Optimising methods for determining RER status in colorectal cancers

Cancer Letters 149 (2000) 15±20 www.elsevier.com/locate/canlet

Optimising methods for determining RER status in colorectal cancers J.G. Stone a, I.P.M. Tomlinson b, R.S. Houlston a,* b

a Section of Cancer Genetics, Institute of Cancer Research, Sutton, SM2 5NG, UK Molecular and Population Genetics Laboratory, Imperial Cancer Research Fund, London WC2A 3PX, UK

Received 9 July 1999; received in revised form 20 August 1999; accepted 22 August 1999

Abstract Approximately 13% of colorectal cancers display microsatellite instability (MSI), a form of replication error repair. Colorectal cancers developing in individuals with constitutional defects in the mismatch repair (MMR) genes hMLH1, hMSH2, hPMS1 and hPMS2 consistently show evidence of this phenomenon. Since MSI is indicative of MMR de®ciency, testing colorectal cancers for MSI provides a method of re®ning the identi®cation of carriers of germline MMR mutations. To assess which microsatellites represent the best reporters of replication error (RER) status we have examined 116 early onset colorectal cancers for MSI. MSI was assessed using eight dinucleotide-and two mononucleotide-repeat ¯uorescently labelled polymerase chain reaction (PCR) markers. The two mononucleotide repeat markers (BAT25 and BAT26) were highly sensitive and typing of either represents an ef®cient strategy for de®ning RER status of colorectal cancers and obviates the requirement of typing numerous microsatellite markers. q 2000 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Microsatellite instability; Replication error; Colorectal cancer

1. Introduction Microsatellites are short tandem repeat sequences of 1±6 bp, occurring throughout the genome [1]. Because of their repetitive nature they are prone to errors caused by slippage or stuttering during replication. Alterations in the length of microsatellite alleles in tumour DNA compared with constitutive DNA from the same individual represents a form of replication error (RER) referred to as microsatellite instability (MSI). Approximately 13% of colorectal cancers display MSI [2±5] but those developing in patients with constitutional defects in the mismatch repair (MMR) * Corresponding author. Tel: 144-181-643-8901; fax: 144-181643-0257. E-mail address: [email protected] (R.S. Houlston)

genes hMLH1, hMSH2, hPMS1 and hPMS2 consistently show evidence of this phenomenon [6,7]. Apart from germline or somatic mutations in the MMR genes, the basis of microsatellite instability in tumours is poorly understood, and the possibility exists that the patterns of instability indicate different underlying mechanisms. Tumours with dramatic changes at a single locus often show widespread instability at other loci. This is a characteristic feature of mutations in hMSH2, hMLH1 or hPMS2 [8]. Less clear are the mechanisms underlying low levels of MSI, though mutations in MSH3 and MSH6 probably play a role [8]. Colorectal cancers developing in carriers of MMR gene mutations are poorly differentiated and frequently multiple [3]. Paradoxically, despite these features, the prognosis appears to be more favourable than in sporadic colorectal cancer cases [9]. Further-

0304-3835/00/$ - see front matter q 2000 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0304-383 5(99)00324-9

16

J.G. Stone et al. / Cancer Letters 149 (2000) 15±20

more, although the genetic basis of hereditary nonpolyposis colorectal cancer (HNPCC) and sporadic cancers with RERs can be different, tumours in the two groups share some common biological characteristics in terms of prognosis: patients with RER-positive cancers have a survival advantage over RERnegative cases [10]. Since MSI is associated with MMR de®ciency, testing colorectal cancers for this phenotype provides a method of re®ning the identi®cation of HNPCC families and carriers of germline MMR gene defects [11]. Furthermore, determining the RER status of colorectal cancers is of prognostic relevance. Different criteria have been proposed for de®ning RER status based on the number of microsatellite loci which should be examined, and the percentage of unstable loci required to classify a tumour as RERpositive. Most of the criteria for de®ning RER status have evolved in a relatively ad hoc fashion and a recent consensus statement is based on an overview of analyses of comparatively small numbers of colorectal cancers by several different research groups [12,13] rather that one single large study. To assess which markers represent the best reporters of RER status, we have used ten microsatellites (eight dinucleotide and two mononucleotide repeats) to assess the frequency of MSI in a large series of colorectal cancers, and assessed which markers provide the most sensitive and ef®cient method for assessing RER status. 2. Patients and methods 2.1. Patients Paired blood-tumour samples were collected from 116 patients with histologically proven colorectal adenocarcinomas. All patients were aged less than 55 years at time of diagnosis and samples were obtained with informed consent and ethical review board approval. 2.2. Methods 2.2.1. DNA extraction DNA was salt-extracted from EDTA±blood samples using a standard sucrose lysis method. Formalin-®xed and paraf®n-embedded sections of

colorectal cancers prepared for routine histopathology were used for DNA preparation. Three 15-mm sections were cut onto double-sided, clear adhesive tape and placed on glass slides. These were lightly stained with toluidine blue and regions containing at least 60% tumour micro-dissected and placed in eppendorf tubes. Tumours were digested in buffer, 10 mM Tris±HCl (pH 7.5), 1 mM EDTA, 15% (w/ v) sodium dodecyl sulphate (SDS), and 500 mg/ml proteinase K, for 16 h at 568C. After phenol±chloroform extraction, DNA was precipitated with sodium acetate±ethanol. 2.2.2. MSI assessment Ten loci, eight dinucleotide (D1S508, D5S346, D11S29, TGFb IIR, DCC, D19S565, D2s123, D15S970) and two mononucleotide markers (BAT25, BAT26) were studied. The markers were selected on the basis of two criteria: (i) ampli®cation of fragments could be undertaken concurrently; and (ii) published studies had reported that the marker was a good reporter of RER status [14±16]. Target DNA sequences were ampli®ed in a 15-ml polymerase chain reaction (PCR) mixture containing 0.5 pmol of each primer (one ¯uorescently labelled), 2 mM each of dATP, dCTP, dTTP, dGTP, 1.5 mM MgCl2, 15 mg BSA, 0.06 ml Taq polymerase (Advanced Bio-technologies, Surrey, UK), and 12.5 ng sample DNA. PCR products were analysed on 6% polyacrylamide denaturing gels in 1 £ TBE buffer and dye-labelled PCR products were detected on ABI 377 DNA sequencers using Genescan and Genotyper software. MSI at each locus was de®ned by the presence of novel bands following PCR ampli®cation of tumour DNA which were not present in PCR products of the corresponding normal DNA. 2.2.3. Statistical analysis The strength of an association between MSI detected at different loci was assessed by means of Kendall's tau statistic, (t), implemented in the STATA program (Version 5., Stata Corp., TX). 3. Results Fig. 1 shows representative stable and unstable allelic pro®les at each of the ten loci. The percentage of

J.G. Stone et al. / Cancer Letters 149 (2000) 15±20

17

distribution of MSI in the tumours examined for these ten markers in the 116 tumours. Also shown is the relationship between MSI status at the two mononucleotide markers and the dinucleotide markers. Eleven of the tumours analysed displayed MSI at both BAT25 and BAT26, and of these, ®ve showed MSI at one or more dinucleotide repeat loci. One showed MSI at BAT25 only and one, MSI at BAT26 only. In both of these cases none of the tumours displayed MSI at any of the dinucleotide markers. one hundred and three tumours showed no evidence of MSI at either BAT25 or BAT26. Of these, only one showed evidence of MSI (at two dinucleotide loci) and none of these tumours displayed MSI at three or more loci. Based on these results, delineation of RER status using either BAT25 or BAT26 is associated with 93% sensitivity and a negative predictive value of 99%. Of the dinucleotide markers studied, instability at D3S1448 and D2S123 both showed a degree of correlation with BAT26 (t ˆ 0:31; P , 0:001 and t ˆ 0:39; P , 0:001, respectively). Instability at the other markers studied was less well correlated with either of the BAT markers. Fig. 1. Representative allelic pro®les of the ten microsatellite nucleotide repeats analysed.

colorectal cancers displaying MSI at each of the ten loci examined ranged from 0.4±10.3%. The most sensitive markers were BAT25 and BAT26. The detection of MSI using BAT25 and BAT26 was highly correlated (t ˆ 0:88; P , 0:001). Figs. 2 and 3 show the distribution of allele sizes for the BAT25 and BAT26 in RER-positive and RERnegative tumours, respectively. The range of the allele sizes for both markers was signi®cantly different between the two tumour groups. The allele size of the BAT25 and BAT26 markers were between 5±11 bp and 5±9 bp, respectively, in the RER-negative tumours, whereas the size of the marker alleles ranged from 10±16 bp and 9±20 bp in RER-positive tumours. Thus, using BAT25 or BAT26, the distribution in allele sizes in RER-positive and RER-negative tumours were suf®ciently different for RER status of cancers to be established solely on the basis of analysis of tumours. Table 1 shows the interrelationship, number and

4. Discussion Testing cancers for MSI provides a method of identifying cancers characterised by genomic instability. In colorectal cancer this provides a means of re®ning the identi®cation of those tumours caused by MMR gene mutations and delineating those cancers which may be associated with a more favourable prognosis [9±11]. Whilst a number of methods of detecting MSI have been reported using a variety of mono-, di-, tri-, tetra-, and other oligonucleotide microsatellite markers, there is debate about the minimal number of loci which should be examined and how many are required to display MSI for a tumour to be classi®ed as RERpositive. It has been suggested that for tumours to be considered RER-positive, 30±50% of microsatellite loci should be reported unstable [14,17±19]. Alternatively Bocker et al. [15] proposed that ®ve microsatellite markers, of different types, should be analysed and if fewer than two loci displayed MSI in this initial screen, a further ®ve loci should be examined. In this

18

J.G. Stone et al. / Cancer Letters 149 (2000) 15±20

Fig. 2. Distribution of BAT25 allele size in RER-negative and RER-positive tumours (PCR fragment size 120 bp).

scheme, tumours would be classi®ed as RER-positive if MSI were detected at more than 10% of loci, or if more than two loci displayed MSI. It is clear that aetiology of MSI is heterogeneous and hence the de®nition of RER status is not straightforward. However, it is clear that any markers used within any de®nition need to offer a high degree of sensitivity and confer a positive predictive value. Simply increasing the number of markers studied

per se ignores the fact that the information entropy may not change by these additional typings. Ideally it should be possible to determine RER status on the basis of a small but sensitive set of robust markers. The National Cancer Institute Research Workshop recently proposed that a panel of microsatellites be used (BAT25, BAT26, D5S346, D2S123 and D17S250) and tumours classi®ed as high frequency MSI if two or more of the ®ve markers show instabil-

Fig. 3. Distribution of BAT26 allele size in RER-negative and RER-positive tumours (PCR fragment size 120 bp).

J.G. Stone et al. / Cancer Letters 149 (2000) 15±20

19

Table 1 Distribution of MSI and relationship between MSI and the mononucleotide and dinucleotide repeats Mononucleotide repeat

Dinucleotide repeat

Bat-25 RER status

Bat-26 RER status

No.

No MSI

MSI at one locus

MSI at two loci

MSI at three or more loci

1 1 2 2

1 2 1 2

11 1 1 103

6 0 0 0

2 0 0 0

1 0 0 1

2 0 0 0

ity and low frequency MSI if only one marker shows MSI [12,13]. We have analysed ten loci for MSI in a large series of colorectal cancers using a combination of eight dinucleotide and two mononucleotide microsatellite markers in order to evaluate which marker combinations offered the best minimum set to establish RER status. The mononucleotide repeat markers, BAT25 and BAT26, were clearly the most sensitive and speci®c of the ten markers analysed and our results are concordant with the recent study reported by Zhou et al. [20] suggesting that typing of the single marker BAT-26 can be used to assign RER status of colorectal cancers in over 90% of cases. We also found that the distributions of BAT26 and BAT25 allele size in RERpositive and RER-negative tumours are suf®ciently different that analysis of blood samples may not be an absolute prerequisite in order to establish RER status. The BAT25 marker is a poly(A) tract located within the intron of the c-kit oncogene [21] and the mononucleotide repeat BAT26 is a poly(A) tract located in the ®fth intron of hMSH2 [16]. These poly(A) repeats appear to be more prone to aberrations than dinucleotide repeats and re¯ect the fact that increased formation of deletions or expansions occurs during DNA replication of these sequences. Our study was based on an analysis of early onset colorectal cancers. These have a higher probability of being caused by a mutation in one of the MMR genes compared with an unselected series of colorectal cancers. Hence the spectrum of MSI in this study may re¯ect this in part. Moreover, recent studies of colorectal cancer suggest high and low forms of microsatellite instability exist (denoted by MSH-L and MSI-H, respectively) [22]. These phenotypes are caused different molecular events, high levels of MSI are indicative of mutations in the MMR genes

associated with HNPCC [22]. In terms of detecting MSI-H in colorectal cancers typing a single mononucleotide marker, either BAT25 or BAT26 offers a high ef®ciency strategy. Our results strongly support the notion that RER status in colorectal cancers can be derived with considerable precision by simply typing one of the BAT markers rather than typing a myriad of markers for MSI. Furthermore, it supports the recently reported consensus statement by National Cancer Institute Research Workshop [12,13]. Acknowledgements This work was supported by the Cancer Research Campaign. References [1] D. Tautz, Notes on the de®nition and nomenclature of tandemly repetitive DNA sequences, Exper. Suppl. (Basel) 67 (1993) 21±28. [2] L.A. Aaltonen, P. Peltomaki, F.S. Leach, P. Sistonen, L. Pylkkanen, J.P. Mecklin, H. Jarvinen, S.M. Powell, J. Jen, S.R. Hamilton, G.M. Peterson, K.W. Kinzler, B. Vogelstein, A. de la Chapelle, Clues to the pathogenesis of familial colorectal cancer, Science 260 (5109) (1993) 812±816. [3] R.A. Lothe, P. Peltomaki, G.I. Meling, L.A. Aaltonen, M. Nystrom-Lahti, L. Pylkkanen, K. Heimdal, T.I. Anderson, P. Moller, T.O. Rognum, S.D. Fossa, T. Haldorsen, A. Brogger, A. de la Chapelle, A.L. Borresen, Genomic instability in colorectal cancer: relationship to clinicopathological variables and family history, Cancer Res. 53 (24) (1993) 5849±5852. [4] S.N. Thibodeau, G. Bren, D. Schaid, Microsatellite instability in cancer of the proximal colon, Science 260 (1993) 816±819. [5] H. Kim, J. Jen, B. Vogelstein, S.R. Hamilton, Clinical and pathological characteristics of sporadic colorectal carcinomas with DNA replication errors in microsatellite sequences, Am. J. Pathol. 145 (1) (1994) 148±156.

20

J.G. Stone et al. / Cancer Letters 149 (2000) 15±20

[6] P. Peltomaki, A. de la Chapelle, Mutations predisposing to hereditary nonpolyposis colorectal cancer, Adv. Cancer Res. 71 (1997) 93±119. [7] B. Liu, R. Parsons, N. Papadopolous, N.C. Nicolaides, H.T. Lynch, P. Watson, J.R. Jass, M. Dunlop, A. Wyllie, P. Peltomaki, A. de la Chapelle, S.R. Hamilton, B. Vogelstein, K.W. Kinzler, Analysis of mismatch repair genes in hereditary nonpolyposis colorectal cancer patients, Nat. Med. 2 (2) (1996) 169±174. [8] A. Percesepe, P. Kristo, L.A. Aaltonen, M. Ponz de Leon, A. de la Chapelle, P. Peltomaki, Mismatch repair genes and mononucleotide tracts as mutation targets in colorectal tumours with different degrees of microsatellite instability, Oncogene 17 (2) (1998) 157±163. [9] R. Sankila, L.A. Aaltonen, H.J. Jarvinen, J.P. Mecklin, Better survival rates in patients with MLH1-associated hereditary colorectal cancer, Gastroenterology 110 (3) (1996) 682±687. [10] V.J. Bubb, L.J. Curtis, C. Cunningham, M.G. Dunlop, A.D. Carothers, R.G. Morris, S. White, C.C. Bird, A.H. Wyllie, Microsatellite instability and the role of hMSH2 in sporadic colorectal cancer, Oncogene 12 (12) (1996) 2641±2649. [11] J.R. Jass, D.S. Cottier, P. Jeevaratnam, V. Pokos, K.M. Holdaway, M.L. Bowden, N.S. Van de Water, P.J. Browett, Diagnostic use of microsatellite instability in hereditary nonpolyposis colorectal cancer, Lancet 346 (1995) 1200± 1201. [12] C.R. Boland, S.N. Thibodeau, S.R. Hamilton, D. Sidransky, J.R. Eshleman, R.W. Burt, S.J. Meltzer, M.A. RodriguezBigas, R. Fodde, G.N. Ranzani, S. Srivastava, National cancer institute workshop on microsatellite instability for cancer detection and familial predisposition: development of international criteria for determination of microsatellite instability in colorectal cancer, Cancer Res. 58 (22) (1988) 5248±5257. [13] Correpsondance re: C.R. Boland, S.N. Thibodeau, S.R. Hamilton, D. Sidransky, J.R. Eshleman, R.W. Burt, S.J. Meltzer, M.A. Rodriguez, R. Fodde, G.N. Ranzani, S. Srivastava.. A national cancer institute workshop on microsatellite instability for cancer detection and familial predisposition: development of international criteria for determination of microsatellite instability in colorectal cancer. Cancer Res. 58 (1988) 5248-5257. Cancer Res. 59 (1999) 249-256. [14] W. Dietmaier, S. Wallinger, T. Bocker, F. Kullmann, R.

[15]

[16]

[17]

[18]

[19]

[20]

[21]

[22]

Fishel, J. Ruschoff, Diagnostic microsatellite instability: de®nition and correlation with mismatch repair protein expression, Cancer Res. 57 (21) (1997) 4749±4756. T. Bocker, J. Diermann, W. Friedl, J. Gebert, E. HolinskiFeder, J. Karner-Hanusch, M. von Knebel-Doeberitz, K. Koelble, G. Moeslein, H.K. Schackert, H.C. Wirtz, R. Fishel, J. Ruschcoff, Microsatellite instability analysis: a multicenter study for reliability and quality control, Cancer Res. 57 (21) (1997) 4739±4743. J.M. Hoang, P.H. Cottu, B. Thuille, R.J. Salmon, G. Thomas, R. Hamelin, Bat 26 is an indicator of the replication error phenotype in colorectal cancers and cell lines, Cancer Res. 57 (2) (1997) 300±303. F. Canzian, R. Salovaara, A. Hemminki, P. Kristo, R.B. Chadwick, L.A. Aaltonen, A. de la Chapelle, Semiautomated assessment of loss of heterozygosity and replication error in tumours, Cancer Res. 56 (14) (1996) 3331±3337. G. Moslein, D.J. Tester, N.M. Lindor, R. Honchel, J.M. Cunningham, A.J. French, K.C. Halling, M. Schwab, P. Goretzki, S.N. Thibodeau, Microsatellite instability and mutation analysis of hMSH2 and hMLH1 in patients with sporadic, familial and hereditary colorectal cancer, Hum. Mol. Genet. 5 (9) (1996) 1245±1252. M.G. Dunlop, S.M. Farrington, A.D. Carothers, A.H. Wyllie, L. Sharp, J. Burn, B. Liu, K.W. Kinzler, B. Vogelstein, Cancer risk associated with germline DNA mismatch repair gene mutations, Hum. Mol. Genet. 6 (1) (1997) 105±110. X.P. Zhou, J.M. Hoang, Y.J. Li, R. Seruca, F. Carneiro, M. Sobrinho-Simoes, R.A. Lothc, C.M. Gleeson, S.E. Russell, F. Muzeau, J.F. Flejou, K. Hoang-Xuan, R. Lidereau, G. Thomas, R. Hamelin, Determination of the replication error phenotype in human tumours without requirement for matching normal DNA by analysis of mononucleotide repeat microsatellites, Genes Chromosomes Cancer 21 (2) (1998) 101± 107. R. Parsons, L. Myeroff, B. Liu, J.K. Wilson, S.D. Markowitz, K.W. Kinzler, B. Vogelstein, Microsatellite instability and mutations of the transforming growth factor b type II receptor gene in colorectal cancer, Cancer Res. 55 (23) (1995) 5548± 5550. I.M. Frayling, Microsatellite instability, Gut 45 (1999) 1±4.