Oncogene (1998) 17, 157 ± 163  1998 Stockton Press All rights reserved 0950 ± 9232/98 $12.00 http://www.stockton-press.co.uk/onc Mismatch repair genes and mononucleotide tracts as mutation targets in colorectal tumors with di€erent degrees of instability

Antonio Percesepe1,2, Paula Kristo1, Lauri A Aaltonen1, Maurizio Ponz de Leon2, Albert de la Chapelle1,3 and PaÈ ivi PeltomaÈ ki1,3

1Department of Medical Genetics, Haartman Institute, University of Helsinki, PO Box 21, Haartmaninkatu 3, FIN-00014 Helsinki, Finland; 2Department of Internal Medicine, University of Modena, Via del Pozzo, 71, 41100 Modena, Italy; 3Division of Human Cancer Genetics, Comprehensive Cancer Center, Ohio State University, 420 W 12th Avenue, Columbus, Ohio 43210, USA

Microsatellite instability occurs in 15% of colorectal genetically unstable phenotype is required for multi- carcinomas and may be due to replication errors (RER). step carcinogenesis (Loeb, 1991). A novel mechanism The pattern of instability ± `severe' vs `mild' ± and the of genetic instability has been elucidated in neoplasms tumorigenic pathway, as re¯ected by the involvement of from di€erent organs, consisting of alterations in functionally important genes, may vary according to the length (insertions or deletions) of repetitive sequences, underlying gene(s). We de®ned `mild' RER as mono- or , interspersed along the whole genome tetranucleotide repeat instability in the absence of (Aaltonen et al., 1993; Ionov et al., 1993; Thibodeau et widespread instability at dinucleotide repeats and studied al., 1993). Such instability is a characteristic conse- 15 colorectal tumors with this phenotype for mutations in quence of mutations in DNA mismatch repair genes the DNA mismatch repair genes MSH2, MLH1, that result in a failure to correct replication errors MSH3, and MSH6. No mutations were found, suggest- (RER) (reviewed in PeltomaÈ ki and de la Chapelle, ing that these genes were not implicated. We then 1997). Homozygous mutations in `major' mismatch compared colorectal cancers with `mild' RER (n=15), repair genes, such as MSH2 and MLH1 may give rise and those with `severe' RER without (n=11) or with to secondary mutations in other genes which contain (n=22) detectable mutations in MSH2 or MLH1 to structural targets for mutations and are selected for assess the involvement of mononucleotide repeats during tumorigenesis (Malkhosyan et al., 1996). contained in the coding regions of MSH3, MSH6, Potential targets that have been proposed to play a BAX, and TGFb RII. The combined mutation rates of role in colorectal tumorigenesis include other mismatch the above mentioned loci varied signi®cantly between the repair genes, for example MSH3 and MSH6 (also three groups of tumors, being 0%, 25% and 52%, called GTBP) (Malkhosyan et al., 1996), as well as respectively. Furthermore, the individual genes showed other genes, such as the gene for transforming growth speci®c patterns of involvement; for example, among factor b receptor II (TGFb RII; Markowitz et al., tumors with `severe' RER, TGFb RII displayed 1995), BAX (Rampino et al., 1997), and the gene for uniformly high mutation rates while MSH3, MSH6, insulin-like growth factor II receptor (IGFIIR; Souza and BAX were more frequently altered in tumors that et al., 1996). also showed MSH2 or MLH1 mutations. Our ®ndings Since the earliest observations it became apparent suggest that di€erent subcategories exist among unstable that two types of microsatellite instability exist, tumors, de®ned by the RER pattern on the one hand and depending on the number of the mutated markers, tumorigenic pathway on the other, and structural and on the degree of the increase or the decrease of the changes of MSH2 and MLH1 are likely to explain only fragment size. Tumors with dramatic changes at a a proportion of these cases. single locus often show a widespread instability at most of the studied microsatellite loci; they share certain Keywords: mismatch repair; microsatellite instability; clinical and pathological features like proximal location ; mutation analysis; HNPCC in the colon, poor di€erentiation, and better prognosis (Thibodeau et al., 1993; Ionov et al., 1993; Aaltonen et al., 1993; Lothe et al., 1993; Kim et al., 1994). This pattern characterizes Hereditary Nonpolyposis Colo- Introduction rectal Carcinoma (HNPCC) tumors and around 10% of sporadic colorectal cancers, and is associated with Current models of cancer development are based on mutations in MSH2, MLH1, or PMS2 in most tumors the experimental evidence that mutations accumulate in from the HNPCC group and in one-third of tumors a stepwise fashion, leading to abnormalities in speci®c from the sporadic group (Bùrresen et al., 1995; Liu et genes and resulting in clonal expansion of neoplastic al., 1996; Konishi et al., 1996; Wu et al., 1997). Less cells (Fearon and Vogelstein, 1990). It has been argued clear are the role and the mechanisms underlying low- that the number of modi®cations in oncogenes and level microsatellite instability. Based on observations in tumor suppressor genes is too high to be accounted for yeast, MSH3 and MSH6 whose protein products only by the spontaneous mutation rate, but a complex with the MSH2 protein in a partially redundant fashion (Marsischky et al., 1996) have been implicated in the explanation of the phenomenon. MSH6 recognizes mainly single-base mispairs and one- Correspondence: A Percesepe Received 2 December 1997; revised 18 February 1998; accepted 18 or two-nucleotide loops, and MSH6 mutant human February 1998 tumors and cell lines show microsatellite instability Genetic basis of microsatellite instability A Percesepe et al 158 primarily at mononucleotide tracts and apparently at cDNA and continued by systematic sequencing of all lower rates than tumor cells with mutations in MSH2 exons and intron ± exon borders from genomic DNA. or MLH1 (Papadopoulos et al., 1995; Akiyama et al., No mutations were found. Mutation analyses were 1997). Likewise, MSH3 mutations have been reported then extended to MSH3 and MSH6 that could even be to be associated with only a moderate increase in considered as better candidates to explain `mild' RER microsatellite instability in yeast, a€ecting only simple based on previous observations (see Introduction). repeats and preferentially leading to deletions (Strand However, no mutations were found in these genes, et al., 1995). Mutations in MSH3 in human either, using the IVSP technique. endometrial tumors have been reported (Risinger et In the course of the above mentioned analyses, a al., 1996). number of single base substitutions were detected that We studied colon cancers showing di€erent levels of were likely to represent innocuous polymorphisms microsatellite instability with the aims (1) to better because they did not generally lead to any coding de®ne low-degree microsatellite instability by screening changes and many have been reported to occur in the `mild' RER positive tumors for mutations in the DNA general population by us and others. The most mismatch repair genes MSH2, MLH1, MSH3 and common variants were (frequencies of the less MSH6, and (2) to study the sequence of events in the common alleles in parentheses): in MSH2, G?A, 12 development of tumors with microsatellite instability, nucleotides downstream of exon 10 (23%) (Wijnen et by analysing mononucleotide tracts in the coding al., 1994), and T?C, six nucleotides upstream of exon regions of MSH3, MSH6, BAX and TGFb RII as 13 (20%) (Leach et al., 1993); and in MLH1, G?A, 93 hotspots for mutations. nucleotides upstream of the initiating methonine (43%) (Aaltonen et al., 1998), and A?G, 19 nucleotides upstream of exon 15 (33%) (Wijnen et al., 1994). The IVSP analysis occasionally yielded extra bands and Results although sequencing of the respective fragments did not reveal any truncating mutations, a homozygous Microsatellite and mutation analyses in tumors with missense change (G?A) at position 1865 of MSH3 `mild' RER was found in tumor cDNA from patient 54, changing We de®ned `mild' RER as instability at mono- or Gly to Glu. The same alteration was present in normal tetranucleotide repeats in the absence of widespread mucosa from this patient and was subsequently found dinucleotide repeat instability. Our panel of six mono- to occur as a polymorphic variant in the population or tetranucleotide repeat markers identi®ed 15 such with the allele frequencies of 81% for A and 19% for tumors among those that had shown instability at one G. dinucleotide repeat locus, but had not satis®ed the de®nition of RER positivity (Table 1). MSH3, MSH6, BAX and TGFb RII mutations in `mild' All 15 tumors were subjected to DNA-based and `severe' RER tumors mutation analyses and with the exception of cases 70, 313 and 5, RNA samples were available for reverse To explore whether the degree of microsatellite transcription-PCR (RT ± PCR) and in vitro synthesized instability or the presence vs absence of detectable protein (IVSP) analyses. The primers and reaction mutations in DNA mismatch repair genes might conditions are speci®ed in NystroÈ m-Lahti et al. (1996), correlate with the rate of mutations in previously Chadwick et al. (1996) and in Table 2 and Materials identi®ed target genes, mononucleotide stretches in and methods. To screen the MSH2 and MLH1 genes MSH3, MSH6, BAX, and TGFb RII were examined in for mutations we started with RT ± PCR analyses of tumors with `mild' RER (n=15), as well as in those

Table 1 Microsatellite instability results in tumors with `mild' RER Dinucleotide repeat markers Mono-and tetranucleotide repeat markers MYCL1 Proportion of Bat 25 BAT 26 BAT 40 APD3 D19S244 tetranucl. Patient unstable loci poly(A) poly(A) poly(A) poly (A) tetranucl. +poly (A) 5 1/11 +a + + + 7b + 51 1/9 7 7 7 7 NDc + 54 1/9 7 7 7 7 ND + 58 1/11 + + + 7 7 7 59 1/8 7 7 7 7 + + 70 1/15 7 7 7 7 ND + 74 1/15 7 7 + + 7 7 77 1/16 7 7 7 7 + 7 92 1/15 7 7 + 7 7 7 93 1/15 7 7 7 7 7 + 97 1/15 7 7 + 7 7 7 313 1/10 7 7 7 7 7 + 369 1/12 7 7 7 7 7 + 395 1/7 + + + + + + 412 1/14 7 7 7 ND ND + aUnstable; bStable; cNot determined Genetic basis of microsatellite instability A Percesepe et al 159 Table 2 Sequences of primers used for the IVSP analysis of MSH3 and MSH6 Primer name Sequence Reference N-MSH3-1Fa,b 5'-ATGTCTCGCCGGAAGCCTGCGTCGG-3' This study MSH3-1F 5'-PGCGAGGCAAGCGGTTTTGAGC-3'd This study MSH3-1Rc 5'-AAAACTCTGTAACTGCCTGGAAAGC-3' Risinger et al. (1996) N-MSH3-2F 5'-AGATGTGAATCCCCTAATCAAGC-3' Malkhosyan et al. (1996) MSH3-2F 5'-PGAGGCGCTCATCCACAGAGCCACA-3' This study MSH3-2R 5'-AGGGAGAAAATGCAGTCAACAGTT-3' Risinger et al. (1996) N-MSH3-3F 5'-ACTCAATGAACAAGCTGCC-3' This study MSH3-3F 5'-PGTGAGCCGCTTTCACTCTCCTTTTATTGTAG-3' Risinger et al. (1996) MSH3-3R 5'-TTTTTAATTCTCCATTTTTGTTCAC-3' Risinger et al. (1996) N-MSH6-1F 5'-GTTGTGACTTCTCACCAGG-3' Nicolaides et al. (1996) MSH6-1F 5'-PGAGGGTTACCCCTGG-3' Papadopoulos et al. (1995) MSH6-1R 5'-ACACTGTAAGTCTGTGTACC-3' Papadopoulos et al. (1995) N-MSH6-2F 5'-TCTGATGTGGAATTTAAGCC-3' This study MSH6-2F 5'-PGGTGAAGGCCTGAACAGCC-3' Papadopoulos et al. (1995) MSH6-2R 5'-AAGTCCAGTCTTTCGAGCC-3' Papadopoulos et al. (1995) N-MSH6-3F 5'-CAGTGACATTAAACAACTTGGAG-3' This study MSH6-3F 5'-PGAGAGGGTTGATACTTGCC-3' Papadopoulos et al. (1995) MSH6-3R 5'-AGAAGTCAACTCAAAGCTTCC-3' Papadopoulos et al. (1995) aN, nesting; bF, forward; cR, reverse; dP, bacteriophage T7 promoter sequence (5'-GGATCCTAA- TACGACTCACTATAGGGAGACCACCATG-3')

Table 3 Mutations in MSH3, MSH6, BAX and TGFbRII in tumors with typical (severe) microsatellite instability with or without mutation in MSH2 and MLH1 genes Tumor code MSH3 MSH6 BAX TGFbRII MSH2 or MLH1 mutation Referencec c17 +a 7b + + none Wu et al., 1997 c18 7 7 7 + none Wu et al., 1997 c39 7 7 7 + none Wu et al., 1997 c60 7 7 7 7 none Wu et al., 1997 c76 7 7 7 + none Wu et al., 1997 c104 + 7 7 + none Wu et al., 1997 c136 7 7 7 + none Wu et al., 1997 c139 7 7 7 + none Wu et al., 1997 c144 7 7 7 7 none Wu et al., 1997 c171 7 7 7 + none Wu et al., 1997 c184 7 7 7 7 none Wu et al., 1997 Total (%) 2/11 (18%) 0/11 (0%) 1/11 (9%) 8/11 (73%) f1e + 7 + + MLH1, exon 16, germlined NystroÈ m-Lahti et al., 1996 f2e 7 7 + + MLH1, exon 16, germlined,f NystroÈ m-Lahti et al., 1996 f7e 7 7 + + MLH1, exon 17, germlinef NystroÈ m-Lahti et al., 1996 f10e 7 7 + + MLH1, exon 16, germlined,f NystroÈ m-Lahti et al., 1996 f11Ae + 7 + 7 MLH1, exon 16, germlined,f NystroÈ m-Lahti et al., 1996 F11Be + 7 + + MLH1, exon 16, germlined NystroÈ m-Lahti et al., 1996 f11Ce + 7 7 7 MLH1, exon 16, germlined NystroÈ m-Lahti et al., 1996 f19e + + + + MLH1, exon 16, germlined NystroÈ m-Lahti et al., 1996 f30e + 7 + + MLH1, exon 16, germlined NystroÈ m-Lahti et al., 1996 f59e + 7 + + MLH1, exon 16, germlined NystroÈ m-Lahti et al., 1996 f83e 7 7 7 + MLH1, exon 17, germlinef NystroÈ m-Lahti et al., 1996 f39e 7 7 + + MLH1, exon 12, germline NystroÈ m-Lahti et al., 1996 c35e + 7 7 + MLH1, exon 16, germlined Aaltonen et al., 1998 c64 + + + + MLH1, exon 6, germline Wu et al., 1997; MLH1, exon 9, somatic Aaltonen et al., 1998 c125e 7 7 7 7 MLH1, exons 3,4,5, germline Nystrom-Lahti et al., 1996 c558e 7 7 7 + MLH1, exon 4, germline Aaltonen et al., 1998 c10 + + 7 7 MLH1, exon 2, somatic Wu et al., 1997 c43 + + + + MLH1, exon 11, somatic Wu et al., 1997 c54 7 7 7 7 MSH2, exon 9, germlinef Wu et al., 1997 c11 + 7 + 7 MSH2, exon 7, somatic Wu et al., 1997 c56 7 7 + + MSH2, exon 13, somaticf Wu et al., 1997 c239 7 7 + 7 MSH2, exon 15, somatic Wu et al., 1997 Total (%) 12/22 (55%) 4/22 (18%) 15/22 (68%) 15/22 (68%) Total over 14/33 (42%) 4/33 (12%) 16/33 (48%) 23/33 (70%) all tumors (%) aMutation present; bMutation absent; cFor MSH2 and MLH1 mutations; dReferred to as mutation 1; eOriginates from a family meeting the Amsterdam criteria for HNPCC (Vasen et al., 1991); fLoss of heterozygosity observed in tumor (Hemminki et al., 1994; Wu et al., 1997) Genetic basis of microsatellite instability A Percesepe et al 160 with `severe' RER (n=33), 22 of which were known to occurred only in tumors with demonstrable mutations have somatic and/or germline mutations in MSH2 or in MSH2 or MLH1 (Table 3 and Figure 1b). When the MLH1 while no mutations in these genes had been four genes were analysed separately in tumors without identi®ed in 11 (NystroÈ m-Lahti et al., 1996; Wu et al., vs with MSH2 or MLH1 mutations (groups II and III 1997; Aaltonen et al., 1998) (Figure 1a). Interestingly, in Figure 1b), TGFb RII showed equally high the overall mutation rates, expressed as the proportions mutation rates in both groups while MSH3, MSH6, of mutated loci among all studied MSH3, MSH6, and BAX were more frequently mutated in the latter BAX, and TGFb RII loci, varied remarkably between group (for BAX, the di€erence was statistically these groups in that `mild' RER tumors showed no signi®cant with P=0.004). There was no apparent mutations at all whereas in tumors with `severe' RER, di€erence in the patterns of mononucleotide tract the rate was signi®cantly higher in those showing instability between MSH2 vs MLH1-associated tu- MSH2 or MLH1 mutations, compared to the mors, but the number of available tumors with remaining tumors with a similar RER phenotype (46/ MSH2 mutations was too small to make signi®cant 88, 52% vs 11/44, 25%, P=0.005) (Table 3). comparisons. Calculated per tumors, frameshift mutations were found in up to 70% of tumors with `severe' RER; furthermore, MSH3, BAX and TGFb RII were Discussion frequently involved (42 ± 70%) whereas MSH6 muta- tions were far less common (12%; P50.001) and There is no consensus as to when a tumor should be designated as RER positive or RER negative. Initial studies suggested a distinction between single locus (41) and multilocus (52) involvement; however, the validity of this distinction strongly depends on the number of loci analysed, and later studies have a typically required that for a tumor to be considered RER positive at least 30 ± 50% of microsatellite loci should be unstable, especially if the RER status is used to determine whether one should screen for germline mutations in DNA mismatch repair genes, i.e. undiagnosed cases of HNPCC (Canzian et al., 1996; Moslein et al., 1996; Dunlop et al., 1997). A further possible source of variation in the detection rates of microsatellite instability is the type of marker used: for example, poly(A) repeats (Hoang et al., 1997) and tri- and tetranucleotide repeats (Mao et al., 1994) have been suggested to be more prone to deletion or expansion than dinucleotide repeats. Apart from HNPCC, the genetic basis of microsatellite instability in various tumors is poorly understood, and the possibility exists that di€erent patterns of instability indicate di€erent underlying genes (see Introduction). We de®ned the RER status based on both the b number of unstable loci and the type of repeats involved. The tumors with `mild' RER were part of a larger series of sporadic colorectal cancers analysed for RER (Canzian et al., 1996; Aaltonen et al., 1998). In regard to clinicopathological characteristics the present series of `mild' RER tumors was closer to the RER negative group than the group with typical RER but did not completely fall into either category. For example, the ratio of proximal to distal tumors was 40 ± 60%, being somewhat higher than in the RER negative tumors (27%/73%; P=0.07) but signi®cantly lower than in the RER positive ones (77%/23%; P50.001). Most tumors with `mild' RER were local representing Dukes stage A or B (67%); the percentage was signi®cantly higher than that for the RER negative tumors (58%, P=0.001) but lower than that for the RER positive ones (71%, P=0.59). The average age at Figure 1 (a) Examples of mononucleotide tract instability in the diagnosis was relatively high (69.9 years) and similar to MSH3, MSH6, TGFb RII and BAX genes. The ampli®ed products represent paired normal mucosa (N) and tumor tissue that calculated for all sporadic tumors, no matter (T) from the indicated samples, including `mild' (59) and `severe' whether RER negative or positive. (f11B, c10, c11, c64) RER phenotypes. For the interpretation of Even though the `mild' RER phenotype was the results, see Table 3. (b) Summary of mutation frequencies (% characterized by a speci®c pattern of microsatellite of tumors) in the above mentioned mononucleotide tracts in `mild' RER tumors (I) as well as in `severe' RER tumors without instability (consisting of alterations at mono- or (II) or with (III) detectable mutations in MSH2 or MLH1 tetranucleotide repeats in the absence of widespread Genetic basis of microsatellite instability A Percesepe et al 161 dinucleotide repeat changes) and a distinct clinico- mismatch repair function, as well as DNA polymerase pathological pro®le, mutation analyses of DNA d (da Costa et al., 1995), are obvious candidates. mismatch repair genes did not result in any Recently, a mechanism of DNA mismatch repair gene distinguishing feature. No mutations were found either inactivation not involving structural changes in coding in genes associated with typical RER (MSH2 and DNA was demonstrated by Kane et al. (1997) who MLH1) or in genes suggested to be responsible for a showed a correlation between hypermethylation of the milder RER phenotype (MSH3 and MSH6). The MLH1 promoter region and lack of expression of this screening strategy used for MSH2 and MLH1, RT ± gene in tumors with no detectable MLH1 mutations. It PCR of cDNA and direct sequencing of genomic is possible that a proportion of sporadic tumors with DNA, was expected to reveal most types of mutations microsatellite instability but no apparent mismatch including large deletions as well as point mutations in repair gene mutations are attributable to inactivation the exons, ¯anking introns, and the promoter region, of these genes by such epigenetic or other mechanisms. and its sensitivity was supported by the detection of Furthermore, phenotypic manifestations of DNA numerous neutral sequence variants (see Results). The mismatch repair gene mutations, including the degree IVSP method used to screen for MSH3 and MSH6 of mismatch repair de®ciency and thus severity of mutations was expected to be somewhat less sensitive, microsatellite instability, may vary according to the being able to detect truncating mutations but not nature of the gene defect, for example, depending on missense changes. Although technical shortcomings the ability of the mutated protein to inactivate the wild cannot be excluded as possible explanations for the type protein by a dominant negative mechanism lack of mutations in cases with `mild' RER it is, (Parsons et al., 1995a; JaÈ ger et al., 1997). Studies to however, noteworthy that mutation analyses have not address these di€erent possibilities in our tumor series been able to identify the defective gene even in a are in progress. majority of sporadic colorectal tumors with typical RER (Eshleman and Markowitz, 1996). It is therefore possible that such as yet unidenti®ed genes that are Materials and methods involved in the latter cases also turn out to be important in the `mild' RER tumors. Alternatively, Tumor selection the `mild' RER phenotype may simply re¯ect back- De®nition of `mild' RER The tumors represented a series ground instability caused by the generally increased of 450 consecutive colorectal tumors previously analysed mutation rate in tumors. for RER status using 6 ± 14 dinucleotide repeat markers Di€erences in mutation frequencies in the mono- (Canzian et al., 1996). Among those, we selected all nucleotide tracts of MSH3, MSH6, BAX and TGFb samples (n=49) that were unstable at one or two loci, RII suggested the presence of two subcategories of but did not meet the working de®nition of RER positivity tumors with typical RER in accordance with a division requiring that at least three markers or at least 30% of the based on the presence vs absence of MSH2 or MLH1 markers studied should be positive. These tumors were mutations in these tumors. Although con®rmation in a then studied with six additional markers detecting poly(A) larger series is necessary, our ®nding of the combined or tetranucleotide repeats. Tumors with none of the markers being positive were considered RER negative mutation rates of MSH3, MSH6, BAX and TGFb RII and were not studied further. Tumors with at least one being clearly higher for tumors with demonstrable of these markers positive (n=15) were designated as having MSH2 or MLH1 mutations as compared to those `mild' RER and were selected for analysis of mutations in without mutations in these genes, implies that in DNA mismatch repair genes. For comparison, we also addition to an MSH2/MLH1-associated pathway, a studied a number of typical (`severe') RER tumors with the parallel `major' pathway is likely to exist, involving as same mono- or tetranucleotide repeat markers and all yet unknown mechanisms. Di€erent tumorigenic path- tumors showed instability in the mono- and tetranucleotide ways in these two groups with high-degree instability repeats (data not shown). were further highlighted by di€erences observed in the Typical RER Thirty-three colorectal tumors were investi- mutation rates between individual target genes. Thus, gated, 22 of which were known to have germline and/or while TGFb RII showed a frequent involvement (70%) somatic mutations in MSH2 or MLH1, whereas no such in all tumors with severe RER, which is in accordance mutations had been detected in 11 (Wu et al., 1997; with previous reports of an important role of this gene NystroÈ m-Lahti et al., 1996; Aaltonen et al., 1998). in unstable colorectal tumors (Markowitz et al., 1995; Parsons et al., 1995b). MSH3, MSH6, and particularly Samples BAX frameshift mutations were much more common RER and mutation analyses were conducted on DNA in the group with MSH2 or MLH1 mutations than in derived from fresh-frozen tumor tissue and paired normal the remaining tumors with severe RER. Furthermore, mucosa or blood. The tissue specimens were cut into 5 mm in keeping with the proposed partial redundancy of sections, and the proportion of tumor cells was estimated MSH3 and MSH6 proteins in DNA mismatch repair by microscopy (all were above 60%). From these samples, (Marsischky et al., 1996) it is of note that MSH6 DNA was extracted according to a non-enzymatic protocol mutations only occurred in the presence of MSH3 (Lahiri and Nurnberger, 1991). The RNeasy Total RNA changes, and mutation of both genes possibly increased kit (Qiagen GmbH, Hilden, Germany) was used for RNA genetic destabilization, as illustrated by the high rate of extraction, according to the conditions suggested by the concomitant BAX and TGFb RII mutations in these manufacturer. tumors (Table 3). Genes that underlie microsatellite instability in cases Assessment of instability at microsatellite sequences with no apparent MSH2 or MLH1 mutations remain Methods for radioactive labeling and analysis of micro- to be identi®ed. Other known genes with DNA satellites have been previously described (Aaltonen et al., Genetic basis of microsatellite instability A Percesepe et al 162 1993). The markers used to de®ne the `mild' RER category reactions were used for all fragments so that a separate were: BAT 25, BAT 26, BAT 40, APD3, MYCL1, D19S244 nesting forward primer was combined with the reverse (Ionov et al., 1993; Parsons et al., 1995b; http:// primer in the ®rst run and the usual forward primer gdbwww.gdb.org/). Frameshift mutations in the polynu- containing the consensus transcription-translation signal

cleotide tracts of MSH3 [(A)8] (Malkhosyan et al., 1996), was combined with the same reverse primer in the second

MSH6 [(C)8] (Malkhosyan et al., 1996), TGFb RII [(A)10] run. The primers used for IVSP are listed in Table 2. The

(Parsons et al., 1995b), and BAX [(G)8] (Rampino et al., PCR reactions were carried out using 35 cycles, each 1997) were studied using the published primer sequences. consisting of 948C for 30 min, 558C for 1 h and 728C for 1 h, followed by a ®nal extension at 728C for 8 h. The TNT T7 Coupled Reticulocyte Lysate System (Promega, Analysis of mutations Madison, WI) was used for the in vitro synthesis of MSH2 and MLH1 To screen for possible splice site protein, according to the manufacturer's instructions, and mutations and large deletions, the MSH2 and MLH1 the resulting proteins were size-separated through 4 ± 20% cDNAs were studied in ®ve and six overlapping fragments, polyacrylamide gradient gels (Novex, San Diego, CA). respectively, by RT ± PCR as previously described (NystroÈ m-Lahti et al., 1996). Genomic DNA samples Statistical analysis were studied by direct exon by exon sequencing of ampli®cation products generated with primers that were Chi-square test was used for assessing di€erences between located in the ¯anking introns approximately 50 bp from the groups. the respective intron/exon borders to detect all possible splice junction mutations (Chadwick et al., 1996). The sequences were determined on an Applied Biosystems 373 Acknowledgements DNA sequencer using ¯uorescently labeled primers and Dr Reijo Salovaara is acknowledged for his expertise in the protocols supplied by the manufacturer (Applied Biosys- processing of pathological specimens. This study was tems, Foster City, CA). supported by grants from the Finnish Cancer Founda- MSH3 and MSH6 The MSH3 and MSH6 cDNAs were tion, the Academy of Finland, the Sigrid Juselius Founda- divided into three overlapping fragments for the IVSP tion, European Commission (contract BMH4-CT96-0772) analysis according to Papadopoulos et al. (1995) and and the US National Institutes of Health grant CA67941. Risinger et al. (1996), respectively, with modi®cations. Part of this study was done at the Folkhalsan Institute of The most important modi®cation was that half-nested Genetics.

References

Aaltonen LA, PeltomaÈ ki P, Leach FS, Sistonen P, PylkkaÈ nen JaÈ ger AC, Bisgaard ML, Myrhoj T, Bernstein I, Rehfeld JF L, Mecklin, J-P, JaÈ rvinen H, Powell SM, Jen J, Hamilton and Nielsen FC. (1997). Am. J. Hum. Genet., 61, 129 ± 133. SR, Petersen GM, Kinzler KW, Vogelstein B and de la Kane MF, Loda M, Gaida GM, Lipman J, Mishra R, Chapelle A. (1993). Science, 260, 812 ± 816. Goldman H, Jessup JM and Kolodner R. (1997). Cancer Aaltonen LA, Salovaara R, Kristo P, Canzian F, Hemminki Res., 57, 808 ± 811. A, PeltomaÈ ki P, Chadwick RB, Percesepe A, KaÈ aÈ riaÈ nen Kim H, Jen J, Vogelstein B and Hamilton SR. (1994). Am. J. H, Ahtola H, Eskelinen M, HaÈ rkoÈ nen N, Julkunen R, Pathol., 145, 148 ± 156. Kangas E, Ojala S, Tulikoura J, Valkamo E, JaÈ rvinen H, Konishi M, Kikuchi Yanoshita R, Tanaka K, Muraoka M, Mecklin J-P and de la Chapelle A. (1998). New Engl. J. Onda A, Okumura Y, Kishi N, Iwama T, Mori T, Koike Med., in press. M, Ushio K, Chiba M, Nomizu S, Konishi F, Utsunomiya Akiyama Y, Sato H, Yamada T, Nagasaki H, Tsuchiya A, J and Miyaki M. (1996). Gastroenterology, 111, 307 ± 317. Abe R and Yuasa Y. (1997). Cancer Res., 57, 3920 ± 3923. Lahiri DK and Nurnberger Jr, JI. (1991). Nucleic Acids Res., Bùrresen AL, Lothe RA, Meling GI, Lystad S, Morrison P, 19, 5444. Lipford J, Kane MF, Rognum TO and Kolodner RD. Leach FS, Nicolaides NC, Papadopoulos N, Liu B, Jen J, (1995). Hum. Mol. Genet., 4, 2065 ± 2072. Parsons R, PeltomaÈ ki P, Sistonen P, Aaltonen LA, Canzian F, Salovaara R, Hemminki A, Kristo P, Chadwick NystroÈ m-Lahti M, Guan X-Y, Zhang J, Meltzer PS, Yu RB, Aaltonen LA and de la Chapelle A. (1996). Cancer J-W, Kao F-T, Chen DJ, Cerosaletti KM, Fournier REK, Res., 56, 3331 ± 3337. Toss S, Lewis T, Leach RJ, Naylor SL, Weissenbach J, Chadwick RB, Conrad MP, McGinnis MD, Johnston-Dow Mecklin J-P, JaÈ rvinen H, Petersen GM, Hamilton SR, L, Spurgeon SL and Kronick MN. (1996). Biotechniques, Green J, Jass J, Watson P, Lynch HT, Trent JM, de la 20, 676 ± 683. Chapelle A, Kinzler KW and Vogelstein B. (1993). Cell, Da Costa LT, Liu B, El Deiry WS, Hamilton SR, Kinzler 75, 1215 ± 1225. KW, Vogelstein B, Markowitz S, Willson JKV, de la Liu B, Parsons R, Papadopoulos N, Nicolaides NC, Lynch Chapelle A, Downey KM and So AG. (1995). Nat. Genet., HT, Watson P, Jass JR, Dunlop M, Wyllie A, PeltomaÈ ki 9, 10 ± 11. P, de la Chapelle A, Hamilton SR, Vogelstein B and Dunlop MG, Farrington SM, Carothers AD, Wyllie AH, Kinzler KW. (1996). Nat. Med., 2, 169 ± 174. Sharp L, Burn J, Liu B, Kinzler KW and Vogelstein B. Loeb LA. (1991). Cancer Res., 51, 3075 ± 3079. (1997). Hum. Mol. Genet., 6, 105 ± 110. Lothe RA, PeltomaÈ ki P, Meling GI, Aaltonen LA, NystroÈ m- Eshleman JR and Markowitz SD. (1996). Hum. Mol. Genet., Lahti M, PylkkaÈ nen L, Heimdal K, Andersen TI, Moller 5, 1489 ± 1499. P, Rognum TO, FossaÈ SD, Haldorsen T, Langmark F, Fearon ER and Vogelstein B. (1990). Cell, 61, 759 ± 767. Brùgger A, de la Chapelle A and Bùrresen AL. (1993). Hemminki A, PeltomaÈ ki P, Mecklin J-P, JaÈ rvinen H, Cancer Res., 53, 5849 ± 5852. Salovaara R, NystroÈ m-Lahti M, de la Chapelle A and Malkhosyan S, Rampino N, Yamamoto H and Perucho M. Aaltonen LA. (1994). Nat. Genet., 8, 405 ± 410. (1996). Nature, 382, 499 ± 500. Hoang JM, Cottu PH, Thuille B, Salmon RJ, Thomas G and Mao L, Lee DJ, Tockman MS, Erozan YS, Askin F and Hamelin R. (1997). Cancer Res., 57, 300 ± 303. Sidransky D. (1994). Proc. Natl. Acad. Sci. USA, 91, Ionov Y, Peinado MA, Malkhosyan S, Shibata D and 9871 ± 9875. Perucho M. (1993). Nature, 363, 558 ± 561. Genetic basis of microsatellite instability A Percesepe et al 163 Markowitz SD, Wang J, Myero€ L, Parsons R, Sun L, PeltomaÈ ki P and de la Chapelle A. (1997). Adv. Cancer Res., Lutterbaugh J, Fan RS, Zborowska E, Kinzler KW, 71, 93 ± 111. Vogelstein B, Brattain M and Wilson JKW. (1995). Rampino N, Yamamoto H, Ionov Y, Li Y, Sawai H, Reed Science, 268, 1336 ± 1338. JC and Perucho M. (1997). Science, 275, 967 ± 969. Marsischky GT, Filosi N, Kane MF and Kolodner R. (1996). Risinger JI, Umar A, Boyd J, Berchuck A, Kunkel TA and Genes Dev., 10, 407 ± 420. Barrett JC. (1996). Nat. Genet., 14, 102 ± 105. Moslein G, Tester DJ, Lindor NM, Honchel R, Cunningham Souza RF, Appel R, Yin J, Wang S, Smolinski KN, JM, French AJ, Halling KC, Schwab M, Goretzki P and Abraham JM, Zou TT, Shi YQ, Lei J, Cottrell J, Cymes Thibodeau SN. (1996). Hum. Mol. Genet., 5, 1245 ± 1252. K, Biden K, Simms L, Leggett B, Lynch PM, Frazier M, Nicolaides NC, Palombo F, Kinzler KW, Vogelstein B and Powell SM, Harpaz N, Sugimura H, Young J and Meltzer Jiricny J. (1996). Genomics, 31, 395 ± 397. PS. (1996). Nat. Genet., 14, 255 ± 257. NystroÈ m-Lahti M, Wu Y, Moisio A-L, Hofstra RMW, Strand M, Earley MC, Crouse GF and Petes TD. (1995). Osinga J and Mecklin J-P. (1996). Hum. Mol. Genet., 5, Proc. Natl. Acad. Sci. USA, 92, 10418 ± 10421. 763 ± 769. Thibodeau SN, Bren G and Schaid D. (1993). Science, 260, Papadopoulos N, Nicolaides NC, Liu B, Parsons R, 816 ± 819. Lengauer C, Palombo F, D'Arrigo A, Markowitz S, Vasen HFA, Mecklin J-P, Meera Khan P and Lynch HT. Willson JKV, Kinzler KW, Jiricny J and Vogelstein B. (1991). Dis. Colon Rectum, 34, 424 ± 425. (1995). Science, 268, 1915 ± 1917. Wijnen J, Fodde R and Meera-Khan P. (1994). Hum. Mol. Parsons R, Li GM, Longley M, Modrich P, Liu B, Berk T, Genet., 3, 2268. Hamilton S, Kinzler KW and Vogelstein B. (1995a). Wu Y, NystroÈ m-Lahti M, Osinga J, Looman MW, Science, 268, 738 ± 740. PeltomaÈ ki P, Aaltonen LA, de la Chapelle A, Hofstra Parsons R, Myero€ LL, Liu B, Willson JK, Markowitz SD, RM and Buys CH. (1997). Genes Chrom. Cancer, 18, 269 ± Kinzler KW and Vogelstein B. (1995b). Cancer Res., 55, 278. 5548 ± 5550.