. Hereditary cancer, oncogenes and anti-oncogenes. Cancer Research 1985;45:1437–43. [36] Knudson AG. Hereditary cancer: two hits revisited. Journal of Cancer Research Clinical Oncology 1996;122:135–40. [37] Hemminki A, Peltomaki P, Mecklin J, Jarvinen H, Salovaara R, Nystrom-Lahti M, et al. Loss of the wild type MLH1 gene is a feature of HNPCC. Nature Genetics 1994;8:405–10. [38] Tomlinson IPM, Ilyas M, Bodmer WF. Allele loss occurs frequently at hMLH1, but rarely at hMSH2, in sporadic colorectal cancers with microsatellite instability. Br J Cancer 1996;74(10): 1514–7. [39] Benachenhou N, Guiral S, Gorska-Flipot I, Labuda D, Sinnett D. High-resolution deletion mapping reveals frequent allelic losses at the DNA mismatch repair loci hMLH1 and hMSH3 in non-small cell lung cancer. Int J Cancer 1998;77(2):173–80.

128

J. Nunn et al. / Oral Oncology 39 (2003) 115–129

[40] Benachenhou N, Guiral S, Gorska-Flipot I, Michalski R, Labuda D, Sinnett D. Allelic losses and DNA methylation at DNA mismatch repair loci in sporadic colorectal cancer. Carcinogenesis 1998;19(11):1925–9. [41] Benachenhou N, Guiral S, Gorska-Flipot I, Labuda D, Sinnett D. Frequent loss of heterozygosity at the DNA mismatch-repair loci hMLH1 and hMSH3 in sporadic breast cancer. Br J Cancer 1999;79(7–8):1012–7. [42] Johannsdottir JT, Bergthorsson JT, Gretarsdottir S, Kristjansson AK, Ragnarsson G, Jonasson JG, et al. Replication error in colorectal carcinoma: association with loss of heterozygosity at mismatch repair loci and clinicopathological variables. Anticancer Res 1999;19(3A):1821–6. [43] Wieland I, Ammermuller T, Bohm M, Totzeck B, Rajewsky MF. Microsatellite instability and LOH at the hMLH1 locus on chromosome 3p21 occur in a subset of Non Small Cell Lung Cancers. Oncology Research 1996;8(1):1–5. [44] Hu X, Guo Z, Pang T, Li Q, Afink G, Ponten J. Immunohistochemical and DNA sequencing analysis on human mismatch repair gene hMLH1 in cervical squamous cell carcinoma with LOH of this gene. Anticancer Res 2000;20(1A):171–5. [45] Fishel R, Lescoe MK, Rao MRS, Copeland NG, Jenkins NA, Garber J, Kane M, et al. The human mutator gene homologue MSH2 and its association with Hereditary Non-Polyposis Colon Cancer. Cell 1993;75:1027–38. [46] Leach FS, Nicolaides NC, Papadopoulos N, Liu B, Jen J, Parsons R, et al. Mutations of a MutS homolog in Hereditary NonPolyposis Colorectal Cancer. Cell 1993;75(6):1215–25. [47] Bronner CE, Baker SM, Morrison PT, Warren G, Smith LG, Lescoe MK, et al. Mutation in the D.N.A. mismatch repair gene homolog hMLH1 is associated with hereditary non-polyposis colon cancer. Nature 6468;1994:368 258–261. [48] Papadopoulos OS, Nicolaides NC, Wei YF, Ruben SM, Carter KC, Rosen CA, et al. Mutation of a MutL homolog in hereditary colon cancer. Science 5153;1994:263 1625–1629. [49] Nicolaides NC, Papadopolous Liu B, Wei Y, Carter KC, Ruben SM, et al. Mutations of two PMS homologues in hereditary nonpolyposis colon cancer. Nature 1994;371:75–80. [50] Risinger JI, Umar A, Boyd J, Berchuck A, Kunkel TA, Barrett JC. Mutation of MSH3 in endometrial cancer and evidence for its functional role in heteroduplex repair. Nat Genet 1996;14(1):102–5. [51] Nystrom- Lahti M, Parsons R, Sistonen P, Pylkkanen L, Aaltonen LA, Leach FS, et al. Mismatch repair genes on chromosome 2p and 3p account for a major share of Hereditary Non-Polyposis Colorectal Cancer families evaluable by linkage. American J Hum Genet 1994;55(4):659–65. [52] Papadopoulos N, Nicolaides NC, Liu B, Parsons RE, Palombo F, D’Arrigo A, et al. Mutations of GTBP in genetically unstable cells. Science 1995;268:1915–7. [53] Palombo F, Gallinari P, Iaccarino I, Lettieri T, Hughes M, D’Arrigo A, et al. GTBP, a 160-kilodalton protein essential for mismatch-binding activity in human cells. Science 1995; 268(5219):1912–4. [54] Lipkin SM, Wang V, Jacoby R, Banerjee-Basu S, Baxevanis AD, Lynch HT, et al. MLH3: a DNA mismatch repair gene associated with mammalian microsatellite instability. Nat Genet 2000; 24(1):27–35. [55] Li GM, Modrich P. Restoration of mismatch repair to nuclear extracts of H6 colorectal tumour cells by a heterodimer of human MutL homologs. Proc Natl Acad Sci USA 1995;92:1950–4. [56] Umar A, Boyer JC, Thomas DC, Nguyen DC, Risinger JI, Boyd J, et al. Defective mismatch repair in extracts of colorectal and endometrial cancer cell lines exhibiting microsatellite instability. J Biol Chem 1994;269(20):14367–70. [57] Liu B, Nicolaides NC, Markowitz S, Wilson JKV, Parsons RE, Jen J, et al. Mismatch repair gene defects in sporadic colorectal cancers with microsatellite instability. Nat Genet 1995;9:48–55.

[58] Boyer JC, Umar A, Risinger JI, Lipford JR, Kane M, Yin S, et al. Microsatellite instability, mismatch repair deficiency, and genetic defects in human cancer cell lines. Cancer Research 1995; 55:6063–70. [59] da Costa LT, Liu B, El-Diery W, Hamilton SR, Kinzler KW, Vogelstein B, et al. Polymerase delta variants in RER colorectal tumours. Nature Genetics 1995;9:10–11. [60] GDB (TM) Human Genome Database [database online]. Toronto (Ontario, Canada): The Hospital for Sick Children, Baltimore (Maryland, USA): John Hopkins University, 1990. Updated daily. http://www.gdb.org/. [61] Dib C, Faure S, Fizames C, Samson D, Drouot N, Vignal A, et al. A comprehensive genetic map of the human genome based on 5, 264 microsatellites. Nature 1996;380(6570):152–4 . [62] Collins A, Frezal J, Teague J, Morton NE. A genetic map of humans; 23, 500 loci in 850 bands. Proc Natl Acad Sci USA 1996; 93:14771–5. [63] Hubbard T, Barker D, Birney E, Cameron G, Chen Y, Clark L, et al. The Ensembl genome database project. Nucleic Acids Research 2002;30(1):38–41. [64] Fink D, Nebel S, Aebi S, Zheng H, Kim HK, Christen RD, et al. Expression of the DNA mismatch repair proteins hMLH1 and hPMS2 in normal human tissues. British Journal of Cancer 1997; 76(7):890–3. [65] Field JK, Kiaris H, Howard P, Vaughan ED, Spandidos DA, Jones AS. Microsatellite instability in squamous cell carcinoma of the head and neck. British Journal of Cancer 1995;71(5):1065– 9. [66] Ishwad CS, Ferrell RE, Rossie KM, Appel BN, Johnson JT, Myers EN, et al. Microsatellite instability in oral cancer. Int J Cancer 1995;64(5):332–5. [67] El-Naggar AK, Hurr K, Huff V, Clayman GL, Luna MA, Batsakis JG. Microsatellite instability in preinvasive and invasive head and neck squamous carcinoma. Am J Pathol 1996;148(6): 2067–72. [68] Mao L, Lee DJ, Tockman MS, Erozan YS, Askin F, Sidransky D. Microsatellite alterations as clonal markers for the detection of human cancer. Proc Natl Acad Sci USA 1994; 91(21):9871–5. [69] Maestro R, Gasparotto D, Vukosavljevic T, Barzan L, Sulfaro S, Boiocchi M. Three discreet regions of deletion at 3p in head and neck cancers. Cancer Res 1993;53:5775–9. [70] Wu CL, Sloan P, Read AP, Harris R, Thakker N. Deletion mapping on the short arm of chromosome 3 in squamous cell carcinoma of the oral cavity. Cancer Res 1994;54(24):6484–8. [71] Partridge M, Emilion G, Langdon JD. Loss of heterozygosity correlates with poor survival in oral squamous cell carcinoma. Br J Cancer 1996;197:73366–71. [72] El-Naggar AK, Hurr K, Batsakis JG, Luna MA, Goepfert H, Huff V. Sequential loss of heterozygosity at microsatellite motifs in preinvasive and invasive head and neck squamous carcinoma. Cancer Res 1995;55:2656–9. [73] Field JK, Tsiriyotis C, Zoumpourlis V, Howard P, Spandidos DA, Jones AS. Allele loss on chromosome 3 in squamous cell carcinoma of the head and neck correlates with poor clinical prognostic indicators. International Journal of Oncology 1994;4: 543–9. [74] Califano J, van der Reit P, Westra W, Nawroz H, Clayman G, Piantadosi S, et al. Genetic progression model for head and neck cancer: implications for field cancerisation. Cancer Research 1996;56:2488–92. [75] Roz L, Wu CL, Porter S, Scully C, Speight P, Read A, et al. Allelic imbalance on Chromosome 3p in oral dysplastic lesions: an early event in oral carcinogenesis. Cancer Res 1996;56(6): 1228–31. [76] Cavanee W. Recessive mutations in the causation of human cancer. Cancer 1991;67:2431–5.

J. Nunn et al. / Oral Oncology 39 (2003) 115–129 [77] Lasko D, Cavanee W. Loss of constitutional heterozygosity in human cancer. Annu Rev Genet 1991;25:281–314. [78] Turker MS, Gage BM, Rose JA, Elroy D, Ponomareva ON, Stambrook PJ, Tischfield JA. A novel signature mutation for oxidative damage resembles a mutational pattern found commonly in human cancers. Cancer Res 1999;59:1837–9. [79] Devilee P, Cleton-Jansen A, Cornelisse CJ. Ever since Knudson. TIGS 2001;17(10):569–73. [80] Shin KH, Shin JH, Kim JH, Park JG. Mutational analysis of promoters of mismatch repair genes hMSH2 and hMLH1 in hereditary nonpolyposis colorectal cancer and early onset colorectal cancer patients: identification of three novel germline mutations in the promoter of the hMSH2 gene. Cancer Res 2002;62:38–42. [81] Lo Muzio L, Nocini P, Mignogna MD, Pannone G, Staibano S, Procaccini M, et al. Immunocytochemical detection of hMSH2 and hMLH1 expression in oral SCC. Anticancer Res 1999; 19(2A):933–40.

129

[82] Thibodeau SN, French AJ, Roche PC, Cunningham JM, Tester DJ, Lindor NM, et al. Altered expression of hMSH2 and hMLH1 in tumours with microsatellite instability and genetic alterations in mismatch repair genes. Cancer Res 1996;56(21): 4836–40. [83] Dietmaier W, Wallinger S, Bocker T, Kullmann F, Fishel R, Ruschoff J. Diagnostic microsatellite instability: definition and correlation with mismatch repair protein expression. Cancer Res 1997;57(21):4749–56. [84] Kim H, Piao Z, Kim JW, Choi JS, Kim NK, Lee JM, et al. Expression of hMSH2 and hMLH1 in colorectal carcinomas with microsatellite instability. Pathol Res Pract 1998; 194:3–9. [85] Marcus VA, Madlensky L, Gryfe R, Kim H, So K, Millar A, et al. Immunohistochemistry for hMLH1 and hMSH2: a practical test for DNA mismatch repair-deficient tumours. Am J Surg Pathol 1999;23(10):1248–55.