1 ASSOCIATIONS OF GUT MICROBIOTA PATTERNS WITH COLORECTAL POLYP PREVALENCE AND DIETARY FIBER INTAKE By TYLER CULPEPPER A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF TH E REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2013
2 2013 Tyler Culpepper
3 To my family members, immediate and extended, who have always provided their unconditional love, and the mentors, peers, pupils, and friend s who never stopped believing I would arrive at this point.
4 ACKNOWLEDGMENTS I owe thanks to many people, whose assistance was indispensable in completing this project. First, I wish to thank my advisor, Dr. Volker Mai, for his guidance and patience t hroughout my degree program, and for allowing me the opportunity and providing me with the resources to complete my project I appreciate all of his contributions of time, ideas, and funding to make my Ph.D. experience productive and stimulating. I am than kful for the excellent example he has provided as a successful scientist. I also thank Dr. Bobbi Langkamp Henken for always being available to discuss results, provide advice, and most importantly, for providing motivation and positive reinforcement throug h some of the most difficult periods of my project I thank my other committee members Dr. Claudio Gonzalez Dr. Nematollah Keyhani and Dr. Joseph Larkin III for always taking time to meet with me and provide constructive criticism and recommendations t o improve the quality of my project Additionally, I thank Dr. Eric Triplet t and the faculty and staff of the Microbiology and Cell Science for their support and assistance. In particular, I thank my fellow graduate students who provided much needed social stimulation and served as an outlet during difficult times. Much of the data generated in this thesis would not have been possible without the support and generous contribution of past and present members of the Mai lab especially Dr. Maria Ukhanova for her indispensible help with experiments and stress management and Dr. Vinata Vedam Mai for her extensive help with coordinating the polyp study Her help and kind words always encouraged me to continue I would like to express my appreciation for the help of Dr. Yijun Sun and Dr. Xiaoyu Wang for their invaluable help with DNA sequencing analysis and figure preparation. I would also like
5 to thank previous undergraduates Madeline Bost, Fleta Netter, and Adeeb Rohani for their generous contributions of time an d effort at the bench. Outside of the lab, I would like to think the entire Henken and Dahl lab research groups for their countless hours working on our clinical trials. I also thank the National Cancer Institute, the Matsutani company, Corn Products, and General Mills for funding this research Lastly, I would like to thank my family for all their encouragement. I thank my parents, who ensured my emotional stability and understood my absence in many familial affairs throughout this challenging process.
6 TA BLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 9 LIST OF FIGURES ................................ ................................ ................................ ........ 10 LIST OF ABBREVIATIONS ................................ ................................ ........................... 14 ABSTRACT ................................ ................................ ................................ ................... 18 CHAPTER 1 FOREWORD ................................ ................................ ................................ .......... 20 2 REVIEW OF LITERATURE ................................ ................................ .................... 23 Background ................................ ................................ ................................ ............. 23 Normal Gut Physiology ................................ ................................ ..................... 24 Normal Commensal Gut Microbiota ................................ ................................ 25 Colorectal Carcinogenesis ................................ ................................ ...................... 28 Methylation ................................ ................................ ................................ ....... 34 Risk Factors for CRC ................................ ................................ ....................... 37 Mucosal Immunology ................................ ................................ .............................. 42 Microbiota, Aberrant Immune Function and IBD ................................ ..................... 44 Evidence for Correlations of Microbiota with CRC ................................ .................. 47 Means of Influencing Gut Microbiota Composition ................................ .................. 52 Conclusions ................................ ................................ ................................ ............ 55 3 M ATERIALS AND METHODS ................................ ................................ ................ 56 Stool Sample Collections ................................ ................................ ........................ 56 Biopsy Collections ................................ ................................ ................................ .. 56 Microbiota Analysis ................................ ................................ ................................ 56 Denaturing Gradient Gel Electrophoresis (DGGE) Analysis ............................. 57 Quantitative PCR ................................ ................................ .............................. 58 454 based 16S rDNA Sequencing ................................ ................................ ... 59 Metagenomic Sequencing ................................ ................................ ................ 60 Fecal pH ................................ ................................ ................................ ................. 61 Statistical Analysis ................................ ................................ ................................ .. 61 4 GUT MICROBIOTA PATTERNS ASSOCIATED WITH COLORECTAL POL YP PREVALENCE ................................ ................................ ................................ ........ 63
7 Differences in Diet and Intestinal Microflora: Potential Associations with Increased CRC Rates in African Americans ................................ ........................ 63 Introduction ................................ ................................ ................................ ....... 63 Study Design ................................ ................................ ................................ .... 63 Results ................................ ................................ ................................ ............. 65 Study population demographic s and dietary habits ................................ .... 65 Fecal microbiota community diversity ................................ ........................ 66 Quantification of targeted bacteria ................................ ............................. 66 454 pyrosequencing analysis of 16S rDNA ................................ ................ 67 Metagenomic shotgun sequencing analysis ................................ .............. 71 Discussion ................................ ................................ ................................ ........ 71 5 EFFECTS OF SPECIFIC COMPLEX CARBOHYDRATES ON FECAL MICROBIOTA COMMUNITY COMPOSITION ................................ ...................... 104 Study #1: Resistant Maltodextrin Increases Bifid obacteria Counts in Healthy Males ................................ ................................ ................................ ................. 104 Introduction ................................ ................................ ................................ ..... 104 Study Design ................................ ................................ ................................ .. 105 R esults ................................ ................................ ................................ ........... 106 Fecal microbiota community diversity ................................ ...................... 106 Bifidobacteria qPCR ................................ ................................ ................. 106 454 pyrosequencing analysis of rDNA ................................ ..................... 107 Conclusion ................................ ................................ ................................ ...... 108 Study #2: The Addition of Whole Grains to the Diets of Middle school Childre n: Effects on Fecal Microbiota Community Structure ................................ ............. 108 Introduction ................................ ................................ ................................ ..... 108 Study Design ................................ ................................ ................................ .. 110 Subject recruitment ................................ ................................ .................. 110 Experimental protocol ................................ ................................ .............. 111 Grain based foods and administration protocol ................................ ........ 11 1 Results ................................ ................................ ................................ ........... 112 Fecal pH ................................ ................................ ................................ ... 112 Fecal microbiota community diversity ................................ ...................... 112 Quantification of targeted bacteria ................................ ........................... 112 454 pyrosequencing of 16S rRNA analysis ................................ .............. 113 Conclusion ................................ ................................ ................................ ...... 113 Study #3: The Effects of Galactooligosaccharides on Fecal Microbiota Composition in 3.1 Undergraduate Students and 3.2 Aged Adults .................... 114 Introduction ................................ ................................ ................................ ..... 114 Protocol 1: Galactooligosaccharides Supplementation in Healthy University Students ................................ ................................ ................................ ............ 115 Study Design ................................ ................................ ................................ .. 115 Subject recruitment ................................ ................................ .................. 115 Experimental protocol ................................ ................................ .............. 116 GOS administration protocol ................................ ................................ .... 117 Results ................................ ................................ ................................ ........... 117
8 Fecal microbiota community diversity ................................ ...................... 117 Quantification of targ eted bacteria ................................ ........................... 118 454 pyrosequencing analysis of 16S rDNA ................................ .............. 118 Protocol 2: Galactooligosaccharides Supplementation in Healthy Aged Adul ts .... 119 Study Design ................................ ................................ ................................ .. 119 Subject recruitment ................................ ................................ .................. 119 Experimental protocol ................................ ................................ .............. 120 GOS administration protocol ................................ ................................ .... 121 Results ................................ ................................ ................................ ........... 121 Fecal pH ................................ ................................ ................................ ... 121 Fecal microbiota community composition ................................ ................ 121 Quantification of targeted bacteria ................................ ........................... 121 454 p yrosequencing analysis of 16S rDNA ................................ .............. 122 Conclusion ................................ ................................ ................................ ...... 123 6 DISCUSSION AND CONCLUSION ................................ ................................ ...... 154 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 187
9 LIST OF TABLES Table page 2 1 Disease specific predisposition genes implicated in CRC. ................................ 31 4 1 Demographic and lifestyle data by case status. ................................ .................. 78 4 2 Dietary differences between case groups and racial groups. ............................. 80
10 LIST OF FIGURES Figure page 2 1 Genetic alterations and the progression of colorectal cancer. ............................ 29 4 1 Diversity ind ices in stool communities between cases and controls. Error bars represent the standard deviation. Total n = 114 (57 cases, 57 controls). ... 81 4 2 qPCR results showing no significant differenc es between demographic groups for lactic acid bacteria and bifidobacteria in stools ................................ .. 82 4 3 qPCR results showing no significant differences between demographic groups for lactic acid bacteria and bifidobacteria in biopsies .............................. 83 4 4 qPCR results showing no significant differences between BMI groups for lactic acid bacteria and bifidobacteria in stools (A) and biopsies (B) .................. 84 4 5 qPCR results showing abundances for lactic acid bacteria and bifidobacteria between racial groups in stools(A) and biopsies (B) ................................ ........... 85 4 6 Rarefaction curves for stool and biopsy communities based on Chao1 ............. 86 4 7 Principal coordinate analysis (PCoA) based on UNIFRAC showing differences between stool and biopsy microbio ta ................................ ............... 87 4 8 Relative abundance of bacteria phyla between cases and controls in stools (A) and biopsies (B). ................................ ................................ ........................... 88 4 9 Heat map showing the distribution of the 20 most significantly differing individual OTUs between all cases and controls in stools at the 95% similarity level ................................ ................................ ................................ .................... 89 4 11 Heat map showing the distribution of the 20 most significantly differing individual OTUs between high risk cases defined conventionally and controls in stools at the 95% similarity leve l ................................ ................................ ..... 91 4 12 Heat map showing the distribution of the 20 most significantly differing individual OTUs between all cases and controls in stools at the 98% similarity level ................................ ................................ ................................ .................... 92 4 13 Heat map showing the di stribution of the 20 most significantly differing individual OTUs between high risk cases and controls in stools at the 98% similarity level ................................ ................................ ................................ ..... 93 4 14 Heat map showing the distribution of the 20 most significantly differing individual OTUs between high risk cases defined by convention and controls in stools at the 98% similarity level ................................ ................................ ..... 94
11 4 15 Heat map showing the distribution of all significantly differing individual OTUs between all cases and controls in biopsies at the 95% similarity level ............... 95 4 16 Heat map showing the distribution of all significantly differing indivi dual OTUs between high risk cases and controls in biopsies at the 95% similarity level ................................ ................................ ................................ .................... 96 4 17 Heat map showing the distribution of all significantly differing individual OTUs between all case s and controls in biops ies at the 98% similarity level ............... 97 4 18 Heat map showing the distribution of all significantly differing individual OTUs between high risk cases and controls in biopsies at the 98% similarity level ..... 98 4 19 Heat map showing the distribution of selected individual OTUs that significantly differed in stools between racial groups at the 95% similarity level ................................ ................................ ................................ .................... 99 4 20 Heat map showing the distribution of selected individual OTUs that significantly differed in stools between racial groups at the 98% similarity level ................................ ................................ ................................ .................. 100 4 21 Heat map based on prevalent OTUs in either case or control groups at the 92% similiarity level ................................ ................................ .......................... 101 4 22 Number of shotgun sequence reads with closest matches to gamma proteobacteria in the case and the control pools ................................ .............. 102 4 23 Number of shotgun sequence reads with closest matches within bacterial metabolic pathways in the case and the control pool s ................................ ...... 103 5 1 DGGE gel showing the enrichment of a particular band as a result of RM treatment ................................ ................................ ................................ .......... 126 5 2 qPCR results showing abundan ces in Bifidobacteria genome equivalents between treatment groups over time ................................ ................................ 127 5 3 Rarefaction curves for RM treatment based on Chao1, which is a measure of estimated diversity if sequenced to completion ................................ ................ 128 5 4 Unifrac based PCoA comparing RM baseline, RM post treatment, and placebo treatment periods ................................ ................................ ................ 129 5 5 He at map showing the distribution of selected individual OTUs that significantly differed in stools between time poi nts at the 95% similarity level .. 130 5 6 Differences in stool pH between WG a nd RG treatment groups over time ...... 131
12 5 7 Diversity indices in stool communities between WG and RG treatment groups over time ................................ ................................ ............................... 132 5 8 qPCR results showing abundances of Bifidobacteria and Lactic acid bacteria (LAB) genome equivalents between WG and RG treatment groups over time 133 5 9 Rarefaction curves for WG tr eatment based on Chao1 ................................ .... 134 5 10 Unifrac based PCoA comparing WG and RG treatment periods ...................... 135 5 11 Relative abundances of bacteri al phyla in stools in WG treated subjects ........ 136 5 12 Heat map showing the distribution of individual OTUs at the 98% similarity level from subjects within the WG treatment group that significant ly differed in stools between time and treatment groups ................................ ....................... 137 5 13 Shannon diversity indices in stool communities between treatment groups over time in the GOS student study population ................................ ................ 138 5 14 Inverse Simpson diversity indices in stool communities between GOS treatment groups over time in the student study population ............................. 139 5 15 qPCR results showing no significant differences in Bifidobacteria genome equivalents between GOS treatment groups over time in the student study population ................................ ................................ ................................ ......... 140 5 16 qPCR results showing no significant differences in Lactic acid bacteria (LAB) genome equivalents between GOS treatment groups over time in the student study population ................................ ................................ ............................... 141 5 17 Relative abundances of bacterial ph yla in stools in GOS treated subjects ...... 142 5 18 Rarefaction curves for GOS treatment in the student study population based on Chao1 ................................ ................................ ................................ .......... 143 5 19 Heat map showing the distribution of individual OTUs that significantly differed in stools between time and GOS treatment groups in the student study population ................................ ................................ ............................... 144 5 20 Measurem ents of stool pH between GOS treatment groups over time in the aged adult study population ................................ ................................ .............. 145 5 21 Diversity indices in stool communities between GOS treatment groups over time in the aged ad ult study population ................................ ............................ 146 5 22 qPCR results showing no significant differences in Bifidobacteria and Lactic acid bacteria (LAB) genome equivalents between GOS treatment groups over time in the aged adult study population ................................ .................... 147
13 5 23 qPCR results showing differences in Bifidobacteria genome equivalent proportions between GOS treatment groups over time in the aged adult study population ................................ ................................ ................................ ......... 148 5 24 The predicted proportion of bifidobacteria in fecal samples after two weeks of supplementation by the proportion of bifidobacteria in baseline samples for A) or <65 y of age (B) ................................ ............ 149 5 25 Relative abundances of bacterial phyla in sto ols in placebo treated subjects and GOS treated subjects post treatment in the aged adult study population .. 150 5 26 Rarefaction curves for GOS treatment in the aged adult study population based on Chao1 ................................ ................................ ............................... 151 5 27 Unifrac based PCoA comparing GO S and placebo treatment periods in the aged adult study population ................................ ................................ .............. 152 5 28 Heat map showing the distribution of individual OTUs that significantly differed in stools between collection period s in the GOS treated group in the aged adult study population ................................ ................................ .............. 153
14 LIST OF ABBREVIATION S C Degree Celsius g Microgram l Microliter M Micromolar B. fragilis Bacteroides fragilis B. lactis Bifidobacterium lactis B. longum Bifidobacterium longum BMI Body mass index bp Base pair AA African A merican ACF Aberrant crypt foci AOM Azoxymethane APC Adenomatous polyposis coli AUC Area under the ROC curve CA Caucasian American CDK Cyclin dependent kinase CIMP CpG island methylator ph enotype COX Cyclooxygenase CRC Colorectal cancer DC Dendritic cell DF Dietary fiber DGGE Denaturing gradient electrophoresis DNA Deoxyribonucleic acid DNMT DNA methyltransferase
15 EDTA Ethylenediaminetetraacetic acid EPIC European Prospective Investigation i nto Cancer E. coli Escherichia coli E. faecalis Enterococcus faecalis FAP Familial adenomatous polyposis g Gram GALT Gut associated lymphoid tissue GI Gastrointestinal GOS Galactooligosaccharides IBD Inflammatory bowel disease IgA Immunoglobin A IL Interle ukin L Liter L. acidophilus Lactobacillus acidophilus L. rhamnosus Lactobacillus rhamnosus LAB Lactic acid bacteria M Molar Mb Megabase pairs mg Milligram MG RAST Metagenomics Rapid Annotation using Subsystems Technology ml Milliliter mm Millimeter mM Mili molar MMR Mismatch repair MSI Microsatellite instability
16 ng Nanogram NK Natural killer NKT Natural killer T OTU Operational taxonomic unit PBS Phosphate buffered saline PCoA Principle coordinate analysis PCR Polymerase chain reaction PSA Polysaccharide A q PCR Quantitative PCR RDP Ribosomal database project ROC Receiver operator characteristic rRNA Ribosomal ribonucleic acid SEM Standard error of the means SFB Segmented filamentous bacteria S. bovis Streptococcus bovia S. gallolyticus Steptococcus gallolytic us SPF Specific pathogen free T1D Type 1 diabetes TAE T ris acetate EDTA TGF Transforming Growth Factor T H T Helper TLR Toll like receptor TNF Tumor necrosis factor T reg Regulatory T cell Tris Tris(hydroxymethyl)aminomethane
17 TST Thiosulfate sulfur transfera se V Volts v / v Volume to volume w / v Weight to volume
18 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy ASSOCIAT IONS OF GUT MICROBIOTA PATTERNS WITH COLORECTAL POLYP PREVALENCE AND DIETARY FIBER INTAKE By Tyler Culpepper August 2013 Chair: Volker Mai Major: Microbiology and Cell Science C olorectal cancers (CRC s ) remain among the mos t frequently detected cancers. H ig her CRC prevalence and mortality are observed in African Americans (AAs). One of the unique characteristics of the colon environment is the presence of a diverse and metabolically active commensal microbiota S equencing technologies allow for an in dept h analysis of associations between microbiota composition and various diseases. This work tested the hypothesis that gut microbiota composition is associated with colorectal polyp presence. Furthermore, the hypothesis that differences in microbiota composi tion are associated with the increased CRC risk observed in AAs was tested Data on dietary and medical history, a fecal sample, and multiple colon biopsy samples w ere collected from 126 subjects undergoing a screening colonoscopy A 16S rDNA sequencing b ased microbiota analysis in 30 individuals presenting with at least one polyp and 30 matched controls revealed that bacterial signatures differed between groups The most significant of these differences were detected in subjects with high risk polyps, str engthening the correlation with carcinogenesis. Furthermore, using a discriminant analysis a microbial pattern predictive of polyp status was detected These data are consistent with the hypothesis that microbiota is associated with CRC risk. We
19 detected d ifferences in specific bacterial signature sequences but not in overall diversity between AAs and Caucasian Americans (CAs). A parallel goal was to determine the effects of fibrous dietary substrate s on gut microbiota composition. Because epidemiological e vidence suggests that dietary fiber intake decreases CRC risk, the effects of various fibrous substrates on gut microbiota composition were determined First, r esistant maltodextrin (RM) was shown to increase some butyrate producing bacteria and also bifi dobacteria, both of which have previously been implied in protection against CRC development. Second, w hole grains (WG) which contain a mixture of dietary fibers, were shown to increase lactic acid bacteria (LAB) which are thought to promote gut health. Lastly, galactooligosaccharides (GOS) increased the proportions of bifidobacteria in aged adults but not young adults. Though no fibrous substrate changed overall gut microbiota composition markedly, all substrates affected the prevalence of several operat ional taxonomic units (OTUs) Each fiber studied may be used as means to influence gut microbiota composition.
20 CHAPTER 1 FOREWORD Colorectal cancers are commonly diagnosed and are a leading cause of cancer related death. In the United States in 2009 13 6,717 new cases were reported with 51,848 deaths (http://apps.nccd.cdc.gov/uscs/toptencancers.aspx) Current s creening techniques, including the colonoscopy procedure are invasive, while others, including the fecal occult blood test, are non specific. Pat ients undergoing a colonoscopy risk perforation of the large bowel. Furthermore, compensation for the team of healthcare providers needed for this procedure increases the economic burden on the patient. Therefore, the development of a non invasive specific screening test is needed. Though many risk factors have been identified, a large proportion of cases are unexplained It is thought that 65 % of CRC arise from environmental fact ors; however, the exact nature of these environmental factors is not fully und erstood (69) Understanding the etiology of CRC will allow for the design of more effective screening and intervention regimen ts. The work presented here is based on the hypothesis that resident members of the normal gut microbiota may contribute to CRC development. While many gut organisms are shared among most individuals, a unique intestinal environment exists within ea ch individual due to the variation in abundance and distribution of these organisms. Individuals belonging to the same demographic group s or individuals who share particular diseases may display similarities in gut microbiota composition based on the parti cular disease present (26, 61 204, 211, 226, 227) Thus, the gut microbiota composition may be an environmental factor that contributes to CRC development The hypothesis will be tested b y determining the differences between the gut microbiota communities of individu als with polyps ( a preneoplastic lesion that is a marker for CRC
21 risk) and individuals without polyps Selecting individuals with early lesions, such as polyps as opposed to fully blown cancer is important because cancers can change the gut environment (particularly immu ne surveillance which likely has a strong impact on microbiota ) Furthermore, cancerous tissue often stimulates angiogenesis and increases blood flow to the affected area. This can change the concentration of nutrients available to the microbiota, and thu s affect the abilities of particular populations of organisms to survive and/or grow. Cancers can also cause blockages and likewise affect the amount of nutrients available. However, a caveat of studying polyps is that not all polyps develop into cancer. R oughly 12% of adenomatous polyps become cancerous and less than 5% lead to invasive cancer (213) AAs suffer from an increased burden of CRC This association led to the hypothesis that differences in the microb iota composition of AAs contribute to increased risk for CRC development The work presented here determined both the differences between the gut microbiota community composition of subjects with polyps and polyp free subjects and the differences between A As and CAs. The differences between racial groups did not relate to the differences found between subjects with polyps and polyp free subjects; however, the devel opment of a stool based test for aberrant microbiota patterns based on the differences between subjects with polyps and poly free subjects would offer a non invasive risk assessment tool that would increase earlier detection and benefit all racial groups Thus, diagnosis and treatment efforts could begin earlier and reduce mortality from this disea se. Beyond the polyp study, other studies were aimed at modifying microbiota compositions towards a beneficial composition E vidence in the literature that suggests
22 lower CRC risk for individuals that consume higher amounts of dietary fiber led to the hypo thesis that potentially beneficial members of the resident gut microbiota interact positively with the fibrous substrates that escape host digestion and reach the colon (22, 143, 205) These interactions, which may include stimulation of growth of potentially beneficial microbes or an increase in certain metabolic activities, may lead to a gut environment that is less conducive to CRC d evelopment. Such favorable interactions may include the production of butyrate, which may contribute to proper cell cycle regulation, and modulation of the immune system away from an inflammatory state. Several studies involving various fibrous substrates showed that microbiota is changed as a result of fiber supplementation. The community compositions resulting from such change were compared with the community composition of polyp free subjects and showed that while each substrate changed gut microbiota co mmunity composition, no substrate was able to shape communities such that they displayed any remarkable similarity to communities in polyp free patients. The discovery of aberrant gut microbiota patterns in subjects with polyps would provide one possible m echanism to the development of polyps and would allow for the development of the screening method based on such microbiota patterns. C oupled with results associating the composition of gut microbiota communities in subjects without polyps to gut microbiota communities after fiber supplementation this work would also demonstrate means of influencing gut microbiota composition Combined, these results would contribute to earlier detection and i ncreased prevention, and would contribute to the curre nt knowledg e of CRC development, r esult ing in a decrease in CRC morbidity and mortality.
23 CHAPTER 2 REVIEW OF LITERATURE Background Colorectal cancers remain among the most frequently observed and fatal malignancies worldwide (http://www.iarc.fr/en/publications/pdfs o nline/wcr/2008/index.php). Mutations in many genes resulting from genetic predispositions or various environmental exposures have been associated with increased CRC risk (31, 62, 167, 200, 232, 234) Dietary habits and the resulting differences in gut microbiota have been suggested to contribute to the increase CRC risk (45, 116, 137, 138, 155, 202) Chronic inflammation and obesity, which both can affect gut microbiota composition, have also been linked to CRC (165, 231, 257) A large body of evidence supports the idea that altering gut microbiota can change host physiology (88, 194, 195, 257) Abnormal activation of immune responses towards commensal microbes is thought to contribute to inflammatory bowel diseases (IBD) that are known to increase CRC risk (231) Many studies also support the idea that changes in host diet, especially the addition of fermentable dietary fiber (DF), can modulate the gut microbiota towards a more protective composition (150, 151) Establishing a clear link between gut microbiota and colorectal carcinogenesis would be crucial for the development of novel prevention approaches that could be directed at both early detect ion of aberrant microbiota patterns indicative of increased CRC risk and shaping microbiota towards a low risk composition by dietary means. Emerging evidence has shown that changes in diet can quickly affect microbiota composition (77) To provide a clear understanding on the development of CRC, aspects of normal colorectal physiology and changes occurring during CRC
24 development will be reviewed first As various links between CRC, inflam mation, and gut microbiota have been established, this review will then describe in some detail host immunity, especially at the mucosal gut interface, as it relates to CRC. Finally, the current evidence for associations between CRC and gut microbiota will be discussed Normal Gut Physiology The gastrointestinal (GI) tract consists of several accessory glands, including salivary glands, the pancreas, and the liver, which are linked to the alimentary canal which stretches from the mouth to the anus. Beginn ing at the most proximal site and continuing distally, these luminal organs consist of the mouth, oropharynx, esophagus, stomach, small intestine and large intestine. The small intestine is further divided into the duodenum, jejunum, ileum, while the large intestine is further divided into the cecum, ascending colon, transverse colon, descending colon, and rectum. Together, the organs of the GI system serve to digest and absorb nutrients, electrolytes, and fluids to maintain metabolic homeostasis (21) The luminal organs along the length of the GI tract beyond the mouth are characterized by the presence of several layers which include, beginning from the most apical layer, the mucosa (which consists of the epithelium and the lamina propria), muscularis mucosae, submucosa, muscularis externa, and serosa. In the healthy gut, commensal microbes are limited to the luminal contents where they interact primarily with the epithelium and lamina propria (21) Both the small and large intestines have increased surface areas due to the three structural features; (1) macroscopic semi lun ar folds know n as folds of Kerckring, (2) finger like project ion s known as villi (found only in the small intestine) surrounded by inwardly juxtaposed structures known as crypts of Lieberkuhn (found in both the small
25 and large intestine), and (3) microvill i present on the membranes of epithelial cells. Intestinal crypts are composed mainly of absorptive epi thelial cells and goblet cells, the latter of which functions in mucin secretion Intestinal crypts throughout the entire gut are characterized by the pr esence of stem/progenitor cells and enteric endocrine cells at their bases (21) Progenitor cells multiply at the base of crypts and migrate up the crypt villous axis, where within at most a few days after reaching the luminal border they are sloughed off as exfoliated colonocytes. This pro cess results in a constant turnover of the large intestinal epithelium (21) Maintenance of structural integrity of the epithelium layer is crucial to avoid leakage of luminal contents. Epithelial permeability is regulated through paracellular resistance, which is dependent upon highly regu lated tight junctions between epithelial cells (21) The colonic epithelium observed in mouse models provides three distinct layers of protection against luminal microbes. A 150 M viscous mucus coating separated into a diffuse outer layer accessible to bacteria and a dense inner layer resi stant to bacterial colonization exists atop a layer of intestinal epithelial cells (120) Mucus primarily consists of heavily O glycosylated proteins known as mucins. Mucins form oligomers which retain water in the glycoprotein matrix (157) Normal Commensal Gut Microbiota The colon harbors up to one hundred trillion bacteria that constitute the commensal human gut microbiota (203) The human microbiome is estimated to contain up to 100 fold more genes than the human genome (111) Microbial communities differ by anatomical site along the colon and their location either within the lumen (fecal samples) or attached to or pene trating the mucin layer (mucosa ) (71) While the stomach and duodenum harbor roughly 10 1 10 3 organisms per milliliter of luminal
26 contents, the small intestine and the large intestine contain an estimated 10 4 10 7 and 10 11 10 12 organisms, respectively (171) While many of the gut microbes represent commensal organisms that live in a beneficial balance within the protected h ost environment, others can become pathogenic (111) The exploration of the microbiome has generated the concept of butes functions that the host could not accomplish alone (87) While microbiota composition differs considerably between individuals, important microbial functions appear more conse rved (57) Various functions have been attributed to the microbiota includi ng diges tion of complex carbohydrates ( such as those in DF), synthesis of vitamins, and modification by deconjugation, dehydrogenation, and dehydroxylation of primary bile acids (5, 86, 163) Through extensive efforts associated with the Human Microbiome Project and related projects, large advances have been made in our understanding of microbiota variation over time and between individuals. These projects have shown that microbiota varies greatly between body sites, and to a lesser extent within body sites between individuals. The gut microbiota is dominated by Firmicutes, Bacteroidetes, Actinobacteria, and Proteobacteria at the phylum level (113) Variation increases at lower taxonomic levels, with thousands of operational taxonomic units (OTUs) detected between individuals (229) However, some studies suggest the presence of a microbial core, or group of organisms found in the majority of individuals (229, 236) Based on the analysis of a large metagenomic data set, the existence of three distinct enterotypes of microbial colonization pattern has been suggested (11, 88)
27 The integrity of tight junctions is compromised in many autoimmune disorders and leads to a condition known as leaky gut synd rome (78) This condition promotes b acterial translocation from the gut into circulation, which results in systemic immune responses and increases the likelihood of sepsis (32) It is unclear as to whether potentially harmful bacteria first breach the epithelial barrier and generate inflammati on which then leads to leaky gut syndrome or if the syndrome develops first facilitating the subsequent translocation of bacteria. Dysbiosis of intestinal microbiota has been associated with autoimmune diseases including eczema and IBD (92, 227) Likewise, many other non autoimmune conditions including type 2 diabetes, obesity, and celiac disease have been linked with aberrant microbiota patterns (65, 136, 237) Recent work has linked distortions in normal gut microbiota establishment in preterm infants with necrotizing enterocolitis, a disease characterized by extensive intestinal inflammation (152) Despite the recently accumulated wealth of evidence for correlations between various diseas e states and microbiota composition the evidence for any causal links is sparse. While observational studies allow for the initial establishment of correlations, showing causality requires microbiota analysis in multiple prospectively collected samples. W ithout such evidence it is not at all obvious if differences in microbiota observed before and after diagnosis or between cases suffering from the respective disease and matched controls are due to the disease or contributed to its development. The large v ariation in gut microbiota composition between individuals and within individuals over time indicates the requirement for a large sample size. A nested case/ control design, with a prospective collection of a series of microbiota samples from
28 populations a t increased risk for the disease of interest offers the most efficient study design. Such studies will be especially difficult for chronic diseases that develop over many years or decades, such as CRC However, without such studies it is unlikely that caus al contributions of the microbiota can be determined Wh ile animal models have served well to prove the general principle that microbes can contribute to carcinogenesis, the specific findings of microbiota contributions cannot likely be extrapolated to hum ans. Colorectal Carcinogenesis Colorectal cancer is thought to initiate in crypt stem cells that form aberrant crypt foci (ACF), regions of cell hyper proliferation at the base of intestinal crypts (24) ACF are foun d more often in rodents with intestinal polyps and are thought to initiate polyp development (228) Polyps, precancerous legions characterized by hyper proliferation and lack of differentiation that vary in their genetic ma keup and potential to progress to CRC, can develop into either hyperplastic polyps or adenomas. Although earlier only adenomas were considered as precancerous, evidence suggests hyperplastic polyps may also be a threat (117) Polyps then develop into cancerous legions known as carcinomas classified into four stages (83) Polyp development is displayed in Figure 2 1 from Volgelstein, et al. (243)
29 Figure 2 1. Genetic alterations and the progression of colorectal ca ncer.
30 Polyps of interest for studying CRC development because 95% of CRCs arise from polyps (115) This risk factor weighs more heavily that lifestyle risk factors and is therefore a desirable marker for prospective cohort studies; however, appr oximately 5% of polyps develop into cancer. Though such low incidence is not ideal for CRC development, it is important to study this environment to determine causal associations. Studies conducted in CRC patients cannot discern whether environmental chang es contributed to CRC development or if CRC contributed to the environmental changes. Like other cancers, CRC arises from abnormalities in proliferation, differentiation, or apoptosis due to either genetic or epigenetic factors that can be inherited or ac quired from environmental exposures. Table 2 1 from Kilpivaara and Aaltonen shows many genes commonly associated with CRC (126) Furthermore, m utations of KRAS and BR AF genes a GTPase and serine/threonine kinase, respectively, which activate the mitogen activated protein kinase (MAPK) pathway, are found abundantly in colorectal cancers (31, 63) PI3KCA encodes the catalytic subunit of phosphatidylinositol 3 k inase (PIK3), which is important for cell proliferation and survival, and has been found to be mutated in 32% of colorectal cancers (199, 200)
31 Table 2 1. Disease specific predisposition genes implicated in CRC. High penetrance colon cancer syndrome Predisposition gene( s) Gene function Lynch syndrome mutL homolog 1 ( MLH1 ) DNA mismatch repair mutS homolog 2 ( MSH2 ) DNA mismatch repair mutS homolog 6 ( MSH6 ) DNA mismatch repair postmeiotic segregation increased 2 ( PMS2 ) DNA mismatch repair Familial adenomatous polypo sis adenomatous polyposis coli ( APC ) Wnt signaling Peutz Jeghers syndrome serine/threonine kinase 11 ( LKB1/STK11 ) Controlling the activity of adenosine monophosphate activated protein kinase (AMPK) family members Juvenile polyposis SMAD family member 4 ( SMAD4 ) Signal transduction of the transforming growth factor (TGF morphogenetic proteins (BMPs) bone morphogenetic protein receptor IA ( BMPR1A ) TGF MYH associated polyposis mutY homolog ( MUTYH ) DNA base excision rep air Colorectal cancer and familial tooth agenesis axin 2 ( AXIN2 ) catenin/Wnt signaling Polymerase proofreading associated polyposis polymerase delta 1, catalytic subunit ( POLD1 ) Catalytic and proofreading polymerase epsil on, catalytic subunit ( POLE ) Catalytic subunit of DNA
32 The activation of the Wnt signaling pathway, important for embryonic development and cell cycle regulation, is often thought to be the initiating event in CRC. Various mutations and deletions in the adenomatous polyposis coli (APC) gene have been detected in CRC models. One study showed that roughly 80% of colorectal adenomas possessed APC mutations, while another study in CRC cell lines showed that about 80% of the cell lines stu died had truncated forms of APC. (118, 219) The condition in which these mutations are inherited is known as familial adenomatous polyposis coli (FAP). FAP patients often display u p to thousands of polyps along the length of their colons (168) This gene acts upstream of many cell cycle regulation proteins through its catenin, a transcription factor to which it binds and targets for degradation (154) Colorectal tumors with intact APC have been shown to have catenin mutations, demonstrating the ability of either type of mutation to ind uce hyper proliferation (167) The expression of the transcription factor c myc, a known oncogene, catenin (102) catenin has been correlated with c myc expression, both of whi ch have been correlated with adenoma size (30) The interactions of c Myc with many genes including cell cycle regulatory genes ha ve been demonstrated; h owever, many of these interactions are not clearly defined (60) c Myc has been shown to induce the expression of cyclin D, a driver of ce ll cycle progression, and its activating counterpart, cyclin dependent kinase 4 (CDK4). Cell culture experiments have demonstrated the dependence of CDK4 expression on c Myc and an increase in cellular proliferation correlated with CDK4 production (104) Furthermore, c Myc has been shown to inhibit the expression of p21, a CDK4 inhibitor (53) Additionally, loss of the CDK inhibitor p27 is associated with
33 decreased stage 3 CRC survival (17) These interactions demonstrate an oncogenic mechanism by which mutations in the Wnt signaling pathway lead to CRC. Mutations in cell cycle regulatory proteins are often detected in CRC. Mutations in the cell cycle arr est gene p53 have been detected in 50% of colorectal cancers (234) Wild type p53 transfected into CRC cell lines has been shown to inhibit cell proliferation with up to 10 fold greater efficiency than mutant p53 (14) A population based analysis revealed that p53 mutations are more common in the distal colon that the proximal colon and are less likely to occur with KRAS mutations or microsatellite instability ( MSI ) (198) Mutation of p53 is thought to be a late transformation event as it is rarely detected in adenomas but can be found abundantly in carcinomas (15) However, mouse models in which human progast rin was over expressed demonstrated an increased number of ACF in p53 mutants, which suggests p53 mutation is an early event (185) These data may conflict due to the difference between models or other genetic factors and demonstrate the involvement of many independent genetic abnormalities in CRC development. Genes involved in growth factor pathways involving prostagland ins are also commonly mutated in CRC (95) Furthermore, mutations of do wnstream MAD related proteins in the TGF (75) Additionally, mutations in epidermal growth factor, i nvolved in the MAPK/ERK pathway, have also been associated with CRC (251) The activation of prostaglandin signaling, notably with COX 2, is also associated with CRC. Prostaglandin E2 contributes to CRC development through activation of PI3K
34 and su catenin (43) The effect of COX 2 is ameliorated in regular aspirin users (44) Studies conducted in CR C cell lines have also demonstrated an increase in COX 2 expression (48) Furthermore, a decrease in the COX 2 inhibitor 15 PGDH, hydroxyprostaglandin dehydrogenase, has been observed in CRC and adenomas (258) Mutatio ns in these CRC associated genes can affect cell turnover, mucin production, secretion of antimicrobial peptides, thus changing luminal conditions affecting gut microbiota composition and activities. Enzymes involved in DNA repair are often mutated or sile nced in CRC. Germ line mutations in the base excision repair gene MUTYH have been observed in CRC (4) Additionally, germ line mutations in DNA mismatch repair genes hMLH1 and hMSH2 lead to a condition known as hereditary nonpolyposis colon cancer, or Lynch syndrome, and lead to an increased risk of CRC (25) This condition can also develop as a result of promoter hypermethylation de scribed in detail below. Adherent E. coli have been shown to decrease expression levels of DNA mismatch repair proteins MLH1 and MSH2 (149, 149) Methylation Epigenetic modifications have been shown to o ccur in response to environmental changes and thus might also correlate with activities of the gut microbiota, especially in the gut epithelium. DNA methyl transferase enzymes, particularly DNMT1, are responsible for CpG methylation through a reaction with the methyl donor S adenosyl methionine (187, 188) CpG islands, frequently located in the promot ers of eukaryotic genes, are usually unmethylated. In contrast, CpG dinucleotides not associated with CpG islands are frequently methylated. (23) The number of CpG islands in the human genome has been estimated to be roughly 45,000 (7) Promoter methylation is generally
35 deoxyazacytidine has been shown to restore gene exp ression (121, 159) According to the Knudson two hit hypothesis, once both alleles of a tumor suppressor gene have become either methylated in their promoter regions or mutated, malignant transformation of the cell may begin (131) Hypermethylation of the promoters of both alleles is common in sporadic colon cancers; however, it is less common in familiar cancers as the germ line mutation is the first hit. (76) Loss of function in tumor suppressor genes due to promoter methylation is observed frequently (121) Several mechanisms exist by which methylation inhibits gene expression, the first of which is through mutation. Methylation of cytosine residues promotes cytosine to ring (191) The remaining mechanisms depend upon protein interactions with the methylated cytosine. Direct binding of methyl cytosine binding transcription repressors can inhibit the form ation of the open promoter complex (214) Additionally, methylated DNA can recruit histone deacetylases, the activity of which subsequently leads to transcriptional r epression, as deacetylation of histones also inhibits the open promoter complex (206) Histone deacetylase inhibitors alone are not able to restore the expression of genes with methylated promoters, but work synergistically with demethylating agents such as deoxyazacytidine to restore gene expression (39) This suggests that methylation, not deacetylation, is the dominant suppressive epigenetic modification. Promoter hypermethylation of several genes has been associated with colorectal carcinogenesis. Inactivation of hMLH1 via p romoter methylation leads to MSI, and is associated with sporadic CRC (164) Reversal of hMLH1 promoter methylation has
36 been shown to restore expression of the HMLH1 protein and subsequently mismatch repair (MMR) gene capacity in MMR deficient ce ll lines (103) CpG island methylator phenotype (CIMP), which has been define d as a condition in which three or more CpG island loci in colorectal tumors are methylated, has been associated with hMLH1 methylation and MSI, and is thought to contribute to MSI development. Though CpG island methylation has been found in normal tissue and is associated with increasing age, some CpG island methylation has been found only in cancerous tissues independent of age (235) Though a molecular mechanism by which CIMP develops has not yet been elucidated, Ogino et al have developed a panel of eight genetic markers that effectively predict CIMP status (174) Similarly, the promoters of several other genes have been shown to be highly methylated in CRC, including tumor suppressor gene p16 serine proteinase inhibitor tissue factor pa thway inhibitor 2 Netrin 1 receptor UNC5C, and E3 ubiquitin ligase HACE1 (94, 106 108) Burri et al reported tumor specific methylation of both p16 and p14 with p16 methylation found primarily in poorly differentiated proximal colon adenocarcinomas and p14 methylation found primarily in well differentiated distal colon adenocarc inomas (37) The occurrence O 6 methylguanine DNA methyltransferase ( MGMT) pr omoter methylation is significantly increased in tumors and adjacent normal tissues of sporadic CRC subjects (210) Hypomethylation of genes, particularly growth related genes, also contributes to the development of CRC. Sharrard et al found that the oncogene c myc was more frequently hypomethylated in adenocarcinomas, m etastatic deposits, adenomatous polyps, and hyperplastic polyps compared to normal mucosal tissue (209) H RAS has
37 also been shown to be hypomethylated in colonic tumors (80) Studies of potential associations between microbiota composition and methylation status are currently ongoing but have not yet been reported in the literature. Once such an association is established the next step would be to determine if changes in microbiota composition, through targeted dietary interventions that might include prebiotics, probiotics or synbiotics, can reverse methylation pattern in the intestinal epithelium. Risk Factors for CRC For average risk indi viduals aged 50 years or older, the American Cancer Society currently recommends one of four routine CRC screening options: (1) fecal occult blood testing every year, (2) flexible sigmoidoscopy every 5 years, (3) double contrast barium enema every five yea rs, or (4) colonoscopy every 10 years (220) Common factors influencing CRC risk include physical activity, diet, family history, obesity, smoking, and aspirin use (232) Many of the dietary factors associated with CRC that are discussed below have been or likely could be linked to microbiota composit ion (151) Obesity has been po sitively associated with CRC risk (165) Data suggest associations between microbiota and obesity, although causality is still not well established (236, 237) Similarly, smoking has been shown to have a positive dose dependent association with CRC risk (46) Inverse associations between CRC and both anti inflammatory medications and physical activity hav e been established (52) A case control study conducted by Sandler et al found that daily aspirin use significantly reduced the occurrence of adenomas in patients who previously had CRC (201) Rofecoxib, another cyclooxygenase inhibitor, has likewise been shown to decrease the
38 occurrence of adenomas (184) Again, both exposures are likely to affect microbiota composition. The evidence regarding associations of dietary fiber intake with CRC risk is not fully consistent. A strong association between DF intake and CRC would be consistent w ith a role of microbiota in CRC as DF can change microbiota composition (150) The analysis of several prospective cohort studies showed that there was an inverse relationship between CRC and dietary fiber intake using an age adjusted model, but no significant relatio nship was seen in fully adjusted models (179) Th e National Institute of Health AARP Diet and Health Study found that a lower consumption of meat and potatoes and a higher consumption of fruits and vegetables was associated with lower CRC risk (82) Data from this study also showed that both whole grain intake and intake of fiber from whole grains were sig nificantly associated with a decrease in CRC risk (205, 205) Larsson et al likewise observed that the consumption of whole grains was associated with decreased CRC risk (137) However, Terry et al showed that the consumption of larger amounts of fruit and vegetables was signif icantly associated with decreased CRC risk while fiber consumption from cereal was not (230) Several studies, including the Netherlands Cohort Study on Diet and Cancer and the Tennessee Colorectal Polyp Study, likewise found a significant association between fruit and vegetable intake and decreased CRC risk (189, 244, 252) showed no significant associations of fiber intake from fruits and vegetables with CRC risk, but showe d a significant association of legume fiber intake with decreased CRC risk (143) Data from the European Prospective Investigation into Cancer and Nutrition (EPIC) study showed that a higher consumption of dietary fiber was associated with
39 lower CRC risk, but the food source of the fiber was not important (22) Methodological differences may account for this discrepancy, as one study found that fiber intake in both an age adjusted model and multivariable models was significantly associated with decreased CRC risk when using data obtained from food diaries, but results using data from food frequency questionnaires provided by the same subjects were not significant (59) Confounding variables may also account for discrepancies as suggested by other prospective cohort studi up Study) in which no significant associations between dietary fiber intake and CRC risk were found (160, 161) The evidence for associations between fruit, vegetable, and whole grain intake and CRC are not much stronger although there are some suppo rtive data (218) Other dietary habits associated with CRC risk include consumption of red meat, fish, vitamins, minerals, and fat. A meta analysis of several prospective studies showed an increase in CRC risk associated with red and processed meat consumption (138) A prospective cohort study conducted by Norat et al showed that increased red meat consumption was significantly associated with increased CRC risk while consumption of fish was significantly associated with decreased risk (170) The method and degree of preparation have been also associated with CRC risk. Butler et al found that consumption of pan fried and well done red meat had the largest associations with CRC risk (38) Calcium and milk consumption were both found to significantly reduce CRC risk in a meta analysi s spanning several cohort studies (51) Another meta analysis conducted by Sanjoaquin et al showed that dietary folate is significantly inversely associated with CRC risk (202) A positive association between CRC risk and energy
40 intake was found in yet another meta analysis conducted by Howe, et al, but no significant associations were found regarding total fat and cholesterol (112) However, data from the Finnish Mobile Clinic Health Examination Survey showed significant positive associations between cholesterol intake and CRC risk, but not other fats (116) Additionally, a study conducted at the National Naval Medical Ce nter showed that increased oleic acid intake was associated with increased CRC risk (155) Gut micr obiota composition may play a role in these dietary associations, as the metabolic activities and by products of the microbiota (discussed below) are affected by the substrates which they receive. Many of these dietary associations are not strongly correla ted with case status, and this may be because of differences in gut microbiota community composition. Some studies show significant dietary associations with CRC risk only in specific regions of the colon. For instance, there was a significant association between high fruit and vegetable intake and low distal CRC risk only (133) Another study showed that rectal cancer was inversely associated with fruit, vegetable, and who le grain intake, and was positively associated with refined grain intake (218) Associations between CRC and red meat intake also appear to vary by location. Chao et al showed that both distal colon and rectal cancer risk were reduced within subjects consuming lower amounts of red meat (45) However, English et al showed only an increase in rectal cancer, but not colon cancer, associated with higher consumption of red meat (74) The data suggesting that some risk factors vary by anatomic location suggests that microbiota might behave in a similar fashion as a risk factor, which would indicate a benefit for stratifying studies of associations b etween CRC and microbiota by location.
41 Some data suggest that race and gender may contribute to dietary associations with CRC. African Americans suffer from an increased incidence and mortality of CRC in comparison to Caucasian Americans and males suffer h igher rates within each racial category (American Cancer Society: Statistics for 2012, http://www.cancer.org/docroot/stt/stt_0.asp?from=fast ). The observed racial differences may be due to variations in the diets between groups contributing to differences in microbiota (151) We showed that AAs consumed significantly more heterocyclic amines, potentially carcinogenic products of food preparation discuss ed later, than did CAs. Gender differences were observed in a study c onducted by Deneo Pellegrini et al which showed that only males experienced a protective effect of plant food intake against CRC (66) Furthermore, d ata from the Japan C ollaborative Cohort Study showed a decrease in CRC risk associated with an increase in fiber intake, but this association was stronger in males (247) Distal, but not proximal, CRC risk was reduced by increasing calcium intake in both men and women (217) Similarly, data from the risk associated with elevated c alcium intake, but no significant associations were seen with fiber or vitamins (212) menopausal women found no significant associations between calcium and vitamin D intake and CRC risk, though another trial did find a significant association of vitamin D intake with decreased CRC risk (93 246) These data did, however, show a significant reduction in CRC risk in subjects treated with estrogen and progesterone (50) Nilsen et al reported a positive association between blood glucose and diabetes with CRC risk in women, while in men the only significant association with increased CRC risk was low
42 physical activity (169) Similarly, a high glycemic load has been associated with CR C risk in women (109) Cellular and molecular effects of dietary intake have also been documented. Slattery et al have shown that high vegetable, whole grain, and fiber intake a re associated with a lower incidence of p53 (a cell cycle regulatory protein) mutations, while only high vegetable and fiber intake are associated with a lower incidence of KRAS (an oncogenic growth factor) mutations (217) Fish and olive oil added to HT 29 and Caco 2 cell cultures increased both apoptosis and cell differentiation and decreased COX 2 expression and subsequently Bcl 2 (an apoptosis regulatory protein) expression; however, only cell differentiation was induced by oleic and linoleic acids. Fish oil also significantly diminished cell proliferation (145) Animal studies provide better mechanistic insight to dietary associations with CRC. Dietary supplementation of either black or navy bean s upon administration of azoxymethane (AOM), an experimental colon carcinogen, in rats led to significantly reduced incidence and multiplicity of adenocarcinomas (101) Mai, et al, have shown ea rlier, that in APC Min mice, diet induced differences in microbiota composition correlate wit h numbers of intestinal polyps. U nique bacterial signatures were also detected in mice with no or few polyps which suggest ed that the gut environment between mice w ith low verses high polyp burden differed (150) Mucosal Immunology Colonic inflammation and CRC are closely associated (231) Therefore, it is important to understand the i nteractions between commensal gut microbiota and the mucosal immune system. Upon initial colonization, the human GI tract is continuously exposed to an abundance of diverse microbes. Through mechanisms of tolerance, and
43 possibly ignorance as long as microb ial contents remain confined to the lumen, the immune system facilitate s a balance that allows for co existence between microbes and host. The epithelium provides a physical barrier against microbes while the lamina propria contains the immune cells that m aintain a tolerogenic response to these microbes. In the lamina propria, microbes that breach the epithelium are eliminated by the action of various immune cells. antibacter ial lectin, in response to pathogen associated molecular patterns (PAMPs) via the MyD88 dependent toll like receptor (TLR) pathway (239) may help compensate for the lack of a dense inner mucus layer. Paneth cells within the defensins which are capable of altering the bacterial composition of the lumen, wher localized to the mucus layer (197, 240) The intestinal lamina propria contains gut associ ated lymphoid tissue (GALT) appendix and colon, and diffuse immune cells throughout the lamina propria. Macrophages produce few pro inflammatory cytokines in response to i nflammatory stimuli but have the ability to induce an anti inflammatory effect via the production of interleukin 10 (IL 10) (67, 221) Dendritic cells (DCs) can project dendrites into the intestinal lumen and present microbial antigens to lymphocytes accompanied by IL 17 secretion to induce antibacterial immunity (49, 67) Upon stimulation by DCs, B cells become antibody producing plasma cells which then home to the lami na propria where they secrete immunoglobin A (IgA) into the intestinal lumen via transcytosis (166) While
4 4 no systemic antibody response is induced, bacteria opsonized with IgA cannot cross the epithelial barrier (148) However, antibodies against commensal bacteria have been detected in serum of healthy donors (98) ileum, participate largely in the crosstalk between the gut microbiota and the host (242) Specialized enterocytes known as M cells reside within the epithelium surrounding delivering antigens to DCs and lymphocyte cells transport free antigens, including bacteria, viruses, protozoa, and macromolecules, in addition to IgA coated antigens from the intestinal lumen, to DCs and other immune cells within the GALT. Analogous lymphoid aggregates which function similarly are found in the colon. Several T helper (T H ) lymphocyte subsets, which include inflammatory T H 1 and T H 17 cells, humoral immunity inducing T H 2 cells, and anti inflammatory T regulatory cells (T regs ), are stimulated by gut microbi ota. In mice, segmented filamentous bacteria (SFB), or Candidatus Savagella, have been shown sufficient to induce signaling of all four T helper subtypes, similar to responses to conventional microbiota (85, 197) Several clostridial strains from the clusters IV and XIVa, which con tain many important butyrate producers discussed later, were shown to induce a T reg phenotype both in the lamina propria and systemically (12) Bacteroides fragilis induces IL 10 production though a polysaccharide A dependent response which prevents the expansion of T H 17 cells (194) Microbiota, Aberrant Immune Function and IBD IBD is a group of infl ammatory conditions of the intestine associated with increased risk of developing CRC and is thought to arise due to an aberrant immune
45 response to gut microbiota in genetically susceptible hosts (257) disease exhibit a higher prevalence of adherent invasive E. coli, similar to CRC patients discussed later (61) Further evidence from murine models demonstrated interplay between gut microbiota and IBD in genetically susceptible hosts, as germ free IL 10 knockout mice did not develop disea se; however, specific pathogen free (SPF) colonized counterparts rapidly developed disease (208) Another study showed that germ free mice compared with SPF colonized mice possessed larger amounts of intestinal natural killer T (NKT) cells and displayed higher levels of oxazolone induced IBD and allergic asthma (177) Conflicting data may be the result of different genetic or IBD induction models between experiments. Additional indirect evidence for a contribution of microbiota comes from the detection in IBD cases of mutations in genes important for gut microbiota i mmune signaling. Mutations in NOD2 whose product binds to ba cterial lipopolysaccharides and interacts with nuclear transcription factor NF humans (175) Increase d production of IL 23, secreted largely by dendritic cells, has been shown in both mice and humans to be associated with active IBD, likely via downstream effects on IL 17 production (73, 110) Mutations in the IL 10 receptor have been shown in humans to be associated with IBD (89) An increase in both T reg numbers and IL 10 expression have been observed in response to PSA from B fragilis in a TLR 2 dependent manner (195) These results together suggest that some members of the microbiota are important for immune homeostasis. Both the addition of potenti ally beneficial bacteria and the removal of potentially harmful bacteria have been demonstrated to improve disease state. Supplementation with the probiotic mixture
46 VSL#3 has been shown to significantly improve disease in pouchitis patients, while patients treated with the antibiotic metronidazole have shown similar improvement (88, 225) Systemic immunity an d autoimmune diseases other than inflammatory bowel diseases have been linked to gut microbiota. SFB promote arthritis and experimental autoimmune encephalitis development, a condition in animal models analogous to human multiple sclerosis (142, 253) PSA producing B. fra gilis induces the expansion of circulating CD4+ T cells (156) MyD88, Myeloid differentiation primary response gene (88), a common adaptor protein in multiple TRL signaling cascades, is important in the development of type 1 diabetes (T1D). MyD88 / germ free mice develop robust T1D, b ut this affect is attenuated in SPF colonized mice (249) Though many positive associations between bacterial groups and disease development have been demonstrated, some comportments of the gut microbiota have been shown to have a protective effect against autoimmunity. Clostridial strains belonging mainly to Clostridium clusters IV and XIVa have been shown to increase T reg numbers in both the lamina propria and systemically and furthermore reduce the intensity of colitis in mice (12) The mechanisms by which inflammation induces CRC have been studied extensively. Many mechanisms involve increased a ctivation of the transcription factor NF et al (96) In humans, Sakamoto et al observed constitutive activatio n of NF lines (196) Tumor tissue examined in parallel with autologous normal tissue in 28
47 patients showed an increase in nuclear NF (132) NF catenin and subsequently up regulates downstream oncogenes including c myc (123) IL 6 and TNF dependent cytokines, are important for the development of colitis associated CRC in the mouse model (97, 181) A cross sectional study conducted by K im et al demonstrated that CRC patients have significantly higher levels of serum IL 6 and TNF (128) Evidence for Correlations of Microbiota with CRC The discovery that Helicobacter pylori is causally associated with non cardia stomach cancers spurred interest in the potential contribution of other infect ious agents in intestinal cancers (180) Increased diversity in the Clostridium coccoides and Clostridium leptum groups has been observed in CRC patients compared with healthy controls ; however these groups of organisms contain many known butyrate producers, whose abundance are thought to be inversely associated with CRC (204) Superoxide producing Enterococcus faecalis a commensal gut organism, has been shown to induce DNA damage in ce ll lines and rat colonic tissue (114) Other bacterial species or groups have been studied more extensively, and both positive associations most notable with E. coli B. fragilis and S. gallolyticus have been found in addition to inverse associations, most notably with Bifidobacteria and butyrate producing bacteria (26, 176, 207, 215, 226) Though not many viral associations have been made with CRC, one study showed a positive association with a mutant JC virus strain possessing a de letion in the transcriptional control region (190) Proteobacteria have also been shown to posi tively associate with CRC, as they were found to be more abundant in patients with adenomas compared with normal controls (211) Swidsinski et al reported an increase in intracellular mu cosal carriage of
48 E. coli in patients with adenomas compared to healthy controls (226) Addition ally, adherent E. coli have been shown to decrease expression levels of DNA mismatch repair proteins MLH1 and MSH2 (149) AOM treated IL 10 knockout mice exhibit greater tumor multiplicity and invasion when colonized with E. coli compared to colonization with E. faecalis (10) Particular strains of E. coli which posses the pks pathogenicity island responsible for the production of the bacterial toxin colibactin have been shown to induce DNA damage in mammalian cells lines and lead to chromosomal instability (58) In an AOM treated IL 10 knock out mouse model, monocolonization with this strain promoted carcinoma formation. When the pks island was deleted, tumor multiplicity and invasion decreased (10) Using subtractive hybridization to remove common sequences between a non pathogenic K 12 E. coli strain and a colon cancer mucosal isolate, Bronows ki et al identified several pathogenicity islands present in 30 40% of isolates from CRC patients that were previously detected only in uropathogenic E. coli strains (33) Streptococcus bovis has been frequently associated with colorectal cancer, but data are conflicting. Feca l carriage of S. bovis has been reported to be significantly higher in carcinoma patients compared to healthy controls (129) How ever, Potter et al reported no significant difference in the stool carriage of colorectal cancer patients and healthy controls (182) This group also reported no significant associations of S. bovis with malignant colon tissue using culture techniques, but a study using quantitative PCR (qPCR) detected significantly more S. bovis DNA in tumor tissue compared to normal tissue in the same host (1) Rats treated with AOM and either live bacterial cells or cell wall antigens via oral gavage displayed a higher incidence of hyper proliferative ACF
49 (72) This study also showed an increase in polyamines associated with an increase in ACF after S. bo vis administration (7 2) Bacterial cell wall antigens have also been shown to increase the expression of COX 2 in vitro, which, as previously stated, is associated with CRC development (20) Durante Mangoni et al reported a significant association between the prevalence of S. bovis endocarditis and increasing age (70) This suggests that S. bovis associations with CRC may be confounded by age, as age is a well es tablished risk factor for CRC. Other discrepancies between these studies may be attributed to small sample sizes, strain level differences among S. bovis species, or the use of animal models. A recent meta analysis conducted by Boleij et al described a spe cific biotype, Streptococcus gallolyticus ( S. bovis Biotype 1), that is more closely associated with CRC (26) Further studies on this strain may provide more consistent results. Enterotoxin producing B. fragilis has been suggested to contribute to intestinal carcinogenesis through the effects of enterotoxin on e cadherin (233, 254) Nuclear catenin has been observed after treating cell lines with B. fragilis enterotoxin, subsequently leading to increased expression of c myc and cellular proliferation (255) Recent data from the same group suggest a T H 17 dependent mechanism for tumorogenesis (256) B. fragilis was also determined by Wang et al to be enriched in the guts of CRC patients compared to healt hy controls. Another mechanism by which B. fragilis may contribute to CRC development is through the metabolism of heterocyclic amines. Heterocyclic amines are generated by cooking meat well done, and their production and mutagenicity vary based on the pr eparation method and type of meat (216) Few studies have examine d associations
50 between gut microbiota and heterocyclic amines; however, B. fragilis cultured in meat extract resulted in a 2 fold increase in extract mutagenicity when applied to rat liver homogenates (130) These results suggest that the microbiota may have an influence on the genotoxicity of nitrogenous compounds. Bacteroides intestinalis among other members of the normal gut microbiota have been shown to produce sec ondary bile acids (84) Furthermore, colonic mucosal proliferation has been positively associated with secondary bile acid levels (173) In vitro studies have demonstrated the ability of secondary bile acids to up regulate NF expression and induce DNA damage through the generation of reactive oxygen species (119) glucuronidase is an important factor for CRC development. Carcinogens are detoxified in the liver via conjugation with gl glucuronidase activity has been shown to be increased in CRC patients compared to healthy controls (127) Sulfate reducing bacteria reside in the gut as part of the normal microbiota. Though sulfate reducing with CRC, the production of hydrogen sulfide has been implicated in CRC development. Hydrogen sulfide was determined to be more abundant in subjects who had previously undergone surgery for CRC a nd developed new neoplasia compared to healthy controls (125) Hydrogen sulfide was also shown to induce DNA damage mediated by oxidative free radicals. Additionally, COX 2 expression was increased as a result of hydrogen sulfide treatment (13) These observations represent two possible mechanisms of the carcinogenic potential of hydrogen sulfide. Furthermore, thiosulfate sulfurtransferase
51 (TST), and important enzyme for detoxifying hydrogen sulfide, was fou nd to be decreased in CRC tissue. In vitro studies showed the abundance of TST was increased in HT 29 cells when cultured with butyrate, suggesting a possible mechanism by which butyrate inhibits CRC development (186) Butyrate producing bacteria are thought to have a protective effect against CRC. The phylogenetic distribution of butyrate producing bacteria lies mostly within Clostridium clusters IV and XIVa, also known as the Clostridium leptum and coccoides groups, respectively (16) Fermentation substrates are obtained from DFs, including inulin and resistant starches that escape host digestion (140, 158) The production of butyrate is primarily accomplished through the reaction of acetate with butyryl CoA via the butyrl CoA: acetate CoA transferase enzyme, but may also be accomplished by replacing CoA with a phosphate group via phosphotransbutyrylase followed by subsequent dephosphorylation via butyrate kinase which has been shown to have both kinase and phosphotase activity (146) In rodent models, administration of both butyrate producing bacteria and butyrate metabolic precursors have been shown to increase colonic levels of butyrate and decrease precancerous lesions (124, 176) Okhawara et al also observed increased glucuronidase activities in fecal cultures, both of which are features associated with protection against CRC. Other protective effects of butyrate against CRC have been proposed to arise through various mechanisms. One such mechanism involves the promotion of apoptosis through the Wnt signaling pathway (28) A linear relationship has been demonstrated between Wnt signaling and apoptosis in CRC cell lines exposed to butyrate (139)
52 Bordonaro et al propose that these data which conflict with numerous experiments s howing an increase in CRC risk as a result of increase Wnt signaling represent a situation in which low levels of signaling result in controlled self renewal, moderate levels promote uncontrolled proliferation, and high levels induce apoptosis. This idea (27) Members of the gut microbiota that produce butyrate may contribute to cell cycle regulation via this mechanism. Further studies a re needed to confirm this hypothesis. Other mechanisms by which butyrate may prevent CRC development are through epigenetic modification. Butyrate is a known histone deacetylase inhibitor and has been shown to promote DNA demethylation of pluripotency ass ociated genes in fibroblasts (153) Another study on CRC cell lines demonstrated the ability of butyrate to promote differentiation, though this was attributed to the interacti on of butyrate with signaling cascade proteins (178) Butyrate has also been shown in tumor cell lines to both increase and decrease genomic methylation, depending on the cell line (55) Means of Influencing Gut Microbiota Co mposition Probiotics have been defined by the World Health Organization and Food and Agriculture Organization of the United Nations a dministered with prebiotics defined as selectively fermented ingredients that allow specific changes, both in the composition and/or activity in the gastrointestinal microflora, that confer benefits upon host well being and health are referred to colle ctively as synbiotics (192) Clinical trials using synbiotics have shown promising results in reducing CRC risk. A study in which polypectomized patients were administered oligofructose enriched inulin, Lactoba cillus rhamnosus, and
53 Bifidobacterium lactis showed a reduction in fecal water genotoxicity (183) Bifidobacterium longum and Lactobacillus acidophilus have also been shown to decrease fecal water genotoxicity (172) Other studi es in which cell lines were incubated in vitro with lactic acid bacteria and/or prebiotics demonstrated a decrease in fecal water genotoxicity (36) Clinical data are lacking regarding the prevention of CRC lesion development as a function of probiotic and synbiotic use as these data would require large prospective cohort studies ove r long periods of time. Animal models provide evidence for the mechanisms by which probiotics and prebiotics reduce CRC risk. B. longum has been shown in rats to decrease tumor incidence, AOM induced cell proliferation, the activity of ornithine decarboxyl ase an important enzyme for the polyamine synthesis and the expression of ras p21 oncoprotein (215) In the same model it was shown that both L. rhamnosus GG and B. lactis in combination with inulin and oligosaccharides prevented the AOM induced suppression of NK cell like activity, stimulated to production of IL 10, and suppressed (193) A significant reduction in tumor incidence an d an increase in cecal propionate and butyrate have been demonstrated using the prebiotic alone and in combination with the probiotic (81) B. lactis when co administered to rats with L. acidophilus and resistant starch, has been shown to increase the acute apoptotic response to a genotoxic carcinogen and is associated with lower pH and coliform counts (141) Galactooligosaccharides (GOS) a re prebiotic dietary fibers consisting of 2 to 8 unit galactose chains with a terminal glucose. Studies in humans examining direct effects of GOS on CRC are lacking, but many studies have shown the ability of GOS to enrich
54 groups of colonic bacteria negati vely associated with CRC. Davis et al demonstrated a dose dependent increase in various species of Bifidobacterium as a result of GOS supplementation (63) A dose dependent effect was also observed in a study in which the GOS were produced using a probiotic Bifidobacterium strain, and this effect was stronger than that of another commercially available GOS supplement (68) An other study demonstrated a bifidogenic effect of not only GOS, but also with fructooligosaccharides and resistant starch (29) van Dokkum et al showed that GOS increased levels of fecal acetate and decreased glucuronidase activity, which suggests a mechanism by which GOS are associated with a decrease in CRC risk (241) Studies conducted by Gopal et al both in vitro and in vivo demonstrated the ability of GOS to stimulate the growth of both lactobacilli and bifidobacteria. This effect on lactobacilli was greater in subjects who began treatment with lower baseline levels of these organisms (90, 91) Another study demonstrated not only the bifidogenic effect of GOS, but also its ability to decrease E. coli and Salmonella enterica serotype Typhimurium adhesion to HT 29 cells (238) GOS were also shown in this study to increase lactobacilli, certain butyrate producing bacteria, and to decrease E. coli Though an increase in bifidobacter ia was seen in the placebo group, a larger increase was seen in the GOS treated group that was significantly higher than the increase in the placebo group. The same study showed an increase in both NK cell activity and anti inflammatory cytokine production while pro inflammatory cytokines were reduced (245) Thou gh mechanistic studies are difficult to conduct in humans, a rat AOM model showed that aberrant crypt multiplicity, but not numbers of ACF, was reduced in rats consuming relatively high amounts of GOS (250)
55 Conclusions Multiple lines of evidence support the notion that gut microbiota can contribute to colorectal carcinogenesis. Diet, particularly dietary fiber intake, which has been associated with re duced CRC risk, affects microbiota composition. Thus, changes in microbiota might represent important mechanisms through which diet reduces CRC risk. Various bacteria have been linked with experimental carcinogenesis in animal models or correlated with CRC in human observational studies. However, causality has not been established in humans. Interventional studies that would introduce potential pathogens in humans are infeasible for ethical reasons. If these bacteria are causally associated with CRC in huma ns, they should be detected at early stages in which polyps have not fully developed into cancer. Multiple microbiota based studies suggest differences in mucosa associated and luminal bacteria in subjects with CRC. Various pro and prebiotic interventions aimed at modifying gut microbiota toward a more beneficial composition have been successful in reducing CRC risk markers, gut inflammation, and preneoplastic lesions. A lthough no single microbial agent has yet been shown to be causally linked to CRC, cont ributions of the gut microbiota to colorectal carcinogenesis are evident. It can be expected that further advances in the field can soon be translated into the development of microbiota based CRC screening and prevention regimens. A more detailed understan ding of how microbiota can be manipulated also might point toward novel means to minimize the detrimental effects of surgery, radiation, and chemotherapy treatment on gut health. Due to the established uniqueness of each individual's microbiota, personaliz ed microbiota manipulations, possibly based on the underlying dominant enterotype, will likely offer the best chance for success.
56 CHAPTER 3 MATERIALS AND METHOD S Stool Sample Collections Subjects collected their own stool s amples which were placed on ice and delivered to the lab within four hours of defecation. The samples were processed immediately upon arrival. Samples were homogenized by k neading in a strong plastic bag and aliquoted for long term storage at 80 C and short term storage pending molecul ar assays at 20 C. Biopsy Collections During the colonoscopy, the attending physician collected multiple biopsies from the ascending colon. If the ascending colon could not be reached, the biopsy was taken from the most proximal location. One biopsy sampl e was placed into a tube of RNA later (Ambion, # AM7024) and fixed overnight at 4 C. Samples in RNA later were then stored at 20 C pending analysis. The remaining biopsy samples for each patient were divided ion (10% glacial acetic acid [MP Biomedicals, #193829] 30% chloroform [Acros, #67 66 2], 60% 200 proof ethanol [Acros, #64 17 5]) and fixed overnight at 4 C Biopsy samples were then transferred to 70% ethanol and submitted to the University of Florida Co llege of Dentistry Molecular Pathology Core lab for microscope slide preparation. Biopsy samples were embedded in paraffin blocks. F our slides in total were prepared 2 unstained slides and 2 slides stained with Hematoxylin and Eosin. Microbiota Analysis A pea sized amount of each stool sample was washed twice in 1 mL sterile phosphate buffered saline (PBS). PBS was prepared with 8.00 g/L sodium chloride
57 (NaCl) (Fisher Scientific, #S671 500), 2.68 g/L potassium chloride (Fisher Scientific, #BP366 500), 14. 20 g/L sodium phosphate dibasic (MP Biomedicals, #194846), and 2.40g potassium phosphate monobasic (Fisher Scientific, # S374 500) adjusted to pH = 7.4 with hydrochloric acid (Acros, #7647 01 0). Bacterial genomic DNA was isolated from fecal samples using the QIAamp DNA Stool kit (Qiagen, #51504) per manufacturer instructions with an added 3 minute bead beating step (3450 oscillations/min) after cell lysis (Zirconia/Silica beads, BioSpec Products, #110791012 ; Mini BeadBeater, Biospec Products #607 ), which efficiently lyses most bacterial cells with little bias (162) Heat resistant 2 mL microcentrifuge tubes were used during the bead beating step to prevent melting or warping of the tubes (Shrstedt #D 51588). Kit buffers were reconstituted using 100% ethanol (Acros, #64 17 5). To detach bacteria from biopsy samples, tissues were first subjected to vigorous agitation in 500 mL PBS using 1mm glass beads. Biopsy tissues were removed and stored for lat er analyses. Bacterial genomic DNA was extracted from the PBS wash via the same method used for stool extractions. Denaturing Gradient Gel E lectrophoresis (DGGE ) A nalysis DGGE profiles were generated as an initial quality control for DNA extraction and as a crude tool for judging diversity in each sample. A 457bp fragment from the V6 to V8 region of the bacterial 16S rDNA was amplified with primers U968 GGG CGC GCC CCG GGC GGG GCG GGG GCA CGG GGG GAA CGC GAA GAA GT GTA CAA GAC CC) as described by Zoetendal et al. (259) PCR reagents were obtained from Qiagen (#201205). Reactions included 1 mM magnesium chloride (Qiagen, #201205) and 4% formamide (Acros, #75 12 7). 15 2 0 L of amplified DNA mixed with an equal volume of loading dye was added to each lane. Loading dye (Bio Rad, DCode Kit) was diluted to 10% in glycerol (70%, Fisher
58 Scientific, #G33 500) and sterile deionized water (20%). DGGE profiling was performed on an 8% (w/v) acrylamide (Acyrlamide/Bis solution, 37.5:1, Bio Rad, #161 0148) gel with a gradient ranging from 40% at the top to 50% at the bottom. 100% denaturing conditions were defined as 7 M urea (Bio Rad, #161 0730) and 40% formamide (Amresco, #0606 500) Polymerization was achieved using ammonium persulfate (0.09% w/v, Bio Rad, #161 tetra methyl ethalinediamine (0.09% v/v, Bio Rad, #161 0800). Gels were run at 60 C for 16 hours at 65V in 0.5 X TAE buffer (Bio Rad, #161 0773) and stain ed with SYBR Gold (Life Technologies, #S11494). Images of the stained gels were scanned with Quantity One software (Biorad) and analyzed with Diversity Database software (Biorad). Quantitative PCR qPCR analysis was performed in duplicate using a QuantiTec t SYBR Green PCR kit (Qiagen, Cat. No. 204143) on a Stratagene MX3000P (La Jolla, CA). Reactions were performed in a final volume of 12.5 l with10 ng DNA template 0.2 M of each primer (Invitrogen). The following primers and annealing temperatures were us ed: 1) eubacteria (V3 F: 5' CCTACGGGAGGCAGCAG 3'; R:5' ATTACCGCGGCTGCTGG TCGCGTC(C/T)GGTGTGAAAG CCACATCCAGC(A/G)TCCAC AGCAGTAGGGAATCTTCCA ATTYCACCGCTACACATG 58C). Standard curves for quantification were constructed using genomic DNA extracted from Bifidobacterium adolescentis (ATC15703D) for bifidobacteria and VSL #3 mix (Sigma Tau Pharmaceuticals, Inc.) for LAB and eubacteria
59 Units are expressed as genome equivalents (1 genome equivalent = mass of 1 genome) per 10 ng DNA, which are derived as follows from the average mass of bifidobacteria and lactic acid bacteria genomes (2.3 Mb): Input mass of DNA for standard curves is converted t o genome equivalents using the multiplicity constant 3.97 X 10 5 derived below: 454 based 16S rD NA S equencing 16S rRNA a mplicons were generated using a published barcoded primer set (100) .The PCR reactions included an initial melting step at 95C for 2 minutes and 30 cycles of 95C for 30 seconds, 55C for 30 seconds, and 68C for 1 minute. Reactions included 1 mM magnesium chloride (Qiagen, #201205) and 4% formamide (Acros, #75 12 7). Sequences with a quality score of less than 20 or with a leng th of less than 150 nucleotides were removed from the analysis. Sequences were initially analyzed using the Ribosomal Database Project (RDP) pyrosequencing pipeline that included features to calculate diversity indices and rarefaction curves. Sequence read s were binned using ESPRIT into OTUs at various similarity level s (223) The QIIME package was used to calculate i) Chao rarefaction diversity, which estimates how many OTU are present in a sample, and ii) UniFrac distances (40, 147) that allow for a comparison of the distribution of OTU among samples.
60 Metagenomic S equencing Bac t erial DNA was isolated using th e GNOME DNA Kit (MP Biomedicals #2010 400 ) per manufact u a ten minute bead beating step (apparatus and materials described above) was used to lyse remaining intact bacterial cells. 15 mg of Polypvinylpolypyrrolidone (PVPP) (Fluka analytical, #9003 39 8) was added to each sample. Samples were mixed and centrifuged for 3 minutes ( Eppendorf 5804R; all centrifugation steps were performed at 20,000 g). The supernatants were removed and the p ellets were washed three times in TNAP buffer (50 mM Tris (Fisher Scientific, #BP152 500), pH 8, 20mM EDTA (Fisher Scientific, #O2793 500) pH 8, 100 mM NaCl, 1% PVPP) Supernatants were pooled. DNA was pelleted from 750 L of pooled supernatant using 1 mL isopropanol (Acros, #67 63 0) for 30 minutes at 20 C. The pellet was dissolved in 400 L water and the Ethanol precipitation occurred at 80 C for one hour. B acterial genomic DNA sheared ran domly via sonication from eight cases with hig h risk polyps was pooled in equal molar amounts as was DNA from eight controls. Both DNA pools were then submitted to 454 pyrosequencing using one half plate for each DNA pool. Sequence reads were deposited and analyzed in MG RAST using the GenBank reference database with 50% identity and a minimum alignment length of 40 bases. 233,141 sequence reads with an average of 180 bases per read were obtained fo r control subjects and 228,814 sequence reads with an avera ge of 176 bases were obtained for case subjects.
61 Fecal pH Approximately 1 g of homogenized stool was added to pre weighed 15 mL conical tubes. Exact stool weight was obtained on a digital scale and a 1:10 w/v dilution with sodium chloride (0.155M) was agi tated with 4 to 5 glass beads (1 mm) to break up particles. Samples were centrifuged (Sorval legend RT+) for 5 minutes at 3716 rcf. Supernatants were measured using a glass electrode and Ultrabasic pH meter (Denver instruments). Statistical Analysis For D GGE profile analysis background was subtracted from each lane and the profiles were subjected to Gauss modeling. The similarity matrix was calculated based clustering meth od (Ward, UPMA). Shannon diversity indices were calculated for each sample using the total number of bands per sample and their relative intensities. For qPCR assays, genome equivalents were used as the unit of comparison in an attempt to correct for diff erences in genome size and copy numbers of 16S rRNA genes. The Student's t test was used to calculate differences in genome equivalents between cases and controls either for raw or log transformed data when the data were normally distributed The Mann Whi tney U test was used on non normally distributed data. For analyzing differences in the prevalence of OTUs, a z score was obtained using a chi square based test For exploratory purposed we used a p<0.05, we corrected for multiple analyses by using a P<0.0 1 and/or FPR rate of 5% ( =0.05). For generating heat maps the most signi fi cantly different OTUs were chose n based upon Z score In the case of more that more than 20 OTUs were significantly enriched, only the 20 with
62 the lowest p values were included in t he heat map The QIIME package was used to calculate p values for differences in UniFrac distances.
63 CHAPTER 4 GUT MICROBIOTA PATTE RNS ASSOCIATED WITH COLORECTAL POLYP PREVALENCE Differences in Diet and Intestinal Microflora: Potential Associations with Increased CRC Rates in African Americans Introduction The work presented first is based on the hypothesis that distortions in gut microbiota composition are associated with colorectal polyp prevalence. A dult s aged 40 years or older who were scheduled to un dergo a screening colonoscopy were recruited It was important to recruit subjects in this age group because the risk of developing CRC and the incidence of polyps increase with age. Furthermore, it is necessary to use subjects only referred for screening as other underlying diseases may be associated with microbiota composition M icrobiota was analyzed using DGGE, qPCR, 16S rRNA pyrosequencing, and metagenomic shotgun sequencing Study Design Volunteers scheduled to undergo a screening colonoscopy were rec ruited from university hospital associated endoscopy clinics in Baltimore, MD and Gainesville, FL. The study was approved by Institutional Review Boards at University of Maryland and the University of Florida. Upon consent to participate, subjects provided dietary and lifestyle information via a demographic questionnaire, food frequency questionnaire (NIH Block 98.2), and the National Institute of Health meats module questionnaire. A 4 day food record reflecting the four days of food and fluid intake prior to stool sample collection was collected by the subjects. Each subject gave written informed consent, and all study procedures were in accordance with the ethical standards of the University
64 of Maryland and the University of Florida Institutional Review Bo ard. Inclusion criteria included the following: Referred for a screening colonoscopy Age of 40 years or older Able to provide informed consent for study procedures Able to provide information on dietary history and other demographic factors (e.g., famil y history) African American of Caucasian American Exclusion criteria included the following: Known colorectal cancer Known or suspected inflammatory bowel disease Acute or chronic diarrhea Short gut syndrome Prior resection of small intestine or ile ocecal valve Known small or large bowel obstruction Hospitalization within the past 4 weeks Current hospitalization for 3 days or longer Use of systemic antibiotics (oral, intravenous, or intramuscular) within the past 4 weeks Diagnosis of malignancy (other than skin cancer within the past five years Chemotherapy treatment within the past three years Colonoscopy and pathology reports were obtained to determine the presence of polyp s their size s and location s Pathology reports were obtained to dete rmine polyp histological classification. Non neoplastic polyps, those commonly thought to be without
65 malignant potential, include hyperplastic polyps and lymphoid aggregates. Neoplastic polyps, or polyps with malignant potential, are mainly adenomas (54) Subjects with polyps were further classified into high risk or low risk groups based on polyp number and size Specifically high risk subjects were defined as those with total polyp diameters equal to 10 mm or greater from one or multiple polyps since this allowed for a nearly equal distribution of subjects between low risk and high risk groups Conventionally, high risk subjects have been defined as those with at least one polyp with a diameter of 10mm or greater (222) Because in this study only few subjects fit this criterion, decreasing s tatistical power, the broader criterion was used 157 subjects consented to the study protocol. Case status was reported for 126 subjects. Any subject displayi ng at least one polyp was considered to be a case subject. S ubjects with out reported case status were excluded Results Study population demographics and dietary habits 115 subjects completed demographic questionnaires (Table 4 1). Control subjects reporte d significantly more exercise than high risk subjects and were less likely to have had polyps removed at previous colonoscopies than both high risk case subjects and all case subjects. No statistically significant differences in dietary habits were detecte d between all cases and controls, but high risk cases consumed significantly less dietary fiber, carbohydrates, fruit, copper, magnesium, potassium, vitamin E and vitamin C than controls (Table 4 2). A As consumed smaller amounts of both dietary and supple mental vitamins and minerals compared to C As A As also consumed smaller amounts of total dietary fiber,
66 and specifically fiber from grains, v itamin E, magnesium and potassium Likewise, A As ate fewer servings of fruits and vegetables per day. Fecal micro biota community diversity DNA was extracted from all available stool and biopsy samples. Initial DGGE profiling confirmed that DNA extractions were successful for all samples and that the samples contained a diverse microbiota community No differences wer e detected in the prevalence of specific DGGE bands between cases and controls, or between the two racial groups in n either stool nor biopsy samples (data not shown). Similarly, no significant differences between case groups were detected for the DGGE base d Shannon and Simpson diversity indices (Figure 4 1). Compared to stool samples, b iopsy sample s exhibited much less diversity (data not shown). Quantification of targeted bacteria A qPCR analysis was performed to compare the amounts of lactic acid bacteri a and bifidobacteria, which are generally thought of as beneficial members of the gut microbiota. No statistically significant differences were detected between cases or controls in both stool (Figure 4 2) and biopsy (Figure 4 3) samples, even after strati fying by body mass index (BMI) or between racial groups. When stratified by BMI independent of case status, no significant differences in either bifidobacteria of lactic acid bacteria abundances were observed in either stool or biopsy samples (Figure 4 4). However, there was a significantly higher abundance of bifidobacteria in stools samples from AA subjects compared to CA subjects independent of case status but this was not seen in biopsy samples (Figure 4 5). Furthermore, in biopsy samples the abundance of lactic acid bacteria was significantly great er than the abundance of bifidobacteria, while in stools samples bifidobacteria were significantly more abundant than lactic acid
67 bacteria. These results further demonstrate the differences between stool and biopsy microbiota communities. 454 pyrosequencing analysis of 16S rDNA 16S rRNA based 454 pyrosequencing was used to analyze 30 case and 30 control samples from stools, and 16 case and 16 control samples from biopsies. After removal of low quality and sho rt reads, a total of 209850 sequence reads with an average length of 216 nucleotides were retained from stools and 42150 sequence reads with an average length of 211 nucleotides were retained from biopsies. Less DNA was sequence d for biopsy samples because of the lower diversity compared to stool samples based on DGGE analysis ESPRIT was used to bin sequences into a total of 1927 OTUs at the 95% similarity level and 6321 OTUs at the 98% similarity level. OTUs were blasted against the RDP database to obtain taxonomic identities. UniFrac, a tool for comparing microbial community diversity, was used to determine if bacterial communities in stool samples differed from those in biopsy samples (99, 223) The unweighted approach was used, which considers only the abs ence or presence of each OTU, as opposed to the weighted approach, which also considers OTU abundance and can be strongly i nfluenced by a few abundant OTU s. Chao1 based rarefaction curves did not significantly differ between case groups, suggesting that ov erall diversity is not significantly affected by case status. (Figure 4 6). As community structures in these two sets of samp les clearly differed (Figure 4 7 a) stool microbiota were analyzed separately from biopsy adherent microbiota When cases were compa red to controls, no differences in UniFrac diversity were dete cted in either stool (Figure 4 7 b) or biops y samples (Figure 4 7 c). There also were no detectable differences in UniFrac diversity between high risk subjects and controls (data
68 not shown) or be tween AAs and CAs (Figure 4 7 d). No significant differences were detected in phylum abundance between cases and controls in both stoo l and biopsy samples (Figure 4 8 ). I n both stool and biopsy samples m ore OTUs were significantly more prevalent in control rather bacteria, rather than the presence of a pathogen, correlates with polyp prevalence To maximize our ability to detect differences, c ontrol samples were compared with all cases, hi gh risk cases as defined in this study (the sum of polyp diameters equaling or exceeding 10mm), and high risk cases as defined by convention (presence of a single polyp with a diameter of 10mm or greater). Figure 4 9 shows the 20 most significantly affect ed OTUs by case status between all cases and controls. Figure 4 10 represents the same analysis between only high risk cases (defined in this study) and controls. Figure 4 11 shows the analysis with high risk cases redefined by convention. The majority of the OTUs detected in these analyses matched within the class Clostridia. C omparisons at the 98% similarity level revealed similar results in that most significantly affected OTUs belonged to the class Clostridia. Figure 4 12 shows the 20 most significantl y affected OTUs by case status between all cases and controls. Figure 4 13 represents the same analysis between only high risk cases (defined in this study) and controls. Figure 4 14 shows the analysis with high risk cases redefined by convention. For biop sy samples, many OTUs that were significantly affected by case status again matched within the class Clostridia; however, some OTUs matched within other phyla including Bacteroidetes and Proteobacteria. The c omparison between controls
69 and all cases at the 95% similarity level is represented in Figure 4 15 Figure 4 16 shows a similar analyses considering only high risk cases defined in this study. Figure 4 17 shows the analysis between controls and all cases at the 98% similarity level Controls and high r isk cases are compared at the 98% similarity level in Figure 4 18. No comparison between controls and high risk subjects defined conventionally was made for biopsy samples because only two case subjects belonged to that category. Differences in OTU preval ence appeared stronger (based on lower p values) when high risk polyp cases were compared with controls giving stronger evidence for those OTUs to tr uly be associated with CRC risk The abundance of OTUs corresponding to buty rate producing bacterial gener a as described by Antharam et al, were compared in both stool and biopsy samples. No ne of the comparisons (controls verses all cases, controls verses high risk cases, and controls verses conventionally defined high risk cases) showed any significant diffe rences in the presence of butyrate producing bacteria (data not shown) (8) V arious OTUs differ ed in prevalence between AAs and CAs These OTUs, similar to those found to be significantly enriched or depleted based on case status, matched mainly to members of the Clostridia class; however, these particular OTUs were different than the significant OT Us observed between case groups. Therefore, although there are differences in gut microbiota composition between race groups, these differences cannot be attributed to increased burden of CRC in AAs. Figure 4 19 displays those OTUs significantly affected b y race at the 95% similarity level. Figure 4 20 displays the OTUs significantly affected by race at the 98% similarity level.
70 Because no t a single OTU could effectively discriminate between cases and controls, a discriminant analysis approach was applied t o develop a fecal microbiota pattern that would allow for such a distinc tion. At the 92% similarity level, a microbiota signature profile was detected based on 27 OTU's that effectively separated cases from control s (Figure 4 21 a). The 92% similarity level is optimal for discriminant analysis accuracy (224) When the similarity level is lower, sequences are grouped into large clusters where discriminating and non discriminating OTUs may be combined. If th e similarity level is higher, the sequencing depth is not great enough to obtain accurate estimates of microbial composition profiles (54, 224) The formula for the discriminant analysis is as follows, where X 1 through X 15 represent the num ber of reads for each OTU, respectively, and Y is the calculated risk assessment. In the formula below the OTUs correspond as follows: Bacteroides eggerthii (X 1 ), Sutterella stercoricanis (X 2 ), Parasutterella excrementihominis (X 3 ), Oscillibacter valericig enes (X 4 ), Alistipes massiliensis (X 5 ), Lactobacillus manihotivorans (X 6 ), Clostridium methylpentosum (X 7 ), Ruminococcus bromii (X 8 ), Clostridium sp. 14774 (X 9 ), Succinispira mobilis(X 10 ), Lutispora thermophila(X 11 ), Oscillibacter valericigenes(X 12 ), human intestinal firmicute CO35 (X 13 ), Ruminococcus sp. CJ60 (X 14 ), Succinivibrio dextrinosolvens (X 15 ) A positive value for Y indicates a case subject and a negative value indicates a control subject. The corresponding receiver operat or characteristic (ROC) curve yields
71 an area under the ROC curve (AUC) value of 0.81, which indicates the sensitivity/specificity for predicting cases status based on microbiota profile (Figure 4 21 b). Metagenomic shotgun sequencing analysis To expand the analysis beyond the limitations of the 16S rRNA based approach, a pilot metagenomic analysis was performed using shotgun sequencing on the 454 pyrosequencing platform. For this analysis equimolar amounts of genomic DNA were pooled from 8 controls and 8 hi gh risk cases. A total of 233,141 and 2 28,814 were generated, respectively. The two datasets were analyzed using the Metagenomics Rapid Annotation using Subsystems Technology (MG RAST) platform with the Greengenes database An enrichment of sequences with closest matches to Klebsiella, Shigellla, and Citrobacter wa s detected in cases (Figure 4 22 ). Klebsiella was also more prevalent in cases when the analysis was limited to 16S rRNA sequences that were detected in the shotgun datasets. Bacterial starch and sucrose metabolic functions were more highly represented in cases while fructose and galactose metabolic functions were more highly represented in controls (Figure 4 23 ). Discussion The results of this study indicate that there were differences in both li festyle, including exercise and diet habits, and gut microbiota composition between healthy subjects and subjects with polyps. Exercise has been associated with reduced polyp development (52) Approximately half of control subjects, compared to only a fifth of subjec ts with polyps exercise d regularly (p < 0.05) Furthermore, high risk subjects consumed significantly less dietary fiber and fruit, which likewise parallels results found in the literature (82, 189, 244, 252)
72 The apparent differences in bacterial phyla abundance did not reach significance likely because of the large amount of variation between subjects (particularly those with a difference in magnitude of 10% or more) R anges of bacterial abundance in the two predominating bacterial phy la, Bacteroidetes and Firmicutes of 32 to 96% and 0.3 to 61%, respectively, in stools and 30 to 82% and 4 to 58%, respecti vely, in biopsies, were detected. The extensive interpersonal variation in this study has been previously observed others (113) Although not significant, the Firmicutes:Bacteroidetes ratio was higher in stools of control subjects. These results parallel those presented by chen, et al. conducted in CRC patients (47) F urthermore, the abundance of Fusobacterium was greater though not significantly, in cases verses controls in both stool and biopsy samples. Fusobacterium nucleatum have been reported to associate with colorectal tumor tissue (42) Differences between case and control subjects were detectable at the OTU level in both stool and biopsy samples Some OTUs were more prevalent in case subjects while other s were more prevalent in control subjects, which suggests that some OTUs may have pathogenic potential while others may be protective. When only high risk case subjects were considered, many of the previously associated OTUs again reached significance in both stool and biopsy communities OTUs that reach significance in multiple analyses can be considered with more confidence as they are less likely to be spurious results. In this study, the combined analysis of all cases and only high risk cases verses controls was used to help eliminate OTUs that may occur by chance. A common problem when dealing with many variables is the detection of significant differences based on chance. With 6321 OTUs in the data set and an alpha level of
73 0.05, one would expect to find 316 significantly different OT Us by chance. One may reduce spurious results by lowering the threshold of significance or using different statistical tests, such as those based on the t distribution verses the chi squared distribution ; however, there is currently no consensus regarding the appropriate statistical tests or analysis method for sequencing data. The false discovery rate can be used to assign a metric (q value) to a range of p values based on the distribution of p values, where lower p values are assigned a lower value. The q value is generally represented by the height of the p value distribution where the tail becomes flat. P values assigned q values that are lower than this threshold are assumed to be true positives. While m ost OTUs associated with polyp presence were detec ted in both the analyses considering all cases and only high risk cases analyses several OTUs only reached significance in the comparison between high risk case s and control s These results suggest that some associations do not arise until after a certain level of disease progression. OTUs that reach significance only in the low risk population may contribute pathogenic physiology which progresses after the pathogen has bee n eliminated. In stools samples, most significantly different OTUs match within the phylum Firmicutes, particularly within the Clas ses Clostridia and Bacilli. Many of these OTUs are known butyrate producers, but while most butyrate producers were associat ed with contr o ls few were associated with cases. Therefore, they cannot be firmly linked with protection against polyps. Whil e some significantly different OTUs in biopsy samples also matched within these groups, there were more matches to organisms in th e phyla
74 Bacteroidetes, Proteobacteria, and Fusobacteria compared with stools These OTUs did not correspond with organisms discussed from the literature in Chapter 2 with the exception of one OTU which was associated with case subjects in biopsies at the 98% similarity level (26, 176, 207, 215, 226) This association is consistent with the results from Swidsinsk,i et al. that dem onstrate increase mucosal carriage of E. coli in CRC patients (226) Though some OTUs did match to streptococci groups, they associated in some analyses with case subjects and in others with control subjects ; therefore results implicating S. gallolyticus in CRC were not supported in this study (26) The most likely reason for th ese inconsisten cies is that this work examined differences be tween subjects with polyps and healthy controls verses subjects with CRC and healthy controls. Because only roughly 5% of subjects with polyps develop CRC, it is not surprising that these results did not parallel those in the CRC literature; however, this particular risk factor is optimal for this prospective cohort study because roughly 95% of CRCs develop from polyps (115) Future studies should examine a much larger samples size and follow the subjects for several years to determine which subjects develop CRC Across the comparisons of individual OTUs stratified by sample site and risk, there were OTUs belonging to the groups Oscillospiraceae, unidentified Firmicutes, unknown butyrate producing bacteria, and Alistipes that consistently associated with control subjects. These results confirm previous findings in which OTUs matching to these groups associated with healthy subjects verses subjects with CRC (248) The inverse association of these OTUs with polyp incidence and CRC suggest that these groups
75 may be protective against CRC development. These OTUs may provide such a benefit through th e production of butyrate (176) Some OTUs showed contradicto ry results. For instance, at the 98% similarity level, some were associated positively with polyp prevalence in biopsy samples but inversely in stool samples suggesting that some OTUs may be harmful when in direct contact with the mucosa but beneficial wh en present in the lumen. At the 95% similarity level, there were no shared OTUs between stool and biopsy communities that associated with case status. There are both advantages and disadvantages to analyzing stool and biopsy communities. The advantage of a nalyzing stool samples is that they can be obtained non invasively; however, analyzing stool samples is disadvantageous in that they contain luminal organisms and not those closely associated with the epithelium, where polyps develop. This work has shown a clear distinction between luminal and mucosa adherent microbiota communities. The organisms in direct contact with the epithelium may have a stronger influence on epithelial cell physiology via stimulation of TLRs on epithelial cells o r activation of dend ritic cells, and thus may represent better candidates for future studies. However, the biopsies in this study were not obtained under physiological conditions. Patients preparing for colonoscopies must undergo a cleansing prep which can remove mucosally as sociated microorganisms. Therefore, in this study stool samples may be more appropriate in that only they can be used for developing a non invasive screening method. Furthermore, stools samples are more representative of normal physiological conditions, as they were collected before patients began their cleansing prep.
76 The comparison between racial groups at the OTU level revealed that many OTUs are associated with racial group; however, most of these racial group associated OTUs did not correlate with poly p prevalence or with organisms previously associated with CRC There is no clear evidence based on OTU distribution that the increased incidence of CRC in AAs is associated with the differences in gut microbiota composition of AAs While the abundance of b ifidobacteria determined by qPCR, was higher in AAs, there was no association between bifidobacteria counts and case status. These results suggest that the increased CRC rate in AAs verses CAs is not microbiota dependent. Further studies examining genetic predispositions to CRC between races may provide better insight. Upon determining differences by case status at the OTU level, a predictive model based on OTU abundance was able to discriminate cases and controls. Several of these OTUs match to groups pr eviously found to be inversely associated with CRC ( Alistipes Oscillobacter unknown Firmicutes) (248) This model was verified using the leave one out cross validation method, in which a model is constructed with a training data set and tested with a validating datum. The training data set includes all but one of the experimental data, while the validating datum is the one excluded This validation is conducted for each individual datum of the experimental data set and t he results are averaged to produce a single model. This model could potentially be used clinically as a preventative s creening method. Further studies with deeper sequencing technologies and larger samples sizes are needed to improve sensitivity and specificity. Differences by case status in stools were also seen using a n exploratory shotgun sequencing approach, though t he bacterial groups associated with p olyp prevalence
77 were different than those detected via 16S sequencing. However, the enrichment of E. coli in the stools of case subjects compliments the significant association of E. coli in biopsies with high risk poly p incidence. The differences in methodologies may account for the ability to detect significant associations with luminal E. coli in this analysis. Metagenomic sequencing may provide better results in that there is no PCR bias, where as 16S sequencing reli es upo n primer sets that may not be truly universal. Polymorphisms in the generally conserved regions of 16S gene bacterial DNA methylation patterns, or DNA binding proteins carried throughout the DNA extraction procedure may influence primer affinity. Fu rthermore, metagenomic sequencing allows for functional assignment verses strictly taxonomic assignment. However, due to the severely limited sample size and the pooling of samples in this analysis, further metagenomic sequencing using more samples without pooling is needed to confirm results. In conclusion there are discernible differences in gut microbiota composition between AAs and CAs, but these differences are not consistent with the differences detected based on case status. Additionally, it was sh own that there are large differences between stool and biopsy communities; however, the analysis of biopsy communities is compromised because all subjects were required to perform a colon cleansing prep prior to their colonoscopy. Therefore, the stool base d analysis may better represent normal physiological conditions F inally these results indicate that a stool based test based at the OTU level is feasible for estimating polyp presence This work may have implications in earlier detection of CRC and thus lower the morbidity and mortality of this disease.
78 Table 4 1. Demographic and lifestyle data by case status. Parameter Study populatio n (N = 115) Controls ( n = 61) All cases (n = 54) High risk cases (n = 24) Mean age 60 9 60 9 59 9 62 10 Gender Female 77 (67) 45 (74) 32 (59) 12 (50) Male 38 (33) 16 (26) 22 (41) 12 (50) Race African Americans 30 (26) 18 (29.5) 12 (22) 4 (17) Caucasian Americans 82 (71) 40 (65.5) 42 (78) 20 (83) Other 3 (3) 3 (5) 0 0 Bo dy mass index (BMI) Mean BMI 29 8 30 10 29 6 28 5 Healthy weight 34 (30) 18 (29) 15 (28) 5 (21) Overweight 43 (37) 22 (36) 22 (40) 13 (54) Obese 37 (32) 21 (35) 16 (29) 5 (21) BMI Not reported 1 (1) 0 1 (3) 1 (4) Education Some high school 13 (11) 6 (10) 7 (13) 3 (13) H igh school diploma 36 (31) 18 (30) 18 (33) 8 (33) Some college 63 (55) 35 (57) 28 (52) 12 (50) Education n ot reported 3 (3) 2 (3) 1 (2) 1 (4) In come per year Less than 30,000 37 (32) 19 (31) 18 (33) 8 (33) 30,000 to 50,000 30 (26) 17 (28) 13 (24) 5 (21) Greater than 50,000 46 (40) 24 (39) 22 (41) 10 (42) Education n ot reported 2 (2) 1 (2) 1 (2) 1 (4) P reviously detected polyps Polyps detected and removed 42 (37) 11 (18) 31*** (57) 16** (67) Polyps not detected 67 (58) 49 (80) 18 (33) 6 (25) Polyps n ot reported 6 (5) 1 (2) 5 (10) 2 (8) Exercise (45 minute activities) At least 3 times per week 47 (41) 30 (49) 17 (31.5) 6 *(25) Fewer than 3 times per week 64 (56) 31 (51) 33 (61) 16 (67) Exercise n ot reported 4 (3) 0 4 (7.5) 2 (8) Anti inflammatory use Daily 51 (44.5) 30 (49 ) 21 (39) 11 (46) Weekly 16 (14) 8 (13) 8 (15) 2 (8) Rarely 37 (32) 18 (29.5) 19 (35) 9 (38) Never 7 (6) 4 (6.5) 3 (5.5) 0 Anti inflammatory use n ot reported 4 (3.5) 1 (2) 3 (5.5) 2 (8)
79 Table 4 1. Continued Para meter Study population (N = 115) Controls ( n = 61) All cases (n = 54) High risk cases (n = 2 4) Laxative use Daily 2 (1.5) 2 (3) 0 0 Weekly 7 (6) 5 (8) 2 (4) 1 (4) Rarely 42 (37) 18 (30) 24 (44) 10 (42) Neve r 61 (53) 36 (59) 25 (46) 12 (50) Laxative use n ot reported 3 (2.5) 0 3 (6) 1 (4) Parentheses indicate the percentage of the responding study population with the given characteristic. BMI is defined as weight in kg divided by height in meter s squared. = p < 0.05, ** = p < 0.001, *** = p <0.0001 compared to the control group. indicates the standard deviation.
80 Table 4 2. Dietary differences between case groups and racial groups. Dietary variable Controls (n = 48) Cases (n = 41) High risk cases (n = 23) AA (n = 25) CA (n =63) Protein (g) 58.8 31.3 59.0 32.4 51.3 20.3 51.4 33.1 61.1 30.5 Total fat (g) 64.1 34.7 69.4 41.8 58.2 24.2 57.4 38.3 69.7 37.8 Saturated fat (g) 18.6 10.4 20.7 13.3 17.3 7.9 16.3 11.7 20.7 11.8 Carbohydrates (g) 192.3 96.6 175.1 86.6 146.1* 53.3 188.4 114.9 182.1 83.0 Fiber (g) 16.4 9.2 12.8 7.1 10.4* 4.4 11.9 7.7 Fiber from fruits and vegetables (g) 7.5 5.2 5.7 4.1 4.6* 2.8 4.9 4.0 7.4 4.9 Daily fruit servings 1.3 0.9 1.1 0.9 0.8* 0.7 0.9 0.7 Vitamin C (mg) 94.7 50.2 79.6 64.9 52.8** 33.5 77.9 59.8 57.7 57.2 Vi tamin E ( tocopherol eq) 8.9 4.9 8.3 4.7 6.6* 2.9 6.8 5.0 Folate (mg) 334.6 178.6 282.7 130.6 248.5* 114.5 286.4 194.6 318.7 145.0 Magnesium (mg) 278.7 137.7 239.5 114.7 201.7* 70.8 212.6 127.9 Pota ssium (mg) 2539.4 1167.7 2315.4 1221.8 1919.2* 627.4 1779.8 1144.2 Copper (mg) 1.3 0.8 1.1 0.6 0.9* 0.5 1.1 0.6 1.3 0.7 Data are presented as means plus or minus the standard deviation. One subject did not provide info rmation on race = p < 0.05, *** = p < 0.0001 compared to the control group. = p < 0.05 compared to AA group.
81 Figure 4 1. Diversity indices in stool communities between cases and controls. Error bars represent the standard deviation. Total n = 114 ( 57 cases, 57 controls). 0 2 4 6 8 10 12 14 16 18 20 Shannon Inverse Simpson Index value Diversity Index Controls Cases
82 Figure 4 2. qPCR results showing no significant differences between demographic groups for lactic acid bacteria and bifidobacteria in stools Error bars represent the standard deviation. BMI is defined as weight in kg divided by height in meters squared. All subjects for whom demographic data were provided are included. Eight subjects did not provide BMI data. Two subjects did not provide race data. Total (n = 111; 55 controls, 56 cases [29 high risk cases]), BMI < 2 5 (n = 30; 15 controls, 15 cases [7 high risk cases]), BMI 25 30 (n = 40; 12 controls, 22 cases [14 high risk cases]), BMI > 30 (n = 33; 18 controls, 15 cases [5 high risk cases]), AA (n = 25, 14 controls, 11 cases [6 high risk cases]), CA (n = 84; 39 co ntrols, 45 cases [23 high risk cases])
83 Figure 4 3. qPCR results showing no significant differences between demographic groups for lactic acid bacteria and bifidobacteria in biopsies Error bars represent the standard deviation. BMI is defined as weight in kg divided by height in meters squared. All subjects for whom demographic data were provided are included. Two subjects did not provide BMI data. Total (n = 78 ; 35 controls, 43 cases [22 high risk cases] ), BMI < 25 (n = 19 ; 8 controls 11 cases [6 high risk cases] ), BMI 25 30 (n = 34 ; 13 controls, 21 cases [13 high risk cases] ), BMI > 30 (n = 23 ; 14 controls, 9 cases [2 high risk cases] ), AA (n = 10 6 controls, 4 cases [2 high risk cases] ), CA (n = 68 ; 29 controls, 39 cases [20 high risk cases] )
84 A B Figure 4 4. qPCR results showing no significant differences between BMI groups for lactic acid bacteria and bifidobacteria in stools (A) and biopsies (B) Error bars represent the standard deviation. BMI is defined as w eight in kg divided by height in meters squared. All subjects that provided BMI data are included. Stools (total n = 103; n = 30 BMI < 25, n = 40 BMI 25 30, n = 33 BMI > 30). Biopsies (total n = 76; n = 19 BMI 25, n = 34 BMI 25 30, n = 23 BMI > 30). 0 1 2 3 4 5 6 Bifidobacteria Lactic acid bacteria log genome equivalents Bacterial target 0 1 2 3 4 5 Bifidobacteria Lactic acid bacteria log genome equivalents Bacterial target BMI < 25 BMI 25 30 BMI > 30
85 A B Figure 4 5. qPCR results showing abundances for lactic acid bacteria and bifidobacteria between racial groups in stools(A) and biopsies (B) Error bars represent the standard deviation. All subjects for whom racial data were p rovided are included. = p < 0.05. AA: n = 10; CA: n = 68. 0 1 2 3 4 5 6 Bifidobacteria Lactic acid bacteria log genome equivalents Bacterial target AA CA
86 Figure 4 6 Rarefaction curves for stool and biopsy communities based on Chao1, which is a measure of estimated diversity if sequenced to completion. Stools (total n = 60; 30 cas es and 30 controls); Biopsies (total n = 32; 16 cases and 16 controls).
87 A B C D Figure 4 7 Principal co ordinate analysis (PCoA) based on UNIFRAC showing differences between stool and biopsy microbiota (A), microbiota by cases status in stool (B) and biopsy samples (C), and microbiota by racial group (D). Stools (total n = 60; 30 cases and 30 controls); Biopsies (total n = 32; 16 cases and 16 controls); Race (total n = 55; 18 AA and 37 CA).
88 A B Figure 4 8 Relative abu ndance of bacteria phyla between cases and controls in stools (A) and biopsies (B).
89 Figure 4 9 Heat map showing the distribution of the 20 most significantly differing individual OTUs between all cases and controls in stools at the 95% similarity level Each column represents data from a single subject. Rows represent individual OTU s Row numbers represent the OTU ID. Percentages represent the percent identity of the representative sequence to the database match. Cell s hade signifies the number of sequence read s matching to the specified OTUs. Labels in parentheses were obtain from the RDP classifier algorithm and used when megablast results did not provide taxonomic information. rb = rumen bacterium, mcrb = mixed cultur e rumen bacterium.
90 Figure 4 10. Heat map showing the distribution of the 20 most significantly differing individual OTUs between high risk cases and controls in stools at the 95% similarity level. Each column represents data from a single subject. Rows represent individual OTUs. Row numbers represent the OTU ID. Percentages represent the percent identity of the representative sequence to the database match. Cell shade signifies the number of sequence reads matching to the specified OTUs. Labels in paren theses were obtain from the RDP classifier algorithm and used when megablast results did not provide taxonomic information. rb = rumen bacterium, mcrb = mixed culture rumen bacterium.
91 Figure 4 11. Heat map showing the distribution of the 20 most signif icantly differing individual OTUs between high risk cases defined conventionally and controls in stools at the 95% similarity level. Each column represents data from a single subject. Rows represent the OTU. Row numbers represent the OTU ID. Percentages re present the percent identity of the representative sequence to the database match. Cell shade signifies the number of sequence reads matching to the specified OTUs. Labels in parentheses were obtain from the RDP classifier algorithm and used when megablast results did not provide taxonomic information. rb = rumen bacterium, mcrb = mixed culture rumen bacterium.
92 Figure 4 12. Heat map showing the distribution of the 20 most significantly differing individual OTUs between all cases and controls in stools at the 98% similarity level. Each column represents data from a single subject. Rows represent the OTU. Row numbers represent the OTU ID. Percentages represent the percent identity of the representative sequence to the database match. Cell shade signifies th e number of sequence reads matching to the specified OTUs. Labels in parentheses were obtain from the RDP classifier algorithm and used when megablast results did not provide taxonomic information. rb = rumen bacterium, bpb = butyrate producing bacterium.
93 Figure 4 13. Heat map showing the distribution of the 20 most significantly differing individual OTUs between high risk cases and controls in stools at the 98% similarity level. Each column represents data from a single subject. Rows represent the OTU. Row numbers represent the OTU ID. Percentages represent the percent identity of the representative sequence to the database match. Cell shade signifies the number of sequence reads matching to the specified OTUs. Labels in parentheses were obtain from the RDP classifier algorithm and used when megablast results did not provide taxonomic information. rb = rumen bacterium, bpb = butyrate producing bacterium
94 Figure 4 14. Heat map showing the distribution of the 20 most significantly differing individual OTU s between high risk cases defined by convention and controls in stools at the 98% similarity level. Each column represents data from a single subject. Rows represent the OTU. Row numbers represent the OTU ID. Percentages represent the percent identity of t he representative sequence to the database match. Cell shade signifies the number of sequence reads matching to the specified OTUs. Labels in parentheses were obtain from the RDP classifier algorithm and used when megablast results did not provide taxonomi c information. rb = rumen bacterium, bpb = butyrate producing bacterium.
95 Figure 4 15 Heat map showing the distribution of all significantly differing individual OTUs between all cases and controls in biopsies at the 95% similarity level. Each column re presents data from a single subject. Rows represent the OTU. Row numbers represent the OTU ID. Percentages represent the percent identity of the representative sequence to the database match. Cell shade signifies the number of sequence reads matching to th e specified OTUs. Labels in parentheses were obtain from the RDP classifier algorithm and used when megablast results did not provide taxonomic information. bpb = butyrate producing bacterium.
96 Figure 4 16 Heat map showing the distribution of all signif icantly differing individual OTUs between high risk cases and controls in biopsies at the 95 % similarity level. Each column represents data from a single subject. Rows represent the OTU. Row numbers represent the OTU ID. Percentages represent the percent i dentity of the representative sequence to the database match. Cell shade signifies the number of sequence reads matching to the specified OTUs. Labels in parentheses were obtain from the RDP classifier algorithm and used when megablast results did not prov ide taxonomic information. uec = unidentified Eubacterium clone, bpb = butyrate producing bacterium.
97 Figure 4 17 Heat map showing the distribution of all significantly differing individual OTUs between all cases and controls in bio psies at the 98 % sim ilarity level. Each column represents data from a single subject. Rows represent the OTU. Row numbers represent the OTU ID. Percentages represent the percent identity of the representative sequence to the database match. Cell shade signifies the number of sequence reads matching to the specified OTUs. Labels in parentheses were obtain from the RDP classifier algorithm and used when megablast results did not provide taxonomic information. hif = human intestinal firmicute.
98 Figure 4 18 Heat map showing the distribution of all significantly differing individual OTUs between high risk cases and controls in biopsies at the 98% similarity level. Each column represents data from a single subject. Rows represent the OTU. Row numbers represent the OTU ID. Percenta ges represent the percent identity of the representative sequence to the database match. Cell shade signifies the number of sequence reads matching to the specified OTUs. Labels in parentheses were obtain from the RDP classifier algorithm and used when meg ablast results did not provide taxonomic information. hif = human intestinal firmicute.
99 Figure 4 19 Heat map showing the distribution of the 20 most significantly differing individual OTUs in stools between racial groups at the 95% similarity level Each column represents data from a single subject. Rows represent the OTU. Row numbers represent the OTU ID. Percentages represent the percent identity of the representative sequence to the database match. Cell shade signifies the number of sequence reads matching to the specified OTUs. Labels in parentheses were obtain from the RDP classifier algorithm and used when megablast results did not provide taxonomic information. sfb = swine fecal bacterium, hif = human intestinal firmicute.
100 Figure 4 20 Heat map showing the distribution of the 20 most significantly differing individual OTUs in stools between racial groups at the 98% similarity level. Each column represents data from a single subject. Rows represent the OTU. Row numbers represent the OTU I D. Percentages represent the percent identity of the representative sequence to the database match. Cell shade signifies the number of sequence reads matching to the specified OTUs. Labels in parentheses were obtain from the RDP classifier algorithm and us ed when megablast results did not provide taxonomic information. sfb = swine fecal bacterium, rb = rumen bacterium, bpb = butyrate producing bacterium.
101 A B Figure 4 21 Heat map based on prevalent OTUs in either case or control groups at the 92% si milar ity level (A) Samples 1 to 30 = control group, samples 31 to 60 = polyp group. Scale represents abundance of OTUs within a sample from 1 (sample with the highest number of stated O T U ) to 0 (sample with stated OTU absent). The AUC demonstr ates the accuracy of the model (B)
102 Figure 4 22 Number of shotgun sequence reads with closest matches to gamma Proteobacteria in the case and the control pools. n = 18; 9 cases and 9 controls. 0 500 1000 1500 2000 2500 3000 3500 4000 Number of sequence reads Cases Controls
103 Figure 4 23 Number of shotgun sequence r eads with closest matches within bacterial metabolic pathways in the case and the control pools. n = 18; 9 cases and 9 controls. 0 100 200 300 400 500 600 700 800 Starch and sucrose metabolism Galactose metabolism Fructose metabolism Number os sequence reads Cases Controls
104 CHAPTER 5 EFFECTS OF SPECIFIC COMPLEX CARBOHYDRATES ON FEC AL MICROBIOTA COMMUNITY COMPOSITIO N Study #1: Resistant M altodextrin Increases Bifidobacteria Counts in Healthy Males Introduction Resistant maltodextrin (RM), a type 3 resistant starch, is a soluble dietary fiber used as a food ingredient to improve qualities such as texture, taste, and health benefits, includi ng reduced blood sugar and improved GI function. The most common form of RM is marketed as Fibersol 2, which consists of 90% RM. The process by which RM is made begins with starch that is hydrolyzed in the presence of heat and acid. The hydrolysis reaction breaks 1 4 and 1 6 glycosidic linkages. Because the amount of water present is limited, transglucosidation occurs between remaining starch polymers randomly forming 1 2, 1 3, 1 4, 1 6, 1 2, and 1 3 linkages. amylase is then added to remove linkages an d resulting glucose monomers are removed, leaving behind the remaining polymer defined as RM (35) RM is considered a dietary fiber by the f ollowing definition (35) : mers not hydrolyzed by endemic digestive enzymes, they are derived from f ood raw materials, and there must be a validated analytical procedure established that measures only the fiber fraction on the ingredient. Also, the material must demonstrate a RM has been shown to induce smaller increases in blood sugar compared with a placebo, allowing it to be considered beneficial to health (34) RM is of interest to food producers i n that it exhibits low levels of GI discomfort and provides many of the other benefits commonly observed with other dietary fibers (35) RM was investigated to determine its effects on energy gain and gut microbiota composition, the latter of which
105 is presented below, under the hypothesis that RM would increase the numbers of bacteria that could selectively ferment RM and induce changes in gut microbiota composition such that similarities arise between RM treated subjects and polyp free subjects Data w ere obtained using DGGE, qPCR, and pyrosequencing as described previously. The data were also compared to those obtained from subjects with poly ps and subjects without po lyps presented in C hapter 4. M icrobiota profiles of RM treated subjects that are consistent with thos e from subjects without polyps suggest that supplementation with RM may offer one avenue for increased protection against CRC. St udy Design Fifteen male volunteers were recruited to participate in a randomized, double blind, placebo controlled, crossover study with three 28 day feeding periods separated by a wash out period of at least two weeks. 14 of the 15 subjects who began the study completed the entire study protocol. One subject failed to complete the study due to scheduling conflicts. Only data from those 14 subjects who completed the study are included in the data analysis Fibersol 2 (Matsutani, Inc., Japan) was used as the RM in this study Subjects were randomly assigned to treatments consisting of 25 g/day of r esistant maltodextrin + 25 g/day maltodextrin (RM25), 50 g/day of r esistant maltodextrin (RM50) and 50g/day maltodextrin (placebo). Prior to the first fecal co llection, volunteers were adapted to a study diet for two weeks. During the three treatment periods, volunteers consumed the same base generic American diet. All foods and beverages were prepared and supplied by the Human Studies Facility at the Beltsvil le Human Nutrition Research Center (Beltsville, MD). Dinner and breakfast were consumed at the center during the week, and carryout lunches and snacks were provided. Weekend meals and treatments were packaged with instructions for home
106 consumption. The stu dy protocol was reviewed and approved by the Medstar Research Institute Institutional Review Board. All subjects provided written informed consent and were compensated for their participation. For microbiota analysis, fecal samples were collected on days 1, 13, and 24 of each intervention for a total of nine fecal samples per subject. Subjects obtained a cooler filled with ice for storage of the sample until delivery to the lab. For the microbiota analysis, all samples were delivered on ice usually within 4 hours of defecation. Results Fecal microbiota community diversity DGGE, a simple but efficient method for initial profiling of fecal microbiota, suggested that each individual harbored a unique microbiota. A distinct band was consistently and dose depend ently increased after 24 days of RM supplementation in 12 out of 14 subjects ( Figure 5 1 ). This increase was also observed, although to a lesser extent, after 13 days of RM supplementation. DNA isolated from two of these bands was cloned and sequenced. Bot h resulting sequences matched closest to sequences in the family Lachnospiraceae within the phylum Firmicutes. Bifidobacteria qPCR Quantitative PCR data was analyzed as described previously using genome equivalents/ng of DNA. Changes in amounts of bacteri al groups of interest were determined by comparing the change of the genome equivalents/ng DNA between Day 1 to Day 24 during the RM period with the change between Day 1 to Day 24 during the placebo period. This approach allowed for consideration of the RM independent effect of the study diet.
107 A statistically significant increase in bifidobacteria from Day 0 to Day 24 was observed during the RM50 period, but not during the placebo period (Figure 5 2 ).13 of the 14 subjects showed an increase in bifidobacter ia from Day 0 to Day 24 during the RM50 period (50 g/day supplementation) while only 5 of 14 subjects showed such an increase during the placebo period (RM0) Furthermore, the mean numbers of bifidobacteria were higher on Day 24 during RM50 period (2.7 x 1 0 5 genome equivalents/ng) compared to the placebo period (1.5 x 10 5 ), although these data did not reach significance. 454 pyrosequencing analysis of rDNA To expand the microbiota analysis from the limited number of targeted groups of gut bacteria, high thr oughput microbial community 16S rRNA sequencing using the 454 titanium technology was performed. After removal of low quality and short reads, a total of 466,622 sequences were retained with an average of 5622 sequences per subject and an average length of 296 nucleotides per sequence read Chao1 based rarefaction curves suggest that o verall numbers of estimated OTU s did not differ between the treatments (Fig. 5 3 ). UniFrac based PCoA analysis also did not detect differences in overall microbiota compositio n between treatments (Fig. 5 4 ). When the distribution of OTUs across treatment was evaluated many OTUs were incr eased or decreased by RM (Fig. 5 5 ). Microbiota was initially compared on days 1 (top block in Figure 5 5 ) and 24 (middle block in Figure 5 5 ) during the RM50 period. However, because study subjects not only added RM to their diet but switched from thei r normal diet to the study diet, the sample collected on day 24 of the placebo tre atment (bottom block in Figure 5 5 ) was included in the analysi s. The OTU s that were increased during RM50 matched
108 closest to clostridia, ruminococci, bacteroidetes and parabacteroides as well as matches to unknown bacteria. Conclusion Changes within bacterial groups were detected with each experimental method; howeve r, each method revealed changes in different bacterial groups. DGGE revealed in most subjects an increase in a bacterial group within the family Lachnospiraceae a group of known butyrate producers, while qPCR showed an increase in bifidobacteria as a resu lt of RM supplementation An increase in bifidobacteria and butyrate production are expected based on previous work (29, 79) OTU based pyrosequencing results showed increases in many OTUs matching to various taxonomic levels within the class Clostridia and decreases in other OTUs within the same class These results indicate that RM supplementation has some relatively non specific effects on higher level taxonomic bacterial groups while simultaneously exhibiting OTU speci fic effects within other groups; however, there were no observed differences in ov erall diversity or community structure. There were no clear associations between OTUs enriched by RM and OTUs that were found to be significantly more prevalent in polyp free subjects. Study #2: The Addition of Whole Grains to the Diets of Middle school Ch ildren: Effects on Fecal Microbiota Community Structure Introduction The Food and Drug Administration defines whole grains to be cereal grains that consist of the intact, ground, cracked or flaked fruit of the grains whose principal components -the starc hy endosperm, germ and bran -are present in the same relative proportions as they exist in the intact grain ( http://www.fda.gov/NewsEvents/Newsroom/PressAnnouncemen ts/2006/ucm108598.ht
109 m ). The bran consists primarily of insoluble fibers including cellulose and hemicelluloses, while the germ consists primarily of phytochemicals and soluble fibers, including oligosaccharides and resistant starch (RS) The endosperm, ho wever, is rich in starch and provides energy for the developing plant (122) The refining process by which whole grains are processed into refined grains involves the removal of the br an and the germ. Although often fortified with B vitamins, this process removes most of the fibers and phytochemicals present in whole grains (3) Phytochemicals, defined as bioactive, non nutrient plant compounds that are associated with reduced risk of chronic diseases incl ude many plant phenolic compounds that often serve as antioxidants (144) The demonstrated esterase activity of fecal homogenates suggests that gut microbiota act ivity may increase bioavailability of phytochemicals, specifically ferulic acids, and thus contribute to the protective effect of whole grains (6, 137, 205) Other protective effects of whole grains may rely on the ability of whole grains to act as prebiotics. Whole grains may be considered prebiotics due to the presence of oligosaccharides and RS. Substrates includ ing RS that enrich bacterial groups, particularly butyrate producers, offer another mechanism by which whole grain consumption promotes GI health. Butyrate production has been shown to be proportional to RS present in grains (105) The associations of whole grain intake with beneficial health outcomes and reduced cancer risk, as well as the ability of their components to modulate gut microbiota composition, led to the hypothesis that whole grain consumption shapes gut microbiota community structures such that they display some similarities with gut microbiota communities found in polyp free patients; however such similarities would
110 likely be subtle due to dietary differences other than grains and factors other than diet that contribute to polyp formation The goal of this study was to determine the changes in gut microbiota composition as a result of whole grain intervention. Stool pH was measured and microbiota analysis will be done using DGGE, qPCR, and pyrosequencing as described previously These data show that gut microbiot a composition may be influenced via dietary fiber intake. Study Design Subject recruitment Adolescents were recruited from a local middle school (ages 11 to 15 years) A total of 196 adolescents were screened and 91 assented to participate with the consent of at least one parent Eight participants were excluded because they did not meet inclusion/exclusion criteria ( n =4) or were no longer interested ( n =4). Eighty three participants began the 1 week pre baseline screen and were randomized to either the refi ned grain group (RG, n =42) or whole grain group (WG, n =41). All study procedures were approved by the University of Florida Institutional Review Board and the School Board of Alachua County. Inclusion criteria included the following: Parental/guardian cons ent Student assent Willingness to eat three different study foods each day for six weeks Willingness to provide two blood samples, 2 saliva samples, and 2 stool samples over the course of the study Exclusion criteria included the following: C urrently t aking medication for constipation or diarrhea Antibiotics within the four weeks before randomization
111 Consumption of probiotics of greater than three servings of yogurt per week Diseases or illnesses such as gastrointestinal disease (gastric ulcers, Croh n's, ulcerative colitis, etc.), other chronic diseases (diabetes, kidney disease, etc.) or immune modulating diseases (HIV, AIDS, autoimmune, hepatitis, cancer, etc.) Food allergies (wheat, soy, egg, milk, gluten, nuts, or any other food or food ingredien t) Experimental protocol Parti cipants were enrolled in this six w ee k, randomized, controll ed, parallel arm study over a three w ee k period in February, 2010. In the week prior to randomization, one stool sample and two targeted 24 h our recalls were collect ed, and height (portable stadiometer), we ight (digital scale), and birth date were obtained to determine BMI percentile for age. Participants were stratified based on pre baseline weight category (<95 th th percentile f randomized to the RG or WG via sealed envelope. The randomiza tion scheme was generated by a statistician who did not have contact with study participants. A grocery bag of a variety of grain based foods was picked up at the school or delivered to the udy a stool sample was obtained. 83 subjects completed the study protocol, and 59 subjects provided both a baseline and a post treatment stool sample. Data from 57 subjects wer e used for 454 pyrosequencing analysis due to barcoding errors in two samples. Grain based foods and administration p rotocol Four rotating weekly food packages containing approximately ten different grain based foods (e.g., cereal, pasta, rice, bread, panc ake mix, snack foods) were provided to participants and their families. Two single serving snack packs (i.e., whole grain cereal [provided by Cereal Partners Worldwide] or 100 calorie cookie packs) were
112 distributed to study participants on school days. The study foods were primarily wheat based but also included oats, rice, and corn. One serving of the study foods provided 0 g (RG) or 161 g (WG) of whole grains. Participants were told to eat the study foods in place of grains consumed as part of their typi cal diet, and they were encouraged to eat three different kinds of study foods each day with the goal of 5 oz eq uivalents of grains per day. Results Fecal pH Fecal pH was measured in all samples and revealed no differences within diet groups over time or between diet g roups post treatment (Figure 5 6 ). Fecal microbiota community diversity The success of DNA extraction in all samples was confirmed by DGGE, but no distinct profiles or individual bands for either diet group or time period were revealed. Usin g the Shannon Weiner and Simpson diversity indices, no significant differences in community diversity were detected (Figure 5 7 ) Profiles within participants remained stable across time periods, suggesting that the intervention did not considerably pertur b microbiota profiles (data not shown). Quantification of targeted bacteria LAB a nd bifidobacteria were measured via qPCR. There was a significant diet group by time period interaction; LAB increased (P = 0.0002) between baseline and final time periods i n stools from participants in the WG group (Figure 5 8 ) There was no difference in bifidobacteria between diet groups; however, there was a significant increase (P < 0.0001 ) in bifidobacteria in both diet groups. There was no effect of baseline weight cat egory on LAB or bifidobacteria.
113 454 pyrosequencing of 16S rRNA analysis After preprocessing, 387,933 16S rRNA sequences from 48 subjects were retained, with a mean of 3043 sequences/sample and a mean length of 498 nucleotides. Sequence readings were binne d by using ESPRIT Tree into 11722 OTUs at the 98% similarity level. Chao1 based rarefaction curves did not significantly differ between diet groups or across time periods, suggesting that overall diversity did not change during the study period (Figure 5 9 ) To evaluate changes in overall microbiota composition, PCoAs based on the unweighted UniFrac metric were generated No distinct clustering by time period or diet group was detected (Figure 5 10 ) The proportions of dominant bacterial phyla did not signi ficantly differ between d i et groups over time (Figure 5 11 ). When the effects of WG consumption on individual OTUs were examined, 6 OTUs showed a strong positive association whereas 3 OTUs were negatively associated with WG consumption ( Figure 5 12 ). Concl usion qPCR revealed an increase in lactic acid bacteria as a result of WG supplementation, which may explain the claimed beneficial effects in the literature regarding WGs. Enrichment of both lactic acid bacteria and bifidobacteria after WG supplementation has been demonstrated previously; however, no significant changes in bifidobacteria were observed in this study (41, 56, 56) Variation in the source of whole grains may account for th is difference. Sequencing analysis revealed no significant differences in the proportions of bacterial phyla as a result of WG supplementation; however, at the individual OTU level, both increases and decreases in the prevalence of specific OTUs were detec ted. The prevalence of some OTUs matching to butyrate producing bacteria was decreased while others were increased as a result of WG
114 supplementation. While these results confirm that gut microbiota composition is affected by WG supplementation, they sugges t that the beneficial effect of WG consumption may not occur at the individual OTU level There are no clear associations between the gut microbiota communities resulting from WG intervention and those found within polyp free subjects. Study #3: The Effect s of Galactooligosaccharides on Fecal Microbiota Composition in 3.1 Undergraduate Students and 3 .2 Aged A dults Introduction GOS, described previously, are of interest due to their ability to enrich butyrate producing bacteria and other bacterial groups tha t negatively associate with CRC. The demonstrated ability of GOS to modulate gut microbiota community structure led to the hypothesis that subjects who consume GOS will develop some gut microbiota community characteristics that are similar those found in s ubjects without polyps Because the diets of study subjects will not be controlled beyond GOS supplementation dramatic similarities are not expected. Furthermore, polyp formation is a multi factorial process that involves parameters other than diet. Gut microbiota composition has been shown to differ between age groups (19) Therefore, two stud ies in two different age groups were conducted to determine the effects of GOS on gut microbiota composition. The first study examined GOS effects in academically stressed healthy undergraduate university students during cold and flu season. The stress mod el required that all subjects be full time students with at least one final exam during the post treatment sample collection period. Because stress is associated with weakened immunity and GI discomfort, this model was used to increase the likelihood of ob serving the beneficial effects of GOS.
115 The second study examined the effects of GOS on gut microbiota composition in aged adults (60 years or more) over cold and flu season. Studying this population is important because CRC risk increases with age. Data f or stool pH and microbiota analyses using DGGE, qPCR, and pyrosequencing were obtained as described previously. Results from both studies were compared to determine both age dependent and age independent effects of GOS. Because both CRC risk and gut microb iota composition are dependent on age, GOS may affect these populations differently. Protocol 1: Galactooligosaccharides Supplementation in Healthy University Students Study Design Subject recruitment Participants from the University of Florida were recrui ted via list servs, flyers, posters, and announcements in early fall of 2009. Each subject gave written informed consent, and all study procedures were in accordance with the ethical standards of the University of Florida Institutional Review Board. Inclus ion criteria included the following: Full time undergraduate student taking at least 12 credit hours Age of 18 years or older Willingness to complete dail y assessment forms via computer Willingness to discontinue any immune enhancing dietary suppl ements (e.g., prebiotics and fib er supplements, probiotics, Echinacea fish oil, vitamin E > 100% of the RDA or > 15 mg/day Willingness to take the fiber daily for 8 weeks Willingness to provide a social security number to receive study payment A cold in the last 12 months At least one final exam during the Fall 2009 exam week, between Saturday,
116 December 12 and Friday, December 18, 2009 Daily access to a computer with Internet access for the entire 8 week study Exclusion criteria included the following: Cur rent smoker Chronic allergies involving the upper respiratory tract Allergy to milk Known illnesses or conditions that may impact perceived health such as HIV/AIDS, diabetes, renal or gastrointestinal diseases Chemotherapy or other immune suppressing t herapy within the last year Antibiotic therapy in the past two months Experimental protocol Subjects (n = 427) were randomly assigned to a supplement group during the first week in November of 2009 and were followed for eight w ee k s including the time of fall final exams. 75 subjects were asked to provide stool samples. Exams were held over a span of one w ee k during the sixth week of the intervention. The number and scheduled time of exams varied subjects completed their last exam s they were on semester break through the remainder of the study. The study was a prospective, randomized, parallel, double blind, placebo controlled trial. Subjects were proportionally stratified based on gender (50/50) and rand omly assigned over a five d ay period via sealed envelopes to receive 0, 2.5, or 5.0 g GOS ( Corn Products Golden, CO). The stratification and randomization schemes were generated by the study statistician who did not have direct contact with any subjects. On the basis of the number of upper respiratory tract infec tions in GOS supplemented infants and the number of colds in academically stressed undergraduate
117 students, which were important for the immunological parameters examined in this study independent o f the microbiota work, a sample size of 140 subjects in each group was calculated to observe a 50% reduction in the proportion of colds in the treatment groups with 80% power, a = 0.05, and a 10% attrition rate (9, 134) 419 subjects completed the study protocol, of which only 69 provided both a baseline and post tr eatment stool sample. Subjects provided baseline stools samples throughout the week before randomization and post treatment samples throughout the week of final exams. GOS administration p rotocol The GOS supplements were provided in coded packets that were similar in size and shape to commercially available single (sucrose) was added to the 0 g and 2.5 g packets so that all packets were the same weight and looked similar. A flow agent (silicon dioxide) was added to all pac kets to improve emptying of the package contents The subjects were instructed to pour the contents of the packet into any beverage, mix well, and consume the beverag e in its entirety each day for eight w ee k s Both the GOS and sucrose had a slightly sweet taste. The subjects were unable to distinguish the GOS packets from the placebo. Regardless of the actual treatment group, the proportions of subjects who thought they were receiving 0, 2.5, and 5.0 g GOS were 44%, 42%, and 14%, respectively. No difference s were found between groups, which suggested successful blinding. Results Fecal microbiota community diversity The success of DNA extraction in all samples was confirmed by DGGE, but no distinct profiles or individual bands for diet group s or time period s were revealed. Using the Shannon Weiner and Simpson diversity indices, no significant differences in
118 community diversity were detected (Figures 5 13 and 5 14 ) Profiles within participants remained stable across time periods, suggesting that the interventi on did not considerably perturb microbiota profiles (data not shown). Quantification of targeted bacteria LAB and bifidobacteria were measured via qPCR. There were no significant diet group by time period interaction s for either LAB or bifidob acteria (Fig ures 5 15 and 5 16 ). 454 pyrosequencing analysis of 16S rDNA After preprocessing, 553,152 16S rD NA sequences from 69 subjects were retained, with a mean of 5885 sequences/sample and a mean length of 394 nucleotides. Sequence readings were binned by using E SPRIT Tree into 25595 OTUs at the 98% similarity level. The proportions of dominant bacterial phyla did not significantly differ between diet groups over time (Figure 5 17). Chao1 based rarefaction curves did not significantly differ between diet groups or across time periods, suggesting that overall diversity did not change during the study period (Figure 5 18) To evaluate changes in overall microbiota composition, PCoAs based on the unweighted UniFrac metric were generated No distinct clustering by time period or diet group (data not shown) was detected. When the effects of GOS consumption on individual OTUs were examined, 50 OTUs showed a stron g positive correlation whereas no OTUs were negatively associated with GOS consumption ( Figure 5 19 ).
119 Protocol 2: Galactooligosaccharides Supplementation in Healthy Aged Adults Study Design Subject recruitment Participants we re healthy, older adults aged 60 or older who suffered at least one cold in the previous year. Each subject gave written informed consent, and all study procedures were in accordance with the ethical standards of the University of Florida Institutional Review Board. Inclusion criteria included the following: Age of 60 years or older Willingness to complete daily and monthly questionnaires Will ingness to receive the fall influenza vaccination as part of the study protocol Willingness to provide three blood samples, three saliva samples, and two stool samples and answer a food frequency questionnaire over the course of the study Willingness to discontinue and immune enhancing dietary supplements (e.g., prebiotics and fiber supplements, probiotics, Echinacea, fish oil, vitamin E > 100% of the RDA or > 15 mg/day) Willingness to take the study fiber for 24 weeks Willingness to provide a social se curity number to receive study payment A cold within the last 12 months Ability to take foods and the study fiber without the aid of another person Eligibility and Willingness to receive the influenza vaccine for the current year Exclusion criteria inc luded the following: Current smoker Chronic allergies involving the upper respiratory tract Currently taking medication for constipation or diarrhea Currently taking any anti inflammatory drugs on a regular basis
120 Current treatment for Alzheimer's disea se Milk or serious egg allergies Current treatment for any known illnesses or conditions that may impact perceived health such as HIV/AIDS, immune modulating diseases (autoimmune, hepatitis, cancer, etc.), diabetes, kidney disease, gastrointestinal disea ses (gastric ulcers, Crohn's, ulcerative colitis, etc.) Chemotherapy or other immune suppressing therapy within the last year Antibiotic therapy within the last two months Current use of supplemental oxygen Experimental protocol Subjects (n = 80 ) were r andomly assigned to a supplement group during the September of 2010 and were followed for six months. The study was a prospective, randomized, parallel, double blind, placebo controlled trial. Subjects were propor tionally stratified based on gender (50/50) and randomly assigned via sealed envelopes to receive 0 or 5.0 g GOS ( Corn Products Golden, CO). The stratification and randomization schemes were generated by the study statistician who did not have direct contact with any subjects. The GOS supplements were provided in 0 g or 2.5 g coded packets and subjects were instructed to consume the GOS in a beverage. Based previous work with this population (135) a significant difference in cold and flu sym ptom intensity score will be observed with 80 subjects (includes 20 % attrition rate). The rationale for using a sample size based on cold and flu symptoms is based on other primary outcomes of the study protocol not related to microbiota analysis 81 subjects completed the study protocol, and 80 provided both a baseline and a post treatment stool samples. Subjects provided a baseline stool sample during the randomization week and a post treatment sample after two weeks of supplementation.
121 GOS administration p rotocol The GOS supplements were provided in coded packets that were similar in size and shape to commercially available single (sucrose) was added to the 0 g and 2.5 g packets so that all packets were the same weight and looked similar. A flow agent (silicon dioxide) was added to all packets to improve emptying of the package contents The subjects were instructed to pour the contents of the packet into any beverage, mix well, and consume the beverage in its entiret y twice daily for 6 months Both the GOS and sucrose had a slightly sw eet taste. The subjects were unable to distinguish the galactooligosaccharide packets from the placebo. Results Fecal pH Stool pH was measured in all samples and revealed no differences between diet groups over time or between diet gr oups post treatment ( Figure 5 20 ). Fecal microbiota community composition The success of DNA extraction in all samples was confirmed by DGGE, but no distinct profiles or individual bands for either diet group or time period were revealed. Using the Shannon Weiner and Simpson d iversity indices, no significant differences in community diversity were detected (Figure 5 21 ) Profiles within participants remained stable across time periods, suggesting that the intervention did not considerably perturb microbiota profiles (data not s hown). Quantification of targeted bacteria LAB and bifidobacteria wer e measured via qPCR. There were no significant diet group by time period interaction s for either LA B or bifidobacteria (Figure 5 22 ) However,
122 proportions of bifidobacterial genome equiv alents were significantly greater in the GOS treated group than the placebo group. (Figure 5 2 3 ). Proportions of genome equivalents detected using qPCR represent the number of target bacterial genome equivalents divided by the number of total genome equiva lents determined by the V3 universal primer set. T he data were then fit in to a model which predicted higher increases in bifidobacteria genome equivalent proportions for subjects with higher baseline values. Bifidobacteria as a proportion of all bacteria m easured in the fecal samples were transformed using arcsin square root and analyzed as a function of the baseline proportion of bifidobacteria (also transformed), intervention group, age group, and the two way interactions of these main effects using a lin ear model. Non significant interactions were removed. GOS treated subjects displayed significantly greater increases than placebo treated subjects with the same baseline levels. (Figure 5 2 4 ). Lactic acid bacteria displayed similar behavior between treatme nt groups (data not shown). Furthermore, subjects aged 65 years or more displayed significantly greater increases than those aged less than 65 years. A similar relationship was observed between baseline value and age group for lactic acid bacteria ( data no t shown ). 454 pyrosequencing analysis of 16S rDNA After preprocessing, 719,946 16S rRNA sequences from 80 subjects were retained, with a mean of 8999 sequences/sample and a mean length of 480 nucleotides. The percentages of dominant bacterial phyla did no t significantly differ between treatme nt groups over time (Figure 5 2 5 ). Sequence readings were binned by using ESPRIT Tree into 27597 OTUs at the 98% similarity level. Chao1 based rarefaction curves did not significantly differ between diet groups or acro ss time periods,
123 suggesting that overall diversity did not change during the study period ( Figure 5 2 6 ) To evaluate changes in overall microbiota composition, a PCoA based on the unweighted UniFrac metric were generated No distinct clustering by time per iod or diet group was detected (Figure 5 2 7 ) When the effects of GOS consumption on individual OTUs were examined, several OTUs were enriched by GOS supplementation (Figure 5 28) Conclusion Both study populations exhibited no significant change in diversi ty indexes as result of GOS treatment. Similarly, qPCR results showed no significant changes in genome equivalents for Bifidobacteria or lactic acid bacteria in either study population; however, when genome equivalents were converted to proportions, there was a significantly higher proportion of bifidobacteria in the aged adult GOS treated subjects compared to their placebo counterparts post treatment, but this was not observed within the student study population A dose dependent increase in GOS was expect ed based on previous studies (63, 64) qPCR is used for bifidobacteria and lactic acid bacteria quantification because the low abundance of these groups. 16S sequencing is not optimized for quantification of organisms with low abundance because the techno logies to date only capture a small percentage of the entire microbial community, and organisms present at low abundances may be missed or underrepresented. In these studies, 16S sequencing data for bifidobacteria did not parallel qPCR results. This is lik ely due to the lack of sufficient sequencing depth needed to accurately quantify lesser abundant organisms. Because of the incongruence between these two methods, it is difficult to determine which bacterial groups were reduced as proportions of bifidobact eria increased.
124 The Firmicutes:Bacteroidetes ratios were consistently higher across diet groups and both time points in the aged adult study compared to both diet groups and time points measured in the student study population; h owever, these ratios were not significantly affected by GOS treatment Together with the bifidobacteria results, these data suggest that gut microbiota composition differs by age group and that GOS affects microbiota community composition differently depending on host age which p arallels the results presented Biagi, et al (18) The results from both studies revealed an increase in the prevalence of several OTUs as a result of GOS intervention. In most cases, GOS altered gut microbiota composition by increasing the prevalence of particular OTUs. In the aged adult study, o nly the prevalence of nine out of forty OTUs signi ficantly affected decreased. These data suggest that GOS mainly acts by enriching certain bacterial groups with little activity towards reducing bacterial groups. The OTUs that were increased in both GOS study cohorts matched to various Ruminococcus s pecie s Faecalibacterium prausnitzii, and unidentified human intestinal Firmicutes The data also show a greater number of OTUs significantly enriched in the student study population compared to the aged adult study population. These results further suggest tha t GOS exhibits different effects within different age groups. However, in both groups GOS significantly increased the prevalence of many butyrate producers which suggests a beneficial effect of GOS independent of age. The gut microbiota profiles of GOS tr eated subjects did not display and strikin g similarities with the gut microbiota profiles found in polyp free subjects. It is difficult to make meaningful comparisons across studies because the study populations were not consistent. Future studies should e xamine the affects of GOS in subjects with
125 polyps previously removed to determine the effects of GOS on subsequent polyp development. Though direct comparisons to the polyp study are difficult, the increase in the proportion of bifidobacteria as a result o f GOS treatment in the aged adult study, combined with results in the literature suggesting the protective effects of bifidobacteria, suggests that GOS may be beneficial in protecting against CRC via this activity.
126 Figure 5 1. DGGE gel showing the enrichment of a particular band as a result of RM treatment. Numbers below lanes represent treatment doses or RM. Letters below lanes represent individual subjects. M = marker.
127 Figure 5 2. qPCR results showing abundances in Bifidobacteria genom e equivalents between treatment groups over time (N = 14) Error bars represent the standard deviation. = p < 0.05.
128 Figure 5 3. Rarefaction curves for RM treatment based on Chao1, which is a measure of estimated d iversity if sequenced to completion (N = 14) Top: day 24 placebo; center: day 1 RM50; bottom: day 24 RM 50.
129 Figure 5 4. UniFrac based PCoA comparing RM baseline, RM post treatment, and placebo treatment periods (N = 14)
130 Figure 5 5. Heat map sh owing the distribution of selected individual OTUs that significantly differed in stools between time points at the 95% similarity level. Each column represents data from a single OTU. Rows represent individual subjects (N = 14) Numbers within cells repre sent the number of sequence reads matching to the specified OTU. Cell shade represents the proportion of a single OTU within a subject summed with the proportion of that Otu in all subjects with darker shades representing greater values. Top block = RM50, day 1; center block = RM50, day 24; bottom block = RM0, day 24.
131 Figure 5 6 Differences in stool pH between WG and RG treatment groups over time (n = 59; 29 WG and 30 RG). Error bars represent the standard deviation
132 Figure 5 7 Diversity indi ces in stool communities between WG and RG treatment groups over time (n = 59; 29 WG and 30 RG) Error bars represent the standard deviation.
133 Figure 5 8 qPCR results showing abundances of Bifidobacteria and Lactic acid bacteria (LAB) genome equivale nts between WG and RG treatment groups over time (n = 59; 29 WG and 30 RG) Error bars represent the standard deviation. = p < 0.001.
134 Figure 5 9 Rarefaction curves for WG treatment based on Chao1, which is a measure of estimated divers ity if sequence d to completion (n = 57; 27 WG and 30 RG)
135 Figure 5 10 UniFrac based PCoA comparing WG and RG treatment periods (n = 57; 27 WG and 30 RG)
136 A B C D Figure 5 11 Relative abundance s of bacteria l phyla in stools in WG treated subjects at base line (A), WG treated subjects post treatment (B), RG treated subjects at baseline (C), and RG treated subjects post treatment (D) (n = 57; 27 WG and 30 RG)
137 Figure 5 12 Heat map showing the distribution of individual OTUs at the 98% similarity le vel from subjects within the WG treatment group that significantly differed in stools between time and treatment groups (n = 27) Each column represents data from a single subject. Rows represent the OTU Cell shade signifies the number of sequence read s m atching to the specified OTU.
138 Figure 5 13 Shannon diversity indices in stool communities between treatment groups over time in the GOS student study population (n = 69; n = 24 (0g), n = 21 (2.5g), n = 24 (5g)) Error bars represent the standard deviat ion. 0 1 2 3 0 2.5 5 Shannon diversity index value Daily GOS supplementation (g) Baseline Post treatment
139 Figure 5 14 Inverse Simpson diversity indices in stool communities between GOS treatment groups over time in the student study population (n = 69; n = 24 (0g), n = 21 (2.5g), n = 24 (5g)) Error bars represent the standard deviation. 0 2 4 6 8 10 12 14 16 18 0 2.5 5 Inverse Simpson diversity index value Daily GOS supplementation (g) Baseline Post treatment
140 Fig ure 5 15 qPCR results showing no significant difference s in Bifidobacteria genome equivalents between GOS treatment groups over time in the student study population (n = 69; n = 24 (0g), n = 21 (2.5g), n = 24 (5g)) Error bars represent the sta ndard deviation. 0 1 2 3 4 5 6 0 2.5 5 log Bifidobacteria Genome Equivlaents Daily GOS supplementation (g) Baseline Post treatment
141 Figure 5 16 qPCR results showing no significant difference s in Lactic acid bacteria (LAB) genome equivalents between GOS treatment groups over time in the student study population (n = 69; n = 24 (0g), n = 21 (2.5g), n = 24 (5g)) Error bars represent the standard deviation. 0 1 2 3 4 0 2.5 5 log LAB Genome Equivalents Daily GOS supplemenation (g) Baseline Post treatment
142 A B C D Figure 5 17 Relative abundance s of bacteria l phyla in stools in GOS treated subjects (5.0g) at baseline (A), GOS treated subjects (5.0g) post treatment (B), placebo treated subje cts at baseline (C), and placebo treated subjects post treatment (D).
143 Figure 5 18 Rarefaction curves for GOS treatment in the student study population based on Chao1, which is a measure of estimated diversity if sequenced to completion (n = 48; n = 2 4 (0g), n = 24 (5g)).
144 Figure 5 19 Heat map showing the distribution of individual OTUs that significantly differed in stools between time and GOS treatment groups in the student study population Each column represents data from a single subject (n = 48; n = 24 (0g), n = 24 (5g)). Rows represent the OUT. Cell shade signifies the number of sequence read s matching to the specified OTU.
145 F igure 5 20 Measurements of stool pH between GOS treatment groups over time in the aged adult study population (n = 80; n = 37 (0g), n = 43 (5g)) Error bars represent the standard deviation. 0 1 2 3 4 5 6 7 8 Placebo GOS pH Baseline Post treatment
146 Figure 5 21 Diversity indices in stool communities between GOS treatment groups over time in the aged adult study population (n = 80; n = 37 (0g), n = 43 (5g)) Er ror bars represent the standard deviation.
147 Figure 5 22 qPCR results showing no significant difference s in Bifidobacteria and Lactic acid bacteria (LAB) genome equivalents between GOS treatment groups over time in the aged adult study population (n = 80; n = 37 (0g), n = 43 (5g)) Error bars represent the standard deviation.
1 48 Figure 5 23 qPCR results showing difference s in Bifidobacteria genome equivalent proportions between GOS treatment groups over time in the aged adult study population (n = 80; n = 37 (0g), n = 43 (5g)) Error bars represent the standard deviation. = p < 0.05.
149 A B Figure 5 24 The predicted proportion of bifidobacteria in fecal samples after two weeks of supplementation by the proportion of bifidobacteria in baseline samples for lines) or placebo (solid lines). Gray lines represent the 95% confidence interval.
150 A B C D Figure 5 25 Relative abundance s of bacteria l phyla in stools in placebo treated subjects (5.0g) at baseline (A), placebo treated subjects (5.0g) post treatment (B), GOS treated subjects at baseline (C), and GOS trea ted subjects post treatment in the aged adult study population (n = 80; 37 (0g), 43 (5g)).
151 Figure 5 2 6 Rarefaction curves for GOS treatment in the aged adult study population based on Chao1, which is a measure of estimated diversity if sequenced to completion (n = 80; n = 37 (0g), n = 43 (5g))
152 Figure 5 27 UniFrac based PCoA comparing GOS and place bo treatment periods in the aged adult study population (n = 80; 37 (0g), 43 (5g))
153 Figure 5 28 Heat map showing the distribution of individual OTUs that significantly differed in stools between collection periods in the GOS treated group in the age d adult study population Each column represents data from a single subject (n = 43). Rows represent the OTU. Cell shade signifies the number of sequence reads matching to the specified OTU.
154 CHAPTER 6 DISCUSSION AND CONCLUSION The primary purpose of this w ork was to test the hypothesis that microbiota contributes to colorectal carcinogenesis. This hypothesis was tested by determining differences in gut microbiota composition between subjects with large intestinal polyps and those without polyps. Previous wo rk has determined gut microbiota associations between CRC patients and healthy counterparts, but has not examined subjects in the pre cancerous polyp stage (2, 204, 211) The results from this study are importa nt in that they provide information about specific bacterial groups that may potentially be linked to polyp formation, as studies involving diseased subjects cannot determine whether the disease developed first and subsequently contributed to the gut micro biota environment or vice versa. In this study differences between racial groups in microbiota composition were detected at the OTU level. Observed differences between cases and controls were also expected as data from the literature on CRC subjects sugges ts an association between the gut microbiota and the CRC microenvironment (26, 207, 215, 226) However, in contrast to previous studies we did not detect differences in bifidobacteria, which has been suggested to be protective against CRC (172, 183) This may be due to differences in methodologi es, study designs, or the disease state of the subjects (pre cancerous vs. developed CRC). The results of this study may be affected by the limitations of each of the methodologies used. However, a variety of microbiota analysis methods were used to minimi ze the effects of method specific biases. DGGE is limited in that it does not provide taxonomic classification of organisms, nor does it determine absolute
155 abundance. The 454 pyrosequencing generated roughly 250,000 sequence reads; however, gut luminal con tents have been reported to contain approximately 10 12 organisms per gram (203) Clearly, only a subset prevalent bacteria present in the gut has been sequenced. Many rare organisms are not detected using this method, though these organisms may not be large contributors to disease development. qPCR is a more specific method, but is limited in that the re are not currently designed primer sets for each gut organism, and individual reactions for every organism, if possible, would be extremely laborious. Furthermore, PCR reactions used to generate amplicons introduce a bias as the universal primer sets, wh en used to calculate proportions of targeted bacterial groups, may have differing affinities for different 16s rD NA sequences. However, bacterial groups that are present at low abundances, such as lactic acid bacteria, are more accurately quantified by qPC R verses 16S sequencing. Fluorescence in situ hybridization techniques can be used, but similar to qPCR, probes for all gut organisms have not been developed. With the increase in sequencing capacities, the bioinformatics of analyzing the massive amounts of sequence data generated has become an important concern. Sequence reads must be grouped by similarity level into bins. Bins can be classified using taxonomy dependent approaches which rely on matching representative sequences to a database. This method introduces a limitation in that only organism s present in the database can be match ed Taxonomy independent methods can be used, but this creates difficulty in comparing results across studies in that no common label is used. Binning algorithms differ betw een programs used to process sequencing data, and data have shown that different results can be obtained from a single dataset
156 based on the binning procedure used. A hierarchical clustering algorithm can be used, which may produce more accurate bin assignm ents, but may require too much computational power for large data sets A greedy heuristic algorithm can be executed with less memory but will produce different results depending on the seed sequence used. Determining the percent similarity between sequen ce reads within the data set can be done different ways, which can also bias results. A multiple sequence alignment can be used, which may produce more accurate similarity assignments across the entire data set, but it is more computationally costly. A pai r wise sequence alignment is faster but does not consider all sequence reads simultaneously Furthermore, using different similarity thresholds (e.g., 95% identity vs. 98% identity) can produce incon sistent binning results. Until a common binning procedure sequence identity procedure, and a comprehensive database are used among studies, specific results at the OTU and phylotype levels may not always be consistent between studies. Another consideration when interpreting these results is geographical locatio n, which has been shown to associate with gut micro bial community composition. These studies were conducted at only two study sites; therefore, generalization from these results applied to the global population would not be appropriate. Combining these res ults with other studies in meta analyses is also difficult due to differences in methodologies. To further elucidate the associations between gut microbiota and CRC development, a prospective cohort study such as this one would need to continue until a num ber of subjects developed CRC such that there was sufficient power to detect
157 differences between these subjects and cancer free controls. Because of the low incidence of CRC of roughly 45 per 100,000 ( http://www.cdc.gov/cancer/colorectal/statistics/ ), the sample size needed to achieve the power to detect meaningful associations would be nearly 20,000 to achieve a power of 0.80. A study of this scale would require years of follow up and an abundance of funding. Further development of next generation DNA sequ encing techniques may provide the tools necessary to detect associations between less abundant members of the gut microbiota and CRC development. The results from the diet studies demonstrate the ability of fibrous dietary substrates to alter microbiota co mmunity composition as expected ; however, these alterations did not clearly shape gut microbiota community composition to resemble that of polyp free subjects. Therefore, the hypothesis that dietary fiber intake contributes to the development of a more pro tective gut microbiota composition cannot be confirmed. A large limitation of comparing the results from the diet studies to those of the polyp study is that the controls were not the same. A more appropriate experiment would examine the changes in microbi ota composition resulting from dietary fiber intake in subjects with previously polyps and matched cases to determine if dietary fibers help prevent future polyp formation Results in the literature often do not report the types of fiber investigated or d ifferentiate between them. This work used specific substrates as opposed to foods that contain various types and amounts of dietary fiber and showed that the affects of fiber on OTU enrichment and depletion are dependent upon the type of fiber supplemente d ; however it is difficult to relate the specific results of this work to the generalized
158 epidemiological results in the literature. Furthermore, potential confounding variables, including dietary intake other than fiber, are often not addressed in the li terature and are difficult to control. In conclusion, gut microbiota appears to be assoc iated with polyp prevalence. Th is work is significant in that the observed differences in microbiota diversity or the presence of indicator bacteria in this study migh t represent a novel target for future CRC screening tests. Diet mediated changes in gut microbiota composition may be one method to reduce CRC risk, but the data from this work are inconclusive regarding the affective varieties and amounts of dietary fiber s Though it i s now clear that there are differences in gut microbiota community composition between subjects with polyps and those without polyps, further studies should be done to determine the effects of dietary fiber on polyp development over time.
159 LIST OF REFERENCES 1. Abdulamir AS, Hafidh RR, Abu Bakar F 2010. Molecular detection, quantification, and isolation of Streptococcus gallolyticus bacteria colonizing colorectal tumors: inflammation driven p otential of carcinogenesis via IL 1, COX 2, and IL 8. Molecular Cancer 9 2. Abdulamir AS, Hafidh RR, Mahdi LK, Al Jeboori T, Abubaker F 2009. Investigation into the controversial association of Streptococcus gallolyticus with colorectal cancer and aden oma. Bmc Cancer 9 3. Adom KK, Sorrells ME, Liu RH 2005. Phytochemicals and antioxidant activity of milled fractions of different wheat varieties. Journal of Agricultural and Food Chemistry 53 :2297 2306. 4. Al Tassan N, Chmiel NH, Maynard J, Fleming N Livingston AL, Williams GT, Hodges AK, Davies DR, David SS, Sampson JR, Cheadle JR 2002. Inherited variants of MYH associated with somatic G : C > T : A mutations in colorectal tumors. Nature Genetics 30 :227 232. 5. Andersson AF, Lindberg M, Jakobsso n H, Bckhed F, Nyrn P, Engstrand L 2008. Comparative analysis of human gut microbiota by barcoded pyrosequencing Public Library of Science 3 :1 8. 6. Andreasen MF, Kroon PA, Williamson G, Garcia Conesa MT 2001. Intestinal release and uptake of phenol ic antioxidant diferulic acids. Free Radical Biology and Medicine 31 :304 314. 7. Antequera F, Bird A 1993. Number of Cpg Islands and Genes in Human and Mouse. Proceedings of the National Academy of Sciences of the United States of America 90 :11995 11999 8. Antharam VC, Li E, Ishmael A, Sharma A, Mai V, Rand KH, Wang GP 2013. Gut dysbiosis and depletion of butyrate producing organisms in Clostridium difficile infection Proceedings of the National Academy of Sciences of the United States of America Su bmitted 9. Arslanoglu S, Moro GE, Boehm G 2007. Early supplementation of prebiotic oligosaccharides protects formula fed infants against infections during the first 6 months of life. Journal of Nutrition 137 :2420 2424. 10. Arthur J C Perez Chanon a E Muhlbauer M Tomkovich S Uronis J M Fan T Campbell B J Abujamel T Dogan B Rogers A B Rhodes J M Stintzi A Simpson K W Hansen J J, Keku T O Fodor A A Jobin C. Intestinal Inflammation Targets Cancer Inducing Activity of the Microbiota. 20 12. Science
160 11. Arumugam M, Raes J, Pelletier E, Le Paslier D, Yamada T, Mende DR, Fernandes GR, Tap J, Bruls T, Batto JM, Bertalan M, Borruel N, Casellas F, Fernandez L, Gautier L, Hansen T, Hattori M, Hayashi T, Kleer ebezem M, Kurokawa K, Leclerc M, Levenez F, Manichanh C, Nielsen HB, Nielsen T, Pons N, Poulain J, Qin JJ, Sicheritz Ponten T, Tims S, Torrents D, Ugarte E, Zoetendal EG, Wang J, Guarner F, Pedersen O, de Vos WM, Brunak S, Dore J, Weissenbach J, Ehrlich SD Bork P 2011. Enterotypes of the human gut microbiome. Nature 473 :174 180. 12. Atarashi K, Tanoue T, Shima T, Imaoka A, Kuwahara T, Momose Y, Cheng GH, Yamasaki S, Saito T, Ohba Y, Taniguchi T, Takeda K, Hori S, Ivanov II, Umesaki Y, Itoh K, Honda K 2 011. Induction of Colonic Regulatory T Cells by Indigenous Clostridium Species. Science 331 :337 341. 13. Attene Ramos MS, Nava GM, Muellner MG, Wagner ED, Plewa MJ, Gaskins HR 2010. DNA Damage and Toxicogenomic Analyses of Hydrogen Sulfide in Human Inte stinal Epithelial FHs 74 Int Cells. Environmental and Molecular Mutagenesis 51 :304 314. 14. Baker SJ, Markowitz S, Fearon ER, Willson JKV, Vogelstein B 1990. Suppression of Human Colorectal Carcinoma Cell Growth by Wild Type P53. Science 249 :912 915. 1 5. Baker SJ, Preisinger AC, Jessup JM, Paraskeva C, Markowitz S, Willson JKV, Hamilton S, Vogelstein B 1990. P53 Gene Mutations Occur in Combination with 17P Allelic Deletions As Late Events in Colorectal Tumorigenesis. Cancer Research 50 :7717 7722. 16. Barcenilla A, Pryde SE, Martin JC, Duncan SH, Stewart CS, Henderson C, Flint HJ 2000. Phylogenetic relationships of butyrate producing bacteria from the human gut. Applied and Environmental Microbiology 66 :1654 1661. 17. Bertagnolli MM, Warren RS, Nie dzwiecki D, Mueller E, Compton CC, Redston M, Hall M, Hahn HP, Jewell SD, Mayer RJ, Goldberg RM, Saltz LB, Loda M 2009. p27(Kip1) in Stage III Colon Cancer: Implications for Outcome following Adjuvant Chemotherapy in Cancer and Leukemia Group B Protocol 8 9803. Clinical Cancer Research 15 :2116 2122. 18. Biagi E, Nylund L, Candela M, Ostan R, Bucci L, Pini E, Nikkila J, Monti D, Satokari R, Franceschi C, Brigidi P, De Vos W 2010. Through Ageing, and Beyond: Gut Microbiota and Inflammatory Status in Senior s and Centenarians. Plos One 5 19. Biagi E, Nylund L, Candela M, Ostan R, Bucci L, Pini E, Nikkila J, Monti D, Satokari R, Franceschi C, Brigidi P, De Vos W 2010. Through Ageing, and Beyond: Gut Microbiota and Inflammatory Status in Seniors and Centena rians. Plos One 5
161 20. Biarc J, Nguyen IS, Pini A, Gosse F, Richert S, Thierse D, Van Dorsselaer A, Leize Wagner E, Raul F, Klein JP, Scholler Guinard M 2004. Carcinogenic properties of proteins with pro inflammatory activity from Streptococcus infantar ius (formerly S.bovis). Carcinogenesis 25 :1477 1484. 21. Binder, H. J., W. F. Boron, and E. L. Boulpaep 2003. Organization of the gastrointestinal system, p. 879 890. In W. R. Schmitt and M. Dudlick (eds.), Medical Physiology. Saunders, Philadelphia. 2 2. Bingham SA, Day EN, Luben R, Ferrari P 2003. Dietary fibre in food and protection against colorectalcancer in the European Prospective Investigation into Cancer and Nutrition (EPIC): an observational study. the Lancet 361 : 1496 1501. 23. Bird A 20 02. DNA methylation patterns and epigenetic memory. Genes & Development 16 :6 21. 24. Bird RP 1995. Role of Aberrant Crypt Foci in Understanding the Pathogenesis of Colon Cancer. Cancer Letters 93 :55 71. 25. Boland CR, Koi M, Chang DK, Carethers JM 20 08. The biochemical basis of microsatellite instability and abnormal immunohistochemistry and clinical behavior in Lynch Syndrome: from bench to bedside. Familial Cancer 7 :41 52. 26. Boleij A, van Gelder MMHJ, Swinkels DW, Tjalsma H 2011. Clinical Impor tance of Streptococcus gallolyticus Infection Among Colorectal Cancer Patients: Systematic Review and Meta analysis. Clinical Infectious Diseases 53 :870 878. 27. Bordonaro M, Lazarova DL, Sartorelli AC 2008. Butyrate and Wnt signaling A possible solut ion to the puzzle of dietary fiber and colon cancer risk? Cell Cycle 7 :1178 1183. 28. Bordonaro M, Mariadason JM, Aslam F, Heerdt BG, Augenlicht LH 1999. Butyrate induced apoptotic cascade in colonic carcinoma cells: Modulation of the beta catenin Tcf p athway and concordance with effects of sulindac and trichostatin a but not curcumin. Cell Growth & Differentiation 10 :713 720. 29. Bouhnik Y, Raskine L, Simoneau G, Vicaut E, Neut C, Flourie B, Brouns F, Bornet FR 2004. The capacity of nondigestible car bohydrates to stimulate fecal bifidobacteria in healthy humans: a double blind, randomized, placebo controlled, parallel group, dose response relation study. American Journal of Clinical Nutrition 80 :1658 1664. 30. Brabletz T, Herrmann K, Jung A, Faller G, Kirchner T 2000. Expression of nuclear beta catenin and c myc is correlated with tumor size but not with proliferative activity of colorectal adenomas. American Journal of Pathology 156 :865 870.
162 31. Breivik J, Meling GI, Spurkland A, Rognum TO, Gaude rnack G 1994. K Ras Mutation in Colorectal Cancer Relations to Patient Age, Sex and Tumor Location. British Journal of Cancer 69 :367 371. 32. Brenchley JM, Douek DC 2012. Microbial Translocation Across the GI Tract. Annual Review of Immunology, Vol 3 0 30 :149 173. 33. Bronowski C, Smith SL, Yokota K, Corkill JE, Martin HM, Campbell BJ, Rhodes JM, Hart CA, Winstanley C 2008. A subset of mucosa associated Escherichia coli isolates from patients with colon cancer, but not Crohn's disease, share pathoge nicity islands with urinary pathogenic E coli. Microbiology Sgm 154 :571 583. 34. Brouns F, Arrigoni E, Langkilde AM, Verkooijen I, Fassler C, Andersson H, Kettlitz B, van Nieuwenhoven M, Philipsson H, Amado R 2007. Physiological and metabolic properties of a digestion resistant maltodextrin, classified as type 3 retrograded resistant starch. Journal of Agricultural and Food Chemistry 55 :1574 1581. 35. Buck, A. W. 2012. Resistant Maltodextrin Overview, p. 279 292. In S. Cho and N. Almeida (eds.), Dietar y Fiber and Health. CRC Press, Boca Raton. 36. Burns AJ, Rowland IR 2004. Antigenotoxicity of probiotics and prebiotics on faecal water induced DNA damage in human colon adenocarcinoma cells. Mutation Research Fundamental and Molecular Mechanisms of Mut agenesis 551 :233 243. 37. Burri N, Shaw P, Bouzourene H, Sordat I, Sordat B, Gillet M, Schorderet D, Bosman FT, Chaubert P 2001. Methylation silencing and mutations of the p14(ARF) and p16(INK4a) genes in colon cancer. Laboratory Investigation 81 :217 22 9. 38. Butler LM, Sinha R, Millikan RC, Martin CF, Newman B, Gammon MD, Ammerman AS, Sandler RS 2003. Heterocyclic amines, meat intake, and association with colon cancer in a population based study. American Journal of Epidemiology 157 :434 445. 39. Ca meron EE, Bachman KE, Myohanen S, Herman JG, Baylin SB 1999. Synergy of demethylation and histone deacetylase inhibition in the re expression of genes silenced in cancer. Nature Genetics 21 :103 107. 40. Caporaso JG, Kuczynski J, Stombaugh J, Bittinger K Bushman FD, Costello EK, Fierer N, Pena AG, Goodrich JK, Gordon JI, Huttley GA, Kelley ST, Knights D, Koenig JE, Ley RE, Lozupone CA, McDonald D, Muegge BD, Pirrung M, Reeder J, Sevinsky JR, Turnbaugh PJ, Walters WA, Widmann J, Yatsunenko T, Zaneveld J, Knight R 2010. QIIME allows analysis of high throughput community sequencing data. Nat.Methods 7 :335 336.
163 41. Carvalho Wells AL, Helmolz K, Nodet C, Molzer C, Leonard C, McKevith B, Thielecke F, Jackson KG, Tuohy KM 2010. Determination of the in vivo p rebiotic potential of a maize based whole grain breakfast cereal: a human feeding study. British Journal of Nutrition 104 :1353 1356. 42. Castellarin M, Warren RL, Freeman JD, Dreolini L, Krzywinski M, Strauss J, Barnes R, Watson P, Allen Vercoe E, Moore RA, Holt RA 2012. Fusobacterium nucleatum infection is prevalent in human colorectal carcinoma. Genome Research 22 :299 306. 43. Castellone MD, Teramoto H, Williams BO, Druey KM, Gutkind JS 2005. Prostaglandin E 2 promotes colon cancer cell growth throu gh a G(s) axin beta catenin signaling axis. Science 310 :1504 1510. 44. Chan AT, Ogino S, Fuchs CS 2007. Aspirin and the risk of colorectal cancer in relation to the expression of COX 2. New England Journal of Medicine 356 :2131 2142. 45. Chao A, Thun M J, Connell CJ, McCullough ML, Jacobs EJ, Flanders WD, Rodriguez C, Sinha R, Calle EE 2005. Meat consumption and risk of colorectal cancer. Jama Journal of the American Medical Association 293 :172 182. 46. Chao A, Thun MJ, Jacobs EJ, Henley SJ, Rodriguez C, Calle EE 2000. Cigarette smoking and colorectal cancer mortality in the cancer prevention study II. Journal of the National Cancer Institute 92 :1888 1896. 47. Chen WG, Liu FL, Ling ZX, Tong XJ, Xiang C 2012. Human Intestinal Lumen and Mucosa Associ ated Microbiota in Patients with Colorectal Cancer. Plos One 7 48. Chen WS, Wei SJ, Liu JM, Hsiao M, Lin JK, Yang WK 2001. Tumor invasiveness and liver metastasis of colon cancer cells correlated with cyclooxygenase 2 (COX 2) expression and inhibited b y a COX 2 selective inhibitor, etodolac. International Journal of Cancer 91 :894 899. 49. Chieppa M, Rescigno M, Huang AYC, Germain RN 2006. Dynamic imaging of dendritic cell extension into the small bowel lumen in response to epithelial cell TLR engagem ent. Journal of Experimental Medicine 203 :2841 2852.
164 50. Chlebowski RT, Wactawski Wende J, Ritenbaugh C, Hubbell FA, Ascensao J, Rodabough RJ, Rosenberg CA, Taylor VM, Harris R, Chen C, Adams Campbell LL, White E, Alving B, Rossouw J, Pottern L, Ludlam S McGowan J, Prentice R, Anderson G, LaCroix A, Patterson R, McTiernan A, Cochrane B, Hunt J, Tinker L, Kooperberg C, McIntosh M, Wang CY, Chen C, Bowen D, Kristal A, Stanford J, Urban N, Weiss N, White E, Shumaker S, Rautaharju P, Prineas R, Naughton M, S tein E, Laskarzewski P, Cummings S, Nevitt M, Dockrell M, Harnack L, Cammarata F, Lindenfelser S, Psaty B, Heckbert S, Wassertheil Smoller S, Frishman W, Wylie Rosett J, Barad D, Freeman R, Hays J, Young R, Anderson J, Lithgow S, Bray P, Manson J, Buring J Gaziano JM, Rexrode K, Chae C, Assaf AR, Wheeler C, Eaton C, Cyr M, Phillips L, Pedersen M, Strickland O, Huber M, Porter V, Beresford SAA, Taylor VM, Woods NF, Henderson M, Kestin M, Hsia J, Gaba N, Ascensao J, Chlebowski R, Detrano R, Nelson A, Heiner J, Marshall J, Ritenbaugh C, Valanis B, Elmer P, Stevens V, Karanja N, Caan B, Sidney S, Bailey G, Hirata J, Kotchen JM, Barnabei V, Kotchen TA, Gilligan MAC, Neuner J, Howard BV, Adams Campbell L, Passaro M, Rainford M, Agurs Collins T, Van Horn L, Greenl and P, Khandekar J, Liu K, Rosenberg C, Black H, Powell L, Mason E, Stefanick ML, Hlatky MA, Chen B, Stafford RS, Giudice LC, Lane D, Granek I, Lawson W, San Roman G, Messina C, Jackson R, Harris R, Paskett E, Mysiw WJ, Blumenfeld M, Lewis CE, Oberman A, S hikany JM, Safford M, Britt BK, Bassford T, Mattox J, Ko M, Lohman T, Wactawski Wende J, Trevisan M, Smit E, Graham S, Chang J, Robbins J, Yasmeen S, Lindfors K, Stern J, Hubbell A, Frank G, Wong N, Greep N, Monk B, Judd H, Heber D, Elashoff R, Langer RD, Criqui MH, Talavera GT, Garland CF, Hanson RE, Gass M, Wernke S, Watts N, Limacher M, Perri M, Kaunitz A, Williams RS, Brinson Y, Curb D, Petrovitch H, Rodriguez B, Masaki K, Sharma S, Wallace R, Torner J, Johnson S, Snetselaar L, VanVoorhis B, Ockene J, R osal M, Ockene I, Yood R, Aronson P, Lasser N, Singh B, Lasser V, Kostis J, O'Sullivan MJ, Parker L, Estape R, Fernandez D, Margolis KL, Grimm RH, Hunninghake DB, LaValleur J, Kempainen S, Brunner R, Graettinger W, Oujevolk V, Heiss G, Haines P, Ontjes D, Sueta C, Wells E, Kuller L, Cauley J, Milas NC, Johnson KC, Satterfield S, Ke RW, Vile J, Tylavsky F, Brzyski R, Schenken R, Trabal J, Rodriguez Sifuentes M, Mouton C, Sarto G, Laube D, McBride P, Mares Perlman J, Loevinger B, Bonds D, Burke G, Crouse R, P arsons L, Vitolins M, Hendrix S, Simon M, McNeely G, Gordon P, Makela P, Allen C, Dougherty S, Carleton R 2004. Estrogen plus progestin and colorectal cancer in postmenopausal women. New England Journal of Medicine 350 :991 1004. 51. Cho E, Smith Warner SA, Spiegelman D, Beeson WL, van den Brandt PA, Colditz GA, Folsom AR, Fraser GE, Freudenheim JL, Giovannucci E, Goldbohm RA, Graham S, Miller AB, Pietinen P, Potter JD, Rohan TE, Terry P, Toniolo P, Virtanen MJ, Willett WC, Wolk A, Wu K, Yaun SS, Zeleniuc h Jacquotte A, Hunter DJ 2004. Dairy foods, calcium, and colorectal cancer: A pooled analysis of 10 cohort studies. Journal of the National Cancer Institute 96 :1015 1022.
165 52. Colbert LH, Mai V, Perkins SN, Berrigan D, Lavigne JA, Wimbrow HH, Alvord WG, Haines DC, Srinivas P, Hursting SD 2003. Exercise and intestinal polyp development in APC(Min) mice. Medicine and Science in Sports and Exercise 35 :1662 1669. 53. Coller HA, Grandori C, Tamayo P, Colbert T, Lander ES, Eisenman RN, Golub TR 2000. Expres sion analysis with oligonucleotide microarrays reveals that MYC regulates genes involved in growth, cell cycle, signaling, and adhesion. Proceedings of the National Academy of Sciences of the United States of America 97 :3260 3265. 54. Colucci PM, Yale SH Rall CJ 2003. Colorectal Polyps. Clinical Medicine and Research 1 :261 262. 55. Cosgrove DE, Cox GS 1990. Effects of Sodium Butyrate and 5 Azacytidine on Dna Methylation in Human Tumor Cell Lines Variable Response to Drug Treatment and Withdrawal. B iochimica et Biophysica Acta 1087 :80 86. 56. Costabile A, Klinder A, Fava F, Napolitano A, Foglian V, Leonard C, Gibson GR, Tuohy KM 2008. Whole grain wheat breakfast cereal has a prebiotic effect on the human gut microbiota: a double blind, placebo con trolled, crossover study. British Journal of Nutrition 99 :110 120. 57. Costello EK, Lauber CL, Hamady M, Fierer N, Gordon JI, Knight R 2009. Bacterial community variation in human body habitats across space and time Science 326 :1694 1697. 58. Cuevas Ramos G, Petit CR, Marcq I, Boury M, Oswald E, Nougayrede JP 2010. Escherichia coli induces DNA damage in vivo and triggers genomic instability in mammalian cells. Proceedings of the National Academy of Sciences of the United States of America 107 :11537 1 1542. 59. Dahm CC, Keogh RH, Spencer EA, Greenwood DC, Key TJ, Fentiman IS, Shipley MJ, Brunner EJ, Cade JE, Burley VJ, Mishra G, Stephen AM, Kuh D, White IR, Luben R, Lentjes MAH, Khaw KT, Rodwell SA 2010. Dietary Fiber and Colorectal Cancer Risk: A Ne sted Case Control Study Using Food Diaries. Journal of the National Cancer Institute 102 :614 626. 60. Dang CV 1999. c myc target genes involved in cell growth, apoptosis, and metabolism. Molecular and Cellular Biology 19 :1 11. 61. Darfeuille Michaud A Boudeau J, Bulois P, Neut C, Glasser AL, Barnich N, Bringer MA, Swidsinski A, Beaugerie L, Colombel JF 2004. High prevalence of adherent invasive Escherichia coli associated with ileal mucosa in Crohn's disease. Gastroenterology 127 :412 421.
166 62. Davie s H, Bignell GR, Cox C, Stephens P, Edkins S, Clegg S, Teague J, Woffendin H, Garnett MJ, Bottomley W, Davis N, Dicks N, Ewing R, Floyd Y, Gray K, Hall S, Hawes R, Hughes J, Kosmidou V, Menzies A, Mould C, Parker A, Stevens C, Watt S, Hooper S, Wilson R, J ayatilake H, Gusterson BA, Cooper C, Shipley J, Hargrave D, Pritchard Jones K, Maitland N, Chenevix Trench G, Riggins GJ, Bigner DD, Palmieri G, Cossu A, Flanagan A, Nicholson A, Ho JWC, Leung SY, Yuen ST, Weber BL, Siegler HF, Darrow TL, Paterson H, Marai s R, Marshall CJ, Wooster R, Stratton MR, Futreal PA 2002. Mutations of the BRAF gene in human cancer. Nature 417 :949 954. 63. Davis LMG, Martinez I, Walter J, Goin C, Hutkins RW 2011. Barcoded Pyrosequencing Reveals That Consumption of Galactooligosac charides Results in a Highly Specific Bifidogenic Response in Humans. Plos One 6 64. Davis LMG, Martinez I, Walter J, Hutkins R 2010. A dose dependent impact of prebiotic galactooligosaccharides on the intestinal microbiota of healthy adults. Internati onal Journal of Food Microbiology 144 :285 292. 65. De Palma G, Nadal I, Medina M, Donat E, Ribes Koninckx C, Calabuig M, Sanz Y 2010. Intestinal dysbiosis and reduced immunoglobulin coated bacteria associated with coeliac disease in children. Bmc Microb iology 10 66. Deneo Pellegrini H, Boffetta P, De Stefani E, Ronco A, Brennan P, Mendilaharsu M 2002. Plant foods and differences between colon and rectal cancers. European Journal of Cancer Prevention 11 :369 375. 67. Denning TL, Wang YC, Patel SR, Wi lliams IR, Pulendran B 2007. Lamina propria macrophages and dendritic cells differentially induce regulatory and interleukin 17 producing T cell responses. Nature Immunology 8 :1086 1094. 68. Depeint F, Tzortzis G, Vulevic J, I'Anson K, Gibson GR 2008. Prebiotic evaluation of a novel galactooligosaccharide mixture produced by the enzymatic activity of Bifidobacterium bifidum NCIMB 41171, in healthy humans: a randomized, double blind, crossover, placebo controlled intervention study. American Journal of C linical Nutrition 87 :785 791. 69. Doll R, Peto R 1981. The Causes of Cancer Quantitative Estimates of Avoidable Risks of Cancer in the United States Today. Journal of the National Cancer Institute 66 :1191 &. 70. Durante Mangoni E, Bradley S, Selton Suty C, Tripodi MF, Barsic B, Bouza E, Cabell CH, Ramos AIDO, Fowler V, Hoen B, Konecny P, Moreno A, Murdoch D, Pappas P, Sexton DJ, Spelman D, Tattevin P, Miro JM, van der Meer JTM, Utili R 2008. Current Features of Infective Endocarditis in Elderly Pati ents Results of the International Collaboration on Endocarditis Prospective Cohort Study. Archives of Internal Medicine 168 :2095 2103.
167 71. Eckburg PB, Bik EM, Bernstein CN, Purdom E, Dethlefsen L, Sargent M, Gill SR, Nelson KE, Relman DA 2005. Diversi ty of the human intestinal microbial flora. Science 308 :1635 1638. 72. Ellmerich S, Scholler M, Duranton B, Gosse F, Galluser M, Klein JP, Raul F 2000. Promotion of intestinal carcinogenesis by Streptococcus bovis. Carcinogenesis 21 :753 756. 73. Elson CO, Cong YZ, Weaver CT, Schoeb TR, McClanahan TK, Fick RB, Kastelein RA 2007. Monoclonal anti interleukin 23 reverses active colitis in a T cell mediated model in mice. Gastroenterology 132 :2359 2370. 74. English DR, MacInnis RJ, Hodge AM, Hopper JL, H aydon AM, Giles GG 2004. Red meat, chicken, and fish consumption and risk of colorectal cancer. Cancer Epidemiology Biomarkers & Prevention 13 :1509 1514. 75. Eppert K, Scherer SW, Ozcelik H, Pirone R, Hoodless P, Kim H, Tsui LC, Bapat B, Gallinger S, An drulis IL, Thomsen GH, Wrana JL, Attisano L 1996. MADR2 maps to 18q21 and encodes a TGF beta regulated MAD related protein that is functionally mutated in colorectal carcinoma. Cell 86 :543 552. 76. Esteller M, Fraga MF, Guo MZ, Garcia Foncillas J, Heden falk I, Godwin AK, Trojan J, Vaurs Barriere C, Bignon YJ, Ramus S, Benitez J, Caldes T, Akiyama Y, Yuasa Y, Launonen V, Canal MJ, Rodriguez R, Capella G, Peinado MA, Borg A, Aaltonen LA, Ponder BA, Baylin SB, Herman JG 2001. DNA methylation patterns in he reditary human cancers mimic sporadic tumorigenesis. Human Molecular Genetics 10 :3001 3007. 77. Faith JJ, McNulty NP, Rey FE, Gordon JI 2011. Predicting a Human Gut Microbiota's Response to Diet in Gnotobiotic Mice. Science 333 :101 104. 78. Fasano A 2012. Leaky Gut and Autoimmune Diseases. Clinical Reviews in Allergy & Immunology 42 :71 78. 79. Fastinger ND, Karr Lilienthal LK, Spears JK, Swanson KS, Zinn KE, Nava GM, Ohkuma K, Kanahori S, Gordon DT, Fahey GC 2008. A novel resistant maltodextrin alt ers gastrointestinal tolerance factors, fecal characteristics, and fecal microbiota in healthy adult humans. Journal of the American College of Nutrition 27 :356 366. 80. Feinberg AP, Vogelstein B 1983. Hypomethylation of Ras Oncogenes in Primary Human C ancers. Biochemical and Biophysical Research Communications 111 :47 54.
168 81. Femia AP, Luceri C, Dolara P, Giannini A, Biggeri A, Salvadori M, Clune Y, Collins KJ, Paglierani M, Caderni G 2002. Antitumorigenic activity of the prebiotic inulin enriched wit h oligofructose in combination with the probiotics Lactobacillus rhamnosus and Bifidobacterium lactis on azoxymethane induced colon carcinogenesis in rats. Carcinogenesis 23 :1953 1960. 82. Flood A, Rastogi T, Wirfalt E, Mitrou PN, Reedy J, Subar AF, Kipn is V, Mouw T, Hollenbeck AR, Leitzmann M, Schatzkin A 2008. Dietary patterns as identified by factor analysis and colorectal cancer among middle aged Americans. American Journal of Clinical Nutrition 88 :176 184. 83. Fodde R, Smits R, Clevers H 2001. AP C, signal transduction and genetic instability in colorectal cancer. Nature Reviews Cancer 1 :55 67. 84. Fukiya S, Arata M, Kawashima H, Yoshida D, Kaneko M, Minamida K, Watanabe J, Ogura Y, Uchida K, Itoh K, Wada M, Ito S, Yokota A 2009. Conversion of c holic acid and chenodeoxycholic acid into their 7 oxo derivatives by Bacteroides intestinalis AM 1 isolated from human feces. Fems Microbiology Letters 293 :263 270. 85. Gaboriau Routhiau V, Rakotobe S, Lecuyer E, Mulder I, Lan A, Bridonneau C, Rochet V, Pisi A, De Paepe M, Brandi G, Eberl G, Snel J, Kelly D, Cerf Bensussan N 2009. The Key Role of Segmented Filamentous Bacteria in the Coordinated Maturation of Gut Helper T Cell Responses. Immunity 31 :677 689. 86. Gibson GR, Roberfroid MB 1995. Dietary modulation of the human colonic microbiota: Introducing the concept of prebiotics The Journal of Nutrition 125 :1401 1412. 87. Gill SR, Pop M, DeBoy RT, Eckburg PB, Turnbaugh PJ, Samuel BS, Gordon JI, Relman DA, Fraser Liggett CM, Nelson KE 2006. Metage nomic analysis of the human distal gut microbiome Science 312 :1355 1359. 88. Gionchetti P, Rizzello F, Helwig U, Venturi A, Lammers KM, Brigidi P, Vitali B, Poggioli G, Miglioli M, Campieri M 2003. Prophylaxis of pouchitis onset with probiotic therapy: A double blind, placebo controlled trial. Gastroenterology 124 :1202 1209. 89. Glocker EO, Kotlarz D, Boztug K, Gertz EM, Schaffer AA, Noyan F, Perro M, Diestelhorst J, Allroth A, Murugan D, Hatscher N, Pfeifer D, Sykora KW, Sauer M, Kreipe H, Lacher M, Nustede R, Woellner C, Baumann U, Salzer U, Koletzko S, Shah N, Segal AW, Sauerbrey A, Buderus S, Snapper SB, Grimbacher B, Klein C 2009. Inflammatory Bowel Disease and Mutations Affecting the Interleukin 10 Receptor. New England Journal of Medicine 361 :2 033 2045.
169 90. Gopal PK, Prasad J, Gill HS 2003. Effects of the consumption of Bifidobacterium lactis HN019 (DR10 (TM)) and galacto oligosaccharides on the microflora of the gastrointestinal tract in human subjects. Nutrition Research 23 :1313 1328. 91. Gopal PK, Sullivan PA, Smart JB 2001. Utilisation of galacto oligosaccharides as selective substrates for growth by lactic acid bacteria including Bifidobacterium lactis DR10 and Lactobacillus rhamnosus DR20. International Dairy Journal 11 :19 25. 92. G ore C, Munro K, Lay C, Bibiloni R, Morris J, Woodcock A, Custovic A, Tannock GW 2008. Bifidobacterium pseudocatenulatum is associated with atopic eczema: A nested case control study investigating the fecal microbiota of infants. Journal of Allergy and Cli nical Immunology 121 :135 140. 93. Gorham ED, Garland CF, Garland FC, Grant WB, Mohr SB, Lipkin M, Newmark HL, Giovannucci E, Wei M Holick M F Vitamin D and prevention of colorectalcancer. The Journal of Steroid Biochem istry and Molecular Biology 97 : 179 194. 94. Goto T, Mizukami H, Shirahata A, Sakata M, Saito M, Ishibashi K, Kigawa G, Nemoto H, Sanada Y, Hibi K 2009. Aberrant Methylation of the p16 Gene Is Frequently Detected in Advanced Colorectal Cancer. Anticancer Research 29 :275 277. 95. Grady WM, Myeroff LL, Swinler SE, Rajput A, Thiagalingam S, Lutterbaugh JD, Neumann A, Brattain MG, Chang J, Kim SJ, Kinzler KW, Vogelstein B, Willson JKV, Markowitz S 1999. Mutational inactivation of transforming growth factor beta receptor type II in microsa tellite stable colon cancers. Cancer Research 59 :320 324. 96. Greten FR, Eckmann L, Greten TF, Park JM, Li ZW, Egan LJ, Kagnoff MF, Karin M 2004. IKK beta links inflammation and tumorigenesis in a mouse model of colitis associated cancer. Cell 118 :285 2 96. 97. Grivennikov S, Karin E, Terzic J, Mucida D, Yu GY, Vallabhapurapu S, Scheller J, Rose John S, Cheroutre H, Eckmann L, Karin M 2009. IL 6 and Stat3 Are Required for Survival of Intestinal Epithelial Cells and Development of Colitis Associated Can cer. Cancer Cell 15 :103 113. 98. Haas A, Zimmermann K, Graw F, Slack E, Rusert P, Ledergerber B, Bossart W, Weber R, Thurnheer MC, Battegay M, Hirschel B, Vernazza P, Patuto N, Macpherson AJ, Gunthard HF, Oxenius A 2011. Systemic antibody responses to g ut commensal bacteria during chronic HIV 1 infection. Gut 60 :1506 1519. 99. Hamady M, Lozupone C, Knight R 2010. Fast UniFrac: facilitating high throughput phylogenetic analyses of microbial communities including analysis of pyrosequencing and PhyloChip data. ISME J 4 :17 27.
170 100. Hamady M, Walker JJ, Harris JK, Gold NJ, Knight R 2008. Error correcting barcoded primers for pyrosequencing hundreds of samples in multiplex. Nat.Methods 5 :235 237. 101. Hangen L, Bennink MR 2002. Consumption of black bea ns and, navy beans (Phaseolus vulgaris) reduced azoxymethane induced colon cancer in rats. Nutrition and Cancer An International Journal 44 :60 65. 102. He TC, Sparks AB, Rago C, Hermeking H, Zawel L, da Costa LT, Morin PJ, Vogelstein B, Kinzler KW 1998. Identification of c MYC as a target of the APC pathway. Science 281 :1509 1512. 103. Herman JG, Umar A, Polyak K, Graff JR, Ahuja N, Issa JPJ, Markowitz S, Willson JKV, Hamilton SR, Kinzler KW, Kane MF, Kolodner RD, Vogelstein B, Kunkel TA, Baylin SB 19 98. Incidence and functional consequences of hMLH1 promoter hypermethylation in colorectal carcinoma. Proceedings of the National Academy of Sciences of the United States of America 95 :6870 6875. 104. Hermeking H, Rago C, Schuhmacher M, Li Q, Barrett JF, Obaya AJ, O'Connell BC, Mateyak MK, Tam W, Kohlhuber F, Dang CV, Sedivy JM, Eick D, Vogelstein B, Kinzler KW 2000. Identification of CDK4 as a target of c MYC. Proceedings of the National Academy of Sciences of the United States of America 97 :2229 2234. 105. Hernot DC, Boileau TW, Bauer LL, Swanson KS, Fahey GC 2008. In Vitro Digestion Characteristics of Unprocessed and Processed Whole Grains and Their Components. Journal of Agricultural and Food Chemistry 56 :10721 10726. 106. Hibi K, Goto T, Kitamur a YH, Yokomizo K, Sakuraba K, Shirahata A, Mizukami H, Saito M, Ishibashi K, Kigawa G, Nemoto H, Sanada Y 2010. Methylation of TFPI2 Gene is Frequently Detected in Advanced Well differentiated Colorectal Cancer. Anticancer Research 30 :1205 1207. 107. Hi bi K, Mizukami H, Shirahata A, Goto T, Sakata M, Saito M, Ishibashi K, Kigawa G, Nemoto H, Sanada Y 2009. Aberrant Methylation of the UNC5C Gene is Frequently Detected in Advanced Colorectal Cancer. Anticancer Research 29 :271 273. 108. Hibi K, Sakata M, Sakuraba K, Shirahata A, Goto T, Mizukami H, Saito M, Ishibashi K, Kigawa G, Nemoto H, Sanada Y 2008. Aberrant methylation of the HACE1 gene is frequently detected in advanced colorectal cancer. Anticancer Research 28 :1581 1584. 109. Higginbotham S, Zh ang ZF, Lee IM, Cook NR, Giovannucci E, Buring JE, Liu SM 2004. Dietary glycemic load and risk of colorectal cancer in the Women's Health Study. Journal of the National Cancer Institute 96 :229 233.
171 110. Holtta V, Klemetti P, Sipponen T, Kociubinski G, W esterholm Ormio M, Solo H, Rasanen L, Kolho KL, Farkkila M, Savilahti E, Vaarala O 2008. IL 23/IL 17 immunity as a hallmark of Crohn's disease. Inflammatory Bowel Diseases 14 :1175 1184. 111. Hooper LV, Gordon JI 2001. Commensal host bacterial relations hips in the gut. Science 292 :1115 1118. 112. Howe GR, Aronson KJ, Benito E, Castelleto R, Cornee J, Duffy S, Gallagher RP, Iscovich JM, DengAo J, Kaaks R, Kune GA, Kune S, Lee HP, Lee M, Miller AB, Peters RK, Potter JD, Riboli E, Slattery ML, Trichopoulo s D, Tuyns A, Tzonou A, Watson LE, Whittemore AS, WuWilliams AH, Shu Z 1997. The relationship between dietary fat intake and risk of colorectal cancer: Evidence from the combined analysis of 13 case control studies. Cancer Causes & Control 8 :215 228.
172 113 Huttenhower C, Gevers D, Knight R, Abubucker S, Badger JH, Chinwalla AT, Creasy HH, Earl AM, FitzGerald MG, Fulton RS, Giglio MG, Hallsworth Pepin K, Lobos EA, Madupu R, Magrini V, Martin JC, Mitreva M, Muzny DM, Sodergren EJ, Versalovic J, Wollam AM, W orley KC, Wortman JR, Young SK, Zeng QD, Aagaard KM, Abolude OO, Allen Vercoe E, Alm EJ, Alvarado L, Andersen GL, Anderson S, Appelbaum E, Arachchi HM, Armitage G, Arze CA, Ayvaz T, Baker CC, Begg L, Belachew T, Bhonagiri V, Bihan M, Blaser MJ, Bloom T, Bo nazzi V, Brooks JP, Buck GA, Buhay CJ, Busam DA, Campbell JL, Canon SR, Cantarel BL, Chain PSG, Chen IMA, Chen L, Chhibba S, Chu K, Ciulla DM, Clemente JC, Clifton SW, Conlan S, Crabtree J, Cutting MA, Davidovics NJ, Davis CC, DeSantis TZ, Deal C, Delehaun ty KD, Dewhirst FE, Deych E, Ding Y, Dooling DJ, Dugan SP, Dunne WM, Durkin AS, Edgar RC, Erlich RL, Farmer CN, Farrell RM, Faust K, Feldgarden M, Felix VM, Fisher S, Fodor AA, Forney LJ, Foster L, Di Francesco V, Friedman J, Friedrich DC, Fronick CC, Fult on LL, Gao HY, Garcia N, Giannoukos G, Giblin C, Giovanni MY, Goldberg JM, Goll J, Gonzalez A, Griggs A, Gujja S, Haake SK, Haas BJ, Hamilton HA, Harris EL, Hepburn TA, Herter B, Hoffmann DE, Holder ME, Howarth C, Huang KH, Huse SM, Izard J, Jansson JK, Ji ang HY, Jordan C, Joshi V, Katancik JA, Keitel WA, Kelley ST, Kells C, King NB, Knights D, Kong HDH, Koren O, Koren S, Kota KC, Kovar CL, Kyrpides NC, La Rosa PS, Lee SL, Lemon KP, Lennon N, Lewis CM, Lewis L, Ley RE, Li K, Liolios K, Liu B, Liu Y, Lo CC, Lozupone CA, Lunsford RD, Madden T, Mahurkar AA, Mannon PJ, Mardis ER, Markowitz VM, Mavromatis K, McCorrison JM, McDonald D, McEwen J, McGuire AL, McInnes P, Mehta T, Mihindukulasuriya KA, Miller JR, Minx PJ, Newsham I, Nusbaum C, O'Laughlin M, Orvis J, P agani I, Palaniappan K, Patel SM, Pearson M, Peterson J, Podar M, Pohl C, Pollard KS, Pop M, Priest ME, Proctor LM, Qin X, Raes J, Ravel J, Reid JG, Rho M, Rhodes R, Riehle KP, Rivera MC, Rodriguez Mueller B, Rogers YH, Ross MC, Russ C, Sanka RK, Sankar P, Sathirapongsasuti JF, Schloss JA, Schloss PD, Schmidt TM, Scholz M, Schriml L, Schubert AM, Segata N, Segre JA, Shannon WD, Sharp RR, Sharpton TJ, Shenoy N, Sheth NU, Simone GA, Singh I, Smillie CS, Sobel JD, Sommer DD, Spicer P, Sutton GG, Sykes SM, Tabb aa DG, Thiagarajan M, Tomlinson CM, Torralba M, Treangen TJ, Truty RM, Vishnivetskaya TA, Walker J, Wang L, Wang ZY, Ward DV, Warren W, Watson MA, Wellington C, Wetterstrand KA, White JR, Wilczek Boney K, Wu YQ, Wylie KM, Wylie T, Yandava C, Ye L, Ye YZ, Y ooseph S, Youmans BP, Zhang L, Zhou YJ, Zhu YM, Zoloth L, Zucker JD, Birren BW, Gibbs RA, Highlander SK, Methe BA, Nelson KE, Petrosino JF, Weinstock GM, Wilson RK, White O 2012. Structure, function and diversity of the healthy human microbiome. Nature 48 6 :207 214. 114. Huycke MM, Abrams V, Moore DR 2002. Enterococcus faecalis produces extracellular superoxide and hydrogen peroxide that damages colonic epithelial cell DNA. Carcinogenesis 23 :529 536.
173 115. Ikeda Y, Mori M, Yoshizumi T, Sugimachi K 1999 Cancer and adenomatous polyp distribution in the colorectum. American Journal of Gastroenterology 94 :191 193. 116. Jarvinen R, Knekt P, Hakulinen T, Rissanen H, Heliovaara M 2001. Dietary fat, cholesterol and colorectal cancer in a prospective study. British Journal of Cancer 85 :357 361. 117. Jass JR 2004. Hyperplastic Polyps and Colorectal Cancer: Is There a Link? Clinical Gastroenterology and Hepatology 2 :1 8. 118. Jen J, Powell SM, Papadopoulos N, Smith KJ, Hamilton SR, Vogelstein B, Kinzler KW 1994. Molecular Determinants of Dysplasia in Colorectal Lesions. Cancer Research 54 :5523 5526. 119. Jenkins GJS, Cronin J, Alhamdani A, Rawat N, D'souza F, Thomas T, Eltahir Z, Griffiths AP, Baxter JN 2008. The bile acid deoxycholic acid has a non lin ear dose response for DNA damage and possibly NF kappa B activation in oesophageal cells, with a mechanism of action involving ROS. Mutagenesis 23 :399 405. 120. Johansson MEV, Larsson JMH, Hansson GC 2011. The two mucus layers of colon are organized by the MUC2 mucin, whereas the outer layer is a legislator of host microbial interactions. Proceedings of the National Academy of Sciences of the United States of America 108 :4659 4665. 121. Jones PA, Baylin SB 2002. The fundamental role of epigenetic even ts in cancer. Nature Reviews Genetics 3 :415 428. 122. Jonnalagadda SS, Harnack L, Liu RH, McKeown N, Seal C, Liu SM, Fahey GC 2011. Putting the Whole Grain Puzzle Together: Health Benefits Associated with Whole Grains Summary of American Society for Nut rition 2010 Satellite Symposium. Journal of Nutrition 141 :1011S 1022S. 123. Kaler P, Godasi BN, Augenlicht L, Klampfer L 2009. The NF k B/AKT dependent Induction of Wnt Signaling in Colon Cancer Cells by Macrophages and IL 1 B Ca ncer Microenvironment 2 :6 9 80. 124. Kameue C, Tsukahara T, Yamada K, Koyama H, Iwasaki Y, Nakayama K, Ushida K 2004. Dietary sodium gluconate protects rats from large bowel cancer by stimulating butyrate production. Journal of Nutrition 134 :940 944. 125. Kanazawa K, Konishi F, Mitsuoka T, Terada A, Itoh K, Narushima S, Kumemura M, Kimura H 1996. Factors influencing the development of sigmoid colon cancer Bacteriologic and biochemical studies. Cancer 77 :1701 1706. 126. Kilpivaara O, Aaltonen LA 2013. Diagnostic cancer ge nome sequencing and the contribution of germline variants. Science 339 :1559 1562.
174 127. Kim DH, Jin YH 2001. Intestinal bacterial beta glucuronidase activity of patients with colon cancer. Archives of Pharmacal Research 24 :564 567. 128. Kim S, Keku TO, Martin C, Galanko J, Woosley JT, Schroeder JC, Satia JA, Halabi S, Sandler RS 2008. Circulating levels of inflammatory cytokines and risk of colorectal adenomas. Cancer Research 68 :323 328. 129. Klein RS, Recco RA, Catalano M T Edberg SC, Casey JI Ste igbigel N H 1977. Association of Streptococcus bovis with carcinoma of the colon. The New England Jo urnal of Medicine 297 :800 802. 130. Knasmuller S, Steinkellner H, Hirschl AM, Rabot S, Nobis EC, Kassie F 2001. Impact of bacteria in dairy products and of the intestinal microflora on the genotoxic and carcinogenic effects of heterocyclic aromatic amines. Mutation Research Fundamental and Molecular Mechanisms of Mutagenesis 480 :129 138. 131. Knudson AG 2001. Two genetic hits (more or less) to cancer. Nature Reviews Cancer 1 :157 162. 132. Kojima M, Morisaki T, Sasaki N, Nakano K, Mibu R, Tanaka M, Katano M 2004. Increased nuclear factor kappa B activation in human colorectal carcinoma and its correlation with tumor progression. Anticancer Research 24 :675 681. 133. Koushik A, Hunter DJ, Spiegelman D, Beeson WL, van den Brandt PA, Buring JE, Calle EE, Cho E, Fraser GE, Freudenheim JL, Fuchs CS, Giovannucci EL, Goldbohm RA, Harnack L, Jacobs DR, Kato I, Krogh V, Larsson SC, Leitzmann MF, Marshall JR, M cCullough ML, Miller AB, Pietinen P, Rohan TE, Schatzkin A, Sieri S, Virtanen MJ, Wolk A, Zeleniuch Jacquotte A, Zhang SM, Smith Warner SA 2007. Fruits, vegetables, and colon cancer risk in a pooled analysis of 14 cohort studies. Journal of the National C ancer Institute 99 :1471 1483. 134. Langkamp Henken, B. 2012. Assessing immune health outcomes following dietary interventions in healthy adults: a model using galactooligosaccharide. In S. Cho (ed.), Dietary Fiber. CRC Press, Boca Raton. 135. Langkamp Henken B, Bender BS, Gardner EM, Herrlinger Garcia KA, Kelley MJ, Murasko DM, Schaller JP, Stechmiller JK, Thomas DJ, Wood SM 2004. Nutritional formula enhanced immune function and reduced days of symptoms of upper respiratory tract infection in seniors. Journal of the American Geriatrics Society 52 :3 12. 136. Larsen N, Vogensen FK, van den Berg FWJ, Nielsen DS, Andreasen AS, Pedersen BK, Abu Al Soud W, Sorensen SJ, Hansen LH, Jakobsen M 2010. Gut Microbiota in Human Adults with Type 2 Diabetes Differs from Non Diabetic Adults. Plos One 5
175 137. Larsson SC, Giovannucci E, Bergkvist L, Wolk A 2005. Whole grain consumption and risk of colorectal cancer: a population based cohort of 60 000 women. British Journal of Cancer 92 :1803 1807. 138. Larsson SC, Wolk A 2006. Meat consumption and risk of colorectal cancer: A meta analysis of prospective studies. International Journal of Cancer 119 :2657 2664. 139. Lazarova DL, Bordonaro M, Carbone R, Sartorelli AC 2004. Linear relationship between WNT activity levels and apoptosis in colorectal carcinoma cells exposed to butyrate. International Journal of Cancer 110 :523 531. 140. Le Blay G, Michel C, Blottiere HM, Cherbut C 1999. Enhancement of butyrate production in the rat caecocolonic tract by long term in gestion of resistant potato starch. British Journal of Nutrition 82 :419 426. 141. Le Leu RK, Brown IL, Hu Y, Bird AR, Jackson M, Esterman A, Young GP 2005. A synbiotic combination of resistant starch and Bifidobacterium lactis facilitates apoptotic dele tion of carcinogen damaged cells in rat colon. Journal of Nutrition 135 :996 1001. 142. Lee YK, Menezes JS, Umesaki Y, Mazmanian SK 2011. Proinflammatory T cell responses to gut microbiota promote experimental autoimmune encephalomyelitis. Proceedings of the National Academy of Sciences of the United States of America 108 :4615 4622. 143. Lin J, Zhang SM, Cook NR, Rexrode KM, Liu S, Manson JE, Lee I, Buring J E. 2005. Dietary intakes of fruit, vegetables, and fiber, and risk of colorectal cancer in a pros pective cohort of women (United States). Cancer Causes and Control 16:225 233. 144. Liu RH 2004. Potential synergy of phytochemicals in cancer prevention: Mechanism of action. Journal of Nutrition 134 :3479S 3485S. 145. Llor X, Pons E, Roca A, Alvarez M, Mane J, Fernandez Banares F, Gassull MA 2003. The effects of fish oil, olive oil, oleic acid and linoleic acid on colorectal neoplastic processes. Clinical Nutrition 22 :71 79. 146. Louis P, Duncan SH, Mccrae SI, Millar J, Jackson MS, Flint HJ 2004. Restricted distribution of the butyrate kinase pathway among butyrate producing bacteria from the human colon. Journal of Bacteriology 186 :2099 2106. 147. Lozupone C, Hamady M, Knight R 2006. UniFrac An online tool for comparing microbial community d iversity in a phylogenetic context. Bmc Bioinformatics 7
176 148. Macpherson AJ, Gatto D, Sainsbury E, Harriman GR, Hengartner H, Zinkernagel RM 2000. A primitive T cell independent mechanism of intestinal mucosal IgA responses to commensal bacteria. Scien ce 288 :2222 +. 149. Maddocks ODK, Short AJ, Donnenberg MS, Bader S, Harrison DJ 2009. Attaching and Effacing Escherichia coli Downregulate DNA Mismatch Repair Protein In Vitro and Are Associated with Colorectal Adenocarcinomas in Humans. Plos One 4 15 0. Mai V, Colbert LH, Perkins SN, Schatzkin A, Hursting SD 2007. Intestinal microbiota: A potential diet responsive prevention target in Apc(Min) mice. Molecular Carcinogenesis 46 :42 48. 151. Mai V, McCrary QM, Sinha R, Glei M 2009. Associations betwe en dietary habits and body mass index with gut microbiota composition and fecal water genotoxicity: an observational study in African American and Caucasian American volunteers. Nutrition Journal 8 152. Mai V, Young CM, Ukhanova M, Wang XY, Sun YJ, Case lla G, Theriaque D, Li N, Sharma R, Hudak M, Neu J 2011. Fecal Microbiota in Premature Infants Prior to Necrotizing Enterocolitis. Plos One 6 153. Mali P, Chou BK, Yen J, Ye ZH, Zou JZ, Dowey S, Brodsky RA, Ohm JE, Yu WN, Baylin SB, Yusa K, Bradley A, Meyers DJ, Mukherjee C, Cole PA, Cheng LZ 2010. Butyrate Greatly Enhances Derivation of Human Induced Pluripotent Stem Cells by Promoting Epigenetic Remodeling and the Expression of Pluripotency Associated Genes. Stem Cells 28 :713 720. 154. Markowitz SD Bertagnolli MM 2009. Molecular Origins of Cancer: Molecular Basis of Colorectal Cancer. New England Journal of Medicine 361 :2449 2460. 155. Mathew A, Peters U, Chatterjee N, Kulldorff M, Sinha R 2004. Fat, fiber, fruits, vegetables, and risk of color ectal adenomas. International Journal of Cancer 108 :287 292. 156. Mazmanian SK, Liu CH, Tzianabos AO, Kasper DL 2005. An immunomodulatory molecule of symbiotic bacteria directs maturation of the host immune system. Cell 122 :107 118. 157. McGuckin MA, Linden SK, Sutton P, Florin TH 2011. Mucin dynamics and enteric pathogens. Nature Reviews Microbiology 9 :265 278. 158. Mcintyre A, Gibson PR, Young GP 1993. Butyrate Production from Dietary Fiber and Protection Against Large Bowel Cancer in A Rat Model Gut 34 :386 391.
177 159. Merlo A, Herman JG, Mao L, Lee DJ, Gabrielson E, Burger PC, Baylin SB, Sidransky D 1995. 5' Cpg Island Methylation Is Associated with Transcriptional Silencing of the Tumor Suppressor P16/Cdkn2/Mts1 in Human Cancers. Nature Medici ne 1 :686 692. 160. Michels KB, Fuchs CS, Giovannucci E, Colditz GA, Hunter DJ, Stampfer MJ, Willett WC 2005. Fiber intake and incidence of colorectal cancer among 76,947 women and 47,279 men. Cancer Epidemiology Biomarkers & Prevention 14 :842 849. 161. Michels KB, Giovannucci E, Joshipura KJ, Rosner BA, Stampfer MJ, Fuchs CS, Colditz GA, Speizer FE, Willett WC 2000. Prospective study of fruit and vegetable consumption and incidence of colon and rectal cancers. Journal of the National Cancer Institute 92 :1740 1752. 162. Miller DN, Bryant JE, Madsen EL, Ghiorse WC 1999. Evaluation and optimization of DNA extraction and purification procedures for soil and sediment samples. Appl.Environ.Microbiol. 65 :4715 4724. 163. Mirande C, Kadlecikova E, Matulova M, Capek P, Bernalier Donadille A, Forano E Bra Maillet C 2009. Dietary fibre degradation and fermentation by twoxylanolytic bacteria Bacteroides xylanisolvens XB1AT and Roseburia intestinalis XB6B4 from the human intestine. Journal of Applied Microbio logy 109 :451 460. 164. Miyakura Y, Sugano K, Konishi F, Ichikawa A, Maekawa M, Shitoh K, Igarashi S, Kotake K, Koyama Y, Nagai H 2001. Extensive methylation of hMLH1 promoter region predominates in proximal colon cancer with microsatellite instability. Gastroenterology 121 :1300 1309. 165. Moghaddam AA, Woodward M, Huxley R 2007. Obesity and risk of colorectal cancer: A meta analysis of 31 studies with 70,000 events. Cancer Epidemiology Biomarkers & Prevention 16 :2533 2547. 166. Mora JR, Iwata M, Eks teen B, Song SY, Junt T, Senman B, Otipoby KL, Yokota A, Takeuchi H, Ricciardi Castagnoli P, Rajewsky K, Adams DH, von Andrian UH 2006. Generation of gut homing IgA secreting B cells by intestinal dendritic cells. Science 314 :1157 1160. 167. Morin PJ, S parks AB, Korinek V, Barker N, Clevers H, Vogelstein B, Kinzler KW 1997. Activation of beta catenin Tcf signaling in colon cancer by mutations in beta catenin or APC. Science 275 :1787 1790. 168. Nagase H, Miyoshi Y, Horii A, Aoki T, Ogawa M, Utsunomiya J, Baba S, Sasazuki T, Nakamura Y 1992. Correlation Between the Location of Germ Line Mutations in the Apc Gene and the Number of Colorectal Polyps in Familial Adenomatous Polyposis Patients. Cancer Research 52 :4055 4057.
178 169. Nilsen TIL, Vatten LJ 200 1. Prospective study of colorectal cancer risk and physical activity, diabetes, blood glucose and BMI: exploring the hyperinsulinaemia hypothesis. British Journal of Cancer 84 :417 422. 170. Norat T, Bingham S, Ferrari P, Slimani N, Jenab M, Mazuir M, Ove rvad K, Olsen A, Tjonneland A, Clavel F, Boutron Ruault MC, Kesse E, Boeing H, Bergmann MM, Nieters A, Linseisen J, Trichopoulou A, Trichopoulos D, Tountas Y, Berrino F, Palli D, Panico S, Tumino R, Vineis P, Bueno De Mesquita HB, Peeters PHM, Engeset D, L und E, Skeie G, Ardanaz E, Gonzalez C, Navarro C, Quiros JR, Sanchez MJ, Berglund G, Mattisson I, Hallmans G, Palmqvist R, Day NE, Khaw KT, Key TJ, San Joaquin M, Hemon B, Saracci R, Kaaks R, Riboli E 2005. Meat, fish, and colorectal cancer risk: The Euro pean prospective investigation into cancer and nutrition. Journal of the National Cancer Institute 97 :906 916. 171. O'Hara AM, Shanahan F 2006. The gut flora as a forgotten organ. Embo Reports 7 :688 693. 172. Oberreuther Moschner DL, Jahreis G, Rechke mmer G, Pool Zobel BL 2004. Dietary intervention with the problotics Lactobacillus acidophilus 145 and Bifidobacterium longum 913 modulates the potential of human faecal water to induce damage in HT29clone19A cells. British Journal of Nutrition 91 :925 932 173. Ochsenkuhn T, Bayerdorffer E, Meining A, Schinkel M, Thiede C, Nussler V, Sackmann M, Hatz R, Neubauer A, Paumgartner G 1999. Colonic mucosal proliferation is related to serum deoxycholic acid levels. Cancer 85 :1664 1669. 174. Ogino S, Kawasaki T, Kirkner GJ, Kraft P, Loda M, Fuchs CS 2007. Evaluation of markers for CpG island methylator phenotype (CIMP) in colorectal cancer by a large population based sample. Journal of Molecular Diagnostics 9 :305 314. 175. Ogura Y, Bonen DK, Inohara N, Nico lae DL, Chen FF, Ramos R, Britton H, Moran T, Karaliuskas R, Duerr RH, Achkar JP, Brant SR, Bayless TM, Kirschner BS, Hanauer SB, Nunez G, Cho JH 2001. A frameshift mutation in NOD2 associated with susceptibility to Crohn's disease. Nature 411 :603 606. 1 76. Ohkawara S, Furuya H, Nagashima K, Asanuma N, Hino T 2005. Oral administration of Butyrivibrio fibrisolvens, a butyrate producing bacterium, decreases the formation of aberrant crypt foci in the colon and rectum of mice. Journal of Nutrition 135 :2878 2883. 177. Olszak T, An DD, Zeissig S, Vera MP, Richter J, Franke A, Glickman JN, Siebert R, Baron RM, Kasper DL, Blumberg RS 2012. Microbial Exposure During Early Life Has Persistent Effects on Natural Killer T Cell Function. Science 336 :489 493.
179 178 Orchel A, Dzierzewicz Z, Parfiniewicz B, Weglarz L, Wilczok T 2005. Butyrate induced differentiation of colon cancer cells is PKC and JNK dependent. Digestive Diseases and Sciences 50 :490 498. 179. Park Y, Hunter DJ, Spiegelman D, Bergkvist L, Berrin o F, van den Brandt PA, Buring JE, Colditz GA, Freudenheim JL, Fuchs CS, Giovannucci E, Goldbohm RA, Graham S, Harnack L, Hartman AM, Jacobs DR, Kato I, Krogh V, Leitzmann MF, McCullough ML, Miller AB, Pietinen P, Rohan TE, Schatzkin A, Willett WC, Wolk A, Zeleniuch Jacquotte A, Zhang SMM, Smith Warner SA 2005. Dietary fiber intake and risk of colorectal cancer A pooled analysis of prospective cohort studies. Jama Journal of the American Medical Association 294 :2849 2857. 180. Polk DB, Peek RM 2010. H elicobacter pylori: gastric cancer and beyond. Nature Reviews Cancer 10 :403 414. 181. Popivanova BK, Kitamura K, Wu Y, Kondo T, Kagaya T, Kaneko S, Oshima M, Fujii C, Mukaida N 2008. Blocking TNF alpha in mice reduces colorectal carcinogenesis associate d with chronic colitis. Journal of Clinical Investigation 118 :560 570. 182. Potter MA, Cunliffe NA, Smith M, Miles RS, Flapan AD, Dunlop MG 1998. A prospective controlled study of the association of Streptococcus bovis with colorectal carcinoma. Journal of Clinical Pathology 51 :473 474. 183. Rafter J, Bennett M, Caderni G, Clune Y, Hughes R, Karlsson PC, Klinder A, O'Riordan M, O'Sullivan GC, Pool Zobel B, Rechkemmer G, Roller M, Rowland I, Salvadori M, Thijs H, Van Loo J, Watzl B, Collins JK 2007. Di etary synbiotics reduce cancer risk factors in polypectomized and colon cancer patients. American Journal of Clinical Nutrition 85 :488 496. 184. Rahme E, Barkun AN, Toubouti Y, Bardou M 2003. The cyclooxygenase 2 selective inhibitors rofecoxib and celec oxib prevent colorectal neoplasia occurrence and recurrence. Gastroenterology 125 :404 412. 185. Ramanathan V, Jin GC, Westphalen CB, Whelan A, Dubeykovskiy A, Takaishi S, Wang TC 2012. P53 Gene Mutation Increases Progastrin Dependent Colonic Proliferati on and Colon Cancer Formation in Mice. Cancer Investigation 30 :275 286. 186. Ramasamy S, Singh S, Taniere P, Langman MJS, Eggo MC 2006. Sulfide detoxifying enzymes in the human colon are decreased in cancer and upregulated in differentiation. American J ournal of Physiology Gastrointestinal and Liver Physiology 291 :G288 G296.
180 187. Rhee I, Bachman KE, Park BH, Jair KW, Yen RWC, Schuebel KE, Cui HM, Feinberg AP, Lengauer C, Kinzler KW, Baylin SB, Vogelstein B 2002. DNMT1 and DNMT3b cooperate to silence g enes in human cancer cells. Nature 416 :552 556. 188. Rhee I, Bachman KE, Park BH, Jair KW, Yen RWC, Schuebel KE, Cui HM, Feinberg AP, Lengauer C, Kinzler KW, Baylin SB, Vogelstein B 2002. DNMT1 and DNMT3b cooperate to silence genes in human cancer cells Nature 416 :552 556. 189. Riboli E, Norat T 2003. Epidemiologic evidence of the protective effect of fruit and vegetables on cancer risk. American Journal of Clinical Nutrition 78 :559S 569S. 190. Ricciardiello L, Chang DK, Laghi L, Goel A, Chang CL, Boland CR 2001. Mad 1 is the exclusive JC virus strain present in the human colon, and its transcriptional control region has a deleted 98 base Pair sequence in colon cancer tissues. Journal of Virology 75 :1996 2001. 191. Rideout WM, Coetzee GA, Olumi A F, Jones PA 1990. 5 Methylcytosine As An Endogenous Mutagen in the Human Ldl Receptor and P53 Genes. Science 249 :1288 1290. 192. Roberfroid M 2007. Prebiotics: The concept revisited. Journal of Nutrition 137 :830S 837S. 193. Roller M, Femia AP, Cadern i G, Rechkemmer G, Watzl B 2004. Intestinal immunity of rats with colon cancer is modulated by oligofructose enriched inulin combined with Lactobacillus rhamnosus and Bifidobacterium lactis. British Journal of Nutrition 92 :931 938. 194. Round JL, Lee SM Li J, Tran G, Jabri B, Chatila TA, Mazmanian SK 2011. The Toll Like Receptor 2 Pathway Establishes Colonization by a Commensal of the Human Microbiota. Science 332 :974 977. 195. Round JL, Mazmanian SK 2010. Inducible Foxp(3+) regulatory T cell develo pment by a commensal bacterium of the intestinal microbiota. Proceedings of the National Academy of Sciences of the United States of America 107 :12204 12209. 196. Sakamoto K, Maeda S, Hikiba Y, Nakagawa H, Hayakawa Y, Shibata W, Yanai A, Ogura K, Omata M 2009. Constitutive NF kappa B Activation in Colorectal Carcinoma Plays a Key Role in Angiogenesis, Promoting Tumor Growth. Clinical Cancer Research 15 :2248 2258. 197. Salzman NH, Ghosh D, Huttner KM, Paterson Y, Bevins CL 2003. Protection against ente ric salmonellosis in transgenic mice expressing a human intestinal defensin. Nature 422 :522 526.
181 198. Samowitz WS, Curtin K, Ma KN, Edwards S, Schaffer D, Leppert MF, Slattery ML 2002. Prognostic significance of P53 mutations in colon cancer at the popu lation level. International Journal of Cancer 99 :597 602. 199. Samuels Y, Diaz LA, Schmidt Kittler O, Cummins JM, DeLong L, Cheong I, Rago C, Huso DL, Lengauer C, Kinzler KW, Vogelstein B, Velculescu VE 2005. Mutant PIK3CA promotes cell growth and invas ion of human cancer cells. Cancer Cell 7 :561 573. 200. Samuels Y, Wang ZH, Bardelli A, Silliman N, Ptak J, Szabo S, Yan H, Gazdar A, Powell DM, Riggins GJ, Willson JKV, Markowitz S, Kinzler KW, Vogelstein B, Velculescu VE 2004. High frequency of mutatio ns of the PIK3CA gene in human cancers. Science 304 :554. 201. Sandler RS, Halabi S, Baron JA, Budinger S, Paskett E, Keresztes R, Petrelli N, Pipas JM, Karp DD, Loprinzi CL, Steinbach G, Schilsky R 2003. A randomized trial of aspirin to prevent colorect al adenomas in patients with previous colorectal cancer. New England Journal of Medicine 348 :883 890. 202. Sanjoaquin MA, Allen N, Couto E, Roddam AW, Key TJ 2005. Folate intake and colorectal cancer risk: A meta analytical approach. International Journ al of Cancer 113 :825 828. 203. Savage D C 2012. Microbial ecology of the gastrointestinal tract. Annual Review of Microbiology 31 : 107 133. 204. Scanlan PD, Shanahan F, Clune Y, Collins JK, O'Sullivan GC, O'Riordan M, Holmes E, Wang YL, Marchesi JR 2 008. Culture independent analysis of the gut microbiota in colorectal cancer and polyposis. Environmental Microbiology 10 :789 798. 205. Schatzkin A, Mouw T, Park Y, Subar AF, Kipnis V, Hollenbeck A, Leitzmann MF Thompson F E 2007. Dietary fiber and whol e grain consumption in relation to colorectal cancer in the NIH AARP Diet and Health Study. American Jo urnal of Clinical Nutrition 85:1353 1360. 206. Sealy L, Chalkley R. 1978. DNA associated with hyperacetylated histone is preferentially digested by DN a se I. Nucleic Acids Research 5: 1863 1876. 207. Sears CL, Pardoll DM 2011. Perspective: Alpha Bugs, Their Microbial Partners, and the Link to Colon Cancer. Journal of Infectious Diseases 203 :306 311. 208. Sellon RK, Tonkonogy S, Schultz M, Dieleman LA, Grenther W, Balish E, Rennick DM, Sartor RB 1998. Resident enteric bacteria are necessary for development of spontaneous colitis and immune system activation in interleukin 10 deficient mice. Infection and Immunity 66 :5224 5231.
182 209. Sharrard RM, Ro yds JA, Rogers S, Shorthouse AJ 1992. Patterns of Methylation of the C Myc Gene in Human Colorectal Cancer Progression. British Journal of Cancer 65 :667 672. 210. Shen LL, Kondo Y, Rosner GL, Xiao LC, Hernandez NS, Vilaythong J, Houlihan PS, Krouse RS, Prasad AR, Einspahr JG, Buckmeier J, Alberts DS, Hamilton SR, Issa JPJ 2005. MGMT promoter methylation and field defect in sporadic colorectal cancer. Journal of the National Cancer Institute 97 :1330 1338. 211. Shen XJ, Rawls JF, Randall TR, Burcal L, M pande CN, Jenkins N, Jovov B, Abdo Z, Sandler RS, Keku T O. 2010. Molecular characterization of mucosal adherent bacteria and associations with colorectal ade nomas. Gut Microbes 1:138 147. 212. Shin A, Li H, Shu XO, Yang G, Gao YT, Zheng W 2006. Dietary intake of calcium, fiber and other micronutrients in relation to colorectal cancer risk: Results from the Shanghai Women's Health Study. International Journal of Cancer 119 :2938 2942. 213. Shinya H, Wolff WI 1979. Morphology, Anatomic Distribution and Cancer Potential of Colonic Polyps Analysis of 7,000 Polyps Endoscopically Removed. Annals of Surgery 190 :679 683. 214. Singal R, van Wert J, Bashambu M 2001. Cytosine methylation represses glutathione S transferase P1 (GSTP1) gene expression in human prostate cancer cells. Cancer Research 61 :4820 4826. 215. Singh J, Rivenson A, Tomita M, Shimamura S, Ishibashi N, Reddy BS 1997. Bifidobacterium longum, a lactic acid producing intestinal bacterium inhibits colon cancer and modulates the intermediate biomarkers of colon carcinogenesis. Carcinogenesis 18 :833 841. 216. Sinha R, Kulldorff M, Chow WH, Denobile J, Rothman N 2001. Dietary intake of heterocyclic amines, meat derived mutagenic activity, and risk of colorectal adenomas. Cancer Epidemiology B iomarkers & Prevention 10 :559 562. 217. Slattery ML, Curtin K, Wolff RK, Herrick JS, Caan BJ, Samowitz W 2010. Diet, physical activity, and body size associations with rectal tumor mutations and epigenetic changes. Cancer Causes & Control 21 :1237 1245. 218. Slattery ML, Curtin KP, Edwards SL, Schaffer DM 2004. Plant foods, fiber, and rectal cancer. American Journal of Clinical Nutrition 79 :274 281. 219. Smith KJ, Johnson KA, Bryan TM, Hill DE, Markowitz S, Willson JKV, Paraskeva C, Petersen GM, Hami lton SR, Vogelstein B, Kinzler KW 1993. The Apc Gene Product in Normal and Tumor Cells. Proceedings of the National Academy of Sciences of the United States of America 90 :2846 2850.
183 220. Smith RA, von Eschenbach AC, Wender R, Levin B, Byers T, Rothenber ger D, Brooks D, Creasman W, Cohen C, Runowicz C, Saslow D, Cokkinides V, Eyre H 2001. American Cancer Society guidelines for the early detection of cancer: Update of early detection guidelines for prostate, colorectal, and endometrial cancers ALSO: Update 2001 Testing for early lung cancer detection. Ca A Cancer Journal for Clinicians 51 :38 75. 221. Smythies LE, Sellers M, Cl ements RH, Mosteller Barnum M, Meng G, Benjamin WH, Orenstein JM, Smith PD 2005. Human intestinal macrophages display profound inflammatory anergy despite avid phagocytic and bacteriocidal activity. Journal of Clinical Investigation 115 :66 75. 222. Summ ers RM 2010. Polyp Size Measurement at CT Colonography: What Do We Know and What Do We Need to Know? Radiology 255 :707 720. 223. Sun Y, Cai Y, Liu L, Yu F, Farrell ML, McKendree W, Farmerie W 2009. ESPRIT: estimating species richness using large collec tions of 16S rRNA pyrosequences. Nucleic Acids Res 37 :e76. 224. Sun YJ, Cai YP, Mai V, Farmerie W, Yu FH, Li JA, Goodison S 2010. Advanced computational algorithms for microbial community analysis using massive 16S rRNA sequence data. Nucleic Acids Rese arch 38 225. Sutherland L, Singleton J, Sessions J, Hanauer S, Krawitt E, Rankin G, Summers R, Mekhjian H, Greenberger N, Kelly M, Levine J, Thomson A, Alpert E, Prokipchuk E 1991. Double Blind, Placebo Controlled Trial of Metronidazole in Crohns Disea se. Gut 32 :1071 1075. 226. Swidsinski A, Khilkin M, Kerjaschki D, Schreiber S, Ortner M, Weber J, Lochs H 1998. Association between intraepithelial Escherichia coli and colorectal cancer. Gastroenterology 115 :281 286. 227. Swidsinski A, Loening Baucke V, Vaneechoutte M, Doerffel Y 2008. Active Crohn's disease and ulcerative colitis can be specifically diagnosed and monitored based on the biostructure of the fecal flora. Inflammatory Bowel Diseases 14 :147 161. 228. Takayama T, Katsuki S, Takahashi Y, Ohi M, Nojiri S, Sakamaki S, Kato J, Kogawa K, Miyake H, Niitsu Y 1998. Aberrant crypt foci of the colon as precursors of adenoma and cancer. New England Journal of Medicine 339 :1277 1284. 229. Tap J, Mondot S, Levenez F, Pelletier E, Caron C, Furet JP Ugarte E, Munoz Tamayo R, Paslier DLE, Nalin R, Dore J, Leclerc M 2009. Towards the human intestinal microbiota phylogenetic core. Environmental Microbiology 11 :2574 2584.
184 230. Terry P, Giovannucci E, Michels KB, Bergkvist L, Hansen H, Holmberg L, Wol k A 2001. Fruit, vegetables, dietary fiber, and risk of colorectal cancer. Journal of the National Cancer Institute 93 :525 533. 231. Terzic J, Grivennikov S, Karin E, Karin M 2010. Inflammation and Colon Cancer. Gastroenterology 138 :2101 U119. 232. T hun MJ, Calle EE, Namboodiri MM, Flanders WD, Coates RJ, Byers T, Boffetta P, Garfinkel L, Heath CW 1992. Risk Factors for Fatal Colon Cancer in A Large Prospective Study. Journal of the National Cancer Institute 84 :1491 1500. 233. Toprak NU, Yagci A, G ulluoglu BM, Akin ML, Demirkalem P, Celenk T, Soyletir G 2010. A possible role of Bacteroides fragilis enterotoxin in the aetiology of colorectal cancer. Clinical Microbiology and Infection 12 :782 786. 234. Tortola S, Marcuello E, Gonzalez I, Reyes G, A rribas R, Aiza G, Sancho FJ, Peinado MA, Capella G 1999. p53 and K ras gene mutations correlate with tumor aggressiveness but are not of routine prognostic value in colorectal cancer. Journal of Clinical Oncology 17 :1375 1381. 235. Toyota M, Ahuja N, Oh e Toyota M, Herman JG, Baylin SB, Issa JPJ 1999. CpG island methylator phenotype in colorectal cancer. Proceedings of the National Academy of Sciences of the United States of America 96 :8681 8686. 236. Turnbaugh PJ, Gordon JI 2009. The core gut microbi ome, energy balance and obesity. Journal of Physiology London 587 :4153 4158. 237. Turnbaugh PJ, Hamady M, Yatsunenko T, Cantarel BL, Duncan A, Ley RE, Sogin ML, Jones WJ, Roe BA, Affourtit JP, Egholm M, Henrissat B, Heath AC, Knight R, Gordon JI 2009. A core gut microbiome in obese and lean twins. Nature 457 :480 4U7. 238. Tzortzis G, Goulas AK, Gee JM, Gibson GR 2005. A novel galactooligosaccharide mixture increases the bifidobacterial population numbers in a continuous in vitro fermentation system an d in the proximal colonic contents of pigs in vivo. Journal of Nutrition 135 :1726 1731. 239. Vaishnava S, Behrendt CL, Ismail AS, Eckmann L, Hooper LV 2008. Paneth cells directly sense gut commensals and maintain homeostasis at the intestinal host micro bial interface. Proceedings of the National Academy of Sciences of the United States of America 105 :20858 20863. 240. Vaishnava S, Yamamoto M, Severson KM, Ruhn KA, Yu XF, Koren O, Ley R, Wakeland EK, Hooper LV 2011. The Antibacterial Lectin RegIII gamm a Promotes the Spatial Segregation of Microbiota and Host in the Intestine. Science 334 :255 258.
185 241. van Dokkum W, Wezendonk B, Srikumar TS, van den Heuvel EGHM 1999. Effect of nondigestible oligosaccharides on large bowel functions, blood lipid concen trations and glucose absorption in young healthy male subjects. European Journal of Clinical Nutrition 53 :1 7. 242. Van Kruiningen HJ, West AB, Freda BJ, Holmes KA 2002. Distribution of Peyer's patches in the distal ileum. Inflammatory Bowel Diseases 8 : 180 185. 243. Vogelstein B, Papadopoulos N, Velculescu VE, Zhou S, Diaz LA, Kinzler KW 2013. Cancer genome landscapes. Science 339 :1546 1558. 244. Voorrips LE, Goldbohm RA, van Poppel G, Sturmans F, Hermus RJJ, van den Brandt PA 2000. Vegetable and f ruit consumption and risks of colon and rectal cancer in a prospective cohort study The Netherlands Cohort Study on Diet and Cancer. American Journal of Epidemiology 152 :1081 1092. 245. Vulevic J, Drakoularakou A, Yaqoob P, Tzortzis G, Gibson GR 2008. Modulation of the fecal microflora profile and immune function by a novel trans galactooligosaccharide mixture (B GOS) in healthy elderly volunteers. American Journal of Clinical Nutrition 88 :1438 1446. 246. Wactawski Wende J, Kotchen JM, Anderson GL, A ssaf AR, Brunner RL, O'Sullivan MJ, Margolis KL, Ockene JK, Phillips L, Pottern L, Prentice RL, Robbins J, Rohan TE, Sarto GE, Sharma S, Stefanick ML, Van Horn L, Wallace RB, Whitlock E, Bassford T, Beresford SAA, Black HR, Bonds DE, Brzyski RG, Caan B, Ch lebowski RT, Cochrane B, Garland C, Gass M, Hays J, Heiss G, Hendrix SL, Howard BV, Hsia J, Hubbell FA, Jackson RD, Johnson KC, Judd H, Kooperberg CL, Kuller LH, LaCroix AZ, Lane DS, Langer RD, Lasser NL, Lewis CE, Limacher MC, Manson JE 2006. Calcium plu s vitamin D supplementation and the risk of colorectal cancer. New England Journal of Medicine 354 :684 696. 247. Wakai K, Date C, Fukui M, Tamakoshi K, Watanabe Y, Hayakawa N, Kojima M, Kawado M, Suzuki K, Hashimoto S, Tokudorne S, Ozasa K, Suzuki S, Toy oshirna H, Ito Y, Tamakoshi A 2007. Dietary fiber and risk of colorectal cancer in the Japan Collaborative Cohort Study. Cancer Epidemiology Biomarkers & Prevention 16 :668 675. 248. Wang TT, Cai GX, Qiu YP, Fei N, Zhang MH, Pang XY, Jia W, Cai SJ, Zhao LP 2012. Structural segregation of gut microbiota between colorectal cancer patients and healthy volunteers. Isme Journal 6 :320 329. 249. Wen L, Ley RE, Volchkov PY, Stranges PB, Avanesyan L, Stonebraker AC, Hu CY, Wong FS, Szot GL, Bluestone JA, Gordon JI, Chervonsky AV 2008. Innate immunity and intestinal microbiota in the development of Type 1 diabetes. Nature 455 :1109 1U10.
186 250. Wijnands MVW, Schoterman HC, Bruijntjes JP, Hollanders VMH, Woutersen RA 2001. Effect of dietary galacto oligosaccharid es on azoxymethane induced aberrant crypt foci and colorectal cancer in Fischer 344 rats. Carcinogenesis 22 :127 132. 251. Wu GY, Hasenberg T, Magdeburg R, Bonninghoff R, Sturm JW, Keese M 2009. Association Between EGF, TGF beta 1, VEGF Gene Polymorphism and Colorectal Cancer. World Journal of Surgery 33 :124 129. 252. Wu H, Dai Q, Shrubsole MJ, Ness RM, Schlundt D, Smalley WE, Chen H, Li M, Shyr Y, Zheng W 2010. Fruit and vegetable intakes are associated with lower risk of colorectal adenomas. (vol 139 pg 340, 2009). Journal of Nutrition 140 :667. 253. Wu HJ, Ivanov II, Darce J, Hattori K, Shima T, Umesaki Y, Littman DR, Benoist C, Mathis D 2010. Gut Residing Segmented Filamentous Bacteria Drive Autoimmune Arthritis via T Helper 17 Cells. Immunity 32 :815 827. 254. Wu SG, Lim KC, Huang J, Saidi RF, Sears CL 1998. Bacteroides fragilis enterotoxin cleaves the zonula adherens protein, E cadherin. Proceedings of the National Academy of Sciences of the United States of America 95 :14979 14984. 255. Wu S G, Morin PJ, Maouyo D, Sears CL 2003. Bacteroides fragilis enterotoxin induces c myc expression and cellular proliferation. Gastroenterology 124 :392 400. 256. Wu SG, Rhee KJ, Albesiano E, Rabizadeh S, Wu XQ, Yen HR, Huso DL, Brancati FL, Wick E, McAllis ter F, Housseau F, Pardoll DM, Sears CL 2009. A human colonic commensal promotes colon tumorigenesis via activation of T helper type 17 T cell responses. Nature Medicine 15 :1016 1U64. 257. Xavier RJ, Podolsky DK 2007. Unravelling the pathogenesis of in flammatory bowel disease. Nature 448 :427 434. 258. Yan M, Rerko RM, Platzer P, Dawson D, Willis J, Tong M, Lawrence E, Lutterbaugh J, Lu SL, Willson JKV, Luo GB, Hensold J, Tai HH, Wilson K, Markowitz SD 2004. 15 hydroxyprostagland in dehydrogenase, a C OX 2 oncogene antagonist, is a TGF beta induced suppressor of human gastrointestinal cancers. Proceedings of the National Academy of Sciences of the United States of America 101 :17468 17473. 259. Zoetendal EG, Akkermans AD, De Vos WM 1998. Temperature g radient gel electrophoresis analysis of 16S rRNA from human fecal samples reveals stable and host specific communities of active bacteria. Appl.Environ.Microbiol. 64 :3854 3859.
187 BIOGRAPHICAL SKETCH Tyler Culpepper was born in Orlando, FL He moved to Ga inesville, FL in 2004 after graduating from high school. He then attended the University of Florida where he earned dual Bachelor of Science degrees in chemistry and microbiology and cell science. After graduation, he matriculated in graduate schoo l in att empt to earn a Doctor of Philosophy in microbiology and cell science. Department of Microbiolog y and Cell Science (MCS) graduate program in August 2008 under the guidance of Dr. Volker Mai. As a graduate student, he has attended research symposiums such as the Florida Genetics Institute Annual Symposium (2010), the Microbiology and Cell Science Grad uate Student Symposium (2009, 2010, 2011, 2012), the Moffit Cancer Center Scientific Retreat (2010), the Emerging Pathogens Institute Research Day (2011), the America Society of Microbiology General Meeting (2010), and the Experimental Biology General Meet ing (2012). He received the Graduate Student Council Travel Award and the MCS Graduate Student Symposium first place poster presentation award in 2012. In addition, he has served as a mentor for three undergraduate students, Madeline Bost, Fleeta Netter, a nd Adeeb Rohani, and has assisted graduate students Varinder Pannu, Venkateswaran Ganesan, Tyler Caton, and Shu Jui Hsu in thesis work or laboratory rotations. He served as the faculty/student liaison and assisted in organizing the MCS Graduate Student Sym posium (2010 2011). He also assisted as a poster judge for the 2011 Undergraduate Microbiology Research Symposium.
188 The work presented in this document has generated five manuscript s some of which are in preparation. Upon completion of his degree, Tyler pl ans to transition into medical science by first pu rsuing a Doctor of Medicine degree while continuing to conduct research and subsequently completing a residency and fellowship to prepare for a career in academic medicine.