Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

A bioinformatics meta-analysis of differentially expressed genes in colorectal cancer Chan, Simon Kit 2007

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2008_spring_chan_simon_kit.pdf [ 24.7MB ]
Metadata
JSON: 24-1.0066251.json
JSON-LD: 24-1.0066251-ld.json
RDF/XML (Pretty): 24-1.0066251-rdf.xml
RDF/JSON: 24-1.0066251-rdf.json
Turtle: 24-1.0066251-turtle.txt
N-Triples: 24-1.0066251-rdf-ntriples.txt
Original Record: 24-1.0066251-source.json
Full Text
24-1.0066251-fulltext.txt
Citation
24-1.0066251.ris

Full Text

A BIOINFORMATICS META-ANALYSIS OF DIFFERENTIALLY EXPRESSED GENES IN COLORECTAL CANCER  by SIMON KIT CHAN B.Sc. First Class Honours, Cell and Molecular Biology, Simon Fraser University, 2005  A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES (Bioinformatics)  THE UNIVERSITY OF BRITISH COLUMBIA December 2007  © Simon Kit Chan, 2007  Abstract BACKGROUND: Elucidation of candidate colorectal cancer biomarkers often  begins by comparing the expression profiles of cancerous and normal tissue by performing high throughput gene expression profiling. While many such studies have been performed, the resulting lists of differentially expressed genes tend to be inconsistent with each other, suggesting that there are some false positives and negatives. One logical solution to this problem is to determine the intersection of the lists of differentially expressed genes from independent studies. It is expected that genes that are biologically relevant to cancer tumorigenesis will be reported most often, while sporadically reported genes are due to the inherent biases and limitations of each of the profiling platforms used. However, the statistical significance of the observed intersection among many independent studies is usually not considered. PURPOSE: To address these issues, we developed a computational meta-analysis method that ranked differentially expressed genes based on the following criteria, which are presented in order of importance: the amount of intersection among studies, total tissue sample sizes, and average fold change in expression. We applied this meta-analysis method to 25 independent colorectal cancer profiling studies that compared cancer versus normal, adenoma versus normal, and cancer versus adenoma tissues. RESULTS: We observed that some genes were consistently reported as differentially expressed with a statistically significant frequency (P < .0001) in the cancer versus normal and adenoma versus normal comparisons, ii  but not in the cancer versus adenoma comparison. We performed a review of some of the high ranking candidates and determined that some have previously been shown to have diagnostic and/or prognostic utility in colorectal cancer. More interestingly, the meta-analysis method also identified genes that had yet to be tested and validated as biomarkers. Thus, these candidates are currently being validated at the protein level on colorectal tissue microarrays. CONCLUSION: Our meta-analysis method identified genes that were consistently reported as differentially expressed. Besides identifying new biomarker candidates, our meta-analysis method also provides another filter to remove false positive genes from further consideration. In conclusion, the genes presented here will aid in the identification of highly sensitive and specific biomarkers in colorectal cancer.  iii  Table of Contents Abstract ^  ii  Table of Contents ^  iv  List of Tables ^  vi  List of Figures ^ Acknowledgements ^ Dedication ^  vii viii ix  Co-Authorship Statement ^ Chapter 1: Introduction ^  1  1.1^INTRODUCTION TO COLORECTAL CANCER ^ 1 1.2^STAGES OF COLORECTAL CANCER ^ 2 1.3^THE UNDERLYING MOLECULAR CHANGES DURING THE PROGRESSION OF NORMAL TO CANCER TISSUE: THE WNT SIGNALING PATHWAY AND THE ADENOMATOUS POLYPOSIS COLI (APC) GENE 3  1.3.1 APC is the Gatekeeper to Colorectal Cancer ^ 4 1.3.2 K-ras Mutations in Intermediate Stage Adenomas^ 5 1.3.3 Loss of Heterozygosity on 18q is associated with Late Stage Adenomas ^ 6 1.3.3.1 The Deleted in Colorectal Cancer (DCC) Gene is a Candidate Tumor Suppressor on 18q 6 1.3.3.2 The SMAD family member 4 (SMAD4) Gene is Another Candidate Tumor Suppressor on 18q ^ 7 1.3.4 p53 Mutation in the Progression of Adenoma to Cancer ^ 8 1.4^TYPES OF COLORECTAL CANCER ^  1.4.1 Familial Adenomatous Polyposis (FAP) ^ 1.4.2 Hereditary Nonpolyposis Colorectal Cancer (HNPCC) ^ 1.4.3 Sporadic Colorectal Cancer ^ 1.5^COLORECTAL CANCER BIOMARKERS ^ 1.6^COLORECTAL CANCER BIOMARKERS DISCOVERY USING TRANSCRIPT BASED GENE EXPRESSION PROFILING METHODS ^ 1.7^ANALYSIS OF MULTIPLE INDEPENDENT STUDIES TO IDENTIFY BIOLOGICALLY RELEVANT GENES 18 1.8^META-ANALYSES OF HIGH THROUGHPUT GENE EXPRESSION STUDIES ^ 1.9^THESIS OVERVIEW ^ CHAPTER 1 REFERENCES ^ Chapter 2: Materials and Methods ^ 2.1^META-ANALYSIS OF PUBLISHED DATASETS OF DIFFERENTIALLY EXPRESSED GENES IN COLORECTAL CANCER ^  9  10 11 13 14 16  20 20 24 37 37  2.1.1 Data Collection and Curation ^ 38 2.1.2 Differentially Expressed Genes Mapped to Entrez Gene IDs ^ 38 2.1.3 Total Gene Lists for Each Study^ 39 2.1.4 Assessment of Significance of Study Overlap using Simulations ^ 39 CHAPTER 2 REFERENCES ^ Chapter 3: Results ^  43 45  3.1^MAPPING SUCCESS RATE AND SIGNIFICANCE OF OVERLAP OBSERVED ^ 45 3.2^OVERLAP RESULTS FOR CANCER VERSUS NORMAL TISSUE COMPARISONS ^ 45 3.3^OVERLAP RESULTS FOR ADENOMA VERSUS NORMAL TISSUE COMPARISONS ^ 47  iv  3.4^OVERLAP RESULTS FOR CANCER VERSUS ADENOMA TISSUE COMPARISONS ^ 48 3.5^OVERLAPPING RESULTS IN THE CANCER VERSUS NORMAL AND ADENOMA VERSUS NORMAL COMPARISONS ^ 48 CHAPTER 3 REFERENCES ^ 73  Chapter 4: Discussion ^  77  4.1^META-ANALYSES TO ELUCIDATE BIOLOGICALLY RELEVANT GENES IN COLORECTAL CANCER TUMORIGENESIS ^ 77 4.2^STATISTICAL SIGNIFICANCE IN OVERLAP IN TWO OUT OF THREE COMPARISONS ^ 78 4.3^GENES REPORTED WITH INCONSISTENT DIRECTION OF DIFFERENTIAL EXPRESSION ^ 79 4.4^LITERATURE REVIEW OF CONSISTENTLY REPORTED DIFFERENTIALLY EXPRESSED GENES ^ 82 4.5^CONSISTENTLY REPORTED DIFFERENTIALLY EXPRESSED GENES IN BOTH CANCER VERSUS NORMAL AND ADENOMA VERSUS NORMAL COMPARISONS ^ 90 CHAPTER 4 REFERENCES ^ 91 Chapter 5: Conclusion and Future Studies ^ 5.1^VALIDATION OF TRANSCRIPT BASED GENE EXPRESSION PROFILING RESULTS ON COLORECTAL TISSUE MICROARRAYS ^ CHAPTER 5 REFERENCES ^  97 97 99  V  List of Tables TABLE 1.1: TWENTY-THREE COLORECTAL CANCER VERSUS NORMAL TISSUE EXPRESSION PROFILING 22 STUDIES INCLUDED IN ANALYSIS ^ TABLE 2.1: SEVEN COLORECTAL ADENOMA VERSUS NORMAL TISSUE EXPRESSION PROFILING STUDIES 41 INCLUDED IN ANALYSIS ^ TABLE 2.2: FIVE CANCER VERSUS ADENOMA TISSUE EXPRESSION PROFILING STUDIES INCLUDED IN ANALYSIS ^ 42 TABLE 3.1: SUMMARY OF COMPARISONS STUDIED FOR OVERLAP SIGNIFICANCE ^ 49 TABLE 3.2: UP-REGULATED GENES MOST COMMONLY REPORTED IN CANCER VERSUS NORMAL EXPRESSION PROFILING STUDIES ^ 50 TABLE 3.3: DOWN-REGULATED GENES MOST COMMONLY REPORTED IN CANCER VERSUS NORMAL EXPRESSION PROFILING STUDIES ^ 54 TABLE 3.4: UP-REGULATED GENES REPORTED IN THREE OR FOUR CANCER VERSUS NORMAL 57 EXPRESSION PROFILING STUDIES ^ TABLE 3.5: DOWN-REGULATED GENES REPORTED IN THREE OR FOUR CANCER VERSUS NORMAL EXPRESSION PROFILING STUDIES ^ 64 TABLE 3.6: UP-REGULATED GENES MOST COMMONLY REPORTED IN ADENOMA VERSUS NORMAL EXPRESSION PROFILING STUDIES ^ 68 TABLE 3.7: DOWN-REGULATED GENES MOST COMMONLY REPORTED IN ADENOMA VERSUS NORMAL EXPRESSION PROFILING STUDIES ^ 70 TABLE 3.8: DIFFERENTIALLY EXPRESSED GENES REPORTED WITH A STATISTICALLY SIGNIFICANT FREQUENCY IN THE ADENOMA VERSUS NORMAL AND CANCER VERSUS NORMAL COMPARISONS . 72  vi  List of Figures FIGURE 1.1: KEY MUTATIONS THAT OCCUR DURING THE PROGRESSION OF COLORECTAL CANCER. ^ 9 FIGURE 2.1: A PIPELINE SUMMARIZING THE STEPS UNDERTAKEN. ^ 40 FIGURE 3.1: OVERLAP ANALYSIS RESULTS FOR THE CANCER VERSUS NORMAL COMPARISON ^ 47  vi i  Acknowledgements I wish to acknowledge my senior supervisor, Dr. Steven Jones, and my committee member, Dr. Isabella Tai. In a letter to Robert Hooke, Sir Isaac Newton wrote "If I have seen further, it is by standing on ye shoulders of Giants." I am grateful to Dr. Jones and Dr. Tai for their patience and guidance throughout this thesis research. If I have gained anything from my graduate training, it is because they have kindly let me stand on their shoulders.  I would also like to acknowledge Dr. Mark Wilkinson for being on my committee, as well as for asking thought provoking questions at my committee meetings.  I wish to thank members of the Genome Sciences Centre where this thesis research was conducted. They provided much support and contributed to a friendly working environment. Special thanks to fellow graduate student, Obi Griffith, for the valuable insight he provided during our countless discussions. I am grateful to the CIHR/MSFHR Strategic Training Program in Bioinformatics for Health Research for providing financial support. Furthermore, the UBC Bioinformatics Coordinator, Sharon Ruschkowski, provided support in applying for funding and registration for classes.  Finally, I wish to thank my parents for their unconditional love and support.  viii  Dedication To my parents  ix  Co-Authorship Statement Dr. Steven Jones, Dr. Isabella Tai, Obi Griffith, and Simon Chan all made important intellectual contributions to this thesis research. Obi Griffith developed an earlier version of the meta-analysis method that was used in this thesis research and contributed a portion of the bioinformatics code. Simon Chan performed the data collection, curation, performed the analyses, produced the figures and tables, and wrote this document. Dr. Isabella Tai supervised the data collection portion of this thesis research and helped interpret results. Dr. Steven Jones conceived of this project, helped interpret the results, and provided financial support. Finally, Rhonda Oshanek edited the document for content and grammar.  x  Chapter 1: Introduction' 1.1 Introduction to Colorectal Cancer Colorectal cancer is defined as cancerous growths in the colon, rectum, or appendix and represents the third most frequent cancer in both men and women in North America [1]. In Canada, colorectal cancer is the second most common cause of cancer-related death. This year, an estimated 20,800 Canadians will be diagnosed with colorectal cancer and approximately 8,700 will die of it [2]. When diagnosed and treated early, the majority of patients show favourable five year survival rates. Among patients diagnosed when the cancer is localized to the colon or rectum, 90% of patients show good survival rates, while the survival rate is 68% among patients diagnosed when the cancer has spread to the surrounding tissues. Unfortunately, approximately 19% of patients are diagnosed when the cancer has metastasized to other organs, such as the liver, and the corresponding survival rate is only 10% [3]. Thus, regular screening remains the most promising strategy to reducing the mortality rates associated with colorectal cancer.  1 A version of each chapter was compiled into a manuscript that has been accepted for publication in the peer reviewed journal Cancer Epidemiology, Biomarkers & Prevention:  Chan SK, Griffith OL, Tai IT, Jones SJM. Meta-analysis of Colorectal Cancer Gene Expression Profiling Studies Identifies Consistently Reported Candidate Biomarkers. (2007). In press. Cancer Epidemiol Biomarkers Prey.  1  1.2 Stages of Colorectal Cancer The transformation of healthy colon epithelia to neoplasia has provided an excellent system to study the histological and molecular changes that are associated with the progression of colorectal cancer [4, 5]. The first step in the development of colorectal cancer is the emergence of aberrant crypt foci (ACF), which are small lesions in the intestinal epithelium [6]. ACF are composed of cells of normal morphology or dysplastic cells, of which the latter are more likely to develop into benign polyps, also known as adenomas, which protrude into the lumen. By the age of 70, approximately 50% of men and women from the western world will have developed an adenoma polyp [5]. Approximately one in ten of these adenomas will progress into the carcinoma stage.  As for many other solid tumors, colorectal cancer is staged mainly on the basis of both the extent of invasion at the primary site and of the presence of metastasis in local lymph nodes and distant organs. The two most common colorectal cancer staging systems are Dukes [7] and TNM [8]. Both systems distinguish between tumors confined to the surface of the colon wall (Dukes A or Stage I), those penetrating the muscle wall but without metastasis (Dukes B or Stage II), those with metastasis to surrounding lymph nodes (Dukes C or Stage III), and finally, those that have metastasized to distant organs, such as the liver (Dukes D or Stave IV).  2  1.3 The Underlying Molecular Changes during the Progression of Normal to Cancer Tissue: The Wnt Signaling Pathway and the Adenomatous Polyposis Coli (APC) Gene Molecular changes in colonic epithelium underlie the histological changes seen in the progression of normal to cancer tissue. One molecular pathway that has been implicated in the development of colorectal cancer is that of the Wnt signaling pathway. This pathway controls cellular division, cell-cell adhesion, and other activities by inducing the expression of genes like myc, ccndl , and mmp-7 [9-12]. This induction is mediated through the binding of the extracellular Wnt signal to the Frizzled and Low density lipoprotein binding protein (LBP) transmembrane proteins. In the absence of the Wnt signal, cytoplasmic 13catenin is destabilized through its interaction with a destruction complex, made up of scaffolding proteins Adenomatous Polyposis Coli (APC), axin (AXIN2), and gylogen synthase kinase 313 (GSK3(3). When bound to the destruction complex, cytoplasmic [3-catenin is ubquinated, leading to its proteasomal degradation. When the Wnt signal is present and binds to Frizzled and LBP, the destruction complex is inactivated via Dishevelled (DSH), thus allowing stabilized 13-catenin to enter the nucleus and interact with DNA-binding proteins of the T-cell-factor (TCF) family. Finally, this complex induces expression of target genes by binding to their respective regulatory regions.  3  1.3.1 APC is the Gatekeeper to Colorectal Cancer The apc gene and the effects of its mutation have been an area of great interest within the colorectal cancer research community. Inactivation of both alleles of apc has been located in dysplastic aberrant crypt foci, which are the earliest  neoplastic lesions and an early precursor to adenomas [5]. Furthermore, mutations in this tumor suppressor gene have been found in up to 80% of colorectal cancer samples [13]. Thus, apc has often been described as the "gatekeeper" to colorectal cancer, as its inactivation leads to the initiation of the disease [14].  As described, the APC protein is a part of the destruction complex within the Wnt pathway. Inactivation of both alleles of apc renders the destruction complex ineffective at promoting the degradation of cytoplasmic 13-catenin, thus allowing it to accumulate and enter the cell nucleus, where it binds with TCF and together induce the expression of the target genes. Thus, the Wnt signal transduction pathway becomes constitutively active, with the net result being an increase in cell division of the colonic epithelial cells. Furthermore, recent studies have also shown that even one mutant allele of apc can contribute to the chromosome instability of colorectal tumor cells [15, 16]. APC protein has been implicated in the proper formation and function of mitotic spindles, as mutations in this gene have resulted in cells with chromosome segregation errors [15, 17, 18]. That only one mutant allele of apc can result in these errors suggests that the mutant  4  allele has a dominant negative effect on the remaining wildtype allele. Consistent with this hypothesis, a recent study demonstrated that a mutant APC protein forms a hetero-oligomer with that of the endogenous APC protein, thus interfering with its proper functioning [16].  1.3.2 K ras Mutations in Intermediate Stage Adenomas -  After development of adenomas from dysplastic ACF, mutations in other genes contribute to colorectal cancer progression. For example, approximately 50% of intermediate-staged adenomas have activating mutations in the k-ras oncogene [4, 19]. K-ras is part of a family of three genes, the other members being h-ras and n-ras, and encodes a small GTPase involved in the transduction of external stimuli across the plasma membrane to the effector molecules that promote cell division. An activation mutation in just one allele of this gene results in the inability of k-ras to be inactivated, thus resulting in uncontrolled cell division and malignant transformation [19]. Furthermore, mutated k-ras has been shown to promote local cell invasion by inducing the expression of metalloproteases, such as mmp-7 [20], to breakdown the underlying basement membrane of the epithelial cells, thus releasing tumor cells from the primary tumor [21].  5  1.3.3 Loss of Heterozygosity on 18q is associated with Late Stage Adenomas 1.3.3.1 The Deleted in Colorectal Cancer (DCC) Gene is a Candidate Tumor Suppressor on 18q Loss of heterozygosity (LOH) on 18q, found in approximately 70% of colorectal cancer samples, is associated with the progression of intermediate-stage adenomas to late-stage adenomas [4, 5, 22]. LOH has been implicated as an important mechanism for the inactivation of tumor suppressor genes [23]. Deleted in Colorectal Cancer (DCC) is one candidate tumor suppressor gene that is in the deleted region [24]. DCC functions as part of a receptor complex for the netrin 1 (NTN1) protein [25-27]. The netrins are a family of secreted proteins with roles in directing the growth and orientation of axons in the developing nervous system [28, 29]. While an obvious role for DCC and NTN1 in the intestine is not evident [28, 30], some studies have suggested that they have roles in regulating apoptosis, in which the expression of DCC in the absence of NTN1 induces apoptosis, while in the presence of NTN1, this expression is blocked [31-33]. Thus, DCC has been proposed to be a member of the Dependence Receptor Family. These receptors promote a certain cellular state in the presence of the ligand, but in its absence, the receptors promote apoptosis [34], which suggests a possible role for DCC in homeostatic regulation of colonic epithelium. This role of DCC is consistent with the concentration gradient of NTN1 along the microvilli axis. The cells in the top portion of the microvilli, where NTN1 levels are low, are in the process of undergoing apoptosis. Conversely,  6  the cells in the bottom of the microvilli, where NTN1 levels are high, undergo proliferation [22].  However, many other experiments have produced results that questioned whether DCC represents the major tumor suppressor gene in the deleted region of 18q. Firstly, germline mutations that segregate with cancer predisposition are convincing evidence that the candidate gene in question is a tumor suppressor. However, unlike other firmly established tumor suppressors such as apc or rbl , there is no evidence that DCC germline mutations play a role in a heritable predisposition to colorectal cancer [22]. Furthermore, there are few reports of somatic mutations in DCC in colorectal cancer [22]. The most common somatic mutation is that of the expansion of a dinucleotide repeat tract located in an intron downstream of the seventh exon [24]. However, whether this expansion actually results in decreased expression of DCC has not been demonstrated [22].  1.3.3.2 The SMAD family member 4 (SMAD4) Gene is Another Candidate Tumor Suppressor on 18q Another candidate tumor suppressor gene in the 18q region is smad4. The smad genes consist of at least nine members [35], some of which play a critical role in the TGFf3 signaling pathway [36]. Binding of the TGF6 ligand to the transmembrane receptor TGF6-R2 activates it, which results in the phosphorylation of TGF6-R1. Next, this transmembrane receptor phosphorylates cytoplasmic proteins SMAD2 or SMAD3, which allows the protein to form a hetero-oliogmer with SMAD4 protein. The resulting complex enters the nucleus  7  to directly or indirectly regulate transcription of target genes, leading to the regulation of cell division.  Germline inactivating mutations of smad4 result an in increased risk of juvenile polyposis syndrome [37], in which affected individuals develop hamartomatous polyps in the intestine and colon and have an increased risk for colorectal and other gastrointestinal cancers. Furthermore, mice that were heterozygous for apc and smad4 inactivating mutations showed an increased susceptibility of  malignant progression compared to mice with only heterozygous inactivating mutations for apc [38]. However, because smad4 is only mutated in about 33% of colorectal cancer samples with chromosome 18q LOH [39-41], it is unlikely to be the major tumor suppressor in the 18q region [22].  1.3.4 p53 Mutation in the Progression of Adenoma to Cancer Mutations in the tp53 tumor suppressor gene occur during the transition from adenoma to cancer [4, 42]. The encoded protein, p53, is a transcription factor that functions as a tetramer [43]. The target genes of p53 are involved in many diverse cellular processes such as apoptosis, cell cycle arrest, DNA repair, cellular differentiation, and many others [44]. This 'guardian of the genome' [45] is mutated in approximately 50% of all human cancers [46, 47]. Heterozygous germ-line mutations in tp53 predispose individuals to a wide range of tumor types at an early age, a condition known as Li-Fraumeni syndrome [48]. Approximately 94% of the tumor-associated mutations of TP53 are point mutations that result in  8  single amino-acid substitutions [49]. These substitutions can result in a protein that is more stable than wildtype p53 and thus can act as dominant negative inhibitors.  While the studies reported in the literature tend to present the occurrence of these mutations as a linear series of events (Figure 1.1), it should be noted that other studies have shown that it is uncommon for all of the described mutations to be found in one colorectal cancer sample [50, 51], thus emphasizing the heterogeneity of the disease and the critical roles that other genes likely play in cancer initiation and progression.  APC  Normal Epithelium  ^  ■•••  Dysplastic ACF  K-RAS 18q LOH  r.■11/  Adenoma  ^  p53  Cancer  loon■O  Metastasis  Figure 1.1: Key mutations that occur during the progression of colorectal cancer. Mutation in K-RAS occurs in the early stages of the adenoma polyp state, while loss of heterozygosity of 18q occurs in the later stages.  1.4 Types of Colorectal Cancer In general, colorectal cancer is divided into two types: hereditary and nonhereditary (sporadic). Common hereditary colorectal cancer syndromes include  9  Familial Adenomatous Polyposis (FAP) and Hereditary Nonpolyposis Colorectal Cancer (HNPCC) [52, 53], both of which are due to the occurrence of highly penetrant mutations that result in the development of cancer by approximately 40 years of age, while onset of sporadic colorectal cancer occurs at approximately 67 years of age [5].  1.4.1 Familial Adenomatous Polyposis (FAP)  FAP, an autosomal dominant disease, is caused by mutations that lead to the inactivation of the apc gene. Patients with FAP have inherited one mutated apc allele from an affected parent and later acquired a somatic mutation of the remaining wildtype apc allele. The majority of these mutations are nonsense mutations, thus resulting in a truncated APC protein [17, 52-54]. By age 16, the development of hundreds to thousands of adenoma polyps in the colon has taken place. Without treatment, the average age of cancer onset is 39 years of age [54] with a life expectancy of 42 years [55]. While these individual adenoma polyps are endoscopically and histologically identical to those resulting from sporadic colorectal cancer, their early onset and sheer number guarantees that at least some will acquire further mutations and develop into colorectal cancer [54].  Due to the aggressiveness of the syndrome, early diagnosis and screening are essential. Currently, genetic testing can be used clinically to identify individuals at risk before the development of polyps. Most often, genetic testing involves  10  DNA sequencing of apc to detect mutations followed by complementary methods to test for protein truncation [56, 57]. Once it has been established that an individual is at risk for FAP, screening and surveillance methods are employed to monitor the patient. Flexible sigmoidoscopy screening is performed annually and continues until 35 years of age, at which point the time interval can be reduced to once every three years if no polyps have been observed [58, 59]. For those with FAP, prophylactic colectomy during adolescence is often the treatment [59].  1.4.2 Hereditary Nonpolyposis Colorectal Cancer (HNPCC)  The other major type of hereditary colorectal cancer is HNPCC. Inherited in an autosomal dominant fashion, HNPCC is the most common form of hereditary colorectal cancer [60, 61]. Mutations in five mismatch repair (MMR) genes are responsible for the development of this disorder. In patients with HNPCC, msh2 and mlhl are the most commonly mutated genes, while mutations in hpmsl, hpms2, and hmsh6 are found less frequently [54]. MMR genes function to  recognize and repair DNA replication errors, typically the insertion or deletion (indel) of one or more nucleotides [56, 62]. Defects in MMR genes can often be detected in microsatellites, which are sequences found throughout the genome that contain many one to three nucleotide repeats. Tumor cells in HNPCC patients demonstrate microsatellite instability (MSI), in which these sequences expand or contract in length due to the inability of the cell to repair these DNA replication errors. Therefore, MSI can impact the regular function of a protein. For example, it has been demonstrated that TGF8-R2, which contains a  11  polyadenine tract, is responsible for relaying the binding of the TGF-I3 signal to the nucleus to induce the expression of target genes. This signal transduction pathway has been shown to inhibit the epithelial cell division, thus loss of this receptor has been associated with tumor growth [63, 64]. In many MSI colorectal cancer cell lines, short indels in both alleles results in frameshift mutations, which leads to this gene's inactivation [65-67].  The adenomas produced in HNPCC patients cannot be distinguished endoscopically from those that result from sporadic colon cancer. Adenomas in HNPCC patients develop between 30 and 40 years of age, are larger, and have more villous and dysplastic traits [68, 69] than those in the general population. In the absence of treatment, the average age of cancer onset is 44 years of age [70].  Similar to FAP, genetic tests are in place to identify subjects at risk for HNPCC. To evaluate families at risk for HNPCC, the Bethesda guideline has been developed [57, 62, 71]. This guideline contains of a series of clinical criteria that should be met before genetic testing for HNPCC is conducted. If the guidelines are met, genetic testing is conducted in which five specific DNA microsatellites are evaluated for mutations. If at least two of the five are abnormal, the tumor is labeled as `MSI-high' [62, 71] and are candidates for msh2 and mlhl genetic testing. If only one of the five is abnormal, the tumor is labeled as `MSI-low.' Finally, if no abnormalities are found, the tumor is labeled as stable. For patients  12  at risk of HNPCC, full colonoscopy is recommended once every one to two years beginning at approximately 20 years of age [58, 72] because this has been demonstrated to be effective at reducing incidence of colorectal cancer and mortality [73]. Similar to FAP, prophylactic colectomy is the recommended treatment for HNPCC patients [54].  1.4.3 Sporadic Colorectal Cancer  Data on the movement of immigrants show that people who move from a low-risk area to high-risk area for colorectal cancers rapidly reach equivalent incidence rates to those of the adopted country [74, 75]. This observation suggests that environmental factors play a role in the development of colorectal cancer. Furthermore, hereditary colorectal cancer has been estimated to account for only 20 to 25% of all colorectal cancer cases [53], suggesting that, in fact, the majority of cases of colorectal cancer result from environmental factors.  Diet is one environmental factor that has been well-studied for its association to colorectal cancer incidence. Diets rich in fiber, fruits, and vegetables have been proposed to protect against colorectal cancer [76] and adoption of such a diet could prevent 50 to 75% of colorectal cancer cases [77]. Conversely, high intake of red meat and saturated fats has been shown to have a positive association with the incidence of colorectal cancer by both retrospective [78-83] and prospective [84-86] studies. Furthermore, a recent meta-analysis of prospective studies has demonstrated that consumption of red meat was associated  13  significantly with increased odds of colorectal cancer [87]. Despite of the abundance of this epidemiological evidence, little is known about the molecular mechanisms that underlie the protective or predisposing effects modulated by diet [88, 89]. However, the molecular mechanisms of initiation and progression of sporadic cases are similar to those described for hereditary cases, with inactivation of apc and MMR genes occurring in sporadic cases as well [5].  1.5 Colorectal Cancer Biomarkers A biomarker is defined as any measurable trait that is indicative of some biological state [90]. Biomarkers exist in many different forms, such as physiological measurements (e.g., weight and blood pressure), images (e.g., mammograms), genetic alterations (e.g., mutations in apc), molecules (e.g., prostate-specific antigen, PSA), and others [91]. However, the term tiomarker is now more commonly used in the context of measuring molecules from patient samples such as serum, urine, and tissue [90].  Elucidation of biomarkers is an active area of research, as they are valuable in many different aspects of disease management [92]. Firstly, biomarkers could be used in the stratification of patients based on cancer risk to identify those most or least likely to benefit from further screening. For example, women with deleterious mutations in BRCA1 and BRCA2 have markedly increased risk of developing ovarian and breast cancer [93]. Therefore, only the women with these mutations need to undergo aggressive risk reduction strategies, thus  14  sparing women without these mutations from unnecessary treatment. Secondly, diagnostic biomarkers could be used in tests to screen asymptomatic people for early stages of cancer. By identifying these people, proper treatment can be undertaken earlier, thus potentially resulting in improved prognosis. Prostatespecific antigen (PSA) in serum is an example of such a biomarker used in the early detection and management of prostate cancer [94-96]. If a patient has consistently elevated levels of serum PSA, then further testing, such as with a prostate biopsy, will take place. Thirdly, biomarkers may serve as surrogate endpoints for a therapeutic response that can potentially shorten the length of clinical trials and thus accelerate drug development. For example, a clinical trial is currently comparing four treatment options for chronic myelogenous leukemia (CML) [92]. The success of a treatment will be defined as a greater than four-log reduction in the break point cluster region-Abelson (BCR-ABL) signal. Using the reduction of this DNA marker as the endpoint, instead of the subjects entering remission, will shorten the clinical trial from several years to approximately one year [92]. Finally, biomarkers can also be used to monitor those patients who have gone into remission for cancer recurrence. For example, Genomic Health, a California-based company, demonstrated that its marketed test, Oncotype Dx, which consists of a 21 gene assay, could predict recurrence of tamoxifen-treated, node-negative breast cancer [97].  15  1.6 Colorectal Cancer Biomarkers Discovery using Transcript based Gene Expression Profiling Methods Discovery and validation of biomarkers in colorectal cancer can potentially lead to better treatment and management of the disease. For example, biomarkers could be used in a test to diagnose colorectal cancer in asymptomatic patients. Currently, the non-invasive fecal-occult blood test (FOBT) is used to screen patients for colorectal cancer. A stool sample is collected onto a pad that will change colour if the presence of blood is detected in the sample, as determined by peroxidase activity found in hemoglobin [98]. Typically, a colonoscopy is used as the follow-up test to a positive FOBT result. While large randomized clinical trials have shown that FOBT is effective in reducing colorectal cancer mortality by up to 30% [99-101], the test has poor sensitivity in detecting early-stage lesions, as adenomas frequently do not bleed [98]. Furthermore, false positives are also common, as the presence of peroxidase in stool can originate from rare red meat or vegetables in the diet [102]. Thus, highly sensitive and specific diagnostic biomarkers will aid in diagnosing colorectal cancer in subjects.  Another example of biomarkers potentially leading to better treatment and management of colorectal cancer involves patients with Dukes' C colorectal cancer. Post surgical Dukes' C patients have a 60% likelihood of recurrence within five years. Treatment of these patients with post surgical adjuvant therapy reduces the recurrence rate to between 40 and 50% and is thus now the standard treatment for these patients [103-105]. However, because it is currently  16  not possible to accurately distinguish between those who are surgically cured and those who will experience disease recurrence, the majority of Dukes' C patients are administered adjuvant therapy even if there will be minimal benefits for them [106]. Thus, the discovery of highly sensitive and specific prognostic biomarkers will allow discrimination between patients who need aggressive post surgical adjuvant therapy and those who do not.  With these issues in mind, many researchers have focused on elucidating biomarkers for colorectal cancer by utilizing gene expression profiling methods on tumor and normal colon mucosa samples, which can be obtained during surgery or colonoscopies. Some examples of such studies include the expression profile of colorectal tumors being compared to that of normal mucosa to identify candidate diagnostic biomarkers (Table 1.1, page 22). Other studies have attempted to identify candidate prognostic biomarkers by comparing the expression profiles of tissue samples from patients of one outcome versus those from patients of another, such as those who will experience disease recurrence and those who will not [106, 107].  In a typical study, the differences in gene expression between tissue samples from different biological states are determined to study the underlying molecular differences. Hybridization based methods, such as oligo-nucleotide and cDNA two channel microarrays, involve the hybridization of fluorescently labeled transcripts to slides spotted with nucleic acid probes [108]. Sequence based  17  methods, such as Serial Analysis of Gene Expression (SAGE), involve sequencing many short tags, each representative of a transcript, from the various tissue types [109]. In both types of methods, statistical and computational methods are applied to the signal, relative hybridization of transcripts in microarrays and relative abundances of tags in SAGE, to determine which genes are differentially expressed between the biological samples under consideration.  1.7 Analysis of Multiple Independent Studies to Identify Biologically Relevant Genes With the increase of available data from transcript based gene expression profiling experiments, many potentially useful biomarkers have been determined that may have diagnostic and/or prognostic utility in colorectal cancer. However, few reliable biomarkers have resulted in useful tests to be performed in a clinical setting. One explanation for this lack of translational success would be the inconsistency of the results generated by independent studies [1, 89, 110]. High throughput expression profiling studies usually result in tens to thousands of differentially expressed genes, only a small fraction of which are reproduced by independent studies. Explanations for this low overlap include utilization of different tissue resection methods, such as microdissection versus laser capture microdissection, different expression profiling technologies, such as cDNA twochannel microarrays, oligonucleotide microarrays, and SAGE, each with their own inherent biases and limitations [111, 112], as well as different analysis methods such as multiple correction tests and fold change thresholds. Given the low overlap between multiple independent studies, it is difficult to determine  18  which differentially expressed genes should be further studied and validated as biomarkers.  One logical solution to this problem would be to take the intersection of multiple independent studies to identify the genes that are consistently reported as differentially expressed. It is expected that genes biologically relevant to colorectal cancer tumorigenesis will be the most commonly reported, while those due to the inherent noise or biases in the different experimental methodologies utilized will be underrepresented [113]. Many different groups have determined the overlap between independent expression profiling studies [1, 89, 110]. While such lists are useful, the statistical significance of the overlap is often not considered. One can imagine randomly choosing genes from each expression profiling platform of each independent study, randomly labeling each gene as upor down-regulated, and observing some overlap due to chance alone. Therefore, it is important to determine which differentially expressed genes were consistently reported in independent colorectal cancer expression profiling studies with a statistically significant frequency because these genes represent the best candidates for further validation as biomarkers with other experimental techniques.  19  1.8 Meta-Analyses of High Throughput Gene Expression Studies With the rapid growth of related high throughput gene expression data, efforts have been made to combine this data from indendent studies to increase the power to detect a certain outcome, such as the differential expression of a gene [114]. This process is known as meta-analysis [115]. Many such meta-analysis studies have been conducted on microarray studies related to cancer research. [113, 116-119]. Generally, such meta-analyses involve re-analyzing this raw high throughput gene expression data in a consistent fashion. Thus, due to the fact that raw data is often not made available, most of the cancer specific metaanalyses utilize at most five studies. This limiting factor can potentially affect the number of consistently reported differentially expressed genes.  1.9 Thesis Overview With these limitations in mind, a novel meta-analysis method was recently developed by our group and applied to published studies of differentially expressed genes in thyroid cancer [120]. This meta-analysis method separated genes demonstrating biological relevance from those that were reported spuriously and has the added advantage of not requiring raw data. The approach involved a vote-counting strategy in which a gene was ranked according to the number of studies reporting its differential expression, the total number of tissue samples utilized in the studies, and the average fold change. That study resulted in a panel of 12 differentially expressed genes reported at a 20  frequency highly unlikely to have occurred by chance. The panel contained both well known thyroid cancer markers as well as some uncharacterized genes, demonstrating the ability of the meta-analysis method to highlight novel candidate biomarkers. With these results in mind, the objective of the current study was to apply the meta-analysis method to colorectal cancer to determine whether a statistically significant level of overlap between studies could be observed and to identify promising biomarkers. Furthermore, we improved the meta-analysis method by dividing genes into semi-quantitative categories based on the number of tissue samples. The categories were lowest (first quartile, Q1), moderate (interquartile range), and greatest (values greater than those in third quartile, Q3) number of tissue samples. This improvement highlighted genes that may have shown the greatest fold changes, but would have been ranked lower by the original meta-analysis method due to fewer tissue samples studied. We curated published lists of differentially expressed genes from 25 independent studies performing high throughput transcript based gene expression profiling to compare colorectal cancer to normal tissue, adenoma to normal tissue, and colorectal cancer to adenoma tissue. We observed that many genes were consistently reported as differentially expressed in multiple studies and that this overlap was highly significant. The list of candidate biomarkers that resulted from this thesis research will be a valuable resource to the colorectal cancer research community for further studies.  21  Table 1.1: Twenty-three Colorectal Cancer versus Normal Tissue Expression Profiling Studies Included in Analysis  Upregulated Features (Mapped)  Downregulated Features (Mapped)  Ref  Study  Platform  Number of Genes/Features  [121]  Habermann JK et al, 2007  Hs-UniGEM2 human cDNA microarray  9128  24 (23)  34 (29)  [122]  Lin YM et al, 2002  Custom cDNA microarray  23,040  63 (53)  375 (321)  [123]  Buckhaults P et al, 2001  SAGE  N/A  153 (106)  246 (201)  [124]  Notterman DA et al, 2001  Affymetrix Human 6500 GeneChip Set  7457  19 (19)  47 (45)  [125]  Galamb 0 et al, 2006  Human Atlas Glass 1.0 cDNA microarray  1090  83 (83)  17 (17)  [126]  Wang JY et al, 2005  TGS s-4k cDNA microarray  3800  23 (23)  0  [127]  Croner RS et al, 2005  Affymetrix HG-U133A  22,283  67 (66)  63 (62)  [128]  Kwon HY et al, 2004  Macrogen MAGIC cDNA microarray  4608  77 (77)  45 (44)  [129]  Bertucci F et al, 2004  Custom nylon cDNA microarray  8074  125 (125)  109 (109)  [130]  Ohmachi T et al, 2006  Agilent cDNA microarray  12,814  84 (82)  0  [131]  Mod D et al, 2005  Human Atlas Glass 1.0 cDNA microarray  1090  32 (32)  0  [132]  Kim H et al, 2004  Oligonucleotide microarray from Compugen/SigmaGenosys  18,861  272 (271)  216 (216)  Zou TT et al, 2002  Custom cDNA microarray  8000  88 (69)  142 (118)  Koehler A et al, 2004  Atlas Human Cancer 1.2 Array  1185  31 (29)  14 (13)  1  [133' [134]  22  Upregulated Features (Mapped)  Downregulated Features (Mapped)  Ref  Study  Platform  Number of Genes/Features  [135]  Ichikawa Y et al, 2002  Custom cDNA microarray  20,784  47 (45)  83 (78)  [136]  Jansova E et al, 2006  Human 19K microarrays (Clinical Genomic Centre)  19,201  31 (29)  163 (162)  [137]  Grade M et al, 2007  National Cancer Institute oligonucleotide arrays (Operon V2 oligo set)  21 , 543  1057 (994)  36 (36)  [138]  Bianchini M et al, 2006  Human 19K microarrays (Clinical Genomic Centre)  19,201  76 (76)  12 (12)  [139]  Agrawal D et al, 2002  Affymetrix Human 6800 GeneChip Set  7129  257 (253)  82 (78)  [140]  Sugiyama Y et al, 2005  Human Cancer Pathway Finder Gene Arrays (Superarray Bioscience)  13 (13)  11 (11)  [141]  Kitahara 0 et al, 2001  Custom cDNA microarray  9216  44 (42)  191 (163)  [142]  Williams NS, et al 2003  Custom cDNA microarray  9592  203 (192)  85 (76)  [143]  Takemasa I et al, 2001  Custom cDNA microarray  4608  22 (22)  36 (36)  3582 (3273)  2955 (2613)  Totals  96  23  Chapter 1 References 1. Shih W, Chetty R, Tsao MS: Expression profiling by microarrays in colorectal cancer (review). Oncol Rep 2005, 13:517-524. 2. Colorectal cancer stats. htte://www.cancerca/ccs/intemet/standard/0,2283,3172 14447 langld-en,00.html Accessed April 2007  3. Jemal A, Siegel R, Ward E, Murray T, Xu J, Thun MJ: Cancer statistics, 2007. CA Cancer J Clin 2007, 57:43-66. 4. Fearon ER, Vogelstein B: A genetic model for colorectal tumorigenesis. Cell 1990, 61:759-767. 5. Kinzler KW, Vogelstein B: Lessons from hereditary colorectal cancer. Cell 1996, 87:159-170. 6. Bird RP: Role of aberrant crypt foci in understanding the pathogenesis of colon cancer. Cancer Lett 1995, 93:55-71. 7. Dukes CE: The classification of cancer of the rectum. J Pathol Bacteriol 1932, 35:323-332. 8. Uicc I international union against cancer -  ratip://www.uicc.orw Accessed April 2007  9. He TC, Sparks AB, Rago C, Hermeking H, Zawel L, da Costa LT, Morin PJ, Vogelstein B, Kinzler KW: Identification of c-MYC as a target of the APC pathway. Science 1998, 281:1509-1512. 10. Tetsu 0, McCormick F: Beta-catenin regulates expression of cyclin D1 in colon carcinoma cells. Nature 1999, 398:422-426. 11. Brabletz T, Jung A, Dag S, Hlubek F, Kirchner T: Beta-catenin regulates the expression of the matrix metalloproteinase-7 in human colorectal cancer. Am J Pathol 1999, 155:1033-1038. 12. Crawford HC, Fingleton BM, Rudolph-Owen LA, Goss KJ, Rubinfeld B, Polakis P, Matrisian LM: The metalloproteinase matrilysin is a target of betacatenin transactivation in intestinal tumors. Oncogene 1999, 18:2883-2891. 13. Schneikert J, Behrens J: The canonical wnt signalling pathway and its APC partner in colon cancer development. Gut 2007, 56:417-425. 14. Lengauer C, Kinzler KW, Vogelstein B: Genetic instability in colorectal cancers. Nature 1997, 386:623-627.  24  15. Green RA, Kaplan KB: Chromosome instability in colorectal tumor cells is associated with defects in microtubule plus-end attachments caused by a dominant mutation in APC. J Cell Biol 2003, 163:949-961. 16. Green RA, Wollman R, Kaplan KB: APC and EB1 function together in mitosis to regulate spindle dynamics and chromosome alignment. Mol Biol Cell 2005, 16:4609-4622. 17. Fodde R, Smits R, Clevers H: APC, signal transduction and genetic instability in colorectal cancer. Nat Rev Cancer 2001, 1:55-67. 18. Kaplan KB, Burds AA, Swedlow JR, Bekir SS, Sorger PK, Nathke IS: A role for the adenomatous polyposis coli protein in chromosome segregation. Nat Cell Biol 2001, 3:429-432. 19. Bos JL: Ras oncogenes in human cancer: A review. Cancer Res 1989, 49:4682-4689. 20. Yamamoto H, Itoh F, Senota A, Adachi Y, Yoshimoto M, Endoh T, Hinoda Y, Yachi A, !mai K: Expression of matrix metalloproteinase matrilysin (MMP-7) was induced by activated ki-ras via AP-1 activation in SW1417 colon cancer cells. J Clin Lab Anal 1995, 9:297-301. 21. Smakman N, Borel Rinkes IH, Voest EE, Kranenburg 0: Control of colorectal metastasis formation by K-ras. Biochim Biophys Acta 2005, 1756:103-114. 22. Mehlen P, Fearon ER: Role of the dependence receptor DCC in colorectal cancer pathogenesis. J Clin Oncol 2004, 22:3420-3428. 23. Knudson AG,Jr: Hereditary cancer, oncogenes, and antioncogenes. Cancer Res 1985, 45:1437-1443. 24. Fearon ER, Cho KR, Nigro JM, Kern SE, Simons JW, Ruppert JM, Hamilton SR, Preisinger AC, Thomas G, Kinzler KW: Identification of a chromosome 18q gene that is altered in colorectal cancers. Science 1990, 247:49-56. 25. Chan SS, Zheng H, Su MW, Wilk R, Killeen MT, Hedgecock EM, Culotti JG: UNC-40, a C. elegans homolog of DCC (deleted in colorectal cancer), is required in motile cells responding to UNC-6 netrin cues. Cell 1996, 87:187195. 26. Keino-Masu K, Masu M, Hinck L, Leonardo ED, Chan SS, Culotti JG, Tessier-Lavigne M: Deleted in colorectal cancer (DCC) encodes a netrin receptor. Cell 1996, 87:175-185.  25  27. Serafini T, Colamarino SA, Leonardo ED, Wang H, Beddington R, Skarnes WC, Tessier-Lavigne M: Netrin-1 is required for commissural axon guidance in the developing vertebrate nervous system. Cell 1996, 87:1001-1014. 28. Tessier-Lavigne M, Goodman CS: The molecular biology of axon guidance. Science 1996, 274:1123-1133. 29. Manitt C, Kennedy TE: Where the rubber meets the road: Netrin expression and function in developing and adult nervous systems. Prog Brain Res 2002, 137:425-442. 30. Fazeli A, Dickinson SL, Hermiston ML, Tighe RV, Steen RG, Small CG, Stoeckli ET, Keino-Masu K, Masu M, Rayburn H, Simons J, Bronson RT, Gordon JI, Tessier-Lavigne M, Weinberg RA: Phenotype of mice lacking functional deleted in colorectal cancer (dcc) gene. Nature 1997, 386:796-804. 31. Forcet C, Ye X, Granger L, Corset V, Shin H, Bredesen DE, Mehlen P: The dependence receptor DCC (deleted in colorectal cancer) defines an alternative mechanism for caspase activation. Proc Nall Acad Sci U S A 2001, 98:3416-3421. 32. Liu J, Yao F, Wu R, Morgan M, Thorburn A, Finley RL,Jr, Chen YQ: Mediation of the DCC apoptotic signal by DIP13 alpha. J Biol Chem 2002, 277:26281-26285. 33. Mehlen P, Rabizadeh S, Snipas SJ, Assa-Munt N, Salvesen GS, Bredesen DE: The DCC gene product induces apoptosis by a mechanism requiring receptor proteolysis. Nature 1998, 395:801-804. 34. Bredesen DE, Ye X, Tasinato A, Sperandio S, Wang JJ, Assa-Munt N, Rabizadeh S: p75NTR and the concept of cellular dependence: Seeing how the other half die. Cell Death Differ 1998, 5:365-371. 35. Zhou S, Kinzler KW, Vogelstein B: Going mad with smads. N Engl J Med 1999, 341:1144-1146. 36. Massague J, Blain SW, Lo RS: TGFbeta signaling in growth control, cancer, and heritable disorders. Cell 2000, 103:295-309. 37. Howe JR, Roth S, Ringold JC, Summers RW, Jarvinen HJ, Sistonen P, Tomlinson IP, Houlston RS, Bevan S, Mitros FA, Stone EM, Aaltonen LA: Mutations in the SMAD4/DPC4 gene in juvenile polyposis. Science 1998, 280:1086-1088. 38. Takaku K, Oshima M, Miyoshi H, Matsui M, Seldin MF, Taketo MM: Intestinal tumorigenesis in compound mutant mice of both Dpc4 (Smad4) and apc genes. Cell 1998, 92:645-656.  26  39. Montgomery E, Goggins M, Zhou S, Argani P, Wilentz R, Kaushal M, Booker S, Romans K, Bhargava P, Hruban R, Kern S: Nuclear localization of Dpc4 (Madh4, Smad4) in colorectal carcinomas and relation to mismatch repair/transforming growth factor-beta receptor defects. Am J Pathol 2001, 158:537-542. 40. Korchynskyi 0, Landstrom M, Stoika R, Funa K, Heldin CH, ten Dijke P, Souchelnytskyi S: Expression of smad proteins in human colorectal cancer. Int J Cancer 1999, 82:197-202. 41. Salovaara R, Roth S, Loukola A, Launonen V, Sistonen P, Avizienyte E, Kristo P, Jarvinen H, Souchelnytskyi S, Sarlomo-Rikala M, Aaltonen LA: Frequent loss of SMAD4/DPC4 protein in colorectal cancers. Gut 2002, 51:56-59. 42. Kinzler KW, Nilbert MC, Su LK, Vogelstein B, Bryan TM, Levy DB, Smith KJ, Preisinger AC, Hedge P, McKechnie D: Identification of FAP locus genes from chromosome 5q21. Science 1991, 253:661-665. 43. de Vries A, Flores ER, Miranda B, Hsieh HM, van Oostrom CT, Sage J, Jacks T: Targeted point mutations of p53 lead to dominant-negative inhibition of wild-type p53 function. Proc Natl Aced Sci U S A 2002, 99:29482953. 44. Levine AJ: P53, the cellular gatekeeper for growth and division. Cell 1997, 88:323-331. 45. Stiewe T: The p53 family in differentiation and tumorigenesis. Nat Rev Cancer 2007, 7:165-168. 46. Greenblatt MS, Bennett WP, Hollstein M, Harris CC: Mutations in the p53 tumor suppressor gene: Clues to cancer etiology and molecular pathogenesis. Cancer Res 1994, 54:4855-4878. 47. Hollstein M, Sidransky D, Vogelstein B, Harris CC: P53 mutations in human cancers. Science 1991, 253:49-53. 48. Malkin D: p53 and the li-fraumeni syndrome. Biochim Biophys Acta 1994, 1198:197-213. 49. Vousden KH, Lu X: Live or let die: The cell's response to p53. Nat Rev Cancer 2002, 2:594-604. 50. Smith G, Carey FA, Beattie J, Wilkie MJ, Lightfoot TJ, Coxhead J, Garner RC, Steele RJ, Wolf CR: Mutations in APC, kirsten-ras, and p53--alternative genetic pathways to colorectal cancer. Proc Natl Aced Sci U S A 2002, 99:9433-9438.  27  51. Conlin A, Smith G, Carey FA, Wolf CR, Steele RJ: The prognostic significance of K-ras, p53, and APC mutations in colorectal carcinoma. Gut 2005, 54:1283-1286. 52. Galiatsatos P, Foulkes WD: Familial adenomatous polyposis. Am J Gastroenterol 2006, 101:385-398. 53. de la Chapelle A: Genetic predisposition to colorectal cancer. Nat Rev Cancer 2004, 4:769-780. 54. Strate LL, Syngal S: Hereditary colorectal cancer syndromes. Cancer Causes Control 2005, 16:201-213. 55. Galle TS, Juel K, Bulow S: Causes of death in familial adenomatous polyposis. Scand J Gastroenterol 1999, 34:808-812. 56. Giardiello FM, Brensinger JD, Petersen GM: AGA technical review on hereditary colorectal cancer and genetic testing. Gastroenterology 2001, 121:198-213. 57. Kaz AM, Brentnall TA: Genetic testing for colon cancer. Nat Clin Pract Gastroenterol Hepatol 2006, 3:670-679. 58. Winawer SJ, Fletcher RH, Miller L, Godlee F, Stolar MH, Mulrow CD, Woolf SH, Glick SN, Ganiats TG, Bond JH, Rosen L, Zapka JG, Olsen SJ, Giardiello FM, Sisk JE, Van Antwerp R, Brown-Davis C, Marciniak DA, Mayer RJ: Colorectal cancer screening: Clinical guidelines and rationale. Gastroenterology 1997, 112:594-642. 59. King JE, Dozois RR, Lindor NM, Ahlquist DA: Care of patients and their families with familial adenomatous polyposis. Mayo Clin Proc 2000, 75:5767. 60. Aaltonen LA, Salovaara R, Kristo P, Canzian F, Hemminki A, Peltomaki P, Chadwick RB, Kaariainen H, Eskelinen M, Jarvinen H, Mecklin JP, de la Chapelle A: Incidence of hereditary nonpolyposis colorectal cancer and the feasibility of molecular screening for the disease. N Engl J Med 1998, 338:1481-1487. 61. Samowitz WS, Curtin K, Lin HH, Robertson MA, Schaffer D, Nichols M, Gruenthal K, Leppert MF, Slattery ML: The colon cancer burden of genetically defined hereditary nonpolyposis colon cancer. Gastroenterology 2001, 121:830-838. 62. Burt R, Neklason DW: Genetic testing for inherited colon cancer. Gastroenterology 2005, 128:1696-1716.  28  63. Bierie B, Moses HL: Tumour microenvironment: TGFbeta: The molecular jekyll and hyde of cancer. Nat Rev Cancer 2006, 6:506-520. 64. Pardali K, Moustakas A: Actions of TGF-beta as tumor suppressor and pro-metastatic factor in human cancer. Biochim Biophys Acta 2007, 1775:2162. 65. Bacon AL, Farrington SM, Dunlop MG: Mutation frequency in coding and non-coding repeat sequences in mismatch repair deficient cells derived from normal human tissue. Oncogene 2001, 20:7464-7471. 66. Parsons R, Myeroff LL, Liu B, Willson JK, Markowitz SD, Kinzler KW, Vogelstein B: Microsatellite instability and mutations of the transforming growth factor beta type II receptor gene in colorectal cancer. Cancer Res 1995, 55:5548-5550. 67. Markowitz S, Wang J, Myeroff L, Parsons R, Sun L, Lutterbaugh J, Fan RS, Zborowska E, Kinzler KW, Vogelstein B: Inactivation of the type II TGF-beta receptor in colon cancer cells with microsatellite instability. Science 1995, 268:1336-1338. 68. Lynch HT, Watson P, Lanspa SJ, Marcus J, Smyrk T, Fitzgibbons RJ,Jr, Kriegler M, Lynch JF: Natural history of colorectal cancer in hereditary nonpolyposis colorectal cancer (lynch syndromes I and II). Dis Colon Rectum 1988, 31:439-444. 69. Vasen HF, Mecklin JP, Watson P, Utsunomiya J, Bertario L, Lynch P, Svendsen LB, Cristofaro G, Muller H, Khan PM: Surveillance in hereditary nonpolyposis colorectal cancer: An international cooperative study of 165 families. the international collaborative group on HNPCC. Dis Colon Rectum 1993, 36:1-4. 70. Aarnio M, Sankila R, Pukkala E, Salovaara R, Aaltonen LA, de la Chapelle A, Peltomaki P, Mecklin JP, Jarvinen HJ: Cancer risk in mutation carriers of DNA-mismatch-repair genes. Int J Cancer 1999, 81:214-218. 71. Grady WM: Genetic testing for high-risk colon cancer patients. Gastroenterology 2003, 124:1574-1594. 72. Burke W, Petersen G, Lynch P, Botkin J, Daly M, Garber J, Kahn MJ, McTiernan A, Offit K, Thomson E, Varricchio C: Recommendations for followup care of individuals with an inherited predisposition to cancer. I. hereditary nonpolyposis colon cancer. cancer genetics studies consortium. JAMA 1997, 277:915-919. 73. Jarvinen HJ, Aarnio M, Mustonen H, Aktan-Collan K, Aaltonen LA, Peltomaki P, De La Chapelle A, Mecklin JP: Controlled 15-year trial on screening for  29  colorectal cancer in families with hereditary nonpolyposis colorectal cancer. Gastroenterology 2000, 118:829-834.  74. Haenszel W, Kurihara M: Studies of japanese migrants. I. mortality from cancer and other diseases among japanese in the united states. J Natl Cancer lnst 1968, 40:43-68. 75. McMichael AJ, Giles GG: Cancer in migrants to australia: Extending the descriptive epidemiological data. Cancer Res 1988, 48:751-756. 76. Bingham SA, Day NE, Luben R, Ferrari P, Slimani N, Norat T, ClavelChapelon F, Kesse E, Nieters A, Boeing H, Tjonneland A, Overvad K, Martinez C, Dorronsoro M, Gonzalez CA, Key TJ, Trichopoulou A, Naska A, Vineis P, Tumino R, Krogh V, Bueno-de-Mesquita HB, Peeters PH, Berglund G, Hallmans G, Lund E, Skeie G, Kaaks R, Riboli E, European Prospective Investigation into Cancer and Nutrition: Dietary fibre in food and protection against colorectal cancer in the european prospective investigation into cancer and nutrition (EPIC): An observational study. Lancet 2003, 361:1496-1501. 77. Doll R, Peto R: The causes of cancer: Quantitative estimates of avoidable risks of cancer in the united states today. J Natl Cancer lnst 1981, 66:1191-1308. 78. Jain M, Cook GM, Davis FG, Grace MG, Howe GR, Miller AB: A casecontrol study of diet and colo-rectal cancer. Int J Cancer 1980, 26:757-768. 79. Potter JD, McMichael AJ: Diet and cancer of the colon and rectum: A case-control study. J Natl Cancer lnst 1986, 76:557-569. 80. Lyon JL, Mahoney AW, West DW, Gardner JW, Smith KR, Sorenson AW, Stanish W: Energy intake: Its relationship to colon cancer risk. J Natl Cancer lnst 1987, 78:853-861. 81. Whittemore AS, Wu-Williams AH, Lee M, Zheng S, Gallagher RP, Jiao DA, Zhou L, Wang XH, Chen K, Jung D: Diet, physical activity, and colorectal cancer among chinese in north america and china. J Natl Cancer Inst 1990, 82:915-926. 82. Miller AB, Howe GR, Jain M, Craib KJ, Harrison L: Food items and food groups as risk factors in a case-control study of diet and colo-rectal cancer. Int J Cancer 1983, 32:155-161. 83. Gerhardsson de Verdier M, Hagman U, Peters RK, Steineck G, Overvik E: Meat, cooking methods and colorectal cancer: A case-referent study in stockholm. Int J Cancer 1991, 49:520-525. 84. Goldbohm RA, van den Brandt PA, van 't Veer P, Brants HA, Dorant E, Sturmans F, Hermus RJ: A prospective cohort study on the relation between 30  meat consumption and the risk of colon cancer. Cancer Res 1994, 54:718723.  85. English DR, Maclnnis RJ, Hodge AM, Hopper JL, Haydon AM, Giles GG: Red meat, chicken, and fish consumption and risk of colorectal cancer. Cancer Epidemiol Biomarkers Prey 2004, 13:1509-1514. 86. Chao A, Thun MJ, Connell CJ, McCullough ML, Jacobs EJ, Flanders WD, Rodriguez C, Sinha R, Calle EE: Meat consumption and risk of colorectal cancer. JAMA 2005, 293:172-182. 87. Larsson SC, Wolk A: Meat consumption and risk of colorectal cancer: A meta-analysis of prospective studies. Int J Cancer 2006, 119:2657-2664. 88. Jacobs ET, Thompson PA, Martinez ME: Diet, gender, and colorectal neoplasia. J Clin Gastroenterol 2007, 41:731-746. 89. Cardoso J, Boer J, Morreau H, Fodde R: Expression and genomic profiling of colorectal cancer. Biochim Biophys Acta 2007, 1775:103-137. 90. Rifai N, Gillette MA, Carr SA: Protein biomarker discovery and validation: The long and uncertain path to clinical utility. Nat Biotechnol 2006, 24:971983. 91. Dalton WS, Friend SH: Cancer biomarkers--an invitation to the table. Science 2006, 312:1165-1168. 92. Hartwell L, Mankoff D, Paulovich A, Ramsey S, Swisher E: Cancer biomarkers: A systems approach. Nat Biotechnol 2006, 24:905-908. 93. Ford D, Easton DF, Stratton M, Narod S, Goldgar D, Devilee P, Bishop DT, Weber B, Lenoir G, Chang-Claude J, Sobol H, Teare MD, Struewing J, Arason A, Scherneck S, Peto J, Rebbeck TR, Tonin P, Neuhausen S, Barkardottir R, Eyfjord J, Lynch H, Ponder BA, Gayther SA, Zelada-Hedman M: Genetic heterogeneity and penetrance analysis of the BRCA1 and BRCA2 genes in breast cancer families. the breast cancer linkage consortium. Am J Hum Genet 1998, 62:676-689. 94. Catalona WJ, Richie JP, deKernion JB, Ahmann FR, Ratliff TL, Dalkin BL, Kavoussi LR, MacFarlane MT, Southwick PC: Comparison of prostate specific antigen concentration versus prostate specific antigen density in the early detection of prostate cancer: Receiver operating characteristic curves. J Urol 1994, 152:2031-2036. 95. Leach FS, Koh MS, Chan YW, Bark S, Ray R, Morton RA, Remaley AT: Prostate specific antigen as a clinical biomarker for prostate cancer: What's the take home message? Cancer Biol Ther 2005, 4:371-375.  31  96. Brawer MK, Chetner MP, Beatie J, Buchner DM, Vessella RL, Lange PH: Screening for prostatic carcinoma with prostate specific antigen. J Urol 1992, 147:841-845. 97. Paik S, Shak S, Tang G, Kim C, Baker J, Cronin M, Baehner FL, Walker MG, Watson D, Park T, Hiller W, Fisher ER, Wickerham DL, Bryant J, Wolmark N: A multigene assay to predict recurrence of tamoxifen-treated, node-negative breast cancer. N Engl J Med 2004, 351:2817-2826. 98. Booth RA: Minimally invasive biomarkers for detection and staging of colorectal cancer. Cancer Lett 2007, 249:87-96. 99. Mandel JS, Bond JH, Church TR, Snover DC, Bradley GM, Schuman LM, Ederer F: Reducing mortality from colorectal cancer by screening for fecal occult blood. minnesota colon cancer control study. N Engl J Med 1993, 328:1365-1371. 100. Mandel JS, Church TR, Ederer F, Bond JH: Colorectal cancer mortality: Effectiveness of biennial screening for fecal occult blood. J Natl Cancer lnst 1999, 91:434-437. 101. Hardcastle JD, Chamberlain JO, Robinson MH, Moss SM, Amar SS, Balfour TW, James PD, Mangham CM: Randomised controlled trial of faecal-occultblood screening for colorectal cancer. Lancet 1996, 348:1472-1477. 102. Walsh JM, Terdiman JP: Colorectal cancer screening: Scientific review. JAMA 2003, 289:1288-1296. 103. Saltz LB, Kelsen DP: Adjuvant treatment of colorectal cancer. Annu Rev Med 1997, 48:191-202. 104. Moertel CG, Fleming TR, Macdonald JS, Haller DG, Laurie JA, Tangen CM, Ungerleider JS, Emerson WA, Tormey DC, Glick JH, Veeder MH, Mailliard JA: Fluorouracil plus levamisole as effective adjuvant therapy after resection of stage Ill colon carcinoma: A final report. Ann Intern Med 1995, 122:321-326. 105. Wolmark N, Rockette H, Mamounas E, Jones J, Wieand S, Wickerham DL, Bear HD, Atkins JN, Dimitrov NV, Glass AG, Fisher ER, Fisher B: Clinical trial to assess the relative efficacy of fluorouracil and leucovorin, fluorouracil and levamisole, and fluorouracil, leucovorin, and levamisole in patients with dukes' B and C carcinoma of the colon: Results from national surgical adjuvant breast and bowel project C-04. J Clin Oncol 1999, 17:3553-3559. 106. Arango D, Laiho P, Kokko A, Alhopuro P, Sammalkorpi H, Salovaara R, Nicorici D, Hautaniemi S, Alazzouzi H, Mecklin JP, Jarvinen H, Hemminki A, Astola J, Schwartz S,Jr, Aaltonen LA: Gene-expression profiling predicts recurrence in dukes' C colorectal cancer. Gastroenterology 2005, 129:874884. 32  107. Wang Y, Jatkoe T, Zhang Y, Mutch MG, Talantov D, Jiang J, McLeod HL, Atkins D: Gene expression profiles and molecular markers to predict recurrence of dukes' B colon cancer. J Clin Oncol 2004, 22:1564-1571. 108. Cowell JK, Hawthorn L: The application of microarray technology to the analysis of the cancer genome. Curr Mol Med 2007, 7:103-120. 109. Velculescu VE, Zhang L, Vogelstein B, Kinzler KW: Serial analysis of gene expression. Science 1995, 270:484-487. 110. Sagynaliev E, Steinert R, Nestler G, Lippert H, Knoch M, Reymond MA: Web-based data warehouse on gene expression in human colorectal cancer. Proteomics 2005, 5:3066-3078. 111. Kuo WP, Jenssen TK, Butte AJ, Ohno-Machado L, Kohane IS: Analysis of matched mRNA measurements from two different microarray technologies. Bioinformatics 2002, 18:405-412. 112. Siddiqui AS, Delaney AD, Schnerch A, Griffith OL, Jones SJ, Marra MA: Sequence biases in large scale gene expression profiling data. Nucleic Acids Res 2006, 34:e83. 113. Rhodes DR, Yu J, Shanker K, Deshpande N, Varambally R, Ghosh D, Barrette T, Pandey A, Chinnaiyan AM: Large-scale meta-analysis of cancer microarray data identifies common transcriptional profiles of neoplastic transformation and progression. Proc Natl Acad Sci USA 2004, 101:93099314. 114. Ghosh D, Barette TR, Rhodes D, Chinnaiyan AM: Statistical issues and methods for meta-analysis of microarray data: A case study in prostate cancer. Funct lntegr Genomics 2003, 3:180-188. 115. Normand SL: Meta-analysis: Formulating, evaluating, combining, and reporting. Stat Med 1999, 18:321-359. 116. Rhodes DR, Barrette TR, Rubin MA, Ghosh D, Chinnaiyan AM: Metaanalysis of microarrays: Interstudy validation of gene expression profiles reveals pathway dysregulation in prostate cancer. Cancer Res 2002, 62:4427-4433. 117. Rhodes DR, Chinnaiyan AM: Integrative analysis of the cancer transcriptome. Nat Genet 2005, 37 Suppl:S31-7. 118. Jiang H, Deng Y, Chen HS, Tao L, Sha Q, Chen J, Tsai CJ, Zhang S: Joint analysis of two microarray gene-expression data sets to select lung adenocarcinoma marker genes. BMC Bioinformatics 2004, 5:81.  33  119. Yang X, Sun X: Meta-analysis of several gene lists for distinct types of cancer: A simple way to reveal common prognostic markers. BMC Bioinformatics 2007, 8:118. 120. Griffith OL, Melck A, Jones SJ, Wiseman SM: Meta-analysis and metareview of thyroid cancer gene expression profiling studies identifies important diagnostic biomarkers. J Clin Oncol 2006, 24:5043-5051. 121. Habermann JK, Paulsen U, Roblick UJ, Upender MB, McShane LM, Korn EL, Wangsa D, Kruger S, Duchrow M, Bruch HP, Auer G, Ried T: Stage-specific alterations of the genome, transcriptome, and proteome during colorectal carcinogenesis. Genes Chromosomes Cancer 2007, 46:10-26. 122. Lin YM, Furukawa Y, Tsunoda T, Yue CT, Yang KC, Nakamura Y: Molecular diagnosis of colorectal tumors by expression profiles of 50 genes expressed differentially in adenomas and carcinomas. Oncogene 2002, 21:4120-4128. 123. Buckhaults P, Rago C, St Croix B, Romans KE, Saha S, Zhang L, Vogelstein B, Kinzler KW: Secreted and cell surface genes expressed in benign and malignant colorectal tumors. Cancer Res 2001, 61:6996-7001. 124. Notterman DA, Alon U, Sierk AJ, Levine AJ: Transcriptional gene expression profiles of colorectal adenoma, adenocarcinoma, and normal tissue examined by oligonucleotide arrays. Cancer Res 2001, 61:3124-3130. 125. Galamb 0, Sipos F, Dinya E, Spisak S, Tulassay Z, Molnar B: mRNA expression, functional profiling and multivariate classification of colon biopsy specimen by cDNA overall glass microarray. World J Gastroenterol 2006, 12:6998-7006. 126. Wang JY, Yeh CS, Tzou WS, Hsieh JS, Chen FM, Lu CY, Yu FJ, Cheng TL, Huang TJ, Lin SR: Analysis of progressively overexpressed genes in tumorigenesis of colorectal cancers using cDNA microarray. Oncol Rep 2005, 14:65-72. 127. Croner RS, Foertsch T, Brueckl WM, Guenther K, Siebenhaar R, Stremmel C, Matzel KE, Papadopoulos T, Kirchner T, Behrens J, Klein-Hitpass L, Stuerzl M, Hohenberger W, Reingruber B: Common denominator genes that distinguish colorectal carcinoma from normal mucosa. Int J Colorectal Dis 2005, 20:353-362. 128. Kwon HC, Kim SH, Roh MS, Kim JS, Lee HS, Choi HJ, Jeong JS, Kim HJ, Hwang TH: Gene expression profiling in lymph node-positive and lymph node-negative colorectal cancer. Dis Colon Rectum 2004, 47:141-152. 129. Bertucci F, Salas S, Eysteries S, Nasser V, Finetti P, Ginestier C, CharafeJauffret E, Loriod B, Bachelart L, Montfort J, Victorero G, Viret F, 011endorff V, 34  Fert V, Giovaninni M, Delpero JR, Nguyen C, Viens P, Monges G, Birnbaum D, Houlgatte R: Gene expression profiling of colon cancer by DNA microarrays and correlation with histoclinical parameters. Oncogene 2004, 23:1377-1391. 130. Ohmachi T, Tanaka F, Mimori K, Inoue H, Yanaga K, Mod M: Clinical significance of TROP2 expression in colorectal cancer. Clin Cancer Res 2006, 12:3057-3063. 131. Mod D, Nakafusa Y, Miyazaki K, Tokunaga 0: Differential expression of janus kinase 3 (JAK3), matrix metalloproteinase 13 (MMP13), heat shock protein 60 (HSP60), and mouse double minute 2 (MDM2) in human colorectal cancer progression using human cancer cDNA microarrays. Pathol Res Pract 2005, 201:777-789. 132. Kim H, Nam SW, Rhee H, Shan Li L, Ju Kang H, Hye Koh K, Kyu Kim N, Song J, Tak-Bun Liu E, Kim H: Different gene expression profiles between microsatellite instability-high and microsatellite stable colorectal carcinomas. Oncogene 2004, 23:6218-6225. 133. Zou TT, Selaru FM, Xu Y, Shustova V, Yin J, Mod Y, Shibata D, Sato F, Wang S, Olaru A, Deacu E, Liu TC, Abraham JM, Meltzer SJ: Application of cDNA microarrays to generate a molecular taxonomy capable of distinguishing between colon cancer and normal colon. Oncogene 2002, 21:4855-4862. 134. Koehler A, Bataille F, Schmid C, Ruemmele P, Waldeck A, Blaszyk H, Hartmann A, Hofstaedter F, Dietmaier W: Gene expression profiling of colorectal cancer and metastases divides tumours according to their clinicopathological stage. J Pathol 2004, 204:65-74. 135. Ichikawa Y, Ishikawa T, Takahashi S, Hamaguchi Y, Morita T, Nishizuka I, Yamaguchi S, Endo I, Ike H, Togo S, Oki S, Shimada H, Kadota K, Nakamura S, Goto H, Nitanda H, Satomi S, Sakai T, Narita I, Gejyo F, Tomaru Y, Shimizu K, Hayashizaki Y, Okazaki Y: Identification of genes regulating colorectal carcinogenesis by using the algorithm for diagnosing malignant state method. Biochem Biophys Res Commun 2002, 296:497-506. 136. Jansova E, Koutna I, Krontorad P, Svoboda Z, Krivankova S, Zaloudik J, Kozubek M, Kozubek S: Comparative transcriptome maps: A new approach to the diagnosis of colorectal carcinoma patients using cDNA microarrays. Clin Genet 2006, 69:218-227. 137. Grade M, Hormann P, Becker S, Hummon AB, Wangsa D, Varma S, Simon R, Liersch T, Becker H, Difilippantonio MJ, Ghadimi BM, Ried T: Gene expression profiling reveals a massive, aneuploidy-dependent transcriptional deregulation and distinct differences between lymph nodenegative and lymph node-positive colon carcinomas. Cancer Res 2007, 67:41-56. 35  138. Bianchini M, Levy E, Zucchini C, Pinski V, Macagno C, De Sanctis P, Valvassori L, Carinci P, Mordoh J: Comparative study of gene expression by cDNA microarray in human colorectal cancer tissues and normal mucosa. Int J Oncol 2006, 29:83-94. 139. Agrawal D, Chen T, Irby R, Quackenbush J, Chambers AF, Szabo M, Cantor A, Coppola D, Yeatman TJ: Osteopontin identified as lead marker of colon cancer progression, using pooled sample expression profiling. J Natl Cancer lnst 2002, 94:513-521. 140. Sugiyama Y, Farrow B, Murillo C, Li J, Watanabe H, Sugiyama K, Evers BM: Analysis of differential gene expression patterns in colon cancer and cancer stroma using microdissected tissues. Gastroenterology 2005, 128:480-486. 141. Kitahara 0, Furukawa Y, Tanaka T, Kihara C, Ono K, Yanagawa R, Nita ME, Takagi T, Nakamura Y, Tsunoda T: Alterations of gene expression during colorectal carcinogenesis revealed by cDNA microarrays after lasercapture microdissection of tumor tissues and normal epithelia. Cancer Res 2001, 61:3544-3549. 142. Williams NS, Gaynor RB, Scoggin S, Verma U, Gokaslan T, Simmang C, Fleming J, Tavana D, Frenkel E, Becerra C: Identification and validation of genes involved in the pathogenesis of colorectal cancer using cDNA microarrays and RNA interference. Clin Cancer Res 2003, 9:931-946. 143. Takemasa I, Higuchi H, Yamamoto H, Sekimoto M, Tomita N, Nakamori S, Matoba R, Monden M, Matsubara K: Construction of preferential cDNA microarray specialized for human colorectal carcinoma: Molecular sketch of colorectal cancer. Biochem Biophys Res Commun 2001, 285:1244-1249.  36  Chapter 2: Materials and Methods 2.1 Meta-analysis of Published Datasets of Differentially Expressed Genes in Colorectal Cancer The meta-analysis method that was created for this thesis was based on a method that was previously developed by our group [1]. This previous method utilized a vote-counting strategy in which a gene was ranked based on three criteria. Presented in order of importance, the three criteria were: 1) the number of studies reporting its differential expression, 2) the total number of tissue samples utilized in the studies, and 3) the average fold change. Originally, the number of tissue samples utilized was deemed more important than the average fold change because many studies do not report a fold change. However, one limitation of this method was that certain genes, while showing a large fold change, were ranked lower due to fewer tissue samples included in the studies. To overcome this, we improved the meta-analysis method by dividing genes into semi-quantitative categories based on the number of tissue samples instead of the absolute numbers. Each gene that was reported in an equal number of independent studies was divided into three semi-quantitative categories based on the number of tissue samples: lowest (first quartile, Q1), moderate (interquartile range), and greatest (values greater than those in third quartile, Q3). This enabled the current version of the meta-analysis method to further emphasize the genes that show the greatest fold changes, which is likely of greater interest.  37  2.1.1 Data Collection and Curation  NCBI PubMed was queried for high throughput colorectal cancer expression profiling studies between 2000 and 2007. Only studies utilizing human tissue samples obtained from the surgical resection of cancerous tumors and/or adenomatous polyps were considered. Studies were divided into three comparison types: cancer versus normal, adenoma versus normal, and cancer versus adenoma. Studies that focused on determining differentially expressed genes between tissues of varying microsatellite stability, specific stages of colorectal cancer, or those comparing cancer to cancer samples to determine prognostic biomarkers were excluded. In total, differentially expressed genes from 25 independent studies were collected. Twenty-three studies performed expression profiling to compare cancer versus normal samples (Table 1.1, page 22), while seven and five studies considered adenoma versus normal (Table 2.1, page 41), and cancer versus adenoma (Table 2.2, page 42), respectively.  2.1.2 Differentially Expressed Genes Mapped to Entrez Gene IDs  In the microarray expression profiling studies, differentially expressed genes were represented by a GenBank accession ID [2], HUGO gene name [3], or Affymetrix probe id [4]. The sequence identifier was mapped to the NCBI Entrez Gene Identifier (Entrez Gene ID) [5] with the aid of custom developed Perl scripts and the Clone/Gene ID Converter tool [6]. For the SAGE study, updated tag to gene mapping data were obtained from SAGE Genie [7].  38  2.1.3 Total Gene Lists for Each Study  In order to estimate the amount of overlap occurring between studies by chance, we obtained the platform-specific annotation file to identify the genes that could potentially be detected as differentially expressed. For commercial platforms, such as Affymetrix and Atlas microarrays, the annotation file was obtained directly from the company website. The identifiers in these annotation files were mapped to the corresponding Entrez Gene ID as above to produce a total gene list for each study. Identifiers that could not be mapped to an Entrez Gene ID were ignored. To obtain a total gene list for the SAGE study, all gene names in the tag to gene mapping data from SAGE Genie were mapped to Entrez Gene IDs. For studies that utilized platforms in which an annotation file could not be obtained, such as the custom cDNA microarrays and some of the oligonucletotide microarrays, an approximation approach was used in which the appropriate number of Entrez Gene IDs was randomly chosen from the combined gene lists from the other platforms. For example, if a study reported 200 differentially expressed genes, then 200 Entrez Gene IDs were randomly chosen from the combined gene list.  2.1.4 Assessment of Significance of Study Overlap using Simulations  To assess the statistical significance of the observed overlap, custom Perl (version 5.8.3) scripts were created to perform Monte Carlo simulations. In each of the 10,000 permutations, the appropriate number of Entrez Gene IDs from the total gene list of each study was randomly chosen. Next, the appropriate number  39  of genes were labeled as 'UP' for up-regulated or 'DOWN' for down-regulated. For example, if a study reported 50 up-regulated and 20 down-regulated genes, 70 Entrez Gene IDs were randomly chosen and 50 would be labeled as 'UP' and 20 would be labeled as 'DOWN.' We utilized an 'all-or-none' approach, in which the level of overlap for a particular gene was only considered if all the independent studies reporting its differential expression agreed on the direction of differential expression. The level of overlap between studies in each permutation was counted as in the real analysis. Upon completion of the permutations, a distribution of overlap results from the simulations was determined and a P-value estimated by comparing the number of instances of a certain level of overlap from the simulations to the actual level of overlap in the real data. Significance was defined at P < 0.05. A pipeline summarizing the steps undertaken is shown in Figure 2.1. a)^b) Processing and Parsing  Obtain platform amotation file  Map IDs to Entrez Gene ID  Determine Overlap and significance  Perform Monte Carlo simulations  Figure 2.1: A pipeline summarizing the steps undertaken. a) PubMed was queried to obtain published studies performing transcript based expression profiling to compare cancer versus normal, adenoma versus normal, and cancer versus adenoma. b) Custom Perl scripts were developed to process and parse the lists of differentially expressed genes. c) The platform specific annotation file for each study was obtained. d) The sequence IDs for the list of differentially expressed genes and the platform specific annotation file were mapped to Entrez Gene IDs. e) Monte Carlo simulations were conducted using custom Perl scripts. f) The observed overlap was compared to that of the simulations and statistical significance was assessed.  40  Table 2.1: Seven Colorectal Adenoma versus Normal Tissue Expression Profiling Studies Included in Analysis Ref  Study  Platform  Number of Genes/ Features  Upregulated Genes/ Features (Mapped)  Downregulated Genes/ Features (Mapped)  [8]  Habermann JK et al, 2007  9128  20 (19)  38 (35)  [9]  Lin YM et al, 2002 Buckhaults P et al, 2001 Notterman DA et al, 2001  Hs-UniGEM2 human cDNA microarray Custom cDNA microarray SAGE  23,040  63 (53)  375 (321)  N/A  247 (208)  246 (180)  7129  20 (20)  0  1090  12 (12)  33 (33)  3800  23 (23)  0  588  15 (11)  9 (5)  400 (346)  701 (640)  [10] [11]  [12]  Galamb 0 et al, 2006  [13]  Wang JY et al, 2005 Lechner S et al, 2003  [14]  Totals  Affymetrix Human 6800 GeneChip Set Human Atlas Glass 1.0 cDNA microarray TGS s-4k cDNA microarray Atlas Human Cancer cDNA microarray  41  Table 2.2: Five Cancer versus Adenoma Tissue Expression Profiling Studies Included in Analysis Ref  Study  [8]  Habermann JK et Hs-UniGEM2 human cDNA al, 2007 microarray Lin YM et al, 2002 Custom cDNA microarray Buckhaults P et al, SAGE 2001 Notterman DA et Affymetrix al, 2001 Human 6500 GeneChip Set Gene Navigator Nosho K et al, 2005 cDNA array filter (Toyobo)  [9] [10] [11]  [15]  Totals  Platform  Number of Genes/ Features  Up-regulated Genes/ Features (Mapped)  Downregulated Genes/ Features (Mapped)  9128  80 (80)  36 (32)  23,040  18 (18)  32 (31)  N/A  163 (75)  167 (137)  7457  20 (20)  0  561  10 (10)  12 (12)  291 (203)  247 (212)  42  Chapter 2 References 1. Griffith OL, Melck A, Jones SJ, Wiseman SM: Meta-analysis and metareview of thyroid cancer gene expression profiling studies identifies important diagnostic biomarkers. J Clin Oncol 2006, 24:5043-5051. 2. Benson DA, Karsch-Mizrachi I, Lipman DJ, Ostell J, Wheeler DL: GenBank. Nucleic Acids Res 2007, 35:D21-5. 3. Eyre TA, Ducluzeau F, Sneddon TP, Povey S, Bruford EA, Lush MJ: The HUGO gene nomenclature database, 2006 updates. Nucleic Acids Res 2006, 34:D319-21. 4. Dalma-Weiszhausz DD, Warrington J, Tanimoto EY, Miyada CG: The affymetrix GeneChip platform: An overview. Methods Enzymol 2006, 410:328. 5. Maglott D, Ostell J, Pruitt KD, Tatusova T: Entrez gene: Gene-centered information at NCBI. Nucleic Acids Res 2007, 35:D26-31. 6. Alibes A, Yankilevich P, Canada A, Diaz-Uriarte R: IDconverter and IDClight: Conversion and annotation of gene and protein IDs. BMC Bioinformatics 2007, 8:9. 7. Boon K, Osorio EC, Greenhut SF, Schaefer CF, Shoemaker J, Polyak K, Morin PJ, Buetow KH, Strausberg RL, De Souza SJ, Riggins GJ: An anatomy of normal and malignant gene expression. Proc Natl Acad Sci U S A 2002, 99:11287-11292. 8. Habermann JK, Paulsen U, Roblick UJ, Upender MB, McShane LM, Korn EL, Wangsa D, Kruger S, Duchrow M, Bruch HP, Auer G, Ried T: Stage-specific alterations of the genome, transcriptome, and proteome during colorectal carcinogenesis. Genes Chromosomes Cancer 2007, 46:10-26. 9. Lin YM, Furukawa Y, Tsunoda T, Yue CT, Yang KC, Nakamura Y: Molecular diagnosis of colorectal tumors by expression profiles of 50 genes expressed differentially in adenomas and carcinomas. Oncogene 2002, 21:4120-4128. 10. Buckhaults P, Rago C, St Croix B, Romans KE, Saha S, Zhang L, Vogelstein B, Kinzler KW: Secreted and cell surface genes expressed in benign and malignant colorectal tumors. Cancer Res 2001, 61:6996-7001. 11. Notterman DA, Alon U, Sierk AJ, Levine AJ: Transcriptional gene expression profiles of colorectal adenoma, adenocarcinoma, and normal tissue examined by oligonucleotide arrays. Cancer Res 2001, 61:3124-3130.  43  12. Galamb 0, Sipos F, Dinya E, Spisak S, Tulassay Z, Molnar B: mRNA expression, functional profiling and multivariate classification of colon biopsy specimen by cDNA overall glass microarray. World J Gastroenterol 2006, 12:6998-7006. 13. Wang JY, Yeh CS, Tzou WS, Hsieh JS, Chen FM, Lu CY, Yu FJ, Cheng TL, Huang TJ, Lin SR: Analysis of progressively overexpressed genes in tumorigenesis of colorectal cancers using cDNA microarray. Oncol Rep 2005, 14:65-72. 14. Lechner S, Muller-Ladner U, Renke B, Scholmerich J, Ruschoff J, Kullmann F: Gene expression pattern of laser microdissected colonic crypts of adenomas with low grade dysplasia. Gut 2003, 52:1148-1153. 15. Nosho K, Yamamoto H, Adachi Y, Endo T, Hinoda Y, lmai K: Gene expression profiling of colorectal adenomas and early invasive carcinomas by cDNA array analysis. Br J Cancer 2005, 92:1193-1200.  44  Chapter 3: Results 3.1 Mapping Success Rate and Significance of Overlap Observed Of the total 8176 differentially expressed genes reported in the 25 studies (4273 up- and 3903 down-regulated), 7287 (89.1%) could be mapped to an Entrez Gene ID (3822 up- and 3465 down-regulated). In the cancer versus normal and adenoma versus normal comparisons, significant overlap was observed. No such significance in overlap was observed in the cancer versus adenoma comparison (Table 3.1, page 49), even though each individual study identified differentially expressed genes.  3.2 Overlap Results for Cancer versus Normal Tissue Comparisons The simulations demonstrated that the amount of overlap in the cancer versus normal tissue comparison was highly significant (P < .0001), with 573 genes reported as differentially expressed with consistent direction of change in at last two studies (multi-study genes) (Figure 3.1). There were 175 multi-study genes that were reported with inconsistent direction of differential expression. Thus, the majority of multi-study genes (76.6%) that were reported as differentially expressed agreed on the direction, even for large numbers of studies.  45  From the cancer versus normal Monte Carlo simulations, an average of 258.30 (95% CI, 258.16 to 258.45) genes would be expected to have an overlap of two, while the actual data contained 410. An average of 18.37 (95% CI, 18.33 to 18.42) genes would be expected to have an overlap of three, compared with 95 in the actual data. For an overlap of four, the simulation produced 1.14 (95% CI, 1.13 to 1.15) genes, while the actual data contained 30 genes. Overlaps of five, six, and seven were observed in the simulations, but with averages of less than one hundredth of a gene. In 10,000 permutations, the simulations never produced an overlap greater than seven, while two genes had an overlap of nine and one gene had an overlap of 11 in the real data. In total, 38 genes were reported in at least five cancer versus normal studies, of which 22 were upregulated (Table 3.2, page 50). and the remaining 16 were down-regulated (Table 3.3, page 54). Also, 125 genes were reported in three or four cancer versus normal studies, of which 77 were up-regulated (Table 3.4, page 57) and the remaining 48 genes were down-regulated (Table 3.5, page 64).  46  450  4^ 10  400  Cl) 0 350  cp 300  ir,  250  3  100  200 t -CI 150 Z  50 0  258.3  95 30 20 [1L7^  Firim  10^5  ri  2^3^4^5^6^7  2^1  ^ ^ 9 11  Number of studies Figure 3.1: Overlap analysis results for the cancer versus normal comparison. The actual overlap values are represented by white bars, while the average overlap values from the Monte Carlo simulations are represented by black bars. The 95% confidence intervals of the simulations were not included as they were too small to visualize. Note that the average overlap values in the simulations never equaled or exceeded that which was observed in the actual data.  3.3 Overlap Results for Adenoma versus Normal Tissue Comparisons In the adenoma versus normal tissue comparison, there were 39 multi-study genes, of which 23 were up-regulated (Table 3.6, page 68) and the remaining 16 were down-regulated (Table 3.7 on page 70). In the simulations, an average of 10.64 (95% CI, 10.61 to 10.68) genes was observed with an overlap of two, while the actual data contained 37. For an overlap of three, an average of 0.07 (95% CI, 0.067 to 0.073) of a gene was observed in the simulations, while two genes were observed with an overlap of three in the real data.  47  3.4 Overlap Results for Cancer versus Adenoma Tissue Comparisons Finally, in the cancer versus adenoma tissue comparison, there were five multistudy genes; all five were reported in only two independent studies. In the 10,000 permutations, there were 775 instances in which five genes were observed with an overlap of two or more, which yields a non-significant P-value of 0.08 (results not shown).  3.5 Overlapping Results in the Cancer versus Normal and Adenoma versus Normal Comparisons The cancer versus normal and adenoma versus normal comparisons produced differentially expressed genes reported at a statistically significant frequency. Thus, we determined the overlap between these two comparisons. The 26 genes are presented in Table 3.8, page 72.  48  Table 3.1: Summary of Comparisons Studied for Overlap Significance Note the overlap observed in the cancer versus adenoma comparison was not significant (P < .05).  Comparison  Cancer versus Normal Adenoma versus Normal Cancer versus Adenoma  Total Number of Studies  Total Number of Differentially Expressed Genes Reported (Mapped)  Total Number of Differentially Expressed Genes Reported in ? 2 Studies  P-value  23  6537 (5886)  573  < .0001  7  1101 (986)  39  < .0001  5  538 (415)  5  .08  49  Table 3.2: Up-regulated Genes Most Commonly Reported in Cancer versus Normal Expression Profiling Studies The 22 up-regulated genes reported in at least five independent studies with consistent direction are presented here. Genes reported by five and six studies were further subdivided into semi-quantitative categories based on the lowest (Q1), moderate (interquartile range), and greatest (values greater than those in Q3) number of tissue samples to give greater importance to the average fold-change criteria for ranking genes when total sample numbers were similar. Validation studies that report a gene as differentially expressed in the opposite direction from that of the meta-analysis are marked with an "" Abbreviations: FC, fold change; RT-PCR, reverse transcriptase polymerase chain reaction; IHC, immunohistochemistry; NB, northern blot; WB, western blot. Gene Name  Description  Studies (with FC)  Total Sample Sizes (with  FC)  TGFr31  IFITMI  MYC  SPARC  GDF15  Mean Fold Change  Range  Validation  Transforming growth factor, beta-induced, 68kDa  9 [1-8] (8)  369 (329)  8.94  1.11 to 32.00  RT-PCR [1, 2, 6, 8, 9]  Interferon induced transmembrane protein 1 (9-27)  9 [2, 3, 6, 8-13] (4)  351 (187)  7.52  3.00 to 12.00  RT-PCR [8, 9]  329 (243)  5.02  1.69 to 7.50  RT-PCR [6, 9, 13, 16]  716' 9 1)3, ] (6  244 (180)  6.30  1.27 to 15.00  IHC [17]*  7 [1, 2, 5, 7, 13, 18] (5)  230 (172)  7.42  1.58 to 12.20  RT-PCR [2, 9]  V-myc myelocytomatosis viral oncogene homolog (avian) Secreted protein, acidic, cysteine-rich (osteonectin) Growth differentiation factor 15  7 [3, 6, 7, 9, 13-15] (4)  7 [2, 3, 5,  50  Gene Name  Description  Studies (with FC)  Chemokine (C-X-C motif) ligand 1 (melanoma growth stimulating activity, alpha)  6 [3, 6, 7, 13, 18, 19] (4)  Cell division cycle 25 homolog B (S. pombe)  6 [3, 7-9, 13, 19] (4)  High-mobility group box 1  6 [1, 5, 6, 10, 12, 13] (3)  Total Sample Samp Sizes (with FC)  Mean Fold Change  Range  Validation  287 (229)  6.54  2.74 to 10.50  RT-PCR [7, 18]  256 (176)  4.93  1.81 to 9.20  RT-PCR [19]  264 (161)  3.27  2.66 to 3.91  WB, IHC [20]  141 (56)  7.09  3.00 to 13.00  RT-PCR [21]  172 (130)  6.93  2.96 to 12.00  None found  6 studies: Greatest sample size  CXCL1  6 studies: Moderate sample size  CDC25B  HMBG1 6 studies: Lowest sample size  Interferon  IFITM2  COL1A2  induced transmembrane protein 2 (1-8D)  6 [1-3, 8, 11, 15] (3)  Collagen, type I, alpha 2  6 [2, 3, 5, 7, 12, 15] (4)  51  Gene Name  Total Sample Sizes (with FC)  Mean Fold Change  Description  Studies (with FC)  CDC28 protein kinase regulatory subunit 2  5 [5-7, 9, 19] (5)  285 (285)  Topoisomerase (DNA) II alpha 170kDa  5 [6, 7, 9, 13, 22] (4)  Ubiquitinconjugating enzyme E2C  Range  Validation  4.21  1.79 to 7.20  RT-PCR [9, 19]  277 (237)  3.61  1.05 to 5.60  NB, WB [23]  5 [3, 5-7, 10] (4)  274 (229)  3.03  1.48 to 5.00  RT-PCR [24]  Cadherin 3, type 1, Pcadherin (placental)  5 [1, 3, 7, 9, 25] (5)  194 (194)  18.16  2.78 to 74.00  WB [26, 27]  Inhibin, beta A (activin A, activin AB alpha polypeptide)  5 [3, 5, 7, 13, 25] (4)  198 (158)  11.05  1.71 to 37.00  RT-PCR [27]  208 (139)  10.58  3.58 to 15.15  RT-PCR [6]  208 (208)  4.22  1.74 to 5.70  WB, IHC [28]  5 studies: Greatest sample size  CKS2  TOP2A  UBE2C 5 studies: Moderate sample size  CDH3  INHBA  SLC12A2  MMP11  Solute carrier family 12 (sodium/ potassium/ chloride transporters), member 2 Matrix metallopeptidase 11 (stromelysin 3)  5 [2, 6, 10, 15, 25] (3)  5 [3, 5, 7, 9, 25] (5)  52  Gene Name  Description  Studies (with FC)  Total Samp Sample Sizes (with  FC)  CSElL  HNRPA1  Mean Fold Change  Range  Validation  1.14 to 5.00  None found  CSE1 chromosome segregation 1-like (yeast)  5 [3, 5, 7, 10, 19] (4)  207 (162)  3.74  Heterogeneo us nuclear ribonucleoprotein Al  5 [2, 5-8] (4)  243 (203)  2.89  5 [2, 3, 7, 22, 25] (5)  150 (150)  13.85  2.66 to 17.59  None found  Collagen, type III, alpha 1 (EhlersDanlos syndrome type IV, autosomal dominant)  5 [4, 5, 7, 12, 13] (3)  178 (120)  4.31  1.24 to 9.38  RT-PCR [12]  Collagen, type IV, alpha 1  5 [3, 5, 7, 12, 15] (3)  168 (126)  2.70  1.01 to 4.50  RT-PCR [29]  5 studies: Lowest sample size  CDK10  COL3A1  COL4A1  Cyclindependent kinase (CDC2-like) 10  1.05 to 4.00  None found  53  Table 3.3: Down-regulated Genes Most Commonly Reported in Cancer versus Normal Expression Profiling Studies The 16 down-regulated genes reported in at least five independent studies with consistent direction are presented here. Genes reported by five and six studies were further subdivided into semi-quantitative categories based on the lowest (Q1), moderate (interquartile range), and greatest (values greater than those in Q3) number of tissue samples to give greater importance to the average fold-change criteria for ranking genes when total sample numbers were similar. Validation studies that report a gene as differentially expressed in the opposite direction from that of the meta-analysis are marked with an `" Abbreviations: FC, fold change; RT-PCR, reverse transcriptase polymerase chain reaction; IHC, immunohistochemistry; NB, northern blot; WB, western blot; MS, mass spectrometry.  Gene Name  Description  CA2  Carbonic anhydrase II  MALL  CEACAM1  Studies (with FC)  Total Sample Sizes (with FC)  Mean Fold Change  Range  Validation  -56.00 to -2.30  RT-PCR [12, 13]  11 [1-3, 5, 7, 8, 12, 13, 15, 19, 30] (7)  474 (352)  -15.51  Mal, T-cell differentiation protein-like  7 [2, 3, 7-9, 15, 19] (5)  244 (180)  -5.34  -10.50 to -1.70  None found  Carcinoembryonic antigenrelated cell adhesion molecule 1 (biliary glycoprotein)  7 [1, 2, 5, 7, 13, 15, 19] (5)  222 (158)  -10.40  -40.00 to -1.38  RT-PCR [13, 19]  Hydroxysteroid (11-(3) dehydrogenase 2  6 [1, 3, 7, 8, 19] (5)  224 (184)  -4.47  -760 . to -2.23  NB [31]  Solute carrier family 26 (sulfate transporter), member 2  6 [1, 3, 5, 7, 15, 18] (4)  190 (148)  -6.78  -9.09 to -4.04  None found  6 studies: Greatest sample size  HSD11B2 6 studies: Moderate sample size  SLC26A2  54  Gene Name  FCGBP  Description  Studies (with FC)  Fc fragment of IgG binding protein  6 [2, 3, 5, 7, 8, 10] (4)  Acyl-Coenzyme A dehydrogenase , C-2 to C-3 short chain  6 [1-3, 5, 13, 19] (5)  Creatine kinase, brain  6 [2, 3, 5, 7, 8, 12] (4)  Clusterin  5 [4, 7, 12, 13, 19] (3)  Carboxylesterase 2 (intestine, liver)  5 [3, 5, 7, 15, 19] (4)  Carbonic anhydrase I  5 [2, 3, 5, 8, 19] (4)  Glycoprotein A 33 (t ransmembrane)  5 [1, 2, 7, 11, 15] (5)  Keratin 20  5 [2 , 5, 7, 8, 19] (4)  Selenium binding protein 1  5 [2, 3, 5, 7, 15] (4)  Total Sample Sizes (with FC)  Mean Fold Change  Range  Validation  215 (130)  -4.88  -7.00 to -1.31  None found  168 (128)  -7.11  -20.00 to -2.00  None found  188 (130)  -3.11  -5.00 to -1.10  WB [32]  178 (120)  -3.83  -5.60 to -1.10  IHC [33]  186 (162)  -3.58  -6.30 to -1.15  None found  146 (106)  -36.90  -59'°° to -5.30  RT-PCR [8]  131 (86)  -12.51  6 studies: Lowest sample size  ACADS  CKB 5 studies: Greatest sample size  CLU  CES2 5 studies: Moderate sample size  CA1  GPA33  KRT20  SELENBP1  tot -3 50 -1.70  None found  176 (136)  -8.31  -20.40 to -1.65  None found  154 (130)  -2.80  -3'45 to -1.11  WB, INC, MS [34]  55  Gene Name  Description  Studies (with FC)  Carbonic anhydrase XII  5 [1, 2, 5, 8, 15] (3)  Total Sample Sizes (with FC)  Mean Fold Change  Range  Validation  126 (62)  -4.41  -7.69 to -2.50  C [351* IH^-^-  116 (34)  -4.28  -5.56 to -3.00  RT-PCR [8]  5 studies: Lowest sample size CAl2  FABP1  Fatty acid binding protein 1, liver  5 [2, 3, 8, 12, 15] (2)  56  Table 3.4: Up-regulated Genes Reported in Three or Four Cancer versus Normal Expression Profiling Studies The 77 up-regulated genes reported by three or four studies were further sub-divided into semi-quantitative categories based on the lowest (Q1), moderate (interquartile range), and greatest (values greater than those in Q3) number of tissue samples to give greater importance to the average fold-change criteria for ranking genes when total sample numbers were similar. Abbreviation: FC, fold change.  Gene Name  Description  Total Sample SampMean Fold Sizes Change (with FC)  Range  249 (249)  3.34  1.20 to 4.80  4 [5-7, 19] (4)  235 (235)  . 303  1.50 to 3.35  4 [5-7, 10] (3)  244 (199)  . 241  2.02 to 2.72  4 [5-7, 10] (3)  280 (235)  236 .  1.89 to 2.25  270 (224)  2.07  1.81 to 2.33  234 (189)  32.64  2.08 to 92.00  162 (162)  826 .  3.47 to 21.00  162 ( 162)  7.28  2.48 to 13.00  Studies (with FC)  4 Studies: Greatest sample size  PCNA  GTF3A  SOX4  SOX9  VEGFA  Proliferating cell nuclear antigen General transcription factor IIIA SRY (sex determining region Y)-box 4 SRY (sex determining region Y)-box 9 (campomelic dysplasia, autosomal sexreversal) Vascular endothelial growth factor A  4 [5-7, 9] (4)  4 [7, 14, 15, 30] (2)  4 Studies: Moderate sample size  ETV4  MMP1  CXCL3  Ets variant gene 4 (E1A enhancer binding protein, E1AF) Matrix metallopeptidase 1 (interstitial collagenase) Chemokine (CX-C motif) ligand 3  6,  4 [5, 9, 1 6.1 (3) '  4 [3, 5 ' 7 ' 19] (4)  4 [3, 5, 7, 19] (4)  57  Gene Name  NPM1  HMGA1  Description  Nucleophosmin (nucleolar phosphorprotein B23, numatrin) High mobility group AT-hook 1  RPS2  Ribosomal protein S2  RPL8  Ribosomal protein L8  EIF3S9  PRKDC  Eukaryotic translation initiation factor 3, subunit 9 eta, 116kDa Protein kinase, DNA-activated, catalytic polypeptide  Total Samp Sample Sizes (with FC)  Mean Fold Change  Range  174 (129)  7.10  4.27 to 12.00  4 [5, 6, 9, 13] (3)  229 (189)  5.50  2.90 to 7.09  4 [1, 2, 6, 8] (3)  169 (129)  5.16  2.39 to 8.13  4[2, 5, 6, 8] ( 3 )  183 (143)  3.50  2.16 to 5.35  4 [5, 6, 8, 9] ( 3 )  229 (189)  3.29  2.53 to 4.30  221 (199)  2.49  1.33 to 3.23  " ' 25] (45) 7  158 (158)  9.39  3.32 to 21.00  4 [1, 4, 15 , ( 31 36] `^'  82 (58)  4.90  4.12 to 5.99  ri^A -101 )(2  157 (72)  4.60  3.30 to 5.90  80 (62)  3.99  3.00 to 5.75  Studies (with FC)  4[ 1 2 6  10 ]3 ( ), '  4 [5-7, 14] (3)  4 Studies: Lowest sample size  MMP7  Matrix metallopeptidase 7 (matrilysin, uterine)  HSPD1  Heat shock 60kDa protein 1 (chaperonin)  RAN  PABPC1  RAN, member RAS oncogene family Poly(A) binding protein, cytoplasmic 1  4 [3  4  4 [1, 2, 5, 12] (3)  58  Gene Name  Description  HSP90AB1  Heat shock protein 90kDa alpha (cytosolic), class B member 1  Total Sample Sizes (with FC)  Mean Fold Change  Range  106 (82)  3.72  2 .87 to 5.34  188 (103)  5.35  5.35 to 5.35  6, 9] 1.^.1 (3)  189 (189)  4.85  2.06 to 6.80  3 [5, 6, 10] (2)  184 (139)  3.96  2.21 to 5.71  3 [5, 6, 8] (2)  179 (139)  3.68  2.62 to 4.74  3 [6, 7, 14] (3)  195 (195)  3.67  2.92 to 5.15  3 [5, 25, 30] (3)  232 (232)  3.34  2.62 to 3.92  3 [5, 6, 13] (2)  179 (139)  3.24  2.92 to 3.55  3[4 , 6, 7] (3)  187 (187)  3.04  1.48 to 4.41  3 [5, 6, 13] (2)  179 (139)  2.51  2.10 to 2.91  3 [5-7] (3)  199 (199)  2.40  1.38 to 3.60  6 3 [5, 6 , 13] (2)  179 (139)  2.28  2.17 to 2.38  199 (199)  2.22  1.30 to 2.98  Studies (with FC)  4 [1, 4, 5, 15] (3)  3 Studies: Greatest sample size RPL29 RRM2 RPS7  MIF  FPRL1  ENC1  BMP4 CPNE1  POLR1D  TGIF1  TRAP1  MCM3  Ribosomal protein L29 Ribonucleotide reductase M2 polypeptide Ribosomal protein S7 Macrophage migration inhibitory factor (glycosylationinhibiting factor) Formyl peptide receptor-like 1 Ectodermalneural cortex (with BTB-like domain) Bone morphogenetic protein 4 Copine I Polymerase (RNA) I polypeptide D, 16kDa TGFB-induced factor (TALE family homeobox) TNF receptorassociated protein 1 MCM3 minichromosome maintenance deficient 3 (S. cerevisiae)  3 [6, 8, 10] (1) 3 r5,  3 [5 _ 71  (3)  i  59  Gene Name  SNRPB  ITGA2  TOMM40  Description  Small nuclear ribonucleoprotein polypeptides B and B1 Integrin, alpha 2 (CD49B, alpha 2 subunit of VLA-2 receptor) Translocase of outer mitochondrial membrane 40 homolog (yeast)  Studies (with FC)  Total Sample Sizes (with FC)  Mean Fold Change  Range  3 [5-71 (3)  199 (199)  2.18  1.01 to 3.03  3 [10, 14 , 30](1)  231 (164)  2.06  2.06 to 2.06  3 [ 5-7 ] (3)  199 (199)  1.94  1.66 to 2.13  312' 7 ' 91 (3)  114 (114)  9.94  1.11 to 15.00  3 r3, 5, L^'^L (3)  126 (126)  6.23  2.23 to 12.00  3 [5 ' 7 ' 25] (3)  128 (128)  6.04  4.53 to 7.84  3 [3, 6, 13] (2)  173 (133)  5.23  4.46 to 6.00  3 [4 ' 7 ' 91 (3)  134 (134)  4.70  1.16 to 6.75  3 [2, 6, 8] (2)  147 (107)  4.59  4.05 to 5.13  3 [3, 7, 14] (2)  112 (90)  4.32  1.64 to 7.00  3 Studies: Moderate sample size FN1  SPP1  TACSTD2  BGN WEE1 RPS19  SERPINE1  Fibronectin 1 Secreted phosphorprotein 1 (osteopontin, bone sialoprotein I, early Tlymphocyte activation 1) Tumorassociated calcium signal transducer 2 Biglycan WEE1 homolog (S. pombe) Ribosomal protein S19 Serpin peptidase inhibitor, Glade E (nexin, plasminogen activator inhibitor type 1), member 1  60  Gene Name  CXCL2  FCGR3A  RPL31 RPS5 HSPE1  EIF3S6  HOMER1 ODC1 HMGB2 PYCR1 RPS18 TRIM28 SORD  HNRPH1  PPIB  IMPDH2  Description  Chemokine (CX-C motif) ligand 2 Fc fragment of IgG, low affinity IIla, receptor (CD16a) Ribosomal protein L31 Ribosomal protein S5 Heat shock 10kDa protein 1 (chaperonin 10) Eukaryotic translation initiation factor 3, subunit 6 48kDa Homer homolog 1 (Drosophila) Ornithine decarboxylase 1 High-mobility group box 2 Pyrroline-5carboxylate reductase 1 Ribosomal protein S18 Tripartite motifcontaining 28 Sorbitol dehydrogenase Heterogeneous nuclear ribonucleoprotein H1 (H) Peptidylprolyl isomerase B (cyclophilin B) IMP (inosine monophosphate) dehydrogenase 2  Total Samp Sample Sizes (with FC)  Mean Fold Change  Range  126 (126)  4.23  2.21 to 8.00  131 (110)  4.22  2.73 to 5.70  166 (103)  4.08  4.08 to 4.08  147 (107)  3.79  2.67 to 4.90  319[ 16 ( 36),  175 (175)  3.74  2.54 to 4.80  3 [1, 6, 10] (2)  170 (125)  3.73  3.26 to 4.20  Studies (with FC)  11. 1 7'' I L''( 3 )' ''  [7, 9 3 ' 11](2) 3 [6, 10, 12] (1) 3 [2, 6, 8] (2)  3 [3, 6, 25] (3) 3 [1, 6, 8] (2) 3 [1, 6, 10] (2)  165 (165)  3.65  2.55 to 5.00  165 (125)  3.59  3.27 to 3.90  170 (125)  3.42  2.30 to 4.54  3 [7 ' 8' 19] (2)  136 (96)  3.27  1.23 to 5.30  131 (131)  3.25  2.02 to 4.80  143 (143)  2.90  2.02 to 4.57  132 (132)  2.86  1.19 to 4.70  (1)  11, 18]  142 (103)  2.81  2.81 to 2.81  3 [5, 6, 37] (2)  149 (139)  2.63  2.50 to 2.75  3 [1, 6, 8] (2)  165 (125)  2.57  2.32 to 2.81  3 [2, 4, 6] (3) 3 [2, , ,6 6] (3) [5, 7 ' 3 19] (3) 3 [6,  61  Gene Name  GSTP1 C2  GGH  CDKN3  MYBL2  Description  Glutathione Stransferase pi Complement component 2 Gammaglutamyl hydrolase (conjugase, folylpolygammaglutamyihydrolase) Cyclindependent kinase inhibitor 3 (CDK2associated dual specificity phosphatase) V-myb myeloblastosis viral oncogene homolog (avian)-like 2  Total Samp Sample Sizes (with FC)  Mean Fold Change  Range  157 (139)  2.43  2.38 to 2.48  130 (90)  2.21  1.42 to 3.00  136 (96)  1.87  1.47 to 2.27  '1 L'' 11] (2)  117 (96)  1.80  1.11 to 2.48  3 [7, 11, 13] (' 1)  121 (60)  1.69  1.69 to 1.69  3 [2 , 10 , 37] (1)  59 (4)  8.30  8.30 to 8.30  88 (64)  8.20  2.39 to 14.00  64 (40)  7.44  3.88 to 11.00  100 (100)  6.82  1.19 to 3.28  70 (7 )  5.78  4.00 to 7.33  98 (58)j  4.26  2.05 to 6.47  80 (22)  4.10  4.10 to 4.10  94 (36)  3.70  3.70 to 3.70  Studies (with FC)  3 [5, 6 , 12] (2) 3 [ 3 , 7, 8 ] (2)  [5, 3 7' 13] (2)  1m  7 '^' '  3 Studies: Lowest sample size  GNB2L1  COL1A1 CLDN2 BST2  AHOY  LDHB RPL30 CFB  Guanine nucleotide binding protein (G protein), beta polypeptide 2like 1 Collagen, type I, alpha 1 Claudin 2 Bone marrow stromal cell antigen 2 S-adenosy lhomocysteine hydrolase Lactate dehydrogenase B Ribosomal protein L30 Complement factor B  3 [2, 7, 15] (2) 3 [2, 5, 15] (2) [2, 5 3 ' 71 (3) 3 r^ 2, 3 [2, 19 (3 0; 3 [1, 5, 8] (2)  3 [1, 8, 12] (1) 3 [13, 16, 18](1)  62  Gene Name  RPL6  BMP7  NME1  EEF1A1  CCT7  RPSA RPL3  Description  Ribosomal protein L6 Bone morphogenetic protein 7 (osteogenic protein 1) Non-metastatic cells 1, protein (NM23A) expressed in Eukaryotic translation elongation factor 1 alpha 1 Chaperonin containing TCP1, subunit 7 (eta) Ribosomal protein SA Ribosomal protein L3  Studies (with FC)  Total Sample Sizes (with FC)  Mean Fold Change  Range  3 [1, 8, 10](1)  107 (22)  3.46  3.46 to 3.46  3 [11, 13 , 25] (1)  93 (32)  3.42  3.42 to 3.42  3 [2, 5, 10] (2)  85 (40)  3.37  2.57 to 4.17  3 [1, 10, 37] (1)  77 (22)  3.12  3.12 to 3.12  2.94  2.40 to 3.47  2.74  2.74 to 2.74  2.33  2.30 to 2.36  3 [1, 5, 8] (2)  98 (58)  3 [2, 10, 37] (1) 3 [1, 2, 8] (2)  59 (4)  66 (26)  63  Table 3.5: Down-regulated Genes Reported in Three or Four Cancer versus Normal Expression Profiling Studies The 48 down-regulated genes reported by three or four studies were further sub-divided into semi-quantitative categories based on the lowest (Q1), moderate (interquartile range), and greatest (values greater than those in Q3) number of tissue samples to give greater importance to the average fold-change criteria for ranking genes when total sample numbers were similar. Abbreviation: FC, fold change.  Gene Name  Description  Studies (with  FC)  Total Sample Sizes (with FC)  Mean Fold Change  Range  -15.09  -30.00 to -6.95  -4.56  -39.00 to -12.5  -2.91  -4.00 to -1.82  -29.80  -4.00 to -1.21  -6.65  -7.69 to -5.60  -4.88  -10.00 to -2.08  -4.50  -8.00 to -2.17  -3.11  -4.00 to -2.21  86 (62)  -4.39  -6.95 to -30.00  109 (40)  -4.37  -5.50 to -3.23  4 Studies: Greatest sample size  CD177 SEPP1 CNN1  CD177 molecule Selenoprotein P, plasma, 1 Calponin 1, basic, smooth muscle  43[3 ) , 5, 6, 15] ( 4 [5, 7, 10, 12] (2)  (196 139) 187 (163) 159 (96)  4 [3, 5, 10, 19] ( 3) 4 [5, 12, 13, 13 19] (2)  (14 0 172) 130 (72)  4 [4-6, 15] (3)  4 Studies: Moderate sample size  CA4 MT1H  ADH1C  VIPR1 MYH11  carbonic anhydrase Metallothionein 1H Alcohol dehydrogenase 1C (class I), gamma polypeptide Vasoactive intestinal peptide receptor 1 Myosin, heavy chain 11, smooth muscle  4 [1, 3, 7, 19] (4)  148 (148)  4 [1, 5, 8, 19] (3) 4 [3, 7, 10, 12] (2)  134 (94) 153 (90)  Integral membrane protein 2C HIG1 domain family, member 1A  4 [1, 2, 5, 15] (3) 4 [2, 5, 10, 15] (2)  4 Studies: Lowest sample size  ITM2C HIGD1A  64  Gene Name  Description  FXYD3  FXYD domain containing ion transport regulator 3  Studies (with FC)  Total Sample Sizes (with  Mean Fold Change  Range  -2.82  -4.00 to -1.21  FC) 4 [2, 3, 7, 15] (3)  118 (94)  3 [4, 9, 10] (2)  (74  -18.57  -33.33 to -3.80  3 [2, 5, 6] (3)  143 (1 43)  -17.11  -36.00 to -7.14  3 [3, 5, 6] (3)  169 (169)  -11.25  -25.00 to -3.03  3 [3, 5, 9] (3)  116 (116)  -7.08  -13.00 to -3.23  33)12, [ 19, 30] (  218 (200) 00)  3 [5,  7, 8] (2) I  136 (96)  -2.90  -3.45 to -2.34  7, 8] (2) '  120 (96)  -1.88  -2.50 to -1.26  3 Studies: Greatest sample size SPIB  MS4Al2  SLC4A4  ANPEP  MT1G  UGT1A6  MYL9 TST PRDX6 3 Studies: Moderate sample size MGLL CHGA  Spi-B transcription factor (Spi-1/PU.1 related) Membrane-spanning 4-domains, subfamily A, member 12 Solute carrier family 4, sodium bicarbonate cotransporter, member 4 Alanyl (membrane) aminopeptidase (aminopeptidase N, aminopeptidase M, microsomal aminopeptidase, CD13, p150) Metallothionein 1G  119)  -3.15  -4.10 to -2.20  UDP glucuronosyltransfer ase 1 family, polypeptide A6 Myosin, light chain 9, regulatory Thiosulfate sulfurtransferase (rhodanese)  3 [5,  3 [5, 7, 8] (2)  136 (96 )  -1.73  -2.27 to -1.19  Peroxiredoxin 6  3 [5, 7, 13] (2)  136 (96)  -1.72  -2.33 to -1.10  Monoglyceride lipase Chromogranin A (parathyroid secretory protein 1)  3 [2, 5, 10] (2)  85 (40)  -19.52  -36.00 to -3.03  3 [3, 10, 19] (2)  111) ASS)  -19.00 to -5.50  65  Gene Name  Description  Studies (with FC)  Total Sample Sizes (with  Mean Fold Change  Range  FC) GSN TSPAN1 HSD17B2  Gelsolin (amyloidosis, Finnish type) Tetraspanin 1 Hydroxysteroid (17beta) dehydrogenase 2  3 [2, 5, 10] (2)  85 (40)  -9.96  -17.54 to -2.38  3 [2, 5, 8] (2)  80 (40)  -9.75  -14.50 to -5.00  3 [3, 5, 19] (3)  1022) (10  -9.59  -18.00 to -5.56  EDN3  Endothelin 3  3 [3, 5, 13] (2)  106 ( 66)  -7.50  -10.00 to -5.00  SMPDL3A  Sphingomyelin phosphodiesterase, acid-like 3A  3 [2, 5, 7] (3)  100 (100 )  -6.55  -15.00 to -1.41  PLS1  Plastin 1 (I isoform)  3 [2, 5, 7] (3)  Metallothionein 1A (functional) Meprin A, alpha (PABA peptide hydrolase) Amiloride binding protein 1 (amine oxidase (coppercontaining)) Amyloid beta (A4) precursor proteinbinding, family A, member 3 (X11-like 2) Tetraspanin 7 Neural cell adhesion molecule 2 3-hydroxy-3methylglutarylCoenzyme A synthase 2 (mitochondria!) Chromosome 1 open reading frame 115 Hypothetical protein FLJ40432  3 [12, 13, 19] (1)  MT1A MEP1A  ABP1  APBA3  TSPAN7 NCAM2  HMGCS2  C1or1115 CCNYL1  100 (100) 94 (36)  -6.32  -15.50 to -1.12  _am  -5.70 to -5.70  3 [2, 5, 19] (3)  76 (76)  -5.56  -7.10 to -4.83  3 [5, 8, 19] (2)  112 (72)  -5.32  -8.20 to -2.44  3 [1, 2, 10] (2)  71 (26)  -3.92  -4.50 to -3.33  3 [1, 5, 15] (2)  82 (58)  -3.45  -3.57 to -3.33  3 [1, 5, 8] (2)  98 (58)  -2.75  -3.33 to -2.17  3 [5, 7, 12] (2)  114 (96)  -2.65  -3.85 to -1.45  3[5, 13, 18] (1)  94 (36)  -2.13  -2.13 to -2.13  3 [5, 8, 18] (1)  94 (36)  -2.08  -2.08 to -2.08  66  Total Sample Sizes (with FC)  Mean Fold Change  Range  Gene Name  Description  Studies (with FC)  ATP5B  ATP synthase, H+ transporting, mitochondria! Fl complex, beta polypeptide  3 [3, 8, 15] (1)  94 (30)  -1.00  -1.00 to -1.00  3 [2, 3, 19] (3)  70 (70)  -27.68  -52.00 to -10.85  3 [2, 5, 12] (2)  58 (40)  -11.02  -19.00 to -3.03  3 [2, 5, 11] (2)  61 (40)  -9.59  -15.00 to -4.17  3 [2, 3, 5] (3)  70 (70)  -8.56  -11.00 to -7.00  3 [2, 5, 15] (2)  64 (40)  -7.38  -10.00 to -4.76  3 [2, 12, 18] (1)  40 (4)  -7.00  -7.00 to -7.00  3 [2, 5, 15] (2)  64 (40)  -4.79  -6.00 to -3.57  3 [2, 12, 15] (1) 3 [1, 2, 12] (2)  46 (4) 44 (26)  -4.22 -3.89  -4.22 to -4.22 -4.45 to -3.33  3 [2, 8, 15] (1)  68 (4)  -3.34  -3.34 to -3.34  3 Studies: Lowest sample size  GUCA2A MT2A MUC2 CLCA1  GCNT3 TMEM54 MUC12 KRT8 KRT17 LGALS4  Guanylate cyclase activator 2A (guanylin) Metallothionein 2A Mucin 2, oligomeric mucus/gel-forming Chloride channel, calcium activated, family member 1 Glucosaminyl (Nacetyl) transferase 3, mucin type Transmembrane protein 54 Mucin 12, cell surface associated Keratin 8 Keratin 17 Lectin, galactosidebinding, soluble, 4 (galectin 4)  67  Table 3.6: Up-regulated Genes Most Commonly Reported in Adenoma versus Normal Expression Profiling Studies. The 23 up-regulated genes reported by two or three studies were further sub-divided into semi-quantitative categories based on the lowest (Q1), moderate (interquartile range), and greatest (values greater than those in Q3) number of tissue samples to give greater importance to the average fold-change criteria for ranking genes when total sample numbers were similar. Abbreviation: FC, fold change. Total Sample Sizes (with FC)  Mean Fold Change  Range  3 [1, 2, 19] (3)  30 (30)  4.345  3.48 to 5.21  2 [2, 16] ( 2)  37 (37)  7.55  2 [2, 16] (2)  37  Gene Name  Description  Studies (with FC)  RPS3  Ribosomal protein S3  EPH receptor B2  2 Studies: Greatest sample size EPHB2  SLC12A2  EEF1A1  Solute carrier family 12 (sodium/potassium/chloride transporters), member 2 Eukaryotic translation elongation factor 1 alpha 1  7.10 to 8.00  f) (37)  5.44  5.30 to 5.57  2 [1, 37] (2)  24 (24)  2.65  2.18 to 3.12  2 [2, 19] (1)  12 (4)  20.00  20.00 to 20.00  2 [1, 2] (2)  22 (22)  12.25  7.50 to 17. .00  2 [1, 2] (2)  22 (22)  11.87  3.23 to 20.50  22 (22)  9 . 12  12 (4)  3.23 to 15.00  7.05  22 (22)  6 . 69  2 Studies: Moderate sample size  ETS2  GDF15  PABPC1 HMGB1 RPS29 RPS27A  V-ets erythroblastosis virus E26 oncogene homolog 2 (avian) Growth differentiation factor 15 Poly(A) binding protein, cytoplasmic 1 High mobility -  group box 1 Ribosomal protein S29 Ribosomal protein S27a  2 [1, 2] (2) 2 [2, 19] )1) (1 2 [1, 2] (2)  7.05 to 7.05 2.60 to 10.77  68  Gene Name  NPM1  RNF43 ZFP36L1  IMPDH2 RPL30 RPS15 RPS4X RPS2 TUBA3 RPS25 RPL3  Description  Nucleophosmin (nucleolar phosphoprotein B23, numatrin) Ring finger protein 43 Zinc finger protein 36, C3H type-like 1 IMP (inosine monophosphate) dehydrogenase 2 Ribosomal protein L30 Ribosomal protein S15 Ribosomal protein S4, X-linked Ribosomal protein S2  Studies (with FC)  Total Sample Sizes (with FC)  Mean Fold Change  Range  2 [1, 2] (2)  22 (22)  6.27  5.03 to 7.50  2 [1, 2] (2)  22 (22)  6.22  4.43 to 8.00  22 (22)  5.91  2.81 to 9.00  22 (22)  5.66  2.81 to 8.50  22 (22)  4.30  4.10 to 4.49  12 (4)  3.90  3.90 to 3.90  22 (22)  3.77  2.93 to 4.60  22 (22)  3.01  2.06 to 3.96  22 (22)  3.00  2.95 to 3.05  12 (4)  2.81  2.81 to 2.81  22 (22)  2.20  2.04 to 2.36  2 [2, 37] (2)  10 (10)  4.65  3.90 to 5.40  2 [2, 37] (2)  10 (10)  3.45  2.31 to 4.58  2 [1, 2] (2)  2 [1, 2] 121  Tubulin, alpha 3  `^i 2 [1, 2] (2) 2 [2, 19] (1) 2 [1, 2] (2) 2 [1, 2] (2) 2 [1, 2J (^ 2)  Ribosomal protein S25 Ribosomal protein L3  2 [2, 19] (1) 2 [1, 2] (2)  2 Studies: Lowest sample size  GNB2L1  RPSA  Guanine nucleotide binding protein (G protein), beta polypeptide 2-like 1 Ribosomal protein SA  69  Table 3.7: Down-regulated Genes Most Commonly Reported in Adenoma versus Normal Expression Profiling Studies The 16 down-regulated genes reported by two or three studies were further sub-divided into semi-quantitative categories based on the lowest (Q1), moderate (interquartile range), and greatest (values greater than those in Q3) number of tissue samples to give greater importance to the average fold-change criteria for ranking genes when total sample numbers were similar. Abbreviation: FC, fold change. Total Sample Sizes (with FC)  Mean Fold Change  Gene Name  Description  Studies (with FC)  KRT8  Keratin 8  3 [2, 19, 38] (3)  38 (38)  -3.45  -5.14 to -2.36  2 [2, 16] (2)  37 (37)  -10.67  -18.00 to -3.33  2 [1, 22] (2)  40 (40)  -2.70  -3.33 to -2.06  2 [1, 16] (2)  51 (51)  -2.67  -3.33 to -2.00  2[1, 16] (2)  51 (51)  -2.50  -3.33 to -1.67  2 [1, 2] (2)  22 (22)  -16.50  -28.00 to -5.00  2[1, 2] (2)  22 (22)  -15.00  -27.50 to -2.50  2 [1, 2] (2)  22 (22)  -12.50  -21.67 to -3.33  2 [1, 2] (2)  22 (22)  -12.17  -21.00 to -3.33  2 [1, 2] (2)  22 (22)  -10.00  -15.00 to -5.00  2 [1, 2] (2) 2 [1, 38] (2)  22 (22) 30 (30)  -6.58 -5.70  -9.83 to -3.33 -8.90 to -2.50  2 [1, 2] (2)  22 (22)  -5.65  -8.80 to -2.50  2 [1, 2] (2)  22 (22)  -4.54  -5.75 to -3.33  Range  2 Studies: Greatest sample size  SEPPI IL12RB1 CETP ITPKB  Selenoprotein P, plasma, 1 Interleukin 12 receptor, beta 1 Cholestenjlester transfer protein, plasma Inositol 1,4,5trisphosphate 3-kinase B  2 Studies: Lowest/ Moderate sample size  CA2 CEACAMI  GPA33 SLC26A3 FKBPIA KRT17 THBS2 PRSS8 KIAA0828  Carbonic anhydrase II Carcinoembryonic antigen-related cell adhesion molecule 1 (biliary glycoprotein) Glycoprotein A33 (transmembrane) Solute carrier family 26, member 3 FK506 binding protein IA, 12kDa Keratin 17 Thrombospondin 2 Protease, serine, 8 (prostasin) Adenosylhomocysteinase 3  70  Gene Name  ACADS UQCR  Description  Acyl-Coenzyme A dehydrogenase, C-2 to C3 short chain Ubiquinol-cytochrome c reductase, 6.4kDa subunit  Studies (with FC)  Total Sample Sizes (with FC)  Mean Fold Change  Range  2 [1, 2] (2)  22 (22)  -4.33  -5.33 to -3.33  2 [1, 2] (2)  22 (22)  -3.36  -4.22 to -2.50  71  Table 3.8: Differentially Expressed Genes Reported with a Statistically Significant Frequency in the Adenoma versus Normal and Cancer versus Normal Comparisons Gene Name  GDF15 HMGB1 SLC12A2  NPM1 IMPDH2 PABPCI RNF43 GNB2L1  EEF1A1 RPL3 RPSA RPL30 RPS3 RPS2 RPS4X RPS15 CA2 CEACAMI  ACADS  GPA33 SEPP1 KRT8 KRT17 ITPKB PRSS8 UQCR  Description  Growth differentiation factor 15 High-mobility group box 1 Solute carrier family 12 (sodium/potassium/chloride transporters), member 2 Nucleophosmin (nucleolar phosphoprotein B23, numatrin) IMP (inosine monophosphate) dehydrogenase 2 Poly(A) binding protein, cytoplasmic 1 Ring finger protein 43 Guanine nucleotide binding protein (G protein), beta polypeptide 2-like 1 Eukaryotic translation elongation factor 1 alpha 1 Ribosomal protein L3 Ribosomal protein SA Ribosomal protein L30 Ribosomal protein S3 Ribosomal protein S2 Ribosomal protein S4, X-linked Ribosomal protein S15 Carbonic anhydrase II Carcinoembryonic antigenrelated cell adhesion molecule 1 (biliary glycoprotein) Acyl-Coenzyme A dehydrogenase, C-2 to C-3 short chain Glycoprotein A33 (transmembrane) Selenoprotein P, plasma, 1 Keratin 8 Keratin 17 Inositol 1,4,5-trisphosphate 3kinase B Protease, serine, 8 (prostasin) Ubiquinol-cytochrome c reductase, 6.4kDa subunit  Direction of Differential Expression  Overlap in Adenoma versus Normal  Overlap in Cancer versus Normal  UP UP UP  2 2 2  7 6 5  UP  2  4  UP  2  4  UP  2  4  UP UP  2 2  3 3  UP  2  3  UP UP UP UP UP UP UP DOWN DOWN  2 2 2 3 2 2 2 2 2  3 3 3 2 2 2 2 11 7  DOWN  2  6  DOWN  2  5  DOWN DOWN DOWN DOWN  2 3 2 2  4 3 3 2  DOWN DOWN  2 2  2 2  72  Chapter 3 References 1. Lin YM, Furukawa Y, Tsunoda T, Yue CT, Yang KC, Nakamura Y: Molecular diagnosis of colorectal tumors by expression profiles of 50 genes expressed differentially in adenomas and carcinomas. Oncogene 2002, 21:4120-4128. 2. Buckhaults P, Rago C, St Croix B, Romans KE, Saha S, Zhang L, Vogelstein B, Kinzler KW: Secreted and cell surface genes expressed in benign and malignant colorectal tumors. Cancer Res 2001, 61:6996-7001. 3. Croner RS, Foertsch T, Brueckl WM, Guenther K, Siebenhaar R, Stremmel C, Matzel KE, Papadopoulos T, Kirchner T, Behrens J, Klein-Hitpass L, Stuerzl M, Hohenberger W, Reingruber B: Common denominator genes that distinguish colorectal carcinoma from normal mucosa. Int J Colorectal Dis 2005, 20:353362. 4. Kwon HC, Kim SH, Roh MS, Kim JS, Lee HS, Choi HJ, Jeong JS, Kim HJ, Hwang TH: Gene expression profiling in lymph node-positive and lymph node-negative colorectal cancer. Dis Colon Rectum 2004, 47:141-152. 5. Kim H, Nam SW, Rhee H, Shan Li L, Ju Kang H, Hye Koh K, Kyu Kim N, Song J, Tak-Bun Liu E, Kim H: Different gene expression profiles between microsatellite instability-high and microsatellite stable colorectal carcinomas. Oncogene 2004, 23:6218-6225. 6. Grade M, Hormann P, Becker S, Hummon AB, Wangsa D, Varma S, Simon R, Liersch T, Becker H, Difilippantonio MJ, Ghadimi BM, Ried T: Gene expression profiling reveals a massive, aneuploidy-dependent transcriptional deregulation and distinct differences between lymph node-negative and lymph node-positive colon carcinomas. Cancer Res 2007, 67:41-56. 7. Agrawal D, Chen T, Irby R, Quackenbush J, Chambers AF, Szabo M, Cantor A, Coppola D, Yeatman TJ: Osteopontin identified as lead marker of colon cancer progression, using pooled sample expression profiling. J Natl Cancer lnst 2002, 94:513-521. 8. Kitahara 0, Furukawa Y, Tanaka T, Kihara C, Ono K, Yanagawa R, Nita ME, Takagi T, Nakamura Y, Tsunoda T: Alterations of gene expression during colorectal carcinogenesis revealed by cDNA microarrays after lasercapture microdissection of tumor tissues and normal epithelia. Cancer Res 2001, 61:3544-3549. 9. Koehler A, Bataille F, Schmid C, Ruemmele P, Waldeck A, Blaszyk H, Hartmann A, Hofstaedter F, Dietmaier W: Gene expression profiling of colorectal cancer and metastases divides tumours according to their clinicopathological stage. J Pathol 2004, 204:65-74.  73  10. Bertucci F, Salas S, Eysteries S, Nasser V, Finetti P, Ginestier C, CharafeJauffret E, Loriod B, Bachelart L, Montfort J, Victorero G, Viret F, 011endorff V, Fert V, Giovaninni M, Delpero JR, Nguyen C, Viens P, Monges G, Birnbaum D, Houlgatte R: Gene expression profiling of colon cancer by DNA microarrays and correlation with histoclinical parameters. Oncogene 2004, 23:1377-1391. 11. Ichikawa Y, Ishikawa T, Takahashi S, Hamaguchi Y, Morita T, Nishizuka I, Yamaguchi S, Endo I, Ike H, Togo S, Oki S, Shimada H, Kadota K, Nakamura S, Goto H, Nitanda H, Satomi S, Sakai T, Narita I, Gejyo F, Tomaru Y, Shimizu K, Hayashizaki Y, Okazaki Y: Identification of genes regulating colorectal carcinogenesis by using the algorithm for diagnosing malignant state method. Biochem Biophys Res Commun 2002, 296:497-506. 12. Jansova E, Koutna I, Krontorad P, Svoboda Z, Krivankova S, Zaloudik J, Kozubek M, Kozubek S: Comparative transcriptome maps: A new approach to the diagnosis of colorectal carcinoma patients using cDNA microarrays. Clin Genet 2006, 69:218-227. 13. Williams NS, Gaynor RB, Scoggin S, Verma U, Gokaslan T, Simmang C, Fleming J, Tavana D, Frenkel E, Becerra C: Identification and validation of genes involved in the pathogenesis of colorectal cancer using cDNA microarrays and RNA interference. Clin Cancer Res 2003, 9:931-946. 14. Sugiyama Y, Farrow B, Murillo C, Li J, Watanabe H, Sugiyama K, Evers BM: Analysis of differential gene expression patterns in colon cancer and cancer stroma using microdissected tissues. Gastroenterology 2005, 128:480-486. 15. Takemasa I, Higuchi H, Yamamoto H, Sekimoto M, Tomita N, Nakamori S, Matoba R, Monden M, Matsubara K: Construction of preferential cDNA microarray specialized for human colorectal carcinoma: Molecular sketch of colorectal cancer. Biochem Biophys Res Commun 2001, 285:1244-1249. 16. Habermann JK, Paulsen U, Roblick UJ, Upender MB, McShane LM, Korn EL, Wangsa D, Kruger S, Duchrow M, Bruch HP, Auer G, Ried T: Stage-specific alterations of the genome, transcriptome, and proteome during colorectal carcinogenesis. Genes Chromosomes Cancer 2007, 46:10-26. 17. Yang E, Kang HJ, Koh KH, Rhee H, Kim NK, Kim H: Frequent inactivation of SPARC by promoter hypermethylation in colon cancers. Int J Cancer 2007, 121:567-575. 18. Zou TT, Selaru FM, Xu Y, Shustova V, Yin J, Mod Y, Shibata D, Sato F, Wang S, Olaru A, Deacu E, Liu TC, Abraham JM, Meltzer SJ: Application of cDNA microarrays to generate a molecular taxonomy capable of distinguishing between colon cancer and normal colon. Oncogene 2002, 21:4855-4862.  74  19. Notterman DA, Alon U, Sierk AJ, Levine AJ: Transcriptional gene expression profiles of colorectal adenoma, adenocarcinoma, and normal tissue examined by oligonucleotide arrays. Cancer Res 2001, 61:3124-3130. 20. Volp K, Brezniceanu ML, Bosser S, Brabletz T, Kirchner T, Gottel D, Joos S, Zornig M: Increased expression of high mobility group box 1 (HMGB1) is associated with an elevated level of the antiapoptotic c-IAP2 protein in human colon carcinomas. Gut 2006, 55:234-242. 21. Andreu P, Colnot S, Godard C, Laurent-Puig P, Lamarque D, Kahn A, Perret C, Romagnolo B: Identification of the IFITM family as a new molecular marker in human colorectal tumors. Cancer Res 2006, 66:1949-1955. 22. Galamb 0, Sipos F, Dinya E, Spisak S, Tulassay Z, Molnar B: mRNA expression, functional profiling and multivariate classification of colon biopsy specimen by cDNA overall glass microarray. World J Gastroenterol 2006, 12:6998-7006. 23. Shibao K, Takano H, Nakayama Y, Okazaki K, Nagata N, Izumi H, Uchiumi T, Kuwano M, Kohno K, Itoh H: Enhanced coexpression of YB-1 and DNA topoisomerase II alpha genes in human colorectal carcinomas. Int J Cancer 1999, 83:732-737. 24. Takahashi Y, Ishii Y, Nishida Y, Ikarashi M, Nagata T, Nakamura T, Yamamori S, Asai S: Detection of aberrations of ubiquitin-conjugating enzyme E2C gene (UBE2C) in advanced colon cancer with liver metastases by DNA microarray and two-color FISH. Cancer Genet Cytogenet 2006, 168:30-35. 25. Ohmachi T, Tanaka F, Mimori K, Inoue H, Yanaga K, Mod M: Clinical significance of TROP2 expression in colorectal cancer. Clin Cancer Res 2006, 12:3057-3063. 26. Hardy RG, Tselepis C, Hoyland J, Wallis Y, Pretlow TP, Talbot I, Sanders DS, Matthews G, Morton D, Jankowski JA: Aberrant P-cadherin expression is an early event in hyperplastic and dysplastic transformation in the colon. Gut 2002, 50:513-519. 27. Wildi S, Kleeff J, Maruyama H, Maurer CA, Buchler MW, Korc M: Overexpression of activin A in stage IV colorectal cancer. Gut 2001, 49:409417. 28. Madoz-Gurpide J, Lopez-Serra P, Martinez-Torrecuadrada JL, Sanchez L, Lombardia L, Casal JI: Proteomics-based validation of genomic data: Applications in colorectal cancer diagnosis. Mol Cell Proteomics 2006, 5:1471-1483.  75  29. Ushigome M, Ubagai T, Fukuda H, Tsuchiya N, Sugimura T, Takatsuka J, Nakagama H: Up-regulation of hnRNP Al gene in sporadic human colorectal cancers. Int J Oncol 2005, 26:635-640. 30. Bianchini M, Levy E, Zucchini C, Pinski V, Macagno C, De Sanctis P, Valvassori L, Carinci P, Mordoh J: Comparative study of gene expression by cDNA microarray in human colorectal cancer tissues and normal mucosa. Int J Oncol 2006, 29:83-94. 31. Takahashi K, Sasano H, Fukushima K, Hirasawa G, Miura H, Sasaki I, Matsuno S, Krozowski ZS, Nagura H: 11 beta-hydroxysteroid dehydrogenase type II in human colon: A new marker of fetal development and differentiation in neoplasms. Anticancer Res 1998, 18:3381-3388. 32. Balasubramani M, Day BW, Schoen RE, Getzenberg RH: Altered expression and localization of creatine kinase B, heterogeneous nuclear ribonucleoprotein F, and high mobility group box 1 protein in the nuclear matrix associated with colon cancer. Cancer Res 2006, 66:763-769. 33. Chen X, Halberg RB, Ehrhardt WM, Torrealba J, Dove WF: Clusterin as a biomarker in murine and human intestinal neoplasia. Proc Natl Acad Sci U S A 2003, 100:9530-9535. 34. Kim H, Kang HJ, You KT, Kim SH, Lee KY, Kim TI, Kim C, Song SY, Kim HJ, Lee C, Kim H: Suppression of human selenium-binding protein 1 is a late event in colorectal carcinogenesis and is associated with poor survival. Proteomics 2006, 6:3466-3476. 35. Kivela AJ, Saarnio J, Karttunen TJ, Kivela J, Parkkila AK, Pastorekova S, Pastorek J, Waheed A, Sly WS, Parkkila TS, Rajaniemi H: Differential expression of cytoplasmic carbonic anhydrases, CA I and II, and membrane-associated isozymes, CA IX and XII, in normal mucosa of large intestine and in colorectal tumors. Dig Dis Sci 2001, 46:2179-2186. 36. Mori D, Nakafusa Y, Miyazaki K, Tokunaga 0: Differential expression of janus kinase 3 (JAK3), matrix metalloproteinase 13 (MMP13), heat shock protein 60 (HSP60), and mouse double minute 2 (MDM2) in human colorectal cancer progression using human cancer cDNA microarrays. Pathol Res Pract 2005, 201:777-789. 37. Wang JY, Yeh CS, Tzou WS, Hsieh JS, Chen FM, Lu CY, Yu FJ, Cheng TL, Huang TJ, Lin SR: Analysis of progressively overexpressed genes in tumorigenesis of colorectal cancers using cDNA microarray. Oncol Rep 2005, 14:65-72. 38. Lechner S, Muller-Ladner U, Renke B, Scholmerich J, Ruschoff J, Kullmann F: Gene expression pattern of laser microdissected colonic crypts of adenomas with low grade dysplasia. Gut 2003, 52:1148-1153. 76  Chapter 4: Discussion 4.1 Meta-analyses to Elucidate Biologically Relevant Genes in Colorectal Cancer Tumorigenesis A logical solution to the problem of lack of agreement between expression profiling studies in colorectal cancer is to determine the overlap between many studies utilizing different platforms and observe which genes are consistently reported as differentially expressed. These genes likely demonstrate biological relevance to the tumorigenesis of colorectal cancer, as opposed to sporadically reported genes, which may be false positives.  Meta-analyses have been previously performed to determine differentially expressed genes in colorectal cancer [1-3]. However, these studies and others usually do not consider whether the level of overlap observed is statistically significant. In the most recent version (3.0) of the cancer profiling database Oncomine [4], a meta-analysis tool was implemented to compare results from independent studies. However, Oncomine presently contains raw data for eight colorectal cancer profiling studies, only two of which would qualify for our study [5, 6] because they were the only studies that performed at least one of the three comparisons of interest. As previously discussed, our meta-analysis method is useful when raw data is unavailable for consistent re-analysis, which is usually the case. Furthermore, in the original meta-analysis study, we were able to obtain raw microarray data from five of the studies included in the meta-analysis  77  [7]. Consistent re-analysis of these raw data showed that there was a highly significant level agreement between the two methods. However, one limitation of our meta-analysis method is that a measure of confidence cannot be assigned at the gene level, such as from calculating a true combined fold-change or P-value. Thus, in order for more powerful meta-analysis methods to be applied to colorectal cancer profiling studies, researchers should be encouraged to make public their raw data so that they may be included in repositories such as Oncomine.  By applying this method to a near comprehensive collection of colorectal cancer expression profiling studies, we were able to determine the genes that were reported with a statistically significant frequency. As an extension of the previous version of the meta-analysis method, we categorized some genes according to their total number of tissue samples as lowest (Q1), moderate (interquartile range), and greatest (values greater than those in Q3), instead of using absolute numbers. This allowed the average fold-change criteria to have a greater effect on the gene rank in cases where total sample sizes were similar. In the original version, fold-change rarely had any impact on rank.  4.2 Statistical Significance in Overlap in Two out of Three Comparisons We observed that for the cancer versus normal and adenoma versus normal comparisons, genes were consistently reported as differentially expressed and 78  that the level of overlap was statistically significant. Conversely, while each of the five cancer versus adenoma studies reported differentially expressed genes, the level of overlap among these studies was not significant, suggesting that the number of multi-study genes in the five studies could have been observed due to chance. Determining the significance of overlap between studies provides another filtering step to remove false positive genes from further consideration. When ignoring the significance of the observed overlap, one may be misled by multi-study genes. For example, without knowledge of the statistical significance, one may reason that the multi-study genes in the cancer versus adenoma comparison are biologically relevant, even though this decision can not be reasonably made because the observed level of overlap may be due to chance alone.  4.3 Genes Reported with Inconsistent Direction of Differential Expression In the cancer versus normal comparison, a total of 748 genes were reported as differentially expressed in at least two independent studies. While the majority of these genes were reported as differentially expressed in the same direction, 175 genes (23.4%) were not. Out of these 175 genes, 132 (75.4%) were reported in two studies, 32 (18.3%) were reported in three studies, eight (4.6%) were reported in four studies, two (1.1%) were reported in five studies, and one (0.6%) was reported in six studies. There are many potential explanations for these observed inconsistencies. Firstly, one limitation with such meta-analyses is the  79  overgeneralization of comparisons. While every effort was made to ensure that each study included in each of the three comparisons were comparable, there are bound to be inconsistencies due to the lack of relevant clinical data being reported in each of the studies. For example, in the cancer versus normal comparison, SLC26A3 was reported as down-regulated in five studies [5, 8-11], but as up-regulated in one study [12]. The five studies that reported this gene as down-regulated did not specify the microsatellite status of the colorectal cancer tissue samples being used, while the one study that reported the up-regulation of this gene utilized a mixture of microsatellite stable and unstable tissue samples. Other than microsatellite stability, other clinical features, such as the specific portion of the colon where the tissue samples were taken [9], may impact the direction of differential expression. Thus, due to the lack of this clinical data, it is difficult to determine whether the results of each independent study are truly comparable with each other. Conversely, if this clinical data was more readily available, more specific comparisons, such as microsatellite stable colorectal tissue samples taken from male patients versus paired normal mucosa, could be performed.  A related explanation for why some genes were reported as differentially expressed in an inconsistent direction is the heterogeneity in the tissue samples utilized. The independent studies experimented on tissue samples taken from vastly different populations, each with different genetic and environmental backgrounds that may contribute to differing expression profiles. Furthermore,  80  the tissue samples utilized by each study themselves will be heterogeneous compared to each other. In order to have adequate quantities of tissue to work with, most studies perform high throughput expression profiling on pooled tissue samples, which results in a gene expression signal that is 'averaged' across all cells in the samples [11]. However, the expression of a gene in a single cell may be drastically different from this average. Therefore, depending on how the tissue samples were isolated and which ones were pooled together, the genes may be reported as differentially expressed in an inconsistent direction. One of the studies included in the cancer versus normal comparison [11] investigated the feasibility of performing pooling tissue samples together by plotting the expression signal of all genes in a pooled sample versus the expression signal of genes from one of the samples in the pool. The authors calculated Pearson correlation coefficients and saw that their values ranged from 0.80 to 0.97, suggesting that the pooling of their specific tissue samples maintained patterns of gene expression representative of each distinct tissue sample. Such an analysis should be performed in studies utilizing pooled samples to ensure that the pooled versus unpooled results are comparable.  Finally, poor study design producing inaccurate results may also explain the presence of these genes. In many cases, these genes were ignored because one lone study reported an inconsistent direction of differential expression, which raises suspicions of the validity of the results of the lone disagreeing study. One concern is that some biologically relevant genes may be omitted due to such a  81  study. Therefore, it may be beneficial to include some genes where the majority the studies agreed on the direction of differential expression, instead of the much more stringent 'all-or-none' approach we have utilized. However, since the majority of these genes (75.4%) were reported in only two studies, including these genes would not alter the identity of the highest ranking candidates greatly.  Despite of these inconsistencies, we remind the reader that the majority of the multi-study genes (76.6%) were consistently reported as differentially expressed in the same direction, which is an encouraging result, given that each independent study utilized diverse experimental techniques and tissue samples.  4.4 Literature Review of Consistently Reported Differentially Expressed Genes To further assess our results, we performed a literature review of the genes reported by at least seven studies in the cancer versus normal comparison to determine if any have been shown to have diagnostic and/or prognostic utility in colorectal cancer. The most consistently reported differentially expressed gene in our meta-analysis was carbonic anhydrase 2 (ca2), which was reported as down-regulated in 11 studies. Along with carbonic anhydrase 1 (cal), ca2 has been shown to have prognostic significance where the expression of the corresponding enzymes was related to the metastatic aggressiveness of colorectal cancer [13]. Similarly, the potential diagnostic utility of ca2 was demonstrated in a study measuring CA2 protein levels in fecal matter. The 82  authors found that the average level of CA2 protein in the fecal matter of colorectal cancer patients was shown to be significantly greater than that which was found in the control group [14], which, interestingly, is inconsistent with the direction reported in the 11 studies. Immunohistochemistry (IHC) has been performed on colorectal tumor and healthy mucosa tissue to monitor the protein levels of four carbonic anhydrases, among them CA2 [15]. That study demonstrated the level of CA2 protein decreased in cancer relative to healthy tissue, thus confirming the transcript based expression profiling results.  Carbonic anhydrases catalyze the reversible reaction CO2 + H2O 4- H + + HCO3. To date, 16 isozymes have been identified in mammals [16]. In general, carbonic anydrases are divided into those that are localized in the cytoplasm (1, 2, 3, 5, 7, and 13) and those that are associated with the plasma membrane (4, 9, 12, 14, and 15) [16].  From our meta-analysis results, we observed that colorectal cancer cells show down-regulation of four carbonic anhydrases (1, 2, 4, and 12). However, we observed some inconsistencies between the results of published studies and that of the meta-analysis. For example, many studies have been conducted to show that the carbonic anhydrase 9 (CA9) enzyme is up-regulated in many different types of cancer, including colorectal [17, 18]. However, only one of the twentythree cancer versus normal studies from our meta-analysis reported the upregulation of ca9 [12]. However, this discrepancy can potentially be explained by  83  the results of a recent study that compared the expression of CA9 at the protein level in HNPCC, MSS, and MSI colorectal samples [17]. This study concluded that CA9 was most strongly up-regulated in HNPCC. Since our meta-analysis focused on sporadic cancer which, as described earlier, makes up the majority of colorectal cancer cases, up-regulation of CA9 may be a unique property of inherited colorectal cancer cases.  The role of ca9 in relation to cancer progression has been well studied. The Oncomine database reports that the expression of this gene has been studied in various types of cancer, such as bladder [19], lung [20], and ovarian [21]. CA9, a transmembrane carbonic anydrase with an extracellular active site, has been reported to promote the acidification of the extracellular environment [18]. This observation suggests that the forward reaction, in which CO2 is hydrated, dominates over the reverse reaction. An acidic extracellular pH (pH e ) has been demonstrated to contribute to cancer progression by many different mechanisms [22, 23], such as by promoting normal cell death by inducing necrosis or caspase mediated activation of p53 dependent apoptosis pathways [24, 25]. It is expected that the death of normal cells produces more space into which cancer cells may proliferate [22]. An acidic pH e also promotes angiogenesis through acid-induced release of vascular endothelial growth factor and interleukin-8 [26, 27]. Finally, an acidic pH e has been shown to indirectly promote extracellular matrix degradation by inducing adjacent normal cells to release proteolytic enzymes, thus promoting cancer invasion [28].  84  Another inconsistency between the results of published studies and that of the meta-analysis involves carbonic anhydrase 12 (cal2). While its expression at the transcript level was reported as down-regulated in four independent studies in the cancer versus normal comparison, another study demonstrated, with IHC, its up-regulation at the protein level [15]. Currently, it is unclear how the cal2 transcript appears to be down-regulated and yet an increased amount of CAl2 protein is observed. Future studies should be conducted to determine if the halflives of cancer and normal transcripts differ significantly. Finally, similar to CA9, CAl2 is a transmembrane protein and has been reported to play a similar role in contributing to an acidic pH e [18].  When considered together, the results from the published studies on carbonic anhydrases and from our meta-analysis suggest that colorectal cancer cells show up-regulation of some carbonic anhydrases, but down-regulation of others. While all known carbonic anhydrases catalyze the reversible reaction, they do so with varying catalytic efficiencies [16]. Furthermore, as described, the isozymes have different subcellular localizations. These two characteristics likely account for the different direction of differential expression of these isozymes. Perhaps CA9 and CAl2 show up-regulation at the protein level because they are both membrane associated, which allows them to contribute to an acidic pH e . Therefore, colorectal cancer cells may be selected for their ability to up-regulate these two isozymes. Conversely, CA1 and CA2, both of which were reported as  85  down-regulated by our meta-analysis, are localized in the cytoplasm of the cancer cell. Due to the hypoxic conditions of most cancer cells, anaerobic respiration is utilized [18, 23], which results in the production of energy, in the form of ATP, and lactic acid. This excess of lactic acid is exported out of the cell, along with H + , to the extracellular space. Furthermore, H + is being exported by transporters, such as the Na + -H + antiporter [18]. Therefore, to maintain homeostasis of the intracellular pH, CA1 and CA2 may favour catalyzing the forward reaction to produce W and HCO3." However, by having these two isozymes down regulated, there will potentially be more CO2 that diffuses out of the cell, where CA9 and CAl2 can catalyze the forward reaction to create an acidic pH e . Thus, colorectal cancer cells may be selected for their ability to down-regulate CA1 and CA2. Therefore, the subcellular localization of an isozyme may impact its function and thus explain the observed direction of differential expression.  Given their potential roles in cancer progression, many carbonic anhydrase inhibitors have been developed. For example, derivatives of sulfonamides, such as methazolamide, currently used to treat glaucoma [29], acetzolamide, ethoxzolamide, indisulam, and others have been shown to be potent inhibitors of CAl2 [29]. Indisulam, a compound that has also been shown to inhibit CA9 [29], is currently in clinical development as an antitumor drug [30, 31]. Thus, future studies should also consider the effectiveness of these inhibitors in colorectal cancer.  86  Transforming Growth Factor, beta-induced, 68 kDa (tgfpi) was reported as upregulated in nine studies. TGF131 is a secreted extracellular matrix protein and was discovered through differential expression analysis of a transforming growth factor-131 (TGF-61)-treated human lung adenocarcinoma cell line [32, 33]. This gene has also been shown to be strongly induced by TGF-61 in many other human cell lines [34, 35]. Despite of the consistent reporting of the overexpression of this gene, as far as we know, no study has focused specifically on its diagnostic and/or prognostic utility or its role in colorectal cancer tumorigenesis. Furthermore, over-expression at the protein level has yet to be validated with IHC.  Interferon induced transmembrane protein 1 (ifitm/) was also reported as upregulated in nine studies. The protein product has been shown to mediate the antiproliferative properties of the interferon cytokines [36] and was observed to be over-expressed in gastric cancer cells, which resulted in tumor cells being more resistant to natural killer cells and produced a more invasive phenotype [37]. As far as we know, IHC on human colorectal cancer tissue have not been performed for IFITMI protein, however, RT-PCR was previously conducted on adenomas in a murine model as well as a human colorectal carcinoma cell line, HT29, and elevated expression of ifitm genes (ifitml, ifitm2, and ifitm3) was observed [38]. No further studies have considered the diagnostic and/or prognostic potential of ifitm/ expression in colorectal cancer.  87  Mal, T-cell differentiation protein-like (mall), reported as down-regulated in seven studies, is a member of the MAL proteolipid family [39] and encodes an integral protein located in glycolipid- and cholesterol-enriched membranes. To the best of our knowledge, its expression at the protein level has not been measured by IHC and diagnostic and/or prognostic utilities have not been studied.  Carcinoembryonic antigen-related cell adhesion molecule 1 (ceacaml), reported as down-regulated in seven studies, has been shown to be a tumor suppressor in which expression is lost in adenomas and carcinomas. Moreover, the absence of ceacaml expression was shown to be correlated with reduced rates of apoptosis in polyps [40]. However, a retrospective study performed IHC on CEACAM1 and showed that 58% of colorectal cancer patients showed an increase in expression and that this increase was not related to overall survival or disease-free survival in colorectal cancer patients [41].  Secreted Protein, Acidic and Rich in Cysteine (sparc) was previously demonstrated by our group to be a putative resistance-reversal gene [42] and was reported as up-regulated in seven studies. Differentially expressed genes between resistant and sensitive human MIP101 colon cancer cells were determined and sparc was shown to be consistently down-regulated in the resistant cell lines. Their sensitivity was restored by re-expression of sparc, suggesting that its expression has prognostic utility. IHC performed on colorectal  88  cancer tissue samples showed increased staining of SPARC protein levels [43]. However, another IHC study [44] demonstrated down-regulation of SPARC due to methylation of its promoter. Further studies related to the role of sparc in colorectal tumorigenesis are currently underway in our group.  Growth Differentiation Factor 15 (gdf15), reported as up-regulated in seven studies, is a member of the TGFI3 superfamily. Diagnostic and prognostic utility of gdf15 in colorectal cancer has been suggested by studies that showed increased serum levels of GDF15 protein in colorectal cancer patients relative to healthy controls [45]. These levels increased during disease progression and may have clinical use in the management of colorectal cancer patients [46].  Myc, reported as up-regulated in seven studies in our meta-analysis, is a transcription factor that regulates various processes such as cell-cycle progression, differentiation, apoptosis, and cell motility [47]. IHC on MYC has shown that its expression increases during disease progression [48] and when combined with nuclear 6-catenin expression, myc expression was demonstrated to have prognostic utility [49].  89  4.5 Consistently Reported Differentially Expressed Genes in both Cancer versus Normal and Adenoma versus Normal Comparisons There were 26 differentially expressed genes that were reported with a statistically significant frequency in both the cancer versus normal and adenoma versus normal results. Up-regulated genes found in both comparisons consisted of those involved in ribosomal function and translation. It has been previously shown that up-regulation of the myc transcription factor induces genes encoding for ribosomal subunits and elongation factors [50]. This is to be expected, given the increased rates of protein synthesis that is needed to sustain the high proliferation rate of tumor cells. Conversely, down-regulated genes found in both comparisons included those involved in dehydrogenase activity and fatty acid oxidation. Elimination of fatty acid metabolic enzymes has been shown to be linked to reduced apoptosis in colonic epithelial cells [51].  90  Chapter 4 References 1. Shih W, Chetty R, Tsao MS: Expression profiling by microarrays in colorectal cancer (review). Oncol Rep 2005, 13:517-524. 2. Cardoso J, Boer J, Morreau H, Fodde R: Expression and genomic profiling of colorectal cancer. Biochim Biophys Acta 2007, 1775:103-137. 3. Sagynaliev E, Steinert R, Nestler G, Lippert H, Knoch M, Reymond MA: Webbased data warehouse on gene expression in human colorectal cancer. Proteomics 2005, 5:3066-3078. 4. Rhodes DR, Kalyana-Sundaram S, Mahavisno V, Varambally R, Yu J, Briggs BB, Barrette TR, Anstet MJ, Kincead-Beal C, Kulkarni P, Varambally S, Ghosh D, Chinnaiyan AM: Oncomine 3.0: Genes, pathways, and networks in a collection of 18,000 cancer gene expression profiles. Neoplasia 2007, 9:166180. 5. Notterman DA, Alon U, Sierk AJ, Levine AJ: Transcriptional gene expression profiles of colorectal adenoma, adenocarcinoma, and normal tissue examined by oligonucleotide arrays. Cancer Res 2001, 61:3124-3130. 6. Zou TT, Selaru FM, Xu Y, Shustova V, Yin J, Mori Y, Shibata D, Sato F, Wang S, Olaru A, Deacu E, Liu TC, Abraham JM, Meltzer SJ: Application of cDNA microarrays to generate a molecular taxonomy capable of distinguishing between colon cancer and normal colon. Oncogene 2002, 21:4855-4862. 7. Griffith OL, Melck A, Jones SJ, Wiseman SM: Meta-analysis and metareview of thyroid cancer gene expression profiling studies identifies important diagnostic biomarkers. J Clin Oncol 2006, 24:5043-5051. 8. Lin YM, Furukawa Y, Tsunoda T, Yue CT, Yang KC, Nakamura Y: Molecular diagnosis of colorectal tumors by expression profiles of 50 genes expressed differentially in adenomas and carcinomas. Oncogene 2002, 21:4120-4128. 9. Buckhaults P, Rago C, St Croix B, Romans KE, Saha S, Zhang L, Vogelstein B, Kinzler KW: Secreted and cell surface genes expressed in benign and malignant colorectal tumors. Cancer Res 2001, 61:6996-7001. 10. Croner RS, Foertsch T, Brueckl WM, Guenther K, Siebenhaar R, Stremmel C, Matzel KE, Papadopoulos T, Kirchner T, Behrens J, Klein-Hitpass L, Stuerzl M, Hohenberger W, Reingruber B: Common denominator genes that distinguish colorectal carcinoma from normal mucosa. Int J Colorectal Dis 2005, 20:353-362.  91  11. Agrawal D, Chen T, Irby R, Quackenbush J, Chambers AF, Szabo M, Cantor A, Coppola D, Yeatman TJ: Osteopontin identified as lead marker of colon cancer progression, using pooled sample expression profiling. J Natl Cancer Inst 2002, 94:513-521. 12. Kim H, Nam SW, Rhee H, Shan Li L, Ju Kang H, Hye Koh K, Kyu Kim N, Song J, Tak-Bun Liu E, Kim H: Different gene expression profiles between microsatellite instability-high and microsatellite stable colorectal carcinomas. Oncogene 2004, 23:6218-6225. 13. Bekku S, Mochizuki H, Yamamoto T, Ueno H, Takayama E, Tadakuma T: Expression of carbonic anhydrase I or II and correlation to clinical aspects of colorectal cancer. Hepatogastroenterology 2000, 47:998-1001. 14. Yokoyama S, Shatney CH, Mochizuki H, Hase K, Johnson DL, Cummings S, Trollope ML, Tamakuma S: The potential role of fecal carbonic anhydrase II in screening for colorectal cancer. Am Surg 1997, 63:243-6; discussion 246-7. 15. Kivela AJ, Saarnio J, Karttunen TJ, Kivela J, Parkkila AK, Pastorekova S, Pastorek J, Waheed A, Sly WS, Parkkila TS, Rajaniemi H: Differential expression of cytoplasmic carbonic anhydrases, CA I and II, and membrane-associated isozymes, CA IX and XII, in normal mucosa of large intestine and in colorectal tumors. Dig Dis Sci 2001, 46:2179-2186. 16. Esbaugh AJ, Tufts BL: The structure and function of carbonic anhydrase isozymes in the respiratory system of vertebrates. Respir Physiol Neurobiol 2006, 154:185-198. 17. Niemela AM, Hynninen P, Mecklin JP, Kuopio T, Kokko A, Aaltonen L, Parkkila AK, Pastorekova S, Pastorek J, Waheed A, Sly WS, Orntoft TF, Kruhoffer M, Haapasalo H, Parkkila S, Kivela AJ: Carbonic anhydrase IX is highly expressed in hereditary nonpolyposis colorectal cancer. Cancer Epidemiol Biomarkers Prey 2007, 16:1760-1766. 18. Thiry A, Dogne JM, Masereel B, Supuran CT: Targeting tumor-associated carbonic anhydrase IX in cancer therapy. Trends Pharmacol Sci 2006, 27:566-573. 19. Bredel M, Bredel C, Juric D, Harsh GR, Vogel H, Recht LD, Sikic BI: Functional network analysis reveals extended gliomagenesis pathway maps and three novel MYC-interacting genes in human gliomas. Cancer Res 2005, 65:8679-8689.  92  20. Bhattacharjee A, Richards WG, Staunton J, Li C, Monti S, Vasa P, Ladd C, Beheshti J, Bueno R, Gillette M, Loda M, Weber G, Mark EJ, Lander ES, Wong W, Johnson BE, Golub TR, Sugarbaker DJ, Meyerson M: Classification of human lung carcinomas by mRNA expression profiling reveals distinct adenocarcinoma subclasses. Proc Natl Acad Sci U S A 2001, 98:1379013795. 21. Hendrix ND, Wu R, Kuick R, Schwartz DR, Fearon ER, Cho KR: Fibroblast growth factor 9 has oncogenic activity and is a downstream target of wnt signaling in ovarian endometrioid adenocarcinomas. Cancer Res 2006, 66:1354-1362. 22. Gatenby RA, Gawlinski ET, Gmitro AF, Kaylor B, Gillies RJ: Acid-mediated tumor invasion: A multidisciplinary study. Cancer Res 2006, 66:5216-5223. 23. Stubbs M, McSheehy PM, Griffiths JR, Bashford CL: Causes and consequences of tumour acidity and implications for treatment. Mol Med Today 2000, 6:15-19. 24. Williams AC, Collard TJ, Paraskeva C: An acidic environment leads to p53 dependent induction of apoptosis in human adenoma and carcinoma cell lines: Implications for clonal selection during colorectal carcinogenesis. Oncogene 1999, 18:3199-3204. 25. Park HJ, Lyons JC, Ohtsubo T, Song CW: Acidic environment causes apoptosis by increasing caspase activity. Br J Cancer 1999, 80:1892-1897. 26. Shi Q, Le X, Wang B, Abbruzzese JL, Xiong Q, He Y, Xie K: Regulation of vascular endothelial growth factor expression by acidosis in human cancer cells. Oncogene 2001, 20:3751-3756. 27. Fukumura D, Xu L, Chen Y, Gohongi T, Seed B, Jain RK: Hypoxia and acidosis independently up-regulate vascular endothelial growth factor transcription in brain tumors in vivo. Cancer Res 2001, 61:6020-6024. 28. Rozhin J, Sameni M, Ziegler G, Sloane BF: Pericellular pH affects distribution and secretion of cathepsin B in malignant cells. Cancer Res 1994, 54:6517-6525. 29. Vullo D, Innocenti A, Nishimori I, Pastorek J, Scozzafava A, Pastorekova S, Supuran CT: Carbonic anhydrase inhibitors. inhibition of the transmembrane isozyme XII with sulfonamides-a new target for the design of antitumor and antiglaucoma drugs? Bioorg Med Chem Lett 2005, 15:963969.  93  30. Brown JM, Wilson WR: Exploiting tumour hypoxia in cancer treatment. Nat Rev Cancer 2004, 4:437-447. 31. Scozzafava A, Owa T, Mastrolorenzo A, Supuran CT: Anticancer and antiviral sulfonamides. Cun Med Chem 2003, 10:925-953. -  32. Schneider D, Kleeff J, Berberat PO, Zhu Z, Korc M, Friess H, Buchler MW: Induction and expression of betaig-h3 in pancreatic cancer cells. Biochim Biophys Acta 2002, 1588:1-6. 33. Skonier J, Neubauer M, Madisen L, Bennett K, Plowman GD, Purchio AF: cDNA cloning and sequence analysis of beta ig-h3, a novel gene induced in a human adenocarcinoma cell line after treatment with transforming growth factor-beta. DNA Cell Biol 1992, 11:511-522. 34. Skonier J, Bennett K, Rothwell V, Kosowski S, Plowman G, Wallace P, Edelhoff S, Disteche C, Neubauer M, Marquardt H: Beta ig-h3: A transforming growth factor-beta-responsive gene encoding a secreted protein that inhibits cell attachment in vitro and suppresses the growth of CHO cells in nude mice. DNA Cell Biol 1994, 13:571-584. 35. LeBaron RG, Bezverkov KI, Zimber MP, Pavelec R, Skonier J, Purchio AF: Beta IG-H3, a novel secretory protein inducible by transforming growth factor-beta, is present in normal skin and promotes the adhesion and spreading of dermal fibroblasts in vitro. J Invest Dermatol 1995, 104:844-849. 36. Akyerli CB, Beksac M, Holko M, Frevel M, Dalva K, Ozbek U, Soydan E, Ozcan M, Ozet G, Ilhan 0, Gurman G, Akan H, Williams BR, Ozcelik T: Expression of IFITMI in chronic myeloid leukemia patients. Leuk Res 2005, 29:283-286. 37. Yang Y, Lee JH, Kim KY, Song HK, Kim JK, Yoon SR, Cho D, Song KS, Lee YH, Choi I: The interferon-inducible 9-27 gene modulates the susceptibility to natural killer cells and the invasiveness of gastric cancer cells. Cancer Lett 2005, 221:191-200. 38. Andreu P, Colnot S, Godard C, Laurent-Puig P, Lamarque D, Kahn A, Perret C, Romagnolo B: Identification of the IFITM family as a new molecular marker in human colorectal tumors. Cancer Res 2006, 66:1949-1955. 39. de Marco MC, Kremer L, Albar JP, Martinez-Menarguez JA, Ballesta J, Garcia-Lopez MA, Marazuela M, Puertollano R, Alonso MA: BENE, a novel raftassociated protein of the MAL proteolipid family, interacts with caveolin-1 in human endothelial-like ECV304 cells. J Biol Chem 2001, 276:23009-23017. 40. Nittka S, Gunther J, Ebisch C, Erbersdobler A, Neumaier M: The human tumor suppressor CEACAMI modulates apoptosis and is implicated in early colorectal tumorigenesis. Oncogene 2004, 23:9306-9313. 94  41. Jantscheff P, Terracciano L, Lowy A, Glatz-Krieger K, Grunert F, Micheel B, Brummer J, Laffer U, Metzger U, Herrmann R, Rochlitz C: Expression of CEACAM6 in resectable colorectal cancer: A factor of independent prognostic significance. J Clin Onco/ 2003, 21:3638-3646. 42. Tai IT, Dai M, Owen DA, Chen LB: Genome-wide expression analysis of therapy-resistant tumors reveals SPARC as a novel target for cancer therapy. J Clin Invest 2005, 115:1492-1502. 43. Lussier C, Sodek J, Beaulieu JF: Expression of SPARC/osteonectin/BM40 in the human gut: Predominance in the stroma of the remodeling distal intestine. J Cell Biochem 2001, 81:463-476. 44. Yang E, Kang HJ, Koh KH, Rhee H, Kim NK, Kim H: Frequent inactivation of SPARC by promoter hypermethylation in colon cancers. Int J Cancer 2007, 121:567-575. 45. Welsh JB, Sapinoso LM, Kern SG, Brown DA, Liu T, Bauskin AR, Ward RL, Hawkins NJ, Quinn DI, Russell PJ, Sutherland RL, Breit SN, Moskaluk CA, Frierson HF,Jr, Hampton GM: Large-scale delineation of secreted protein biomarkers overexpressed in cancer tissue and serum. Proc Natl Acad Sci U S A 2003, 100:3410-3415. 46. Brown DA, Ward RL, Buckhaults P, Liu T, Romans KE, Hawkins NJ, Bauskin AR, Kinzler KW, Vogelstein B, Breit SN: MIC-1 serum level and genotype: Associations with progress and prognosis of colorectal carcinoma. Clin Cancer Res 2003, 9:2642-2650. 47. Vita M, Henriksson M: The myc oncoprotein as a therapeutic target for human cancer. Semin Cancer Biol 2006, 16:318-330. 48. Xie D, Sham JS, Zeng WF, Lin HL, Che LH, Wu HX, Wen JM, Fang Y, Hu L, Guan XY: Heterogeneous expression and association of beta-catenin, p16 and c-myc in multistage colorectal tumorigenesis and progression detected by tissue microarray. Int J Cancer 2003, 107:896-902. 49. Bondi J, Bukholm G, Nesland JM, Bukholm IR: Expression of nonmembranous beta-catenin and gamma-catenin, c-myc and cyclin D1 in relation to patient outcome in human colon adenocarcinomas. APMIS 2004, 112:49-56. 50. Kim S, Li Q, Dang CV, Lee LA: Induction of ribosomal genes and hepatocyte hypertrophy by adenovirus-mediated expression of c-myc in vivo. Proc Natl Acad Sci USA 2000, 97:11198-11202.  95  51. Augenlicht LH, Anthony GM, Church TL, Edelmann W, Kucherlapati R, Yang K, Lipkin M, Heerdt BG: Short-chain fatty acid metabolism, apoptosis, and apc-initiated tumorigenesis in the mouse gastrointestinal mucosa. Cancer Res 1999, 59:6005-6009.  96  Chapter 5: Conclusion and Future Studies The results of this thesis research have demonstrated that the application of the meta-analysis method that was developed can identify consistently reported differentially expressed genes in colorectal cancer. That some of the high ranking candidates have already been shown to have diagnostic and/or prognostic utility in colorectal cancer is a very encouraging result and they collectively act as a positive control.  5.1 Validation of Transcript based Gene Expression Profiling Results on Colorectal Tissue Microarrays Tissue microarrays have emerged as a powerful tool to validate the candidate biomarkers identified in transcript based expression profiling studies [1]. They consist of a glass slide that contains tissue samples arranged in an array format. Construction of tissue microarrays begin with collection of archival paraffinembedded, formalin-fixed tissues. Next, representative areas from each tissue are selected by a pathologist. Tissue cores, approximately 0.6 to 2.0 mm in diameter, are removed and then transferred to a recipient block. The resulting tissue microarray is cut into sections and then transferred to glass slides. At this point, IHC with fluorescently labeled antibodies can be performed to quantify the expression levels of the proteins of interest [2, 3]. The intensity and localization of the antibody signal within the tissues are examined by pathologists and scored based on a predefined semi-quantitative scale. 97  One area of future work would be to have a portion of these consistently reported differentially expressed genes validated on tissue microarrays. Antibodies specific to 21 of the high ranking candidates have been ordered and screening the protein expression of those candidates will begin shortly. These candidates were chosen based on their ranking, antibody availability, and their relationship to progression in various types of cancer. Another area of future work involves determining whether the protein expression of the candidates can subdivide patients into prognostic subgroups using various clustering and classification methods. Clinical data from the patients in which the tissue microarrays were generated will be compiled.  In conclusion, the results of this meta-analysis identified genes already shown to have diagnostic and/or prognostic potential in colorectal cancer. Perhaps more interesting are the genes, such as tgfpi and ifitm/, that were consistently reported but have yet to be studied specifically as biomarkers. Also, the genes further down the list (i.e., those identified as differentially expressed by four, five, six, etc. independent studies) warrant further investigation. Further studies focused on these genes will aid in determining a panel of diagnostic and prognostic colorectal cancer biomarkers with sufficient sensitivity and specificity.  98  Chapter 5 References 1. Giltnane JM, Rimm DL: Technology insight: Identification of biomarkers with tissue microarray technology. Nat Clin Pract Oncol 2004, 1:104-111. 2. Wan WH, Fortuna MB, Furmanski P: A rapid and efficient method for testing immunohistochemical reactivity of monoclonal antibodies against multiple tissue samples simultaneously. J Immunol Methods 1987, 103:121129. 3. Kononen J, Bubendorf L, Kallioniemi A, Barlund M, Schraml P, Leighton S, Torhorst J, Mihatsch MJ, Sauter G, Kallioniemi OP: Tissue microarrays for high-throughput molecular profiling of tumor specimens. Nat Med 1998, 4:844-847.  99  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0066251/manifest

Comment

Related Items