Absence of mutation in the putative tumor-suppressor gene ... - Nature

Oncogene (2005) 24, 7253–7256

& 2005 Nature Publishing Group All rights reserved 0950-9232/05 $30.00 www.nature.com/onc

Absence of mutation in the putative tumor-suppressor gene KLF6 in colorectal cancers Astrid Lie`vre1, Bruno Landi2, Jean-François Coˆte´1, Nicolas Veyrie1, Jessica Zucman-Rossi3, Anne Berger4 and Pierre Laurent-Puig*,1,5 1

INSERM U490 Laboratoire de Toxicologie Molećulairei, Universite´ Rene´ Descartes, 45 rue des Saints-Pe`res, 75006 Paris, France; Assistance Publique Hoˆpitaux de Paris, Hoˆpital Europeén Georges Pompidou, service de gastroente´rologie, 20 rue Leblanc, 75015 Paris, France; 3INSERM U674 Geńomique des tumeurs solides, 27 rue Juliette Dodu, 75010 Paris, France; 4Assistance Publique Hoˆpitaux de Paris, Hoˆpital Europeén Georges Pompidou, service de chirurgie geńe´rale digestive et oncologique, 20 rue Leblanc, 75015 Paris, France; 5Assistance Publique Hoˆpitaux de Paris, Hoˆpital Europeén Georges Pompidou, service de Biochimie, 20 rue Leblanc, 75015 Paris, France 2

The KLF6 gene encodes the Kru¨ppel-like factor 6, a transcription factor that has been individualized as a tumor-suppressor gene involved in the regulation of cell proliferation and differentiation. Recently, high frequency (42%) of KLF6 mutations have been reported in colorectal cancers (CRC) as in prostate cancers, astrocytic gliomas and hepatocellular carcinomas. The aims of the study was to confirm the frequency of KLF6 mutations in a larger series of CRC than that previously published by using DNA extracted from frozen tissue samples, which have been proved to generate less mutational artefact than that extracted from formalin-fixed paraffin-embedded tissue samples, in order to compare KLF6 mutation frequency with that of other common genetic alterations and to determine genotype–phenotype correlations. Amplification and direct sequencing of KLF6 exon 2 of 76 CRC and matched normal frozen tissues was performed. Polymorphisms were observed in 14 cases, among which two (T35T and S116S) had not already been reported. No KLF6 somatic mutation was observed. Our data suggest a minor role of KLF6 mutation in colorectal carcinogenesis and underline the fact that the validity of sequence informations obtained from DNA extracted from formalin-fixed tissues may be limited. Oncogene (2005) 24, 7253–7256. doi:10.1038/sj.onc.1208867; published online 25 July 2005 Keywords: colorectal cancer; KLF6; mutation

Colorectal cancer (CRC) is one of the most common human malignancies with more than 300 000 cases both in the United States and in the European Union each year. Significant progress has been made to elucidate the molecular mechanisms involved in colorectal tumor *Correspondence: P Laurent-Puig, INSERM UMR-S490 Laboratoire de Toxicologie Molećulaire, Universite´ Rene´ Descartes, 45 rue des Saints-Pe`res, 75006 Paris, France; E-mail: [email protected] Received 14 February 2005; revised 21 April 2005; accepted 23 May 2005; published online 25 July 2005

development and progression. As in most solid tumors the malignant transformation of colonic epithelial cells is thought to arise through a multistep process during which they acquire genetic changes involving the activation of proto-oncogenes and the loss of tumor supressor genes (Vogelstein et al., 1988). Two different types of genetic alteration mechanisms have been identified in colorectal carcinomas: microsatellite instability that results of an alteration of DNA mismatch repair genes such as MLH1 and MSH2 (Peltomaki and Vasen, 1997) and chromosomal instability characterized by a loss of heterozygosity (LOH) at loci harboring tumor-suppressor genes such as APC and TP53. Kru¨ppel-like factors are core transcription factors that regulate numerous mammalian genes (Bieker, 2001). The KLF6 gene encodes one of these transcription factors, the Kru¨ppel-like factor 6 (KLF6), which has been individualized as a tumor-suppressor gene involved in the regulation of cell proliferation and differentiation. Indeed, KLF6 mediates growth inhibition by an upregulation of the cell cycle inhibitor CDKN1A (p21WAF1/CIP1) through an interaction with cyclin D1 in a TP53-independent manner (Narla et al., 2001; Benzeno et al., 2004). Tumor-derived KLF6 mutants fail to induce CDKN1A or suppress proliferation in prostate and CRC cells (Narla et al., 2001; Reeves et al., 2004). Recently, high frequency (42%) of KLF6 somatic mutations have been reported in CRC (Reeves et al., 2004). In order to (i) confirm this result that may be potentially important to better understand colorectal carcinogenesis, (ii) characterize the frequency of KLF6 mutations in CRC compared to other common genetic alterations and (iii) determine genotype–phenotype correlations, we looked for KLF6 mutations in a larger series of CRC than that previously reported by using DNA extracted from frozen tissue samples that have been proved to generate less mutational artefact than that extracted from formalin-fixed paraffinembedded tissue samples (Williams et al., 1999). In the present study, 76 tumors were collected from patients who underwent a surgical resection of CRC in Laennec Hospital (Paris, France) between 1997 and

Absence of KLF6 mutation in Colorectal cancers A Lie`vre et al

7254

1999. The sex ratio (M : F) was 1 : 1 and the mean age of the patients was 70.5 years old. The tumor location was proximal, distal and rectal in 32.5, 44.5 and 23% of the cases, respectively. The tumor stage was I or II in 51% of the cases, III and IV in 26 and 23% of the cases, respectively. All DNA were extracted from tumor and matched normal frozen tissue sections after a histological HES staining and only tumors with more than 70% of tumor cells were retained. Amplification and direct sequencing of KLF6 exon 2 was performed as almost all mutations previously found in tumors, including CRC, were confined to this exon that encodes three fourths of the wild-type protein (Table 1). The results of DNA sequence analysis were compared with the published reference KLF6 sequence (GenBank, access number NM_1300) to identify sequence variants. The analysis of both colorectal tumor and matched normal tissues DNA allowed us to distinguish germline polymorphisms from somatic mutations. Each DNA sequence variant was checked against databases and those not recorded were categorized as novel. Nucleotide polymorphisms were observed in 14 cases, a G>A polymorphism in the intron 1 in position 3023 (27 upstream to exon 2) (n ¼ 10) and three synomymous coding polymorphisms were found: T35T (n ¼ 1), S116S (n ¼ 1) and R201R (n ¼ 2) (Table 2, Figure 1). Two of these polymorphims (G3023A and R201R) had been already reported (Kohler et al., 2004; Koivisto et al., 2004a, b; Montanini et al., 2004; Reeves et al., 2004; Boyault et al., 2005). We did not find mutation in KLF6 exon 2 in any of the 76 tumor samples analysed. As it was reported that KLF6 could also be inactivated by LOH in a large number of CRC (Reeves et al., 2004), we analysed allelic loss with four microsatellite markers surrounding KLF6 locus on chromosome 10p. A total of 55 of the 72 tumors (76%) without microsatellite instability phenotype were informative and a loss of KLF6 locus was found in five (9%) of them (Figure 2). In these series of tumors, LOH

at chromosome 17p, 18q, 8p, 5q and 22q was observed respectively in 63, 59, 30, 37 and 25%, which is consistent with previous studies (Halling et al., 1999;

Table 1 Primer sequences used for PCR amplification and sequencing of KLF6 exon 2 Primer name

Primer sequence (50 –30 )

KLF6EX2F KLF6EX2R KLF6EX2F1a KLF6EX2R1a

CTGTCCTTCATTCTTGCATGT GACCACATCCTGTGCAGCC CTTAGAGACCAACAGCCTGA GTTGAGAAGGTTCCCTGCTC

F: forward, R: reverse. aPrimers used for the sequencing only

Table 2

Germline polymorphisms in KLF6 exon 2 from colorectal cancers

DNA sample number L6, L27, L49, L61, L78, L74, L71, L82, L94, L69 L15, L96 L3 L20 R: arginine, S: serine, T: threonine Oncogene

Figure 1 Three representative chromatograms of DNA sequence variants located in the same tumor by direct sequencing analysis: (a) G3023A germline polymorphism in the intron 1. The KLF6 exon 2 was amplified and sequenced by using primers referenced in Table 1. For amplification, 50 ng DNA was used in a final 20 ml mix containing: 1 final concentration of 10 PCR buffer (Qiagen, Coutaboeuf, France), 200 mM of each dNTP, 0.3 mM of each primer, 0.2 ml of HotStartTaq DNA Polymerase (Qiagen) under the following step-down PCR protocol: 951C for 15 min, 2 cycles; 951C for 15 s, 601C for 30 s, 721C for 2 min; 2 cycles; 951C for 15 s, 591C for 30 s, 721C for 2 min, 2 cycles; 951C for 15 s, 581C for 30 s, 721C for 2 min, 3 cycles; 951C for 15 s, 571C for 30 s, 721C for 2 min, 3 cycles; 951C for 15 s, 561C for 30 s, 721C for 2 min, 4 cycles; 951C for 15 s, 551C for 30 s, 721C for 2 min, 4 cycles; 951C for 15 s, 541C for 30 s, 721C for 2 min, 5 cycles; 951C for 15 s, 531C for 30 s, 721C for 2 min, 10 cycles; 951C for 15 s, 521C for 30 s, 721C for 2 min and a final extension at 721C for 5 min. Each successfully amplified fragment was purified and sequenced using the Big Dye Terminator V3.1 kit (Applied Biosystems, Foster City, CA, USA) on an ABI Prisms 3100 DNA Analyser automated sequencer (Applied Biosystems). Two ml of purified PCR product were used in a final 20 ml mix containing: 1 ml of Big Dye Terminator V 3.1, 0.875 final concentration of 5 Big Dye Terminator sequencing buffer and 0.5 mM of each primer. DNA was then submitted to the following cycle conditions: 961C for 4 min; 25 cycles, 961C for 10 s, 551C for 5 s and 601C for 4 min. Germline polymorphisms were defined by sequence variations found in both normal and tumor tissues DNA whereas somatic mutations were defined by sequence variations found in tumor tissue DNA but not in matched normal tissue DNA (b) TP53 R213X somatic mutation and (c) KRAS G12D somatic mutation

Nucleotide position 3023 (intron 1) 3550 3295 3052

Nucleotide change

Codon position

Novel

G-A G-A C-T C-G

R201R S116S T35T

No No Yes Yes


7255

Figure 2 Data of the five tumors that had loss of at least two markers, consistent with KLF6 LOH. LOH status of the KLF6 locus was performed by using four fluorescently labeled microsatellite markers (D10S591, D10S189, D10S1745 and D10S249) located on chromosome 10. The relative positions of these markers and of KLF6 on chromosome 10 are indicated. Amplification of the four microsatellite markers was performed by a multiplex PCR using 50 ng DNA in a final 15 ml mix containing: 1 final concentration of 10 PCR buffer (Qiagen), 1 final concentration of 5 Q solution (Qiagen), 200 mM of each dNTP, 0.4 mM of each primer and 0.12 ml of HotStartTaq DNA Polymerase (Qiagen). PCR was performed under the following step-down protocol: 951C for 15 min, 35 cycles; 941C for 30 s, 531C for 30 s, 721C for 1 min. Amplified fragments were analysed after dilution on an ABI 3700 DNA Sequencer (Applied Biosystems). Allelic loss was analysed by Genotyper software (Applied Biosystems) and LOH was defined for each tumor as a ¼ (NL/NS)/(TL/TS) where L is the intensity of the larger allele and S the intensity of the smaller allele in normal DNA (N) or tumor DNA (T). An a-score p0.5 was defined as LOH. Black circle, LOH; open circle, noninformative marker

Watanabe et al., 2001; Choi et al., 2002; Zhou et al., 2002). Our results are unexpected and appear to be highly discordant with that reported by Reeves et al. (2004) who found KLF6 mutations in 21 of the 50 CRCs (42%) they analysed (Po0.0001). Interestingly, such discrepancies in incidence of KLF6 mutations have also been observed in prostate cancer. Narla et al. (2001) originally found KLF6 mutations in 18 of 33 (55%) prostate carcinomas. Chen et al. (2002) identified somatic KLF6 mutations in only 14 of 96 prostate cancers (15%), whereas Muhlbauer et al. (2003) did not found any mutations in 32 prostate tumors. Moreover, none of the mutations found by Chen et al. (2002) were identical to those reported in the previous study. The same discrepant result was obtained in hepatocellular carcinomas initially reported to be mutated in 15% of the cases by the same authors than those who published firstly on prostate and colorectal carcinomas (Kremer-Tal et al., 2004). Indeed, no mutation was found in a larger series of hepatocellular carcinoma by another group (Boyault et al., 2005). Finally, astrocytic gliomas show the same difference with no mutation

found in three different studies (Kohler et al., 2004; Koivisto et al., 2004b; Montanini et al., 2004) after a first report that described 9% of KLF6 mutations in the exon 2 (Jeng and Hsu, 2003). The reasons of the difference of KLF6 mutation frequency between the study of Reeves et al. (2004) and the present study remain to be elucidated. It is not likely to be due to a failure in our mutation screening method since its sensitivity has been previously validated in a larger series of tumor DNA samples including those presently reported by the identification of TP53 and KRAS2 gene mutations in 51 and 37% of the cases (Ferraz et al., 2004). These frequencies of TP53 and KRAS mutations are similar to that reported in previous published series of CRC (Hamelin et al., 1994; Andreyev et al., 1998; Smith et al., 2002). One possible reason of the difference observed could be that the study by Reeves et al. (2004) was carried out using DNA extracted from microdissected formalin-fixed paraffinembedded tissue samples, a procedure that has previously been shown to enhance the risk of false-positive mutations (Williams et al., 1999). The formalin fixation is known to lower the success of polymerase chain reaction (PCR) but the exact mechanism for modification of DNA in formalin-fixed samples remain to be elucidated. Formaldehyde forms monoethylol adducts with nucleotide rings which can be irreversible (Feldman, 1973). These base damages may impair PCR by halting polymerase elongation, but another possibility is error-prone translesion synthesis across sites of damage responsible for artefactual mutations (Goodman, 2002). AmpliTaq Gold DNA polymerase (Applied Biosystems) commonly used for PCR as it was the case in the study by Reeves et al. (2004) is from A-family of DNA polymerases that lack 30 exonuclease proofreading activity (Tindall and Kunkel, 1988) and exhibit translesion synthesis with modified DNA templates (Smith et al., 1998; Duarte et al., 2000). Thus, if formalin causes DNA damages and Taq DNA polymerase performs errorprone translesional misinterpretation of these damages, analysis of fixed human tissues may be altered. Quach et al. (2004) have also noted three to fourfold more mutations when formalin-fixed specimens are examined by PCR compared to fresh tissues, although they used 100 ng of template DNA like Reeves et al. (2004) did. Other studies found high frequencies of mutations in ovarian and lung cancer that have finally been proved to be artefact due to formalin-fixation (Wong et al., 1998; Yngveson et al., 1999). Thus, the validity of sequence information obtained from DNA isolated from chemically fixed tissues must be questioned in the absence of corroborating data from frozen tissues. In this way, it is interesting to note that no mutation were identified in four CRC cell lines (SW480, HCT115, HCT116 and HT29) analysed in the study by Reeves et al. (2004). This argument of artefactual mutations caused by formalin-fixation has already been raised to explain the discrepant results obtained for KLF6 mutations in astrocytic gliomas (Kohler et al., 2004). KLF6 allelic loss was found in 9% of the cases and may be another mechanism of KLF6 inactivation as it was reported Oncogene


7256

previously (Reeves et al., 2004; Cho et al., 2005), just as alternative splicing that has been shown to be enhanced by the intron 1 G3023A germline polymorphism (Narla et al., 2005) that was found in 13% of the cases (10/76) in the present series. It is interesting to note that none of the tumors harboring this polymorphism was associated with the loss of KLF6 locus. In conclusion, our findings do not support a major role of KLF6 mutations in colorectal carcinogenesis. Furthermore, our results underline the fact that the

validity of sequence informations obtained from DNA extracted from formalin-fixed tissues may be limited, which clearly indicates the significance of setting up fresh-frozen tissue banks in future studies for mutational screening of genes suspected to be involved in a carcinogenesis pathway.

Acknowledgements Le Ligue Nationale de Lutte Contre le Cancer.

References Andreyev HJ, Norman AR, Cunningham D, Oates JR and Clarke PA. (1998). J. Natl. Cancer Inst., 90, 675–684. Benzeno S, Narla G, Allina J, Cheng GZ, Reeves HL, Banck MS, Odin JA, Diehl JA, Germain D and Friedman SL. (2004). Cancer Res., 64, 3885–3891. Bieker JJ. (2001). J. Biol. Chem., 276, 34355–34358. Boyault S, Herault A, Balabaud C and Zucman-Rossi J. (2005). Hepatology, 41, 681–682; author reply 682–683. Chen HK, Liu XQ, Lin J, Chen TY, Feng QS and Zeng YX. (2002). Ai Zheng, 21, 1047–1050. Cho YG, Kim CJ, Park CH, Yang YM, Kim SY, Nam SW, Lee SH, Yoo NJ, Lee JY and Park WS. (2005). Oncogene. 11 April [E-pub ahead of print]. Choi SW, Lee KJ, Bae YA, Min KO, Kwon MS, Kim KM and Rhyu MG. (2002). Clin. Cancer Res., 8, 2311–2322. Duarte V, Gasparutto D, Jaquinod M and Cadet J. (2000). Nucleic Acids Res., 28, 1555–1563. Feldman MY. (1973). Prog. Nucleic Acid Res. Mol. Biol., 13, 1–49. Ferraz JM, Zinzindohoue F, Lecomte T, Cugnenc PH, Loriot MA, Beaune P, Stucker I, Berger A and Laurent-Puig P. (2004). Int. J. Cancer, 110, 183–187. Goodman MF. (2002). Annu. Rev. Biochem., 71, 17–50. Halling KC, French AJ, McDonnell SK, Burgart LJ, Schaid DJ, Peterson BJ, Moon-Tasson L, Mahoney MR, Sargent DJ, O’Connell MJ, Witzig TE, Farr Jr GH, Goldberg RM and Thibodeau SN. (1999). J. Natl. Cancer Inst., 91, 1295–1303. Hamelin R, Laurent-Puig P, Olschwang S, Jego N, Asselain B, Remvikos Y, Girodet J, Salmon RJ and Thomas G. (1994). Gastroenterology, 106, 42–48. Jeng YM and Hsu HC. (2003). Int. J. Cancer, 105, 625–629. Kohler B, Wolter M, Blaschke B and Reifenberger G. (2004). Int. J. Cancer, 111, 644–645. Koivisto PA, Hyytinen ER, Matikainen M, Tammela TL, Ikonen T and Schleutker J. (2004a). J. Urol., 172, 506–507. Koivisto PA, Zhang X, Sallinen SL, Sallinen P, Helin HJ, Dong JT, Van Meir EG, Haapasalo H and Hyytinen ER. (2004b). Int. J. Cancer, 111, 642–643. Kremer-Tal S, Reeves HL, Narla G, Thung SN, Schwartz M, Difeo A, Katz A, Bruix J, Bioulac-Sage P, Martignetti JA and Friedman SL. (2004). Hepatology, 40, 1047–1052. Montanini L, Bissola L and Finocchiaro G. (2004). Int. J. Cancer, 111, 640–641.

Oncogene

Muhlbauer KR, Grone HJ, Ernst T, Grone E, Tschada R, Hergenhahn M and Hollstein M. (2003). Br. J. Cancer, 89, 687–690. Narla G, Difeo A, Reeves HL, Schaid DJ, Hirshfeld J, Hod E, Katz A, Isaacs WB, Hebbring S, Komiya A, McDonnell SK, Wiley KE, Jacobsen SJ, Isaacs SD, Walsh PC, Zheng SL, Chang BL, Friedrichsen DM, Stanford JL, Ostrander EA, Chinnaiyan AM, Rubin MA, Xu J, Thibodeau SN, Friedman SL and Martignetti JA. (2005). Cancer Res., 65, 1213–1222. Narla G, Heath KE, Reeves HL, Li D, Giono LE, Kimmelman AC, Glucksman MJ, Narla J, Eng FJ, Chan AM, Ferrari AC, Martignetti JA and Friedman SL. (2001). Science, 294, 2563–2566. Peltomaki P and Vasen HF. (1997). Gastroenterology, 113, 1146–1158. Quach N, Goodman MF and Shibata D. (2004). BMC Clin. Pathol., 4, 1. Reeves HL, Narla G, Ogunbiyi O, Haq AI, Katz A, Benzeno S, Hod E, Harpaz N, Goldberg S, Tal-Kremer S, Eng FJ, Arthur MJ, Martignetti JA and Friedman SL. (2004). Gastroenterology, 126, 1090–1103. Smith CA, Baeten J and Taylor JS. (1998). J. Biol. Chem., 273, 21933–21940. Smith G, Carey FA, Beattie J, Wilkie MJ, Lightfoot TJ, Coxhead J, Garner RC, Steele RJ and Wolf CR. (2002). Proc. Natl. Acad. Sci. USA, 99, 9433–9438. Tindall KR and Kunkel TA. (1988). Biochemistry, 27, 6008–6013. Vogelstein BFE, Hamilton SR, Kern SE, Preisinger AC, Leppert M, Nakamura Y, White R, Smits AMM and Bos JL. (1988). N. Engl. J. Med., 319, 525–532. Watanabe T, Wu TT, Catalano PJ, Ueki T, Satriano R, Haller DG, Benson III AB and Hamilton SR. (2001). N. Engl. J. Med., 344, 1196–1206. Williams C, Ponten F, Moberg C, Soderkvist P, Uhlen M, Ponten J, Sitbon G and Lundeberg J. (1999). Am. J. Pathol., 155, 1467–1471. Wong C, DiCioccio RA, Allen HJ, Werness BA and Piver MS. (1998). Cancer Genet. Cytogenet., 107, 21–27. Yngveson A, Williams C, Hjerpe A, Lundeberg J, Soderkvist P and Pershagen G. (1999). Cancer Epidemiol. Biomarkers Prev., 8, 433–438. Zhou CZ, Peng ZH, Zhang F, Qiu GQ and He L. (2002). World J. Gastroenterol., 8, 668–673.