Molecular classification of colorectal cancer · 2017-03-17 · predisposing to colorectal cancer largely remain to be characterized. Though the main relevance of hereditary cancer

MOLECULAR CLASSIFICATION OF COLORECTAL CANCER

Päivi Laiho

Department of Medical Genetics Molecular and Cancer Biology Programme

Haartman Institute and Biomedicum Helsinki University of Helsinki

Finland

Academic dissertation

To be publicly discussed, with the permission of the Medical Faculty of the University of Helsinki, in the lecture hall 3, Biomedicum Helsinki, Haartmaninkatu 8, on 30 September

2005, at 12 noon.

Helsinki 2005

2

Supervised by Lauri A Aaltonen, MD, PhD Research Professor of the Finnish Academy of Sciences Department of Medical Genetics Haartman Institute and Biomedicum Helsinki University of Helsinki Finland Virpi Launonen, PhD Docent Department of Medical Genetics Haartman Institute and Biomedicum Helsinki University of Helsinki Finland Reviewed by Jorma Isola, MD, PhD Professor of Cancer Biology Institute of Medical Technology University of Tampere Finland Robert WM Hofstra, PhD University Medical Center Groningen University of Groningen The Netherlands Official opponent Ritva Karhu, PhD Docent Institute of Medical Technology University of Tampere Finland ISBN 952-91-8989-3 (paperback) ISBN 952-10-2584-0 (PDF) http://ethesis.helsinki.fi Yliopistopaino Helsinki 2005

3

TABLE OF CONTENTS

ABBREVIATIONS........................................................................................................... 5

LIST OF ORIGINAL PUBLICATIONS........................................................................ 8

ABSTRACT....................................................................................................................... 9

REVIEW OF THE LITERATURE .............................................................................. 11

1. GENES AND CANCER...........................................................................................................................11 1.1 Oncogenes............................................................................................................................................12 1.2 Tumor suppressor genes ......................................................................................................................13

1.2.1 Gatekeepers.................................................................................................................................13 1.2.2 Caretakers....................................................................................................................................13 1.2.3 Landscapers..................................................................................................................................13

1.3 Epigenetic mechanisms .......................................................................................................................14 1.3.1 Methylation ..................................................................................................................................14 1.3.2 Loss of imprinting (LOI) .............................................................................................................14

2. COLORECTAL CANCER......................................................................................................................15 2.1 Non-polyposis syndromes ...................................................................................................................15 2.2 Polyposis syndromes ...........................................................................................................................17

2.2.1 Adenomatous polyposis syndromes.............................................................................................17 2.2.2 Hamartomatous polyposis syndromes .........................................................................................18

2.3 Low-penetrance alleles predisposing to colorectal cancer ..................................................................19 3. PATHWAYS LEADING TO COLORECTAL CANCER...................................................................19

3.1 Suppressor pathway (The adenoma-carcinoma sequence)..................................................................19 3.2 Mutator pathway ..................................................................................................................................22 3.3 Serrated neoplasia pathway .................................................................................................................23

4. MICROSATELLITE INSTABILITY....................................................................................................24 4.1 Mismatch repair system.......................................................................................................................26

4.1.1 Defects in mismatch repair system ..............................................................................................28

AIMS OF THE STUDY ................................................................................................. 30

MATERIALS AND METHODS ................................................................................... 31

1. PATIENT SAMPLES (I, II, III, IV) .......................................................................................................31 1.1 DNA (I, II, III, IV) and RNA (IV) extraction......................................................................................31

2. MSI ANALYSIS (I, II) .............................................................................................................................31 3. IMMUNOHISTOCHEMISTRY AND MUTATION ANALYSIS OF MLH1, MSH2 AND MSH6 (I)........................................................................................................................................................................32 4. LOCUS SPECIFIC LOH ANALYSIS (II).............................................................................................32 5. MUTATION ANALYSIS OF KRAS (II)................................................................................................34 6. MGMT, MLH1, AND EPHB2 PROMOTER HYPERMETHYLATION ANALYSIS (II, IV).........34

6.1 Bisulfite modification ..........................................................................................................................34 6.2 Methylation specific PCR....................................................................................................................34

7. TOPOGRAPHIC MUTATION HETEROGENEITY ANALYSIS (II)..............................................35 8. ALLELIC IMBALANCE (AI) ANALYSIS (III) ..................................................................................35 9. COMPARATIVE GENOMIC HYBRIDIZATION (CGH) (III).........................................................36 10. REAL–TIME QUANTITATIVE PCR.................................................................................................36 11. STATISTICAL ANALYSES (II, III) ...................................................................................................36

4

12. ARRAY ANALYSIS (IV) ......................................................................................................................36 13. ARRAY DATA ANALYSIS (IV) ..........................................................................................................37

13.1 Normalization and filtering................................................................................................................37 13.2 Unsupervised hierarchical clustering and identification of differentially expressed genes ..............37 13.3 Class prediction..................................................................................................................................37

14. IMMUNOHISTOCHEMISTRY (IV)...................................................................................................38 15. MUTATION SCREENING OF EPHRIN TYPE-B RECEPTOR 2 (EPHB2) (IV) ...........................38

15.1 Direct sequencing ..............................................................................................................................38

RESULTS ........................................................................................................................ 40

1. COMPARISON BETWEEN BAT26 AND THE BETHESDA PANEL (I) ........................................40 1.1 MSI analysis using the Bethesda panel ...............................................................................................40

2. ANALYSIS OF LOW-LEVEL MSI (II) ................................................................................................40 2.1 MSI analysis ........................................................................................................................................40 2.2 Comparison between putative MSI-L and MSS groups......................................................................41

3. ALLELIC IMBALANCE IN FAMILIAL AND SPORADIC TUMORS (III)...................................42 3.1 AI analysis ...........................................................................................................................................42 3.2 Comparative genomic hybridization and DNA copy number measurement.......................................42

4. GENE EXPRESSION PROFILE OF SERRATED CRCs (IV)...........................................................43 4.1 Immunohistochemistry ........................................................................................................................44 4.2 EPHB2 mutation screening and promoter hypermethylation analysis................................................44

DISCUSSION .................................................................................................................. 45

1. MSI AS A DIAGNOSTIC TOOL (I) ......................................................................................................45 2. THE ROLE OF MSI-L IN COLORECTAL TUMORS (II)................................................................46 3. CHARACTERIZATION OF CHROMOSOMAL REGIONS CONTAINING PUTATIVE LOW-RISK ALLELES FOR CRC (III) ...............................................................................................................49 4. THE GENE EXPRESSION PROFILE OF SERRATED COLORECTAL TUMORS (IV) ............51

SUMMARY AND CONCLUSIONS ............................................................................. 53

ACKNOWLEDGEMENTS ........................................................................................... 54

REFERENCES................................................................................................................ 56

5

ABBREVIATIONS A adenine ACF aberrant crypt foci AFAP attenuated familial adenomatous polyposis AI allelic imbalance AMPK AMP-activated protein kinase APC adenoamtous polyposis coli AURKA aurora kinase A bp base pair BAX BCL2-associated X-protein BCL2/Bcl-2 B-cell lymphoma 2 gene/protein BLM bloom syndrome gene BMPR1A bone morphogenetic protein receptor-1A BRAF v-raf murine sarcoma viral oncogene homolog B1 BRCA1/2 breast and ovarian cancer gene-1/2 BRG1(SMARCA4) SWI/SNF related, matrix associated, actin dependent regulator of

chromatin, subfamily a, member 4 C cytosine CCND1 cyclin D1 CCNT2 cyclin T2 CD95(Fas) tumor necrosis factor receptor superfamily, member 6 CDK9 cyclin-dependent kinase-9 CDKN2B cyclin-dependent kinase inhibitor 2B cDNA complementary deoxyribonucleic acid CGH comparative genomic hybridization CHRPE congenital hypertrophy of the retinal pigment CIN chromosomal instability cM centiMorgan C-MYC myelocytomatosis viral oncogene homolog CRC colorectal cancer cRNA comprementary ribonucleic acid CTNNB1 β-catenin DCC deleted in colorectal cancer dCTP deoxycytocine triphosphate DNA deoxyribonucleic acid DPC4 deleted in pancreatic cancer-4 dUTP deoxyuracil triphosphate EGFR epidermal growth factor receptor EPHB2/EphB2 ephrin type-B receptor-2 gene/protein EST expressed sequence tag EXO1 exonuclease-1 FAP familial adenomatosis polyposis G guanine GDP guanosine 5’diphospate GSK-3β glycogen synthase kinase-3β

6

GTP guanosin 5’triphosphate HFE hemochromatosis HH hedgehog signaling HIF1α hypoxia-inducible factor α HNPCC hereditary non-poyposis colorectal cancer H-RAS Harvey rat sarcoma viral oncogene homolog IDL insertion-deletion loop IGFII insulin-like growth factor-2 IGFRII insulin-like growth factor receptor-2 JP juvenile polyposis KRAS Kirsten rat sarcoma viral oncogene homolog LKB1 serine/threonine kinase defective in PJS LOH loss of heterozygosity LOI loss of imprinting MAP mitogen activated protein MBD4 methyl-CpG binding domain protein 4 MGMT o6-methylguanine-DNA methyltransferase MLH1/3 human mutL homolog-1/3 MMR mismatch repair MSH2-5 human mutS homolog-2-5 MSI microsatellite instability MSI-H high microsatellite instability MSI-L low microsatellite instability MSS microsatellite stable MTA1 metastasis associated-1 MTS Muir-Torre syndrome MYH human mutY homolog p short arm of a chromosome q long arm of a chromosome SMAD2/4 human homologs of Drosophila melanogaster Mad gene-2/4 smo smoothened PCR polymerase chain reaction PJS Peutz-Jeghers syndrome PMS1/2 human postmeiotic segregation increased-1/2 PTCH/ptc patched gene/protein RB retinoblastoma gene RNA ribonucleic acid SNP single nucleotide polymorphism SSCP single strand conformation polymorphism STK11 serine/threonine kinase-11 STK15 serine/threonine kinase-15 T thymidine TCF T-cell factor TGFβ transforming growth factor-β TGFβRI/II transforming growth factor-β receptor type I/II TP53(p53) tumor protein 53

7

TS Turcot syndrome UV ultraviolet VEGF vascular endothelial growth factor VHL Von Hippel-Lindau syndrome gene wnt wingless signaling pathway WT Wilm’s tumor

8

LIST OF ORIGINAL PUBLICATIONS This thesis is based on the following original articles. In the text they will be referred to by their Roman numericals. I Loukola A*, Eklin K*, Laiho P*, Salovaara R, Kristo P, Järvinen H, Mecklin JP, Launonen V, and Aaltonen LA. Microsatellite marker analysis in screening for hereditary nonpolyposis colorectal cancer (HNPCC). Cancer Research 61:4545-4549, 2001 IIa Laiho P, Launonen V, Lahermo P, Esteller M, Guo M, Herman JG, Mecklin J-P, Järvinen H, Sistonen P, Kim K-M, Shibata D, Houlston RS, and Aaltonen LA. Low-level microsatellite instability in most colorectal carcinomas. Cancer Research 62: 1166-1170, 2002 IIb Laiho P, Aaltonen LA. Correspondence re: P. Laiho et al. Low-level microsatellite instability in most colorectal carcinomas. Cancer res., 62:1166-1170, 2002. Cancer Research 62: 5988-5990, 2002 III Laiho P*, Hienonen T*, Karhu A, Lipton L, Aalto Y, Thomas H, Birkenkamp-Demtroder K, Hodgson S, Salovaara R, Mecklin J-P, Järvinen H, Knuutila S, Halford S, Orntoft TF, Tomlinson I, Launonen V, Houlston R, and Aaltonen LA. Genome-wide allelotyping of 104 Finnish colorectal cancers reveals an excess of allelic imbalance in chromosome 20q in familial cases. Oncogene 22: 2206-2214, 2003 IV Laiho P*, Kokko A*, Vanharanta S, Salovaara R, Sammalkorpi H, Järvinen H, Mecklin J-P, Karttunen TJ, Tuppurainen K, Arango D, Mäkinen MJ, and Aaltonen LA. Serrated carcinomas form a subclass of colorectal cancer with distinct molecular basis. Submitted * Equal contribution

9

ABSTRACT Colorectal cancer (CRC) is the third most common cancer in the Western countries. Some genetic alterations behind the colorectal tumor progression are well known, such as loss of chromosomal regions 5q, 17p and 18q. Genomic instability plays a key role in cancer development and multiple mutations in several genes eventually lead to invasive cancer. Despite that our knowledge of cancer development has increased during the past decades, many of the genetic alterations underlying cancer remains to be clarified. Which of these alterations are the key players in the initial development of colorectal cancer is not entirely known. Much of our information about the molecular pathogenesis of cancer has arisen from studies on hereditary tumors. Lynch syndrome is a hereditary colorectal cancer syndrome, which accounts for 1-5% of all CRCs. The underlying gene defects causing Lynch syndrome are well characterized. The moderate and low penetrance genes predisposing to colorectal cancer largely remain to be characterized. Though the main relevance of hereditary cancer studies relates to increasing understanding on the mechanisms of cancer development, such studies also benefit the cancer families. Knowledge of the underlying molecular features can lead to more accurate diagnosis, better treatment and prevention of cancer. Approximately 15% of CRCs display a phenomenon called microsatellite instability (MSI). MSI tumors show several clinical and molecular features differing from microsatellite stable (MSS) tumors. This classification can be utilized in CRC diagnostics. A panel of five microsatellite markers known as the Bethesda panel has been proposed for screening for MSI. To test a hypothesis that the use of the mononucleotide marker BAT26 alone is feasible in screening for MLH1/MSH2 mutation positive cases, 494 colorectal cancer patients were studied. BAT26 status was compared to results obtained using the Bethesda panel. BAT26 was able to detect all mutation positive individuals in this series. To study the proportion and significance of tumors belonging to a third subgroup, MSI-low (MSI-L), 90 BAT26 stable CRC samples were analyzed with the five Bethesda markers and 372 additional microsatellite markers. Several molecular and clinical features were scrutinized, to examine the previously proposed differences between MSI-L and MSS tumors. Convincing differences between putative MSI-L and MSS groups were not observed and the results suggest that MSS and MSI-L tumors have a common molecular background. To identify new chromosomal regions for low and moderate penetrance genes predisposing to CRC a genome-wide allelotyping was performed on 29 familial and 75 sporadic CRCs. The study revealed differences in AI pattern between sporadic and familial cases, and several chromosomal loci that displayed more AI in familial tumors were identified. The most promising region was located in chromosome 20q which was amplified more often in familial tumors compared to sporadic ones. Chromosome 20q may harbor a novel CRC predisposing gene.

10

Serrated colorectal tumors are a fairly recently characterized entity among CRCs. They are morphologically different from conventional adenocarcinomas but whether they are biologically different as well has not been clear. To study the molecular background of serrated CRCs the gene expression profile of 8 serrated CRCs was compared to 29 conventional adenocarcinomas using microarrays. Several differentially expressed genes were identified and the expression differences were validated in a larger dataset using immunohistochemistry. The results establish serrated CRCs as a distinct entity with unique molecular basis.

11

REVIEW OF THE LITERATURE

1. GENES AND CANCER Cancer is a serious health problem in the Western countries. In Finland over 20 000 people develop cancer every year and almost 10 000 people die of it. Since the first cancer statistics from 1950 the cancer incidence has continuously increased. The most common cancer types today are prostate-, breast-, and colorectal cancer which account for 43% of all cancers (Table 1) (Finnish Cancer Registry. Cancer Statistics at www.cancerregistry.fi last updated on 30 Aug 2004). Table 1. The most common cancer types in Finland in 2002 Cancer type Number of cases (% of all cancers) Prostate 3930 (17%) Breast 3791 (16%) Colorectal 2236 (10%) Lung 1861 (8%) Skin 1518 (7%)

Several environmental and lifestyle factors increase cancer risk. Exposure to UV-radiation predisposes to skin cancer, and cigarette smoking is a known risk factor for lung cancer. Environmental factors do not, however, directly cause cancer since cancer is essentially a genetic disease. Malignant tumors arise when cells have acquired several mutations in DNA and are no longer able to control their own growth. Hanahan and Weinberg (2000) have proposed that six essential alterations in cell physiology are responsible for malignant growth (Figure 1). Tumor cells lose their capability to control cell cycle, differentiation, and programmed cell death. They are able to induce the formation of blood vessels and eventually invade to surrounding tissues and metastasize to other organs. Majority of cancers arise sporadically. In sporadic cancer mutations occur by chance in a single cell giving it a growth advantage. Cancer can, also, run in a family. In hereditary cancer a defective gene is present in every cell of its carrier. The age of onset in hereditary cancer is usually lower than in sporadic cancer and these families are characterized by accumulation of cancer in the family. Whether cancer is sporadic or hereditary, progression of a normal tissue to a cancer is characterized by accumulation of genetic alterations in the target tissue. The genes that underlie cancer are called oncogenes and tumor suppressor genes.

12

Figure 1. Alterations in cell physiology in cancer (modified from Hanahan & Weinberg, 2000)

1.1 Oncogenes Oncogenes are altered forms of normal cellular components called proto-oncogenes. In normal cells proto-oncogene expression results in increased cell birth or decreased cell death. Oncogenes are dominant at the cellular level. An inactivating mutation in either of the two alleles is sufficient to encourage increased proliferation. Oncogenes can be classified into different subgroups according to their function in cells. They can be e.g. growth factors, growth factor receptors, intracellular signaling proteins, cell cycle regulators and transcription factors. Proto-oncogenes become oncogenes through gain-of-function mutations such as an activating point mutation, amplification or chromosomal translocation (Bishop, 1991). The first report of mutated proto-oncogene was presented by Reddy et al. in 1982. They observed a point mutation in HRAS codon 12 that converted glycine to valine in a bladder carcinoma cell line. Since then mutations have been found in over a hundred other proto-oncogenes (Futreal et al., 2004).

Cancer

Self sufficiency in signals stimulating growth

Insensitivity to negative growth signals

Avoidance of apoptosis

Capasity for sustained proliferation

Induced angiogenesis

Tissue invasion and metastasis

13

1.2 Tumor suppressor genes Tumor suppressor genes inhibit cell growth. Tumor suppressor genes act in a recessive manner, meaning that both alleles must be inactive before the effect of the gene product is lost. This was first suggested in 1971 when Knudson presented his famous two-hit hypothesis. He proposed that two genetic hits were sufficient to lead to retinoblastoma. Later, Cavenee et al. (1983) showed that when the first hit in retinoblastoma was a small change, the second one was usually a gross chromosomal alteration, such as a large deletion, resulting in loss of heterozygosity (LOH) in retinoblastoma (RB) locus. Most tumor suppressor genes act like RB. In hereditary cancer the first hit is the inherited mutation, while in sporadic cancer it is the acquired one. A somatic mutation, LOH or an epigenetic mechanism, such as promoter hypermethylation, then inactivates the other allele.

1.2.1 Gatekeepers

Gatekeepers are tumor suppressor genes that directly regulate tumor growth by inhibiting growth or by promoting death. The role of a gatekeeper gene varies in different tissues. Loss of function of a certain gene leads to a specific form of cancer so that inactivation of APC leads to colon cancer but not renal cancer, whereas mutations in VHL cause renal cancer but not colon cancer (Latif et al., 1993; Kinzler and Vogelstein, 1997).

1.2.2 Caretakers

Unlike gatekeepers, inactivation of a caretaker gene does not directly promote tumor growth. Inactivation of a caretaker leads to genetic instability that indirectly promotes tumor formation by causing an increased mutation rate in all genes in a cell, including oncogenes and tumor suppressor genes. Since several mutations are required for the full development of cancer, the inactivation of a caretaker can greatly increase the mutation rate in cells and thereby accelerate the tumor formation (Kinzler and Vogelstein, 1997). Similar to gatekeepers inactivation of a caretaker leads to tissue a specific tumor. For example mutations in DNA repair genes MLH1 and MSH2 underlie colorectal cancer, while BRCA1 and BRCA2 are mutated in breast cancer.

1.2.3 Landscapers

A third group of tumor suppressor genes, landscapers, are proposed to play a role in neoplastic transformation indirectly by creating an abnormal microenvironment. Patients affected with juvenile polyposis and ulcerative colitis develop hamartomatous polyps in which the proliferating cells are mostly stromal. The epithelial cells associated with abnormal stroma are more likely to become neoplastic due to an abnormal microenvironment (Kinzler and Vogelstein, 1998).

14

1.3 Epigenetic mechanisms Besides mutations, oncogenes and tumor suppressor genes can be activated/inactivated through epigenetic modifications. Epigenetic modifications are alterations in the genome that do not involve the DNA sequence itself (Verma and Srivastava, 2002). Epigenetic modifications are dynamic. In somatic cells they are passed on to the progeny of a cell through cell division, in germline, however, epigenetic modifications are reversible. The first indication that epigenetics play a role in cancer was the discovery of altered methylation in colorectal tumors (Feinberg and Vogelstein; 1983). In addition to aberrant DNA methylation, another epigenetic change associated with cancer is loss of imprinting (LOI).

1.3.1 Methylation

Methylation takes place at cytocine bases that are located 5' to a guanosine in a CpG dinucleotide. These dinucleotides are enriched in promoter regions of genes, and are called CpG islands (Larsen et al., 1992). In normal cells CpG islands are generally unmethylated. However, in cancer the hypermethylation of these promoter regions is a common epigenetic change, and is associated with the transcriptional silencing of tumor suppressor genes (Baylin and Herman, 2000). Genes that cause familial forms of cancer when mutated in the germ line are known to undergo methylation-associated silencing in sporadic cancer. Mismatch-repair gene MLH1 underlies hereditary colorectal cancer, and is methylated in sporadic microsatellite instable (MSI) colorectal cancers (Kane et al., 1997; Herman et al., 1998). Additionally, there are a growing number of tumor suppressor genes that are silenced by promoter hypermethylation but seem not to be frequently mutated (Esteller et al. 2000).

1.3.2 Loss of imprinting (LOI)

Genetic imprinting is a form of epigenetic inheritance that distinguishes maternal and paternal alleles. Usually genes are expressed from both chromosomes, but imprinted genes show preferential expression of either a maternal or a paternal allele. Usually imprinting is mediated through a specific methylation pattern of an imprinted allele (Feinberg, 2002). Loss of imprinting leads to loss of the normal pattern of parental origin-specific gene expression. In cancer it was first detected in an embryonal kidney cancer Wilm's tumor (WT). In normal kidney cells a growth-promoting gene insulin-like growth factor II (IGFII) is imprinted so that it is expressed preferentially from the paternal allele. LOI causes biallelic expression of IGFII increasing its concentration in a cell, and thus giving a growth advantage (Ogawa et al., 1993; Rainier et al., 1993). In addition to WT, several adult tumors including cervical, colorectal and liver tumors display LOI as well (Douc-Rasy et al., 1996; Takeda et al., 1996; Cui et al., 2003).

15

2. COLORECTAL CANCER Colorectal cancer (CRC) is the third most common cancer in Finland (Finnish Cancer Registry. Cancer Statistics at www.cancerregistry.fi last updated on 30 Aug 2004). It is common also in other Western world countries, as well as in some developing countries. Differences between races and ethnic groups, however, exist (Boyle and Leon, 2002). Several environmental factors may increase CRC risk. High body-mass index, increased meat consumption, and tobacco smoking are associated with higher CRC risk (Martinez et al., 1997; Willett et al., 1990; Giovannucci et al., 1996; Giovannucci 2001). A family history of CRC represents an independent risk factor. CRC in a family confers a 2-fold to 6-fold increase in risk. The risk associated with family history varies greatly, and depends on the age of onset of CRC in family members, the number of affected relatives, and whether cancers have occurred in multiple generations (Goldgar et al., 1994; Dong and Hemminki, 2001; Risch 2001; Slattery et al., 2003). The proportion of CRCs considered to be familial depends on the definition. Approximately 5% of CRCs are caused by a highly penetrant inherited mutation in a cancer predisposing gene. If the low-penetrant susceptibility alleles are taken into account the proportion of familial CRC rises up to 20-35%. The rest of CRCs are considered sporadic (Lichtenstein et al., 2000; de la Chapelle, 2004). Hereditary CRC syndromes are divided into two groups: polyposis and non-polyposis syndromes. Polyposis syndromes are further divided into adenomatous and hamartomatous polyposis syndromes.

2.1 Non-polyposis syndromes Lynch syndrome (Hereditary non-polyposis colorectal cancer) Lynch syndrome, also known as hereditary non-polyposis colorectal cancer (HNPCC), is an autosomal dominant condition characterized by familial aggregation of early-onset, right-sided, synchronous and metachronous CRC. The incidence of Lynch syndrome is estimated to be 2-3% (Aaltonen et al., 1998; Salovaara et al., 2000). People affected with Lynch syndrome are in an increased risk for several other cancers besides CRC, such as endometrial, ovarian, small bowel, stomach, urinary tract, renal pelvis and brain (Watson and Lynch, 1993; Vasen et al., 1999). The lifetime risk of CRC of a person with Lynch syndrome is as great as 70-85%. Women have approximately 50% risk of endometrial cancer. The risk of other malignancies is below 20% (Watson and Lynch, 1993; Aarnio et al., 1995). Diagnosis of Lynch syndrome is based on the Amsterdam criteria. The Amsterdam criteria were created in 1991 to establish diagnostic guidelines for this syndrome, when the genetic basis was not yet clear. The criteria have since been revised to include extracolonic malignancies (Vasen et al., 1991; Vasen et al., 1999). The Amsterdam criteria are listed in Table 2.

16

Table 2. Amsterdam criteria for Lynch syndrome diagnosis • At least three affected relatives with Lynch syndrome associated cancer • One affected person is a first-degree relative of the other two affected persons • At least two successive generations should be affected • At least one cancer should be diagnosed before age 50 • Familial adenomatous polyposis should be excluded • Pathological verification of tumors The genes underlying Lynch syndrome are DNA mismatch repair (MMR) genes. Four genes, MLH1, MSH2, MSH6, and PMS2 have been identified that, when mutated in germline cause susceptibility to Lynch syndrome (Fishel et al., 1993; Leach et al., 1993; Bronner et al., 1994; Nicolaides et al., 1994; Papadopoulos et al., 1994; Miyaki et al., 1997). The function of MMR genes in a cell is to correct errors, such as base-base mismatches and insertion-deletion loops (IDLs) that occur during DNA replication. Base-base mismatches lead to single base substitutions, whereas IDLs affect microsatellites, leading to insertion or deletion of repetitive units. This phenomenon is known as microsatellite instability (MSI). Approximately 90% of CRCs associated with Lynch syndrome display MSI (Aaltonen et al., 1994; Pedroni et al., 1999). Most Lynch syndrome families have a mutation in MLH1 and MSH2, while MSH6 and PMS2 are less frequently involved (Peltomäki, 2003). MLH1 and MSH2 mutations often generate typical Lynch syndrome families which fulfil the Amsterdam criteria, and display high degree of MSI (Wijnen et al., 1997). Less typical Lynch syndrome families often display mutations in MSH6. In these families the age of onset is higher, and penetrance is lower (Miyaki et al., 1997). It has been proposed that mutations in MMR genes MLH3 and EXO1 may also underlie Lynch syndrome. However, the available data are limited and controversial, thus making a reliable assessment of their role in Lynch syndrome predisposition is difficult. (Wu et al., 2001; 2001b; Hienonen et al., 2003; Jagmohan-Changur et al., 2003). MSI and MMR genes are discussed later in more detail. Muir-Torre syndrome Muir-Torre syndrome (MTS) is a variant of Lynch syndrome. In addition to Lynch syndrome associated tumors, it includes sebaceous skin tumors (Hall et a., 1994; Schwartz and Torre, 1995). About half of MTS patients are affected with CRC. Both MLH1 and MSH2 have been observed to harbour mutations in MTS tumors, the predominant gene, however, being MSH2 (Kruse et al., 1998). Turcot syndrome Turcot syndrome (TS) could be placed under polyposis or non-polyposis syndromes depending on the underlying gene. It is a rare disorder which includes colorectal polyposis, CRC, and brain tumors; either medulloblastoma or glioblastoma. Families with colon polyposis and medulloblastoma are likely to have a mutation in APC gene, whereas families with CRC and glioblastoma harbour mutations in MLH1 or PMS2 genes

17

(Hamilton et al., 1995; Mori et al., 1994). PMS2 is seldom mutated in classical Lynch syndrome families, and it seems that it is primarily associated with TS (Hamilton et al., 1995; Liu et al., 2001).

2.2 Polyposis syndromes

2.2.1 Adenomatous polyposis syndromes

Familial adenomatous polyposis and Attenuated familial adenomatous polyposis Familial adenomatous polyposis (FAP) is a well-described autosomal dominant condition in which affected individuals develop hundreds to thousands of adenomatous polyps in the colon and rectum typically after the first decade of life. If untreated, approximately 90% of individuals with FAP develop CRC by the time they reach their 40s (Talbot et al., 2000). FAP patients are also at risk for several extracolonic malignancies such as cancers of the thyroid, small intestine, stomach, liver and brain. Benign extracolonic features associated with FAP include congenital hypertrophy of the retinal pigment epithelium (CHRPE), dental abnormalities, jaw cysts, and osteomas of the skull, mandible, and long bones (Lal and Gallinger, 2000). The term Gardner syndrome has been used to describe FAP patients who in addition to colorectal polyposis display osteomas, epidermoid cysts, and skin fibromas (Gardner and Richards, 1953). Now it's known that features previously associated with Gardner syndrome are, with variable expression, observed in FAP patients, thus Gardner syndrome should not be considered a genetic variant of FAP. Most individuals affected with FAP have a mutation in tumor suppressor gene APC (Groden et al., 1991; Nishisho et al., 1991). In normal cells APC together with other proteins forms a complex which binds to β-catenin and directs it to proteosomal degradation. When APC is mutated, β-catenin accumulates in a cell and activates transcription of several target genes which promotes tumor formation (Giles et al., 2003). The clinical phenotype of FAP patients appears to correlate with the type and location of mutation. Classic FAP with the occurrence of thousands of polyps is observed in patients who have mutations in APC between codons 169 and 1600. CHRPE is mostly seen in individuals with mutations between codons 463-1387 (de la Chapelle, 2004). Attenuated familial adenomatous polyposis (AFAP) is a milder variant of FAP and is characterized by a far fewer polyps (<100) and an age of onset 10 to 15 years later than in classical FAP patients. However, the lifetime risk for CRC is still very high (Leppert et al., 1990; Spirio et al., 1992). Mutations causing AFAP are mainly located in the extreme 5', and 3' ends of APC. MYH-associated polyposis MYH-associated polyposis is a recently reported autosomal recessively inherited condition, which is caused by mutations in DNA repair gene MYH (Al-Tassan et al.,

18

2002). Biallelic inactivation, either a homozygous mutation or a compound heterozygote, of MYH can lead to a disease that resembles classic FAP, AFAP or Lynch syndrome (Enholm et al., 2003; Sieber at al., 2003). MYH is a component of the base-excision-repair system. MYH deficient tumors display an excess of somatic G:C to T:A transversions in APC gene, and this is thought to cause the phenotype of adenomatous polyps followed by cancer (Al-Tassan et al., 2002).

2.2.2 Hamartomatous polyposis syndromes

Peutz-Jeghers syndrome The main feature of Peutz-Jeghers syndrome (PJS) is hamartomatous polyps in the gastrointestinal tract. Polyps are mainly located in the small intestine, but can be found in colon as well. PJS polyps show a characteristic tree-like structure, which arises when the muscular mucosa extends into the polyp. Besides polyps, PJS patients display mucocutaneous pigmentation of lips and oral area (Jeghers et al., 1949; Spigelman et al., 1995). PJS patients are at an increased risk for colon cancer, as well as several other malignancies, such as breast, pancreatic, stomach, and ovarian cancer (Giardiello et al., 2000). The gene predisposing to PJS is LKB1 (STK11), which encodes a serine-threonine kinase and appears to function as a tumor suppressor (Hemminki et al., 1998). LKB1 is involved in BRG1-dependent chromatin remodeling and in p53 dependent apoptosis pathway and VEGF signaling. It also contributes to cell stress sensoring through activation of AMPK kinases (Karuman et al., 2001; Marignani et al., 2001; Ylikorkala et al., 2001; Hawley et al., 2003). Juvenile polyposis Juvenile polyposis (JP) is a child-onset syndrome which is characterized by hamartomatous polyps throughout the gastrointestinal tract. Usually juvenile polyps are benign, but malignant transformation can occur. Individuals affected with JP have an elevated risk mainly for CRC, but also stomach, duodenal, and pancreatic cancers have been associated with JP (Järvinen and Franssila, 1984; Järvinen et al., 1993). A variety of malformations have been observed in JP patients including congenital heart disease, cleft lip/palate, and mental retardation (Järvinen et al., 1993). To date, germline mutations in two genes, SMAD4 (DPC4) and BMPR1A, have been detected in JP families. Both are components of the transforming-growth factor-β (TGF-β) superfamily (Howe et al., 1998; Howe et al., 2001).

19

2.3 Low-penetrance alleles predisposing to colorectal cancer

APC I1307K Six to seven % of Ashkenazi Jewish individuals carry APC I1307K allele. This variant is associated with a 10-20% lifetime risk of CRC (Laken et al., 1997; Gryfe et al., 1999). I1307K variant involves a T->A transversion which creates a stretch of eight adenines. This may lead to errors in the replication process, causing predisposition to single-nucleotide insertions or deletions in the (A)8 stretch leading to frameshift mutations. Tumor progression is likely initiated when the wild-type allele is somatically mutated in APC I1307K carriers (Laken et al., 1997; Zauber et al., 2003). TGFβRI*6A The tumor suppressor Transforming growth factor beta (TGFβ) is an efficient inhibitor of cell growth. It mediates the growth inhibitory signal through two receptors TGFβRI and TGFβRII. TGFβRI has a stretch of 9 alanines within a coding sequence of exon 1. Approximately 86% of Caucasian populations are 9A/9A homozygotes, while 14% have an allele that is 3 alanines shorter (genotype 9A/6A or 6A/6A) (Pasche et al., 2004). Several studies have shown that individuals with at least one 6A allele have a higher risk for several cancer types, including CRC (Chen et al., 1999; Pasche et al., 1999; Baxter et al., 2002; Pasche et al., 2004). The mechanism by which TGFβ 6A causes a cancer predispostion is not known. In a cell line experiments, however, 6A allele have been shown to be impaired as a mediator of TGFβ growth inhibitory signals (Chen et al., 1999; Pasche et al., 1999).

3. PATHWAYS LEADING TO COLORECTAL CANCER

3.1 The suppressor pathway (The adenoma-carcinoma sequence) The progression of colorectal carcinoma is a multistep process and may take years or even decades (Muto et al., 1975). The exceptional feature of CRC is that it is fairly easy to collect lesions which precede carcinoma. This has enabled the molecular characterization of CRCs at different stages and led to the development of a stepwise model of colorectal tumorigenesis referred to as the adenoma-carcinoma sequence (Figure 2) (Fearon and Vogelstein, 1990; Kinzler and Vogelstein, 1996).

20

Figure 2. The adenoma-carcinoma sequence. Mutations in the APC gene initiate the neoplastic process, and tumor progression results from mutations in the other genes, such as KRAS, TP53, and SMAD4 (Modified from Kinzler and Vogelstein 1996). The adenoma-carcinoma sequence has also been called the suppressor pathway, since the tumor initiation and progression requires inactivation of several tumor suppressor genes. These tumor suppressor genes are located in chromosomal regions which often appear to be deleted in tumor cells. The most frequently deleted chromosome arms in colorectal tumors are 5q, 17p, and 18q containing tumor suppressor genes APC, TP53, and SMAD4, respectively (Vogelstein et al., 1988; Kinzler and Vogelstein, 1996). APC has been proposed to function as a "gatekeeper" gene, regulating the entry of colon epithelial cells into the neoplastic process. Already the earliest precursor lesions, aberrant crypt foci (ACF), display mutations in APC (Jen et al., 1994). APC protein is a part of Wingless (Wnt) signaling pathway. In normal cells, in the absence of Wnt ligand, APC forms a complex together with axin and glycogen synthase kinase-3B (GSK-3β). This complex binds and phosphorylates β-catenin, and directs it to subsequent ubiquitination and degradation (Orford et al., 1997; Giles et al., 2003). When APC is mutated β-catenin degradation complex does not take form, and β-catenin accumulates in the nucleus. β-catenin is able to bind to the T-cell factor (TCF) family of transcription factors which again activate transcription of other tumor promoting genes, such as C-MYC and cyclin D1 (CCND1) (He et al., 1998; Wong and Pignatelli, 2002). Wnt signaling pathway may be disrupted also by mutated β-catenin. Approximately half of the tumors with wild type APC display mutations in β-catenin encoding gene CTNNB1, suggesting that CTNNB1 mutation can substitute APC mutation in colon carcinogenesis (Morin et al., 1997; Sparks et al., 1998). It, however, appears that mutations in APC and CTNNB1 are not functionally identical, since adenomas with CTNNB1 mutations seem not to progress to malignant tumors as likely as adenomas with APC mutation (Samowitz et al., 1999a). An ACF, which results from the initial mutations, eventually acquires additional mutations in other genes. Another genetic alteration which occurs early in the adenoma-

TP53 APC KRAS

SMAD4

SMAD2

Normal epithelium

Aberrant crypt foci

Early adenoma

Intermediate adenoma Late

adenoma Carcinoma Metastasis

21

carcinoma sequence is activation of KRAS oncogene (Vogelstein et al., 1988; Shibata et al., 1993). Kras is a guanosine 5'-triphosphate (GTP)-binding protein. It is located at the cytoplasmic site of a cell membrane, and is involved in transducing signals from receptor tyrosine kinases, like epidermal growth factor receptor (EGFR). Downstream elements of this pathway include cytoplasmic RAF serine-threonine kinase and mitogen activated protein (MAP) kinase (Vogelstein and Kinzler, 2004). K-ras is active, when it is bound to GTP but becomes inactive when GTP is hydrolysed to guanosine 5'-diphosphate (GDP). Carcinogenic mutations in KRAS affect codons in the GTP-binding domain, resulting in a constitutively active protein (Bos, 1988; McCormick, 1989). KRAS mutations occur in small adenomas, but it is more common in larger adenomas (Vogelstein et al., 1988; Scott et al.., 1993; Rashid et al., 1999). It seems that while mutated KRAS gives cells a growth advance it is unlikely an initiating factor in colorectal tumorigenesis. Cells with only KRAS mutation form foci of hyperproliferating cells, which have a normal cellular organization. These cells appear to have little or no potential to form clinically important tumors (Pretlow et al., 1993; Jen et al., 1994). Chromosome arm 18q is deleted in approximately 50% of colorectal adenomas and 70% of carcinomas (Vogelstein et al., 1988; Fearon and Vogelstein 1990). Initially the candidate tumor suppressor gene in this region was deleted in colorectal cancer (DCC). However, mutant alleles of DCC are seldom observed in CRCs showing 18q loss, and mouse studies involving DCC homologue have not been able to demonstrate the tumor suppressor function of DCC in cells (Cho et al., 1994; Fazeli et al., 1997). Two other tumor suppressor genes, SMAD2 and SMAD4 also reside in chromosome 18q. Smad proteins are components of the transforming growth factor-β (TGF-β) signaling pathway (Massague et al., 1998). TGF-β signaling has various inhibitory effects in cells, including regulation of cell growth, differentiation, and apoptosis (Heldin et al., 1997). Germline mutations in SMAD4 predispose to familial colon cancer syndrome juvenile polyposis (Howe et al., 1998), and somatic mutations in both SMAD2 and SMAD4 have been observed in CRC, as well as in several other cancers (Riggins et al., 1997). Furthermore, mutations in TGF-β type II receptor (TGF-βRII) have been detected in 90% of microsatellite instable CRCs, supporting the importance of aberrant TGF-β signaling in colorectal tumorigenesis (Parsons et al., 1995). TP53 is frequently altered in various human cancers (Caron de Fromentel and Soussi, 1992). It resides in the short arm of chromosome 17, which is often deleted also in CRC (Vogelstein et al., 1988). p53 is important in maintaining DNA integrity. In the presence of damaged DNA, p53 blocks cell proliferation until the damage has been repaired. If the damage is too great, p53 is able to induce apoptosis. Loss of function of p53 results in accumulation of mutations throughout the genome, karyotypic instability, and reduced apoptosis (Lane 1992; Donehower and Bradley, 1993; Carder et al., 1993; May and May, 1999). TP53 mutation represents a late event in colorectal tumorigenesis, as it tends to occur at the late adenoma stage. It, however, seems not to be an absolute requirement for malignant transformation, since CRCs develop also without detectable TP53 mutation (Baker et al., 1990).

22

3.2 The mutator pathway Colorectal tumors evolving through the suppressor pathway often show gross chromosomal instability (CIN) e.g. losses and amplifications of whole chromosomes. A mutator pathway is driven by a defect in DNA repair caused by the inactivation of DNA MMR genes (Boland et al., 1998a). CRCs evolving through mutator pathway are usually diploid but the deficient DNA repair causes accumulation of genetic alterations especially in the repetitive sequences. This phenomenon is called microsatellite instability (MSI) (Ionov et al., 1993; Thibodeau et al., 1993). Almost all colorectal tumors from Lynch syndrome patients display MSI. However, MSI occurs in approximately 15% of unselected CRCs as well. In these tumors MSI is usually due to MLH1 promoter hypermethylation (Kane et al., 1997; Cunningham et al., 1998; Herman et al., 1998). Figure 3. The mutator pathway. The genes mutated in MSI tumors often possess repetitive tracts in the coding sequence. Tumor suppressor genes are targets also in the mutator pathway, and a tumor progresses from adenoma to carcinoma, but most of the genes underlying CIN and MSI tumors are different (Figure 3). The repertoire of tumor suppressor genes mutated in MSI tumors at least in part includes genes that possess short repetitive tracts within their coding sequences. One such gene is TGF-βRII. It contains an (A)10 repeat in exon 3, which is mutated in up to 90% of MSI CRCs (Markowitz et al., 1995; Parsons et al., 1995). TGF-βRII mutations are present in late adenomas, and in adenomas containing invasive carcinoma, suggesting that inactivation of TGF-βRII is an important step in MMR-deficient CRC progression (Grady et al., 1998). BAX and IGFIIR show frequent mutation rates in MSI tumors as well (Rampino et al., 1997, Calin et al., 2000). The target for mutations in both BAX and IGFIIR is a (G)8 repeat. BAX is mutated in approximately 50% of MSI CRCs, while 10-20% of MSI tumors display mutations in IGFIIR (Schwartz et al., 1999; Calin et al., 2000; Mori et al., 2001). Bax belongs to a bcl-2 multiprotein family, which is a major regulator of apoptosis and consists of both inducers and repressors (Hirose et al., 1997). Bcl-2 prolongs cell survival by blocking apoptosis. Bax forms heterodimers with bcl-2, and thereby accelerates apoptosis (Krajewski et al., 1994).

BAX

BRAF

Metastasis Normal epithelium

Aberrant crypt foci

Early adenoma

Intermediate adenoma Late

adenoma Carcinoma

TGFβRII IGFIIR

MSH3

MSH6

23

IGFIIR is a growth factor receptor, which plays an important role in regulating cell growth and apoptosis. It participates in lysosomal degradation of a growth stimulating protein IGFII, and is required for the activation of TGFβ, an important growth inhibitor (Dennis et al., 1991; Ellis et al., 1996). Besides TGF-βRII, BAX, and IGFIIR, repeat mutations have been reported in mismatch repair genes MSH3 and MSH6 (Malkhosyan S, et al., 1996). Additionally, several other genes have been proposed to be mutated in MSI tumors, but the studies on these genes are currently not extensive enough to confirm their role in MSI tumorigenesis (Duvald and Hamelin, 2002). Approximately 30-40% of MSI tumors display mutations in BRAF oncogene (Rajagopalan et al., 2002; Domingo et al., 2004). BRAF is a member of RAF family of serine/threonine kinases. It mediates responses to growth signals through the RAS-RAF-MAP kinase pathway (Peyssonnaux and Eychene, 2001).

3.3 The serrated neoplasia pathway The serrated neoplasia pathway may represent a novel route that leads to malignant colorectal tumors. Serrated CRCs differ from conventional adenocarcinomas on morphology. Epithelium of serrated CRCs show saw-toothed structure similar to hyperplastic polyps (Figure 4), thus hyperplastic polyp is thought to be the precursor lesion of serrated CRCs. Hyperplastic polyp then progresses to serrated adenoma and further to serrated carcinoma (Hawkins et al., 2002) (Figure 5). Serrated CRCs appear to be quite rare, accounting for approximately 6% of all CRCs (Mäkinen et al., 2001). Their contribution to overall CRC burden thus equals CRCs associated with highly penetrant hereditary syndromes.

Figure 4. Hematoxylin-eosin stained sections from conventional (A) and serrated (B) adenocarcinomas showing the characteristic saw-toothed epithelium of the latter tumors. Recent studies have tried to clarify molecular changes underlying the serrated pathway. The serrated structure may be due to inhibition of apoptosis. Tateyama et al. (2002)

24

observed decreased expression of CD95 (Fas) in cells of the upper crypt. CD95 is a transmembrane glycoprotein that belongs to the tumor necrosis factor (TNF) superfamily. Binding of Fas-ligand to CD95 is able to induce apoptosis in normal colon cells (Strater et al., 1997). Charcteristic serrated morphology takes form as cells continue to proliferate but are blocked from ascending to their right location in the colonic epithelium. Approximately 20% of serrated adenomas display Wnt pathway abnormalities suggesting that a subset of serrated tumors develop along pathways involving changes in APC/β-catenin (Sawyer et al., 2002). Mutations in BRAF have been observed in serrated adenomas and tumors (Kambara et al., 2004). BRAF appears to be mutated early in the serrated pathway (Kambara et al., 2004). KRAS mutations, however, appear to be rare in serrated lesions since only 15% of serrated adenomas display mutations in KRAS (Sawyer et al., 2002). Loss of chromosome 1p may also represent an early event in the serrated pathway, since it can be detected already in hyperplastic polyps (Rashid et al., 2000). Figure 5. The serrated neoplasia pathway leading to CRC.

4. MICROSATELLITE INSTABILITY Microsatellites are short tandem repeats of simple sequences that are scattered throughout the genome. They typically consist of 10-50 copies of 1 to 6 bp motifs, and are characterized by a high degree of polymorphism. The most common repeats in humans are mononucleotide repeats (A)n/(T)n and dinucleotide repeats (CA)n/(GT)n. Microsatellites usually occur in non-coding regions (Weber and Wong, 1993; Strachan and Read, 1999). Microsatellite instability (MSI) is defined as “a change of length due to either insertions or deletions of repeat units in a microsatellite within a tumor compared to normal tissue” (Boland et al., 1998b). MSI occurs when cells are not able to repair base-base mismatches and insertion/deletion loops (IDLs) that arise as a consequence of DNA polymerase slippage during DNA replication. Normally mismatch repair (MMR) machinery repairs majority of damages in DNA, but in cells with defective MMR system the microsatellite mutation rate is highly accelerated (Bhattacharyya et al., 1994).

Metastasis Serrated carcinoma

Normal epithelium

Hyperplastic polyp

Serrated adenoma

1p loss? APC? BRAF?

CD95?

25

Approximately 90% of CRCs from Lynch syndrome patients show high degree of MSI (MSI-H), and it occurs also in 10-15% of sporadic CRCs (Aaltonen et al., 1993; Thibodeau et al., 1993; Aaltonen et al., 1994; Liu et al., 1996). Besides previously mentioned molecular differences, MSI-H and microsatellite stable (MSS) tumors display several clinical and pathological differences. MSI-H tumors are more likely to be poorly differentiated, and have a mucinous phenotype. They are more often located in the right side of colon, and are associated with female sex. MSI-H tumors also have a better prognosis (Ward et al., 2001). In addition to CRC, MSI has also been observed in other cancer types such as endometrium, breast, pancreatic, gastric and prostate cancer (Han et al., 1993; Burks et al., 1994; Yee et al., 1994; Suzuki et al., 1995). In 1997 National Cancer Institute workshop on microsatellie instability defined the uniform criteria for MSI, and proposed technical guidelines for MSI detection. The recommended method to analyze MSI is to study tumor and respective normal tissue DNA using a panel of five microsatellite markers; BAT25, BAT26, D2S123, D5S346 and D17S250. If two or more markers display instability, a sample is designated MSI-H. If none of the markers show instability a sample is MSS. If one marker shows instability a sample is defined as MSI-L (low-level MSI). In this case additional markers should be analyzed to distinguish between MSI-H and MSI-L. If more than five markers are analyzed a sample is MSI-H if more than 30% of markers show MSI. The remaining cases fall into MSI-L category (Boland et al., 1998). Since MSI is a potential marker for Lynch syndrome it is utilized in CRC diagnostics together with immunohistochemistry. Tumors from putative Lynch syndrome patients can be analyzed for MSI. If a tumor turns out to be MSI-H, the underlying germline MMR mutation can be detected using immunohistochemistry and direct sequencing of the gene (Table 3). Table 3. The Bethesda guidelines for the situations when colorectal tumors should be tested for MSI 1 CRC diagnosed before age 50 2 Presence of synchronous, metachronous CRC or other Lynch syndrome associated

tumor regardless of age 3 CRC with typical MSI-H histology (e.g. presence of tumor infiltrating lymphocytes,

Crohn's-like lymphocytic reaction, mucinous/signet ring differentiation, or medullary growth pattern)

4 CRC diagnosed in one or more first-degree relatives with an Lynch syndrome associated tumor, with one of the cancers being diagnosed under age 50

5 CRC diagnosed in two or more first- or second degree relatives with Lynch syndrome related tumors regardless of age

26

Whilst MSI-H and MSS tumors clearly form separate CRC subgroups, the role of MSI-L tumors have been obscure. MMR genes MLH1, MSH2, and MSH6 appear not to play a role in MSI-L tumor development (Percesepe et al., 1998; Thibodeau et al., 1998), and clinicopathological features does not differ between MSI-L and MSS tumors. Thus, many studies support the combining of these two groups (Mirabelli-Primdahl et al., 1999; Gonzales-Garcia et al., 2000; Parc et al., 2001; Ward et al., 2001). Some studies have, however, found molecular differences between MSI-L and MSS tumors. Reduced Bcl-2 expression, excess of KRAS mutations, lower deletion frequency in chromosome 5q, higher apoptotic activity and lymphocyte infiltration have been associated with MSI-L tumors (Biden et al., 1999; Jass et al., 1999; Michael-Robinson et al., 2001). Methylation of MGMT may also play a role in MSI-L tumor formation (Whitehall et al., 2001). The usefulness of MSI-L group was controversial already at the beginning. Its existence was, however, considered necessary since all genes responsible for MSI are not yet known, and some of them may be associated with more attenuated phenotype, like MSH6 (Akiyama et al., 1997; Wu et al., 1999). Despite that the workshop held in 2002 proposed that MSI-L and MSS tumors should not be separated for clinical purposes (Umar et al., 2004).

4.1 Mismatch repair system MMR system is well conserved in evolution. In Eschricia coli MutS recognises base-base mismatches and small IDLs (Modrich and Lahue, 1996). MutL homodimer then forms a complex with MutS, and couples mismatch recognition and downstream MMR events (Hall and Matson, 1999). Endonuclease mutH makes the discrimination between the template and newly replicated strands by introducing a nick in the new hemi-methylated strand (Buermeyer et al., 1999). DNA-specific exonucleases, such as RecJ, ExoVII, ExoI, and ExoX excise the incorrect DNA sequence (Burdett et al., 2001), and finally DNA polymerase III holoenzyme and DNA ligase take care of resynthesis and ligation (Modrich and Lahue, 1996). Eukaryotic MMR system functions largely in a similar manner, but is more complex than the system in E. coli. So far five MutS and four MutL homologs have been identified in human cells (Table 4). For mismatch recognition MSH2 forms a heterodimer with either MSH6 or MSH3 depending on the damage that needs to be repaired. The MSH2-MSH6 complex (MutSα) takes care of mismatches and single base IDLs, while the larger 2-4 bp IDLs are recognized by either MSH2-MSH3 complex (MutSβ) or MutSα. (Acharya et al., 1996; Marsischky et al., 1996; Edelmann et al., 2000; Jiricny 2000). A heterodimer of MLH1 and PMS2 (MutLα) interacts with mismatch recognition complexes (MutSα) and (MutSβ), and mediates the actual mismatch repair (Prolla et al., 1994; Wang et al., 1999) (Figure 6).

27

Table 4. MMR gene homologs in E.coli and humans, and the distribution of mutations detected in Lynch syndrome families (http://www.insight-group.org; The pathogenicity of each mutation has not been confirmed) E.coli H.sapiens No of detected mutations MutS MSH2 175 (38.5%) MSH3 - MSH4 - MSH5 - MSH6 32 (7.0%) MutL MLH1 225 (49.5%) MLH3 16 (3.5%) PMS1 1 (<0.1%) PMS2 5 (1.1%) Total 454 (100%) Human MMR complex most likely contains several other proteins. Like the repair system in E.coli endonuclease, exonuclease, DNA polymerase, and ligase are probably required in human MMR as well. The mechanism for strand discrimination in humans is unclear. No MutH endonuclease homologs have been identified in humans so far, and the only exonuclease known to act in MMR is exonuclease 1 (EXO1) (Schmutte et al., 1998; Tischkoff et al., 1998). MSH4 and MSH5 are not needed in the actual MMR, but are necessary for meiotic recombination (de Vries et al., 1999; Edelmann et al., 1999). The role of PMS1 in MMR awaits further research. Besides DNA repair, the MMR system plays an important role in signaling the presence of DNA damage to the apoptotic machinery (D’Atri et al., 1998; Hickman and Samson, 1999). This may explain why MMR deficient tumors are more resistant to DNA modifying drugs than their MMR proficient counterparts (Karran and Hampson, 1996; Fink et al., 1999).

28

Mismatches Insertion/deletion loops ATGGACCTAA CACACACACACACA TACCGGGATT GTGTGTGTGTGTGT GT ATGGCCCTAA CACACACACACACA TACCGGGATT GTGTGTGTGTGTGT Figure 6. Human MMR system

4.1.1 Defects in the mismatch repair system

The majority of MMR gene defects that predispose to Lynch syndrome are in either MLH1 or MSH2 (approximately 50 and 40%, respectively) (International Collaborative Group on Hereditary Non-Polyposis Colorectal Cancer http://www.insight-group.org). Mutations in both genes are located in all regions without obvious hotspots. In Finland two founder mutations account for the majority of Lynch syndrome mutations. Mutation 1 is a 3.5 kb genomic deletion causing an in-frame 165 bp deletion comprising MLH1 exon 16. Mutation 2 is a splice acceptor mutation in MLH1 leading to the deletion of exon 6, and causing frame-shift and a premature stop codon (exon 6, G->A at 454-1) (Nyström-Lahti et al., 1995). Approximately 10% of Lynch syndrome patients have mutation in MSH6 (http://www.insight-group.org). MSH6 mutations are associated with atypical Lynch syndrome with later age of onset, lower penetrance, and lower degree of MSI in tumors. In females truncating mutation of MSH6 seems more likely to predispose to endometrium cancer than CRC (Wijnen et al., 1999). PMS2 mutations are rare. To date

MSH2 MSH6

MSH2 MSH6

MLH1

PMS2

MSH2

MSH2 MSH3

MSH3

PMS2

MLH1

29

only five mutations have been reported in the HNPCC mutation database (http://www.insight-group.org). The role of MMR genes MLH3 and EXO1 in Lynch syndrome is not clear. Germline mutations have been found in both genes but subsequent studies have proven that many of the reported mutations occur in healthy controls as well. Further studies are needed to thoroughly clarify the role of these genes in cancer predisposition (Wu et al., 2001; Wu et al 2001b exo; Hienonen et al., 2003; Jamoghan-Changur et al., 2003).

30

AIMS OF THE STUDY 1. To study the capability of the mononucleotide marker BAT26 to detect mutation

positive HNPCC patients compared to the recommended Bethesda microsatellite marker panel.

2. To clarify the molecular character of MSI-L tumors and the functionality of the

Bethesda panel in detecting MSI-L tumors. 3. To identify new chromosomal regions for low and moderate penetrance genes

predisposing to CRC. 4. To study the molecular background of CRCs displaying serrated histology.

31

MATERIALS AND METHODS

1. PATIENT SAMPLES (I, II, III, IV) Approximately 1500 fresh frozen CRC samples have been collected in hospitals mainly located in Eastern Finland between May 1994 and December 2004. Of these 1042 tumors were a part of consecutively collected sample set that has been described by Aaltonen et al., 1998 and Salovaara et al., 2000. Samples selected for each study were a part of this collection. All samples have been analyzed for MSI using BAT26 and TGFβRII. All MSI samples have been analyzed for the two most common MMR gene mutations in Finland (see page 29). If neither mutation has been detected, mutation analysis of MLH1 and MSH2 has been performed using direct sequencing of the coding exons. Background information, such as sex, age at diagnosis, tumor stage, grade, and location of all patients has been carefully documented. First degree relatives have been identified through parish and population registries and their cancer status obtained from Finnish Cancer Registry. In study IV additional samples were used from a collection described by Mäkinen et al., 2001. The study was approved by Ministry of Social Affairs and Health and the Ethics committee of the Department of Medical Genetics, University of Helsinki.

1.1 DNA (I, II, III, IV) and RNA (IV) extraction DNA was extracted from fresh frozen tumors. A pathologist evaluated the proportion of tumor tissue in samples prior to DNA extraction to confirm that the maximal amount of tumor tissue was used. Samples typically displayed over 60% of tumor tissue. None of the samples displayed less than 50% of tumor tissue. Respective normal DNA was extracted either from normal colon mucosa or peripheral blood. A standard non-enzymatic method was used (Lahiri and Nürnberg, 1991). For RNA extraction the tissue was first homogenized using Ultra Turrax homogenizer. Samples were subsequently treated with Trizol reagent (Gibco BRL, Long Island, NY) and cleaned with RNeasy spin columns (Qiagen, Valencia, CA). The quality of the RNA was confirmed using Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA).

2. MSI ANALYSIS (I, II) The MSI status of samples used in study I and II was analyzed using the five microsatellite markers BAT25, BAT26, D2S123, D5S346, and D17S250 recommended by the International Workshop (Boland et al., 1998). In study II MSI analysis was

32

continued with 372 markers (ABI PRISM Linkage Mapping Set MD-10, P/N 450067). All markers were run in ABI PRISM 377 DNA Sequencer (Applied Biosystems (AB) Division, Foster City, CA), and the data was analyzed using either GeneScan 3.1 or Genotyper 2.5 softwares. The results were always evaluated by more than one individual. In cases of any ambiguity a previously presented mathematical model was used to score for MSI (Canzian et al., 1996).

3. IMMUNOHISTOCHEMISTRY AND MUTATION ANALYSIS OF MLH1, MSH2 AND MSH6 (I) Immunohistochemical staining of MLH1, MSH2, and MSH6 were performed to study the protein expression in 20 MSI-L and 2 novel MSI-H cases appearing in the microsatellite analysis. The details of immunohistochemistry are explained in study I. The 2 novel MSI-H cases were sequenced for mutations in MLH1 and MSH2 genes using previously described primers (Chadwick et al., 2001). To exclude the possibility of large deletions MLH1 and MSH2 were analyzed using Southern blotting. Genomic DNA was digested using EcoRI and analyzed with two cDNA probes designed for MLH1 and MSH2, respectively. The two new MSI-H cases were also analyzed for mutations in MSH6. The primer sequences and PCR conditions used to amplify the 10 exons of MSH6 are listed in Table 5. Exon 1 was amplified in 2, and exon 4 in 8 overlapping fragments. Primer sequences were designed using the Primer3 server (http://Frodo.wi.mit.edu/cgi-bin/primer3/primer3_www.cgi). Direct sequencing of all three genes was performed using Big Dye 3 Terminator chemistry, and reactions were run on ABI 3100 capillary sequencer (AB). The frequency of the variant observed in MSH6 exon 2 in one of the MSI-H cases was analyzed in 182 cancer-free control individuals and 83 CRC patients by single strand conformational polymorphism (SSCP). The SSCP procedure is described in study I.

4. LOCUS SPECIFIC LOH ANALYSIS (II) LOH at three different loci in chromosomes 5q, 17p, and 18q was analyzed using five microsatellite markers: D5S318, D5S346, TP53, D18S1156, and D18S363. The primer sequences for markers were: D5S318F: AGGATCTTCCCTCTTTCTCTCTG R: GGCATCTATGTTGATGGGATCTATC TP53F: ACTGCCACTCCTTGCCCCATTC R: AGGGATACTATTCAGCCCGAGGTC D18S1156F: CCTGCAAGTTTACTGGC R: CAATGACAACCTGTTGTTGG D18S363F: TTGGGAACTGCTCTACATTC R: GCTTCATTCTCTCACTGGAT. Primer sequences for D5S346 are listed in Table 1 in study I. PCR conditions as well as ABI PRISM 377 run were similar to the marker D5S346, and are described in study I.

33

Table 5. Primer sequences and PCR annealing temperature for the MSH6 gene Exon Primer sequence Annealing temperature Exon 1aF tcc gtc cga cag aac ggt tg Exon 1aR ttc gcg tga ggc cct ggc cga

56°C

Exon 1bF cgc tga gtg atg cca aca ag Exon 1bR caa ccc cct gtg cga gcc tc

56°C

Exon 2F tgc cag aag act tgg aat tg Exon 2R cac aca cac atg gca gta gtg a

57°C

Exon 3F tgc tgg gat tac agt cgt ga Exon 3R tcc ccc atc acc cta aca ta

59°C

Exon 4aF gaa aaa cag tgg ctg cac g Exon 4aR gag cca ttt cca gtc acc at

57°C

Exon 4bF tgg agt ggg gga tag tga ga Exon 4bR cac ttc ctc atc cca gga gt

57°C

Exon 4cF atc acc ccg att ttg atg c Exon 4cR gga tca cct tcc agc aca ct

55°C

Exon 4dF gga tca tta cca agg gta cac a Exon 4dR gcc atc act tag ctt ttc cc

55°C

Exon 4eF ata ccc ggc tcc cag ttt t Exon 4eR cct ctc tag tag ggt tcc ttc agt

55°C

Exon 4fF gcc tat caa cga atg gtg ct Exon 4fR ttt gaa tc ttc cag agc aga

55°C

Exon 4gF ccc aga cag cag ggc tat aa Exon 4gR gtt cct acc aat ccc cca cca at

55°C

Exon 4hF tga aca gag cct cct gga at Exon 4hR agc tgg caa aca gca cta ctt

57°C

Exon 5F taa aac ccc caa acg atg aa Exon 5R gga gta att tcc ctt tgc ttc c

55°C

Exon 6F gtt tat gaa act gtt act acc Exon 6R gca aat atc ttt tat cac at

53°C

Exon 7F gag tat tca ttt gtg att tt Exon 7R cgc cca tgt ttt taa gat agt agt ctt c

55°C

Exon 8F ccg atg ttg ctt ttc tgt cc Exon 8R cag aag tgc cct ctc aaa aa

55°C

Exon 9F gag agg gca ctt ctg ttg ct Exon 9R cac tag cca ggc aaa ctt cc

55°C

Exon 10F gga agg gat gat gca cta tga Exon 10R tgt tgt ctg aat tta cc acct ttg

55°C

34

5. MUTATION ANALYSIS OF KRAS (II) Mutation hotspots in KRAS codons 12, 13, and 61 were screened by direct sequencing of exons 1 (codons 12 and 13) and 2 (codon 61) as described previously (Servomaa et al., 2000). Direct sequencing was performed using Big Dye 3 Terminator chemistry, and reactions were run on ABI 3100 capillary sequencer (AB). Sequence chromatograms were analyzed using EditView 1.0.1 Software (AB).

6. MGMT, MLH1, AND EPHB2 PROMOTER HYPERMETHYLATION ANALYSIS (II, IV) DNA methylation patterns in the CpG islands of MGMT, MLH1, and EPHB2 promoter regions were determined by chemical modification of unmethylated, but not methylated cytocines to uracil, and subsequent PCR using primers specific for either methylated or the modified unmethylated DNA.

6.1 Bisulfite modification One µg DNA was denatured by 0.2 M NaOH for 10 min at 37°C. Thirty µl of 10 mM hydroquinone (Sigma, St. Louis, MO) and 520 µl of 3 M sodium bisulfate (Sigma) at pH 5, both freshly prepared, were added and mixed, and samples were incubated under mineral oil at 50°C for 16 hours. Modified DNA was purified using Wizard DNA purification resin (Promega, Madison, WI) and eluted into 50 µl of water. Modification was completed by 0.3 M NaOH treatment for 5 min at room temperature, followed by ethanol precipitation. DNA was resuspended in water and used immediately or stored at -20°C.

6.2 Methylation specific PCR Methylation specific primers for MGMT and MLH1 have been described previously (Herman et al., 1998; Esteller et al., 1999), primers for EPHB2 are described in study IV. Primers have been designed to amplify either methylated or unmethylated bisulfite modified DNA. CRC cell line SW48 was used as a positive control and DNA from normal lymphocytes as a negative control for methylated alleles of MLH1 and MGMT. For EPHB2 in vitro methylated DNA (Chemicon International, Temecula, CA) was used as a positive control for methylated alleles, while paired normal tissue DNA was used as a negative control. Ten µl of each PCR reaction was directly loaded onto non-denaturing 6% polyacrylamide gels, stained with ethidium bromide, and visualized under UV illumination.

35

7. TOPOGRAPHIC MUTATION HETEROGENEITY ANALYSIS (II) To study the possible topographic mutation heterogeneity of the observed MSI shifts, stained microscopic sections from three MSI-L cancers were subdivided by microdissection into 6-10 small subregions. The sizes of the microsatellite alleles present in each subregion were determined by electrophoresis with 6% acrylamide sequencing gels after incorporating [33P]dCTP during PCR. In addition, DNA from whole tumors was diluted to essentially single molecules prior to PCR to see if intratumor microsatellite mutation heterogeneity was present.

8. ALLELIC IMBALANCE (AI) ANALYSIS (III) The analysis was done in three parts: First, the 104 CRC samples (29 familial and 75 sporadic cases, sample set S1) were analyzed with 372 CA-repeat markers (ABI PRISM Linkage Mapping Set MD-10, P/N 450067, 10 cM density, Applied Biosystems (AB) Division, Foster City, CA). PCR reactions for all these markers were carried out in 5 µl reaction volume containing 20 ng of genomic DNA, 1 × AmpliTaq Gold PCR buffer (AB), 250 µM each deoxynucleotide triphosphate (Amersham Pharmacia Biotech, NJ), 1,7 pmol each primer, 2,5 mM MgCl2 and 0.4 units of AmpliTaq Gold polymerase (AB). PCR cycles were 95°C for 12 min, 10 cycles at 94°C for 30 s, 55°C for 30s and 72°C for 30s, 20 cycles at 89°C for 30s, 55°C for 30s and 72°C for 30s. The final extension was 72°C for 10 min. PCR products were run in ABI PRISM 377 DNA Sequencer (AB) according to manufacturer´s protocols and the collected data was analyzed by GeneScan 3.1 and Genotyper 2.5 software (AB). To score for AI, a previously presented mathematical model was utilized. AI was scored if an allele was relatively decreased or increased by more than 40% (Canzian et al., 1996). To further evaluate the initial findings the sample set was extended with 59 familial and 64 sporadic cases (sample set S2). Loci were selected for further analyses when the following criteria were fulfilled: The AI percentage at a marker locus was 10 units or greater (e.g. 20% versus 30% AI cases) in the familial group, and at least one adjacent marker showed such difference as well. From these clusters the locus showing the greatest difference was further analyzed. Also two loci which showed 30 percent units more AI in the familial group were included. Third, 47 additional sporadic and 11 familial cases (Sample set S3) were analyzed with the markers showing significant AI difference after combined analysis of sample sets S1 and S2. S1, S2, and S3 comprised in total 285 CRC samples, 99 of which were familial and 186 sporadic. One locus was further evaluated in CRC samples from Danish (n=42) and UK (n=163) populations.

36

9. COMPARATIVE GENOMIC HYBRIDIZATION (CGH) (III) To characterize genomic imbalances in familial CRCs, 26 tumors were examined by CGH. Tumor DNA and normal reference DNA were labeled by nick translation with fluorescein-isothiocyanate-conjugated dCTP and dUTP (DuPont, Boston, MA) and Texas-red-conjugated dCTP and dUTP (DuPont). The hybridization was analyzed using an Olympus fluorescence microscope mounted to a CCD camera and the ISIS digital image analysis system (MetaSystems, Altlussheim, Germany). Three-color images (green for tumor DNA, red for reference DNA, and blue for chromosome counterstaining) were acquired from 8-10 metaphases with strong uniform hybridization. Chromosome regions were interpreted as over represented when the green-to red ratio was higher than 1.17 (gains) and underrepresented when the ratio was lower than 0.85 (losses).

10. REAL–TIME QUANTITATIVE PCR To further study the proportion of 20q amplification in both familial and sporadic tumors, the 26 familial tumors included in the CGH study, and 26 sporadic cases were analyzed by real-time quantitative PCR. The relative DNA copy numbers were determined as previously described by Ginzinger et al. (2000). For reference, three microsatellite loci, D10S586, D11S1315, and D21S1904) were chosen in genomic regions that usually did not show alterations in the genome-wide study.

11. STATISTICAL ANALYSES (II, III) The statistical analyses used in studies II and III are described in the respective works.

12. ARRAY ANALYSIS (IV) RNA from 8 serrated and 29 conventional adenocarcinomas was first transcribed into double stranded cDNA. Biotin labeled cRNA was synthesized from cDNA using a labeling kit containing biotin-labeled UTP and CTP in addition to unlabeled ribonucleotides. Purified and fragmented cRNA was hybridized onto oligonucleotide arrays overnight in constant rotation. After hybridization, arrays were washed, stained with streptavidin-phycoerythrin, and scanned. The arrays were initially analyzed with Microarray Suite 5.0 software (Affymetrix Inc., Santa Clara, CA). The array analysis is described in detail in study IV.

37

13. ARRAY DATA ANALYSIS (IV)

13.1 Normalization and filtering The expression data was normalized in three steps: 1) data transformation where values below 0.01 were set to 0.01 to enable logarithmic transformation, 2) centering of arrays where each measurement was divided by the median (50th percentile) of all measurements in that sample, and 3) centering of gene intensities where each gene was divided by the median of its measurements in all samples. After normalization the data was filtered using Affymetrix flag calls (Absent, Present Marginal), so that the program allowed inclusion of all genes that had Present or Marginal flag in at least 29 samples. After filtering 7928 probe sets remained in the data analysis. Normalization and filtering were performed using GeneSpring 6.2.1 software (Silicon Genetics, Redwood City, CA).

13.2 Unsupervised hierarchical clustering and identification of differentially expressed genes To study whether serrated CRCs had a different overall gene expression when compared to conventional adenocarcinomas an unsupervised hierarchical clustering of the filtered 7928 probes was done using Spearman’s Rank Correlation as a similarity measure. A Student’s t-test and Benjamini and Hochberg multiple testing correction was used to analyze differentially expressed genes between serrated and conventional CRCs. Both of the above mentioned analyses were done using GeneSpring.

13.3 Class prediction Class prediction method was used to find a set of differentially expressed genes (probes) between the serrated and non-serrated CRCs that could predict whether an unknown sample was a serrated or a conventional tumor. The predictor was created using 4413 probes that had a Present or Marginal flag call (MAS5) in all of the 37 samples. The predictor was built and validated in GeneCluster2 software using a K-Nearest Neighbors (KNN) algorithm. Validation of the predictor was performed using a leave-one-out cross-validation method.

38

14. IMMUNOHISTOCHEMISTRY (IV) The array results were validated in a larger dataset by immunohistochemistry. The sample set was divided into two separate sets: training set (TS) consisted of the 37 samples hybridized on arrays (8 serrated and 29 conventional CRCs), validation set (VS) consisted of additional 37 serrated and 86 conventional CRCs. Six genes, ephrin type-B receptor 2 (EPHB2), patched (PTCH), hypoxia-inducible factor 1-alpha (HIF1α), Cyclin T2 (CCNT2), metastasis associated-1 (MTA1), and methyl-CpG binding domain protein 4 (MBD4) that were differentially expressed between serrated and conventional CRCs were chosen for staining. Besides the difference in expression level these genes were biologically relevant, and the antibody was commercially available. The details of the immunohistochemistry are provided in study IV.

15. MUTATION SCREENING OF EPHRIN TYPE-B RECEPTOR 2 (EPHB2) (IV)

15.1 Direct sequencing Twenty-four serrated samples were screened for mutations in the coding region of EPHB2. Primers were designed using Primer3 server (http://Frodo.wi.mit.edu/cgi-bin/primer3/primer3_www.cgi). The primer sequences and PCR conditions are listed in Table 6. Exon 3 was amplified in two fragments, and exons 16 and 17 in one fragment. Purified PCR products were run on ABI 3730 sequencer (AB), and sequence chromatograms analyzed using BioEdit Sequence Alignment Editor 7.0.1.

39

Table 6. Primer sequences and PCR annealing temperature for EPHB2 gene Exon Primer sequence Annealing temperature Exon 2F ccgggagagaccactgct Exon 2R cctcctcctggcacagtca

TD 64°C->61°C*

Exon 3aF gggttccctaaggcagagtc Exon 3aR cacttgcggtagaagacacg

TD 62°C->60°C*

Exon 3bF tgaaggtggataccattgcag Exon 3bR agtctgcagccaagaccagt

TD 64°C->61°C*

Exon 4F cctagacagctcacggacaag Exon 4R tccatttctcttcccaccag

62°C

Exon 5F tttctctggtggggagacag Exon 5R ggaccgatcacatgaaaagc

62°C

Exon 6F ttgatgtggcctgagcttc Exon 6R cagtgctctttctgctgcac

62°C

Exon 7F agggaggggatgaagaattg Exon 7R acaacgggaacagaaacagg

62°C

Exon 8F aggattggagcagttcatgg Exon 8R tcggttcaccagcactgtag

62°C

Exon 9F gacttgaggtgggaggagtg Exon 9R cttgcaaaggcctagaggtg

62°C

Exon 10F ctcagctgctccagaaaacc Exon 10R tcccacatcctaccttctgc

62°C

Exon 11F atgagtgggccattgagaag Exon 11R ctgccatggataagggaatg

59°C

Exon 12F ccactctgaagttggcctct Exon 12R ggagagatggacactcatgc

TD 61°C->59°C*

Exon 13F aaatgcccagtggtctgaag Exon 13R gaagacgtgctgatgtgtgg

62°C

Exon 14F ccaggaggtgaaagtggttg Exon 14R cccaagctgcagtaaggaag

62°C

Exon 15F acgtgcacattcacgcatac Exon 15R gtgagaaaacggaggctgag

59°C

Exon 16-17F cttgctttgccatcttcctc Exon 16-47R gacacatcgcatgaatctcc

TD 62°C->60°C*

*TD = Touch down from one temperature to another, decreasing 0.5°C/cycle

40

RESULTS

1. COMPARISON BETWEEN BAT26 AND THE BETHESDA PANEL (I)

1.1 MSI analysis using the Bethesda panel Out of 494 patients analyzed 73 had shown MSI when previously analyzed with BAT26 only, and 27 patients had an underlying MMR gene mutation (Aaltonen et al., 1998; Salovaara et al., 2000). All of these classified as MSI-H after applying all the five Bethesda markers. Altogether 95 of 494 patients classified as MSI when the Bethesda panel was used: 75 (15.2%) of patients classified as MSI-H, and 20 (4.0%) patients classified as MSI-L. Sequencing of coding regions, exon-intron boundaries and promoter regions of MLH1 and MSH2 genes of the 2 novel MSI-H cases revealed no mutations. Neither did we detect any aberrant bands in Southern blot which was used to study the possibility that large deletions would have caused the MSI-H phenotype. One of the novel MSI-H patients, however, displayed a previously reported missense mutation in MSH6 exon 2: S144I (AGC->ATC). We analyzed 83 CRC patients and 182 cancer-free controls with SSCP for the change and found it in one control sample. The tumor harboring the variant showed no loss of heterozygosity at this change. Additionally, the 22 novel MSI cases appearing in this study were studied for MLH1, MSH2, and MSH6 protein expression using immunohistochemistry. All samples scored positive (++, +++, or ++++) in the staining. BAT26 was able to identify all 27 MMR gene mutation positive individuals in the sample set. It failed to identify 2 MSI-H cases and 20 MSI-L cases. No MMR deficiency was observed in these samples.

2. ANALYSIS OF LOW-LEVEL MSI (II)

2.1 MSI analysis Altogether 90 samples were analyzed for MSI using 377 microsatellite markers. After the initial analysis with the Bethesda markers 6 of the 90 tumors (7%) displayed MSI. The addition of eight markers listed in the Bethesda reference panel resulted in 9 of 90 (10%) tumors showing MSI. After analysis with all 377 markers 71 tumors had displayed MSI. Out of 90 tumors 19 did not show MSI at all. The number of mutated loci varied from 1

41

to 11 (see figure 2 in study IIa). Only dinucleotide markers displayed MSI, none of the samples was unstable with mononucleotide markers BAT25 or BAT26. The most sensitive markers were D6S441 and D10S217. They detected six and five positive lesions, respectively. Topographical distribution of the microsatellite mutations were studied in three tumors showing instability with 9-11 markers. The lack of widespread allelic or topographical mutation suggests that most microsatellite mutations were acquired before final clonal expansion. Topographical heterogeneity was present only at two of six loci from a single cancer. When tumor DNA was diluted before PCR no variation in mutation size was observed. Thus, it appears that only single somatic mutation was present throughout the tumor.

2.2 Comparison between putative MSI-L and MSS groups Several molecular and clinical differences, including locus specific LOH frequency, KRAS mutation frequency, promoter methylation of MGMT and MLH1, sex, age, family history of CRC, tumor site, stage, and grade, were studied to analyze if there were differences between MSI-L and MSS tumors. All eleven possible cutoff levels for MSI were used to see whether significant associations existed (see Table 1 in study IIa). None of the examined molecular features were statistically significantly different between the two groups. When MSI cutoff was set between zero and one or more mutated markers, chromosome 5q was deleted in 16% of MSS cancers vs. 36% in MSI-L cancers (p=0.10), chromosome 17p was deleted in 38% of MSS cancers vs. 53% in MSI-L cancers (p=0.40), and chromosome 18q was deleted in 60% of MSS cancers vs. 27% in MSI-L cancers (p=1.0). KRAS mutation frequency was 37% among MSS tumors and 31% among MSI-L tumors (p=0.26), giving no implication that MSI-L tumors would have higher KRAS frequency. Neither did we observe any association between MSI frequency and KRAS mutation status (p=0.78). MGMT promoter was methylated in 11% of MSS tumors and 27% of MSI-L tumors, while MLH1 promoter was methylated in 5% of MSS tumors and 10% of MSI-L tumors (p=1.0). There was no evidence for a relationship between MSI-L status and either tumor stage, location or family history of cancer. The only statistically significant relationship between grade and MSI-L status was observed in tumor grade, when the cutoff for MSI was set to <3 (p=0.01 using 0 as a cutoff). When patient age at tumor removal was plotted with respect to the relative frequency of microsatellife mutation, there was a significant relationship between age and frequency of MSI (p=0.012) (see Figure 3 in study IIa). It, thus, appears that older patients tend to have more microsatellite mutations.

42

3. ALLELIC IMBALANCE IN FAMILIAL AND SPORADIC TUMORS (III)

3.1 AI analysis In step one, we analyzed 29 familial and 75 sporadic CRCs (sample set S1) for AI , using a genome-wide panel of 372 microsatellite markers. Overall, AI varied from 0 to 59% between markers. AI was especially frequently observed in chromosomes 5q, 17p, and 18q. Previous studies have shown that these regions are often deleted in CRCs. After the first step several loci appeared, in which familial tumors showed more AI when compared to sporadic tumors. Odds ratios (ORs) were computed for each of the 22 autosome markers. Figure 2 in study III shows crude ORs and associated 95% confidence intervals for individual marker loci. After analysis of the first 104 samples ORs were significantly increased for five chromosome regions in the familial group: 7q21.3, 14q24.3, 20p12.2, 20q13.1, and 20q13.3. The most promising regions were in chromosome 20. In step two to see if the observed differences held in a larger data set the sample set was extended with an additional 59 familial and 64 sporadic CRCs (sample set S2). Ten promising loci identified in step one were analyzed and the OR associated with each marker recomputed. The only significant relationship between AI and family history was seen with the three chromosome 20 loci. These markers were chosen for step 3. In the third step the three chromosome 20 markers, D20S186 in 20p12.2, D20S178 in 20q13.1 and D20S171 in 20q13.3 were analyzed with set S3, which consisted of an additional 11 familial and 47 sporadic CRCs. After the third step the OR for D20S186 was 1.72 (p=0.06; 95%CI: 0.98-3.03), for D20S171 the OR was 1.97 (p=0.078; 95%CI: 0.93-4.18), and for D20S178 the OR was 2.74 (p=0.002; 95%CI: 1.44 - 5.20). After analyzing 99 familial and 186 sporadic CRCs altogether, the relationship between AI in familial risk remained significant in chromosome 20q13.1. The difference was present in a small Danish sample set. Out of 10 informative familial cases 8 showed AI at D20S178 when compared with 1 out of 11 in sporadic tumors (p=0.002). However, no difference in AI was detected in a larger sample set from UK.

3.2 Comparative genomic hybridization and DNA copy number measurement To study whether interesting chromosomal regions were lost or amplified, 26 familial CRCs were analyzed by CGH. The CGH results are presented in Table 7. It appeared that the mechanism underlying AI in chromosome 20 was amplification.

43

Table 7. The most frequently altered chromosomal regions detected by CGH Chromosome Minimal overlapping Proportion of tumors arm region showing alteration Lost chromosomal regions 18q 18q22-qter 81% 14q 14q12-23, 14q24-32 39% 8p 8p21-pter 23% 4q 4q21-qter 23% Amplified chromosomal regions 20q 20q11.2-q12 77% 13q 13q13-q21 50% 7p 7p21-pter 39% 7q 7cen-q21 39% 8q 8q24.1-qter 27% The DNA copy numbers in the sample set was studied using quantitative real-time PCR. The copy numbers in the familial group varied from 2 to 5 and in the sporadic group from 1 to 14. The mean values in the respective groups were 3 and 4. In all, 72% (18/26) of familial and 73% (19/26) of sporadic cases were defined as amplified.

4. GENE EXPRESSION PROFILE OF SERRATED CRCs (IV) The unsupervised hierarchical clustering was performed using 7928 probes. All serrated samples except for one clustered in the same branch. The only feature that clearly separated samples was the histology, no clustering according to any other parameter, like stage, grade or sex was observed (see Figure 1 in study IV). Multiple test corrected Student’s T-test revealed 226 probes (201 genes, and 7 hypothetical proteins or ESTs) that were statistically significantly differentially expressed between serrated and conventional CRCs (p<0.05). The most significantly enriched functional categories in the list of 226 differentially expressed genes were morphogenesis (p=0.0002), organogenesis (p=0.0002), integral to membrane (p=0.0035), and membrane (p=0.0049). We also created a predictor that could predict the subtype (serrated or conventional CRC) of an unknown sample. The predictor consisted of 27 probes, which are listed in Supplementary table 5 in study IV. Using these probes the cross-validation procedure was able to predict the subtype of an unknown tumor with 100% accuracy.

44

4.1 Immunohistochemistry To validate the array results, four differentially expressed genes, EPHB2, PTCH, HIF1α, and CCNT2 were immunohistochemically stained in 160 samples (45 serrated and 115 conventional tumors) that were divided into a training set (TS) and a validation set (VS). Additionally, the whole TS but only a subset of the VS (in total 59 tumors) was stained for MBD4 and MTA1. Normal colon mucosa showed positive nuclear staining for Cyclin T2, MTA, and MBD4, and positive cytoplasmic staining for EphB2 and Patched. In normal mucosa HIF1α was negative. Serrated CRCs typically showed weak immunoreaction for EPHB2 (in TS p=0.001, in VS p=0.029, chi-square), diffuse and weak immunoreaction for PTCH (staining pattern p=0.01, staining intensity p=0.693 in TS, respectively, staining pattern p=0.0003, staining intensity p=0.006 in VS, respectively) and frequent positivity for HIF1α (in TS p=0.021, in VS p=0.00006). Additionally, serrated CRCs showed a tendency to preserved nuclear immunoreaction for Cyclin T2. The difference was statistically significant in the training set (p=0.004) but failed to reach significant level in the validation set (p=0.21) (see table 2 in study IV). The expression of MTA1 and MBD4 did not show correlation between cancer types.

4.2 EPHB2 mutation screening and promoter hypermethylation analysis Altogether 24 serrated tumors were sequenced for somatic mutations in EPHB2. After sequencing 98% of the coding region no somatic mutations were detected. LOH was assessed using intronic SNPs. Three out of 12 (25%) informative samples showed LOH at EPHB2 locus. Promoter was hypermethylated in 5 out of 8 (63%) tumors.

45

DISCUSSION

1. MSI AS A DIAGNOSTIC TOOL (I) The characteristic feature of Lynch syndrome is an early onset CRC. Affected patients often develop tumors before the reach the age of 50, and they often have synchronous and metachronous tumors. Many Lynch syndrome associated CRCs have been prevented because of an accurate diagnosis, regular colonoscopies and removal of early lesions. Järvinen et al., (2000) reported that the risk for CRC decreases by 50-60% if a colonoscopy is performed every three years. It is, therefore, very important to identify Lynch syndrome tumors, so that patients and their families can be directed to counseling and screening procedures. MSI is a practical tool in Lynch syndrome diagnostics. Up to 90% of Lynch syndrome tumors display MSI (Aaltonen et al., 1994; Pedroni et al., 1999). In 1997 the National Cancer Institute workshop on microsatellie instability suggested a panel of five microsatellite markers (BAT25, BAT26, D2S123, D5S346, D17S250), referred to as the Bethesda panel, to be used in MSI analysis (Boland et al., 1998). Back then there was little experience how practical the panel would be in everyday diagnostics. Quasimonomorphic mononucleotide markers, like BAT26, have some advantages when compared to dinucleotide markers. Allelic variations in the Caucasian population are rare (Samowitz et al., 1999), which obviates the use of the corresponding normal tissue in MSI analysis. In addition, interpretation of MSI from mononucleotide graphs is usually unequivocal. In study I we examined the ability of BAT26 to identify tumors harboring germline MMR gene mutations compared to the whole set of Bethesda markers. Out of 494 samples included in this study 73 had previously been designated as MSI when analyzed with BAT26 alone, 27 samples had a germline MMR gene mutation. BAT26 detected all mutation positive samples, and 97% (93/95) of all MSI-H cases. The two novel MSI-H cases arising after analysis with the Bethesda panel did not harbor a mutation in MLH1, MSH2 or MSH6 genes, and showed instability only with dinucleotide markers. BAT26, thus, appears to be a sensitive marker when prescreening putative Lynch syndrome patients. It has been suggested that the Bethesda set would underestimate the number of MSI-H tumors and overestimate the number of MSI-L tumors because the panel consists of two mononucleotide markers and three dinucleotide markers. Adding more mononucleotide markers would improve the sensitivity of the panel (Umar et al., 2004). Suraweera et al. (2002) has shown that a set of five quasimonomorphic mononucleotide markers evaluates tumor MSI status accurately. All five markers can be pooled and amplified in a single PCR reaction, it is, therefore, simpler to use than the Bethesda panel and is more specific than using BAT26 alone. Using an extended set of mononucleotide markers may be

46

arguable. Although BAT26 is sensitive in detecting MSI, and in our study it identified all mutation positive cases, it has also failed to detect mutation positive cases (Shia et al., 2005). BAT26 is quasimonomorphic in Caucasian populations, but 12.6% of African Americans show allelic variation in BAT26 locus (Pyatt et al., 1999; Samowitz et al., 1999b), therefore using BAT26 only in more mixed populations may lead to false positive samples. The use of BAT26 only may also become a problem in MSH2 deletion carriers as BAT26 is located in MSH2. Using dinucleotide markers in diagnostic purposes has been shown to be fairly impractical. Dinucleotide markers are often polymorphic, thus their use in MSI analysis requires the analysis of the respective normal DNA. This is often laborious and time consuming. Moreover, interpretation of MSI in dinucleotide repeats can be difficult which leads to misclassifications. The role of MSI-L in tumors is not clear. It has been suggested that only MSI-H characterizes tumors with true MMR defect (Laiho et al., 2002; Tomlinson et al., 2002). The National Cancer Institute workshop suggested in 2004 that MSI-L classification for clinical purposes should be reconsidered. In our study the two novel MSI-H cases showed instability only with two dinucleotide markers in the absence of BAT26 instability. Both samples expressed MLH1, MSH2, and MSH6, and no germline mutation in MMR genes were detected. As a diagnostic tool MSI alone is not sufficient. It is specific but not very sensitive since approximately 15% of unselected CRCs display MSI as well, and only a subset is Lynch syndrome cases. Potential Lynch syndrome tumors are selected to MSI analysis based on evaluation of the family history. Positive MSI test indicates a possible gene defect but it does not pinpoint the actual mutated gene. Immunohistochemical staining of MMR genes is often used to identify the underlying gene, and the mutation is confirmed by direct sequencing. Since the first recommendations of the National Cancer Institute workshop in 1998, several research groups have studied whether MSI analysis can be replaced using immunohistochemistry alone. It appears that immunohistochemistry does not quite reach the sensitivity of MSI analysis. Since it fails to detect some mutation positive cases, immunohistochemistry alone is not a sufficient tool either (Lindor et al., 2002; Chapusot et al., 2004; Halvarsson et al., 2004; Jover et al., 2004; Muller et al., 2004; Shia et al., 2005). However, MSI analysis together with immunohistochemistry appears to be highly sensitive and reliable method to identify Lynch syndrome patients.

2. THE ROLE OF MSI-L IN COLORECTAL TUMORS (II) Classifying colorectal tumors according to MSI status has clear impact on patient diagnosis and treatment. MSI-H tumors are molecularly and clinically different from both MSS and MSI-L tumors; one of the most significant differences being a more favorable prognosis (Gryfe et al., 2000; Ward et al., 2001). Prognosis, besides that it is of pivotal importance for patients, it also guides the nature and intensity of treatment, such as chemotherapy. The National Cancer Institute workshop in 1998 suggested that CRCs should be divided into three subgroups, MSI-H, MSI-L, and MSS depending on the

47

proportion of markers that show MSI (Boland et al., 1998). The role of MSI-L tumors have been ambiguous, since they in many ways resemble MSS tumors (Lothe et al., 1993; Cawkwell et al., 1997; Gonzalez-Garcia et al., 2000; Ward et al., 2001). While some studies have found molecular differences between MSI-L and MSS tumors, others have failed to distinguish the two groups. The aim in the study II was to examine if MSI-L tumors truly form a relevant CRC subgroup, and to study whether the division of CRCs into three subgroups for clinical purposes is necessary. We also wanted to establish a cut-off level for MSI-L, above which the tumors would display molecular and/or clinical features different from the tumors under the cut-off level. The markers selected in the Bethesda panel have been proposed to be sensitive in detecting MSI (Dietmaier et al., 1997; Boland et al., 1998). However, in our study they appeared to be quite insensitive. Thirteen markers listed in the Bethesda panel identified 9 MSI-L tumors, thus leaving 62 tumors displaying similar degree of MSI undetected. It appears that the marker set used in MSI-L analysis so far is inadequate, since it only identifies a subset of MSI-L tumors. After completion of MSI analysis with 377 markers almost 80% of the tumors had displayed MSI. It is therefore conceivable to hypothesize that if even more markers would have been analyzed, all tumors would have scored as MSI-L. Jass et al., (2002) question the value of employing a large number of relatively stable markers, and suggest that only the markers that have proven to be sensitive in MSI-L analysis should be applied. However, the 11 recommended Bethesda dinucleotide markers were in no way more sensitive than other dinucleotide markers in the set. The MSI sensitivity of the 11 most sensitive markers in our set of 364 randomly chosen dinucleotide markers was equal or better than the most sensitive Bethesda marker D2S123. If the MSI-L analysis of this tumor set would have been performed first using the 11 markers listed in the Bethesda panel, and subsequently using the 11 markers that were the most sensitive among the 364 randomly chosen dinucleotide markers, the two sets of detected MSI-L tumors would have been almost completely different (see Figure 1 in study IIb). The reported differences between MSI-L and MSS tumors include higher KRAS mutation frequency, apoptotic activity, and lymphocytic infiltration, lower LOH frequency in chromosome 5q, and reduced Bcl-2 expression in MSI-L tumors (Biden et al., 1999; Jass et al., 1999; Michael-Robinson et al., 2001). It has also been reported that MGMT gene promoter methylation correlates with KRAS mutations in MSI-L tumors (Whitehall et al., 2001). In the study II we analyzed KRAS mutation frequency, promoter hypermethylation of MGMT and MLH1, LOH frequency at APC, p53, and SMAD4 loci, as well as several clinical features e.g. sex, stage, grade, site, and family history of cancer. The tumors displayed MSI at 0-11 loci. When examining all the possible cut-off levels for MSI-L we did not detect any convincing differences between the putative MSS and MSI-L tumors in terms of the above mentioned features (see Table 1 in study IIa). The distribution of microsatellite mutations was not, however, random, since tumors from older patients tended to have more mutations. Mutation rates observed in this data set are similar to the median rate of mutations observed in normal human fibroblasts (1.3x10-7 mutations per division) (Boyer and Farber, 1998), suggesting that the mutations detected in samples reflect the number of cell division and stochastic mutation (see Figure 3 in study IIa). Nor

48

did tumors show genetic heterogeneity in the microsatellite mutation patterns, suggesting that the mutations have occurred before the final clonal expansion. The study by Halford et al. (2002) also attempted to define the role of MSI-L. The study yielded highly similar results to ours; no differences between MSI-L and MSS tumors were observed, yet the distribution of microsatellite mutations was not random. Unfortunately Halford et al. (2002) did not proportion the number of mutations to patients’ age. Thus, it is not known if older patients displayed more microsatellite mutations in that data set as well. Halford et al. (2002) hypothesized that the difference in mutation frequency could be due to variation in the natural history of tumors. If a tumor had developed through a pathway that involved few cell divisions, fewer microsatellite mutations would have taken place. Another explanation could be polymorphisms in DNA repair genes that have only subtle affect on MMR system.

Figure 6. D1S413 shows an example of a minor extra band, which failed to qualify as MSI. D2S151 illustrates an example where the RER-score was calculated to confirm the MSI status (RER-score 25%). PCR artefacts may also complicate MSI scoring. This could have hampered the assessment of MSI in many studies. Many of the DNA polymerases routinely used in

49

PCR reactions do not have a proofreading activity, and gains and losses of repetitive units occur also during PCR reaction. So-called stutter bands are frequently observed in sequencer outputs. In dinucleotide repeats losses of repeat units are much more frequent, but also gains are detected (Hite et al., 1996). Distinguishing PCR artifacts from real insertions in the cases which show small insertion bands is particularly difficult (Figure 6). The mathematical model known as a RER-score can be utilized in determining the MSI status of visually ambiguous samples (Canzian et al., 1996). However, the RER-score appear not to be widely applied, and in many studies only visual criteria are used in scoring MSI. In conclusion, it appears that MSS and MSI-L tumors have a common molecular background, and that distribution of CRCs into three subgroups does not seem to be justified at least for clinical purposes. The mechanisms underlying colorectal cancer are, however, not clear, and further studies may prove the existence of a true MSI-L pathway. Therefore maintenance of MSI-L subgroup for research purposes may be entitled, but the current method to detect MSI-L tumors should be revised.

3. CHARACTERIZATION OF CHROMOSOMAL REGIONS CONTAINING PUTATIVE LOW-RISK ALLELES FOR CRC (III) Much of our information about the molecular pathogenesis of cancer has arisen from studies on hereditary tumors. Dominantly inherited CRC syndromes account for approximately 5% of all CRCs (Mecklin, 1987; Aaltonen et al., 1998). The underlying gene defects causing CRC syndromes are well characterized. Besides that these genes are key players in the respective syndromes they also have a major role in sporadic colon tumorigenesis, exemplified by APC and SMAD4. Except the hereditary syndromes, the distinction between sporadic and familial CRCs is not unequivocal. The proportion of familial cases is often estimated to be 20-25%, however, it may be that up to 35% of CRCs can be ascribed to inherited genetic susceptibility (Lichtenstein et al., 2000). As is the case in hereditary syndromes, the moderate and low-penetrance alleles are likely to play an important role in sporadic tumorigenesis as well, and therefore cause a major burden to public health. Several such alleles involved in CRC predisposition have been identified (Table 7). The high penetrance genes predisposing to CRC have been identified through linkage analysis, but when the penetrance of an allele is low, the locus is difficult to identify by linkage, because the slightly increased risk that a low-penetrance gene confers may not be sufficient to generate significant familial clustering. Additionally, many of the unaffected family members carry the disease allele, which confuses calculations. Thus, alternative methods to identify predisposing loci must be applied. We compared AI frequency in familial tumors to their sporadic counterparts, and identified several chromosomal regions that displayed higher frequency of AI in the familial group. The most promising regions after genotyping all the three sample sets were located in chromosome 20: 20p12.2 defined by D20S186, and 20q13.1-q13.3 defined by D20S178 and D20S171.

50

Table 7. Low-penetrance alleles predisposing to CRC (Laken et al., 1997; Houlston and Tomlinson, 2001; Gruber et al., 2002; Shaheen et al., 2003; Pashe et al., 2004) Gene Allele Population frequency Relative risk

of CRC TGFβR1 6Ala (Polyalanine polymorphism) Caucasian, 14% 1.2 APC I1307K Ashkenazi Jews, 5-7% 1.5-2.2 HRAS1 variable number of tandem repeats

(VNTR) Caucasian, 1-6% 2.5

BLM 2281 delATCTGA insTAGATTC Ashkenazi Jews, 1% 2.4 HFE H63D, C282Y USA, North Carolina,

9%, 4%, respectively 1.4

Comparative genomic hybridization showed that the mechanism creating AI was amplification. Despite that most cancer susceptibility genes act as tumor suppressor or caretaker genes, oncogenic predisposition can occur as well. In hereditary papillary renal cell cancer the mutated MET proto-oncogene is amplified in tumors (Fischer et al., 1998). In 2003 Ewart-Toland et al., proposed the role of AURKA/(STK15) as a low-penetrance CRC susceptibility gene. AURKA is located in chromosome 20q13. It is a member of Aurora/Ipl1p family of serine-threonine kinases which are important regulators of chromosome segregation (Bischoff and Plowman, 1999). Misregulation of this process results in aneuploidy, a characteristic feature of many cancers. Ewart-Toland et al. (2003) showed that an A91 (Ile31) allele of AURKA was preferentially amplified in their set of 162 colorectal tumors. Furthermore, they showed that tumors from individuals with even one copy of the AURKA A91 allele show a higher degree of aneuploidy than tumors from T91 homozygotes, and that in cultured cells an A allele induced more rapid cell growth. Besides CRC, the role of Ile31 of AURKA as a low-penetrance gene has since been proposed in breast, ovarian, and stomach cancer (Dicioccio et al., 2004; Miao et al., 2004; Sun et al., 2004). In the allelotyping study, no differences in 20q amplification were observed between familial and sporadic CRCs using quantitative PCR. The difference was, however, clear using microsatellites. Quantitative PCR can not separate between the two alleles, whereas microsatellite markers examine the status of alleles independently. This suggests that allele-specific amplification underlies the AI difference between familial and sporadic tumors. Allele-specific amplification of AURKA A91 has since been evaluated in 125 familial and 110 sporadic CRCs from our sample collection. Consistent with the study by Ewart-Toland et al. (2003), a preferential amplification of A91 allele was observed, and the allele specific amplification difference was more pronounced in the familial group (Hienonen et al., 2005). The preferential amplification of an A91 allele in the familial group suggests that the variant could confer a higher risk for CRC. Further studies are, however, needed to confirm the role of AURKA in colorectal tumor predisposition.

51

4. THE GENE EXPRESSION PROFILE OF THE SERRATED COLORECTAL TUMORS (IV) Two pathways that lead to CRC are well-established: MSS cancers evolve through the suppressor pathway and MSI cancers through the mutator pathway. Tumors arising through different pathways display mutations in different genes; APC, KRAS, and TP53 are frequently mutated in MSS tumors, while MSI tumors have mutations in TGFβRII and BRAF. However, tumors that do not follow either of these pathways exist. Tumors arising through a serrated pathway may represent a new entity among CRCs. Serrated lesions combine the architectural features of hyperplastic polyps and the cytological features of conventional adenocarcinomas. Thus, they look morphologically different, but whether they are biologically different as well has not been clear. Previous studies have tried to tie together MSI and the serrated pathway, suggesting that serrated polyps give rise to MSI tumors (Iino et al., 1999; Jass et al., 1999; Jass et al., 2000; Hawkins and Ward, 2001). On the contrary, in two recent studies serrated lesions have very rarely displayed MSI (Sawyer et al., 2002; Dong et al., 2005). Inconsistent results suggest that serrated CRCs form a heterogenous group of lesions and that among serrated tumors there are lesions that follow either the suppressor or the mutator pathway. In our dataset out of 45 serrated tumors 4 (8.9%) were MSI. Though it has fairly recently been proposed that serrated pathway exists, it has already been shown that serrated lesions have molecular features that distinguish them from conventional CRCs (Rashid et al., 2000; Sawyer et al., 2002; Kambara et al., 2004; Dong et al., 2005). The studies have, however, concentrated on few molecular features at a time; extensive genome-wide studies have not been performed so far. The difference between serrated and conventional adenocarcinomas was clearly present already in the unsupervised clustering. All the serrated samples, except for one, clustered together in the same branch (see Figure 1 in the study IV). Altogether 201 genes and 7 ESTs were differentially expressed between the two groups. Using the predictor consisting of 27 probes it was possible to predict an unknown sample with 100% accuracy (see Supplementary table 5 in the study IV). Since these probes were repeatedly present in cross-validations, it is likely that at least some of them play a role in serrated tumorigenesis. The array results were validated by immunohistochemical staining of six differentially expressed genes: EPHB2, PTCH, HIF1α, CCNT2, MTA1, and MBD4. Staining of EphB2, Patched, and HIF1α showed a statistically significant association with tumor morphology. EphB2 is a receptor tyrosine kinase which has been shown to be critical for the right patterning of intestinal crypt-villus axis (Batlle et al., 2002). In prostate cancer EPHB2 acts as a tumor suppressor gene (Huusko et al., 2004). Due to its decreased expression in serrated samples we sequenced 24 serrated tumors for mutations in EPHB2, but found no somatic mutations. In 63% (5/8) of serrated tumors EPHB2 promoter was hypermethylated, and 25% (3/12) displayed LOH at EPHB2 locus. It appears that EPHB2 is silenced by LOH and methylation. Progressive methylation has been observed in

52

serrated lesions (Dong et al., 2005). Similarly, loss of chromosome 1p is observed already in hyperplastic polyps (Rashid et al., 2000). Patched (Ptch) is a receptor in the Hedgehog (HH) signaling pathway. Correct HH signaling is crucial during development of gastrointestinal tract, as well as in maintaining stem cells (Oldak et al., 2001). In the absence of the HH signal Ptch represses Smoothened (Smo). Binding of a HH protein to Ptch releases the repression of Smo which transmits the HH signal through Gli proteins. Inactivation of Ptch is equivalent to constant HH signaling. Patients affected with Gorlin syndrome have a germline mutation in PTCH. The syndrome is characterized by early onset multiple basal cell carcinomas and higher rate of other tumors e.g. medulloblastomas (Hahn et al., 1996; Johnson et al., 1996). Loss of PTCH have been reported in sporadic forms of the same tumor types (Taipale and Bleachy, 2001), but not in primary CRCs. Disruption of the HH signaling pathway in serrated tumors may represent a novel route to CRC. HIF1α is expressed in lack of oxygen. HIF1α is a transcription factor which is able to induce transcription of genes needed in hypoxia (Ratcliffe et al., 1998). Some tumors displayed focal HIF1α positivity, while others showed diffuse positivity throughout the tumor. HIF1α expression was higher in serrated tumors, and serrated tumors displayed more often diffuse staining pattern in immunohistochemistry. Cyclin T2 showed statistically significant difference in the expression pattern in the training set which confirms the expression difference for the respective samples on array. The result did not reach statistically significant level in the validation set, thus lower expression of Cyclin T2 can not be generalized as a marker for serrated tumors. Cyclins are proteins that act together with cyclin-dependent kinases (CDKs) in the cell cycle regulation. Cyclin T2 is able to phosphorylate RNA polymerase II, and regulate transcriptional elongation of RNA by forming a complex with CDK9 (De Luca et al., 2002). Both MTA1 and MBD4 showed significant expression difference in arrays. Despite that neither of the genes showed correlation between cancer types in immunohistochemistry. However, changes in RNA level do not necessarily reflect changes in protein level. It is also possible that the change in the protein level is too small to be detected in immunohistochemistry. In all, the results of the immunohistochemistry support the array results. This study shows that serrated CRCs differ from conventional CRCs not only morphologically but at the molecular level as well. This work has brought out some potential candidates playing a role in serrated tumorigenesis and further studies should establish the mechanisms and key molecules underlying serrated CRCs.

53

SUMMARY AND CONCLUSIONS In the current study several molecular methods to classify colorectal cancer were studied. The practicality of the Bethesda microsatellite marker panel in hereditary CRC diagnostics was compared to mononucleotide marker BAT26, the role and extent of low-level MSI was analyzed, new chromosomal regions containing putative low-predisposition alleles were localized, and molecular differences between serrated and conventional CRCs were studied. From the results of the study the following conclusions were made:

1. BAT26 was found to be a sensitive tool in Lynch syndrome diagnostics. It identified all mutation positive samples in the dataset, but failed to identify two MSI-H and 20 MSI-L samples. These samples did not display known MMR defects. Analysis of additional four markers added workload and costs, and increased difficulties in MSI scoring but did not increase sensitivity. BAT26 alone can be used in Lynch syndrome diagnostics especially if resources are limited.

2. When a large number of microsatellites are studied, most CRCs display low-level

MSI. The commonly used 5-15 markers are inadequate to define MSI-L tumors, since they fail to detect majority of lesions displaying similar degree of MSI-L. Neither clinical nor molecular differences were observed between MSI-L and MSS CRCs. However, the distribution of microsatellite mutations was not random; older patients tended to have more mutations. MSI-L and MSS tumors appear to have a common molecular background, and the observed instability difference reflects number of cell divisions and stochastic mutation.

3. Allelotyping of familial and sporadic CRCs revealed several chromosomal

regions that potentially harbor novel CRC predisposition genes. The most promising region was located on chromosome 20q where familial tumors showed more AI compared to sporadic ones. The mechanism creating AI was shown to be amplification. Genome-wide AI analysis provides an alternative strategy to identify cancer predisposition genes, since it obviates the requirement to collect multiple samples from families.

4. Serrated CRCs differed from conventional CRCs on gene expression level and

formed a distinct branch in the hierarchical clustering tree. Several differentially expressed genes were identified, and the differences were shown at protein level as well. The study supports the existence of serrated neoplasia pathway, and shows that serrated tumors are different from conventional CRCs not only morphologically but also molecularly.

54

ACKNOWLEDGEMENTS This study was carried out at the Department of Medical Genetics, at Haartman Institute and Biomedicum Helsinki, University of Helsinki during 2000-2005. I sincerely want to thank all the people who have been involved: My supervisor, Lauri Aaltonen, for his guidance, encouragement and support during the project. Besides that he is a skilled and enthusiastic scientist I admire his talent as a leader. As a boss he is fair, understanding, and extremely nice, yet not shiftless. Lauri’s sociable personality and a great sense of humor are the basis for a fun and relaxed atmosphere in the lab. It has been a privilege to work in his group. My other supervisor Virpi Launonen for all help during these years. During the first years, when the world of genetics seemed sometimes little blurred it was nice to have someone to turn into with thousands of questions. Thank you, also, for being a friend when one was needed. The Aaltonen (in some occasions also Kere) group: My room-mates, Sini, Sakari, and Rainer for understanding that to be able to work, music is absolutely needed. Sini for all things from dealing with samples to sending FedEx packages, and the most important of all: the numerous gossip-moments, dancing company, and friendship, Rainer for help in all computer-associated things, as well as being a soul of every party. My fellow Ph.D students Maija, Susa (already postdocs), Tuija, Sanna, both Helis, Pia, Antti, and Marianna. Special thanks to Tuija for the moments spent staring at the weird-looking microsatellite graphs in the AI project, Antti for taking care of almost all of the array data analysis in the serrated project when I wasn’t much of a help, and Sanna for excellent riding company. The postdocs Auli, Diego, and Pia, Auli for kicking people’s butts by giving sudden fillips, and Diego for his expertise in the array field and adding some Spanish color to our group, the lab team Anniina, Iina, Inga-Lill, and Mikko for superb technical assistance, Anniina especially for the endlessly positive and joyful personality, as well as friendship, Reijo for pathology expertise, and last but not least the former members of our group; Stina for guiding me through my first steps in the lab, Anu and Kirsi for showing an exemplary party mentality, and Siv, Sinikka, Taija, Annika, and Nina. The former and present heads of the Department of Medical Genetics: Pertti Aula, Anna-Elina Lehesjoki, Leena Palotie, Kristiina Aittomäki, and Päivi Peltomäki for providing excellent working facilities. Heikki Järvinen and Jukka-Pekka Mecklin for initiating the HNPCC research in Finland and providing a continuous flow of tumor samples. The collaborators and co-authors Richard Houlston, Ian Tomlinson, Darryl Shibata, Torben Ørntoft, Manel Esteller, Markus Mäkinen, Sakari Knuutila, Katja Eklin,

55

Mingzhou Guo, James Herman, Pertti Sistonen, Kyoung-Mee Kim, Lara Lipton, Yan Aalto, Huw Thomas, Karin Birkenkamp-Demtröder, Shirley Hodgson, Sarah Halford, Tuomo Karttunen, and Karoliina Tuppurainen. Doctors Robert Hofstra and Jorma Isola, the pre-examiners of this thesis, for valuable comments, Anna-Elina Lehesjoki, and Aarno Palotie for the thesis committee. Janna Saarela, Outi Monni and the whole Biomedicum Biochip Center crew for help with all kind of array-related issues. Susanna Anjala, Heli Keränen, Pekka Ellonen, Paula Kristo, and Elvi Karila for sequencing and fragment analysis. Third floor girls Saila and Katja for help with sample associated matters and Miina for nice chitchat moments. Minna Maunula, Ilpo Vilhunen, Pirjo Koljonen, and Merja Voutilainen for help in practical matters. My dear friends Annika, Eeva, Keiju, and Heidi for all kind of free-time activities. Nothing adjusts one’s brain to zero better than a proper girls’ night. Unfortunately they have become more middle-age –like lately… Special thanks to Eeva for the years as flat-mates in Lauttasaari. My fellow students Virpi, Inari, Tero, Tiina, Anu, Panu, Santtu, and Jouni for friendship that lasted beyond lecture halls. Thanks for sharing the anxiety when life as a biologist (or life in general) did not always turn out to be what we as freshmen expected ☺. Virpi especially for the time as flat-mates as well as being an excellent climbing tutor. My parents Marjatta and Yrjö, brother Pekka and sister Elina for support and keeping my feet on the ground. Perttu for his love and care, and sharing both good and bad days with me. Finally, I would like to express my gratitude to all the cancer patients who have volunteered to participate in this study. This study was financially supported by the Helsinki Graduate School in Biotechnology and Molecular Biology, the Finnish Cultural Foundation, Emil Aaltonen Foundation, Ida Montin Foundation, Maud Kuistila Foundation, Jalmari and Rauha Ahokas Foundation, Finnish Cancer Society, the Finnish Oncology Foundation, the Research and Science Foundation of Farmos, and the Finnish-Norwegian Medical Foundation.

56

REFERENCES

Aaltonen LA, Peltomaki P, Leach FS, Sistonen P, Pylkkanen L, Mecklin JP, Jarvinen H, Powell SM, Jen J and Hamilton SR. Clues to the pathogenesis of familial colorectal cancer. Science 260: 812-816, 1993.

Aaltonen LA, Peltomaki P, Mecklin JP, Jarvinen H, Jass JR, Green JS, Lynch HT, Watson P, Tallqvist G and Juhola M. Replication errors in benign and malignant tumors from hereditary nonpolyposis colorectal cancer patients. Cancer Res. 54: 1645-1648, 1994.

Aaltonen LA, Salovaara R, Kristo P, Canzian F, Hemminki A, Peltomaki P, Chadwick RB, Kaariainen H, Eskelinen M, Jarvinen H, Mecklin JP and de la Chapelle A. Incidence of hereditary nonpolyposis colorectal cancer and the feasibility of molecular screening for the disease. N.Engl.J.Med. 338: 1481-1487, 1998.

Aarnio M, Mecklin JP, Aaltonen LA, Nystrom-Lahti M and Jarvinen HJ. Life-time risk of different cancers in hereditary non-polyposis colorectal cancer (HNPCC) syndrome. Int.J.Cancer 64: 430-433, 1995.

Acharya S, Wilson T, Gradia S, Kane MF, Guerrette S, Marsischky GT, Kolodner R and Fishel R. hMSH2 forms specific mispair-binding complexes with hMSH3 and hMSH6. Proc.Natl.Acad.Sci.U.S.A. 93: 13629-13634, 1996.

Akiyama Y, Sato H, Yamada T, Nagasaki H, Tsuchiya A, Abe R and Yuasa Y. Germ-line mutation of the hMSH6/GTBP gene in an atypical hereditary nonpolyposis colorectal cancer kindred. Cancer Res. 57: 3920-3923, 1997.

Al-Tassan N, Chmiel NH, Maynard J, Fleming N, Livingston AL, Williams GT, Hodges AK, Davies DR, David SS, Sampson JR and Cheadle JP. Inherited variants of MYH associated with somatic G:C-->T:A mutations in colorectal tumors. Nat.Genet. 30: 227-232, 2002.

Baker SJ, Preisinger AC, Jessup JM, Paraskeva C, Markowitz S, Willson JK, Hamilton S and Vogelstein B. p53 gene mutations occur in combination with 17p allelic deletions as late events in colorectal tumorigenesis. Cancer Res. 50: 7717-7722, 1990.

Batlle E, Henderson JT, Beghtel H, van den Born MM, Sancho E, Huls G, Meeldijk J, Robertson J, van de Wetering M, Pawson T and Clevers H. Beta-catenin and TCF mediate cell positioning in the intestinal epithelium by controlling the expression of EphB/ephrinB. Cell 111: 251-263, 2002.

Baxter SW, Choong DY, Eccles DM and Campbell IG. Transforming growth factor beta receptor 1 polyalanine polymorphism and exon 5 mutation analysis in breast and ovarian cancer. Cancer Epidemiol.Biomarkers Prev. 11: 211-214, 2002.

Baylin SB and Herman JG. DNA hypermethylation in tumorigenesis: epigenetics joins genetics. Trends Genet. 16: 168-174, 2000.

Bhattacharyya NP, Skandalis A, Ganesh A, Groden J and Meuth M. Mutator phenotypes in human colorectal carcinoma cell lines. Proc.Natl.Acad.Sci.U.S.A. 91: 6319-6323, 1994.

Biden KG, Simms LA, Cummings M, Buttenshaw R, Schoch E, Searle J, Gobe G, Jass JR, Meltzer SJ, Leggett BA and Young J. Expression of Bcl-2 protein is decreased in colorectal adenocarcinomas with microsatellite instability. Oncogene 18: 1245-1249, 1999.

Bischoff JR and Plowman GD. The Aurora/Ipl1p kinase family: regulators of chromosome segregation and cytokinesis. Trends Cell Biol. 9: 454-459, 1999.

Bishop JM. Molecular themes in oncogenesis. Cell 64: 235-248, 1991.

57

Boland CR, Sato J, Saito K, Carethers JM, Marra G, Laghi L and Chauhan DP. Genetic instability and chromosomal aberrations in colorectal cancer: a review of the current models. Cancer Detect.Prev. 22: 377-382, 1998.

Boland CR, Thibodeau SN, Hamilton SR, Sidransky D, Eshleman JR, Burt RW, Meltzer SJ, Rodriguez-Bigas MA, Fodde R, Ranzani GN and Srivastava S. A National Cancer Institute Workshop on Microsatellite Instability for cancer detection and familial predisposition: development of international criteria for the determination of microsatellite instability in colorectal cancer. Cancer Res. 58: 5248-5257, 1998.

Bos JL. The ras gene family and human carcinogenesis. Mutat.Res. 195: 255-271, 1988.

Boyer JC and Farber RA. Mutation rate of a microsatellite sequence in normal human fibroblasts. Cancer Res. 58: 3946-3949, 1998.

Boyle P and Leon ME. Epidemiology of colorectal cancer. Br.Med.Bull. 64: 1-25, 2002.

Bronner CE, Baker SM, Morrison PT, Warren G, Smith LG, Lescoe MK, Kane M, Earabino C, Lipford J and Lindblom A. Mutation in the DNA mismatch repair gene homologue hMLH1 is associated with hereditary non-polyposis colon cancer. Nature 368: 258-261, 1994.

Buermeyer AB, Deschenes SM, Baker SM, Liskay RM. Mammalian DNA mismatch repair. Annu. Rev. Genet. 33: 533-564, 1999.

Burdett V, Baitinger C, Viswanathan M, Lovett ST and Modrich P. In vivo requirement for RecJ, ExoVII, ExoI, and ExoX in methyl-directed mismatch repair. Proc.Natl.Acad.Sci.U.S.A. 98: 6765-6770, 2001.

Burks RT, Kessis TD, Cho KR and Hedrick L. Microsatellite instability in endometrial carcinoma. Oncogene 9: 1163-1166, 1994.

Calin GA, Gafa R, Tibiletti MG, Herlea V, Becheanu G, Cavazzini L, Barbanti-Brodano G, Nenci I, Negrini M and Lanza G. Genetic progression in microsatellite instability high (MSI-H) colon cancers correlates with clinico-pathological parameters: A study of the TGRbetaRII, BAX, hMSH3, hMSH6, IGFIIR and BLM genes. Int.J.Cancer 89: 230-235, 2000.

Calin GA, Gafa R, Tibiletti MG, Herlea V, Becheanu G, Cavazzini L, Barbanti-Brodano G, Nenci I, Negrini M and Lanza G. Genetic progression in microsatellite instability high (MSI-H) colon cancers correlates with clinico-pathological parameters: A study of the TGRbetaRII, BAX, hMSH3, hMSH6, IGFIIR and BLM genes. Int.J.Cancer 89: 230-235, 2000.

Canzian F, Salovaara R, Hemminki A, Kristo P, Chadwick RB, Aaltonen LA and de la Chapelle A. Semiautomated assessment of loss of heterozygosity and replication error in tumors. Cancer Res. 56: 3331-3337, 1996.

Carder P, Wyllie AH, Purdie CA, Morris RG, White S, Piris J and Bird CC. Stabilised p53 facilitates aneuploid clonal divergence in colorectal cancer. Oncogene 8: 1397-1401, 1993.

Caron de Fromentel C and Soussi T. TP53 tumor suppressor gene: a model for investigating human mutagenesis. Genes Chromosomes Cancer 4: 1-15, 1992.

Cavenee WK, Dryja TP, Phillips RA, Benedict WF, Godbout R, Gallie BL, Murphree AL, Strong LC and White RL. Expression of recessive alleles by chromosomal mechanisms in retinoblastoma. Nature 305: 779-784, 1983.

Cawkwell L, Li D, Lewis FA, Martin I, Dixon MF and Quirke P. Microsatellite instability in colorectal cancer: improved assessment using fluorescent polymerase chain reaction. Gastroenterology 109: 465-471, 1995.

58

Chadwick RB, Pyatt RE, Niemann TH, Richards SK, Johnson CK, Stevens MW, Meek JE, Hampel H, Prior TW and de la Chapelle A. Hereditary and somatic DNA mismatch repair gene mutations in sporadic endometrial carcinoma. J.Med.Genet. 38: 461-466, 2001.

Chapusot C, Martin L, Puig PL, Ponnelle T, Cheynel N, Bouvier AM, Rageot D, Roignot P, Rat P, Faivre J and Piard F. What is the best way to assess microsatellite instability status in colorectal cancer? Study on a population base of 462 colorectal cancers. Am.J.Surg.Pathol. 28: 1553-1559, 2004.

Chen T, de Vries EG, Hollema H, Yegen HA, Vellucci VF, Strickler HD, Hildesheim A and Reiss M. Structural alterations of transforming growth factor-beta receptor genes in human cervical carcinoma. Int.J.Cancer 82: 43-51, 1999.

Cho KR, Oliner JD, Simons JW, Hedrick L, Fearon ER, Preisinger AC, Hedge P, Silverman GA and Vogelstein B. The DCC gene: structural analysis and mutations in colorectal carcinomas. Genomics 19: 525-531, 1994.

Cui H, Cruz-Correa M, Giardiello FM, Hutcheon DF, Kafonek DR, Brandenburg S, Wu Y, He X, Powe NR and Feinberg AP. Loss of IGF2 imprinting: a potential marker of colorectal cancer risk. Science 299: 1753-1755, 2003.

Cunningham JM, Christensen ER, Tester DJ, Kim CY, Roche PC, Burgart LJ and Thibodeau SN. Hypermethylation of the hMLH1 promoter in colon cancer with microsatellite instability. Cancer Res. 58: 3455-3460, 1998.

Cunningham JM, Kim CY, Christensen ER, Tester DJ, Parc Y, Burgart LJ, Halling KC, McDonnell SK, Schaid DJ, Walsh Vockley C, Kubly V, Nelson H, Michels VV and Thibodeau SN. The frequency of hereditary defective mismatch repair in a prospective series of unselected colorectal carcinomas. Am.J.Hum.Genet. 69: 780-790, 2001.

D'Atri S, Tentori L, Lacal PM, Graziani G, Pagani E, Benincasa E, Zambruno G, Bonmassar E and Jiricny J. Involvement of the mismatch repair system in temozolomide-induced apoptosis. Mol.Pharmacol. 54: 334-341, 1998.

de la Chapelle A. Genetic predisposition to colorectal cancer. Nat.Rev.Cancer. 4: 769-780, 2004.

De Luca A, De Falco M, Baldi A and Paggi MG. Cyclin T: Three forms for different roles in physiological and pathological functions J Cell Physiol 194: 101-107, 2002

de Vries SS, Baart EB, Dekker M, Siezen A, de Rooij DG, de Boer P and te Riele H. Mouse MutS-like protein Msh5 is required for proper chromosome synapsis in male and female meiosis. Genes Dev. 13: 523-531, 1999.

Dennis PA and Rifkin DB. Cellular activation of latent transforming growth factor beta requires binding to the cation-independent mannose 6-phosphate/insulin-like growth factor type II receptor. Proc.Natl.Acad.Sci.U.S.A. 88: 580-584, 1991.

Dicioccio RA, Song H, Waterfall C, Kimura MT, Nagase H, McGuire V, Hogdall E, Shah MN, Luben RN, Easton DF, Jacobs IJ, Ponder BA, Whittemore AS, Gayther SA, Pharoah PD and Kruger-Kjaer S. STK15 polymorphisms and association with risk of invasive ovarian cancer. Cancer Epidemiol.Biomarkers Prev. 13: 1589-1594, 2004.

Dietmaier W, Wallinger S, Bocker T, Kullmann F, Fishel R and Ruschoff J. Diagnostic microsatellite instability: definition and correlation with mismatch repair protein expression. Cancer Res. 57: 4749-4756, 1997.

Domingo E, Laiho P, Ollikainen M, Pinto M, Wang L, French AJ, Westra J, Frebourg T, Espin E, Armengol M, Hamelin R, Yamamoto H, Hofstra RM, Seruca R, Lindblom A, Peltomaki P, Thibodeau SN,

59

Aaltonen LA and Schwartz S,Jr. BRAF screening as a low-cost effective strategy for simplifying HNPCC genetic testing. J.Med.Genet. 41: 664-668, 2004.

Donehower LA and Bradley A. The tumor suppressor p53. Biochim.Biophys.Acta 1155: 181-205, 1993.

Dong C and Hemminki K. Modification of cancer risks in offspring by sibling and parental cancers from 2,112,616 nuclear families. Int.J.Cancer 92: 144-150, 2001.

Dong SM, Lee EJ, Jeon ES, Park CK and Kim KM. Progressive methylation during the serrated neoplasia pathway of the colorectum. Mod.Pathol. 18: 170-178, 2005.

Douc-Rasy S, Barrois M, Fogel S, Ahomadegbe JC, Stehelin D, Coll J and Riou G. High incidence of loss of heterozygosity and abnormal imprinting of H19 and IGF2 genes in invasive cervical carcinomas. Uncoupling of H19 and IGF2 expression and biallelic hypomethylation of H19. Oncogene 12: 423-430, 1996.

Duval A and Hamelin R. Mutations at coding repeat sequences in mismatch repair-deficient human cancers: toward a new concept of target genes for instability. Cancer Res. 62: 2447-2454, 2002.

Edelmann W, Cohen PE, Kneitz B, Winand N, Lia M, Heyer J, Kolodner R, Pollard JW and Kucherlapati R. Mammalian MutS homologue 5 is required for chromosome pairing in meiosis. Nat.Genet. 21: 123-127, 1999.

Ellis MJ, Leav BA, Yang Z, Rasmussen A, Pearce A, Zweibel JA, Lippman ME and Cullen KJ. Affinity for the insulin-like growth factor-II (IGF-II) receptor inhibits autocrine IGF-II activity in MCF-7 breast cancer cells. Mol.Endocrinol. 10: 286-297, 1996.

Enholm S, Hienonen T, Suomalainen A, Lipton L, Tomlinson I, Karja V, Eskelinen M, Mecklin JP, Karhu A, Jarvinen HJ and Aaltonen LA. Proportion and phenotype of MYH-associated colorectal neoplasia in a population-based series of Finnish colorectal cancer patients. Am.J.Pathol. 163: 827-832, 2003.

Esteller M, Hamilton SR, Burger PC, Baylin SB and Herman JG. Inactivation of the DNA repair gene O6-methylguanine-DNA methyltransferase by promoter hypermethylation is a common event in primary human neoplasia. Cancer Res. 59: 793-797, 1999.

Esteller M, Toyota M, Sanchez-Cespedes M, Capella G, Peinado MA, Watkins DN, Issa JP, Sidransky D, Baylin SB and Herman JG. Inactivation of the DNA repair gene O6-methylguanine-DNA methyltransferase by promoter hypermethylation is associated with G to A mutations in K-ras in colorectal tumorigenesis. Cancer Res. 60: 2368-2371, 2000.

Ewart-Toland A, Briassouli P, de Koning JP, Mao JH, Yuan J, Chan F, MacCarthy-Morrogh L, Ponder BA, Nagase H, Burn J, Ball S, Almeida M, Linardopoulos S and Balmain A. Identification of Stk6/STK15 as a candidate low-penetrance tumor-susceptibility gene in mouse and human. Nat.Genet. 34: 403-412, 2003.

Fazeli A, Dickinson SL, Hermiston ML, Tighe RV, Steen RG, Small CG, Stoeckli ET, Keino-Masu K, Masu M, Rayburn H, Simons J, Bronson RT, Gordon JI, Tessier-Lavigne M and Weinberg RA. Phenotype of mice lacking functional Deleted in colorectal cancer (Dcc) gene. Nature 386: 796-804, 1997.

Fearon ER and Vogelstein B. A genetic model for colorectal tumorigenesis. Cell 61: 759-767, 1990.

Feinberg AP and Vogelstein B. Hypomethylation distinguishes genes of some human cancers from their normal counterparts. Nature 301: 89-92, 1983.

Feinberg AP. Genomic imprinting and cancer. In: The genetic basis of human cancer. Eds. Vogelstein B, and Kinzler KW. McGraw-Hill Companies Inc., USA, 43-55, 2002.

Fink D, Aebi S and Howell SB. The role of DNA mismatch repair in drug resistance. Clin.Cancer Res. 4: 1-6, 1998.

60

Fischer J, Palmedo G, von Knobloch R, Bugert P, Prayer-Galetti T, Pagano F and Kovacs G. Duplication and overexpression of the mutant allele of the MET proto-oncogene in multiple hereditary papillary renal cell tumours. Oncogene 17: 733-739, 1998.

Fishel R, Lescoe MK, Rao MR, Copeland NG, Jenkins NA, Garber J, Kane M and Kolodner R. The human mutator gene homolog MSH2 and its association with hereditary nonpolyposis colon cancer. Cell 75: 1027-1038, 1993.

Futreal PA, Coin L, Marshall M, Down T, Hubbard T, Wooster R, Rahman N and Stratton MR. A census of human cancer genes. Nat.Rev.Cancer. 4: 177-183, 2004.

Gardner EJ and Richards RC. Multiple cutaneous and subcutaneous lesions occurring simultaneously with hereditary polyposis and osteomatosis. Am.J.Hum.Genet. 5: 139-147, 1953.

Giardiello FM, Brensinger JD, Tersmette AC, Goodman SN, Petersen GM, Booker SV, Cruz-Correa M and Offerhaus JA. Very high risk of cancer in familial Peutz-Jeghers syndrome. Gastroenterology 119: 1447-1453, 2000.

Giles RH, van Es JH and Clevers H. Caught up in a Wnt storm: Wnt signaling in cancer. Biochim.Biophys.Acta 1653: 1-24, 2003.

Ginzinger DG, Godfrey TE, Nigro J, Moore DH,2nd, Suzuki S, Pallavicini MG, Gray JW and Jensen RH. Measurement of DNA copy number at microsatellite loci using quantitative PCR analysis. Cancer Res. 60: 5405-5409, 2000.

Giovannucci E. An updated review of the epidemiological evidence that cigarette smoking increases risk of colorectal cancer. Cancer Epidemiol.Biomarkers Prev. 10: 725-731, 2001.

Giovannucci E, Colditz GA, Stampfer MJ and Willett WC. Physical activity, obesity, and risk of colorectal adenoma in women (United States). Cancer Causes Control 7: 253-263, 1996.

Goldgar DE, Easton DF, Cannon-Albright LA and Skolnick MH. Systematic population-based assessment of cancer risk in first-degree relatives of cancer probands. J.Natl.Cancer Inst. 86: 1600-1608, 1994.

Gonzalez-Garcia I, Moreno V, Navarro M, Marti-Rague J, Marcuello E, Benasco C, Campos O, Capella G and Peinado MA. Standardized approach for microsatellite instability detection in colorectal carcinomas. J.Natl.Cancer Inst. 92: 544-549, 2000.

Grady WM, Rajput A, Myeroff L, Liu DF, Kwon K, Willis J and Markowitz S. Mutation of the type II transforming growth factor-beta receptor is coincident with the transformation of human colon adenomas to malignant carcinomas. Cancer Res. 58: 3101-3104, 1998.

Groden J, Thliveris A, Samowitz W, Carlson M, Gelbert L, Albertsen H, Joslyn G, Stevens J, Spirio L and Robertson M. Identification and characterization of the familial adenomatous polyposis coli gene. Cell 66: 589-600, 1991.

Gruber SB, Ellis NA, Scott KK, Almog R, Kolachana P, Bonner JD, Kirchhoff T, Tomsho LP, Nafa K, Pierce H, Low M, Satagopan J, Rennert H, Huang H, Greenson JK, Groden J, Rapaport B, Shia J, Johnson S, Gregersen PK, Harris CC, Boyd J, Rennert G and Offit K. BLM heterozygosity and the risk of colorectal cancer. Science 297: 2013, 2002.

Gryfe R, Di Nicola N, Lal G, Gallinger S and Redston M. Inherited colorectal polyposis and cancer risk of the APC I1307K polymorphism. Am.J.Hum.Genet. 64: 378-384, 1999.

Gryfe R, Kim H, Hsieh ET, Aronson MD, Holowaty EJ, Bull SB, Redston M and Gallinger S. Tumor microsatellite instability and clinical outcome in young patients with colorectal cancer. N.Engl.J.Med. 342: 69-77, 2000.

61

Hahn H, Wicking C, Zaphiropoulous PG, Gailani MR, Shanley S, Chidambaram A, Vorechovsky I, Holmberg E, Unden AB, Gillies S, Negus K, Smyth I, Pressman C, Leffell DJ, Gerrard B, Goldstein AM, Dean M, Toftgard R, Chenevix-Trench G, Wainwright B and Bale AE. Mutations of the human homolog of Drosophila patched in the nevoid basal cell carcinoma syndrome. Cell 85: 841-851, 1996.

Halford S, Sasieni P, Rowan A, Wasan H, Bodmer W, Talbot I, Hawkins N, Ward R and Tomlinson I. Low-level microsatellite instability occurs in most colorectal cancers and is a nonrandomly distributed quantitative trait. Cancer Res. 62: 53-57, 2002.

Hall MC and Matson SW. The Escherichia coli MutL protein physically interacts with MutH and stimulates the MutH-associated endonuclease activity. J.Biol.Chem. 274: 1306-1312, 1999.

Hall NR, Williams MA, Murday VA, Newton JA and Bishop DT. Muir-Torre syndrome: a variant of the cancer family syndrome. J.Med.Genet. 31: 627-631, 1994.

Halvarsson B, Lindblom A, Rambech E, Lagerstedt K and Nilbert M. Microsatellite instability analysis and/or immunostaining for the diagnosis of hereditary nonpolyposis colorectal cancer? Virchows Arch. 444: 135-141, 2004.

Hamilton SR, Liu B, Parsons RE, Papadopoulos N, Jen J, Powell SM, Krush AJ, Berk T, Cohen Z and Tetu B. The molecular basis of Turcot's syndrome. N.Engl.J.Med. 332: 839-847, 1995.

Han HJ, Yanagisawa A, Kato Y, Park JG and Nakamura Y. Genetic instability in pancreatic cancer and poorly differentiated type of gastric cancer. Cancer Res. 53: 5087-5089, 1993.

Hanahan D and Weinberg RA. The hallmarks of cancer. Cell 100: 57-70, 2000.

Hawkins NJ, Bariol C and Ward RL. The serrated neoplasia pathway. Pathology 34: 548-555, 2002.

Hawkins NJ and Ward RL. Sporadic colorectal cancers with microsatellite instability and their possible origin in hyperplastic polyps and serrated adenomas. J.Natl.Cancer Inst. 93: 1307-1313, 2001.

Hawley SA, Boudeau J, Reid JL, Mustard KJ, Udd L, Makela TP, Alessi DR and Hardie DG. Complexes between the LKB1 tumor suppressor, STRAD alpha/beta and MO25 alpha/beta are upstream kinases in the AMP-activated protein kinase cascade. J.Biol. 2: 28, 2003.

He TC, Sparks AB, Rago C, Hermeking H, Zawel L, da Costa LT, Morin PJ, Vogelstein B and Kinzler KW. Identification of c-MYC as a target of the APC pathway. Science 281: 1509-1512, 1998.

Heldin CH, Miyazono K and ten Dijke P. TGF-beta signalling from cell membrane to nucleus through SMAD proteins. Nature 390: 465-471, 1997.

Hemminki A, Markie D, Tomlinson I, Avizienyte E, Roth S, Loukola A, Bignell G, Warren W, Aminoff M, Hoglund P, Jarvinen H, Kristo P, Pelin K, Ridanpaa M, Salovaara R, Toro T, Bodmer W, Olschwang S, Olsen AS, Stratton MR, de la Chapelle A and Aaltonen LA. A serine/threonine kinase gene defective in Peutz-Jeghers syndrome. Nature 391: 184-187, 1998.

Herman JG, Umar A, Polyak K, Graff JR, Ahuja N, Issa JP, Markowitz S, Willson JK, Hamilton SR, Kinzler KW, Kane MF, Kolodner RD, Vogelstein B, Kunkel TA and Baylin SB. Incidence and functional consequences of hMLH1 promoter hypermethylation in colorectal carcinoma. Proc.Natl.Acad.Sci.U.S.A. 95: 6870-6875, 1998.

Hickman MJ and Samson LD. Role of DNA mismatch repair and p53 in signaling induction of apoptosis by alkylating agents. Proc.Natl.Acad.Sci.U.S.A. 96: 10764-10769, 1999.

Hienonen T, Laiho P, Salovaara R, Mecklin JP, Jarvinen H, Sistonen P, Peltomaki P, Lehtonen R, Nupponen NN, Launonen V, Karhu A and Aaltonen LA. Little evidence for involvement of MLH3 in colorectal cancer predisposition. Int.J.Cancer 106: 292-296, 2003.

62

Hienonen T, Salovaara R, Mecklin J-P, Järvinen H, Karhu A and Aaltonen LA. Preferential amplification of AURKA 91A (Ile31) in familial colorectal cancers. Int. J. Cancer, 2005, In press.

Hirose Y, Yoshimi N, Suzui M, Kawabata K, Tanaka T and Mori H. Expression of bcl-2, bax, and bcl-XL proteins in azoxymethane-induced rat colonic adenocarcinomas. Mol.Carcinog. 19: 25-30, 1997.

Hite JM, Eckert KA and Cheng KC. Factors affecting fidelity of DNA synthesis during PCR amplification of d(C-A)n.d(G-T)n microsatellite repeats. Nucleic Acids Res. 24: 2429-2434, 1996.

Houlston RS and Tomlinson IP. Polymorphisms and colorectal tumor risk. Gastroenterology 121: 282-301, 2001.

Howe JR, Bair JL, Sayed MG, Anderson ME, Mitros FA, Petersen GM, Velculescu VE, Traverso G and Vogelstein B. Germline mutations of the gene encoding bone morphogenetic protein receptor 1A in juvenile polyposis. Nat.Genet. 28: 184-187, 2001.

Howe JR, Roth S, Ringold JC, Summers RW, Jarvinen HJ, Sistonen P, Tomlinson IP, Houlston RS, Bevan S, Mitros FA, Stone EM and Aaltonen LA. Mutations in the SMAD4/DPC4 gene in juvenile polyposis. Science 280: 1086-1088, 1998.

Huusko P, Ponciano-Jackson D, Wolf M, Kiefer JA, Azorsa DO, Tuzmen S, Weaver D, Robbins C, Moses T, Allinen M, Hautaniemi S, Chen Y, Elkahloun A, Basik M, Bova GS, Bubendorf L, Lugli A, Sauter G, Schleutker J, Ozcelik H, Elowe S, Pawson T, Trent JM, Carpten JD, Kallioniemi OP and Mousses S. Nonsense-mediated decay microarray analysis identifies mutations of EPHB2 in human prostate cancer. Nat.Genet. 36: 979-983, 2004.

Iino H, Jass JR, Simms LA, Young J, Leggett B, Ajioka Y and Watanabe H. DNA microsatellite instability in hyperplastic polyps, serrated adenomas, and mixed polyps: a mild mutator pathway for colorectal cancer? J.Clin.Pathol. 52: 5-9, 1999.

Ionov Y, Peinado MA, Malkhosyan S, Shibata D and Perucho M. Ubiquitous somatic mutations in simple repeated sequences reveal a new mechanism for colonic carcinogenesis. Nature 363: 558-561, 1993.

Jagmohan-Changur S, Poikonen T, Vilkki S, Launonen V, Wikman F, Orntoft TF, Moller P, Vasen H, Tops C, Kolodner RD, Mecklin JP, Jarvinen H, Bevan S, Houlston RS, Aaltonen LA, Fodde R, Wijnen J and Karhu A. EXO1 variants occur commonly in normal population: evidence against a role in hereditary nonpolyposis colorectal cancer. Cancer Res. 63: 154-158, 2003.

Järvinen H. Juvenile gastrointestinal polyposis. Problems in general surgery 10: 749-57, 1993

Jarvinen H and Franssila KO. Familial juvenile polyposis coli; increased risk of colorectal cancer. Gut 25: 792-800, 1984.

Jarvinen HJ, Aarnio M, Mustonen H, Aktan-Collan K, Aaltonen LA, Peltomaki P, De La Chapelle A and Mecklin JP. Controlled 15-year trial on screening for colorectal cancer in families with hereditary nonpolyposis colorectal cancer. Gastroenterology 118: 829-834, 2000.

Jass JR, Biden KG, Cummings MC, Simms LA, Walsh M, Schoch E, Meltzer SJ, Wright C, Searle J, Young J and Leggett BA. Characterisation of a subtype of colorectal cancer combining features of the suppressor and mild mutator pathways. J.Clin.Pathol. 52: 455-460, 1999.

Jass JR, Iino H, Ruszkiewicz A, Painter D, Solomon MJ, Koorey DJ, Cohn D, Furlong KL, Walsh MD, Palazzo J, Edmonston TB, Fishel R, Young J and Leggett BA. Neoplastic progression occurs through mutator pathways in hyperplastic polyposis of the colorectum. Gut 47: 43-49, 2000.

Jeghers H, McKusick VA and Katz KH. Generalized intestinal polyposis and melanin spots of the oral mucosa, lips and digits; a syndrome of diagnostic significance. N.Engl.J.Med. 241: 1031-1036, 1949.

63

Jen J, Powell SM, Papadopoulos N, Smith KJ, Hamilton SR, Vogelstein B and Kinzler KW. Molecular determinants of dysplasia in colorectal lesions. Cancer Res. 54: 5523-5526, 1994.

Jiricny J. Mediating mismatch repair. Nat.Genet. 24: 6-8, 2000.

Johnson RL, Rothman AL, Xie J, Goodrich LV, Bare JW, Bonifas JM, Quinn AG, Myers RM, Cox DR, Epstein EH,Jr and Scott MP. Human homolog of patched, a candidate gene for the basal cell nevus syndrome. Science 272: 1668-1671, 1996.

Jover R, Paya A, Alenda C, Poveda MJ, Peiro G, Aranda FI and Perez-Mateo M. Defective mismatch-repair colorectal cancer: clinicopathologic characteristics and usefulness of immunohistochemical analysis for diagnosis. Am.J.Clin.Pathol. 122: 389-394, 2004.

Kambara T, Simms LA, Whitehall VL, Spring KJ, Wynter CV, Walsh MD, Barker MA, Arnold S, McGivern A, Matsubara N, Tanaka N, Higuchi T, Young J, Jass JR and Leggett BA. BRAF mutation is associated with DNA methylation in serrated polyps and cancers of the colorectum. Gut 53: 1137-1144, 2004.

Kane MF, Loda M, Gaida GM, Lipman J, Mishra R, Goldman H, Jessup JM and Kolodner R. Methylation of the hMLH1 promoter correlates with lack of expression of hMLH1 in sporadic colon tumors and mismatch repair-defective human tumor cell lines. Cancer Res. 57: 808-811, 1997.

Karran P and Hampson R. Genomic instability and tolerance to alkylating agents. Cancer Surv. 28: 69-85, 1996.

Karuman P, Gozani O, Odze RD, Zhou XC, Zhu H, Shaw R, Brien TP, Bozzuto CD, Ooi D, Cantley LC and Yuan J. The Peutz-Jegher gene product LKB1 is a mediator of p53-dependent cell death. Mol.Cell 7: 1307-1319, 2001.

Kinzler KW and Vogelstein B. Lessons from hereditary colorectal cancer. Cell 87: 159-170, 1996.

Kinzler KW and Vogelstein B. Cancer-susceptibility genes. Gatekeepers and caretakers. Nature 386: 761, 763, 1997.

Kinzler KW and Vogelstein B. Landscaping the cancer terrain. Science 280: 1036-1037, 1998.

Knudson AG,Jr. Mutation and cancer: statistical study of retinoblastoma. Proc.Natl.Acad.Sci.U.S.A. 68: 820-823, 1971.

Krajewski S, Krajewska M, Shabaik A, Miyashita T, Wang HG and Reed JC. Immunohistochemical determination of in vivo distribution of Bax, a dominant inhibitor of Bcl-2. Am.J.Pathol. 145: 1323-1336, 1994.

Kruse R, Rutten A, Lamberti C, Hosseiny-Malayeri HR, Wang Y, Ruelfs C, Jungck M, Mathiak M, Ruzicka T, Hartschuh W, Bisceglia M, Friedl W and Propping P. Muir-Torre phenotype has a frequency of DNA mismatch-repair-gene mutations similar to that in hereditary nonpolyposis colorectal cancer families defined by the Amsterdam criteria. Am.J.Hum.Genet. 63: 63-70, 1998.

Lahiri DK and Nurnberger JI,Jr. A rapid non-enzymatic method for the preparation of HMW DNA from blood for RFLP studies. Nucleic Acids Res. 19: 5444, 1991.

Laiho P, Launonen V, Lahermo P, Esteller M, Guo M, Herman JG, Mecklin JP, Jarvinen H, Sistonen P, Kim KM, Shibata D, Houlston RS and Aaltonen LA. Low-level microsatellite instability in most colorectal carcinomas. Cancer Res. 62: 1166-1170, 2002.

Laken SJ, Petersen GM, Gruber SB, Oddoux C, Ostrer H, Giardiello FM, Hamilton SR, Hampel H, Markowitz A, Klimstra D, Jhanwar S, Winawer S, Offit K, Luce MC, Kinzler KW and Vogelstein B. Familial colorectal cancer in Ashkenazim due to a hypermutable tract in APC. Nat.Genet. 17: 79-83, 1997.

64

Lal G and Gallinger S. Familial adenomatous polyposis. Semin.Surg.Oncol. 18: 314-323, 2000.

Lane DP. Cancer. p53, guardian of the genome. Nature 358: 15-16, 1992.

Larsen F, Gundersen G, Lopez R and Prydz H. CpG islands as gene markers in the human genome. Genomics 13: 1095-1107, 1992.

Latif F, Tory K, Gnarra J, Yao M, Duh FM, Orcutt ML, Stackhouse T, Kuzmin I, Modi W and Geil L. Identification of the von Hippel-Lindau disease tumor suppressor gene. Science 260: 1317-1320, 1993.

Leach FS, Nicolaides NC, Papadopoulos N, Liu B, Jen J, Parsons R, Peltomaki P, Sistonen P, Aaltonen LA and Nystrom-Lahti M. Mutations of a mutS homolog in hereditary nonpolyposis colorectal cancer. Cell 75: 1215-1225, 1993.

Leppert M, Burt R, Hughes JP, Samowitz W, Nakamura Y, Woodward S, Gardner E, Lalouel JM and White R. Genetic analysis of an inherited predisposition to colon cancer in a family with a variable number of adenomatous polyps. N.Engl.J.Med. 322: 904-908, 1990.

Lichtenstein P, Holm NV, Verkasalo PK, Iliadou A, Kaprio J, Koskenvuo M, Pukkala E, Skytthe A and Hemminki K. Environmental and heritable factors in the causation of cancer--analyses of cohorts of twins from Sweden, Denmark, and Finland. N.Engl.J.Med. 343: 78-85, 2000.

Lindor NM, Burgart LJ, Leontovich O, Goldberg RM, Cunningham JM, Sargent DJ, Walsh-Vockley C, Petersen GM, Walsh MD, Leggett BA, Young JP, Barker MA, Jass JR, Hopper J, Gallinger S, Bapat B, Redston M and Thibodeau SN. Immunohistochemistry versus microsatellite instability testing in phenotyping colorectal tumors. J.Clin.Oncol. 20: 1043-1048, 2002.

Liu B, Parsons R, Papadopoulos N, Nicolaides NC, Lynch HT, Watson P, Jass JR, Dunlop M, Wyllie A, Peltomaki P, de la Chapelle A, Hamilton SR, Vogelstein B and Kinzler KW. Analysis of mismatch repair genes in hereditary non-polyposis colorectal cancer patients. Nat.Med. 2: 169-174, 1996.

Liu T, Yan H, Kuismanen S, Percesepe A, Bisgaard ML, Pedroni M, Benatti P, Kinzler KW, Vogelstein B, Ponz de Leon M, Peltomaki P and Lindblom A. The role of hPMS1 and hPMS2 in predisposing to colorectal cancer. Cancer Res. 61: 7798-7802, 2001.

Lothe RA, Peltomaki P, Meling GI, Aaltonen LA, Nystrom-Lahti M, Pylkkanen L, Heimdal K, Andersen TI, Moller P and Rognum TO. Genomic instability in colorectal cancer: relationship to clinicopathological variables and family history. Cancer Res. 53: 5849-5852, 1993.

Makinen MJ, George SM, Jernvall P, Makela J, Vihko P and Karttunen TJ. Colorectal carcinoma associated with serrated adenoma--prevalence, histological features, and prognosis. J.Pathol. 193: 286-294, 2001.

Malkhosyan S, Rampino N, Yamamoto H and Perucho M. Frameshift mutator mutations. Nature 382: 499-500, 1996.

Marignani PA, Kanai F and Carpenter CL. LKB1 associates with Brg1 and is necessary for Brg1-induced growth arrest. J.Biol.Chem. 276: 32415-32418, 2001.

Markowitz S, Wang J, Myeroff L, Parsons R, Sun L, Lutterbaugh J, Fan RS, Zborowska E, Kinzler KW and Vogelstein B. Inactivation of the type II TGF-beta receptor in colon cancer cells with microsatellite instability. Science 268: 1336-1338, 1995.

Marsischky GT, Filosi N, Kane MF and Kolodner R. Redundancy of Saccharomyces cerevisiae MSH3 and MSH6 in MSH2-dependent mismatch repair. Genes Dev. 10: 407-420, 1996.

65

Martinez ME, Giovannucci E, Spiegelman D, Hunter DJ, Willett WC and Colditz GA. Leisure-time physical activity, body size, and colon cancer in women. Nurses' Health Study Research Group. J.Natl.Cancer Inst. 89: 948-955, 1997.

Massague J. TGF-beta signal transduction. Annu.Rev.Biochem. 67: 753-791, 1998.

May P and May E. Twenty years of p53 research: structural and functional aspects of the p53 protein. Oncogene 18: 7621-7636, 1999.

McCormick F. ras GTPase activating protein: signal transmitter and signal terminator. Cell 56: 5-8, 1989.

Mecklin JP. Frequency of hereditary colorectal carcinoma. Gastroenterology 93: 1021-1025, 1987.

Miao X, Sun T, Wang Y, Zhang X, Tan W and Lin D. Functional STK15 Phe31Ile polymorphism is associated with the occurrence and advanced disease status of esophageal squamous cell carcinoma. Cancer Res. 64: 2680-2683, 2004.

Michael-Robinson JM, Biemer-Huttmann A, Purdie DM, Walsh MD, Simms LA, Biden KG, Young JP, Leggett BA, Jass JR and Radford-Smith GL. Tumour infiltrating lymphocytes and apoptosis are independent features in colorectal cancer stratified according to microsatellite instability status. Gut 48: 360-366, 2001.

Mirabelli-Primdahl L, Gryfe R, Kim H, Millar A, Luceri C, Dale D, Holowaty E, Bapat B, Gallinger S and Redston M. Beta-catenin mutations are specific for colorectal carcinomas with microsatellite instability but occur in endometrial carcinomas irrespective of mutator pathway. Cancer Res. 59: 3346-3351, 1999.

Miyaki M, Konishi M, Tanaka K, Kikuchi-Yanoshita R, Muraoka M, Yasuno M, Igari T, Koike M, Chiba M and Mori T. Germline mutation of MSH6 as the cause of hereditary nonpolyposis colorectal cancer. Nat.Genet. 17: 271-272, 1997.

Modrich P and Lahue R. Mismatch repair in replication fidelity, genetic recombination, and cancer biology. Annu.Rev.Biochem. 65: 101-133, 1996.

Mori T, Nagase H, Horii A, Miyoshi Y, Shimano T, Nakatsuru S, Aoki T, Arakawa H, Yanagisawa A and Ushio Y. Germ-line and somatic mutations of the APC gene in patients with Turcot syndrome and analysis of APC mutations in brain tumors. Genes Chromosomes Cancer 9: 168-172, 1994.

Mori Y, Yin J, Rashid A, Leggett BA, Young J, Simms L, Kuehl PM, Langenberg P, Meltzer SJ and Stine OC. Instabilotyping: comprehensive identification of frameshift mutations caused by coding region microsatellite instability. Cancer Res. 61: 6046-6049, 2001.

Morin PJ, Sparks AB, Korinek V, Barker N, Clevers H, Vogelstein B and Kinzler KW. Activation of beta-catenin-Tcf signaling in colon cancer by mutations in beta-catenin or APC. Science 275: 1787-1790, 1997.

Muller A, Edmonston TB, Dietmaier W, Buttner R, Fishel R and Ruschoff J. MSI-testing in hereditary non-polyposis colorectal carcinoma (HNPCC). Dis.Markers 20: 225-236, 2004.

Muto T, Bussey HJ and Morson BC. The evolution of cancer of the colon and rectum. Cancer 36: 2251-2270, 1975.

Nakamura Y, Nishisho I, Kinzler KW, Vogelstein B, Miyoshi Y, Miki Y, Ando H, Horii A and Nagase H. Mutations of the adenomatous polyposis coli gene in familial polyposis coli patients and sporadic colorectal tumors. Princess Takamatsu Symp. 22: 285-292, 1991.

Nicolaides NC, Papadopoulos N, Liu B, Wei YF, Carter KC, Ruben SM, Rosen CA, Haseltine WA, Fleischmann RD and Fraser CM. Mutations of two PMS homologues in hereditary nonpolyposis colon cancer. Nature 371: 75-80, 1994.

66

Nishisho I, Nakamura Y, Miyoshi Y, Miki Y, Ando H, Horii A, Koyama K, Utsunomiya J, Baba S and Hedge P. Mutations of chromosome 5q21 genes in FAP and colorectal cancer patients. Science 253: 665-669, 1991.

Nystrom-Lahti M, Kristo P, Nicolaides NC, Chang SY, Aaltonen LA, Moisio AL, Jarvinen HJ, Mecklin JP, Kinzler KW and Vogelstein B. Founding mutations and Alu-mediated recombination in hereditary colon cancer. Nat.Med. 1: 1203-1206, 1995.

Ogawa O, Becroft DM, Morison IM, Eccles MR, Skeen JE, Mauger DC and Reeve AE. Constitutional relaxation of insulin-like growth factor II gene imprinting associated with Wilms' tumour and gigantism. Nat.Genet. 5: 408-412, 1993.

Oldak M, Grzela T, Lazarczyk M, Malejczyk J and Skopinski P. Clinical aspects of disrupted Hedgehog signaling (Review). Int.J.Mol.Med. 8: 445-452, 2001.

Orford K, Crockett C, Jensen JP, Weissman AM and Byers SW. Serine phosphorylation-regulated ubiquitination and degradation of beta-catenin. J.Biol.Chem. 272: 24735-24738, 1997.

Papadopoulos N, Nicolaides NC, Wei YF, Ruben SM, Carter KC, Rosen CA, Haseltine WA, Fleischmann RD, Fraser CM and Adams MD. Mutation of a mutL homolog in hereditary colon cancer. Science 263: 1625-1629, 1994.

Parc YR, Halling KC, Wang L, Christensen ER, Cunningham JM, French AJ, Burgart LJ, Price-Troska TL, Roche PC and Thibodeau SN. HMSH6 alterations in patients with microsatellite instability-low colorectal cancer. Cancer Res. 60: 2225-2231, 2000.

Parsons R, Myeroff LL, Liu B, Willson JK, Markowitz SD, Kinzler KW and Vogelstein B. Microsatellite instability and mutations of the transforming growth factor beta type II receptor gene in colorectal cancer. Cancer Res. 55: 5548-5550, 1995.

Pasche B, Kaklamani V, Hou N, Young T, Rademaker A, Peterlongo P, Ellis N, Offit K, Caldes T, Reiss M and Zheng T. TGFBR1*6A and cancer: a meta-analysis of 12 case-control studies. J.Clin.Oncol. 22: 756-758, 2004.

Pasche B, Kolachana P, Nafa K, Satagopan J, Chen YG, Lo RS, Brener D, Yang D, Kirstein L, Oddoux C, Ostrer H, Vineis P, Varesco L, Jhanwar S, Luzzatto L, Massague J and Offit K. TbetaR-I(6A) is a candidate tumor susceptibility allele. Cancer Res. 59: 5678-5682, 1999.

Pedroni M, Tamassia MG, Percesepe A, Roncucci L, Benatti P, Lanza G,Jr, Gafa R, Di Gregorio C, Fante R, Losi L, Gallinari L, Scorcioni F, Vaccina F, Rossi G, Cesinaro AM and Ponz de Leon M. Microsatellite instability in multiple colorectal tumors. Int.J.Cancer 81: 1-5, 1999.

Peltomaki P. Role of DNA mismatch repair defects in the pathogenesis of human cancer. J.Clin.Oncol. 21: 1174-1179, 2003.

Percesepe A, Kristo P, Aaltonen LA, Ponz de Leon M, de la Chapelle A and Peltomaki P. Mismatch repair genes and mononucleotide tracts as mutation targets in colorectal tumors with different degrees of microsatellite instability. Oncogene 17: 157-163, 1998.

Peyssonnaux C and Eychene A. The Raf/MEK/ERK pathway: new concepts of activation. Biol.Cell. 93: 53-62, 2001.

Pretlow TP, Brasitus TA, Fulton NC, Cheyer C and Kaplan EL. K-ras mutations in putative preneoplastic lesions in human colon. J.Natl.Cancer Inst. 85: 2004-2007, 1993.

Prolla TA, Pang Q, Alani E, Kolodner RD and Liskay RM. MLH1, PMS1, and MSH2 interactions during the initiation of DNA mismatch repair in yeast. Science 265: 1091-1093, 1994.

67

Pyatt R, Chadwick RB, Johnson CK, Adebamowo C, de la Chapelle A and Prior TW. Polymorphic variation at the BAT-25 and BAT-26 loci in individuals of African origin. Implications for microsatellite instability testing. Am.J.Pathol. 155: 349-353, 1999.

Rainier S, Johnson LA, Dobry CJ, Ping AJ, Grundy PE and Feinberg AP. Relaxation of imprinted genes in human cancer. Nature 362: 747-749, 1993.

Rajagopalan H, Bardelli A, Lengauer C, Kinzler KW, Vogelstein B and Velculescu VE. Tumorigenesis: RAF/RAS oncogenes and mismatch-repair status. Nature 418: 934, 2002.

Rampino N, Yamamoto H, Ionov Y, Li Y, Sawai H, Reed JC and Perucho M. Somatic frameshift mutations in the BAX gene in colon cancers of the microsatellite mutator phenotype. Science 275: 967-969, 1997.

Rashid A, Houlihan PS, Booker S, Petersen GM, Giardiello FM and Hamilton SR. Phenotypic and molecular characteristics of hyperplastic polyposis. Gastroenterology 119: 323-332, 2000.

Rashid A, Zahurak M, Goodman SN and Hamilton SR. Genetic epidemiology of mutated K-ras proto-oncogene, altered suppressor genes, and microsatellite instability in colorectal adenomas. Gut 44: 826-833, 1999.

Ratcliffe PJ, O'Rourke JF, Maxwell PH and Pugh CW. Oxygen sensing, hypoxia-inducible factor-1 and the regulation of mammalian gene expression. J.Exp.Biol. 201: 1153-1162, 1998.

Reddy EP, Reynolds RK, Santos E and Barbacid M. A point mutation is responsible for the acquisition of transforming properties by the T24 human bladder carcinoma oncogene. Nature 300: 149-152, 1982.

Riggins GJ, Kinzler KW, Vogelstein B and Thiagalingam S. Frequency of Smad gene mutations in human cancers. Cancer Res. 57: 2578-2580, 1997.

Risch N. The genetic epidemiology of cancer: interpreting family and twin studies and their implications for molecular genetic approaches. Cancer Epidemiol.Biomarkers Prev. 10: 733-741, 2001.

Salovaara R, Loukola A, Kristo P, Kaariainen H, Ahtola H, Eskelinen M, Harkonen N, Julkunen R, Kangas E, Ojala S, Tulikoura J, Valkamo E, Jarvinen H, Mecklin JP, Aaltonen LA and de la Chapelle A. Population-based molecular detection of hereditary nonpolyposis colorectal cancer. J.Clin.Oncol. 18: 2193-2200, 2000.

Samowitz WS, Powers MD, Spirio LN, Nollet F, van Roy F and Slattery ML. Beta-catenin mutations are more frequent in small colorectal adenomas than in larger adenomas and invasive carcinomas. Cancer Res. 59: 1442-1444, 1999.

Samowitz WS, Slattery ML, Potter JD and Leppert MF. BAT-26 and BAT-40 instability in colorectal adenomas and carcinomas and germline polymorphisms. Am.J.Pathol. 154: 1637-1641, 1999.

Sawyer EJ, Cerar A, Hanby AM, Gorman P, Arends M, Talbot IC and Tomlinson IP. Molecular characteristics of serrated adenomas of the colorectum. Gut 51: 200-206, 2002.

Schmutte C, Marinescu RC, Sadoff MM, Guerrette S, Overhauser J and Fishel R. Human exonuclease I interacts with the mismatch repair protein hMSH2. Cancer Res. 58: 4537-4542, 1998.

Schwartz RA and Torre DP. The Muir-Torre syndrome: a 25-year retrospect. J.Am.Acad.Dermatol. 33: 90-104, 1995.

Schwartz S,Jr, Yamamoto H, Navarro M, Maestro M, Reventos J and Perucho M. Frameshift mutations at mononucleotide repeats in caspase-5 and other target genes in endometrial and gastrointestinal cancer of the microsatellite mutator phenotype. Cancer Res. 59: 2995-3002, 1999.

68

Scott N, Bell SM, Sagar P, Blair GE, Dixon MF and Quirke P. p53 expression and K-ras mutation in colorectal adenomas. Gut 34: 621-624, 1993.

Servomaa K, Kiuru A, Kosma VM, Hirvikoski P and Rytomaa T. p53 and K-ras gene mutations in carcinoma of the rectum among Finnish women. Mol.Pathol. 53: 24-30, 2000.

Shaheen NJ, Silverman LM, Keku T, Lawrence LB, Rohlfs EM, Martin CF, Galanko J and Sandler RS. Association between hemochromatosis (HFE) gene mutation carrier status and the risk of colon cancer. J.Natl.Cancer Inst. 95: 154-159, 2003.

Shia J, Klimstra DS, Nafa K, Offit K, Guillem JG, Markowitz AJ, Gerald WL and Ellis NA. Value of immunohistochemical detection of DNA mismatch repair proteins in predicting germline mutation in hereditary colorectal neoplasms. Am.J.Surg.Pathol. 29: 96-104, 2005.

Shibata D, Schaeffer J, Li ZH, Capella G and Perucho M. Genetic heterogeneity of the c-K-ras locus in colorectal adenomas but not in adenocarcinomas. J.Natl.Cancer Inst. 85: 1058-1063, 1993.

Shpitz B, Hay K, Medline A, Bruce WR, Bull SB, Gallinger S and Stern H. Natural history of aberrant crypt foci. A surgical approach. Dis.Colon Rectum 39: 763-767, 1996.

Sieber OM, Lipton L, Crabtree M, Heinimann K, Fidalgo P, Phillips RK, Bisgaard ML, Orntoft TF, Aaltonen LA, Hodgson SV, Thomas HJ and Tomlinson IP. Multiple colorectal adenomas, classic adenomatous polyposis, and germ-line mutations in MYH. N.Engl.J.Med. 348: 791-799, 2003.

Slattery ML, Levin TR, Ma K, Goldgar D, Holubkov R and Edwards S. Family history and colorectal cancer: predictors of risk. Cancer Causes Control 14: 879-887, 2003.

Sparks AB, Morin PJ, Vogelstein B and Kinzler KW. Mutational analysis of the APC/beta-catenin/Tcf pathway in colorectal cancer. Cancer Res. 58: 1130-1134, 1998.

Spigelman AD, Arese P and Phillips RK. Polyposis: the Peutz-Jeghers syndrome. Br.J.Surg. 82: 1311-1314, 1995.

Spirio L, Otterud B, Stauffer D, Lynch H, Lynch P, Watson P, Lanspa S, Smyrk T, Cavalieri J and Howard L. Linkage of a variant or attenuated form of adenomatous polyposis coli to the adenomatous polyposis coli (APC) locus. Am.J.Hum.Genet. 51: 92-100, 1992.

Strachan T and Read AP. Organization of the human genome. In: Human Molecular Genetics. BIOS Scientific Publishers Ltd., Oxford, UK, pp. 139-168, 1999

Strater J, Wellisch I, Riedl S, Walczak H, Koretz K, Tandara A, Krammer PH and Moller P. CD95 (APO-1/Fas)-mediated apoptosis in colon epithelial cells: a possible role in ulcerative colitis. Gastroenterology 113: 160-167, 1997.

Sun T, Miao X, Wang J, Tan W, Zhou Y, Yu C and Lin D. Functional Phe31Ile polymorphism in Aurora A and risk of breast carcinoma. Carcinogenesis 25: 2225-2230, 2004.

Suraweera N, Duval A, Reperant M, Vaury C, Furlan D, Leroy K, Seruca R, Iacopetta B and Hamelin R. Evaluation of tumor microsatellite instability using five quasimonomorphic mononucleotide repeats and pentaplex PCR. Gastroenterology 123: 1804-1811, 2002.

Suzuki H, Komiya A, Aida S, Akimoto S, Shiraishi T, Yatani R, Igarashi T and Shimazaki J. Microsatellite instability and other molecular abnormalities in human prostate cancer. Jpn.J.Cancer Res. 86: 956-961, 1995.

Taipale J and Beachy PA. The Hedgehog and Wnt signalling pathways in cancer. Nature 411: 349-354, 2001.

69

Takeda S, Kondo M, Kumada T, Koshikawa T, Ueda R, Nishio M, Osada H, Suzuki H, Nagatake M, Washimi O, Takagi K, Takahashi T, Nakao A and Takahashi T. Allelic-expression imbalance of the insulin-like growth factor 2 gene in hepatocellular carcinoma and underlying disease. Oncogene 12: 1589-1592, 1996.

Talbot IC, Burt R, Järvinen H, Thomas G. Familial adenomatous polyposis. In: Pathology and genetics of tumors of the digestive system, World Health Organization classification of tumors. Eds. Hamilton SR, Aaltonen LA. Lyon: IARC Press, 120-5, 2000

Tateyama H, Li W, Takahashi E, Miura Y, Sugiura H and Eimoto T. Apoptosis index and apoptosis-related antigen expression in serrated adenoma of the colorectum: the saw-toothed structure may be related to inhibition of apoptosis. Am.J.Surg.Pathol. 26: 249-256, 2002.

Thibodeau SN, Bren G and Schaid D. Microsatellite instability in cancer of the proximal colon. Science 260: 816-819, 1993.

Thibodeau SN, French AJ, Cunningham JM, Tester D, Burgart LJ, Roche PC, McDonnell SK, Schaid DJ, Vockley CW, Michels VV, Farr GH,Jr and O'Connell MJ. Microsatellite instability in colorectal cancer: different mutator phenotypes and the principal involvement of hMLH1. Cancer Res. 58: 1713-1718, 1998.

Tishkoff DX, Amin NS, Viars CS, Arden KC and Kolodner RD. Identification of a human gene encoding a homologue of Saccharomyces cerevisiae EXO1, an exonuclease implicated in mismatch repair and recombination. Cancer Res. 58: 5027-5031, 1998.

Tomlinson I, Halford S, Aaltonen L, Hawkins N and Ward R. Does MSI-low exist? J.Pathol. 197: 6-13, 2002.

Umar A, Boland CR, Terdiman JP, Syngal S, de la Chapelle A, Ruschoff J, Fishel R, Lindor NM, Burgart LJ, Hamelin R, Hamilton SR, Hiatt RA, Jass J, Lindblom A, Lynch HT, Peltomaki P, Ramsey SD, Rodriguez-Bigas MA, Vasen HF, Hawk ET, Barrett JC, Freedman AN and Srivastava S. Revised Bethesda Guidelines for hereditary nonpolyposis colorectal cancer (Lynch syndrome) and microsatellite instability. J.Natl.Cancer Inst. 96: 261-268, 2004.

Vasen HF, Mecklin JP, Khan PM and Lynch HT. The International Collaborative Group on Hereditary Non-Polyposis Colorectal Cancer (ICG-HNPCC). Dis.Colon Rectum 34: 424-425, 1991.

Vasen HF, Watson P, Mecklin JP and Lynch HT. New clinical criteria for hereditary nonpolyposis colorectal cancer (HNPCC, Lynch syndrome) proposed by the International Collaborative group on HNPCC. Gastroenterology 116: 1453-1456, 1999.

Verma M and Srivastava S. Epigenetics in cancer: implications for early detection and prevention. Lancet Oncol. 3: 755-763, 2002.

Vogelstein B, Fearon ER, Hamilton SR, Kern SE, Preisinger AC, Leppert M, Nakamura Y, White R, Smits AM and Bos JL. Genetic alterations during colorectal-tumor development. N.Engl.J.Med. 319: 525-532, 1988.

Vogelstein B and Kinzler KW. Cancer genes and the pathways they control. Nat.Med. 10: 789-799, 2004.

Wang TF, Kleckner N and Hunter N. Functional specificity of MutL homologs in yeast: evidence for three Mlh1-based heterocomplexes with distinct roles during meiosis in recombination and mismatch correction. Proc.Natl.Acad.Sci.U.S.A. 96: 13914-13919, 1999.

Watson P and Lynch HT. Extracolonic cancer in hereditary nonpolyposis colorectal cancer. Cancer 71: 677-685, 1993.

Weber JL and Wong C. Mutation of human short tandem repeats. Hum.Mol.Genet. 2: 1123-1128, 1993.

70

Whitehall VL, Walsh MD, Young J, Leggett BA and Jass JR. Methylation of O-6-methylguanine DNA methyltransferase characterizes a subset of colorectal cancer with low-level DNA microsatellite instability. Cancer Res. 61: 827-830, 2001.

Wijnen J, de Leeuw W, Vasen H, van der Klift H, Moller P, Stormorken A, Meijers-Heijboer H, Lindhout D, Menko F, Vossen S, Moslein G, Tops C, Brocker-Vriends A, Wu Y, Hofstra R, Sijmons R, Cornelisse C, Morreau H and Fodde R. Familial endometrial cancer in female carriers of MSH6 germline mutations. Nat.Genet. 23: 142-144, 1999.

Wijnen J, Khan PM, Vasen H, van der Klift H, Mulder A, van Leeuwen-Cornelisse I, Bakker B, Losekoot M, Moller P and Fodde R. Hereditary nonpolyposis colorectal cancer families not complying with the Amsterdam criteria show extremely low frequency of mismatch-repair-gene mutations. Am.J.Hum.Genet. 61: 329-335, 1997.

Willett WC, Stampfer MJ, Colditz GA, Rosner BA and Speizer FE. Relation of meat, fat, and fiber intake to the risk of colon cancer in a prospective study among women. N.Engl.J.Med. 323: 1664-1672, 1990.

Wong NA and Pignatelli M. Beta-catenin--a linchpin in colorectal carcinogenesis? Am.J.Pathol. 160: 389-401, 2002.

Wu Y, Berends MJ, Mensink RG, Kempinga C, Sijmons RH, van Der Zee AG, Hollema H, Kleibeuker JH, Buys CH and Hofstra RM. Association of hereditary nonpolyposis colorectal cancer-related tumors displaying low microsatellite instability with MSH6 germline mutations. Am.J.Hum.Genet. 65: 1291-1298, 1999.

Wu Y, Berends MJ, Sijmons RH, Mensink RG, Verlind E, Kooi KA, van der Sluis T, Kempinga C, van dDer Zee AG, Hollema H, Buys CH, Kleibeuker JH and Hofstra RM. A role for MLH3 in hereditary nonpolyposis colorectal cancer. Nat.Genet. 29: 137-138, 2001.

Wu Y, Berends MJ, Post JG, Mensink RG, Verlind E, Van Der Sluis T, Kempinga C, Sijmons RH, van der Zee AG, Hollema H, Kleibeuker JH, Buys CH and Hofstra RM. Germline mutations of EXO1 gene in patients with hereditary nonpolyposis colorectal cancer (HNPCC) and atypical HNPCC forms. Gastroenterology 120: 1580-1587, 2001b.

Yee CJ, Roodi N, Verrier CS and Parl FF. Microsatellite instability and loss of heterozygosity in breast cancer. Cancer Res. 54: 1641-1644, 1994.

Ylikorkala A, Rossi DJ, Korsisaari N, Luukko K, Alitalo K, Henkemeyer M and Makela TP. Vascular abnormalities and deregulation of VEGF in Lkb1-deficient mice. Science 293: 1323-1326, 2001.

Zauber NP, Sabbath-Solitare M, Marotta SP and Bishop DT. The characterization of somatic APC mutations in colonic adenomas and carcinomas in Ashkenazi Jews with the APC I1307K variant using linkage disequilibrium. J.Pathol. 199: 146-151, 2003.

Documents

Molecular classification of colorectal cancer · 2017-03-17 · predisposing to colorectal cancer largely remain to be characterized. Though the main relevance of hereditary cancer