Isolation of an hMSH2-p160 heterodimer that restores DNA mismatch repair to tumor cells

Molecular basis of HNPCC: mutations of MMR genes

Hereditary nonpolyposis colorectal cancer (HNPCC) is inherited as a dominant disorder caused by germline defects in one of at least four mismatch repair (MMR) genes. Two of these genes, hMSH2 and hMLH1, account for the vast majority of the germline mutations in HNPCC kindreds, whereas hPMS1 and hPMS2 are mutated in only few families. MMR genes also are susceptible to somatic mutations in sporadic tumors. The mutational spectrum of the MMR genes shows no predominant type of mutation. Furthermore, the mutations are spread throughout the length of the genes, with no significant hot spots. Identification of MMR genes as the cause of HNPCC made presymptomatic diagnosis a reality. However, the presence of multiple genes and the heterogeneity of mutations present challenges to the development of diagnostic tests for this disease.

DHFR/MSH3 amplification in methotrexate-resistant cells alters the hMutSalpha/hMutSbeta ratio and reduces the efficiency of base-base mismatch repair

The level and fate of hMSH3 (human MutS homolog 3) were examined in the promyelocytic leukemia cell line HL-60 and its methotrexate-resistant derivative HL-60R, which is drug resistant by virtue of an amplification event that spans the dihydrofolate reductase (DHFR) and MSH3 genes. Nuclear extracts from HL-60 and HL-60R cells were subjected to an identical, rapid purification protocol that efficiently captures heterodimeric hMutSalpha (hMSH2. hMSH6) and hMutSbeta (hMSH2.hMSH3). In HL-60 extracts the hMutSalpha to hMutSbeta ratio is roughly 6:1, whereas in methotrexate-resistant HL-60R cells the ratio is less than 1:100, due to overproduction of hMSH3 and heterodimer formation of this protein with virtually all the nuclear hMSH2. This shift is associated with marked reduction in the efficiency of base-base mismatch and hypermutability at the hypoxanthine phosphoribosyltransferase (HPRT) locus. Purified hMutSalpha and hMutSbeta display partial overlap in mismatch repair specificity: both participate in repair of a dinucleotide insertion-deletion heterology, but only hMutSalpha restores base-base mismatch repair to extracts of HL-60R cells or hMSH2-deficient LoVo colorectal tumor cells.

Conditional mutator phenotypes in hMSH2-deficient tumor cell lines

Two human tumor cell lines that are deficient in the mismatch repair protein hMSH2 show little or no increase in mutation rate relative to that of a mismatch repair-proficient cell line when the cells are maintained in culture conditions allowing rapid growth. However, mutations accumulate at a high rate in these cells when they are maintained at high density. Thus the mutator phenotype of some mismatch repair-deficient cell lines is conditional and strongly depends on growth conditions. These observations have implications for tumor development because they suggest that mutations may accumulate in tumor cells when growth is limited.

Molecular pathogenesis of colorectal cancer

The enormous progress made in the identification of genes that are involved in colon carcinogenesis has provided the foundation for further understanding the biology of both normal and cancer cells and for targeted therapeutic strategies. In one sense, the genes described in this review are only the building blocks of a larger puzzle that constitutes the integrated metabolic function of a cell. The current challenge is to understand the functional role of these genes in normal cellular physiology and make the connections between pathways that knit together integrated cellular homeostasis. A complete understanding of the regulatory pathways, and the synthesis and modifications of the proteins involved, will provide novel targets for therapeutic agents.

Functional domains of the Saccharomyces cerevisiae Mlh1p and Pms1p DNA mismatch repair proteins and their relevance to human hereditary nonpolyposis colorectal cancer-associated mutations

The MutL protein is an essential component of the Escherichia coli methyl-directed mismatch repair system but has no known enzymatic function. In the yeast Saccharomyces cerevisiae, the MutL equivalent, an Mlh1p and Pms1p heterodimer, interacts with Msh2p bound to mismatch-containing DNA. Little is known of the functional domains of Mlh1p and Pms1p. In this report, we define the Mlh1p and Pms1p domains required for Mlh1p-Pms1p interaction. The Mlh1p-interactive domain of Pms1p is comprised of 260 amino acids near the carboxyl terminus while the Pms1p-interactive domain of Mlh1p resides in the final 212 residues. The two domains are sufficient for Mlh1p-Pms1p interaction, as determined by the two-hybrid assay and by in vitro protein affinity chromatography. Deletions within the domains completely eliminated Mlh1p-Pms1p interaction. Using site-directed mutagenesis, we altered a number of highly conserved residues in the Mlh1p and Pms1p proteins, including some alterations that mimic germline mutations observed for human hereditary nonpolyposis colorectal cancer. Alterations either in the consensus MutL box located in the amino-terminal portion of each protein or in the carboxyl-terminal homology motif of Mlh1p eliminated DNA mismatch repair function but had no effect on Mlh1p-Pms1p interaction. In addition, certain MLH1 and PMS1 mutant alleles caused a dominant negative mutator effect when overexpressed. We discuss the implications of these findings for the structural organization of the Mlh1p and Pms1p proteins and the importance of Mlh1p-Pms1p interaction.

Frequent somatic mutations of hMSH3 with reference to microsatellite instability in hereditary nonpolyposis colorectal cancers

hMSH3 is one of the human DNA mismatch repair genes but has not yet been reported to be associated with hereditary nonpolyposis colorectal cancer. Recently, somatic mutation at a polyadenine tract, i.e., (A)8, in hMSH3 was reported in cancers with microsatellite instability (MI). To clarify the tumorigenetic role of hMSH3, we screened for somatic mutations at the hMSH3 (A)8 repeat in 29 tumors from 23 hereditary nonpolyposis colorectal cancer patients. One or two A deletions in the (A)8 repeat were found in 11 (57.9%) of the 19 MI-positive tumors but not in 10 MI-negative ones, indicating secondary mutations after germline mutations of other mismatch repair genes. Moreover, the MI frequency of three or more nucleotide repeats was higher in hMSH3 (A)8-mutated tumor cells than in nonmutated ones (p<0.05). These data suggest that a mutation of a mismatch repair gene enhances the frequency of another mismatch repair gene mutation, such as of hMSH3, resulting in severe MI.

Genetic instability in patients with metachronous colorectal cancers

BACKGROUND

Nearly 7 per cent of patients who undergo resection for colorectal cancer develop metachronous cancers several years later. A molecular marker that could identify patients susceptible to metachronous cancers would be of clinical importance.

METHODS

Twenty-four colorectal cancers from 15 individuals with metachronous colorectal cancer were investigated for microsatellite instability at five loci by single stranded conformational polymorphism analysis. A control group of 14 colorectal cancers from individuals who had only developed one sporadic colorectal cancer each was analysed similarly.

RESULTS

Microsatellite instability was demonstrated in 17 of 24 cancers from individuals with metachronous cancer compared with one of 14 cancers from individuals with a single colorectal cancer.

CONCLUSION

These results suggest that testing for microsatellite instability may be useful in recognizing patients at high risk of developing metachronous colorectal cancers.

DNA mismatch repair in mammals: role in disease and meiosis

The understanding of mammalian mismatch repair (MMR) gene function has been accelerated as a result of progress on several fronts. First, the biochemical analysis of MMR has been advanced by the production of purified human MMR proteins which will eventually allow reconstitution of MMR activity in vitro. Second, a wealth of clinical studies on colon cancer patients have begun to allow correlations to be made among MMR mutations, tumor types, therapeutic approaches and clinical outcomes. Finally, new unexpected meiotic phenotypes have been associated with mutations in certain mouse MMR genes.

Genetic and biochemical analysis of Msh2p-Msh6p: role of ATP hydrolysis and Msh2p-Msh6p subunit interactions in mismatch base pair recognition

Recent studies have shown that Saccharomyces cerevisiae Msh2p and Msh6p form a complex that specifically binds to DNA containing base pair mismatches. In this study, we performed a genetic and biochemical analysis of the Msh2p-Msh6p complex by introducing point mutations in the ATP binding and putative helix-turn-helix domains of MSH2. The effects of these mutations were analyzed genetically by measuring mutation frequency and biochemically by measuring the stability, mismatch binding activity, and ATPase activity of msh2p (mutant msh2p)-Msh6p complexes. A mutation in the ATP binding domain of MSH2 did not affect the mismatch binding specificity of the msh2p-Msh6p complex; however, this mutation conferred a dominant negative phenotype when the mutant gene was overexpressed in a wild-type strain, and the mutant protein displayed biochemical defects consistent with defects in mismatch repair downstream of mismatch recognition. Helix-turn-helix domain mutant proteins displayed two different properties. One class of mutant proteins was defective in forming complexes with Msh6p and also failed to recognize base pair mismatches. A second class of mutant proteins displayed properties similar to those observed for the ATP binding domain mutant protein. Taken together, these data suggested that the proposed helix-turn-helix domain of Msh2p was unlikely to be involved in mismatch recognition. We propose that the MSH2 helix-turn-helix domain mediates changes in Msh2p-Msh6p interactions that are induced by ATP hydrolysis; the net result of these changes is a modulation of mismatch recognition.

DNA polymerase delta is required for human mismatch repair in vitro

HeLa nuclear extract was resolved into a depleted fraction incapable of supporting mismatch repair in vitro, and repair activity was restored upon the addition of a purified fraction isolated from HeLa cells by in vitro complementation assay. The highly enriched complementing activity copurified with a DNA polymerase, and the most pure fraction contained DNA polymerase delta but was free of detectable DNA polymerases alpha and epsilon. Calf thymus DNA polymerase delta also fully restored mismatch repair to the depleted extract, indicating DNA polymerase delta is required for mismatch repair in human cells. However, due to the presence of DNA polymerases alpha and epsilon in the depleted extract, potential involvement of one or both of these activities in the reaction cannot be excluded.

Elevated levels of mutation in multiple tissues of mice deficient in the DNA mismatch repair gene Pms2

The Pms2 gene has been implicated in hereditary colon cancer and is one of several mammalian homologs of the Escherichia coli mutL DNA mismatch repair gene. To determine the effect of Pms2 inactivation on genomic integrity in vivo, hybrid transgenic mice were constructed that carry targeted disruptions at the Pms2 loci along with a chromosomally integrated mutation reporter gene. In the absence of any mutagenic treatment, mice nullizygous for Pms2 showed a 100-fold elevation in mutation frequency in all tissues examined compared with both wild-type and heterozygous litter mates. The mutation pattern in the nullizygotes was notable for frequent 1-bp deletions and insertions within mononucleotide repeat sequences, consistent with an essential role for PMS2 in the repair of replication slippage errors. Further, the results demonstrate that high rates of mutagenesis in multiple tissues are compatible with normal development and life and are not necessarily associated with accelerated aging. Also, the finding of genetic instability in all tissues tested contrasts with the limited tissue distribution of cancers in the animals, raising important questions regarding the role of mutagenesis in carcinogenesis.

DNA excision repair pathways

The major DNA excision repair pathways of base excision repair for endogenous DNA lesions and nucleotide excision repair for DNA damage inflicted by ultraviolet light have been reconstructed with purified mammalian proteins and details of these repair mechanisms are emerging. Similar data are becoming available with regard to mismatch repair for correction of replication errors. Deletion of individual DNA repair proteins in knockout mice provides information on the roles of such factors in vivo and recent three-dimensional structures of several repair enzymes explain their detailed modes of action.

DNA mismatch repair in lymphoblastoid cells from hereditary non-polyposis colorectal cancer (HNPCC) patients is normal under conditions of rapid cell division and increased mutational load

Patients with hereditary non-polyposis colorectal cancer (HNPCC) have germ-line mutations in one of four DNA mismatch repair genes (hMSH2, hMLH1, hPMS1 and hPMS2). It is supposed that a single functional copy of these genes is sufficient for normal mismatch repair, but it is not certain that this is the case under abnormal conditions such as rapid cell division or an increased tendency to DNA replication errors (RERs). We have analysed mismatch repair by examining replication errors in immortalised lymphoblastoid cells derived from two HNPCC patients heterozygous for mismatch repair defects (one hMSH2 mutant and one hMLH1 mutant), and from control individuals. Three conditions of cell culture have been used: (i) relatively slow cell growth and division; (ii) relatively fast growth and division; and (iii) chronic perturbation of the intracellular dNTP pool to promote a increased frequency of replication errors. No significant differences in microsatellite instability were found between HNPCC patients and controls in any of these environments. Lymphoblastoid cells from hMSH2 and hMLH1 mutant/wild-type heterozygotes appear, therefore, to have normal levels of mismatch repair, even under conditions that increase the requirement for repair. The pool bias cultures from the HNPCC patients and controls did, however, show similar, increased frequencies of RERs, suggesting that the mismatch repair capacity of the cells had been overloaded, but that the number of normal HNPCC alleles was not the limiting factor.

Mapping the minimal domain of hMSH-2 sufficient for binding mismatched oligonucleotides

The human MSH-2 gene product is a member of a highly conserved family of proteins involved in post-replication mismatch repair. Germline mutations in this gene have been implicated in hereditary non-polyposis colorectal cancer (HNPCC). Alterations in the coding region of the hMSH-2 gene result in a mutator phenotype with marked instability of microsatellite sequences, indicative of a deficiency in DNA repair. We have previously shown that a region of high homology between MutS proteins of different species containing a nucleotide binding domain, is sufficient to bind DNA containing specific mismatched residues. In order to determine the minimal domain of hMSH-2 necessary for binding mismatch-containing oligonucleotides, deletion analysis of the C-terminal region was performed. We have constructed a 5' and 3' deletion series, expressed each deletion as a bacterial fusion protein and assessed it for ATPase activity and its ability to identify mismatch containing DNA. Here we demonstrate that a 585 bp fragment encoding 195 amino acids within the C-terminal domain of hMSH-2 is sufficient to bind to DNA containing mismatches.

The role of DNA repair in development

Several pathways of DNA repair are essential for maintaining genomic integrity in mammalian cells. Mismatch repair is the final line of defense against polymerase errors during normal cellular replication. Base excision repair removes endogenous DNA damage resulting from normal cellular metabolism. Nucleotide excision repair removes bulky, transcription blocking, lesions resulting from endogenous and environmental insults to the DNA. The role of DNA repair in mammalian development is not well understood. Nevertheless, clues to the essential nature of these processes are evident in the human DNA repair syndromes, in the nature of the interactions between DNA repair and other proteins, and in the phenotypes of genetically engineered, knockout mice lacking functional repair genes. Questions remain: what is the relative importance of endogenous vs. environmental DNA damage and is repair itself critical for normal development or are transcription-repair interactions more crucial?

Mismatch repair provokes chromosome aberrations in hamster cells treated with methylating agents or 6-thioguanine, but not with ethylating agents

O6-Methylguanine (O6MeG) is important in induction of chromosome aberrations (abs), with the unusual property that new abs are produced in the second cycle after treatment; cells lacking repair by O6-alkylguanine DNA-alkyltransferase (AGT) have more abs at the second division (M2) than at the first (M1). These second-cycle abs are likely caused by attempted correction by mismatch repair (MMR) of O6MeG:T mispairs, since we previously showed that MMR-deficient human cells (MT1 lymphoblasts) treated with SN-1 methylating agents do not produce new abs at M2 and are resistant to killing. Here we used MMR-deficient rodent cells to examine ab induction by alkylators and by incorporated 6-thioguanine (6-tG) which produces mispairs. BrdUrd labeling was used to identify cells at first, second and third metaphase after treatment (M1, M2 and M3). MMR-deficient Chinese hamster Clone B cells were 10-fold more resistant to ab induction by methyl nitrosourea and 1-methyl-3-nitro-1-nitrosoguanidine compared to their MMR-proficient parent cells, CHO MT+. Both cell lines express AGT and can remove the methyl group from O6MeG. Clone B has twice the AGT activity of CHO MT+, but inhibition of AGT with O6-benzylguanine did not change ab induction, indicating that methylation tolerance of Clone B cells was due to defective MMR and not to increased repair of O6MeG. Confirming the importance of O6MeG in inducing abs, even when it is a minor component of the adducts induced, Clone B cells were 2-fold more resistant to ab induction by methyl methanesulfonate and dimethylsulfate, whereas they had normal sensitivity to ethyl nitrosourea and 1-ethyl-3-nitro-1-nitrosoguanidine. Clone B cells are also resistant to killing by 6-tG, and 6-tG induced few abs in MMR-deficient Clone B (6-fold lower than CHO MT+ cells). Since mispairs do not occur until the cell cycle following incorporation of 6-tG, new abs in MMR-proficient cells are expected one cell cycle later than with the methylators, i.e., at M3. As expected, in normal CHO MT+, high ab levels were seen at M3, but there was also ab induction at M2. Similarly, with methylating agents we saw higher levels of abs at M1 in the MMR-proficient CHO MT+ cells than in Clone B cells, suggesting that in the rodent cells, MMR is involved in ab formation from mispairs or modified base pairs induced in the first S-phase, such as O6MeG:C. These rodent cells thus differ from human MT1 lymphoblasts which had similar ab levels to their normal parent cells at the first metaphase after treatment with methylators.

MutS homologs in mammalian cells

Alterations of the human mismatch repair genes have been linked to hereditary non-polyposis colon cancer (HNPCC) as well as to sporadic cancers that exhibit microsatellite instability. The human mismatch repair genes are highly conserved homologs of the Escherichia coli MutHLS system. Six MutS homologs have been identified in Saccharomyces cerevisiae and four MutS homologs have been identified in human cells. At least three of these eukaryotic MutS homologs are involved in the recognition/binding of mispaired nucleotides and nucleotide lesions. MSH2 plays a fundamental role in mispair recognition whereas MSH3 and MSH6 appear to modify the specificity of this recognition. The redundant functions of MSH3 and MSH6 explain the greater prevalence of hmsh2 mutations in HNPCC families.

Recognition and repair of compound DNA lesions (base damage and mismatch) by human mismatch repair and excision repair systems

Nucleotide excision repair and the long-patch mismatch repair systems correct abnormal DNA structures arising from DNA damage and replication errors, respectively. DNA synthesis past a damaged base (translesion replication) often causes misincorporation at the lesion site. In addition, mismatches are hot spots for DNA damage because of increased susceptibility of unpaired bases to chemical modification. We call such a DNA lesion, that is, a base damage superimposed on a mismatch, a compound lesion. To learn about the processing of compound lesions by human cells, synthetic compound lesions containing UV photoproducts or cisplatin 1,2-d(GpG) intrastrand cross-link and mismatch were tested for binding to the human mismatch recognition complex hMutS alpha and for excision by the human excision nuclease. No functional overlap between excision repair and mismatch repair was observed. The presence of a thymine dimer or a cisplatin diadduct in the context of a G-T mismatch reduced the affinity of hMutS alpha for the mismatch. In contrast, the damaged bases in these compound lesions were excised three- to fourfold faster than simple lesions by the human excision nuclease, regardless of the presence of hMutS alpha in the reaction. These results provide a new perspective on how excision repair, a cellular defense system for maintaining genomic integrity, can fix mutations under certain circumstances.

Selective recognition of a cisplatin-DNA adduct by human mismatch repair proteins

The antitumor agent cis-diamminedichloroplatinum(II) (cisplatin) introduces cytotoxic DNA damage predominantly in the form of intrastrand crosslinks between adjacent purines. Binding assays using a series of duplex oligonucleotides containing a single 1,2 diguanyl intrastrand crosslink indicate that human cell extracts contain factors that preferentially recognise this type of damage when the complementary strand contains T opposite the 3', and C opposite the 5'guanine in the crosslink. Under the conditions of the band-shift assay used, little binding is observed if the positions of the T and C are reversed in the complementary strand. Similarly, duplexes containing CC or TT opposite the crosslink are recognised relatively poorly. The binding activity is absent from extracts of the colorectal carcinoma cell lines LoVo and DLD-1 in which the hMutSalpha mismatch recognition complex is inactivated by mutation. Extensively purified human hMutSalpha exhibits the same substrate preference and binds to the mismatched platinated DNA at least as well as to an identical unplatinated duplex containing a single G.T mismatch. It is likely, therefore, that human mismatch repair may be triggered by 1,2 diguanyl intrastrand crosslinks that have undergone replicative bypass.

Cloning of the cDNA encoding rat homologue of the mismatch repair gene MSH2 and its expression during spermatogenesis

A rat cDNA clone encoding the mismatch repair protein MSH2 has been isolated and characterized. The cDNA has an open reading frame of 2802 nucleotides in length coding for a protein of 933 amino acids (100 kDa). It shows significant homology to human and mouse MSH2. Northern blot analysis of rat MSH2 in the testes of rats of different ages showed maximum expression at 20 days, at which time the germ cells are undergoing premeiotic DNA replication. We observed down-regulation in the expression of rat MSH2 beyond 25 days by which time the germ cells have entered meiotic prophase.

Presence of human cellular protein(s) that specifically binds and cleaves 8-hydroxyguanine containing DNA

8-hydroxyguanine (oh8Gua) is a major form of oxygen free radical-induced DNA damage. The oh8Gua nucleotide can pair with cytosine (C) and adenine (A) nucleotides which can cause G:C to T:A transversions. It is known that multiple repair systems for the correction of the oh8Gua exist in both mammalian and bacterial cells. Using the technique of gel mobility shift assay, protein(s) bound to the oh8Gua:C base pair in short fragments of DNA was detected in cell-free extracts of a human small-cell lung cancer cell line. This DNA binding activity was specific, since it was poorly detected with an unmodified G:C base pair containing oligonucleotide duplex and was affected by neither the unmodified G:C base pair nor an oh8Gua:A base pair containing oligonucleotide duplex. The partially purified protein which selectively binds to the oh8Gua:C base pair was shown by gel filtration column chromatography to have an apparent molecular mass of 52 kDa. The column fraction which showed the highest binding activity to the oh8Gua:C base pair was found to possess an enzymatic activity that specifically cleaves the oh8Gua containing oligonucleotide strand at both the 5' and 3' sides of the oh8Gua residue. These results indicate the presence of a protein(s) that is involved in a DNA repair pathway for the correction of the oh8Gua residue in human cells.

Saccharomyces cerevisiae MSH2, a mispaired base recognition protein, also recognizes Holliday junctions in DNA

Genetic and biochemical studies have suggested that mismatch repair proteins interact with recombination intermediates to prevent recombination, or to limit the extent of formation of heteroduplex DNA during recombination between divergent DNA sequences. To test the idea that mismatch repair proteins regulate recombination by interacting with recombination intermediates, we investigated whether the Saccharomyces cerevisiae MutS homolog MSH2 could interact with Holliday junctions. Both filter-binding and electron-microscopic analysis showed that MSH2 bound to duplex DNA molecules containing Holliday junctions with a higher affinity than to control duplex DNA, single-stranded DNA or a control duplex DNA containing a mispaired base. The MSH2-Holliday junction complexes were also more stable than MSH2-duplex DNA complexes. This observation suggests that MSH2 protein could directly coordinate the interaction between mismatch repair and genetic recombination observed in genetic studies.

Mutations predisposing to hereditary nonpolyposis colorectal cancer

Since 1993 four genes have been identified that, when mutated, confer predisposition to a form of hereditary colon cancer (hereditary nonpolyposis colorectal cancer [HNPCC]). These genes belong to the Mut-related family of DNA mismatch repair genes whose protein products are responsible for the recognition and correction of errors that arise during DNA replication. Mutational inactivation of both copies of a DNA mismatch repair gene results in a profound repair defect demonstrable by biochemical assays, and in vivo this defect is presumed to lead to progressive accumulation of secondary mutations throughout the genome, some of which affect important growth-regulatory genes and, hence, give rise to cancer. To date, more than 70 different germline mutations have been detected in DNA mismatch repair genes and shown to be associated with HNPCC. Current evidence suggests that two genes, MSH2 and MLH1, account for roughly equal proportions of HNPCC kindreds, together being responsible for a majority of these families, but striking interethnic differences occur. Most mutations lead to truncated protein products. Mutation screening is quite demanding in HNPCC since, with a few exceptions, the predisposing mutations typically vary from kindred to kindred and individual mutations are scattered throughout the genes. Knowledge of the predisposing mutations allows genotype-phenotype correlations and forms the basis for further studies clarifying the pathogenesis of this disorder. In at-risk individuals, it allows predictive testing for cancer susceptibility and, consequently, appropriate clinical management of mutation carriers and noncarriers.

DNA repair and colorectal cancer

The mismatch repair system plays a major role in the processing of recombination intermediates and in the repair of errors made during DNA replication or resulting from chemical damage to DNA. Human homologues of the bacterial and yeast mismatch repair genes have been recently identified, and mutations in these genes have been found to show risk for tumor development in hereditary nonpolyposis colorectal cancer syndrome (HNPCC). Colorectal tumors bearing homozygous mutations in these mismatch repair genes show a hypermutable phenotype, mainly at microsatellite regions of DNA. The temporal relationship between the loss of mismatch repair activity and the cascades of mutations in critical genes involved in the carcinogenesis of HNPCC tumors is unknown.

Microsatellite instability in early onset and familial colorectal cancer

Hereditary non-polyposis colorectal cancer syndrome (HNPCC) is often considered to be the most common form of inherited colorectal cancer, although its precise incidence is unknown. The clinical diagnosis of HNPCC relies on a combination of family history and young age of onset of colorectal cancer, but as many familial aggregations of colorectal cancer do not fulfil the strict diagnostic criteria, HNPCC might be underdiagnosed. The majority of HNPCC families have germline mutations in mismatch repair (MMR) genes, such as MSH2 or MLH1, so that HNPCC cancers characteristically exhibit DNA replication errors (RERs) at microsatellite loci. Although an RER positive phenotype in tumours can also result from somatic mutations in an MMR gene, the prevalence of RER + tumours should provide a maximum estimate of the incidence of germline MMR gene mutations in patients with early onset and familial colorectal cancer. We investigated colorectal cancers for RERs from (1) a population based study of 33 patients with colorectal cancer aged 45 years or less, (2) 65 kindreds with familial colorectal cancer which only partially fulfilled the criteria for the diagnosis of HNPCC, and (3) 18 cancers from 12 HNPCC kindreds. Seven of 33 patients (21%) with colorectal cancer aged 45 years or less had an RER + cancer, with only two of these having a clear family history of HNPCC. A greater proportion of RER + tumours (5/7) occurred proximal to the splenic flexure than RER - tumours (4/26; chi2 = 6.14, p < 0.025). RERs were detected in all 18 cancers from HNPCC patients but in only six of 65 non-HNPCC familial colorectal cancer kindreds (9%; chi2 = 52.2, p < 0.0005). These findings suggest that most cancers in patients diagnosed at 45 years of age or less and familial aggregations of colorectal cancer which do not fulfil HNPCC diagnostic criteria do not have germline mutations in MSH2 and MLH1. Hence population screening for germline mutations in these genes is unlikely to be an efficient strategy for identifying people at high risk of developing colorectal cancer.

Extracolonic features of familial adenomatous polyposis in patients with sporadic colorectal cancer

We have investigated the occurrence of attenuated extracolonic manifestations (AEMs) of familial adenomatous polyposis (FAP) in patients with non-polyposis colorectal cancer. In a prospective case-control study, we observed that significantly more colorectal cancer patients exhibited AEM than did age and sex-matched controls (19.5% vs 7.5%, P < 0.004). However patients with AEMs do not have occult FAP, as we found no heterozygous adenomatous polyposis coli (APC) gene mutations despite extensive analysis of constitutional DNA. Genome-wide DNA replication errors (RERs) occur in a proportion of colorectal cancers, particularly right-sided lesions and in almost all tumours from hereditary non-polyposis colorectal cancer (HNPCC) patients. As AEMs have been reported in familial colon cancer cases, we investigated the relationship of AEMs to tumour RER phenotype. There was indeed an excess of AEMs in patients with right-sided tumours (30.2% of 53 patients vs 14.7% of 116 patients, P < 0.03) and in those with RER tumours (3 out of 12 patients with RER tumours vs none out of 21 patients with non-RER tumours, P < 0.05). Two patients with AEM were from HNPCC families compared with none of those without AEM (P < 0.05). The association of AEMs with colorectal cancer is intriguing, and we speculate that it may be a manifestation of mutational mosaicism of the APC gene, perhaps associated with a constitutional defect in DNA mismatch pair.

Analysis for microsatellite instability and mutations of the DNA mismatch repair gene hMLH1 in familial gastric cancer

We examined 30 gastric-cancer patients with a varying degree of family history of stomach cancer and/or synchronous gastric tumors for microsatellite instability. We observed microsatellite instability at at least 1 of 8 loci tested in tumors of 14/30 patients; of these 14, 8 had single locus alterations and 6 had alterations at at least half of the 8 loci. Among the patients with microsatellite instability at > or = 4 loci, 3 patients showed a strong familial clustering of gastric cancer. Mutation analysis of the DNA mismatch repair gene hMLHl on paired non-tumorous and tumor DNA from 10 patients, 6 with microsatellite instability at > or = 4 loci and 4 with an alteration at one locus, revealed a novel missense mutation, present in the normal and tumor DNA of one patient with microsatellite instability at multiple loci in his tumor. His family history of cancer included one second-degree relative affected with gastric cancer. These data suggest that germline mutations in the hMLHl gene occur in some gastric-cancer patients and that in the majority of cases microsatellite instability in gastric tumors may be due to defects in other genes responsible for DNA replication fidelity than the hMLHl.

hMSH2 forms specific mispair-binding complexes with hMSH3 and hMSH6

The genetic and biochemical properties of three human MutS homologues, hMSH2, hMSH3, and hMSH6, have been examined. The full-length hMSH6 cDNA and genomic locus were isolated and characterized, and it was demonstrated that the hMSH6 gene consisted of 10 exons and mapped to chromosome 2p15-16. The hMSH3 cDNA was in some cases found to contain a 27-bp deletion resulting in a loss of nine amino acids, depending on the individual from which the cDNA was isolated. hMSH2, hMSH3, and hMSH6 all showed similar tissue-specific expression patterns. hMSH2 protein formed a complex with both hMSH3 and hMSH6 proteins, similar to protein complexes demonstrated by studies of the Saccharomyces cerevisiae MSH2, MSH3, and MSH6. hMSH2 was also found to form a homomultimer complex, but neither hMSH3 nor hMSH6 appear to interact with themselves or each other. Analysis of the mismatched nucleotide-binding specificity of the hMSH2-hMSH3 and hMSH2-hMSH6 protein complexes showed that they have overlapping but not identical binding specificity. These results help to explain the distribution of mutations in different mismatch-repair genes seen in hereditary nonpolyposis colon cancer.

Evidence for involvement of yeast proliferating cell nuclear antigen in DNA mismatch repair

DNA mismatch repair plays a key role in the maintenance of genetic fidelity. Mutations in the human mismatch repair genes hMSH2, hMLH1, hPMS1, and hPMS2 are associated with hereditary nonpolyposis colorectal cancer. The proliferating cell nuclear antigen (PCNA) is essential for DNA replication, where it acts as a processivity factor. Here, we identify a point mutation, pol30-104, in the Saccharomyces cerevisiae POL30 gene encoding PCNA that increases the rate of instability of simple repetitive DNA sequences and raises the rate of spontaneous forward mutation. Epistasis analyses with mutations in mismatch repair genes MSH2, MLH1, and PMS1 suggest that the pol30-104 mutation impairs MSH2/MLH1/PMS1-dependent mismatch repair, consistent with the hypothesis that PCNA functions in mismatch repair. MSH2 functions in mismatch repair with either MSH3 or MSH6, and the MSH2-MSH3 and MSH2-MSH6 heterodimers have a role in the recognition of DNA mismatches. Consistent with the genetic data, we find specific interaction of PCNA with the MSH2-MSH3 heterodimer.

Genomic instability in colorectal cancers in Turkey

Microsatellite instability in a subset of colorectal cancers from North America, Europe and Japan has been reported. We examined 88 colorectal cancers from Turkish patients. Five different microsatellite loci (1 mono- and 4 dinucleotide repeat regions) were tested. Eight tumors displayed replication errors (RERs) with at least 2 different markers. Right-sided tumors showed significantly higher frequency of microsatellite instability compared with left-sided tumors. The frequency of RER phenotype was slightly higher in tumors occurring in younger (<50 years old) than in older patients (13% vs. 8%), and there was no association between sex and genomic instability. The frequency of genomic instability in our study group was 9%, whereas the reported frequencies in tumors from other countries varied between 12% and 16%.

Loss of heterozygosity and base substitution at the APRT locus in mismatch-repair-proficient and -deficient colorectal carcinoma cell lines

We determined the nature of mutations occurring at the autosomal APRT locus in mismatch-repair-proficient and -deficient colorectal carcinoma cell lines. The analysis of mutations that result in APRT deficiency in a mismatch-repair-deficient strain of DLD-1 heterozygous for this locus enabled us to measure the rate of loss of the wild-type gene through deletion, recombination, or gene conversion as well as the rate of point mutation. The overall rate of mutation at the APRT locus in DLD-1 was elevated 100-fold compared with the mismatch-repair-proficient colorectal carcinoma cell line SW620. Loss of heterozygosity (LOH) at APRT accounted for only 4 to 9% of mutant strains derived from DLD-1, indicating a rate for these types of events of 4 x 10(-7) to 9 x 10(-7). In SW620 the rate of LOH at APRT was about 10-fold higher. LOH was not found at polymorphic markers within the same chromosome subband as APRT, indicating that only a limited portion of the chromosome was affected by these alterations. Chromosome painting of SWS620 mutants revealed that the loss of APRT occurred together with a substantial portion of the long arm of chromosome 16. Differences in the nature of base substitutions at APRT (e.g., the proportion of mutations resulting from transitions or transversions) in these tumor cell lines were also detected. There was also an important similarity---the presence of a mutant APRT gene with multiple base substitutions that may be the result of some sort of error-prone DNA synthesis.

Drug-related killings: a case of mistaken identity

Abortive attempts at DNA repair can contribute to the effects of DNA damage inflicted by cytotoxic drugs. DNA methylation damage, 6-thioguanine and cisplatin adducts all owe their cytotoxicity in part to the intervention of DNA mismatch repair.

Processing of O6-methylguanine by mismatch correction in human cell extracts

Human cell extracts perform an aberrant form of DNA synthesis on methylated plasmids [1], which represents processing of O6-methylguanine (O6-meG). Here, we show that extracts of colorectal carcinoma cells with defects in the mismatch repair proteins that normally correct replication errors do not carry out this synthesis. hMSH2-defective LoVo cell extracts (hMSH for human MutS homologue) performed O6-meG-dependent DNA synthesis only after the addition of the purified hMutS alpha mismatch recognition complex. Processing of O6-meG by mismatch correction requires PCNA and therefore probably DNA polymerase delta and/or epsilon. Mismatch repair-defective cells withstand O6-meG in their DNA [2], making them tolerant to methylating agents. Methylation-tolerant HeLaMR clones, with a mutator phenotype and a defect in either mismatch recognition or correction in vitro, also performed little O6-meG-dependent DNA synthesis. Assays of pairwise combinations of tolerant and colorectal carcinoma cell extracts identified hMLH1 as the missing mismatch repair function in a group of tolerant clones. The absence of processing by extracts of methylation-tolerant cells provides the first biochemical evidence that lethality of DNA O6-meG derives from its interaction with mismatch repair.

Requirement for PCNA in DNA mismatch repair at a step preceding DNA resynthesis

A two-hybrid system was used to screen yeast and human expression libraries for proteins that interact with mismatch repair proteins. PCNA was recovered from both libraries and shown in the case of yeast to interact with both MLH1 and MSH2. A yeast strain containing a mutation in the PCNA gene had a strongly elevated mutation rate in a dinucleotide repeat, and the rate was not further elevated in a strain also containing a mutation in MLH1. Mismatch repair activity was examined in human cell extracts using an assay that does not require DNA repair synthesis. Activity was inhibited by p21WAF1 or a p21 peptide, both of which bind to PCNA, and activity was restored to inhibited reactions by addition of PCNA. The data suggest a PCNA requirement in mismatch repair at a step preceding DNA resynthesis. The ability of PCNA to bind to MLH1 and MSH2 may reflect linkage between mismatch repair and replication and may be relevant to the roles of mismatch repair proteins in other DNA transactions.

The Saccharomyces cerevisiae Msh2 and Msh6 proteins form a complex that specifically binds to duplex oligonucleotides containing mismatched DNA base pairs

The yeast Saccharomyces cerevisiae encodes six proteins, Msh1p to Msh6p, that show strong amino acid sequence similarity to MutS, a central component of the bacterial mutHLS mismatch repair system. Recent studies with humans and S. cerevisiae suggest that in eukaryotes, specific MutS homolog complexes that display unique DNA mismatch specificities exist. In this study, the S. cerevisiae 109-kDa Msh2 and 140-kDa Msh6 proteins were cooverexpressed in S. cerevisiae and shown to interact in an immunoprecipitation assay and by conventional chromatography. Deletion analysis of MSH2 indicated that the carboxy-terminal 114 amino acids of Msh2p are important for Msh6p interaction. Purified Msh2p-Msh6p selectively bound to duplex oligonucleotide substrates containing a G/T mismatch and a +1 insertion mismatch but did not show specific binding to +2 and +4 insertion mismatches. The mismatch binding specificity of the Msh2p-Msh6p complex, as measured by on-rate and off-rate binding studies, was abolished by ATP. Interestingly, palindromic substrates that are poorly repaired in vivo were specifically recognized by Msh2p-Msh6p; however, the binding of Msh2p-Msh6p to these substrates was not modulated by ATP. Taken together, these studies suggest that the repair of a base pair mismatch by the Msh2p-Msh6p complex is dependent on the ability of the Msh2p-Msh6p-DNA mismatch complex to use ATP hydrolysis to activate downstream events in mismatch repair.

Fibroblast growth factor-1 stimulation of quiescent NIH 3T3 cells increases G/T mismatch-binding protein expression

Polypeptide growth factors promote cell-cycle progression in part by the transcriptional activation of a diverse group of specific genes. We have used an mRNA differential-display approach to identify several fibroblast growth factor (FGF)-1 (acidic FGF)-inducible genes in NIH 3T3 cells. Here we report that one of these genes, called FGF-regulated (FR)-3, is predicted to encode G/T mismatch-binding protein (GTBP), a component of the mammalian DNA mismatch correction system. The murine GTBP gene is transiently expressed after FGF-1 or calf serum treatment, with maximal mRNA levels detected at 12 and 18 h post-stimulation. FGF-1-stimulated NIH 3T3 cells also express an increased amount of GTBP as determined by immunoblot analysis. These results indicate that elevated levels of GTBP may be required during the DNA synthesis phase of the cell cycle for efficient G/T mismatch recognition and repair.

3-aminobenzamide and/or O6-benzylguanine evaluated as an adjuvant to temozolomide or BCNU treatment in cell lines of variable mismatch repair status and O6-alkylguanine-DNA alkyltransferase activity

O6-benzylguanine (O6-BG) and 3-aminobenzamide (3-AB) inhibit the DNA repair proteins O6-alkylguanine-DNA alkyltransferase (AGT) and poly(ADP-ribose) polymerase (PARP) respectively. The effect of O6-BG and/or 3-AB on temozolomide and 1,3-bis(2-chloroethyl)-nitrosourea (BCNU) cytotoxicity, was assessed in seven human tumour cell lines: six with an AGT activity of > 80 fmol mg-1 protein (Mer+) and one with an AGT activity of < 3 fmol mg-1 protein (Mer-). Three of the Mer+ cell lines (LS174T, DLD1 and HCT116) were considered to exhibit resistance to methylation by a mismatch repair deficiency (MMR-), each being known to exhibit microsatellite instability, and DLD1 and HCT116 having well-characterised defects in DNA mismatch binding. Potentiation was defined as the ratio between an IC50 achieved without and with a particular inhibitor treatment. Temozolomide or BCNU cytotoxicity was not potentiated by either inhibitor in the Mer- cell line. Preincubation with O6-BG (100 microM for 1 h) was found to potentiate the cytotoxicity of temozolomide by 1.35- to 1.57-old in Mer+/MMR+ cells, but had no significant effect in Mer+/MMR- cells. In comparison, O6-BG pretreatment enhanced BCNU cytotoxicity by 1.94- to 2.57-fold in all Mer+ cell lines. Post-incubation with 3-AB (2 mM, 48 h) potentiated temozolomide by 1.35- to 1.59-fold in Mer+/MMR+ cells, and when combined with O6-BG pretreatment produced an effect which was at least additive, enhancing cytotoxicity by 1.97- to 2.16-fold. 3-AB treatment also produced marked potentiation (2.20- to 3.12-fold) of temozolomide cytotoxicity in Mer+/MMR- cells. In contrast, 3-AB produced marginal potentiation of BCNU cytotoxicity in only three cell lines (1.19- to 1.35-fold), and did not enhance the cytotoxicity of BCNU with O6-BG treatment in any cell line. These data suggest that the combination of an AGT and PARP inhibitor may have a therapeutic role in potentiating temozolomide activity, but that the inhibition of poly(ADP-ribosyl)ation has little effect on the cytotoxicity of BCNU.

Mutation screening of MSH2 and MLH1 mRNA in hereditary non-polyposis colon cancer syndrome

Germline mutations in four human mismatch repair genes (MSH2, MLH1, PMS1, and PMS2) have been reported to cause hereditary non-polyposis colon cancer syndrome (HNPCC). The identification of germline mutations in HNPCC kindreds allows precise diagnosis and accurate predictive testing. To investigate further the genetic epidemiology of HNPCC and the nature and frequency of germline mutations in this disorder, we studied 17 English HNPCC kindreds for germline mutations in MSH2 and MLH1. A previous genetic linkage study had suggested that most English HNPCC families will have mutations in one of these genes. Mutation analysis was performed in a three step process. (1) mRNA extracted from lymphoblastoid cell lines was analysed for gross rearrangements, (2) the in vitro transcription-translation (IVTT) assay was then performed to detect protein truncating mutations, and (3) partial cDNA sequencing of MSH2 or MLH1 was undertaken in families (n = 6) linked to MSH2 or MLH1 but without a detectable mutation. Seven different germline mutations were identified in eight of 17 (47%) kindreds (five in MSH2 and three in MLH1). In three cases there was a deletion of a single exon in MSH2 mRNA, three mutations resulted in a truncated protein product, and two missense mutations were identified by direct sequencing. Six mutations were novel. No precise correlation between genotype and phenotype was observed, although a MSH2 missense (Thr905Arg) mutation was associated with a susceptibility to multiple colorectal polyps. Age related risks for colorectal and uterine cancer were similar for MSH2 and MLH1 mutations.

Binding of insertion/deletion DNA mismatches by the heterodimer of yeast mismatch repair proteins MSH2 and MSH3

DNA-mismatch repair removes mismatches from the newly replicated DNA strand. In humans, mutations in the mismatch repair genes hMSH2, hMLH1, hPMS1 and hPMS2 result in hereditary non-polyposis colorectal cancer (HNPCC) [1-8]. The hMSH2 (MSH for MutS homologue) protein forms a complex with a 160 kDa protein, and this heterodimer, hMutSalpha, has high affinity for a G/T mismatch [9,10]. Cell lines in which the 160 kDa subunit of hMutSalpha is mutated are specifically defective in the repair of base-base and single-nucleotide insertion/deletion mismatches [9,11]. Genetic studies in S. cerevisiae have suggested that MSH2 functions with either MSH3 or MSH6 in mismatch repair, and, in the absence of the latter two genes, MSH2 is inactive [12,13]. MSH6 encodes the yeast counterpart of the 160 kDa subunit of hMutSalpha [12,13]. As in humans, yeast MSH6 forms a complex with MSH2, and the MSH2-MSH6 heterodimer binds a G/T mismatch [14]. Here, we find that MSH2 and MSH3 form another stable heterodimer, and we purify this heterodimer to near homogeneity. We show that MSH2-MSH3 has low affinity for a G/T mismatch but binds to insertion/deletion mismatches with high specificity, unlike MSH2-MSH6.

Mismatch repair defects in human carcinogenesis

Mismatch repair defects are carcinogenic. This conclusion comes some 80 years after the original description of a type of familial colorectal cancer in which mismatch repair defects are involved, and from decades of dedicated basic science research into fundamental mechanisms cells use to repair their DNA. Mismatch repair (MMR) was described first in bacteria, later in yeast and finally in higher eukaryotes. In bacteria, one of its roles is the rapid repair of replicative errors thereby providing the genome with a 100-1000-fold level of protection against mutation. It also guards the genome by preventing recombination between non-homologous regions of DNA. The information gained from bacteria suddenly became relevant to human neoplasia in 1993 when the RER phenotype of microsatellite instability was discovered in human cancers and was rapidly shown to be due to defects in mismatch repair. Evidence supporting the role of MMR defects in carcinogenesis comes from a variety of independent sources including: (i) theoretical considerations of the requirement for a mutator phenotype as a step in multistage carcinogenesis; (ii) discovering that MMR defects cause a 'mutator phenotype' destabilizing endogenous expressed genes including those integral to carcinogenesis; (iii) finding MMR defects in the germline of HNPCC kindred members; (iv) finding that such defects behave as classic tumor suppressor genes in both familial and sporadic colorectal cancers; (v) discovering that MMR 'knockout' mice have an increased incidence of tumors; and (vi) discovering that genetic complementation of MMR defective cells stabilizes the MMR deficiency-associated microsatellite instability. Models of carcinogenesis now must integrate the concepts of a MMR defect induced mutator phenotype (Loeb) with the concepts of multistep colon carcinogenesis (Fearon and Vogelstein) and clonal heterogeneity/selection (Nowell).

hMutSbeta, a heterodimer of hMSH2 and hMSH3, binds to insertion/deletion loops in DNA

In human cells, mismatch recognition is mediated by a heterodimeric complex, hMutSalpha, comprised of two members of the MutS homolog (MSH) family of proteins, hMSH2 and GTBP [1,2]. Correspondingly, tumour-derived cell lines defective in hMSH2 and GTBP have a mutator phenotype [3,4], and extracts prepared from these cells lack mismatch-binding activity [1]. However, although hMSH2 mutant cell lines showed considerable microsatellite instability in tracts of mononucleotide and dinucleotide repeats [4,5], only mononucleotide repeats were somewhat unstable in GTBP mutants [4,6]. These findings, together with data showing that extracts of cells lacking GTBP are partially proficient in the repair of two-nucleotide loops [2], suggested that loop repair can be GTBP-independent. We show here that hMSH2 can also heterodimerize with a third human MSH family member, hMSH3, and that this complex, hMutSbeta, binds loops of one to four extrahelical bases. Our data further suggest that hMSH3 and GTBP are redundant in loop repair, and help explain why only mutations in hMSH2, and not in GTBP or hMSH3, segregate with hereditary non-polyposis colorectal cancer (HNPCC) [7].

Mutation of MSH3 in endometrial cancer and evidence for its functional role in heteroduplex repair

Many human tumours have length alterations in repetitive sequence elements. Although this microsatellite instability has been attributed to mutations in four DNA mismatch repair genes in hereditary nonpolyposis colorectal cancer (HNPCC) kindreds, many sporadic tumours exhibit instability but no detectable mutations in these genes. It is therefore of interest to identify other genes that contribute to this instability. In yeast, mutations in several genes, including RTH and MSH3, cause microsatellite instability. Thus, we screened 16 endometrial carcinomas with microsatellite instability for alterations in FEN1 (the human homolog of RTH) and in MSH3 (refs 12-14). Although we found no FEN1 mutations, a frameshift mutation in MSH3 was observed in an endometrial carcinoma and in an endometrial carcinoma cell line. Extracts of the cell line were deficient in repair of DNA substrates containing mismatches or extra nucleotides. Introducing chromosome 5, encoding the MSH3 gene, into the mutant cell line increased the stability of some but not all microsatellites. Extracts of these cells repaired certain substrates containing extra nucleotides, but were deficient in repair of those containing mismatches or other extra nucleotides. A subsequent search revealed a second gene mutation in HHUA cells, a missense mutation in the MSH6 gene. Together the data suggest that the MSH3 gene encodes a product that functions in repair of some but not all pre-mutational intermediates, its mutation in tumours can result in genomic instability and, as in yeast, MSH3 and MSH6 are partially redundant for mismatch repair.

APC mutations in colorectal tumors with mismatch repair deficiency

We have investigated the influence of genetic instability [replication error (RER) phenotype] on APC (adenomatous polyposis coli), a gene thought to initiate colorectal tumorigenesis. The prevalence of APC mutations was similar in RER and non-RER tumors, indicating that both tumor types share this step in neoplastic transformation. However, in a total of 101 sequenced mutations, we noted a substantial excess of APC frameshift mutations in the RER cases (70% in RER tumors versus 47% in non-RER tumors, P < 0.04). These frameshifts were characteristic of mutations arising in cells deficient in DNA mismatch repair, with a predilection for mononucleotide repeats in the RER tumors (P < 0.0002), particularly (A)n tracts (P < 0.00007). These findings suggest that the genetic instability that is reflected by the RER phenotype precedes, and is responsible for, APC mutation in RER large bowel tumors and have important implications for understanding the very earliest stages of neoplasia in patients with tumors deficient in mismatch repair.

Cisplatin and adriamycin resistance are associated with MutLalpha and mismatch repair deficiency in an ovarian tumor cell line

In contrast to parental A2780 ovarian tumor cells, extracts of one doxorubicin-resistant and two independent cis-diamminedichloroplatinum(II)-resistant derivatives are defective in strand-specific mismatch repair. The repair defect of the three hypermutable, drug-resistant cell lines is only evident when the strand break that directs the reaction is located 3' to the mismatch, and in each case repair is restored to extracts by addition of purified MutLalpha heterodimer. As judged by immunological assay, drug resistance is associated with the virtual absence of the MutLalpha MLH1 subunit and greatly reduced levels of the PMS2 subunit. These findings implicate a functional mismatch repair system in the cytotoxic effects of these antitumor drugs and may have ramifications for their clinical application.

The mismatch-repair protein hMSH2 binds selectively to DNA adducts of the anticancer drug cisplatin

BACKGROUND

The antitumor drug cis-diamminedichloroplatinum(II) (cis-DDP or cisplatin) exerts its cytotoxic effects through the formation of covalent DNA adducts. A family of proteins possessing a common HMG box motif that binds specifically to cisplatin DNA adducts has been previously suggested to be important in the clinical efficacy of the drug.

RESULTS

We have shown that the human mismatch-repair protein, hMSH2, also binds specifically to DNA containing cisplatin adducts and displays selectivity for the DNA adducts of therapeutically active platinum complexes. Moreover, hMSH2 is overexpressed in testicular and ovarian tissue; tumors in these tissues are most effectively treated by cisplatin.

CONCLUSIONS

Our results suggest a role for hMSH2 in mediating cisplatin toxicity. Supporting this view, previous studies in Escherichia coli dam- strains demonstrate that mutations in mismatch-repair proteins confer resistance to cisplatin toxicity. Mismatch-repair deficiency is also correlated with tolerance to O6-methylguanine, a cytotoxic DNA lesion formed by methylating agents. A current model ascribes O6-methylguanine toxicity to unsuccessful attempts at repair of this lesion by mismatch-repair proteins, resulting in a futile cycle of incision and synthesis, leading ultimately to lethal DNA-strand breaks. We propose that mismatch repair may contribute to cisplatin toxicity by a similar mechanism. Alternatively, hMSH2 may shield cisplatin adducts from repair, allowing adducts to persist, thus enhancing lethality.

Competency in mismatch repair prohibits clonal expansion of cancer cells treated with N-methyl-N'-nitro-N-nitrosoguanidine

The phenomenon of alkylation tolerance has been observed in cells that are deficient in some component of the DNA mismatch repair (MMR) system. An alkylation-induced cell cycle arrest had been reported previously in one MMR-proficient cell line, whereas a MMR-defective clone derived from this line escapes from this arrest. We examined human cancer cell lines to determine if the cell cycle arrest were dependent upon the MMR system. Growth characteristics and cell cycle analysis after MNNG treatment were ascertained in seven MMR-deficient and proficient cell lines, with and without confirmed mutations in hMLH1 or hMSH2 by an in vitro transcription/translation assay. MMR-proficient cells underwent growth arrest in the G2 phase of the cell cycle after the first S phase, whereas MMR-deficient cells escaped an initial G2 delay and resumed a normal growth pattern. In the HCT116 line corrected for defective MMR by chromosome 3 transfer, the G2 phase arrest lasted more than five days. In another MMR-proficient colon cancer cell line, SW480, cell death occurred five days after MNNG treatment. A competent MMR system appears to be necessary for G2 arrest or cell death after alkylation damage, and this cell cycle checkpoint may allow the cell to repair damaged DNA, or prevent the replication of mutated DNA by prohibiting clonal expansion.