The Roles of MLH1 and MSH2 in Growth and Drug Resistance in Human Colorectal Cancer Cells

The roles of MLH1 and MSH2 in growth and drug resistance in human colorectal cancer cells

Amanda Barber

A Thesis presented to The University of Guelph

In partial fulfilment of requirements for the degree of Master of Science in Biomedical Sciences

Guelph, Ontario, Canada

The role of MLH1 and MSH2 in tumor growth and drug resistance in human colorectal cancer cells

Amanda Barber Advisor: University of Guelph, 2012 Dr. B. L. Coomber

Loss of genomic stability is associated with a variety of diseases, particularly cancer. Of the many proteins which maintain genomic integrity, two of the most important are MLH1 and

MSH2, which participate in DNA mismatch repair. Previous work established derivatives of the

CaCo2 human colorectal cancer cell line with siRNA-mediated knockdown of these proteins.

When xenografted into mice, tumors with reduced MLH1 or MSH2 expression grew faster than controls. Following growth in vivo, clonal cell lines were established from the tumors and used to examine the effects that knockdown of MSH2 had on other members of the DNA mismatch repair system. Clonal survival following exposure to 5-fluorouracil was also evaluated, and those cells with reduced MLH1 and MSH2 levels were found to be resistant. This study has implications for the importance of knowing the MMR status of a given tumor when deciding on a course of treatment, and of the compounding effects of the loss of one MMR protein on others in the family. ACKNOWLEDGEMENTS

I would firstly like to acknowledge my advisor, Dr. Brenda Coomber for her guidance, wisdom and support throughout the course of my Masters degree. Her input has always been invaluable and has helped shaped the way I think about both science and research.

I would also like to thank my advisory committee, Dr. David Josephy and Dr. Geoffrey

Wood, for their suggestions and assistance throughout this process. Thanks also goes to Kristen

Lacombe for providing the jumping-off point for my project.

As well, I would like to acknowledge all of my past and present labmates, both in the

Department of Biomedical Sciences, as well as those in the Department of Molecular and

Cellular Biology. I would like to thank the laboratory of Dr. Ian Tetlow, specifically Dr. Amina

Makhmoudova, Dr. Nadya Romanova, Dr. Wendy Allen and Dr. Fushan Liu for giving me a strong early foundation in molecular biology techniques when I was a wide-eyed student fresh off my second year of undergraduate courses. Thank you for your infinite patience, and thank you also to Ian for allowing me to get my feet wet in his laboratory. I would also like to thank the members of the laboratory of Dr. Ray Lu, especially Dr. Mandi Martyn, Tegan Williams, and

Sally Burtenshaw. Thank you for being great coworkers and mentors throughout the course of my fourth-year research project, but more importantly, thank you for being fantastic friends and for making that experience worthwhile. While at times it was undoubtedly challenging, I ended up learning more about science and research in that year than I would have ever thought possible.

Of course, I must also give thanks to those I worked alongside throughout the course of my masters research. Specifically, I would like to thank Dr. Una Adamcic-Bistrivoda, Leanne

Delaney, Nathan Farias, Richard Gilbert, Nelson Ho, Mai Jarad, Lizzie Kuczynski, Jodi

iii

Morrison, Amy Richard, Dr. Kaya Skowronski, Dr. Jen Thompson, and Sonja Zours. Thank you for listening to me rant, and for providing one of the best work environments I’ve been lucky enough to experience. I would like to also thank Noel Darling Hill for her artistic and creative input into the various posters I’ve presented over the years – “bajis!”.

Finally, I cannot conclude this series of acknowledgements without mentioning my parents and family. Thank you for your financial and emotional support throughout the course of my post-graduate studies – I honestly could not have done this without you. You have always believed in me and my abilities, and provided an excellent launching-off point for my forays into scientific research, and that has meant more to me than I will ever be able to express. Thank you also for your infinite attempts to understand “exactly what it is that I do” – hopefully reading this thesis can give you some idea.

DECLARATION OF WORK PERFORMED

I declare that all work reported in this thesis was performed by me, with the exception of items indicated below.

Kristen Lacombe developed the initial MLH1 and MSH2 knockdown cell lines used for this project. Dr. Brenda Coomber xenografted these cells into mice and collected the tumor tissue to establish the cell lines examined in this thesis. Genomic DNA samples from these cell lines were sent to Sequenom, Inc. who then performed the MassArray analysis. This analysis was also performed a second time by Cheryl Crozier under the supervision of Dr. Victoria Sabine of the

Ontario Institute for Cancer Research. Nelson Ho helped maintain cell cultures when I was absent from the laboratory.

TABLE OF CONTENTS

ACKNOWLEDGEMENTS ...... iii DECLARATION OF WORK PERFORMED ...... v TABLE OF CONTENTS ...... vi LIST OF TABLES ...... viii LIST OF FIGURES ...... ix LIST OF ABBREVIATIONS ...... x INTRODUCTION ...... 1 LITERATURE REVIEW ...... 4 Colorectal cancer ...... 4 Genetic instability as an enabling hallmark of cancer ...... 4 Types of genetic instability ...... 7 DNA repair mechanisms ...... 9 Nucleotide excision repair and cancer ...... 10 Base excision repair and cancer ...... 12 Double strand break repair and cancer ...... 13 Mismatch repair and cancer: MSH and MLH proteins ...... 15 Familial colorectal cancer syndromes ...... 16 Familial Adenomatous Polyposis ...... 16 Lynch Syndrome ...... 18 Microsatellite Instability ...... 19 Diagnostic approaches ...... 21 MSI and chemotherapy ...... 22 Models of microsatellite instability ...... 24 In vitro models ...... 24 In vivo models ...... 25 RATIONALE ...... 27 MATERIALS AND METHODS ...... 29 Tissue culture ...... 29 DNA isolation and quantification from cells ...... 29 Protein isolation and quantification from cells ...... 30

SDS-PAGE and Western blotting analysis ...... 30 Detection of somatic mutations with Sequenom OncoCarta 1.0 panel ...... 32 Determination of BAT-25 and BAT-26 repeat lengths...... 32 Detection of BRAF V600E mutations in clonal cell lines ...... 33 Determination of clonal cell survival following treatment with 5-fluorouracil ...... 33 Statistical analysis ...... 34 RESULTS ...... 35 Knockdown of MLH1 or MSH2 increases tumorigenicity of Caco-2 cells in vivo ...... 35 Knockdown of MLH1 is lost following growth in vivo ...... 35 Knockdown of MSH2 is maintained following growth in vivo ...... 36 No point mutations detected in key oncogenes following growth in vivo in knockdown cells ...... 37 No changes in BAT-25 or BAT-26 repeat length in clones following growth in vivo ...... 44 Loss of MSH2 is accompanied by loss of PMS2, MSH3 and MSH6 ...... 47 Cells with reduced MLH1 or MSH2 show resistance to chronic exposure to low doses of 5-fluorouracil ...... 51 DISCUSSION ...... 55 Implications...... 65 Limitations ...... 66 Future Directions ...... 68 SUMMARY AND CONCLUSIONS ...... 69 REFERENCES ...... 71 APPENDIX I – Chemical List and Suppliers ...... 88 APPENDIX II – Preparation of Materials Used ...... 90 APPENDIX III – Primer Sequences ...... 92

vii

LIST OF TABLES

TABLE

1 Amino acid changes and small insertion/deletion mutations queried by Sequenom OncoCarta v1.0 panel ……………………………………………………………..42

viii

LIST OF FIGURES

FIGURE

1. Tumor growth of MLH1 and MSH2 knockdown cells in vivo…………………….. 38 2. MLH1 protein levels following growth in vivo……………………………………. 39 3. MSH2 protein levels following growth in vivo……………………………….…….40 4. BRAF V600 mutation screening in knockdown clones……………………….…… 41 5. BAT-25 tRFLP analysis for microsatellite instability………………………….….. 45 6. BAT-26 tRFLP analysis for microsatellite instability………………………….….. 46 7. Effects of MSH2 knockdown on PMS2 levels………………………………….…. 48 8. Effects of MSH2 knockdown on MSH3 levels………………………………….….49 9. Effects of MSH2 knockdown on MSH6 levels………………………………….….50 10. Colony formation in MLH1 and MSH2 knockdown cells exposed to chronic 5 μM 5-fluorouracil……………………………………………………..……….… 52 11. Colony formation in MLH1 and MSH2 knockdown cells exposed to chronic 2.5 μM 5-fluorouracil………………………………………………………………. 53 12. Colony formation in MLH1 and MSH2 knockdown cells exposed to 10 μM 5-fluorouracil for 24 hours………………………………………….……………… 54

LIST OF ABBREVIATIONS

5-FU 5-fluorouracil 6-FAM 6-carboxyfluorescein 6-TG 6-thioguanine APC Adenomatous polyposis coli ATM Ataxia-telangiectasia mutated BAT-25 Big adenine tract 25 BAT-26 Big adenine tract 26 BER Base excision repair BRAF v-raf murine sarcoma viral oncogene homolog B1 CIMP CpG island methylator phenotype CIN Chromosomal instability CRC Colorectal carcinoma DMEM Dulbecco’s Modified Eagle Media DMSO Dimethyl sulfoxide DSB Double strand break EDTA Ethylene diamine tetraacetic acid FBS Fetal bovine serum HIF Hypoxia inducible factor HPRT Hypoxanthine-guanine phosphoribosyltransferase HNPCC Hereditary non-polyposis colorectal cancer IDL Insertion/deletion mutation KD Knockdown KO Knockout KRAS V-Ki-ras2 Kirsten rat sarcoma viral oncogene homolog miRNA micro RNA MLH1 MutL homolog 1 MMR DNA mismatch repair MNNG N-methyl-N’-nitro-N-nitrosoguanidine MSI Microsatellite instability MSH2 MutS homolog 2 MSH3 MutS homolog 3 MSH6 MutS homolog 6 MSS Microsatellite stable NER Nucleotide excision repair PBS Phosphate buffered saline PCR Polymerase chain reaction PMS1 Postmeiotic segregation increased 1 PMS2 Postmeiotic segregation increased 2 PVDF Polyvinylidene difluoride RFLP Restriction fragment length polymorphism RNAi RNA interference SCID Severe combined immunodeficiency SDS Sodium dodecyl sulfate

SEM Standard error of the mean SHO SCID hairless outbred shRNA Short hairpin RNA siRNA Short interfering RNA TAE Tris-acetate-EDTA buffer TBS Tris buffered saline TBST Tris buffered saline with 0.1% Tween-20 TEMED N’,N’,N’,N’-tetramethylethylene diamine

INTRODUCTION

Cancer is the leading cause of death worldwide, accountable for 7.6 million fatalities in 2008 (WHO). This rate is expected to rise and is projected to exceed 11 million by 2030. Colorectal cancer in particular is the second leading cause of cancer-related deaths in Canadians, with 8900 predicted deaths in 2011, and is the fourth most common cancer diagnosis overall (www.cancer.ca/statistics).

While each individual tumour has its own unique characteristics, a list of the hallmarks of cancer has been proposed by Hanahan and Weinberg (Hanahan and Weinberg,

2000). These hallmarks include evading apoptosis, sustained angiogenesis, self-sufficiency in growth signals, insensitivity to anti-growth signals, limitless growth potential, and tissue invasion and metastasis. Recently revised in 2011, this list provides a framework of the traits that distinguish cancerous cells from normal cells (Hanahan and Weinberg, 2011).

These hallmarks allow for opportunities to study the disease, as well as development of therapeutics which target these distinct features. One of these newly-identified hallmarks is genomic instability and mutation, which is classified as an “enabling characteristic”, and serves to promote the development of many of the other features on the list. By acquiring mutations, neoplastic cells are able to evade apoptosis, proliferate incessantly, and otherwise continue to grow in the patient.

Genomic instability can be broadly divided into three phenotypes: chromosomal instability (CIN), microsatellite instability (MSI), and CpG island methylator phenotype

(CIMP). While potentially related, each of these categories produces different phenotypes in the affected cells in terms of the types of mutations observed and the effect that these

1 mutations have on cellular behavior. Microsatellite instability is the main focus of this thesis, and arises from the loss of DNA mismatch repair (MMR) proteins (Aquilina et al.,

1994). DNA mismatch repair is the cellular process responsible for repairing single base mismatches, as well as small insertions and deletions.

Loss of DNA mismatch repair proteins can occur through a variety of mechanisms.

These can include mutations in the coding sequences, rendering the proteins non-functional

(Peltomaki et al., 2001), or epigenetic modifications in the promoter regions, resulting in loss of expression (Cunningham et al., 1998). These modifications can occur sporadically throughout an individual’s lifetime (Yoon et al., 2011), or they may be inherited germline mutations, a familial condition known as Lynch Syndrome (Lynch et al., 1977). Regardless of the mechanism by which these proteins are lost, understanding the effect that this has on cancer development, progression, and response to therapeutics is an important avenue of research.

Previous work in our laboratory has established two subclones of the CaCo2 human colorectal cancer cell line with siRNA-mediated knockdown of the two most commonly lost DNA mismatch repair proteins, MLH1 and MSH2 (Lacombe, 2009). When xenografted into mice, these cell lines showed increased tumor take rate, increased tumor growth rate, and decreased tumor latency when compared to controls, implicating these two proteins in maintaining a stable genome and preventing cancerous growth.

This thesis aims to explore the effect that the loss of MLH1 and MSH2 proteins may have had on the development of microsatellite instability during growth in vivo, as well as any effect on the development of known mutations associated with cancer which may explain the increased tumorgenicity of these cells. The response of these cells to the

2 chemotherapeutic drug 5-fluorouracil will be evaluated, as cells with deficient MMR have been shown to be resistant to this drug (de Angelis et al., 2004; Laiho et al., 2002).

LITERATURE REVIEW

Colorectal cancer

While rates of colorectal cancer have been declining over the past two decades, the overall number of new cases is increasing as a result of population growth and aging

(www.cancer.ca). As of January 1, 2007, it is estimated that nearly 100,000 Canadians had been diagnosed with this kind of cancer in the preceding 10 years. Compared to other types of cancer, colorectal cancer has a moderate survivability, with a five-year relative survival ratio of 63%. This ranks higher than some carcinomas, such as lung (16%) and pancreatic

(6%), though lower than others, such as prostate (96%) and breast (88%) (www.cancer.ca).

Incidence of colorectal cancer in an individual can be divided into two broad categories – that which is sporadic, and that which is associated with a familial cancer syndrome. Sporadic colorectal cancer represents the majority, accounting for approximately 80% of cases (Al-Sukhni et al., 2008).

Genetic instability as an enabling hallmark of cancer

The hallmarks of cancer are a list comprised of unique biological characteristics and capabilities that are acquired by a tumor cell during its malignant progression. The six classical hallmarks consist of resisting cell death, sustaining proliferative signalling, resisting inhibitory signals, unlimited replicative potential, the ability to metastasize, and inducing angiogenesis (Hanahan and Weinberg, 2000). However, underlying all of these features is the development of genomic instability. This generates the genetic diversity necessary to allow the acquisition of the various hallmarks, and as such has been identified as an “enabling” hallmark (Hanahan and Weinberg, 2011). This increase in mutability can

4 be due to a variety of factors, including increased susceptibility to mutagens, problems with

DNA repair systems, or both. As well, the accumulation of mutations can be accelerated by the breakdown of genomic surveillance systems which aim to prevent the further cycling of genetically compromised cells.

The study of colorectal cancer in particular provides an opportunity to examine the various stages of the disease progression. Other models also exist which allow for the exploration of cancer progression. These models include the Solt-Farber model to initiate hepatocellular carcinomas in rats (Solt et al., 1977), or the SENCAR (SENsitivity to

CARcinogens) mouse model of skin tumor induction (Fischer et al., 1987). The Solt-Farber model traditionally involves the treatment with a single carcinogenic dose of diethylnitrosamine (DEN), followed by short-term dietary exposure to 2- acetylaminofluorene (2-AAF), and finally a partial hepatectomy. Within 1 week of this treatment regime, the altered hepatocytes begin to grow out and are histologically indistinguishable from liver nodules described in the literature. The use of this model allows for the study of the carcinogenic process in the liver, and can potentially be applied to other tissues.

Tumorigenesis has long been thought to be a multi-step progression (Foulds, 1958), but it was not until the early 1990s that a genetic model of colorectal tumorigenesis was proposed (Fearon and Vogelstein, 1990). This model identified key regions of the genome that were lost throughout the course of the adenoma-to-carcinoma progression observed in colorectal cancer, and identified several new tumor suppressor genes and oncogenes located in these regions. As well, the development of this model provided a new way of

5 thinking about cancer progression and expansion as a multi-step process owing to the accumulation of both mutations and epigenetic changes in the affected cells.

Generally speaking, the first mutations observed are those in APC, found on chromosome 5q. This event is then followed by global DNA hypomethylation, which has been shown in some cases to lead to mitotic nondisjunction by inhibiting chromosome condensation (Silverman et al., 1989). The development of a mutation in KRAS seems to help mediate the transition from an early adenoma to an intermediate adenoma. KRAS mutations occur at a relatively high frequency (approximately 50%) in adenomas and carcinomas greater than 1 cm in size (Bos et al., 1987; Vogelstein et al., 1988). In contrast,

KRAS mutations have been observed in fewer than 10% of adenomas smaller than one centimetre (Vogelstein et al., 1988), implying that the development of a mutation in this gene may be at least partially responsible for the progression of a small adenoma to a larger one (Fearon and Vogelstein, 1990). Following this, mutations in DCC and p53 aid in progression to a carcinoma and eventually, in conjunction with other mutations and cellular alterations, to metastasis. Examples of further mutations commonly observed in colorectal cancer include those affecting transforming growth factor β and PIK3CA (Fearon and

Vogelstein, 1990).

The generation of this model introduced some key underlying principles to the study of cancer as a disease of genetic alterations. Firstly, a cancerous phenotype results from both the activation of oncogenes, as well as the loss of tumor suppressor genes.

Notably, not all tumor suppressors are truly recessive, and in some cases, may function as dominant negative instead, as is the case with p53 (Herskowitz, 1987). In the case of this tumor suppressor gene, the highly charged basic region at the C terminus of the protein is

6 vital for the formation of tetrameric p53 complexes (Iwabuchi et al., 1993). It is believed that this binding greatly facilitates interactions with p53 response elements (Prives, 1994).

Mutations of p53 are most commonly missense mutations (accounting for approximately

85% of reported gene alterations), though deletions and nonsense mutations can also occur

(Cariello et al., 1994). Recent studies have shown that a mutation in one member of this tetramer can render the entire protein complex non-functional as it is unable to bind to and activate transcription of target DNA sequences (Aramayo et al., 2011). This phenomenon implies that mutation in only one allele can negate the activity of the p53 tetramer.

Mutations in other genes may behave in a similar fashion, depending on the function of the gene product.

Also of note is the observation that it is the total accumulated effect of all the mutations in a given cell which determines the malignancy of the tumor. While most colorectal cancers are postulated to acquire their mutations in a specific order, this order does not have as much of an effect on the development of the cancer as does the accumulation of a series of mutations. However, in a few select cases, this order appears to be quite specific. For example, mutational activation of KRAS cannot initiate cancer in vivo, and only when combined with a mutation in APC does mutant KRAS promote tumor progression (Haigis et al., 2008).

Types of genetic instability Genetic instability can be broadly divided into three phenotypes – chromosomal instability (CIN), CpG island methylator phenotype (CIMP) and microsatellite instability

(MSI). There is some overlap between these phenotypes – up to 25% of MSI colorectal cancers exhibit chromosomal abnormalities characteristic of CIN (Sinicrope et al., 2006).

Similarly, as many as 33% of CIMP-positive tumors also exhibit CIN characteristics

(Cheng et al., 2008).

CIN has been implicated in more than 80% of sporadic colorectal cancer cases

(Pino and Chung, 2010), and is characterized by the loss and rearrangement of large regions of chromosomes, resulting in aneuploidy and polyploidy, as well as loss of heterozygosity. CIN+ status in a cancer has been shown to positively correlate with poor prognosis and drug resistance, potentially through its promotion of genetic plasticity

(Walther et al., 2008). This phenotype can be caused by problems with chromosomal segregation, telomere stability, or the DNA damage response (Cahill et al., 1998; Gertler et al., 2004; Peddibhotla et al., 2009). The development of this phenotype has been associated with altered kinetochore dynamics in the cell, leading to improper chromosomal segregation during mitosis. In normal cells, merotelic attachments (kinetochore attachments to both mitotic spindle poles which may lead to mitotic nondisjunction if not corrected) commonly form during early mitosis, but are eventually corrected due to the rapidly forming and dissolving microtubule-kinetochore associations (Zhai et al., 1995).

Cancer cells which are CIN+ have been shown to have more stable kinetochore microtubule attachments, which becomes a hindrance to proper chromosomal segregation as they allow for the persistence of merotelic attachments (Bakhoum et al., 2009).

The CIMP phenotype is associated with altered methylation patterns of CpG islands in the genome. CpG islands are regions of the genome rich in cytosine and guanine residues which are commonly the target of DNA methylation. These regions are typically found in promoter regions and 5’ untranslated regions (UTR) of genes. Approximately 70% of mammalian genes contain a CpG island in either their promoter or 5’-UTR sequences

(Takai and Jones, 2002), and more than 90% of these genes are believed to be unmethylated in normal cells, the obvious exceptions being those which are on the inactive

X chromosome (Eckhardt et al., 2006).

A significant event observed to occur in cancer progression is the loss of nearly one third of all of the genomic methylation when compared to the DNA of normal colonic mucosa, representing approximately 10-20 million methyl groups (Feinberg et al., 1988).

However, it is worth noting that methylation-induced silencing of tumor suppressor genes is also a possibility, as a few specific regions of the genome have been observed to become hypermethylated (Silverman et al., 1989). This is commonly the case with the gene encoding the mismatch repair protein MLH1 (Herman et al., 1998). Thus, the development of CIMP serves as a mechanism of both gene activation and inactivation in cancer.

Finally, the microsatellite instability phenotype is observed in approximately 15% of CRC and is the main focus of this thesis (Geiersbach and Samowitz, 2011). MSI is characterized by changes in the lengths of repetitive regions of the genome, termed microsatellites. These regions are especially prone to mutation due to slippage of DNA polymerase during DNA replication. However, these lesions are normally repaired by the

DNA mismatch repair (MMR) pathway. This phenotype is associated with increased rates of point mutations and insertions/deletions in the genome. Should these mutations affect a tumor suppressor gene or an oncogene, this can lead to tumorigenesis.

DNA repair mechanisms

A mammalian cell is subject to a variety of destructive forces, both external and internal, which can result in the introduction of mutations into the genome. It is estimated

9 that 50,000 – 100,000 DNA modifications occur in one cell per day (Hubscher and Maga,

2011). Given the potentially disastrous consequences of unrepaired mutations, a variety of

DNA repair mechanisms exist within the cell to respond to this challenge. These include nucleotide excision repair (NER), base excision repair (BER), double strand break (DSB) repair, and mismatch repair (MMR).

Nucleotide excision repair and cancer Nucleotide excision repair acts in response to DNA lesions such as bulky adducts - nucleotides to which chemical groups have become attached. These lesions interfere with vital DNA functions, such as replication, recombination, and transcription, and therefore must be repaired to allow for cellular survival. This repair pathway was first characterized in the 1960s in E. coli when it was observed that thymine dimers caused by UV exposure are enzymatically removed, resulting in increased survival (Setlow and Carrier, 1964).

There are two main types of eukaryotic NER – global genome NER (GG-NER) and transcription-coupled NER (TC-NER). As the names suggest, GG-NER occurs at all stages of the cell cycle, while TC-NER only occurs when the damaged DNA is being transcribed into RNA.

The primary difference between these two pathways is the protein responsible for the initial recognition of the base(s) which need to be repaired. In the case of GG-NER, this recognition is mediated by XPC-HR23B and UV-DDB (Sugasawa et al., 1998), whereas in

TC-NER, stalled RNA polymerase II is believed to serve as the lesion sensor (Woudstra et al., 2002). Following recognition, the DNA is unwound by the enzymes XPB and XPD

(Coin et al., 2007). Two endonucleases, XPG and XPF-ERCC1, then incise the DNA on either side of the damaged site, resulting in the removal of 25-30 nucleotides (Mu et al.,

1996). The resulting gap is then filled by DNA polymerases, and the ends sealed by DNA ligase.

The consequences of defective NER can be observed in several genetic disorders, most notably xeroderma pigmentosum (XP), which is associated with mutations in the XP family of genes mentioned earlier. Problems with NER are also observed in those with

Cockayne Syndrome (CS) and Trichothiodystrophy (TTD), though these patients do not experience the dramatic 1000-fold increase in skin cancer incidence that is seen in XP patients (de Boer and Hoeijmakers, 2000). Interestingly, while CS and TTD are both associated with problems with proteins involved in NER, it is hard to explain some of the symptoms of these diseases based on what is known about the proteins affected (Stefanini et al., 2010).

Patients with XP are extremely sun-sensitive, with symptoms such as increased freckling and increased propensity to sunburn first appearing at 1-2 years of age (Kraemer et al., 1987). The median age for the development of skin cancer in these individuals is 8 years of age, whereas in the normal Caucasian population, the median is 60 years (Kraemer and Slor, 1985). Beyond skin cancer, XP patients are also at higher risk for other types of cancer, such as lung cancer (Kraemer et al., 1994). It is believed that tobacco smoke can serve as an “internal sun” in some cases, generating lesions which also require NER to be repaired. XP follows an autosomal recessive mode of inheritance, and while heterozygous carriers are believed to be phenotypically normal, some early studies have suggested an increased risk for skin cancer in these patients (Swift and Chase, 1979).

Base excision repair and cancer Base excision repair is the pathway responsible for removing most of the endogenous DNA damage occurring in the cell. These modifications include depurinations, deaminations, alkylations, and oxidative damages resulting from reactive oxygen species produced by cellular metabolism. It is estimated that approximately 30,000 assaults of this type occur per day in the cell (Friedberg et al., 2006). The BER pathway was first characterized in the 1970s in E. coli based on its role in removing misincorporated uracil bases from DNA (Lindahl, 1974). The initial step in this repair process is the recognition of the damaged base by a DNA glycosylase, which cleaves the N-glycosidic bond between the damaged base and the deoxyribose, producing an apyrimidinic or apurinic (AP) site (Laval,

1977). The DNA backbone is then cleaved by either the glycosylase or a secondary DNA apurinic endonuclease (Doetsch and Cunningham, 1990). It is at this point that the two types of BER, short-patch and long-patch, diverge. In short-patch BER, DNA polymerase β inserts the missing base, and the nick is sealed by a DNA ligase (Sobol et al., 1996). In long-patch BER, a DNA polymerase fills in the gap and then continues to synthesize beyond this gap, producing a “flap” of DNA (Dianov et al., 1999). This flap is then cleaved by the flap endonuclease FEN1, removing 2-13 nucleotides (Klungland and Lindahl, 1997).

Similar to short-patch BER, a DNA ligase seals the nick, restoring the intact sequence.

One important protein involved in facilitating BER in mammalian cells is MUTYH

(MutY Homolog). This glycosylase is responsible for recognizing one of the most common forms of oxidative DNA damage, the oxidation of guanine to form 8-oxoG (8-oxo-7,8- dihydro-2’-deoxyguanosine) (Shibutani et al., 1991). If left unrepaired, this oxidation will go on to result in a G:C to T:A transversion following the next round of DNA replication,

12 due to the misrecognition of the modified guanine residue as thymine by DNA polymerase

(Shibutani et al., 1991). Mutations in MUTYH are associated with a type of familial cancer referred to as MAP, or MUTYH-associated polyposis, which was initially characterized in

2002 (Al-Tassan et al., 2002).

Double strand break repair and cancer Double strand breaks (DSBs) are among the most cytotoxic forms of DNA damage and must be repaired with high fidelity to ensure both cell survival and genomic integrity.

DSBs can occur through exposure to exogenous agents, such as ionizing radiation (IR) and certain chemotherapeutic drugs (Resnick and Moore, 1979). They can also occur during endogenous cellular processes, including meiosis and lymphocyte development (Klein et al., 1996; Soulas-Sprauel et al., 2007). There are two main pathways employed by cells to repair DSBs – nonhomologous end joining (NHEJ) and homologous recombination (HR).

NHEJ involves direct ligation of the broken ends, while HR requires the presence of an undamaged homologous sequence to mediate the repair. HR primarily utilizes the sister chromatid to provide the correct DNA sequence for repair (Heyer, 1994). Thus, this type of

DSB repair is limited to the G2 and S phases of the cell cycle, when fully formed sister chromatids are available, or as DNA replication duplicates the genome, respectively

(Tarsounas et al., 2003).

In contrast, NHEJ is a more error-prone option for DSB repair, as the free ends of

DNA are recognized by the Ku protein complexes, which then act to recruit DNA ligases and accessory proteins to repair the lesion. Due to the ability of HR to faithfully restore a

DNA molecule which has been damaged, by using the homologous chromosome, it is the superior of the two options for accurate DSB repair. Therefore, it is interesting to note that

13 the preferred option more commonly used for this type of repair is actually NHEJ, even when the cells being examined are halted in G2 (Kinner et al., 2008).

The BRCA1 gene is involved in the repair of DSBs and is commonly mutated in breast cancer patients. This gene, found on chromosome 17q, encodes a protein of 1863 amino acids (Miki et al., 1994). Almost all of the over 500 different sequence variants reported are germline mutations – somatic mutations of this gene in ovarian or breast cancer patients are exceedingly rare (Merajver et al., 1995). Mutations in BRCA1 are found in almost 70% of patients when the median onset of breast cancer in the family is less than

45 years of age (Easton et al., 1993).

There are several different domains to this protein, each with unique biological function, all of which can be potentially implicated in cancer development and progression.

There is an N-terminal RING domain that is thought to bind to DNA, suggesting a possible role in transcriptional regulation, though no DNA-binding-partners have been identified to date (Bork et al., 1997). BRCA1 is able to form a variety of different heterodimers, and has been shown to dimerize with BAP1 to potentially mediate cell growth suppression (Jensen et al., 1998). Of particular relevance to this thesis, however, is the role of BRCA1 in the

DNA damage response. BRCA1 has dimerizes with Rad51 and BRCA2, both of which have implicated roles in the DNA DSB response (Scully et al., 1997). Rad51 is a homolog of the yeast RecA protein, which has been implicated in DSB repair and recombination

(Shinohara et al., 1992). BRCA2 mutant mice have been shown to be hypersensitive to agents which cause DNA double strand breaks, implicating the importance of the BRCA2 protein in this repair process (McAllister et al., 1997). BRCA1 nullizygous mouse embryonic stem cells have also been shown to be hypersensitive to ionizing radiation and

14 hydrogen peroxide, and are unable to repair oxidative DNA lesions (Gowen et al., 1998).

This data establishes BRCA1 as a caretaker gene responsible for helping to maintain genomic integrity, and loss of proper BRCA1 function contributes to breast cancer progression.

Mismatch repair and cancer: MSH and MLH proteins The DNA mismatch repair pathway was first characterized in E. coli and analogues of the bacterial proteins have been identified in eukaryotes. This pathway helps to maintain

DNA integrity during replication by repairing insertion/deletion mutations caused by DNA polymerase slippage, as well as repairing DNA mismatch mutations that are missed by the innate proofreading activity of the DNA polymerases (Charames and Bapat, 2003).

In E. coli, four main proteins are responsible for DNA mismatch repair – MutS,

MutH, UvrD and MutL. The MutS homodimer acts to recognize DNA mismatches, due to the bulges in the DNA which result (Cox et al., 1972; Wagner and Meselson, 1976).

Following binding of MutS, the MutL homodimer is recruited to the site of the mismatch

(Robertson et al., 2006). Formation of this tertiary complex results in the ATP-dependent activation of MutH, which is an endonuclease (Modrich, 1991). Using the methylation pattern of the DNA to distinguish the parental and daughter strands, the MutL-MutS-MutH complex incises the newly synthesized strand of DNA (Langle-Rouault et al., 1987). The helicase UvrD then unwinds the ends of the error-containing strand, allowing exonucleases to access the DNA and degrade the error-containing region (Oeda et al., 1982). Once this degradation is complete, a DNA polymerase resynthesizes the daughter strand and the nicks in the DNA are then sealed by DNA ligase.

In eukaryotes, the DNA mismatch repair system is more complex but still follows the same framework as in prokaryotes. To date, five homologs of the MutS protein have been identified, three of which have been found to play a role in eukaryotic mismatch repair – MSH2, MSH3 and MSH6. The other two homologs, MSH4 and MSH5, are involved in meiotic recombination (Hollingsworth et al., 1995; Kneitz et al., 2000). The

MutS homologues involved in MMR form two different heterodimers – MutSα and MutSβ.

MutSα is composed of MSH2 and MSH6 (Iaccarino et al., 1996), and acts to recognize base mismatches as well as small insertion/deletion (IDL) mutations (Alani, 1996). MutSβ, composed of MSH2 and MSH3, primarily recognizes larger IDLs (Habraken et al., 1996).

There is some redundancy between the two complexes, with both being able to recognize small IDLs.

As with MutS, a variety of homologues of MutL have been identified in mammalian mismatch repair. These include MLH1, MLH3, PMS2 and PMS1. The primary

MutL complex, MutLα, is a heterodimer of MLH1 and PMS2 (Li and Modrich, 1995).

Secondary MutL complexes include MutLβ, a heterodimer of MLH1 and PMS1 (Raschle et al., 1999) and MutLγ, a heterodimer of MLH1 and MLH3 (Lipkin et al., 2000). A variety of studies, using both animal models and in vitro work, have identified that the formation of a functional MutL complex is necessary for mismatch repair proficiency in mammalian systems, though the exact role the MutL complex plays is unknown.

Familial colorectal cancer syndromes

Familial Adenomatous Polyposis Familial adenomatous polyposis (FAP) is an autosomal dominant disease which has an estimated frequency of 1 in 8,000 to 1 in 15,000 births (Gryfe, 2006). Without a

16 prophylactic colectomy following diagnosis, the risk of developing cancer in these patients is nearly 100% by the fourth or fifth decade of life (Merg et al., 2005). Patients with FAP present with more than 100 colorectal polyps, which typically develop during puberty and may progress into adenomas.

Genetic linkage studies have associated FAP with a loss of the APC gene, which is encoded on chromosome 5q (Galiatsatos and Foulkes, 2006). APC, or adenomatous polyposis coli, is a tumor suppressor gene that works through the Wnt signalling pathway.

Briefly, APC forms a multi-protein complex with glycogen synthase kinase 3β (GSK-3β) and β-catenin, allowing GSK-3β to phosphorylate β-catenin (Su et al., 1993). This phosphorylation then leads to the degradation of β-catenin through the ubiquitin- proteasome pathway (Aberle et al., 1997). Without sufficient levels of functional APC protein, β-catenin levels can accumulate in the cytosol of cells, leading to continous activation of this pathway, even in the absence of Wnt.

Within the cell, β-catenin functions in two main ways – signalling and structural.

When levels in the cytosol are sufficient, it can translocate to the nucleus where it engages the transcription factors LEF/Tcf-4 to activate expression of target genes (Korinek et al.,

1997; Miyazawa et al., 2000). Downstream target genes of β-catenin include c-Myc (Wolf et al., 2002), cyclin D1 (Shtutman et al., 1999) and MMP-7 (Brabletz et al., 1999), ultimately promoting cell cycle progression, survival and invasion. A second function of β- catenin is its role in cell-cell junctions. It is an integral structural component in the formation of cadherin-based adherens junctions (Ozawa et al., 1989). The formation of these junctions contribute to the maintenance of polarized epithelial tissues (Meng and

Takeichi, 2009).

Several hotspots in the APC gene are commonly mutated, the two most prevalent being codons 1061 and 1309, which account for 11% and 13% of all germline mutations

(Al-Sukhni et al., 2008). It appears that some linkages exist between phenotype of the disease and the region of the APC gene which is affected, though there are compounding factors such as modifier genes, which affect this correlation (Nieuwenhuis and Vasen,

2007).

Lynch Syndrome A second major familial form of colorectal cancer arises from the genetic disease known as Lynch syndrome, or hereditary nonpolyposis colon cancer (HNPCC). The former nomenclature is preferred as these patients do develop polyps, and though CRC is common, it is not always the primary malignancy affecting a patient with the disease (Jass, 2006).

Other cancer types known to occur in patients with Lynch Syndrome include endometrial, ovarian, stomach and urinary tract (Lin et al., 1998; Maul et al., 2006; Peltomaki et al.,

2001; Watson and Lynch, 1993; Watson et al., 2008). This disease was first characterized in the 1970s as “Cancer Family Syndrome” by Dr. Henry Lynch and follows an autosomal dominant mode of transmission (Lynch and Krush, 1971).

Lynch syndrome accounts for approximately 2% of all CRC cases (Lynch et al.,

2006). The cardinal features of this disease are identified in the Amsterdam criteria, which is a list of criteria which, when met by a patient, indicates that he or she should be tested for Lynch Syndrome. These criteria include an early age of onset (<50 years), at least three family members who have developed a cancer known to be associated with the disease, and the exclusion of FAP as a potential explanation (Park et al., 1999).

Lynch syndrome is caused by germline mutations in the DNA mismatch repair

(MMR) family of genes. This family of proteins is responsible for recognizing DNA mismatches as well as small insertion/deletion mutations. Specifically, MLH1and MSH2 are mutated in 80-90% of patients with Lynch syndrome (Al-Sukhni et al., 2008). The functional redundancy of the other MMR proteins results in less severe phenotypes when these proteins are lost. Loss of these MMR proteins results in a phenotype referred to as microsatellite instability, presenting as changes in the length of repetitive regions of the genome, termed microsatellites. The lifetime risk for patients with Lynch syndrome to develop CRC is approximately 80%, as compared to 6.7% for the general population

(Chung and Rustgi, 2003).

Microsatellite Instability

Microsatellite instability was first identified as a unique mechanism in tumor development in the early 1990s, challenging the earlier discussed “Vogelstein model” as being a different pathway by which a cell may become cancerous. A group of researchers from California in the laboratory of Dr. Manuel Perucho began using arbitrarily-primed

PCR to amplify thousands of sequences from both tumors and normal tissue in patients.

This technique resulted in reproducible electrophoretic “signatures” for each tumor sample.

Initially the intent of this research was to determine regions of the genome which were missing in the tumors, but were present in the patient’s normal tissue, to potentially identify new tumor suppressor genes which may be contained within these lost regions. Deleted

DNA bands were noted in most of the colorectal tumor samples examined (Ionov et al.,

1993). Interestingly, however, approximately 12% of the tumors contained bands that were not deleted, but were shorter in length. Analysis of these bands revealed that they contained

19 simple repetitive sequences, termed microsatellites, and that the tumors with altered microsatellite regions had a distinct phenotype.

Microsatellites account for approximately 3% of the human genome, and can be found in both coding and noncoding regions (Subramanian et al., 2003). Microsatellites are commonly single-base or dinucleotide repeats, which makes them prone to insertion/deletion mutations due to slippage of DNA polymerases during replication. In a cell with functional DNA repair mechanisms, these IDLs are recognized and repaired by the cellular MMR. However, in a cell which has lost its MMR capabilities, these IDLs go unrepaired. Depending on where the specific microsatellite affected is located within the genome, this may or may not have a direct effect on the cell’s behaviour. In this sense, changes in the length of microsatellite regions of the genome are not necessarily a cause of tumorigenicity of the cell, but rather a symptom of a problem with MMR.

Most commonly, the MMR proteins lost in tumors displaying MSI are MLH1 or

MSH2 (Soreide et al., 2006). As discussed previously, these two proteins are arguably the most important in the repair pathway, due to the redundancy of the others. Loss of these proteins may occur in one of two ways: though an inherited germline mutation (most commonly occurring in a familial predisposition to cancer referred to as Lynch Syndrome) followed by loss of the second copy, most commonly due to promoter hypermethulation.

Such cases are referred to as inherited MSI, which occurs in approximately 92% of all

Lynch Syndrome tumors (Liu et al., 1996). Alternatively, a combination of mutation, gene deletion or gene silencing may result in the loss of MMR activity in patients without Lynch

Syndrome, referred to as sporadic MSI.

Diagnosing a given tumor as microsatellite stable or unstable has several implications for both the patient and their family. Microsatellite instability is most commonly observed in colorectal tumors, though it has also been reported in gastric (Kim et al., 2010), prostate (Bauer et al., 2011), breast (da Silva et al., 2010) and endometrial cancers (Clarke and Cooper, 2012). Tumors which are shown to be microsatellite unstable have a better prognosis than those which are microsatellite stable (Popat et al., 2005). As well, knowing the microsatellite status of a tumor has implications for the effectiveness of some chemotherapeutics, particularly those which act by inducing apoptosis in tumor cells by introducing bulky adducts or interfering with the fidelity of DNA replication. Finally, the development of a microsatellite-unstable tumor may be an indication that family members of the patient should be tested for this syndrome.

Diagnostic approaches In 1996, a set of criteria was developed to identify tumors which should be tested for microsatellite instability (Boland et al., 1998). These criteria are referred to as the

Bethesda Guidelines, and were revised in 2002 to reflect how the field had evolved since their inception (Umar et al., 2004). Among these guidelines are the recommendation for

MSI testing in colorectal cancer patients who are under 50 years of age, or if the tumor displays typical histology of a MSI-high tumor. A second set of diagnostic criteria associated with the diagnosis of MSI are referred to as the Amsterdam Criteria. In contrast to the Bethesda criteria, which are used to determine if an individual tumor should be tested for MSI, the Amsterdam criteria are used to determine whether a family should be tested for Lynch syndrome (Boland et al., 1998).

A panel of five microsatellite regions is tested to determine microsatellite status of a given tumor. These regions include two mononucleotide repeats, BAT-25 and BAT-26, and three dinucleotide repeats, D2S123, D5S346, and D17S250. If changes in repeat length in two or more of these markers are observed, the tumor is classified as highly microsatellite unstable, or MSI-H. If one marker shows changes, the tumor is classified as MSI-L, or microsatellite unstable-low. If no markers show changes, the tumor is classified as MSS, or microsatellite stable (Boland et al., 1998).

MSI and chemotherapy As mentioned previously, the microsatellite status of a tumor can have a large impact on the effectiveness of various chemotherapeutic drugs, particularly those which aim to induce apoptosis in tumor cells by activating the cellular responses to irreparable

DNA damage. One such drug is the chemotherapeutic agent widely used in the treatment of colon cancer, 5-fluorouracil (5FU).

5FU was first developed as an anti-cancer therapy in the 1950s, based on the observation that rat hepatomas utilized more uracil than normal tissues, implicating uracil metabolism as a potential therapeutic target (Rutman et al., 1954). Clinically, 5FU is a component of the common therapeutic regime FOLFOX, alongside folinic acid (leucovorin) and oxaliplatin. This regime is the standard of care in adjuvant colorectal cancer therapy

(Zaanan et al., 2011). Used alone, 5FU has a response rate (as determined by measuring tumor volume of human patients) of approximately 10-20% (Gailani et al., 1972), although combination therapies increase this to as much as 40% in some studies (Douillard et al.,

2000).

5FU is an analogue of the base uracil, with the replacement of the hydrogen atom at the C-5 position with fluorine. This close similarity allows the drug to enter the cell using the same transporters as does uracil (Wohlhueter et al., 1980). Once inside the cell, 5FU is converted into several different metabolites: fluorodeoxyuridine monophosphate (FdUMP), fluorodeoxyuridine triphosphate (FdUTP), and fluorouridine triphosphate (FUTP). These metabolites act in two major ways to prevent cell growth and induce apoptosis: they inhibit the nucleotide synthetic enzyme thymidylate synthase (TS), and they misincorporate into

DNA and RNA (Danenberg et al., 1974).

Thymidylate synthase is the enzyme responsible for the reductive methylation of deoxyuridine monophosphate (dUMP) to deoxythymidine monophosphate (dTMP), with a metabolite of the B vitamin folate serving as a methyl donor. Depletion of cellular stores of dTMP results in subsequent depletion of TTP. This results in a form of cellular death referred to as “thymineless death”, which was initially characterized in bacteria (Rolfe,

1967). This type of apoptosis is associated with strand breaks in the DNA, which may be part of the initiating events of 5FU-mediated apoptosis (Goulian et al., 1986). Alongside the depletion of dTTP in the cell, there is also an accumulation of dUTP. This leads to increased uracil misincorporation into DNA during DNA synthesis, which is recognized by the nucleotide excision repair enzyme uracil-DNA-glycosylase (Lindahl, 1974). However, given the high intracellular dUTP concentration in the 5FU-treated cells, these repair attempts are futile and eventually lead to DNA strand breaks and apoptosis (Ladner, 2001).

Models of microsatellite instability

In vitro models Initial studies of mismatch repair were performed in Escherichia coli and

Saccharomyces cerevisiae, and were expanded to studies of human cells and systems in the

1990s. One of the first models used was the colorectal cancer cell line HCT116. This cell line was observed to introduce insertion and deletion mutations into a vector containing a

(CA)n repeat tract (Parsons et al., 1993). This DNA mismatch repair deficiency is due to a premature stop codon in the MLH1 gene, which is located on chromosome 3

(Papadopoulos et al., 1994). DNA mismatch repair proficiency can be restored in this cell line through introduction of chromosome 3, allowing for its use as a model of the role of

MLH1 in DNA mismatch repair (Koi et al., 1994).

A second colorectal cancer cell line, LoVo, allows for the study of the other major mismatch repair protein, MSH2. This cell line harbors a large deletion in MSH2, spanning from exon 3 to exon 8 (Liu et al., 1995). Much like with HCT116, complementation studies with LoVo have been performed, and stable transfer of chromosome 2 (containing MSH2) has restored MSH2 expression in this cell line, resulting in the use of LoVo as a model for loss of this protein (Watanabe et al., 2000). These in vitro models have been used to investigate a variety of phenomena related to DNA mismatch repair proficiency, including chemosensitivity, mutation rate, and interactions and signalling of the various MMR proteins (Meyers et al., 2001; Mure and Rossman, 2001).

A limitation of both of these models is that an entire chromosome is reintroduced, rather than just the gene of interest. This may complicate the interpretation of the observed results due to the introduction of another copy of the other genes on the chromosome.

Some studies have attempted to overcome this problem by just introducing the cDNA for the gene of interest, as has been done with Mlh1 (Jacob et al., 2001). This approach introduces problems of its own, as the promoters on these plasmids are not endogenous and therefore MLH1 expression is not at endogenous levels. In fact, overexpression of MLH1 and MSH2 has been shown to induce apoptosis in some cases (Zhang et al., 1999).

In vivo models Nullizygous mouse models for all of the major DNA mismatch repair genes have been developed, including Pms2 (Baker et al., 1995), Mlh1 (Baker et al., 1996), Msh2

(Reitmair et al., 1995), Msh3 (Edelmann et al., 2000) and Msh6 (Edelmann et al., 1997).

Comparisons of these models demonstrated the relative importance of the different MMR genes, with the Mlh1 and Msh2 knockout mice having the most severe phenotypes, and

Msh3 knockout mice having the least severe (Hegan et al., 2006). These studies also demonstrated the specific types of mismatch repair in which each of these genes was involved, such as repairing transitions and transversions, or IDLs. Crossing of these mouse models has also allowed development of animals missing two or more MMR genes, producing in some cases even more severe phenotypes, such as Msh3-/- Msh6-/- mice having mutation frequencies >100-fold higher than wild-type (Hegan et al., 2006), demonstrating the importance of a functional MutS complex in vivo.

One feature common to all of these mouse models is their propensity to develop lymphoma rather than colorectal tumors. While in some cases, mice do develop intestinal neoplasms associated with Apc inactivation and increased age, the predominance of a lymphoma phenotype has limited the use of these animals as preclinical models (Reitmair et al., 1996). The development of a tissue-specific Msh2 knockout mouse model allowed

25 for the study of animals which developed colorectal neoplasms, rather than lymphomas

(Kucherlapati et al., 2010). In this case, the MMR deficiency is limited to the intestinal epithelium, with 89% of animals developing intestinal tumors by 10 months of age.

Development of this model is a better recapitulation of what may occur in Lynch syndrome patients, though it is still not a perfect system. For example, these tissue-specific knockout mice primarily develop tumors in the small intestine, while patients with Lynch syndrome typically develop tumors of the colon.

RATIONALE

This thesis aimed to follow previous work in the laboratory, which established derivatives of the MMR-proficient CaCo-2 human colorectal cancer cell line with siRNA- mediated knockdown of the mismatch repair proteins MLH1 or MSH2. The observation of increased tumor growth rate in these knockdown cells when xenografted into mice prompted investigation into the types of alterations that may have occurred in these cells during growth in vivo, to explain their increased tumorigenicity when compared to controls.

We hypothesize that growth in vivo was a more stressful environment for the cells and may have led to the development of point mutations and microsatellite instability in cells in which mismatch repair capabilities had already been compromised. Therefore, this led to the first objective:

Characterize clonal cell lines isolated from MLH1/MSH2 knockdown tumors

for the development of point mutations and microsatellite instability following

growth in vivo

Loss of one member of the DNA mismatch repair family of proteins is a key initiating event in some forms of cancer development. This is especially true for patients with Lynch syndrome, in which one allele of a DNA mismatch repair gene is compromised with a germline mutation. Given that most DNA MMR proteins function as heterodimers, it is possible that loss of one protein may have effects on its binding partners. We hypothesized that reduction of MLH1 or MSH2 may have also resulted in reduction of their major binding partners, namely PMS2, MSH3, and MSH6. From this, the second objective of this thesis was developed:

Characterize levels of other DNA MMR proteins in cells in which MSH2 or

MLH1 has been knocked down

Sufficient DNA mismatch repair mechanisms are believed to be required for the efficacy of various chemotherapy drugs, including the antimetabolite 5-fluorouracil. We hypothesized that cells in which MMR capability had been compromised due to reduction of MLH1 or MSH2 would be resistant to the common chemotherapeutic drug 5- fluorouracil. For this reason, my final objective is:

Measure with a clonogenic survival assay the repopulating ability of colorectal

cancer cells with reduced MLH1 or MSH2 expression when exposed to

5-fluorouracil

MATERIALS AND METHODS

All supplies and materials used are detailed in Appendix I, and preparation of materials is detailed in Appendix II.

Tissue culture

Cells were grown in standard culture conditions. Cells were maintained at 37°C in a humidified atmosphere containing 5% CO2, in complete medium consisting of DMEM

(Sigma-Aldrich) supplemented with 10% FBS (Invitrogen), 50 μg/ml gentamicin

(Invitrogen) and 1 mM sodium pyruvate (Sigma-Aldrich). Medium was changed every two to three days and cells were passaged when confluency reached 90% or above.

Passaging consisted of a rinse with PBS (Sigma-Aldrich) and lifting of cells using trypsin. Three millilitres of pre-warmed trypsin was added for a tissue culture dish with a diameter of 10 cm, and cells were moved to the incubator for approximately 5 minutes to allow for sufficient detachment. Trypsin was then inactivated using complete medium and cells were pelleted at 350 x g for 4 minutes prior to replating in sufficient amounts of complete medium. Cells were typically split at a 1 in 5 ratio.

DNA isolation and quantification from cells

Genomic DNA was isolated from cells using the DNeasy Blood and Tissue Kit

(Qiagen) according to the manufacturer’s protocol. Following DNA isolation, DNA concentration and purity were determined using a Nanodrop ND-1000 (Thermo Scientific) to measure absorbance at 230, 260 and 280 nm.

Protein isolation and quantification from cells

To collect total cellular protein, cell pellets were resuspended with vortexing and pipetting in appropriate volumes (300 μL per confluent 10 cm plate) of cell lysis buffer containing 1% Triton X-100 (Cell Signalling). Following resuspension, samples were incubated on ice for 10 minutes and then centrifuged for 15 minutes at 4°C at 12500 x g.

Supernatent was collected, aliquoted, and stored at -80C until use.

Protein quantification was done using a Bio-Rad protein quantification system. To quantify samples, an aliquot of 5 μl of each sample was loaded into a 96-well plate in duplicate and combined with 25 μl of a 1:50 dilution of Reagent S with Reagent A. To this,

200 μl of Reagent B was added. Samples were the incubated at room temperature with shaking for 10 minutes prior to reading the absorbance at 630 nm using an ELx800

Universal Microplate Reader (BIO-TEK Instruments Inc). Solutions of BSA ranging from

0 to 5 mg/ml were prepared in lysis buffer and used to generate a standard curve, from which sample concentrations were interpolated.

SDS-PAGE and Western blotting analysis

Samples of 10-50 μg of protein were combined with water and protein sample buffer containing SDS and β-mercaptoethanol to allow for denaturation of protein.

Samples were then incubated at 95°C for 5 minutes to allow for denaturation of proteins.

Following this, samples were briefly vortexed and then quickly spun down to collect condensation. Lysates were then loaded onto a 7.5% acrylamide gel and run at 120V using the BioRad Mini Protean system until the dye front reached the bottom of the gel. The

BioRad PageRuler Plus Prestained Protein Ladder was also included (5 μL) to allow for protein size determination.

Following electrophoresis, gels were incubated in cold transfer buffer for 10 minutes prior to transfer to a PVDF membrane. Membranes were activated for 60 seconds in methanol and also allowed to incubate in transfer buffer for 10 minutes prior to cassette assembly. Transfer occurred at 100V for 2 hours using a wet transfer system, and successful transfer was determined by staining of membranes with amido black until bands appeared, followed by destaining using methanol. Membranes were then blocked for 30 minutes at room temperature with shaking in 5% skim milk (w/v) prepared in 0.1% TBST.

Following blocking, membranes were incubated with primary antibody overnight at 4°C with shaking. Primary antibodies were prepared in 5% skim milk (w/v) prepared in 0.1%

TBST at the following dilutions: anti-α-tubulin (Sigma-Aldrich) 1:600,000, anti-MSH2

(Calbiochem) 1:1000, anti-MLH1 (BD Biosciences) 1:1000, anti-MSH3 (Santa Cruz)

1:1000, anti-PMS2 (Abnova) 1:1000, anti-MSH6 (Bethyl Laboratories) 1:2000. The next day, membranes were washed three times for 10 minutes each wash with 0.1% TBST with shaking, and then incubated with HRP-labelled secondary antibodies diluted in 5% skim milk prepared in 0.1% TBST at the following concentrations: anti-mouse HRP (Sigma-

Aldrich) 1:20,000, anti-rabbit HRP (Sigma-Aldrich) 1:10,000. Incubations with secondary antibody were performed at room temperature for one hour with shaking. Membranes were washed again as following primary antibody incubation.

To visualize bands, membranes were incubated in enhanced chemiluminescent

(ECL) substrate (Roche) for 1 minute. Bands were then exposed using X-ray film (Kodak) and an X-ray film processor (Konica Minolta) was used for development. Following processing, blots were scanned and densitometric analysis performed using ImageJ software (National Institutes of Health).

Detection of somatic mutations with Sequenom OncoCarta 1.0 panel

Genomic DNA samples were diluted to 10 ng/μl in nuclease-free water (Ambion), then shipped on dry ice to Sequenom, Inc. in San Diego, California. Analysis was performed at that facility. Reanalysis of duplicate samples was done using the same technology at the Ontario Institute of Cancer Research in Toronto, Ontario.

Determination of BAT-25 and BAT-26 repeat lengths

BAT-25 and BAT-26 repeat lengths were determined using tRFLP analysis.

Primers were designed to encapsulate these repeat regions, and PCR was performed to amplify these regions. Primer sequences can be found in Appendix III, and the forward primer of each set was labelled with the fluorescent dye 6-FAM. In one reaction (50 μl volume), 10 ng of genomic DNA was combined with 0.2 mmol/L deoxynucleotide triphosphates (Fermentas), 0.5 μmol/L of each primer (Sigma-Aldrich) and 0.02 U/μl of

Phusion High-Fidelity DNA polymerase (New England Biolabs). The amplification program involved an initial denaturation cycle of 98°C for 2 minutes, followed by 36 cycles of 98°C for 30 seconds, 58°C for 30 seconds, and 72°C for 30 seconds. This was followed by a final elongation step of 72°C for 5 minutes. PCR reactions were stored at

4°C until further analysis was performed.

Amplicons were run out using standard gel electrophoresis techniques in a 2% agarose gel made in 1xTAE buffer. PCR reaction volumes were combined with DNA sample buffer consisting of 50% glycerol combined with Orange G dye before being run at

100V until adequate separation was obtained. Successful amplification was confirmed by visualizing bands using ethidium bromide staining of DNA. PCR products were then excised from the gel and isolated using a commercially available PCR purification kit

(Qiagen). Quantification of amplicons was done using a Nanodrop ND-1000 reading absorbance at 260 nm. Samples then underwent tRFLP analysis at the University of

Guelph’s Laboratory Services Division.

Detection of BRAF V600E mutations in clonal cell lines

Detection of BRAF V600E mutations was done using PCR analysis. PCR was used to amplify a region of the genome where the mutation would be found. Conditions for PCR amplification, agarose gel electrophoresis, and fragment purification are the same as was described for BAT-25 and BAT-26 amplification. Following successful amplification,

DNA fragments were sequenced at the University of Guelph’s Laboratory Services

Division. Genomic DNA from the melanoma cell line WM115 was also analyzed as a positive control, as this cell line is known to harbour the V600E mutation (Maddodi et al.,

2010).

Determination of clonal cell survival following treatment with 5-fluorouracil

Sensitivity of MLH1 and MSH2 knockdown cells to apoptosis caused by exposure to 5-fluorouracil was examined using a clonogenic survival assay. Cells were trypsinized as described previously and pelleted with centrifugation before resuspension in sterile PBS.

Cell suspensions were passed through a 20 g needle four times to ensure a single-cell suspension. Cells were then counted using the Countess® Automated Cell Counter (Life

Sciences), and a range of 500 to 5,000 cells were plated in 10 cm tissue culture dishes in standard DMEM medium. The following day, media was changed to standard DMEM supplemented with 2.5 or 5 μM 5-fluorouracil, with dishes containing 500 and 1000 cells left untreated to control for plating efficiency of the cell lines examined.

For acute exposures of the drug, the medium was changed 24 hours after exposure to standard complete DMEM. For chronic exposures, medium was changed every 3 days with medium containing fresh drug. Cultures were maintained until visible colonies of >50 cells formed, approximately 3 weeks. Colonies were visualized by first rinsing plates three times with PBS, and then stained and fixed by addition of a solution of 0.5% crystal violet

(w/v) prepared in 25% methanol. Following 10 minutes of staining, crystal violet solution was removed, and plates were washed with tap water. Plates were then scanned and images were adjusted using Paint Shop Pro 7 (Jasc Software Inc) to improve contrast. Colonies were counted using NICE (NIST’s Integrated Colony Enumerator) software (Clarke et al.,

2010). Cell counts of treated plates were normalized to counts for untreated controls.

Statistical analysis

Statistical analysis was done using PRISM (GraphPad Software). Means from each sample were plotted and standard error was used to plot error bars. For correlation experiments, statistical tests to determine if the slope of the regression line was significantly different from zero were performed. Results were considered statistically significant if the p-value was less than 0.05.

RESULTS

Knockdown of MLH1 or MSH2 increases tumorigenicity of Caco-2 cells in vivo

The MLH1 or MSH2 knockdown cells generated by Kristen Lacombe were xenografted into nude mice by Dr. Brenda Coomber and tumor volume was measured over a period of up to 118 days (Figure 1). Caco-2 cells transfected with a scrambled siRNA sequence were also used for xenografts to serve as a negative control. When compared to the scrambled controls, both MLH1 and MSH2 knockdown cells showed increased tumor take rates - 60% for MLH1 knockdown, 80% for MSH2 knockdown, versus 20% for scrambled. The knockdown cell-derived tumors reached a palpable stage much more quickly than scrambled controls, indicating decreased tumor latency. The slope of the lines for the MLH1 and MSH2 knockdown-derived tumors appear to be steeper than for the scrambled control cell line, indicating an increased tumor growth rate. It is also worth noting that the growth characteristics observed for the scrambled control tumor follow the typical wildtype Caco-2 pattern observed previously in our laboratory (Shahrzad et al.,

2005), indicating that the inclusion of the vector containing the siRNA sequences likely did not influence the growth patterns observed.

Knockdown of MLH1 is lost following growth in vivo

Following growth in vivo, clonal cell lines were isolated from the tumors to allow for further analysis. These cell lines were first examined to check for the maintenance of the expected MMR protein knockdown. This was done to ensure that the cells responsible for the tumor growth observed were those with reductions of either MLH1 or MSH2 protein levels, and to ensure that the knockdown was not lost during growth in vivo.

In the case of the clonal cell lines derived from the MLH1 knockdown tumors, the

MLH1 knockdown was not maintained following growth in vivo, as determined by western blotting analysis of whole cell lysates (Figure 2). Lysates from wildtype CaCo2 cells were also included to demonstrate MLH1 levels in cells which had not been grown in vivo. As well, clonal cell lines were derived from the tumor resulting from xenografting the CaCo2 cells containing the scrambled siRNA sequence. These were also included in the western blot analysis to examine any potential effects that growth in vivo may have had on the levels of these MMR proteins, as well as any effects that the introduction of the plasmid and generation of siRNA may have had on cellular behaviour. Densitometric analysis

(Figure 2B) was performed on the western blots to quantify MLH1 levels in the various tumor-derived cell lines relative to α-tubulin. There was variation of MLH1 levels observed in the different clones (ranging from 21-90% of wildtype levels), but similar levels of variation were also seen in the cell lines derived from the scrambled control tumors (27% and 64% of wildtype levels). Therefore, it is reasonable to conclude that the surviving cells isolated from the tumors have lost their MLH1 knockdown.

Knockdown of MSH2 is maintained following growth in vivo

As with the MLH1 knockdown tumors, MSH2 levels were examined in the clonal cell lines derived from the MSH2-knockdown tumors (Figure 3). Figure 3A shows a representative western blot of MSH2 levels in the knockdown clones, and is accompanied by the densitometric analysis, normalized to tubulin with wildtype CaCo-2 levels set to 1

(Figure 3B). In contrast to the MLH1 knockdown tumors, clones derived from these tumors show that MSH2 knockdown was maintained following growth in vivo in all clones examined, ranging from 0.6% to 30% of wildtype levels. No obvious changes in MSH2

36 levels in clones derived from scrambled controls were observed, indicating that this effect is likely not due to growth in vivo, but rather to the transfected plasmid in these cells.

No point mutations detected in key oncogenes following growth in vivo in knockdown cells

The development of mutations in key oncogenes associated with colorectal cancer progression was examined in clones isolated from MLH1 and MSH2 knockdown tumors following growth in vivo. BRAF V600E mutations were detected using traditional sequencing methods following PCR amplification (Figure 4). As well, a large panel of oncogenes was examined for point mutations by the Sequenom Mass Array OncoCarta v1.0 panel, with a detection threshold of 10%. A full list of the genes and mutations examined by this panel can be seen in Table 1. No point mutations were detected in the 16 clones examined by either method. This result was confirmed with reanalysis of select clones at the Ontario Institute for Cancer Research (OICR) in Toronto.

Figure 1: Growth of MLH1 and MSH2 knockdown cells when xenografted into mice. Tumor growth was measured over a period of up to 120 days. For each cell line, five mice were injected with 5 x 106 cells and points represent average tumor volume +/- SEM. Percentages represent percent of mice in which tumors successfully developed (tumor take rate).

Figure 2: MLH1 knockdown is lost following growth in vivo. A) Western blot analysis of MLH1 levels in various clonal cell lines. Cell lines were isolated from tumors formed when MLH1 knockdown cell lines were used for xenografts. Wildtype Caco-2 (Lane 1) and cell lines established from scrambled control tumors (Lanes 13 and 14) represent controls. B) Densitometric analysis of the above western, showing no substantial variation in cell lines derived from MLH1 knockdown tumors when compared to cell lines derived from scrambled control tumor. Only one replicate was completed as cells were untreated, and thus no statistical analysis was performed.

Figure 3: MSH2 knockdown is maintained following growth in vivo. A) Western blot analysis of MSH2 levels in various clonal cell lines. Cell lines were isolated from tumors formed when MSH2 knockdown cell lines were used for xenografts. Wildtype Caco-2 (Lane 1) and cell lines established from scrambled control tumors (Lanes 13 and 14) represent negative controls. Values represent percent knockdown as compared to wildtype Caco-2. B) Densitometric analysis of the above western, showing decrease of MSH2 levels in cell lines isolated from tumors when compared to wildtype cells and scrambled control cell lines.

Figure 4: BRAF V600 mutation screening in knockdown clones. Alignment of sequencing results of the BRAF gene to detect mutations in codon 600. * indicates alignment between nucleotides in all samples. Negative control (CaCo2) shows the wildtype sequence, and the melanoma cell line WM115 was used as a positive control, as it has been shown to harbor a mutation in BRAF (Benlloch et al., 2006).

Table 1: Amino acid changes and small insertion/deletion mutations queried by Sequenom OncoCarta v1.0 panel. Gene Mutation Gene Mutation Gene Mutation ABL1 G250E BRAF G469V ERBB2 P780_Y781insGSP Q252H D594V A775_G776insYVMA Y253F CDK R24C P780_Y781insGSP Y253H R24H S779_P780insVGS E255K EGFR R108K FGFR1 S125L E255V S768I P252T D276G V769_D770insASV FGFR3 K650Q F311L V769_D770insCV K650E T315I D770_N771>AGG G370C F317L D770_N771insG Y373C M351T N771_P772>SVDNR A391E E355G P772_H773insV K650Q F359V H773>NPY K650E H396R H773_V774insNPH K650T AKT1 rs11555435 H773_V774insPH K650M rs11555431 V774_C775insHV FLT3 I836del rs11555432 T263P D835H rs12881616 T790M D835Y rs11555433 L858R HRAS G12V rs11555436 L861Q G12D rs34409589 A289V G13C AKT2 S302G G598V G13R R371H E709K Q61H BRAF G464R E709H Q61L F595L E709A Q61P G596R E709G Q61R L597S E709V Q61K L597R G719S JAK2 V617F L597Q G719C KIT V559del T599I G719A D816H V600E M766_A767insAI D816Y V600K E746_T751del V560D V600L A750P V560G V600R T751A M552L K601N T751I V559I K601E E746_T751del, I ins Y568D G464V E746_A750del V559A G464E E746_T751del, V ins D52N G466R L747_T750del, P ins V559_V560del F468C L747_S752del, Q ins V560del G469S S752_I759del P551_V555del G469A ERBB2 L755P Y553_Q556del G469E G766S Y570_L576del G469R G766LC E561K G469S G766VC L576P

Gene Mutation Gene Mutation KIT D579del NRAS G12A Y503_F504insAY G12D K642E G12R D816V G12C D816H G12S K550_K558del G13S V825A G13C E839K G13R F584S Q61H W557R Q61E W557G Q61K V559D PDGFRA V561D V559G I843_S847>T V559I D846Y K558_V560del D842V K558_E562del T674I KRAS Q61E F808L Q61K N870S G12V D1071N G12A D842_H845del G12D I843_D846del G12S S566_E571>K G12R PIK3CA C420R G12C P539R G13V R88Q G13D H1047Y A59T R38H Q61L C901F Q61P M1043I Q61R N345K Q61H E542K MET R970C E545K T992I Q546K Y1230C H701P Y1235D H1047R M1250T H1047L NRAS Q61L RET C634Y Q61R C634R Q61P C634W G13V E632_L633del G13A M918T G13D A664D A18T G12V

No changes in BAT-25 or BAT-26 repeat length in clones following growth in vivo

The development of microsatellite instability following growth in vivo was examined in the clonal cell lines isolated from the tumors. BAT-25 and BAT-26 repeat lengths were determined using tRFLP analysis following PCR amplification with fluorescently labelled primers. Genomic DNA isolated from the colorectal cancer cell line

HCT116 was used as a positive control (Figure 5C, Figure 6C), as this cell line has been shown to be microsatellite unstable at these two loci (Liu et al., 2004). Wildtype CaCo-2 cell genomic DNA was used as a negative control (Figure 5A, 6A), as a cell line which does not show microsatellite instability. The original knockdown cell lines were also examined to control for any effects due to knockdown of MLH1 or MSH2 (Figure 5B, 6B), but not further exacerbated by growth in vivo. Finally, clones from the tumors arising from cells containing the scrambled siRNA sequence were also examined to control for any changes in microsatellite length due to growth in vivo that were unrelated to the knockdown (Figure 5E, 6E). Of all clones examined, no changes in microsatellite length in either BAT-25 or BAT-26 were observed. Representative results are shown in Figure 5D,

5F, 6D, and 6F. Previous work in the laboratory has shown that the mutation detection threshold of this technique is 5% (Lacombe, 2009).

A B

C D

E F

Figure 5: BAT-25 repeat length in clonal cell lines developed from MLH1 and MSH2 knockdown tumors. Figure shows peaks produced from BAT-25 PCR products labelled with a 6-FAM fluorescent tag. The top axis shows size of products in base pairs, peaks represent fluorescent intensity. A) shows microsatellite-stable Caco-2 cells. B) shows representative result from knockdown cells prior to growth in vivo. C) shows HCT116, which is MSI and shows a shift in BAT-25 repeat length. D), E) and F) show representative results from MLH1, MSH2, and scrambled tumor-derived cell lines. No noticeable shifts in peaks were observed in any of the clonal cell lines examined.

A B

C D

E F

Figure 6: BAT-26 repeat length in clonal cell lines developed from MLH1 and MSH2 knockdown tumors. Figure shows peaks produced from BAT-26 PCR products labelled with a 6-FAM fluorescent tag. The top axis shows size of products in base pairs, peaks represent fluorescent intensity. A) shows microsatellite-stable Caco-2 cells. B) shows representative result from knockdown cells prior to growth in vivo. C) shows HCT116, which is MSI and shows a shift in BAT-26 repeat length. D), E) and F) show representative results from MLH1, MSH2, and scrambled tumor- derived cell lines. No noticeable shifts in peaks were observed in any of the clonal cell lines examined.

Loss of MSH2 is accompanied by loss of PMS2, MSH3 and MSH6

Following the negative results for both point mutation and microsatellite instability development in the knockdown cells following growth in vivo, other members of the DNA mismatch repair protein family were examined for potential compensatory upregulation.

Normalized MSH2 levels in the various MSH2 knockdown clonal cell lines positively correlated with normalized PMS2 levels in those same cells (Figure 7B) with a r-squared value of 0.4708 and a significant p-value for the slope of the regression line of 0.0412.

The two major binding partners of MSH2 are MSH6 and MSH3. Together, these functional heterodimers form the MutS protein complexes, with MSH6 being the more abundant (Boland and Goel, 2010). As shown with western blotting, loss of MSH2 in the various clones isolated from the MSH2 knockdown tumors is accompanied by a proportionate loss in MSH3 (Figure 8) and MSH6 (Figure 9). The correlations between

MSH2 levels and MSH3 and MSH6 levels in these clones were shown to be statistically significant, with p-values of 0.0017 and 0.0427, respectively.

y = 1.112x + 1.048 p-value = 0.0412 R-squared = 0.4708

Figure 7: Correlation of MSH2 and PMS2 levels in MSH2 knockdown clones following growth in vivo. Western blotting analysis (A) was done to determine PMS2 levels in MSH2 knockdown clones. These levels were normalized to tubulin levels, and then correlated with normalized MSH2 levels in the same clones and the resulting data graphed (B). For both MSH2 and PMS2, wildtype CaCo2 levels were set to 1. Linear regression analysis was performed, and the slope was found to be significantly different from 0, with a p-value of 0.0412.

y = 1.004x – 0.0523 p-value = 0.0017 R-squared = 0.7747

Figure 8: Correlation of MSH2 and MSH3 levels in MSH2 knockdown clones following growth in vivo. Western blotting analysis (A) was done to determine MSH3 levels in MSH2 knockdown clones. These levels were normalized to tubulin levels, and then correlated with normalized MSH2 levels in the same clones and the resulting data graphed (B). For both MSH2 and MSH3, wildtype CaCo2 levels were set to 1. Linear regression analysis was performed, and the slope was found to be significantly different from 0, with a p-value of 0.0017.

y = 1.029x + 0.0856 p-value = 0.0457 R-squared = 0.4113

Figure 9: Correlation of MSH2 and MSH6 levels in MSH2 knockdown clones following growth in vivo. Western blotting analysis (A) was done to determine MSH6 levels in MSH2 knockdown clones. These levels were normalized to tubulin levels, and then correlated with normalized MSH2 levels in the same clones and the resulting data graphed (B). For both MSH2 and MSH6, wildtype CaCo2 levels were set to 1. Linear regression analysis was performed, and the slope was found to be significantly different from 0, with a p-value of 0.0457.

Cells with reduced MLH1 or MSH2 show resistance to chronic exposure to low doses of 5-fluorouracil

Following chronic low-dose (5 μM) exposure to the chemotherapeutic drug 5- fluorouracil, the derivatives of CaCo-2 with reduced MLH1 and MSH2 showed increased colony formation as compared to cells transfected with the scrambled control plasmid, with p-values < 0.05 for Tukey’s comparisons of control vs MLH1 knockdown, and control vs

MSH2 knockdown (Figure 12). Experiments with lower doses of 5-fluorouracil (2.5 μM) produced a trend towards resistance, but this was not found to be statistically significant following three biological replicates (p-value = 0.1053 for one-way ANOVA) (Figure 11).

Acute exposures of 24 hours to higher doses of 5-fluorouracil (10 μM) did not appear to produce any discernible trends in resistance (Figure 10).

Figure 10: Clonogenic cell survival in MLH1 and MSH2 knockdown cells exposed to continuous 5 μM 5-fluorouracil. Clonogenic cell survival rate was calculated for each cell line following treatment by normalizing to untreated controls. Colonies were counted using NICE (NIST’s Integrated Colony Enumerator) software (Clarke et al., 2010). Values are shown as mean ± SE, n = 3 for each cell line. ANOVA analysis showed that at least one group was significantly different from the others, and a Tukey’s multiple comparison test showed that both MSH2 and MLH1 knockdown cell clonogenic survival rates were higher than scrambled controls, with p-values less than 0.05.

Figure 11: Clonogenic cell survival in MLH1 and MSH2 knockdown cells exposed to continuous 2.5 μM 5-fluorouracil. Clonogenic cell survival rate was calculated for each cell line following treatment by normalizing to untreated controls. Colonies were counted using NICE (NIST’s Integrated Colony Enumerator) software (Clarke et al., 2010). Values are shown as mean ± SE, n=3 for each cell line. ANOVA analysis did not show any significant differences between cell lines.

Figure 12: Clonogenic cell survival in MLH1 and MSH2 knockdown cells exposed to 10 μM 5-fluorouracil for 24 hours. Clonogenic cell survival rate was calculated for each cell line following treatment by normalizing to untreated controls. Colonies were counted using NICE (NIST’s Integrated Colony Enumerator) software (Clarke et al., 2010). Values are shown as mean ± SE, n=3 for each cell line. ANOVA analysis did not show any significant differences between cell lines.

DISCUSSION

In this study, we aimed to explore the effect that knockdown of mismatch repair proteins may have played in tumor progression in vivo. Previous work had established derivatives of the Caco-2 human colorectal cancer cell line with siRNA-mediated knockdown of the mismatch repair proteins MLH1 and MSH2 (Lacombe, 2009). These two proteins were chosen because they are the most commonly lost DNA mismatch repair proteins in cancer which develops with a microsatellite unstable phenotype (Soreide et al.,

2006). The Caco-2 cell line was used for the development of this model for a variety of reasons, including that it is microsatellite stable and does not have mutations in genes that are commonly mutated in microsatellite-unstable colorectal cancers (Oliveira et al., 2003).

As well, its phenotype closely resembles normal human intestinal mucosa and it has been frequently used to model this in vitro (Eilers et al., 1989).

When characterized following growth in vitro, these knockdown cells displayed no signs of developing microsatellite instability, a type of genomic instability that results from loss of DNA mismatch repair proteins. MSI status was evaluated by looking for changes in repetitive regions of the genome, termed microsatellites, due to slippage of DNA polymerases during DNA replication. In this case, BAT-25 and BAT-26 mononucleotide repeat regions were examined. Previous work has shown that these two markers are able to detect microsatellite instability when the tumor cell content of the sample is as low as 10%

(Brennetot et al., 2005).

While repetitive DNA regions are more obviously susceptible to mutation in cells with defective DNA mismatch repair, mutations can also occur in non-repetitive regions.

Activating point mutations in either BRAF or KRAS were detected using standard DNA

55 sequencing techniques. These two oncogenes are likely targets of point mutations in colorectal cancer, and in some cases are used alongside microsatellite testing when determining the MSI status of a tumor, and whether it is familial in origin (Loughrey et al.,

2007). As these two proteins are members of the same family and signal through the same pathway, mutations are generally mutually exclusive (Rajagopalan et al., 2002), suggesting that the two genes provide the same selective advantage when mutated. Both are involved in the activation of the RAS/RAF/MAPK pathway, which ultimately is involved in regulating cellular proliferation, and is a crucial pathway in colorectal carcinogenesis

(Davies et al., 2002).

The most common mutation observed in BRAF is V600E, which accounts for over

90% of mutations observed in this gene in CRC (Deng et al., 2004). This mutation has also been reported to be present in ~35% of sporadic MSI-H cases (Loughrey et al., 2007), but has as of yet, never been seen in a case of Lynch Syndrome MSI-H CRC (McGivern et al.,

2004). This has led to proposed inclusion of screening for this mutation as a negative indicator of Lynch Syndrome status (Domingo et al., 2004). Specifically, the BRAF V600E mutation has been found to be positively correlated with hMLH1 promoter hypermethylation (Koinuma et al., 2004), making it an attractive target for mutation in the

MLH1 knockdown cells in particular. Mutations in KRAS have been reported to occur in

~20% of sporadic MSI-H tumors, as well as in ~35% of MSS tumors negative for BRAF mutations (Losi et al., 1997). KRAS mutations have also been observed to occur at approximately equal frequencies in familial and sporadic MSI-H tumors (Aaltonen et al.,

1993), a stark contrast to the pattern seen with BRAF mutations.

Both MLH1 and MSH2 knockdown Caco-2 cells tested negative for the BRAF

V600E mutation, as well as mutations in codons 12 or 13 of KRAS, following growth in standard tissue culture conditions, or following low-oxygen and low-glucose exposures in vitro (Lacombe, 2009). Combined, the lack of BAT-25 and BAT-26 repeat lengths, along with lack of expected mutations in BRAF and KRAS, led to the expectation that these cells would likely not form tumors very quickly or aggressively when grown in vivo. However, when xenografted into mice, the MLH1 and MSH2 knockdown Caco-2 cells showed increased tumor take rate, increased tumor growth, and decreased tumor latency.

Following growth in vivo, a variety of clonal cell lines were developed from the tumors and used to investigate potential effects that growth in vivo may have had to explain the increased tumorigenicity of the knockdown cells compared to controls. All of the previous assays to measure the development of the MSI phenotype were repeated, with the expectation that growth in vivo had challenged these cells in some way. It was hypothesized that growth in vivo would lead to the development of MSI and the subsequent development of point mutations in these cells, and that this would explain the dramatic increase in tumorigenicity observed when compared to controls.

An initial experiment aimed to test whether the clonal cell lines collected from the tumors maintained the expected knockdown of either MLH1 or MSH2. This was done with western blotting analysis of whole cell lysates collected from the cells, once cell lines had been established in vitro. Knockdown of MSH2 was maintained in the cell lines to different degrees, the variation presumably owing to the heterogeneity of solid tumors. The variation in MSH2 levels seen in the clones underscored the fact that all cells in a given tumor are not equal, potentially due to factors involving the tumor microenvironment and varied

57 oxygen and nutrient concentrations. It has been well-established that cancer cells in vivo are exposed to hypoxia and hypoglycaemia due to altered blood flow (Raghunand et al.,

2003), and depending on the proximity of the cells to blood vessels, they may experience this phenomenon to varying degrees. Hypoxia has been shown to repress expression of both MLH1 and MSH2 in a c-Myc dependent manner, and independently of HIF1α expression (Bindra and Glazer, 2007).

However, knockdown of MLH1 was not maintained in the surviving clones examined, following growth in vivo. There are several potential explanations for this phenomenon. One hypothesis is that knockdown of MLH1 in the cells used for the xenograft allowed for initial development of the tumor, but perhaps loss of MLH1 slowed cellular growth in vivo, allowing the subset of the population with higher MLH1 levels to overtake the tumor. In a study examining patients with sporadic endometrial adenocarcinoma, loss of MLH1 was shown to be significantly correlated with less aggressive clinopathological features (Ju et al., 2006). This may indicate that cells with reduced MLH1 levels are not as aggressive and do not grow as quickly as those with higher

MLH1 levels. This would allow the MLH1-low subpopulation to be overtaken quickly in vivo.

Knockdown of MLH1 in SW480 cells results in decreased levels of cyclin E, a key protein involved in cell cycle progression (Chung et al., 2010). Cyclin E binds to the cyclin-dependent kinase CDK2 (Koff et al., 1992). This complex then aims to promote cell cycle progression from G1 to S phase by phosphorylating p27Kip1, tagging it for degradation and promoting expression of Cyclin A (Toyoshima and Hunter, 1994).

Expression of Cyclin A allows the cell cycle to progress (Pagano et al., 1992). A similar

58 phenomenon may happen in our MLH1 knockdown cells, allowing cells with higher

MLH1 levels to dominate the tumor or to dominate growth in vitro, as they are able to grow faster and overtake the cells with lower MLH1 levels. Evaluating the population doubling time or cyclin E levels of the MLH1 knockdown cells we produced may provide further support for this hypothesis.

Given the known role of mismatch repair proteins in preventing the development of microsatellite instability in human cancers (Soreide et al., 2006), we investigated the potential effect that growth in vivo may have had in the development of MSI in our knockdown cells. Growth in a tumor, rather than in a monolayer in vitro, would presumably have led to some of the cells in the tumor being exposed to areas of ischemia.

Ischemia has been shown to reduce levels of MSH2 and induce mutations in the KRAS oncogene in vitro, and this effect becomes even more pronounced when cells are grown in xenografts in vivo (Shahrzad et al., 2005). Other studies have shown that hypoxia results in reduction of MLH1 levels, and increases point mutation rate and the development of microsatellite instability in stem cells (Rodriguez-Jimenez et al., 2008). This led to the hypothesis that growth in vivo would further reduce levels of both MLH1 and MSH2 in the xenografted cells, allowing for the development of both point mutations and changes in the length of repetitive microsatellite regions of the genome, owing to the vital role these proteins play in maintaining genomic stability. Depending on the point mutations developed, this would then provide a growth advantage to mutated cells, allowing for them to dominate the tumor, explaining the increased tumorigenicity seen.

However, when BAT-25 and BAT-26 regions were examined using tRFLP analysis, no obvious shifts in repeat lengths were observed. Similar studies have shown that

59 reduction of MMR proteins is not always sufficient to induce changes in the length of the

BAT-26 region – it appears that the proteins must be lost entirely in order for this to occur

(Shin and Park, 2000). Given that we only reduced MLH1 or MSH2 levels, rather than knocking them out entirely, it appears that the residual levels of these proteins were able to prevent changes in these regions.

As well, no detectable point mutations developed in key oncogenes following growth in vivo as determined by a broad-spectrum genomic screen, examining 19 different oncogenes for over 260 different common point mutations. This screen included a variety of different mutations in both BRAF and KRAS, as well as other oncogenes commonly implicated in other forms of cancer. As mentioned previously, BRAF and KRAS mutations are commonly implicated in colorectal cancer pathogenesis (Fearon and Vogelstein, 1990), and we expected to observe mutations in one of these genes. However, it appears that the small amount of MLH1 and MSH2 remaining in the cells was sufficient to prevent these mutations from developing.

These observations make it difficult to explain the dramatic increase in tumor growth seen in the knockdown cells. One potential explanation for this is that there was enough residual MLH1 or MSH2 remaining in the cells to keep the genome stable, even when challenged with ischemia and interactions with the host environment in vivo.

Alternatively, mutations in other regions of the genome which were not examined may have occurred. This may have included mutations in promoter regions of oncogenes leading to their overexpression, or in promoter regions of tumor suppressors leading to their silencing. One example of such a mutation is found to occur in the G-protein HRAS.

The HRAS promoter contains two G-rich regions, which form quadruplexes in vivo.

Mutations in these regions which prevented quadruplex formation resulted in a 5-fold increase of transcription of this gene (Membrino et al., 2011). Such changes would not have any effect on the coding region of the genes themselves, and as such would not have been detected with the techniques used.

It is also worth noting that the different MutS and MutL complexes that are formed in vivo are better at repairing certain types of mutations. The relative repair frequencies are influenced by a variety of factors, which include motif size, length, and sequence composition, as well as any secondary structures the DNA may form which may cause

DNA polymerase stalling during replication (Eckert and Hile, 2009). Some of the secondary binding partners of the major MMR proteins are better at repairing certain types of mutations, for example, loss of MSH6 (a binding partner of MSH2) causes an increase in mononucleotide repeat mutations, but not dinucleotide repeat mutations (Papadopoulos et al., 1995). However, given the importance of both MLH1 and MSH2 in recognizing all kinds of MMR-repairable mutations, it is not likely that these factors can explain why no point mutations were observed.

Following these observations, we then aimed to explore the potential effects of

MSH2 reduction on levels of other members of the DNA mismatch repair family. The two major binding partners of MSH2 are MSH6 and MSH3, forming the MutSα and MutSβ complexes, respectively. Initially, we hypothesized that loss of MSH2 may have resulted in upregulation of these other proteins, allowing for genomic stability to be maintained and preventing the development of point mutations and microsatellite instability when grown in vivo. However, the opposite was found to be true. MSH2 levels in the various knockdown

61 clones isolated from the tumors correlated positively with their levels of both MSH6 and

MSH3.

There are several potential explanations for this phenomenon. Firstly, it could be due to off-target effects of the siRNA inadvertently silencing MSH3 and MSH6, along with the targeted silencing of MSH2. However, the more likely explanation is that the dimerization of MSH2 with either MSH3 or MSH6 contributes to the stability of these proteins and prevents their degradation. Immunohistochemical studies have shown that in tumors where MSH2 levels are reduced or lost entirely, MSH6 levels are also drastically reduced (Stormorken et al., 2005). This matches other findings, for in a tissue-specific inducible MSH2 knockout mouse model, when MSH2 is knocked out, MSH6 levels are dramatically reduced (Kucherlapati et al., 2010). Similar results have also been seen correlating MSH2 and MSH3 levels (Plaschke et al., 2012). The known stoichiometry of the MutS family of MMR proteins also supports this hypothesis, as the combined levels of

MSH3 and MSH6 in the HCT116 colorectal cancer cell line have been shown to equal the levels of MSH2 (Chang et al., 2000), indicating that there are likely no other binding partners for these proteins, and that only those proteins which are in a complex are maintained within the cell (Hayes et al., 2009). This is supported by our findings, showing that in an MSH2-dose-dependent manner, the two major binding partners of this protein,

MSH3 and MSH6, are also reduced. Given that loss of just MSH3 or MSH6 can produce a mutator phenotype in mice (Hegan et al., 2006), patients with reduced MSH2 may show an even more dramatic phenotype, as these other MMR proteins are lost as well.

Previous work has shown that there is a dramatic difference between complete knockout of mismatch repair proteins, when compared to severely reduced levels like those

62 seen in our model. Mouse embryonic stem cells in which MSH2 levels have been reduced to approximately 10% of wild-type show an interesting phenotype which may help explain the disparity of our results. These MSH2-low cells more closely resemble our own knockdown cells, and provide a better basis for comparison than cells in which MSH2 has been completely knocked out.

When MSH2 levels are drastically reduced, but not lost, mouse embryonic stem cells maintain stability at microsatellite (CA)n regions (Claij and Te Riele, 2002), with a mutation rate of 2%, compared to 32% in complete MSH2 knockout cells. The MSH2-low cells showed mutation frequencies very similar to those in wildtype cells, indicating that low amounts of MSH2 are sufficient to maintain stability at some microsatellite regions, and prevent the development of point mutations (Claij and Te Riele, 2002). This phenomenon may explain why we were unable to detect any point mutations or signs of microsatellite instability in our MSH2 knockdown cells, regardless of their exposure to in vivo conditions.

However, the one phenotype in which the murine MSH2-low cells did show a dramatic significant difference when compared to wildtype cells was their susceptibility to various DNA-damaging chemicals, such as the DNA alkylating agent MNNG (N-methyl-

N’-nitro-N-nitrosoguanidine) (Claij and Te Riele, 2002). This chemical acts to add bulky

O6-meG (O6-methylguanine) lesions to DNA, which poses a problem during DNA replication, and is most commonly mispaired with thymine. This then activates the MMR system to excise the improperly incorporated residue (Berardini et al., 2000). This repair cycle is futile, however, and may result in either the development of double strand breaks in the DNA (Karran and Bignami, 1992), or direct signalling of the activated MMR

63 machinery (Fishel, 1999), resulting in apoptosis. In this case, the murine MSH2-low cells behaved similarly to MSH2-knockout cells, indicating that more abundant MMR machinery is necessary to combat the increased mutation rates caused by this chemotherapeutic drug.

Similar to the results shown by Claij and Te Riele, we showed that cells with dramatically reduced levels of the DNA mismatch repair enzymes MLH1 or MSH2 are resistant to the chemotherapeutic drug 5-fluorouracil. The use of a clonogenic survival assay, rather than a more indirect measure of cell growth or survival, increases the confidence in this result, as this assay allows us to distinguish between senescent cells as well as cells which change their growth rate in response to the drug. 5-Fluorouracil is a pyrimidine analogue that acts to inhibit cell growth and induce apoptosis by misincorporating into DNA and RNA, and by inhibiting the enzyme thymidylate synthase which converts TTP to dUTP in the cell (Danenberg et al., 1974). Other cell lines with complete loss of different mismatch repair proteins have also been shown to be resistant to this drug, such as HCT116 (no MLH1) and LoVo (no MSH2) (Adamsen et al., 2011;

Plasencia et al., 2003). However, until recently it was not known whether this effect was due to 5FU’s role in preventing proper RNA synthesis, DNA synthesis, or its thymidylate synthase inhibitory effects. Recent work has demonstrated that one of this drug’s effects that MMR-deficient cells are able to evade is its misincorporation into DNA. Researchers measured cell growth, clonogenic survival, and apoptotic phenotype when cell lines deficient in MLH1, MSH2, or MSH6 were transfected with plasmids containing 5FU induced lesions. MMR-deficient cell lines displayed signs of resistance to 5FU in all of the assays used when challenged with error-containing DNA (Iwaizumi et al., 2011). Our work

64 demonstrates that this resistance to 5FU can occur in cells in which MLH1 or MSH2 levels have been reduced, rather than completely knocked out, implicating the dose-dependent aspect of MMR protein levels in determining phenotype and response to chemotherapy.

Implications

This study demonstrates the importance of determining the expression levels of various MMR proteins when determining what type of chemotherapeutic approach will be the most successful for a patient. We have shown that cells with reduced levels of the mismatch repair proteins MLH1 or MSH2 will not test positive for the development of microsatellite instability with traditional methods – measuring of the BAT-25 and BAT-26 repeat regions. In a clinical setting, this would often be sufficient evidence for a cancer to be diagnosed as MSS. We have shown that dramatic reduction of MMR proteins is not accompanied by the development of an MSI phenotype, even when tested following the more complex environment of in vivo growth.

Although these cells did not test positive for MSI, they still portrayed one very important phenotype of cells which are MSI positive – they were shown to be resistant to a

DNA-damaging chemotherapeutic drug, 5-fluorouracil. Had these cells been clinically treated with 5-fluorouracil, the tumor may not have responded as well as a tumor with sufficient DNA MMR systems, based on our clonogenic survival assays showing resistance to chronic treatment with 5 μM 5-fluorouracil. Our resistance studies were done in vitro without the compounding factors of ischemia, interaction with the host, and three- dimensional growth which may change the results we would see if these studies were repeated in vivo, but the use of a clonogenic survival assay provided a more accurate representation of the dynamics of cellular senescence, apoptosis, and resistance in vitro.

Therefore, it is a reasonable conclusion that cells which do not show any other signs of

MSI, may still have reduced levels of MMR proteins and therefore be resistant to certain

DNA-damaging drugs.

These observations demonstrated the importance of the “dose” of mismatch repair capabilities that given tumor cells may have in determining their phenotype. There may be sufficient amount of repair capabilities to prevent the development of microsatellite instability or point mutations in the genes examined with the OncoCarta panel, but there may not be enough to allow for DNA damaging drugs, such as 5FU, to exert their apoptosis-inducing effects as expected. Should these in vitro studies recapitulate what happens in vivo, this may partially explain why 5FU efficacies are so low clinically – perhaps some of the cells being treated have reduced levels of MMR proteins, which results in their phenotype resembling what we were able to produce in vitro.

As well, these observations demonstrated that the loss of one MMR protein in a dimer may have effects on its various binding partners. This has implications for patients in which one MMR protein has been lost, as others may also follow, compounding the effects and producing cells in which the genome is even more prone to developing mutations. The secondary loss of other proteins also makes it more difficult to determine exactly which protein was lost first in a cell, and which lost as a consequence of the loss of their binding partners.

Limitations

One of the limitations of this study is the use of a mouse xenograft model to examine the growth of the knockdown cells in vivo. The mice used for this study were the

SHO strain, which is an immunocompromised hairless mouse lacking both T cells and B

66 cells (Sharpless and Depinho, 2006). The use of an animal model lacking a properly functioning immune system is necessary to allow the growth of human cells in a different species. However, the lack of an immune system in these mice may have influenced the growth characteristics seen. As well, when using a xenograft model, millions of cells are injected into the animal at once, whereas when a tumor develops in a patient, it typically starts from only a few transformed cells. This allows for more clonal selection to occur in a patient, a facet of tumor development that is lost when using a xenograft.

Another limitation of this study is the inability of standard tissue culture techniques to fully recapitulate what occurs when a tumor is growing in vivo. When grown in vitro for these experiments, cells were grown in monolayer, rather than in three-dimensions. This prevents regions of ischemia from developing in vitro, though they would be expected to develop in the tumor in vivo. As well, oxygen concentrations in vitro are much higher than standard oxygen concentrations that would be seen in tissue in vivo. Given the role of hypoxia in downregulating levels of mismatch repair proteins and inducing genomic instability (Rodriguez-Jimenez et al., 2008; Shahrzad et al., 2005), this may have also influenced our results.

The use of Caco-2 cells as the cell line in which this model was developed also presents some potential problems. While Caco-2 cells are used by many as a model of normal colonic epithelium, they may have a variety of unknown mutations as they are a cancer cell line. One known mutation in Caco-2 is a mutation of the tumor suppressor protein p53 (Djelloul et al., 1997), though many others may exist, perhaps in tumor- suppressor genes or other genes not examined by the Sequenom array, which may complicate and influence the results seen.

Future Directions

Future work on this project could explore the effects that knockdown of MLH1 or

MSH2 may have had on the doubling time of these cells, as that may explain why following growth in vivo the MLH1 knockdown seems to have been lost. This may be due to the apparent role of MLH1 in regulating cyclin E expression (Chung et al., 2010), but this effect should be confirmed in our cell lines.

Another interesting avenue would be to examine the effects that reduced levels of

MLH1 and MSH2 have on cellular responses to other chemicals or chemotherapeutics, particularly those in which MSI cells have been shown to be resistant, such as MNNG, alkylating agents, and platinum-containing compounds (Aebi et al., 1997; Carethers et al.,

1996; Fink et al., 1996; Koi et al., 1994). We showed that reduction of DNA MMR proteins results in resistance to the chemotherapeutic 5-fluorouracil, and it would be interesting to see if similar results are seen with other drugs.

An important facet of drug effectiveness studies is the role that the in vivo environment plays. Two such concerns are drug metabolism in a whole organism, as well as the effect the drug may have on other non-cancerous components of the tumor, such as vascular endothelial cells. This was not directly examined in our studies, and showing that our MMR-knockdown cells are resistant to 5-fluorouracil both in vitro and in vivo would strengthen our findings tremendously.

SUMMARY AND CONCLUSIONS

In this study, we aimed to explore the effects that growth in vivo may have had on the development of microsatellite instability in a derivative of the Caco-2 human colorectal cancer cell line. The cell lines used for this study had reduced levels of one of two important mismatch repair proteins, MLH1 and MSH2. Prior to growth in vivo, these knockdown cells did not show any signs of MSI – BAT-25 and BAT-26 repeat lengths were unchanged, and no point mutations were detected in BRAF or KRAS oncogenes.

These results did not change following growth in vivo, indicating that the residual levels of

MLH1 and MSH2 in the cells were probably sufficient to maintain genomic stability in the regions examined. In fact, a broader panel of oncogenes was examined for the development of point mutations, and all loci examined maintained the wildtype sequence. Following this, we explored the potential effects that loss of MSH2 may have had on other MMR proteins, namely MSH3 and MSH6. We demonstrated that loss of MSH2 results in proportionate loss of its major binding partners, and did not cause upregulation of other MMR proteins, such as PMS2. Given that another phenotype of cancers which are MSI is their resistance to certain chemotherapeutics, we aimed to explore if reductions in MLH1 and MSH2 contribute to this resistance. Cells with knockdown of MLH1 or MSH2 showed resistance to the common chemotherapy drug 5-fluorouracil in a clonogenic survival assay following chronic treatment at 5 μM. The drug 5-fluorouracil was chosen as it exerts its cytotoxic effects by both directly and indirectly inducing DNA damage.

Overall, these results indicate that there is a dose-dependent effect of MLH1 and

MSH2 in the cell. Extremely low levels of these proteins appear to be sufficient to maintain genomic stability, as no point mutations or microsatellite instability were observed.

However, this knockdown did appear to have an effect on the susceptibility of these cells to the chemotherapeutic 5-fluorouracil, demonstrating the importance of wildtype levels of these proteins in allowing this drug to induce apoptosis. The changes in these cells that lead to the dramatic differences we observed in the xenograft model remain to be discovered.

REFERENCES

Aaltonen, L.A., Peltomaki, P., Leach, F.S., Sistonen, P., Pylkkanen, L., Mecklin, J.P., Jarvinen, H., Powell, S.M., Jen, J., Hamilton, S.R., et al. (1993). Clues to the pathogenesis of familial colorectal cancer. Science 260, 812-816.

Aberle, H., Bauer, A., Stappert, J., Kispert, A., and Kemler, R. (1997). beta-catenin is a target for the ubiquitin-proteasome pathway. EMBO J 16, 3797-3804.

Adamsen, B.L., Kravik, K.L., and De Angelis, P.M. (2011). DNA damage signaling in response to 5-fluorouracil in three colorectal cancer cell lines with different mismatch repair and TP53 status. Int J Oncol 39, 673-682.

Aebi, S., Fink, D., Gordon, R., Kim, H.K., Zheng, H., Fink, J.L., and Howell, S.B. (1997). Resistance to cytotoxic drugs in DNA mismatch repair-deficient cells. Clin Cancer Res 3, 1763- 1767.

Al-Sukhni, W., Aronson, M., and Gallinger, S. (2008). Hereditary colorectal cancer syndromes: familial adenomatous polyposis and Lynch syndrome. Surg Clin North Am 88, 819-844, vii.

Al-Tassan, N., Chmiel, N.H., Maynard, J., Fleming, N., Livingston, A.L., Williams, G.T., Hodges, A.K., Davies, D.R., David, S.S., Sampson, J.R., et al. (2002). Inherited variants of MYH associated with somatic G:C→T:A mutations in colorectal tumors. Nat Genet 30, 227-232.

Alani, E. (1996). The Saccharomyces cerevisiae Msh2 and Msh6 proteins form a complex that specifically binds to duplex oligonucleotides containing mismatched DNA base pairs. Mol Cell Biol 16, 5604-5615.

Aquilina, G., Hess, P., Branch, P., MacGeoch, C., Casciano, I., Karran, P., and Bignami, M. (1994). A mismatch recognition defect in colon carcinoma confers DNA microsatellite instability and a mutator phenotype. Proc Natl Acad Sci U S A 91, 8905-8909.

Aramayo, R., Sherman, M.B., Brownless, K., Lurz, R., Okorokov, A.L., and Orlova, E.V. (2011). Quaternary structure of the specific p53-DNA complex reveals the mechanism of p53 mutant dominance. Nucleic Acids Res 39, 8960-8971.

Baker, S.M., Bronner, C.E., Zhang, L., Plug, A.W., Robatzek, M., Warren, G., Elliott, E.A., Yu, J., Ashley, T., Arnheim, N., et al. (1995). Male mice defective in the DNA mismatch repair gene PMS2 exhibit abnormal chromosome synapsis in meiosis. Cell 82, 309-319.

Baker, S.M., Plug, A.W., Prolla, T.A., Bronner, C.E., Harris, A.C., Yao, X., Christie, D.M., Monell, C., Arnheim, N., Bradley, A., et al. (1996). Involvement of mouse Mlh1 in DNA mismatch repair and meiotic crossing over. Nat Genet 13, 336-342.

Bakhoum, S.F., Genovese, G., and Compton, D.A. (2009). Deviant kinetochore microtubule dynamics underlie chromosomal instability. Curr Biol 19, 1937-1942.

Bauer, C.M., Ray, A.M., Halstead-Nussloch, B.A., Dekker, R.G., Raymond, V.M., Gruber, S.B., and Cooney, K.A. (2011). Hereditary prostate cancer as a feature of Lynch syndrome. Fam Cancer 10, 37-42.

Berardini, M., Mazurek, A., and Fishel, R. (2000). The effect of O6-methylguanine DNA adducts on the adenosine nucleotide switch functions of hMSH2-hMSH6 and hMSH2-hMSH3. J Biol Chem 275, 27851-27857.

Bindra, R.S., and Glazer, P.M. (2007). Co-repression of mismatch repair gene expression by hypoxia in cancer cells: role of the Myc/Max network. Cancer Lett 252, 93-103.

Boland, C.R., and Goel, A. (2010). Microsatellite instability in colorectal cancer. Gastroenterology 138, 2073-2087 e2073.

Boland, C.R., Thibodeau, S.N., Hamilton, S.R., Sidransky, D., Eshleman, J.R., Burt, R.W., Meltzer, S.J., Rodriguez-Bigas, M.A., Fodde, R., Ranzani, G.N., et al. (1998). 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.

Bork, P., Hofmann, K., Bucher, P., Neuwald, A.F., Altschul, S.F., and Koonin, E.V. (1997). A superfamily of conserved domains in DNA damage-responsive cell cycle checkpoint proteins. FASEB J 11, 68-76.

Bos, J.L., Fearon, E.R., Hamilton, S.R., Verlaan-de Vries, M., van Boom, J.H., van der Eb, A.J., and Vogelstein, B. (1987). Prevalence of ras gene mutations in human colorectal cancers. Nature 327, 293-297.

Brabletz, T., Jung, A., Dag, S., Hlubek, F., and Kirchner, T. (1999). beta-catenin regulates the expression of the matrix metalloproteinase-7 in human colorectal cancer. Am J Pathol 155, 1033- 1038.

Brennetot, C., Buhard, O., Jourdan, F., Flejou, J.F., Duval, A., and Hamelin, R. (2005). Mononucleotide repeats BAT-26 and BAT-25 accurately detect MSI-H tumors and predict tumor content: implications for population screening. Int J Cancer 113, 446-450.

Cahill, D.P., Lengauer, C., Yu, J., Riggins, G.J., Willson, J.K., Markowitz, S.D., Kinzler, K.W., and Vogelstein, B. (1998). Mutations of mitotic checkpoint genes in human cancers. Nature 392, 300- 303.

Carethers, J.M., Hawn, M.T., Chauhan, D.P., Luce, M.C., Marra, G., Koi, M., and Boland, C.R. (1996). Competency in mismatch repair prohibits clonal expansion of cancer cells treated with N- methyl-N'-nitro-N-nitrosoguanidine. J Clin Invest 98, 199-206.

Cariello, N.F., Cui, L., Beroud, C., and Soussi, T. (1994). Database and software for the analysis of mutations in the human p53 gene. Cancer Res 54, 4454-4460.

Chang, D.K., Ricciardiello, L., Goel, A., Chang, C.L., and Boland, C.R. (2000). Steady-state regulation of the human DNA mismatch repair system. J Biol Chem 275, 18424-18431.

Charames, G.S., and Bapat, B. (2003). Genomic instability and cancer. Curr Mol Med 3, 589-596.

Cheng, Y.W., Pincas, H., Bacolod, M.D., Schemmann, G., Giardina, S.F., Huang, J., Barral, S., Idrees, K., Khan, S.A., Zeng, Z., et al. (2008). CpG island methylator phenotype associates with low-degree chromosomal abnormalities in colorectal cancer. Clin Cancer Res 14, 6005-6013.

Chung, D.C., and Rustgi, A.K. (2003). The hereditary nonpolyposis colorectal cancer syndrome: genetics and clinical implications. Ann Intern Med 138, 560-570.

Chung, H., Chaudhry, J., Lopez, C.G., and Carethers, J.M. (2010). Cyclin E and histone H3 levels are regulated by 5-fluorouracil in a DNA mismatch repair-dependent manner. Cancer Biol Ther 10, 1147-1156.

Claij, N., and Te Riele, H. (2002). Methylation tolerance in mismatch repair proficient cells with low MSH2 protein level. Oncogene 21, 2873-2879.

Clarke, B.A., and Cooper, K. (2012). Identifying lynch syndrome in patients with endometrial carcinoma: shortcomings of morphologic and clinical schemas. Adv Anat Pathol 19, 231-238.

Clarke, M.L., Burton, R.L., Hill, A.N., Litorja, M., Nahm, M.H., and Hwang, J. (2010). Low-cost, high-throughput, automated counting of bacterial colonies. Cytometry A 77, 790-797.

Coin, F., Oksenych, V., and Egly, J.M. (2007). Distinct roles for the XPB/p52 and XPD/p44 subcomplexes of TFIIH in damaged DNA opening during nucleotide excision repair. Mol Cell 26, 245-256.

Cox, E.C., Degnen, G.E., and Scheppe, M.L. (1972). Mutator gene studies in Escherichia coli: the mutS gene. Genetics 72, 551-567.

Cunningham, J.M., Christensen, E.R., Tester, D.J., Kim, C.Y., Roche, P.C., Burgart, L.J., and Thibodeau, S.N. (1998). Hypermethylation of the hMLH1 promoter in colon cancer with microsatellite instability. Cancer Res 58, 3455-3460. da Silva, F.C., de Oliveira, L.P., Santos, E.M., Nakagawa, W.T., Aguiar Junior, S., Valentin, M.D., Rossi, B.M., and de Oliveira Ferreira, F. (2010). Frequency of extracolonic tumors in Brazilian families with Lynch syndrome: analysis of a hereditary colorectal cancer institutional registry. Fam Cancer 9, 563-570.

Danenberg, P.V., Langenbach, R.J., and Heidelberger, C. (1974). Structures of reversible and irreversible complexes of thymidylate synthetase and fluorinated pyrimidine nucleotides. Biochemistry 13, 926-933.

Daniotti, M., Vallacchi, V., Rivoltini, L., Patuzzo, R., Santinami, M., Arienti, F., Cutolo, G., Pierotti, M.A., Parmiani, G., and Rodolfo, M. (2007). Detection of mutated BRAFV600E variant in circulating DNA of stage III-IV melanoma patients. Int J Cancer 120, 2439-2444.

Davies, H., Bignell, G.R., Cox, C., Stephens, P., Edkins, S., Clegg, S., Teague, J., Woffendin, H., Garnett, M.J., Bottomley, W., et al. (2002). Mutations of the BRAF gene in human cancer. Nature 417, 949-954. de Angelis, P.M., Fjell, B., Kravik, K.L., Haug, T., Tunheim, S.H., Reichelt, W., Beigi, M., Clausen, O.P., Galteland, E., and Stokke, T. (2004). Molecular characterizations of derivatives of HCT116 colorectal cancer cells that are resistant to the chemotherapeutic agent 5-fluorouracil. Int J Oncol 24, 1279-1288. de Boer, J., and Hoeijmakers, J.H. (2000). Nucleotide excision repair and human syndromes. Carcinogenesis 21, 453-460.

Deng, G., Bell, I., Crawley, S., Gum, J., Terdiman, J.P., Allen, B.A., Truta, B., Sleisenger, M.H., and Kim, Y.S. (2004). BRAF mutation is frequently present in sporadic colorectal cancer with methylated hMLH1, but not in hereditary nonpolyposis colorectal cancer. Clin Cancer Res 10, 191- 195.

Dianov, G.L., Prasad, R., Wilson, S.H., and Bohr, V.A. (1999). Role of DNA polymerase beta in the excision step of long patch mammalian base excision repair. J Biol Chem 274, 13741-13743.

Djelloul, S., Forgue-Lafitte, M.E., Hermelin, B., Mareel, M., Bruyneel, E., Baldi, A., Giordano, A., Chastre, E., and Gespach, C. (1997). Enterocyte differentiation is compatible with SV40 large T expression and loss of p53 function in human colonic Caco-2 cells. Status of the pRb1 and pRb2 tumor suppressor gene products. FEBS Lett 406, 234-242.

Doetsch, P.W., and Cunningham, R.P. (1990). The enzymology of apurinic/apyrimidinic endonucleases. Mutat Res 236, 173-201.

Domingo, E., Laiho, P., Ollikainen, M., Pinto, M., Wang, L., French, A.J., Westra, J., Frebourg, T., Espin, E., Armengol, M., et al. (2004). BRAF screening as a low-cost effective strategy for simplifying HNPCC genetic testing. J Med Genet 41, 664-668.

Douillard, J.Y., Cunningham, D., Roth, A.D., Navarro, M., James, R.D., Karasek, P., Jandik, P., Iveson, T., Carmichael, J., Alakl, M., et al. (2000). Irinotecan combined with fluorouracil compared with fluorouracil alone as first-line treatment for metastatic colorectal cancer: a multicentre randomised trial. Lancet 355, 1041-1047.

Easton, D.F., Bishop, D.T., Ford, D., and Crockford, G.P. (1993). Genetic linkage analysis in familial breast and ovarian cancer: results from 214 families. The Breast Cancer Linkage Consortium. Am J Hum Genet 52, 678-701.

Eckert, K.A., and Hile, S.E. (2009). Every microsatellite is different: Intrinsic DNA features dictate mutagenesis of common microsatellites present in the human genome. Mol Carcinog 48, 379-388.

Eckhardt, F., Lewin, J., Cortese, R., Rakyan, V.K., Attwood, J., Burger, M., Burton, J., Cox, T.V., Davies, R., Down, T.A., et al. (2006). DNA methylation profiling of human chromosomes 6, 20 and 22. Nat Genet 38, 1378-1385.

Edelmann, W., Umar, A., Yang, K., Heyer, J., Kucherlapati, M., Lia, M., Kneitz, B., Avdievich, E., Fan, K., Wong, E., et al. (2000). The DNA mismatch repair genes Msh3 and Msh6 cooperate in intestinal tumor suppression. Cancer Res 60, 803-807.

Edelmann, W., Yang, K., Umar, A., Heyer, J., Lau, K., Fan, K., Liedtke, W., Cohen, P.E., Kane, M.F., Lipford, J.R., et al. (1997). Mutation in the mismatch repair gene Msh6 causes cancer susceptibility. Cell 91, 467-477.

Eilers, U., Klumperman, J., and Hauri, H.P. (1989). Nocodazole, a microtubule-active drug, interferes with apical protein delivery in cultured intestinal epithelial cells (Caco-2). J Cell Biol 108, 13-22.

Fearon, E.R., and Vogelstein, B. (1990). A genetic model for colorectal tumorigenesis. Cell 61, 759-767.

Feinberg, A.P., Gehrke, C.W., Kuo, K.C., and Ehrlich, M. (1988). Reduced genomic 5- methylcytosine content in human colonic neoplasia. Cancer Res 48, 1159-1161.

Fink, D., Nebel, S., Aebi, S., Zheng, H., Cenni, B., Nehme, A., Christen, R.D., and Howell, S.B. (1996). The role of DNA mismatch repair in platinum drug resistance. Cancer Res 56, 4881-4886.

Fischer, S.M., O'Connell, J.F., Conti, C.J., Tacker, K.C., Fries, J.W., Patrick, K.E., Adams, L.M., and Slaga, T.J. (1987). Characterization of an inbred strain of the SENCAR mouse that is highly sensitive to phorbol esters. Carcinogenesis 8, 421-424.

Fishel, R. (1999). Signaling mismatch repair in cancer. Nat Med 5, 1239-1241.

Foulds, L. (1958). The natural history of cancer. J Chronic Dis 8, 2-37.

Friedberg, E.C., Walker, G.C., Siede, W., Wood, R.D., Schultz, R.A., and Ellenberger, T. (2006). DNA Repair and Mutagenesis (Washington DC, ASM Press).

Gailani, S., Holland, J.F., Falkson, G., Leone, L., Burningham, R., and Larsen, V. (1972). Comparison of treatment of metastatic gastrointestinal cancer with 5-fluorouracil (5-FU) to a combination of 5-FU with cytosine arabinoside. Cancer 29, 1308-1313.

Galiatsatos, P., and Foulkes, W.D. (2006). Familial adenomatous polyposis. Am J Gastroenterol 101, 385-398.

Geiersbach, K.B., and Samowitz, W.S. (2011). Microsatellite instability and colorectal cancer. Arch Pathol Lab Med 135, 1269-1277.

Gertler, R., Rosenberg, R., Stricker, D., Friederichs, J., Hoos, A., Werner, M., Ulm, K., Holzmann, B., Nekarda, H., and Siewert, J.R. (2004). Telomere length and human telomerase reverse transcriptase expression as markers for progression and prognosis of colorectal carcinoma. J Clin Oncol 22, 1807-1814.

Goulian, M., Bleile, B.M., Dickey, L.M., Grafstrom, R.H., Ingraham, H.A., Neynaber, S.A., Peterson, M.S., and Tseng, B.Y. (1986). Mechanism of thymineless death. Adv Exp Med Biol 195 Pt B, 89-95.

Gowen, L.C., Avrutskaya, A.V., Latour, A.M., Koller, B.H., and Leadon, S.A. (1998). BRCA1 required for transcription-coupled repair of oxidative DNA damage. Science 281, 1009-1012.

Gryfe, R. (2006). Clinical implications of our advancing knowledge of colorectal cancer genetics: inherited syndromes, prognosis, prevention, screening and therapeutics. Surg Clin North Am 86, 787-817.

Habraken, Y., Sung, P., Prakash, L., and Prakash, S. (1996). Binding of insertion/deletion DNA mismatches by the heterodimer of yeast mismatch repair proteins MSH2 and MSH3. Curr Biol 6, 1185-1187.

Haigis, K.M., Kendall, K.R., Wang, Y., Cheung, A., Haigis, M.C., Glickman, J.N., Niwa-Kawakita, M., Sweet-Cordero, A., Sebolt-Leopold, J., Shannon, K.M., et al. (2008). Differential effects of oncogenic K-Ras and N-Ras on proliferation, differentiation and tumor progression in the colon. Nat Genet 40, 600-608.

Hanahan, D., and Weinberg, R.A. (2000). The hallmarks of cancer. Cell 100, 57-70.

Hanahan, D., and Weinberg, R.A. (2011). Hallmarks of cancer: the next generation. Cell 144, 646- 674.

Hayes, A.P., Sevi, L.A., Feldt, M.C., Rose, M.D., and Gammie, A.E. (2009). Reciprocal regulation of nuclear import of the yeast MutSalpha DNA mismatch repair proteins Msh2 and Msh6. DNA Repair (Amst) 8, 739-751.

Hegan, D.C., Narayanan, L., Jirik, F.R., Edelmann, W., Liskay, R.M., and Glazer, P.M. (2006). Differing patterns of genetic instability in mice deficient in the mismatch repair genes Pms2, Mlh1, Msh2, Msh3 and Msh6. Carcinogenesis 27, 2402-2408.

Herman, J.G., Umar, A., Polyak, K., Graff, J.R., Ahuja, N., Issa, J.P., Markowitz, S., Willson, J.K., Hamilton, S.R., Kinzler, K.W., et al. (1998). Incidence and functional consequences of hMLH1 promoter hypermethylation in colorectal carcinoma. Proc Natl Acad Sci U S A 95, 6870-6875.

Herskowitz, I. (1987). Functional inactivation of genes by dominant negative mutations. Nature 329, 219-222.

Heyer, W.D. (1994). The search for the right partner: homologous pairing and DNA strand exchange proteins in eukaryotes. Experientia 50, 223-233.

Hollingsworth, N.M., Ponte, L., and Halsey, C. (1995). MSH5, a novel MutS homolog, facilitates meiotic reciprocal recombination between homologs in Saccharomyces cerevisiae but not mismatch repair. Genes Dev 9, 1728-1739.

Hubscher, U., and Maga, G. (2011). DNA replication and repair bypass machines. Curr Opin Chem Biol 15, 627-635.

Iaccarino, I., Palombo, F., Drummond, J., Totty, N.F., Hsuan, J.J., Modrich, P., and Jiricny, J. (1996). MSH6, a Saccharomyces cerevisiae protein that binds to mismatches as a heterodimer with MSH2. Curr Biol 6, 484-486.

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

Iwabuchi, K., Li, B., Bartel, P., and Fields, S. (1993). Use of the two-hybrid system to identify the domain of p53 involved in oligomerization. Oncogene 8, 1693-1696.

Iwaizumi, M., Tseng-Rogenski, S., and Carethers, J.M. (2011). DNA mismatch repair proficiency executing 5-fluorouracil cytotoxicity in colorectal cancer cells. Cancer Biol Ther 12, 756-764.

Jacob, S., Aguado, M., Fallik, D., and Praz, F. (2001). The role of the DNA mismatch repair system in the cytotoxicity of the topoisomerase inhibitors camptothecin and etoposide to human colorectal cancer cells. Cancer Res 61, 6555-6562.

Jass, J.R. (2006). Hereditary Non-Polyposis Colorectal Cancer: the rise and fall of a confusing term. World J Gastroenterol 12, 4943-4950.

Jensen, D.E., Proctor, M., Marquis, S.T., Gardner, H.P., Ha, S.I., Chodosh, L.A., Ishov, A.M., Tommerup, N., Vissing, H., Sekido, Y., et al. (1998). BAP1: a novel ubiquitin hydrolase which binds to the BRCA1 RING finger and enhances BRCA1-mediated cell growth suppression. Oncogene 16, 1097-1112.

Ju, W., Park, H.M., Lee, S.N., Sung, S.H., and Kim, S.C. (2006). Loss of hMLH1 expression is associated with less aggressive clinicopathological features in sporadic endometrioid endometrial adenocarcinoma. J Obstet Gynaecol Res 32, 454-460.

Karran, P., and Bignami, M. (1992). Self-destruction and tolerance in resistance of mammalian cells to alkylation damage. Nucleic Acids Res 20, 2933-2940.

Kim, S.H., Ahn, B.K., Nam, Y.S., Pyo, J.Y., Oh, Y.H., and Lee, K.H. (2010). Microsatellite instability is associated with the clinicopathologic features of gastric cancer in sporadic gastric cancer patients. J Gastric Cancer 10, 149-154.

Kinner, A., Wu, W., Staudt, C., and Iliakis, G. (2008). Gamma-H2AX in recognition and signaling of DNA double-strand breaks in the context of chromatin. Nucleic Acids Res 36, 5678-5694.

Klein, S., Zenvirth, D., Sherman, A., Ried, K., Rappold, G., and Simchen, G. (1996). Double-strand breaks on YACs during yeast meiosis may reflect meiotic recombination in the human genome. Nat Genet 13, 481-484.

Klungland, A., and Lindahl, T. (1997). Second pathway for completion of human DNA base excision-repair: reconstitution with purified proteins and requirement for DNase IV (FEN1). EMBO J 16, 3341-3348.

Kneitz, B., Cohen, P.E., Avdievich, E., Zhu, L., Kane, M.F., Hou, H., Jr., Kolodner, R.D., Kucherlapati, R., Pollard, J.W., and Edelmann, W. (2000). MutS homolog 4 localization to meiotic chromosomes is required for chromosome pairing during meiosis in male and female mice. Genes Dev 14, 1085-1097.

Koff, A., Giordano, A., Desai, D., Yamashita, K., Harper, J.W., Elledge, S., Nishimoto, T., Morgan, D.O., Franza, B.R., and Roberts, J.M. (1992). Formation and activation of a cyclin E-cdk2 complex during the G1 phase of the human cell cycle. Science 257, 1689-1694.

Koi, M., Umar, A., Chauhan, D.P., Cherian, S.P., Carethers, J.M., Kunkel, T.A., and Boland, C.R. (1994). Human chromosome 3 corrects mismatch repair deficiency and microsatellite instability and reduces N-methyl-N'-nitro-N-nitrosoguanidine tolerance in colon tumor cells with homozygous hMLH1 mutation. Cancer Res 54, 4308-4312.

Koinuma, K., Shitoh, K., Miyakura, Y., Furukawa, T., Yamashita, Y., Ota, J., Ohki, R., Choi, Y.L., Wada, T., Konishi, F., et al. (2004). Mutations of BRAF are associated with extensive hMLH1 promoter methylation in sporadic colorectal carcinomas. Int J Cancer 108, 237-242.

Korinek, V., Barker, N., Morin, P.J., van Wichen, D., de Weger, R., Kinzler, K.W., Vogelstein, B., and Clevers, H. (1997). Constitutive transcriptional activation by a beta-catenin-Tcf complex in APC-/- colon carcinoma. Science 275, 1784-1787.

Kraemer, K.H., Lee, M.M., and Scotto, J. (1987). Xeroderma pigmentosum. Cutaneous, ocular, and neurologic abnormalities in 830 published cases. Arch Dermatol 123, 241-250.

Kraemer, K.H., Levy, D.D., Parris, C.N., Gozukara, E.M., Moriwaki, S., Adelberg, S., and Seidman, M.M. (1994). Xeroderma pigmentosum and related disorders: examining the linkage between defective DNA repair and cancer. J Invest Dermatol 103, 96S-101S.

Kraemer, K.H., and Slor, H. (1985). Xeroderma pigmentosum. Clin Dermatol 3, 33-69.

Kucherlapati, M.H., Lee, K., Nguyen, A.A., Clark, A.B., Hou, H., Jr., Rosulek, A., Li, H., Yang, K., Fan, K., Lipkin, M., et al. (2010). An Msh2 conditional knockout mouse for studying intestinal cancer and testing anticancer agents. Gastroenterology 138, 993-1002 e1001.

Lacombe, K. (2009). The role of mismatch repair and ischemia in development of microsatellite instability in human colorectal cancer. In Biomedical Sciences (Guelph, University of Guelph).

Ladner, R.D. (2001). The role of dUTPase and uracil-DNA repair in cancer chemotherapy. Curr Protein Pept Sci 2, 361-370.

Laiho, P., Launonen, V., Lahermo, P., Esteller, M., Guo, M., Herman, J.G., Mecklin, J.P., Jarvinen, H., Sistonen, P., Kim, K.M., et al. (2002). Low-level microsatellite instability in most colorectal carcinomas. Cancer Res 62, 1166-1170.

Langle-Rouault, F., Maenhaut-Michel, G., and Radman, M. (1987). GATC sequences, DNA nicks and the MutH function in Escherichia coli mismatch repair. EMBO J 6, 1121-1127.

Laval, J. (1977). Two enzymes are required from strand incision in repair of alkylated DNA. Nature 269, 829-832.

Li, G.M., and Modrich, P. (1995). Restoration of mismatch repair to nuclear extracts of H6 colorectal tumor cells by a heterodimer of human MutL homologs. Proc Natl Acad Sci U S A 92, 1950-1954.

Lin, K.M., Shashidharan, M., Ternent, C.A., Thorson, A.G., Blatchford, G.J., Christensen, M.A., Lanspa, S.J., Lemon, S.J., Watson, P., and Lynch, H.T. (1998). Colorectal and extracolonic cancer variations in MLH1/MSH2 hereditary nonpolyposis colorectal cancer kindreds and the general population. Dis Colon Rectum 41, 428-433.

Lindahl, T. (1974). An N-glycosidase from Escherichia coli that releases free uracil from DNA containing deaminated cytosine residues. Proc Natl Acad Sci U S A 71, 3649-3653.

Lipkin, S.M., Wang, V., Jacoby, R., Banerjee-Basu, S., Baxevanis, A.D., Lynch, H.T., Elliott, R.M., and Collins, F.S. (2000). MLH3: a DNA mismatch repair gene associated with mammalian microsatellite instability. Nat Genet 24, 27-35.

Liu, B., Nicolaides, N.C., Markowitz, S., Willson, J.K., Parsons, R.E., Jen, J., Papadopolous, N., Peltomaki, P., de la Chapelle, A., Hamilton, S.R., et al. (1995). Mismatch repair gene defects in sporadic colorectal cancers with microsatellite instability. Nat Genet 9, 48-55.

Liu, B., Parsons, R., Papadopoulos, N., Nicolaides, N.C., Lynch, H.T., Watson, P., Jass, J.R., Dunlop, M., Wyllie, A., Peltomaki, P., et al. (1996). Analysis of mismatch repair genes in hereditary non-polyposis colorectal cancer patients. Nat Med 2, 169-174.

Liu, W.H., Kaur, M., Wang, G., Zhu, P., Zhang, Y., and Makrigiorgos, G.M. (2004). Inverse PCR- based RFLP scanning identifies low-level mutation signatures in colon cells and tumors. Cancer Res 64, 2544-2551.

Losi, L., Ponz de Leon, M., Jiricny, J., Di Gregorio, C., Benatti, P., Percesepe, A., Fante, R., Roncucci, L., Pedroni, M., and Benhattar, J. (1997). K-ras and p53 mutations in hereditary nonpolyposis colorectal cancers. Int J Cancer 74, 94-96.

Loughrey, M.B., Waring, P.M., Tan, A., Trivett, M., Kovalenko, S., Beshay, V., Young, M.A., McArthur, G., Boussioutas, A., and Dobrovic, A. (2007). Incorporation of somatic BRAF mutation testing into an algorithm for the investigation of hereditary non-polyposis colorectal cancer. Fam Cancer 6, 301-310.

Lynch, H.T., Boland, C.R., Gong, G., Shaw, T.G., Lynch, P.M., Fodde, R., Lynch, J.F., and de la Chapelle, A. (2006). Phenotypic and genotypic heterogeneity in the Lynch syndrome: diagnostic, surveillance and management implications. Eur J Hum Genet 14, 390-402.

Lynch, H.T., Harris, R.E., Organ, C.H., Jr., Guirgis, H.A., Lynch, P.M., Lynch, J.F., and Nelson, E.J. (1977). The surgeon, genetics, and cancer control: the Cancer Family Syndrome. Ann Surg 185, 435-440.

Lynch, H.T., and Krush, A.J. (1971). The cancer family syndrome and cancer control. Surg Gynecol Obstet 132, 247-250.

Maddodi, N., Bhat, K.M., Devi, S., Zhang, S.C., and Setaluri, V. (2010). Oncogenic BRAFV600E induces expression of neuronal differentiation marker MAP2 in melanoma cells by promoter demethylation and down-regulation of transcription repressor HES1. J Biol Chem 285, 242-254.

Maul, J.S., Warner, N.R., Kuwada, S.K., Burt, R.W., and Cannon-Albright, L.A. (2006). Extracolonic cancers associated with hereditary nonpolyposis colorectal cancer in the Utah Population Database. Am J Gastroenterol 101, 1591-1596.

McAllister, K.A., Haugen-Strano, A., Hagevik, S., Brownlee, H.A., Collins, N.K., Futreal, P.A., Bennett, L.M., and Wiseman, R.W. (1997). Characterization of the rat and mouse homologues of the BRCA2 breast cancer susceptibility gene. Cancer Res 57, 3121-3125.

McGivern, A., Wynter, C.V., Whitehall, V.L., Kambara, T., Spring, K.J., Walsh, M.D., Barker, M.A., Arnold, S., Simms, L.A., Leggett, B.A., et al. (2004). Promoter hypermethylation frequency and BRAF mutations distinguish hereditary non-polyposis colon cancer from sporadic MSI-H colon cancer. Fam Cancer 3, 101-107.

Membrino, A., Cogoi, S., Pedersen, E.B., and Xodo, L.E. (2011). G4-DNA formation in the HRAS promoter and rational design of decoy oligonucleotides for cancer therapy. PLoS One 6, e24421.

Meng, W., and Takeichi, M. (2009). Adherens junction: molecular architecture and regulation. Cold Spring Harb Perspect Biol 1, a002899.

Merajver, S.D., Pham, T.M., Caduff, R.F., Chen, M., Poy, E.L., Cooney, K.A., Weber, B.L., Collins, F.S., Johnston, C., and Frank, T.S. (1995). Somatic mutations in the BRCA1 gene in sporadic ovarian tumours. Nat Genet 9, 439-443.

Merg, A., Lynch, H.T., Lynch, J.F., and Howe, J.R. (2005). Hereditary colon cancer--part I. Curr Probl Surg 42, 195-256.

Meyers, M., Wagner, M.W., Hwang, H.S., Kinsella, T.J., and Boothman, D.A. (2001). Role of the hMLH1 DNA mismatch repair protein in fluoropyrimidine-mediated cell death and cell cycle responses. Cancer Res 61, 5193-5201.

Miki, Y., Swensen, J., Shattuck-Eidens, D., Futreal, P.A., Harshman, K., Tavtigian, S., Liu, Q., Cochran, C., Bennett, L.M., Ding, W., et al. (1994). A strong candidate for the breast and ovarian cancer susceptibility gene BRCA1. Science 266, 66-71.

Miyazawa, K., Iwaya, K., Kuroda, M., Harada, M., Serizawa, H., Koyanagi, Y., Sato, Y., Mizokami, Y., Matsuoka, T., and Mukai, K. (2000). Nuclear accumulation of beta-catenin in intestinal-type gastric carcinoma: correlation with early tumor invasion. Virchows Arch 437, 508- 513.

Modrich, P. (1991). Mechanisms and biological effects of mismatch repair. Annu Rev Genet 25, 229-253.

Mu, D., Hsu, D.S., and Sancar, A. (1996). Reaction mechanism of human DNA repair excision nuclease. J Biol Chem 271, 8285-8294.

Mure, K., and Rossman, T.G. (2001). Reduction of spontaneous mutagenesis in mismatch repair- deficient and proficient cells by dietary antioxidants. Mutat Res 480-481, 85-95.

Nieuwenhuis, M.H., and Vasen, H.F. (2007). Correlations between mutation site in APC and phenotype of familial adenomatous polyposis (FAP): a review of the literature. Crit Rev Oncol Hematol 61, 153-161.

Oeda, K., Horiuchi, T., and Sekiguchi, M. (1982). The uvrD gene of E. coli encodes a DNA- dependent ATPase. Nature 298, 98-100.

Oliveira, C., Pinto, M., Duval, A., Brennetot, C., Domingo, E., Espin, E., Armengol, M., Yamamoto, H., Hamelin, R., Seruca, R., et al. (2003). BRAF mutations characterize colon but not gastric cancer with mismatch repair deficiency. Oncogene 22, 9192-9196.

Ozawa, M., Baribault, H., and Kemler, R. (1989). The cytoplasmic domain of the cell adhesion molecule uvomorulin associates with three independent proteins structurally related in different species. EMBO J 8, 1711-1717.

Pagano, M., Pepperkok, R., Verde, F., Ansorge, W., and Draetta, G. (1992). Cyclin A is required at two points in the human cell cycle. EMBO J 11, 961-971.

Papadopoulos, N., Nicolaides, N.C., Liu, B., Parsons, R., Lengauer, C., Palombo, F., D'Arrigo, A., Markowitz, S., Willson, J.K., Kinzler, K.W., et al. (1995). Mutations of GTBP in genetically unstable cells. Science 268, 1915-1917.

Papadopoulos, N., Nicolaides, N.C., Wei, Y.F., Ruben, S.M., Carter, K.C., Rosen, C.A., Haseltine, W.A., Fleischmann, R.D., Fraser, C.M., Adams, M.D., et al. (1994). Mutation of a mutL homolog in hereditary colon cancer. Science 263, 1625-1629.

Park, J.G., Vasen, H.F., Park, K.J., Peltomaki, P., Ponz de Leon, M., Rodriguez-Bigas, M.A., Lubinski, J., Beck, N.E., Bisgaard, M.L., Miyaki, M., et al. (1999). Suspected hereditary nonpolyposis colorectal cancer: International Collaborative Group on Hereditary Non-Polyposis Colorectal Cancer (ICG-HNPCC) criteria and results of genetic diagnosis. Dis Colon Rectum 42, 710-715; discussion 715-716.

Parsons, R., Li, G.M., Longley, M., Modrich, P., Liu, B., Berk, T., Hamilton, S.R., Kinzler, K.W., and Vogelstein, B. (1995). Mismatch repair deficiency in phenotypically normal human cells. Science 268, 738-740.

Parsons, R., Li, G.M., Longley, M.J., Fang, W.H., Papadopoulos, N., Jen, J., de la Chapelle, A., Kinzler, K.W., Vogelstein, B., and Modrich, P. (1993). Hypermutability and mismatch repair deficiency in RER+ tumor cells. Cell 75, 1227-1236.

Peddibhotla, S., Lam, M.H., Gonzalez-Rimbau, M., and Rosen, J.M. (2009). The DNA-damage effector checkpoint kinase 1 is essential for chromosome segregation and cytokinesis. Proc Natl Acad Sci U S A 106, 5159-5164.

Peltomaki, P., Gao, X., and Mecklin, J.P. (2001). Genotype and phenotype in hereditary nonpolyposis colon cancer: a study of families with different vs. shared predisposing mutations. Fam Cancer 1, 9-15.

Pino, M.S., and Chung, D.C. (2010). The chromosomal instability pathway in colon cancer. Gastroenterology 138, 2059-2072.

Plaschke, J., Preussler, M., Ziegler, A., and Schackert, H.K. (2012). Aberrant protein expression and frequent allelic loss of MSH3 in colorectal cancer with low-level microsatellite instability. Int J Colorectal Dis.

Plasencia, C., Rooney, P.H., Taron, M., Martinez-Balibrea, E., McLeod, H.L., and Abad, A. (2003). Chromosomal imbalance maps of human 5FU-resistant colorectal cancer cell lines: implications in the analysis of 5FU-acquired resistance mechanisms. Int J Oncol 22, 945-953.

Popat, S., Hubner, R., and Houlston, R.S. (2005). Systematic review of microsatellite instability and colorectal cancer prognosis. J Clin Oncol 23, 609-618.

Prives, C. (1994). How loops, beta sheets, and alpha helices help us to understand p53. Cell 78, 543-546.

Raghunand, N., Gatenby, R.A., and Gillies, R.J. (2003). Microenvironmental and cellular consequences of altered blood flow in tumours. Br J Radiol 76 Spec No 1, S11-22.

Rajagopalan, H., Bardelli, A., Lengauer, C., Kinzler, K.W., Vogelstein, B., and Velculescu, V.E. (2002). Tumorigenesis: RAF/RAS oncogenes and mismatch-repair status. Nature 418, 934.

Raschle, M., Marra, G., Nystrom-Lahti, M., Schar, P., and Jiricny, J. (1999). Identification of hMutLbeta, a heterodimer of hMLH1 and hPMS1. J Biol Chem 274, 32368-32375.

Reitmair, A.H., Cai, J.C., Bjerknes, M., Redston, M., Cheng, H., Pind, M.T., Hay, K., Mitri, A., Bapat, B.V., Mak, T.W., et al. (1996). MSH2 deficiency contributes to accelerated APC-mediated intestinal tumorigenesis. Cancer Res 56, 2922-2926.

Reitmair, A.H., Schmits, R., Ewel, A., Bapat, B., Redston, M., Mitri, A., Waterhouse, P., Mittrucker, H.W., Wakeham, A., Liu, B., et al. (1995). MSH2 deficient mice are viable and susceptible to lymphoid tumours. Nat Genet 11, 64-70.

Resnick, M.A., and Moore, P.D. (1979). Molecular recombination and the repair of DNA double- strand breaks in CHO cells. Nucleic Acids Res 6, 3145-3160.

Robertson, A., Pattishall, S.R., and Matson, S.W. (2006). The DNA binding activity of MutL is required for methyl-directed mismatch repair in Escherichia coli. J Biol Chem 281, 8399-8408.

Rodriguez-Jimenez, F.J., Moreno-Manzano, V., Lucas-Dominguez, R., and Sanchez-Puelles, J.M. (2008). Hypoxia causes downregulation of mismatch repair system and genomic instability in stem cells. Stem Cells 26, 2052-2062.

Rolfe, R. (1967). On the mechanism of thymineless death in Bacillus subtilis. Proc Natl Acad Sci U S A 57, 114-121.

Rutman, R.J., Cantarow, A., and Paschkis, K.E. (1954). Studies in 2-acetylaminofluorene carcinogenesis. III. The utilization of uracil-2-C14 by preneoplastic rat liver and rat hepatoma. Cancer Res 14, 119-123.

Scully, R., Chen, J., Plug, A., Xiao, Y., Weaver, D., Feunteun, J., Ashley, T., and Livingston, D.M. (1997). Association of BRCA1 with Rad51 in mitotic and meiotic cells. Cell 88, 265-275.

Setlow, R.B., and Carrier, W.L. (1964). The disappearance of thymine dimers from DNA: an error- correcting mechanism. Proc Natl Acad Sci U S A 51, 226-231.

Shahrzad, S., Quayle, L., Stone, C., Plumb, C., Shirasawa, S., Rak, J.W., and Coomber, B.L. (2005). Ischemia-induced K-ras mutations in human colorectal cancer cells: role of microenvironmental regulation of MSH2 expression. Cancer Res 65, 8134-8141.

Sharpless, N.E., and Depinho, R.A. (2006). The mighty mouse: genetically engineered mouse models in cancer drug development. Nat Rev Drug Discov 5, 741-754.

Shibutani, S., Takeshita, M., and Grollman, A.P. (1991). Insertion of specific bases during DNA synthesis past the oxidation-damaged base 8-oxodG. Nature 349, 431-434.

Shin, K.H., and Park, J.G. (2000). Microsatellite instability is associated with genetic alteration but not with low levels of expression of the human mismatch repair proteins hMSH2 and hMLH1. Eur J Cancer 36, 925-931.

Shinohara, A., Ogawa, H., and Ogawa, T. (1992). Rad51 protein involved in repair and recombination in S. cerevisiae is a RecA-like protein. Cell 69, 457-470.

Shtutman, M., Zhurinsky, J., Simcha, I., Albanese, C., D'Amico, M., Pestell, R., and Ben-Ze'ev, A. (1999). The cyclin D1 gene is a target of the beta-catenin/LEF-1 pathway. Proc Natl Acad Sci U S A 96, 5522-5527.

Silverman, A.L., Park, J.G., Hamilton, S.R., Gazdar, A.F., Luk, G.D., and Baylin, S.B. (1989). Abnormal methylation of the calcitonin gene in human colonic neoplasms. Cancer Res 49, 3468- 3473.

Sinicrope, F.A., Rego, R.L., Halling, K.C., Foster, N., Sargent, D.J., La Plant, B., French, A.J., Laurie, J.A., Goldberg, R.M., Thibodeau, S.N., et al. (2006). Prognostic impact of microsatellite instability and DNA ploidy in human colon carcinoma patients. Gastroenterology 131, 729-737.

Sobol, R.W., Horton, J.K., Kuhn, R., Gu, H., Singhal, R.K., Prasad, R., Rajewsky, K., and Wilson, S.H. (1996). Requirement of mammalian DNA polymerase-beta in base-excision repair. Nature 379, 183-186.

Solt, D.B., Medline, A., and Farber, E. (1977). Rapid emergence of carcinogen-induced hyperplastic lesions in a new model for the sequential analysis of liver carcinogenesis. Am J Pathol 88, 595-618.

Soreide, K., Janssen, E.A., Soiland, H., Korner, H., and Baak, J.P. (2006). Microsatellite instability in colorectal cancer. Br J Surg 93, 395-406.

Soulas-Sprauel, P., Rivera-Munoz, P., Malivert, L., Le Guyader, G., Abramowski, V., Revy, P., and de Villartay, J.P. (2007). V(D)J and immunoglobulin class switch recombinations: a paradigm to study the regulation of DNA end-joining. Oncogene 26, 7780-7791.

Stefanini, M., Botta, E., Lanzafame, M., and Orioli, D. (2010). Trichothiodystrophy: from basic mechanisms to clinical implications. DNA Repair (Amst) 9, 2-10.

Stormorken, A.T., Bowitz-Lothe, I.M., Noren, T., Kure, E., Aase, S., Wijnen, J., Apold, J., Heimdal, K., and Moller, P. (2005). Immunohistochemistry identifies carriers of mismatch repair gene defects causing hereditary nonpolyposis colorectal cancer. J Clin Oncol 23, 4705-4712.

Su, L.K., Vogelstein, B., and Kinzler, K.W. (1993). Association of the APC tumor suppressor protein with catenins. Science 262, 1734-1737.

Subramanian, S., Mishra, R.K., and Singh, L. (2003). Genome-wide analysis of microsatellite repeats in humans: their abundance and density in specific genomic regions. Genome Biol 4, R13.

Sugasawa, K., Ng, J.M., Masutani, C., Iwai, S., van der Spek, P.J., Eker, A.P., Hanaoka, F., Bootsma, D., and Hoeijmakers, J.H. (1998). Xeroderma pigmentosum group C protein complex is the initiator of global genome nucleotide excision repair. Mol Cell 2, 223-232.

Swift, M., and Chase, C. (1979). Cancer in families with xeroderma pigmentosum. J Natl Cancer Inst 62, 1415-1421.

Takai, D., and Jones, P.A. (2002). Comprehensive analysis of CpG islands in human chromosomes 21 and 22. Proc Natl Acad Sci U S A 99, 3740-3745.

Tarsounas, M., Davies, D., and West, S.C. (2003). BRCA2-dependent and independent formation of RAD51 nuclear foci. Oncogene 22, 1115-1123.

Toyoshima, H., and Hunter, T. (1994). p27, a novel inhibitor of G1 cyclin-Cdk protein kinase activity, is related to p21. Cell 78, 67-74.

Umar, A., Boland, C.R., Terdiman, J.P., Syngal, S., de la Chapelle, A., Ruschoff, J., Fishel, R., Lindor, N.M., Burgart, L.J., Hamelin, R., et al. (2004). Revised Bethesda Guidelines for hereditary nonpolyposis colorectal cancer (Lynch syndrome) and microsatellite instability. J Natl Cancer Inst 96, 261-268.

Vogelstein, B., Fearon, E.R., Hamilton, S.R., Kern, S.E., Preisinger, A.C., Leppert, M., Nakamura, Y., White, R., Smits, A.M., and Bos, J.L. (1988). Genetic alterations during colorectal-tumor development. N Engl J Med 319, 525-532.

Wagner, R., Jr., and Meselson, M. (1976). Repair tracts in mismatched DNA heteroduplexes. Proc Natl Acad Sci U S A 73, 4135-4139.

Walther, A., Houlston, R., and Tomlinson, I. (2008). Association between chromosomal instability and prognosis in colorectal cancer: a meta-analysis. Gut 57, 941-950.

Watanabe, Y., Haugen-Strano, A., Umar, A., Yamada, K., Hemmi, H., Kikuchi, Y., Takano, S., Shibata, Y., Barrett, J.C., Kunkel, T.A., et al. (2000). Complementation of an hMSH2 defect in human colorectal carcinoma cells by human chromosome 2 transfer. Mol Carcinog 29, 37-49.

Watson, P., and Lynch, H.T. (1993). Extracolonic cancer in hereditary nonpolyposis colorectal cancer. Cancer 71, 677-685.

Watson, P., Vasen, H.F., Mecklin, J.P., Bernstein, I., Aarnio, M., Jarvinen, H.J., Myrhoj, T., Sunde, L., Wijnen, J.T., and Lynch, H.T. (2008). The risk of extra-colonic, extra-endometrial cancer in the Lynch syndrome. Int J Cancer 123, 444-449.

Wohlhueter, R.M., McIvor, R.S., and Plagemann, P.G. (1980). Facilitated transport of uracil and 5- fluorouracil, and permeation of orotic acid into cultured mammalian cells. J Cell Physiol 104, 309- 319.

Wolf, D., Rodova, M., Miska, E.A., Calvet, J.P., and Kouzarides, T. (2002). Acetylation of beta- catenin by CREB-binding protein (CBP). J Biol Chem 277, 25562-25567.

Woudstra, E.C., Gilbert, C., Fellows, J., Jansen, L., Brouwer, J., Erdjument-Bromage, H., Tempst, P., and Svejstrup, J.Q. (2002). A Rad26-Def1 complex coordinates repair and RNA pol II proteolysis in response to DNA damage. Nature 415, 929-933.

Yoon, Y.S., Yu, C.S., Kim, T.W., Kim, J.H., Jang, S.J., Cho, D.H., Roh, S.A., and Kim, J.C. (2011). Mismatch repair status in sporadic colorectal cancer: immunohistochemistry and microsatellite instability analyses. J Gastroenterol Hepatol 26, 1733-1739.

Zaanan, A., Flejou, J.F., Emile, J.F., Des, G.G., Cuilliere-Dartigues, P., Malka, D., Lecaille, C., Validire, P., Louvet, C., Rougier, P., et al. (2011). Defective mismatch repair status as a prognostic biomarker of disease-free survival in stage III colon cancer patients treated with adjuvant FOLFOX chemotherapy. Clin Cancer Res 17, 7470-7478.

Zhai, Y., Kronebusch, P.J., and Borisy, G.G. (1995). Kinetochore microtubule dynamics and the metaphase-anaphase transition. J Cell Biol 131, 721-734.

Zhang, H., Richards, B., Wilson, T., Lloyd, M., Cranston, A., Thorburn, A., Fishel, R., and Meuth, M. (1999). Apoptosis induced by overexpression of hMSH2 or hMLH1. Cancer Res 59, 3021-3027.

APPENDIX I – Chemical List and Suppliers

0.5 M Tris-HCl Buffer pH 6.8 BioRad, Mississauga, ON 1.5 M Tris-HCl Buffer pH 8.8 BioRad, Mississauga, ON 40% Acrylamide/Bis solution BioRad, Mississauga, ON 5-fluorouracil Hospira, Montreal, QC 50 bp DNA Ladder – Generuler MBI Fermentas, Burlington, ON Agarose Low EEO Electrophoresis Grade Fisher Scientific, Nepean, ON Amido Black BioRad, Mississauga, ON Ammonium Persulfate BioRad, Mississauga, ON Aprotinin Sigma- Aldrich, Oakville, ON Biorad DC Protein Assay BioRad, Mississauga, ON Caco-2 (CRC) American Type Culture Collection (ATCC), Manassas, VA Cell culture dishes Sarstedt, Newton, NC Cell lysis buffer (10X) Cell Signalling, Danvers, MA Chemiluminescence blotting substrate A Roche, Mississauga, ON Chemiluminescence blotting substrate B Roche, Mississauga, ON Crystal violet Fisher Scientific, Nepean, ON Dc Protein Assay BioRad, Mississauga, ON Dimethyl Sulfoxide (DMSO) Fisher Scientific, Nepean, ON DMEM Sigma- Aldrich, Oakville, ON DNA Loading Dye Life Technologies, Burlington, ON DNeasy Blood and Tissue Kit Qiagen, Mississauga, ON dNTPs MBI Fermentas, Burlington, ON DTT Fisher Scientific, Nepean, ON Ethidium Bromide BioRad, Mississauga, ON Fetal Bovine Serum Life Sciences, Burlington, ON Gentamycin Life Sciences, Burlington, ON Glycine Fisher Scientific, Nepean, ON Goat anti-mouse peroxidase conjugated antibody Sigma-Aldrich, Oakville, ON Goat anti-rabbit peroxidase conjugated antibody Sigma-Aldrich, Oakville, ON illustra* GFX PCR and DNA Purification Kit GE Healthcare, Piscataway, NJ Isopropanol Fisher Scientific, Nepean, ON Methanol Fisher Scientific, Nepean, ON Mouse anti-MSH2 antibody Calbiochem, San Diego, California PCR primers Sigma- Aldrich, Oakville, ON Phosphate buffered saline (10X) Sigma- Aldrich, Oakville, ON Phosphate buffered saline (1X) Sigma- Aldrich, Oakville, ON Phusion DNA polymerase NEB, Pickering, ON Phusion HF Buffer NEB, Pickering, ON PMSF Sigma-Aldrich, Oakville, ON PVDF membrane Roche, Mississauga, ON Rabbit anti-MLH1 antibody BD Biosciences, Mississauga, ON Rabbit anti-MSH6 antibody Betyl, Montgomery, TX Rabbit anti-PMS2 antibody Abcam, Cambridge, MA

Sodium chloride Fisher Scientific, Nepean, ON Sodium dodecyl sulfate (SDS), 20% Biorad, Missisauga, ON Sodium pyruvate Sigma- Aldrich, Toronto, ON TEMED (N,N,N,N-Tetramethylethylene diamine) Roche, Missisauga, ON Tris Base Fisher Scientific, Nepean, ON Triton X-100 Biorad, Missisauga, ON Trypsin (ATV) Veterinary College, Guelph, ON Tween-20 Fisher Scientific, Nepean, ON X-ray film Kodak, Rochester, NY

APPENDIX II – Preparation of Materials Used

0.1% TBST 10X TBS 100 mL Tween-20 1 mL H2O 900 mL Mix and store at room temperature

7.5% Resolving Gel H2O 11 mL 40% acrylamide-bis solution 3.72 mL 1.5 M Tris-HCl (pH 8.8) 5 mL 10% SDS 200 uL 10% APS 240 uL TEMED 10 uL Gel components are combined in the order listed above, mixed well, and then added to gel cassettes. A thin layer of isopropanol alcohol is then layered on top, and gels are allowed to set.

5% Stacking Gel H2O 3.2 mL 40% acrylamide-bis solution 500 uL 0.5M Tris-HCl (pH 6.8) 1.26 mL 10% SDS 50 uL 10% APS 50 uL TEMED 5 uL Once the resolving gel (prepared above) has set, the isopropanol is poured off and Whatman paper is used to blot the glass dry. The stacking gel mixture is prepared by adding components in the order listed above, and then poured on top of the resolving gel. A comb is then added to form lanes and the gel is allowed to set.

10X TBS Tris base 24.25 g NaCl 80 g H2O 600 mL Mix all components and adjust pH to 7.6 with HCl, then top up to 1L with H2O

50X TAE Tris base 242 g Glacial acetic acid 57.1 mL 0.5M EDTA 100 mL H2O to 1L Mix all components and store at room temperature

1X TAE 50X TAE 20 mL H2O 980 mL Mix all components and store at room temperature

10X Electrophoresis Buffer Tris base 30.2 g Glycine 144.2 g 20% SDS 50 mL H2O 600 mL Mix all components and adjust to pH 8.3 with HCl, then top up to 1 L with H2O

Amido Black Stain Amido black 0.1 g Acetic acid 2 mL Methanol 10 mL H2O 88 mL Mix all components and store at room temperature

Crystal Violet Stain Crystal violet 0.5 g Methanol 25 mL H2O 75 mL Mix all components and store at room temperature

Towbin’s Solution Tris Base 30.25 g Glycine 141.1 g H2O to 1 L Mix and store at room temperature

Wet Transfer Buffer H2O 768 mL 2% SDS 2 mL Towbin’s Solution 80 mL Methanol 150 mL Mix all components and store at 4°C

APPENDIX III – Primer Sequences

Name Primer Sequence (5’ – 3’) Product Size Tag Reference BAT-25 Forward: 120 bp 6-FAM on 5’ (Parsons et al., TCG CCT CCA AGA of forward 1995) ATG TAA GT primer

Reverse: TCT GCA TTT TAA CTA TGG CTC BAT-26 Forward: 116 bp 6-FAM on 5’ (Parsons et al., TGA CTA CTT TTG ACT of forward 1995) TCA GCC primer

Reverse: AAC CAT TCA ACA TTT TTA ACC C BRAF Forward: 283 bp None (Daniotti et al., CTC TAA GAG GAA 2007) AGA TGA AGT ACT AT

Reverse: AAT AGC CTC AAT TCT TAC CAT CC