Extensive characterization of genetic alterations in a series of human colorectal cancer cell lines

Transcription

1 (2001) 20, 5025 ± 5032 ã 2001 Nature Publishing Group All rights reserved 0950 ± 9232/01 $ Extensive characterization of genetic alterations in a series of human colorectal cancer cell lines Jacqueline Gayet 1, Xiao-Ping Zhou 1, Alex Duval 1, Sandra Rolland 1, Jean-Marc Hoang 1, Paul Cottu 1 and Richard Hamelin*,1 1 INSERM U434 - CEPH, Paris, France A number of genetic alterations have been described in colorectal cancers. They include allelic losses on speci c chromosomal arms, mutations of oncogenes, tumor suppressor genes and mismatch repair genes, microsatellite instability in coding repeat sequences of target genes and methylation defects in gene promoters. Since these alterations have been reported by di erent groups on di erent tumors and cell lines, the complete repertoire of genetic alterations for any given tumor sample remains unknown. In the present study, we analysed a series of 22 colorectal cancer cell lines for 31 di erent genetic alterations. We found signi cant correlations between mutational pro les in these colorectal cell lines associated with di erences in mismatch repair status. This panel of colon cancer cell lines is representative of the genetic heterogeneity occurring in sporadic colorectal carcinoma. Our results may prove to be very useful for understanding the di erent biological pathways involved in the development of colon cancer, and for groups studying cellular biology and pharmacology on the same cell lines. (2001) 20, 5025 ± Keywords: colorectal cancer; cell lines; genetic alterations Introduction Carcinogenesis is known to be a multistage process involving a number of genes giving growth advantage to the cells when altered. Among the di erent neoplasms, colorectal cancer is one of the most frequent in human and is also the best characterized for genetic alterations associated with tumor progression. It was the rst tumor type for which a model of carcinogenesis was proposed (Fearon and Vogelstein, 1990). Colon cancer develops as a result of the transformation of normal colonic epithelium to an adenomatous polyp and ultimately to an invasive cancer. The number of genetic defects described as playing a potential role during the development and progression of colorectal cancer has *Correspondence: R Hamelin, INSERM U434-CEPH 27 rue Juliette Dodu Paris, France; richard.hamelin@cephb.fr Received 24 January 2001; revised 19 March 2001; accepted 9 May 2001 been increasing steadily in recent years (Ilyas et al., 1999; Chung, 2000). These genetic alterations include inactivation of tumor suppressor genes by mutation and/or loss of heterozygosity (LOH), activation of oncogenes, mismatch repair defects with, as a consequence, frameshift mutations in target genes containing nucleotide repeats, and epigenic silencing by methylation of gene promoters. It appears however that the repertoire of genetic alterations in colorectal cancer is still incomplete since, for example, frequent LOH has been described at particular chromosome arms where the putative tumor suppressor gene(s) remains undetermined (Lerebours et al., 1999). The percentage of colorectal tumors showing a particular genetic alteration varies depending on the alteration which is assessed, because all tumors do not present the same genetic alterations. This is consistent with the view that colon tumors represent a group of heterogeneous neoplasms. Given the genetic diversity known to be present in primary colorectal tumors, it is di cult to compare the mutational pro les of genetic alterations published by di erent groups. In addition, no-one has performed an extensive analysis of genetic alterations in the same series of primary tumors. There are many established cancer cell lines which are widely used as in vitro model systems to study biological pathways in the development and treatment of cancer. Cell lines present the same repertoire of genetic alterations as primary tumors, but here again, each cell line has been characterized for only a limited number of genetic alterations. In the present study, we analysed a series of 22 colorectal cancer cell lines for 31 di erent genetic alterations known to be involved in colorectal cancers. This study represents the most complete genetic analysis reported to date. Since these cell lines represent the genetic diversity encountered among primary colorectal tumors, our analysis should prove useful for relating biological and pharmacological changes with particular genetic alterations in these cell lines. Results We have presented in four consecutive tables, using bold characters, the genetic alterations observed in a series of 22 colorectal cancer cell lines. Some of these alterations were already reported in the literature, and

2 5026 the references (including published results from our laboratory) are indicated in the tables. These references generally give results in agreement with our own results except when the reference is underlined. Microsatellite analysis of instability and LOH The characterization of the Microsatellite Instability (MSI-H) status of the cell lines has already been described using Bat-26 and Bat-25 mononucleotide repeats (Cottu et al., 1996; Hoang et al., 1997). The HCT15 cell line, which is known to be stable on dinucleotide microsatellites and unstable on mononucleotide repeats, was considered as showing MSI in the present study. Taken together, we analysed 8 MSI-H and 14 microsatellite-stable (MSS) cell lines (Table 1). LOH analysis was performed on chromosomes 2, 3, 5, 17 and 18 which contain the hmsh2, hmlh1, APC, p53 and DCC/SMAD2/SMAD4 genes respectively, all involved in colorectal carcinogenesis. In the absence of matching normal DNA for each cell line, ve dinucleotide repeats with an average heterozygosity index of 0.7 were analysed for each chromosome. We have indicated in Table 1 the number of homozygous loci for each chromosome in each cell line. A cell line was considered as showing LOH on a chromosome when all the ampli ed microsatellites of this chromosome were homozygous. With ve microsatellites, such an event has a probability of (1 ± 0.7) 5 (= )of being observed without LOH. Due to the presence of frequent microsatellite instability at dinucleotide repeats, allelotyping cannot be interpreted unambiguously in 7/8 MSI-H cell lines. It appeared, however, that none of these cell lines showed clear evidence of LOH on any chromosome. The last MSI cell line, HCT15, also did not present LOH on any chromosome. In contrast, all 14 MSS cell lines showed LOH on at least one chromosome, and therefore could be classi ed as LOH+. Chromosomes 17, 18, 5, 3 and 2 were lost in 13, seven, seven, three and one cases out of the 14 LOH/MSS cell lines respectively. Allelic imbalance was observed for chromosome 3 in the Isreco-2, Isreco-3 and LS1034 cell lines for all markers. This is probably due to a duplication of this chromosome. It has to be noted that in Isreco2 and Isreco3 cells which have been established from hepatic and peritoneal metastasis, respectively from the same patient whose primary tumor was established as Isreco1, di erent parental chromosome 3 were duplicated. Moreover Isreco3 presented a LOH on chromosome 18 absent in Isreco1 and 2. p53 Most of these cell lines have been previously screened for p53 mutations in exons 5 ± 8 (Cottu et al., 1996) and results are indicated in Table 2. All MSS/LOH cell lines except EB and LS513 have a p53 mutation. Dinucleotide microsatellites used for LOH analysis on chromosome 17 were chosen in the vicinity of the p53 gene and in every case, MSS cell lines with a p53 mutation showed LOH on chromosome 17 indicating that the p53 wild-type allele was lost. In MSI-H cell lines, only TC71 and HCT15 presented a mutation on codons 176 and 241, respectively. As indicated by microsatellite analysis and con rmed by the DGGE pro les, the wild-type p53 allele was not lost in these two cases. Table 1 Microsatellite analysis of colorectal cancer cell lines Cell line Sex/age MSI Ch.2 Ch.3 Ch.5 Ch.17 Ch.18 Co115 F77 + a 2/5 d 2/5 d 4/5 d 1/5 d 2/5 d LS174T F58 + a 1/5 d 1/5 d 1/5 d 1/5 d 0/5 d Lovo M56 + a 0/5 d 0/5 d 1/5 d 2/5 d 1/5 d HCT116 M? + a 2/5 d 2/5 d 0/5 d 1/5 d 1/5 d SW48 F82 + b 0/5 d 0/5 d 3/5 d 2/5 d 2/5 d TC71? + b 2/5 d 1/5 d 1/5 d 1/5 d 1/5 d TC7 F? + b 2/5 d 2/5 d 0/4 d 1/5 d 1/5 d HCT15 M? + a,c 3/5 2/5 4/5 3/5 2/5 Isreco1? 7 a 1/5 0/5 0/5 5/5 0/5 Isreco2? 7 a 0/5 0/5 e 0/5 5/5 0/5 Isreco3? 7 a 1/5 0/5 e 0/4 5/5 5/5 FET? 7 a 5/5 5/5 5/5 5/5 3/5 LS1034 M54 7 a 2/5 0/5 e 5/5 5/5 4/5 V9P? 7 a 2/5 2/5 5/5 5/5 5/5 Colo320 F55 7 a 1/5 2/4 3/3 5/5 4/5 FRI? 7 a 4/5 2/4 1/5 5/5 5/5 ALA? 7 a 3/5 0/5 5/5 5/5 4/4 SW1116 M73 7 a 1/5 5/5 0/4 5/5 5/5 SW480 M51 7 a 1/5 5/5 4/4 5/5 5/5 HT29 F44 7 a 1/5 2/5 1/5 5/5 5/5 EB? 7 a 1/5 2/5 2/5 4/4 1/5 LS513 M63 7 a 1/5 1/4 5/5 3/5 0/5 a Cottu et al., 1996; b Hoang et al., 1997; c cell line unstable on mononucleotide, but not on dinucleotide microsatellites; d di cult to interpret due to microsatellite instability; e allelic imbalance APC Since most described APC mutations are truncating ones and localized in the rst half of the protein, the present series of cell lines was screened for APC mutations by an in vitro transcription and translation (IVTT) assay using a RT ± PCR product encompassing exons 1 ± 14 and two overlapping PCR products in the rst half of exon 15. Truncated APC translation products were detected in all MSS/LOH cases except the LS513 cell line (Table 2). In some cases, indicated by an asterisk, no full-length APC product was detected besides the truncated product, indicating a biallelic inactivation as con rmed by LOH analysis with microsatellite markers chosen in the vicinity of the APC gene. In cases without LOH on chromosome 5, and showing APC mutations, two di erent translation products were truncated with the exception of the FRI tumor cell line. In the 14 MSS cell lines, the APC gene was thus biallelically inactivated by double mutation in six cases, and by a single mutation and a LOH in six cases. Among the eight MSI-H cell lines, no APC mutation was observed in Co115, LS174T, HCT116 and SW48. A double APC inactivation was observed for Lovo, TC71 and HCT15, while TC7 presented an APC mutation on a single allele.

3 Table 2 Mutations of oncogenes and tumor suppressor genes in colorectal cancer cell lines P53 APC b-catenin Ki-ras P16 Cell line Mutation LOH 17 IVTT b,c LOH 5 Mutation c Deletion Mutation Promoter 1 Mutation 5027 Co115 7 a M 7 LS174T 7 a 7 7 d,e 7 +*,e 7 G12D j,k U 7 o Lovo 7 a 7 55 e 7 7 g 7 G13D j,k M P48L HCT116 7 a 7 7 d,e 7 + e,g,h 7 G13D UM m 23insG o SW d,e 7 + e,g,h 7 7 M n 7 TC71 C176Y G12D M 7 TC G12D U 7 HCT15 R241F a 7 5*,d,e 7 7 e,g,h,i 7 G13D j,k M 7 o Isreco1 Y163H a G12D M 7 Isreco2 Y163H a G12D UM 7 Isreco3 Y163H a G12D UM 7 FET C176F a + 5* G12D M 7 LS1034 G245D a + 5*,f UM 7 V9P G245D a + 5* U 7 Colo320 R248W a + 5* G12D M n 7 FRI C277F a G13D U 7 ALA 301delC a + 5* G12D M 7 SW1116 A159D a G12D j M n 7 SW480 R273H a + 5*,d,e,f + 7 e,h 7 G12V j M n 7 HT29 R273H a + 55 d,e 7 7 e,i 7 7 M n 7 EB 7 a G12D M 7 LS513 7 a G12D M 7 a Cottu et al., b For IVTT, 5 indicates a single shortened translation product and 55 indicates two shortened products with di erent primer pairs; c, *indicates the absence of wild-type product or wild-type sequence; d Heinen et al., 1995; e Ilyas et al., References are underlined when discordant with our own results; f Cottrell et al., 1992; g Sparks et al., 1998; h Morin et al., 1997; i Kitaeva et al., 1997; j Peinado et al., 1993; k Casares et al., 1995; l M: Methylated, U: unmethylated; m MyoÈ haè nen et al., 1998; n Herman et al., 1995; o Okamoto et al., 1994 b-catenin It has been reported that the b-catenin gene can be mutated in the GSK-b phosphorylation site localized in exon 3. We thus screened this exon and adjacent splice junctions for mutations by DGGE. Abnormal DGGE pro les were detected in the LS174T, HCT116 and SW48 MSI-H cell lines and in none of the 14 MSS cell lines (Table 2). To extend our analysis, we looked for potential interstitial deletions involving exon 3. Using primers complementary to sequences in exons 2 and 4, there was no evidence of a shorter ampli cation product of exon 3 in any of the 22 cell lines. Ki-ras Mutations in exons 1 and 2 of the Ki-ras gene were screened by DGGE. No altered pro le was detected within exon 2 in any cell line indicating the absence of mutations on codon 61. Six MSI-H and 11 MSS cell lines showed DGGE alterations in exon 1 which was then sequenced. All cell lines were mutated in codon 12 (13 cases) or 13 (four cases) (Table 2). In 16 out of the 17 mutated cases, glycine at position 12 or 13 was changed to aspartic acid. p16 The CDKN2/p16/MTS1 inhibitor of the cyclin D/ CDK4 complex has been shown to be frequently inactivated in colorectal cancer by promoter methylation. We performed methylation-speci c PCR (MSP) for all cell lines. In the majority of cases, p16 promoter was methylated on both alleles if present (Table 2). The LS174T, TC7, V9P and FRI cell lines showed unmethylated promotors on both alleles (if present). Finally, in HCT116, LS1034, Isreco2 and Isreco3, one allele was methylated, and the other was not. Exons 1 and 2 of the p16 gene were also directly sequenced in all cell lines. No mutation was detected except a C ± T transition at the second nucleotide of codon 48 (pro 4 leu) within the second ankyrin domain of p16 in LoVo, and a G insertion in codon 23 in HCT116 leading to a frameshift. In addition, a known G to A polymorphism at codon 148 was detected once in the EB cell line. It has to be noted that in any of the 22 cell lines, we observed an homozygous p16 deletion as it has been reported for primary tumors of other organs (Kamb et al., 1994). Mismatch repair genes Truncating mutations were screened for in the hmsh2, hmlh1, hpms1 and hpms2 mismatch repair genes by IVTT (Table 3). The hmsh2 gene was ampli ed from cdna into two fragments encompassing exons 1 ± 11 and 9 ± 16, respectively. Full-length polypeptides were translated in all cell lines except Lovo and TC71 showing shortened truncated polypeptides in the rst fragment. No truncated polypeptide was detected with any of the 22 cell lines when two overlapping cdna fragments of hmlh1 were transcribed and translated. In spite of numerous attempts, we were unable to amplify hmlh1 from cdna of the HCT116, Co115, TC7 and SW48 cell lines. We then analysed the methylation status of

4 5028 Table 3 Mismatch repair genes defects in colorectal cancer cell lines hmsh2 hmlh1 hpms1 hpms2 Cell line IVTT a,b IVTT f Promoter i IVTT IVTT Co115 7 NA M 7 7 LS174T 7 c 7 c,g U c,g 7 7 Lovo 5*,c,d 7 c,d U c,g,j nd l,d 7 d HCT116 7 c NA c,h U c,g,k 7 7 SW48 7 c,d NA c,d M c,g,j 7 d 7 d TC U 7 7 TC7 7 NA U 7 7 HCT15 7 c,e 7 c,e U c,j 7 e 7 e Isreco1 7 7 U 7 7 Isreco2 7 7 U 7 7 Isreco3 7 7 U 7 7 FET 7 7 U 7 7 LS c 7 c U c 7 7 V9P 7 7 U 7 7 Colo320 7 c 7 c U c,g 7 7 FRI 7 7 U 7 7 ALA 7 7 U 7 7 SW c 7 c U c,g 7 7 SW U j,k 7 7 HT29 7 c 7 c U c,g,j 7 7 EB 7 7 U 7 7 LS U 7 7 a For IVTT, 5 indicates a single shortened translation product. b, *Indicates the absence of wild-type product; c Wheeler et al., 1999; d Liu et al., 1995; e da Costa et al., 1995; f NA: not ampli able; g Deng et al., 1999; h Papadopoulos et al., 1994; i M: Methylated, U: unmethylated; j Herman et al., 1998; k Veigl et al., 1998; l nd: not done the hmlh1 promoter in all cell lines, and found it to be methylated in Co115 and SW48. hpms1 was translated in three overlapping polypeptides with no alteration in any cell line. For hpms2, we screened only the second half of the gene, and no changes were detected in any cell line. Nucleotide repeats containing genes Mononucleotide runs in coding sequences are known to be frequently altered in MSI-H tumors, and stable in MSS tumors. Twenty- ve genes containing coding mononucleotide repeats were selected from human genome databases and analysed for size alterations in our series of 22 colorectal cancer cell lines. Based on the analysis of the same 25 genes in a large series of primary colorectal tumors, we de ned a cut-o value of 12% to distinguish mutations within target genes for instability and those mutations in genes considered as bystander events without relationship to tumoral progression (Duval et al., 2001). A total of 10 genes were de ned as target genes including the already described TGF-b receptor type II (10A), caspase-5 (10A), TCF-4 (9A), IGFIIR (8G), BAX (8G), hmsh3 (8A), and hmsh6 (8C) genes, and three new candidates named RAD50 (9A, accession no. U63139), RHAMM (9A, accession no U29343), and GRB14 (9A, accession no L76687). With the exception of TGFbRII in the LS1034 cell line showing a 1 bp insertion, none of the 14 MSS cell lines showed mutation within any of these 10 genes. In contrast, all 8 MSI-H cell lines showed mutations in these coding repeats (Table 4). The mutations were 1 ± 2 bp deletions or insertions leading to frameshifts. We have indicated by an asterisk genes for which the normal size allele was not detectable in the di erent cell lines. Mutational Indexes at Real Target genes (MIRT) for each MSI-H cell line are shown in Table 4 and vary from 0.9 for the Co115 cell line to 0.2 for HCT15. The 15 other coding mononucleotide repeats, localized in genes with no role in tumoral progression remained unaltered in all MSS cell lines. They were however more or less mutated in MSI-H cell lines de ning a Mutational Index for ByStander events (MIBS) ranging from 0.5 for the Lovo cell line to 0 for HCT15. The E2F-4 gene contains a trinucleotide coding repeat which has also been examined in this study. All MSI-H cell lines, except HCT15, as well as the SW480 MSS cell line showed varying numbers of this repeat as compared to other cell lines. However these alterations did not change the open reading frame. Discussion In this study we report the analysis of a series of 22 colorectal cancer cell lines for MSI status, LOH on ve chromosomes (2, 3, 5, 17 and 18), mutations in nine genes (p53, APC, b-catenin, Ki-ras, p16, hmsh2, hmlh1, hpms1 and hpms2), interstitial deletions in one gene (b-catenin), methylation defects in the promoters of two genes (p16 and hmlh1), mutational indexes at both real target (MIRT) and bystander (MIBS) genes for instability, mutations in a trinucleotide coding repeat (E2F-4) and mutations in 10 mononucleotide coding repeats in target genes for instability. That makes a total of 31 genetic alterations for a total of 682 informations (which can be increased to 46 genetic alterations for 1012 informations if we also consider the analysis of the 15 coding repeats within bystander genes). Of these, only 185 informations (27 or 18%) were reported in 25 di erent references, as indicated in Tables 1 ± 4. The MSI status was established using mono and dinucleotide repeat microsatellites. Eight cell lines were MSI-H and the remaining 14 cell lines were MSS. This does not represent the real percentage of MSI-H tumors in colorectal cancer since we intentionally enriched the proportion of this phenotype. All MSS cell lines showed LOH on at least one out of the ve analysed chromosomes. Although microsatellite instability makes interpretation di cult, the eight MSI- H cell lines did not demonstrate LOH. This allowed us to exclusively classify cell lines either as MSI-H or MSS/LOH. Chromosomes 2, 3, 5, 18 and 17 were lost in 7, 21, 50, 50 and 93% of MSS/LOH cell lines, respectively. The high frequency of LOH on chromosomes 17, 18 and 5 indicates that tumor suppressor genes localized on these chromosomes should be frequently altered. Indeed, we nd p53 and APC mutations in 12 and 13

5 Table 4 Mutations in coding repeats in colorectal cancer cell lines TGFbRII a IGFIIR BAX a TCF-4 a caspase 5 a MSH3 a MSH6 GRB14 RHAMM RAD50 MIRT MIBS E2F-4 a Cell line 10A 8G 8G 9A 10A 8A 8C 9A 9A 9A values values tri-n Co115 +*,b + b + 7 e /10 1/15 +* LS174T +*,b 7 b +* + e +*,g + h + h /10 2/14 +* Lovo +*,b 7 b +*,c +*,e + g 7 7 h /10 8/15 +* HCT116 +*,b 7 b + c 7 e 7 g +* + h /10 5/15 + SW48 +* + b 7 c 7 e,f + g 7 + h /10 1/14 +* TC71 + b 7 b 7 +*,e /10 5/15 +* TC7 +*,b 7 b 7 + e h 7 na d 7 5/9 2/15 + HCT15 + b 7 b 7 c 7 e + g 7 7 h,i /10 0/15 7 Isreco1 7 b 7 b 7 7 e /10 0/15 7 Isreco2 7 b 7 b 7 7 e 7 7 na d /9 0/14 7 Isreco3 7 b 7 b 7 7 e /10 0/14 7 FET 7 b 7 b 7 7 e,f /10 0/15 7 LS b 7 b 7 7 e,f /10 0/15 7 V9P 7 b 7 b 7 7 e /10 0/14 7 Colo320 7 b 7 b 7 7 e,f /10 0/15 7 FRI 7 b 7 b 7 7 e /10 0/15 7 ALA 7 b 7 b 7 7 e /10 0/15 7 SW b 7 b 7 7 e /10 0/14 7 SW480 7 b 7 b 7 7 e /10 0/15 +* HT29 7 b 7 b 7 7 e /10 0/13 7 EB 7 b 7 b 7 7 e /10 0/15 7 LS513 7 b 7 b 7 7 e /10 0/15 7 a, *Indicates the absence of wild-type sequence. b Zhou et al., 1997; c Rampino et al., 1997; d na: not ampli ed; e Duval et al., 1999; f These cell lines have TCF-4 mutations in other sequences than the (A)9 repeat (Duval et al., 2000); g Schwartz et al., 1999; h Malkhosyan et al., 1996; i HCT-15 cell line is mutated on codons 222 and 1103 of the MSH6 (GTBP) gene (Papadopoulos et al., 1995) cases of 14 MSS/LOH cell lines, respectively. In all cases but one, biallelic inactivation by mutation and LOH, or by double mutation was observed. The only exception was the FRI cell line for which we detected only one hit for the APC gene. In MSI-H cell lines, p53 mutations were observed only in two out of eight cell lines (TC71 and HCT15), but only one hit was observed in each case. Four MSI-H cell lines were mutated in APC with a two-hit mechanism in three cases. MSI-H cell lines were signi cantly less mutated in p53 (P=0.017) than MSS/LOH cell lines and showed a tendency for the APC gene (P=0,07). Di erences between MSI-H and MSS/LOH cell lines were more signi cant when the number of hits for p53 and APC were compared (P= and respectively). Although interstitial deletion of exon 3 of the b-catenin gene has been reported in a subset of primary colorectal carcinomas (Iwao et al., 1998), we found none in the 22 cell lines used in this study. DNA variants of the b-catenin gene were however observed by DGGE in three out of eight MSI-H cell lines and in none of the 14 MSS/LOH cell lines. These three cases had a wild-type APC gene. There was no di erence in frequency of mutations for the Ki-ras gene between MSI-H (6/8) and MSS/ LOH (11/14) cell lines. P16 expression was also inhibited by methylation of its promoter as frequently in MSI-H as in MSS cell lines. Two mutations potentially inactivating the p16 gene were detected, both in MSI-H cell lines. It is known that the MSI-H phenotype is due to defects in mismatch repair (MMR) genes (Leach et al., 1993; Fishel et al., 1993; Bronner et al., 1994; Papadopoulos et al., 1994; Nicolaides et al., 1994). We thus looked for inactivation of the hmsh2, hmlh1, hpms1 and hpms2 genes in all cell lines. As expected, no alteration was found in any of the MSS cell lines. hmsh2 was truncated in Lovo and TC71 cell lines, and the hmlh1 promoter was methylated in Co115 and SW48 preventing hmlh1 expression. We were unable to amplify by RT ± PCR the hmlh1 gene in HCT116 and TC7 cell lines. Both cell lines are known to have nonsense mutations in hmlh1 (Papadopoulos et al., 1994, and M Muleris, personal communication) and the degradation of the corresponding mrna could be accelerated by the nonsense-mediated mrna decay mechanism implicated in mrna turnover (Belgrader et al., 1994). The LS174T cell line contains a missense mutation within the hmlh1 gene (Deng et al., 1999) which cannot be detected by IVTT. Since it is known that HCT15 is mutated on hmsh6 (Papadopoulos et al., 1995), all eight MSI-H cell lines presented in this study have defects in one of the known MMR genes. The direct consequence of MMR defect is the high frequency of deletions/insertions in repeated sequences such as microsatellites. Some genes contain mononucleotide repeats within their coding sequences and are thus frequent targets for mutations in MSI-H tumors (Parsons et al., 1995; Souza et al., 1996; Rampino et al., 1997; Malkhosyan et al., 1996; Schwartz et al., 1999; Duval et al., 1999). We screened our series of colorectal cancer cell lines for such alterations in 10 genes supposed to be involved in MSI-H tumor progression. Only one mutation was found for 139 events in MSS cell lines compared with 44 mutations for 79 events in MSI-H cell lines (P=0.0001).

6 5030 TGFbRII was mutated in all eight MSI-H cell lines and RHAMM in one MSI-H cell line out of the seven cases analysed for this gene. The other target genes for instability had an intermediate mutation rate. Biallelic inactivation was observed in some cases for the TGFbRII, BAX, TCF-4, caspase-5 and MSH3 genes. We de ned a MIRT showing that MSI-H cell lines were generally mutated in about 50% of the target genes, with a maximum of nine out of 10 for the Co115 cell line. MIBS values were de ned by mutations in a series of 15 genes containing coding repeats and rarely mutated in primary colorectal tumors (Duval et al., 2001). It has to be noted that, as observed for other common genetic alterations, MIRT and MIBS values are higher for cell lines than for primary tumors. In an international meeting on MSI-H tumors held in Bethesda (Boland et al., 1998), consensus guidelines have been proposed to gauge whether an a ected gene is a true target of inactivation involved in tumorigenesis. Among these guidelines is a high frequency of inactivation, and biallelic inactivation. Our results show that cell lines should not be used to demonstrate that a gene is a real target gene for instability in MSI- H tumors since some genes never mutated in MSI-H primary tumors are mutated in as much as 50% and/or biallelically inactivated in our series of eight MSI-H cell lines. It is known that there are at least two alternative pathways in primary colorectal tumors de ned as MSI- H and LOH and showing microsatellite instability or frequent allelic losses, respectively (Aaltonen et al., 1993; Thibodeau et al., 1993; Ionov et al., 1993; Olschwang et al., 1997). In the present study, we have unambiguously classi ed the 22 colorectal cancer cell lines within two distinct groups, MSI-H/LOH7 and MSS/LOH+, showing that these groups were exclusive. Some genetic alterations such as Ki-ras mutation or p16 inactivation were equally present in the two groups. Mutations of p53 and APC were more frequently involved in MSS/LOH+ than in MSI-H cell lines characterized by mutations in genes containing coding repeats. b-catenin mutations were also only observed in the MSI-H group in cell lines without APC mutations. These results con rm the di erent repertoires of genetic alterations involved in both pathways of colorectal tumorogenesis. It is known that LOH cell lines continue to show chromosomal instability during in vitro passages (Lengauer et al., 1997). The same is true for MSI-H cell lines which can have additional mutations within coding repeats (Zhang et al., 2000). In four cases, there were discrepancies between our own results and published ones, and these could be due to di erent genetic changes with time when cell lines are grown in vitro. By comparative genomic hybridization and ow cytometry analysis, Georgiades et al. (1999) suggested the existence of a third subset of colorectal cancers distinct from the LOH and MSI types. None of the 22 cell lines of the present series could belong to this third putative type, with maybe the only exception of the LS513 cell line which does not present genetic alterations except a LOH on chromosome 5 and a Ki-ras mutation. It has also been shown that a number of primary colorectal tumors have a so-called CpG island methylator phenotype (CIMP) (Toyota et al., 1999). This phenotype is not exclusive from the MSI-H phenotype since about 70% of the MSI-H primary tumors are CIMP+. The correlations between the CIMP and LOH phenotypes are presently unknown, but since there are many CIMP+ tumors which are not MSI-H, a number of them should be CIMP+/ LOH. The CIMP phenotype was not really determined in the cell lines of this study although we analysed methylation defects in the promoters of two genes. The Co115 and SW48 cell lines could be tentatively de ned as MSI-H/CIMP+ since both of them present hmlh1 promoter methylation. We have performed an extensive analysis of genetic alterations in a series of 22 colorectal cancer cell lines. Although this study is not completely exhaustive, it is the most complete report of this type to date. It should be a very useful starting point to establish a database of Genetic Alterations in Colorectal Cancer Cell lines which could be progressively updated by other authors ( This database should be most useful for the understanding of the cellular biology and pharmacology of the same cell lines. Materials and methods Cell lines Twenty-two cell lines were obtained from Dr Sordat (Epalinges, Switzerland), Dr Zweibaum (Villejuif, France), Dr Poupon (Curie, France) or purchased from the American Type Culture Collection. DNA was extracted by proteinase K digestion and phenol-chloroform extraction, and total RNA was extracted with Trizol LS (Gibco BRL Life Technology) according to manufacturer instructions. Microsatellite analysis The poly-ca microsatellite loci on chromosomes 2, 3, 5, 17 and 18 (list available on request) were selected from the Genethon database (Gyapay et al., 1994). Primers speci c for each locus were used to amplify the repeat and short anking sequences from template DNA using multiplex PCR as described (Muzeau et al., 1997). PCR products were electrophoresed on 7M urea/32% formamide/6% polyacrylamide denaturing gel, transferred onto Hybond nylon membranes (Amersham) and revealed by hybridization with a 32 P-labeled (CA) 12 probe. Mononucleotide repeat microsatellites analysis Primers used to amplify BAT-26 and BAT-25, and PCR conditions were as described (Parsons et al., 1995; Zhou et al., 1997). The TGFbRII, IGFRII, BAX, MSH3, MSH6, caspase-5 and TCF-4 genes containing coding repeats were ampli ed using primer sets as described (Parsons et al., 1995; Souza et al., 1996; Rampino et al., 1997; Malkhosyan et al., 1996; Schwartz et al., 1999; Duval et al., 1999). Eighteen other genes containing coding mononucleotide repeats were

7 also ampli ed with speci c primers (sequences available on request). All PCR products were separated on 5.6M urea/ 32% formamide/7% polyacrylamide gels, transferred overnight onto Hybond N+ nylon membranes (Amersham) and hybridized with a 32 P-labeled probe made with one of the primers used in the corresponding PCR reaction. The trinucleotide repeat of the E2F-4 gene was examined by the same technique using a primer set as described (Yoshitaka et al., 1996). Denaturing gradient gel electrophoresis DGGE running conditions were de ned using the Lerman and Silverstein program (1987). Primers and DGGE conditions for p53 were as described (Hamelin et al., 1993). Exons 1 and 2 of Ki-ras were ampli ed using primers as described (RidanpaÈ aè and Husgafvel-Pursiainen, 1993). PCR products of exon 1 were run at 160 V for 4 h in a 10 ± 60% denaturing gradient gel maintained at 608C, while exon 2 was run in a 25 ± 40% gradient for 2.5 h under the same conditions. For the exon 3 of the b-catenin gene, primers were as described (de la Coste et al., 1998) with a GC-clamp, and PCR product was run 3.5 h in a 20 ± 70% denaturing gradient gel. Moreover, we also performed PCR experiments using genomic DNA and primer pairs as described (Iwao et al., 1998) to detect interstitial deletions involving exon 3 of the b- catenin gene. Protein truncation assays Total RNA was reverse transcribed using random hexamers and a Perkin-Elmer kit (Roche Molecular Systems). RT ± PCR products for the hmsh2, hmlh1, hpms1 and hpms2 genes were in vitro transcribed and translated using the TnT coupled Reticulocyte Lysate System (Promega) as indicated by the manufacturer. All sense primers contained a 29-base T7 promoter sequence followed by an 8-base translational initiation signal (Sarkar and Sommer, 1989). Speci c primers were as described by Luce et al. (1995) for hmsh2 and hmlh1, and by Nicolaides et al. (1994) for hpms1 and hpms2. The APC gene was ampli ed by RT ± PCR for exons 1 ± 14, and by PCR on genomic DNA for exon 15 with primers as described (Powell et al., 1993), and in vitro transcribed and translated with the same TnT kit. Analysis of promoter methylation Genomic DNA samples were digested with HpaII (methylation-sensitive) and MspI (methylation-non sensitive) restriction enzymes before PCR ampli cation of the hmlh1 promoter with a primer set as described (Kane et al., 1997). The methylation of the p16 promoter was examined by methylation-speci c PCR (MSP) after bisul te DNA treatment as described (Herman et al., 1996) DNA sequencing Cell lines showing altered DGGE pro les for p53 and Ki-ras were sequenced using an ABI Prism TM 377 sequencer. The p16 gene was directly sequenced using primer sets as described (Li et al., 1995). Acknowledgments We thank Dr Elizabeth Newcomb for critical reading of the manuscript and Emmanuel Tubachei for the gacc WEB page. This work was partly supported by grants from the Association pour la Recherche contre le Cancer. References Aaltonen LA, PeltomaÈ ki P, Leach FS, Sistonen P, PylkkaÈ nen L, Mecklin JP, JaÈ rvinen GM, Kinzler KW, Vogelstein B and de la Chapelle A. (1993). Science, 260, 812 ± 816. Belgrader P, Cheng J, Zhou X, Stephenson LS and Maquat LE. (1994). Mol. Cell. Biol., 14, 8219 ± Boland CR, Thibodeau SN, Hamilton SR, Sidransky D, Eshleman JR, Burt RW, Meltzer SJ, Rodriguez-Bigas MA, Fodde R, Ranzani GN and Srivastava S. (1998). Cancer Res., 58, 5248 ± Bronner CE, Baker SM, Morrison PT, Warren G, Smith LG, Lescoe MK, Kane M, Earabino C, Lipford J, Lindblom A, Tannergard P, Bollag RJ, Godwin AR, Ward DC, Nordenskjold M, Fishel R, Kolodner R and Liskay RM. (1994). Nature, 368, 258 ± 261. Casares S, Ionov Y, Ge HY, Stanbridge E and Perucho M. (1995)., 11, 2303 ± Chung DC. (2000). Gastroenterology, 119, 854 ± 865. Cottrell S, Bicknell D, Kakamanis L and Bodmer WF. (1992). Lancet, 340, 626 ± 630. Cottu PH, Muzeau F, Estreicher A, Flejou JF, Iggo R, Thomas G and Hamelin R. (1996)., 13, 2727 ± da Costa LT, Liu B, El-Deiry WF, Hamilton S, Kinzler KW, Vogelstein B, Markowitz S, Willson JKW, de la Chapelle A, Downey KM and So AG. (1995). Nat. Genet., 9, 10 ± 11. de la Coste A, Romagnolo B, Billuart P, Renard CA, Buendia MA, Soubrane O, Fabre M, Chelly J, Beldjord C, Kahn A and Perret C. (1998). Proc. Natl. Acad. Sci. USA, 95, 8847 ± Deng G, Chen A, Hong J, Chae HS and Kim YS. (1999). Cancer Res., 59, 2029 ± Duval A, Gayet J, Zhou XP, Iacopetta B, Thomas G and Hamelin R. (1999). Cancer Res., 59, 4213 ± Duval A, Rolland S, Tubacher E, Bui H, Thomas G and Hamelin R. (2000). Cancer Res., 60, 3872 ± Duval A, Rolland S, Compoint A, Tubacher E, Iacopetta B, Thomas G and Hamelin R. (2001). Hum. Mol. Genet., 10, 513 ± 518. Fearon ER and Vogelstein B. (1990). Cell, 61, 759 ± 767. Fishel R, Lescoe MK, Rao MRS, Copeland NG, Jenkins NA, Garber J, Kane M and Kolodner R. (1993). Cell, 75, 1027 ± Georgiades IB, Curtis LJ, Morris RM, Bird CC and Wyllie AH. (1999)., 18, 7933 ± Gyapay G, Morissette J, Vignal A, Dib C, Fizames C, Millasseau P, Marc S, Bernardi G, Lathrop M and Weissenbach J. (1994). Nat. Genet., 7, 246 ± 338. Hamelin R, Jego N, Laurent-Puig P, Vidaud M and Thomas G. (1993)., 8, 2213 ± HeinenCD,RichardsonD,WhiteRandGrodenJ.(1995). Proc. Natl. Acad. Sci. USA, 55, 4797 ± Herman JG, Merlo A, Mao L, Lapidus RG, Issa JPJ, Davidson NE, Sidransky D and Baylin SB. (1995). Cancer Res., 55, 4525 ± Herman JG, Gra JR, MyoÈ haè nens,nelkinbdandbaylin SB. (1996). Proc. Natl. Acad. Sci. USA, 93, 9821 ± 9826.

8 5032 Herman JG, Umar A, Polyak K, Gra JR, Ahuja N, Issa JPJ, Markowitz S, Willson JKV, Hamilton SR, Kinzler KW, Kane MF, Kolodner R, Vogelstein B, Kunkel TA and Baylin SB. (1998). Proc. Natl. Acad. Sci. USA, 95, 6870 ± Hoang JM, Cottu PH, Thuille B, Salmon RJ, Thomas G and Hamelin R. (1997). Cancer Res., 57, 300 ± 303. Ilyas M, Tomlinson IPM, Rowan A, Pignatelli M and Bodmer WF. (1997). Proc. Natl. Acad. Sci. USA, 94, ± Ilyas M, Straub J, Tomlinson IPM and Bodmer WF. (1999). Eur. J. Cancer, 35, 335 ± 351. Ionov Y, Peinado MA, Malkhosyan S, Shibata D and Perucho M. (1993). Nature, 363, 558 ± 561. Iwao K, Nakamori S, Kameyama M, Kinoshita M, Fukui T, Ishiguro S, Nakamura Y and Miyoshi Y. (1998). Cancer Res., 58, 1021 ± Kamb A, Gruis NA, Weaver-Feldhaus J, Liu Q, Harshman K, Tavtigian SV, Stockert E, Day III RS, Johnson BE and Skolnick MH. (1994). Science, 264, 436 ± 440. Kane MF, Loda M, Gaida GM, Lipman J, Mishra R, Goldman H, Jessup JM and Kolodner R. (1997). Cancer Res., 57, 808 ± 811. Kitaeva MN, Grogan L, Williams JP, Dimond E, Nakahara K, Hausner P, DeNobile JW, Soballe PW and Kirsch IR. (1997). Cancer Res., 57, 4478 ± Leach FS, Nicolaides NC, Papadopoulos N, Liu B, Jen J, Parsons R, PeltoÈ maki P, Sistonen P, Aaltonen P, NystroÈ m-lahtim,guanxj,zhangj,metlzerps,yu JW,KaoFT,ChenDJ,CerosalettiKM,KeithFournier RE,ToddS,LewisT,LeachRJ,NaylorSL,Weissenbach J, Mecklin JP, JaÈ rvinen H, Petersen GM, Hamilton SR, GreenJ,JassJ,WatsonP,LynchHT,TrentJM,dela Chapelle A, Kinzler KW and Vogelstein B. (1993). Cell, 75, 1215 ± Lengauer C, Kinzler KW and Vogelstein B. (1997). Nature, 386, 623 ± 627. Lerebours F, Olschwang S, Thuille B, Schmitz A, Fouchet P, Laurent-Puig P, Boman F, Flejou JF, Monges G, Paraf F, Bedossa P, Sabourin JC, Salmon RJ, Parc R and Thomas G. (1999). Genes Chromosomes Cancer, 25, 147 ± 153. Lerman LS and Silverstein K. (1987). Meth. Enzymol., 155, 482 ± 501. LiYJ,Hoang-XuanK,DelattreJY,PoissonM,ThomasG and Hamelin R. (1995)., 11, 597 ± 600. Liu B, Nicolaides NC, Markowitz S, Willson JKV, Parsons RE, Jen J, Papadopoulos N, PeltomaÈ ki P, de la Chapelle A, Hamilton SR, Kinzler KW and Vogelstein B. (1995). Nat. Genet., 9, 48 ± 55. LuceMC,MarraG,ChauhanDP,LaghiL,CarethersJM, Cherian SP, Hawn M, Binnie CG, Kam-Morgan LNW, Cayouette MC, Koi M and Boland RC. (1995). Gastroenterology, 109, 1368 ± Malkhosyan S, Rampino N, Yamamoto H and Perucho M. (1996). Nature, 382, 499 ± 500. Morin PJ, Sparks AB, Korinek V, Barker N, Clevers H, Vogelstein B and Kinzler KW. (1997). Science, 275, 1787 ± Muzeau F, Flejou JF, Belghiti J, Thomas G and Hamelin R. (1997). Br.J.Cancer,75, 1336 ± MyoÈ haè nen SK, Baylin SB and Herman JG. (1998). Cancer Res., 58, 591 ± 593. Nicolaides NN, Papadopoulos N, Liu B, Wei YF, Carter KC, Ruben SM, Rosen CA, Haseltine WA, Fleischmann RD, Fraser CM, Adams MD, Venter JC, Dunlop MG, Hamilton SR, Petersen GM, de la Chapelle A, Vogelstein B and Kinzler KW. (1994). Nature, 371, 75 ± 80. Okamoto A, Demetrick DJ, Spillare EA, Hagiwara K, Hussain SP, Bennett WP, Forrester K, Gerwin B, Serrano M, Beach DH and Harris CC. (1994). Proc. Natl. Acad. Sci. USA, 91, ± Olschwang S, Hamelin R, Laurent-Puig P, Thuille B, de Rycke Y, Li YJ, Muzeau F, Girodet J, Salmon RJ and Thomas G. (1997). Proc. Natl. Acad. Sci. USA, 94, ± Papadopoulos N, Nicolaides NC, Wei YF, Ruben SM, CarterKC,RosenCA,HaseltineWA,FleischmannRD, Fraser CM, Adams MD, Venter JC, Hamilton SR, Petersen GM, Watson P, Lynch HT, PeltomaÈ ki P, Mecklin JP, de la Chapelle A, Kinzler KW and Vogelstein B. (1994). Science, 263, 1625 ± Papadopoulos N, Nicolaides NC, Liu B, Parsons R, Lengauer C, Palombo F, D'Arrigo A, Markowitz S, Willson JKV, Kinzler KW, Jiricny J and Vogelstein B. (1995). Science, 268, 1915 ± Parsons R, Myero LL, Liu B, Willson JKV, Markowitz SD, Kinzler KW and Vogelstein B. (1995). Cancer Res., 55, 5548 ± Peinado MA, Fernandez-Renart M, Capella G, Wilson L and Perucho M. (1993). Int. J. Onc., 2, 123 ± 134. Powell SM, Petersen GM, Krush AJ, Booker S, Jen J, Giardello FM, Hamilton SR, Vogelstein B and Kinzler KW. (1993). N.Engl.J.Med.,329, 1982 ± RampinoN,YamamotoH,IonovY,LiY,SawaiH,Reed JC and Perucho M. (1997). Science, 275, 967 ± 969. RidanpaÈ aè M and Husgafvel-Pursiainen K. (1993). Hum. Mol. Genet., 2, 639 ± 644. Sarkar S and Sommer SS. (1989). Science, 244, 331 ± 334. Schwartz S, Yamamoto H, Navarro M, Maestro M, Reventos J and Perucho M. (1999). Cancer Res., 59, 2995 ± Souza RF, Appel R, Yin J, Wang S, Smolinski KN, AbrahamJM,ZouTT,ShiYQ,LeiJ,CottrellJ,Cymes K, Biden K, Simms L, Leggett B, Lynch PM, Frazier M, Powell SM, Harpaz N, Sugimura H, Young J and Meltzer SJ. (1996). Nat. Genet., 14, 255 ± 257. Sparks AB, Morin PJ, Vogelstein B and Kinzler KW. (1998). Cancer Res., 58, 1130 ± Thibodeau SN, Bren G and Schaid D. (1993). Science, 260, 816 ± 819. Toyota M, Ahuja N, Ohe-Toyota M, Herman JG, Baylin SB and Issa JPJ. (1999). Proc. Natl. Acad. Sci. USA, 96, 8681 ± VeiglML,KasturiL,OlechnowiczJ,MaA,LutterbaughJD, Periyasamy S, Li GM, Drummond J, Modrich P, Sedwick WD and Markowitz SD. (1998). Proc. Natl. Acad. Sci. USA, 95, 8698 ± Wheeler JMD, Beck NE, Kim HC, Tomlinson IPM, Mortensen NJMcC and Bodmer WF. (1999). Proc. Natl. Acad.Sci.USA,96, ± Yoshitaka T, Matsubara N, Ikeda M, Tanino M, Hanafusa H, Tanaka N and Shimizu K. (1996). Biochem. Biophys. Res. Com., 227, 553 ± 557. Zhang L, Yu J, Park BO, Kinzler KW and Vogelstein B. (2000). Science, 290, 989 ± 992. Zhou XP, Hoang JM, Cottu P, Thomas G and Hamelin R. (1997)., 15, 1713 ± 1718.