Daniel W. Rosenberg - US grants

Based on technique developed by the Principal Investigator that allows the muscularis mucosae (the innermost muscle layer of the intestine) from different anatomical locations to be studied in a uniform manner it has been shown that this muscle is not homogeneous. It exhibits clearly defined differences in its innervation and pharmacologic behavior along the length of the gut and in corresponding intestinal regions in different species. The proposed project utilizes three species, the North American opossum, the rabbit and the rat, to investigate potential mechanisms that may explain these observation. Specific aim 1 is to characterize the nature of the neurotransmitters released by the nerves of the submucosal plexus innervating the muscularis mucosae in different intestinal regions. This is to test the hypothesis that disparate responses of different muscularis mucosae preparations are a function of their respective intrinsic innervation. This will be studied by the established in vitro technique of electrical field stimulation, in association with the use of selective pharmacologic agents that mimic or block the observed responses via receptor occupation or inhibition of transmitter synthesis or release. Specific aim 2 is to comprehensively classify the receptor subtypes that mediate muscularis mucosae responses to the endogenous substances identified in specific aim 1. This is designed to test the hypothesis that differences in the pharmacologic behavior of the muscularis mucosae from different anatomical locations or species are a function of the receptor subtypes that the muscle possesses. This will be determined through the use of selective agonists and antagonists and the derivation of values such as the pA2 and the dissociation constant. Specific aim 3 is to determine which second messengers are utilized by the muscle cells of the muscularis mucosae once endogenous compounds have interacted with the receptors identified in specific aim 2. This is to test the hypothesis that stimulation of the same receptor subtype in different preparations elicits contrasting effects because the receptor is not coupled to the same intracellular second messenger system. This will be achieved by quantitating agonist-evoked responses of the muscularis mucosae in terms of their respective phasic and tonic components and reevaluating these parameters following exposure to agents that selectively block different intracellular signal transduction pathways. These studies will, for the first time, allow region- and species-dependent differences in the characteristics of the muscularis mucosae to be quantitatively explained.

Aberrant crypt foci (ACF) is a precursor for colon cancer. Human and mouse ACR are either hyperplastic are either hyperplastic or dysplastic, a classification that is based on morphology and their potential to form tumors. Understanding mechanisms that govern formulation of ACFs and their conversion to fully malignant colonic lesions is the focus of this proposal. It is known that heritable characteristics of inbred mice lead to either susceptibility or resistance to formulation of colon tumors after carcinogen treatment. We will test two hypotheses that attempt to define genetic mechanisms that determine differential susceptibility. Specific aim 1: Does malignant potential of ACFs, formed in mice of differing colon tumor susceptibilities, result from the specific complement of gene mutations that target cells acquire? Using the colon carcinogen, azoxymethane, we will generate populations of ACF in mice. We have shown that hyperplastic ACF are produced in resistant strains, but fail to progress in carcinomas. We will conduct a detailed morphometric and genetic analysis of sub-populations of ACFs that will enable us to understand why azoxymethane-induced lesions fail to progress to tumors in resistant AKR mice. Laser capture micro-dissection will be used to isolated DNA from single crypts for mutational analyses of tumor related genes. 1.1 What is the time course of zoxymethane- induced ACF formation in tumor susceptible and resistant mice? 1.2 What functional characteristics distinguish hyperplastic and dysplastic ACF? 1.3 Are there differences between hyperplastic and dysplastic ACFs in the frequency of mutations in K-ras, A[C, beta-catenin and p53? Specific aim 2: Is malignant potential of ACFs controlled by expression of, and signalling by, the secretory phospholipase A2, encoded by the Moml locus? Among genetic factors that affect tumorigenesis factors that affect tumorigenesis in the multiple intestinal neoplasia (Min) model is the gene product of Moml, a calcium-dependent non-pancreatic secretory phospholipase (Pla2g2a [Pla2]). sPla2 plays a role in inflammation and hydrolyzes the sn-2 position of glycerolipids, releasing arachidonic acid for prostaglandin synthesis. We present evidence that sPla2 is differentially expressed in mouse clone in a patter inversely correlated with tumor susceptibility to azoxymethane: A/J < SWR < AKR. This suggests that colonic sPla2 also plays a role in chemically-induced tumorigenesis equivalent to its role in the Min model. We propose to explore the mechanism by which sPla2 affects tumor formation in sensitive and resistant mice. 2.1 What colon cell types produce sPla2/ Are there changes in sPla2 levels or function within and adjacent to carcinogen-induced ACF and tumors? 2.2 Can AKR resistances to carcinogen by reversed with the use of specific inhibitors of sPla2? 2.3 Can over- expression of sPla2 in transgenic A/J mice protect against azoxymethane- induced tumorigenesis?

A long term goal of our research program is to identify mechanisms that underlie genetic susceptibility to colon tumorigenesis induced by chemical carcinogens. In previous studies, we have examined the colon carcinogen, azoxymethane (AOM), which induces tumors in inbred mice and rats. In the following studies, we propose to evaluate the carcinogenic potential of an important drinking water disinfection bi-product (DBP), bromodichloromethane (BDCM). Using aberrant crypt foci (ACF) as an early biomarker, we will evaluate the ability of BDCM to induce malignant transformation within the intestinal epithelium of Fischer 344/N rats and C3B6F1 mice at early, intermediate and late time points after exposure through the drinking water. In order to assess malignant potential of colonic lesions, tissue sections containing ACF will be subjected to a complete morphometric analysis. To gain a further understanding of the mechanisms of tumorigenesis, tissue sections will be further subjected to immunohistochemical analyses using a panel of tumor biomarkers that are routinely used in our laboratory, including the following markers of cell proliferaton: PCNA, K-ras, cyclin D1 and CDK 4. In addition, we will evaluate the status of the following genes are associated with colon tumor progression, including p53, APC and beta-catenin. We will also measure cellular differentiation of colonoctyes using both dolichos biflorus agglutinin and hexosaminidase. It will also be important to assess the effects of BDCM on the rate of cell death within the colonic epithelium. We will therefore assay apoptotic response of carcinogen-exposed colon epithelium using commercially available reagents, including the TUNEL assay. Immunochemical analyses will be complemented by mutational analysis of candidate tumor-related genes, using PCR-SSCP analysis and direct DNA sequencing of cDNA prepared from freshly isolated tissue as well as samples procured by laser capture microscopy, techniques that are routinely used in our laboratory. To learn more about the potential role of these genes in the malignant conversion of BDCM-induced ACF, we will examine their mRNA expression using RT-PCR analysis. Depending on the outcome of our studies of BDCM, further analysis of related DBPs, including dibromacetic (DBA) acid, may also be undertaken. It is anticipated that our comparative morphometric and genetic analyses of target epithelial cell populations in rats and mice will provide important new insights into potential mechanisms by which DBPs induce precancerous lesions and tumors within the large intestine.

DESCRIPTION (provided by applicant): The primary thematic idea that the local and systemic oxidant balance is a key determinant for the initiation, progression and metastatic conversion of colon and breast cancers. Potential to modulate local and systemic oxidant balance with diet will be investigated to determine the genetic and functional consequences for cancer progression. Selective cancer models which appropriately define in situ transformation, proliferation, invasion and metastasis will be used in conjunction with genetic and functional analyses to clearly define these events in cancer progression. Part of the program will be directed towards defming diet-induced changes in cancer gene expression that correspond to particular steps in the cancer progression pathway. Specific gene regulating events may be delineated as positive or negative effecters of tumorigenesis, invasion or metastasis. Reproducible gene expression changes observed in tumor cells or the host and can be correlated to pro- or anti-oxidant status, will then be further developed as independent biomarkers of oxidant balance-dependent tumorigenic events. Biomarkers of oxidant status will be determined to establish baseline oxidant balance at the site of tumor initiation and progression as well as in the circulation over the entire time course of each tumor progression model. Dietary interventions will be employed as pro- and anti-oxidant treatments to determine their contribution or alteration to both the level of oxidant balance observed and the corresponding tumor progression events. The overall programatic approach will necessarily define the particular positive and negative events for which dietary constituents, especially those that affect local or overall oxidant balance, will have on select tumorigenic events. Application of oxidant biomarker screening methods defined in these models will then be applied to human tumors (i.e. local oxidant balance), as well as circulating serum levels will be correlated to tumor status and diagnosis. Thus, a long-term goal of this collaboration will be to define baseline oxidant status for human patients depending on their tumor type, current treatment, diet and lifestyle factors. In addition, continued oxidant balance monitoring through the course of treatments, such as chemo and radiotherapies, new therapeutics and changes in diet in lifestyle which may be employed are additive long term goals of the program.

pharmacogenetics; gene expression; drug hypersensitivity; paclitaxel; antineoplastics; gene expression profiling; patient oriented research; clinical research; tissue /cell culture; human subject;

DESCRIPTION (provided by applicant): Colon cancer develops from a combination of genetic and environmental influences. Aberrant crypt loci (ACF) represent a morphologically identifiable premalignant lesion that is associated with colon cancer pathogenesis. We and others have identified a number of alterations in tumor-related genes that may predict their outcome. We intend to fully characterize ACF in terms of their significance in the dysplasia-adenoma-carcinoma sequence. To accomplish this goal, we will examine potential genetic aberrations and expression profiles of ACF at their earliest stages of formation. We will also test how luminal factors induce their growth, focusing on bile acid (BA) composition and bacterial colonization. Finally, we will extend our recent studies on dysregulation of the ARF-p53 circuit, using pdmary epithelial tumor cells (SP-2 cells) recently established from NJ mice. Aim 1 will determine whether genomic instability contributes to colon tumor pathogenesis in AOM-treated mice. We will use CGH and SKY to analyze colon tumors from a panel AOM sensitive strains (A,SWR, FVB/N,Balb/c). We will then use focused BAC arrays and global mutational analyses to study ACF that have been stratified by histomorphology. We will confirm altered genes by TaqMan/IHC analysis and representative known genes involved in colon tumorigenesis will be selected for analysis from regions of gain/loss and mutational alterations. Parallel gene profiling of ACF will be undertaken to determine molecular signatures that predict cancer risk in ACF stratified by tumorigenic potential. RNA from laser-captured ACF of varying morphology and predicted growth characteristics will be linear amplified and applied to high-density arrays. Finally, we will use focused BAC arrays to characterize human ACF using candidate regions identified above. Aim 2 will establish the luminal conditions that favor ACF progression, focusing on chemical exposures, bile acids and bacterial colonization. We will test whether bile acids induce altered luminal redox status and inactivate p53 and whether ACF that form reflect these early alterations. We will also test the effects of bacterial colonization on bile acid metabolism and ACF promotion. Aim 3 will determine the mechanism for functional inactivation of p53 that we recently reported in NJ tumors. This is the first in vivo model demonstrating loss of p53 function (DNA binding, target gene regulation) with sequence-normal cDNA. We will use our epithelial tumor cells (SP-2 cells) to study mechanisms and consequences of p53 inactivation. We will examine the responsiveness of p53 in AOM-induced colon cancer cells to p53 activators and determine the posttranslational modification status of p53. We will determine which proteins associate with p53 in the SP-2 cells and determine the influence of identified p53 modifications on p53 activity. Finally, we will evaluate the effects of BAs on p53 function both YAMC and SP-2 cells.

[unreadable] DESCRIPTION (provided by applicant): Arachidonic acid (AA) metabolism and downstream eicosanoid signaling pathways are well-established targets for colon cancer chemopreventive agents. AA is generated in tumor cells by the actions of phospholipases. One phospholipase of particular interest is cytoplasmic phospholipase A2 (cPLA2). This enzyme contributes to the production of cancer-promoting prostaglandins (PGs) via the Cox-2-catalyzed conversion of AA. Additionally, cPLA2 regulation of intracellular AA levels affects apoptotic signaling, as AA not consumed by Cox-2 controls sphingomyelin conversion to ceramide, a key death effector. Using BALB/c mice that are deficient in cPLA2, we obtained evidence that the growth inhibitory role of cPLA2 in colon via apoptosis regulation may predominate over its role as a tumor promoter through prostaglandin synthesis. We propose a series of in vivo studies in mice and in vitro studies in colon cancer cells to evaluate the role of cPLA2 in the pathogenesis of colorectal cancer (CRC). Aim 1 will test the hypothesis that cPLA2 suppresses colon carcinogenesis through its actions on AA metabolism and the sphingomyelinase (Smase) - ceramide apoptotic pathway. We will evaluate the cPLA2 effect in AOM- and ApcMin-initiated tumors. These studies will provide a comprehensive view of how cPLA2 status impacts intestinal tumorigenesis. Aim 2 will explore mechanisms by which cPLA2 controls tumor cell growth. We will harvest tumor cells from cPLA2-null mice and test their growth in vitro and in an orthotopic model using syngeneic mice. Mechanistic studies will be performed using a conditional Tet-off cPLA2 expression system to evaluate a panel of apoptotic effectors. Aim 3 will test the influence of cPLA2 on chemoprevention by the Cox inhibitor, sulindac. Efficacy will be evaluated in mice with normal or reduced levels of cPLA2. Since Cox inhibitors are only effective on a subset of lesions, we will use a novel chromendoscopic imaging approach to predict and track the effects of chemoprevention on the biological fate of individual AOM-induced adenomas with different cPLA2 genetic backgrounds. Genetic signatures of responsive and non-responsive adenomas will be developed that may be used to target chemoprevention strategies to responsive individuals at the greatest risk for CRC. Predicting drug response through molecular profiling of adenomas will become an important consideration in cancer prevention. Our goal would be to improve targeting of individuals most likely to respond to chemopreventive agents that impact the AA metabolic pathway. [unreadable] [unreadable] [unreadable]

DESCRIPTION (provided by applicant): Colorectal cancer (CRC) is the second leading cause of cancer deaths in the U.S. Screening high-risk individuals for the early detection of colon lesions is an important approach to improving treatment and survival. This application is intended to develop and refine a novel combined endoscopic and proteomic method for evaluating precancerous aberrant crypt foci (ACF) and adenomatous polyps. This methodology aims to establish molecular features that predict, and potentially confer, the efficacy of specific chemoprevention agents. We will use sulindac, currently being tested in an NCI multi-center pilot chemoprevention clinical trial (CPN), as a model chemopreventive agent to develop this methodology. Our proposed studies will address the following issues: (1) The efficacy of sulindac against ACF and adenomas will be tracked in situ using a novel endoscopic imaging/lesion mapping protocol in mice;(2) The proteomic and genomic features of adenomas that predict and/or confer their response to sulindac will be identified;(3) The predictive value of these molecular targets for sulindac efficacy will be evaluated in human colon tissues in a six-week sulindac pilot study. Our hypothesis is that the molecular features of early precancerous lesions and adenomas will predict and potentially confer the efficacy of chemoprevention. Ultimately we envision generating a molecular profile of colon lesions to serve as the basis for assessing risk, designing cancer- prevention strategies customized to the individual and identifying targets for the development of future chemopreventive agents. Although our proposed studies focus on the response to sulindac, this general strategy could be adapted to chemoprevention agents that function through different modes of action. A more comprehensive understanding of how the molecular profile of an individual's colon lesions relates to their response to specific chemopreventive agents could ultimately be used to develop safe and effective strategies that fully realize the promise of chemoprevention for reducing mortality and morbidity related to colon cancer. PUBLIC HEALTH RELEVANCE: We are developing an approach to view the events of colon carcinogenesis and chemoprevention in 'real-time'. It will be possible in principal to determine which subpopulations of early colon lesions develop into tumors, and whether chemoprevention agents suppress the rate of ACF formation, or promote their regression. Our approach is predicted to recapitulate potential clinical situations, in which protein markers can be used to identify individuals with 'high-risk'ACF or adenomas. In addition, a long-term goal of our approach is to customize chemoprevention in human populations based on expression of predictive proteins or genes uncovered in precancerous lesions. Although our proposed studies focus on the response to sulindac, this general strategy could be adapted to chemoprevention agents that function through different modes of action. A more comprehensive understanding of how the molecular profile of an individual's colon lesions relates to their response to specific chemopreventive agents could ultimately be used to develop safe and effective strategies that fully realize the promise of chemoprevention for reducing mortality and morbidity related to colon cancer.

DESCRIPTION (provided by applicant): We propose a multi-disciplinary approach combining molecular analyses with funded clinical studies to evaluate aberrant crypt foci (ACF) and other diminutive lesions (<5 mm) for their significance in cancer development in the distal and proximal colon. We propose that distal colon ACF are self-limiting lesions and that few if any of these mainly hyperplastic lesions progress to neoplasia. Although ACF possess oncogenic mutations (e.g. BRAF or KRAS), we hypothesize that these lesions mobilize signaling pathways comparable to oncogene-induced senescence (OIS) pathways described in other cell and tissue types. We will approach this problem by using confocal imaging techniques and sensitive nanofluidic proteomics to examine the relationship between oncogene activation and OIS within ACF (Specific Aim 1). In contrast to the distal colon, we propose in Specific Aim 2 that ACF in the proximal colon (particularly those with activated BRAF) are at higher risk for progression. We will determine whether BRAF-activated proximal ACF show less efficient OIS, possibly through aberrant DNA methylation, and more frequent microsatellite instability, relative to distal colon lesions. In addition to assessing regional differences in colon cancer progression, we will also evaluate regional differences in chemoprevention using a mouse model of proximal colon cancer driven by KRAS activation (Specific Aim 3). We will focus initially on atorvastatin, a chemoprevention agent that has shown promise but has recently become a concern for potentially increasing proximal colon cancer risk. We will also determine the impact of statin usage on molecular markers expressed in human ACF (obtained from a completed trial). In addition, we will develop and evaluate a novel BRAF activated mouse model that may be particularly useful for proximal cancer chemoprevention studies. Finally, in Specific Aim 4, we will test the hypothesis that the frequency and/or molecular features of ACF will provide an index marker for future or synchronous colon cancer risk. Successful completion of these studies is anticipated to provide important insight into the cancer risk associated with the presence of ACF and other diminutive lesions in the proximal and distal colon. The Translational Significance of our studies is that the paradigm for screening on the right side of the colon may change from identifying large lesions for removal to the inclusion of ACF and other diminutive lesions as biomarkers for cancer risk, or even as potential cancer precursors. This information could potentially shape clinical practice, with additional attention to (or removal of) ACF and other small lesions in the proximal colon and depending on the molecular features identified, a recommendation of more frequent surveillance colonoscopies.

PROJECT SUMMARY The levels of dietary methyl donors vary widely in the population and impact numerous cancer-related pathways. However, the role of folate and other methyl donors in the etiology of colon cancer is complex and controversial. We recently reported that dietary methyl donor depletion (MDD) affords dramatic protection against intestinal tumor formation in several Apc mouse cancer models, even after mice are returned to a normal methyl donor replete diet. Our current data suggest that epigenetic reprogramming contributes to this dramatic and durable cancer protection. Among the metabolic changes we observed were dramatic increases in the levels of the DNMT inhibitor, SAH, reducing DNA methylation among a panel of cancer-related genes, including Wnt targets. In response to PQ11, we will evaluate the possibility that a methyl donor restricted diet, when combined with standard FOLFOX chemotherapy, will enhance tumor cell killing and enable a durable and lasting protection to the colon, in part by preventing epigenetic-driven drug resistance. For our pre-clinical model, we will use the conditional Lrig1xApc mouse, in which distal colon tumors are induced by tamoxifen. The distribution of the tumors in this model to the distal colon will enable us to track their response to treatment in real-time using our mouse colonoscopy system. Overall, the proposed studies will help to elucidate the interplay between dietary methyl donors and colon cancer therapy. The extent to which cancer cell killing is enhanced by methyl donor restriction following FOLFOX administration will be determined. In addition, the possibility that methyl donor restriction affects epigenetic changes and gene expression patterns for long-lasting tumor suppression will be explored. We hypothesize that MDD causes gene-specific changes to DNA methylation and that sustained tumor protection results in part from lasting changes to the epigenome of cancer stem cells. This exploratory proposal could establish a novel paradigm for a nutritional intervention during colon cancer therapy to improve response rates and prevent the development of treatment-resistant cancers.

Project Summary Activation of the COX-2/mPGES-1/PGE2 signaling axis is a hallmark of many cancers, including colorectal cancer (CRC), prompting the implementation of prevention strategies targeting COX-2 activity1-3. Despite their demonstrated chemopreventive efficacy, long-term treatment with COX inhibitors poses significant health risks associated with global suppression of physiological prostanoids, including GI and cardiovascular toxicities. We have shown that targeting the downstream terminal PGE2 synthase, mPGES-1, specifically reduces inducible PGE2 formation without disrupting synthesis of other essential prostanoids4. We have also shown that abrogation of inducible PGE2 formation confers dramatic protection against colon carcinogenesis in both azoxymethane (AOM)-treated Strain A mice and Apc?14/+ mice, with or without associated DSS-induced inflammation4-6. Thus our data and the work of others point to mPGES-1 as a drugable target with potentially high chemopreventive benefit. In order to accelerate its development as a viable drug target, and to better understand the mechanisms underlying this protection, further investigation of mPGES-1 is required. In this exploratory project, we propose to examine how mPGES-1 is incorporated into intracellular signaling pathways involved in the maintenance of the mucosal barrier, and the growth and progression of cancerous lesions. Towards this end, we have recently created an mPGES-1 conditional knockout mouse (cKO) in which mPGES-1 can be inactivated in a cell type-specific manner. In the proposed studies, we will test the cell type-dependent contribution of inducible PGE2 to colon carcinogenesis in the AOM-DSS colis-associated cancer model. Our goal is to determine the relative contribution of PGE2 synthesis generated from its primary cellular sources (epithelial vs. myeloid) to colon cancer development. In Aim 1, we will inactivate mPGES-1 in epithelial cells by crossing cKO with Car1-Cre mice (cKO:Car1). Following AOM- DSS treatment, we will examine the impact of inducible epithelial-derived PGE2 on mucosal injury, repair and multi-stage CRC development. In Aim 2, we will conditionally inactivate mGES-1 in the myeloid-derived lineage, using Lyz2-Cre (cKO:Lyz2) mice and determine the impact of AOM/DSS on mucosal injury and cancer development. Endpoints analyzed in each genetic model will include neoplastic lesion formation and growth, activation status of the COX-2/mPGES-1/PGE2 signaling axis, Wnt signaling and the mutation status of Apc and CTNNB1. We will also study epithelial-stromal cross-talk with a focus on the recruitment and activation of the potent immune-suppressing Tregs and MDSC cell populations. Results from these studies will define the critical sites of inducible PGE2 formation and its relative contribution to colon carcinogenesis. In addition to providing fundamental insight into the process of CRC development, these studies will provide information for identifying patient populations that are most likely to benefit from mPGES-1 inhibition and at what stages of cancer development these inhibitors would be most effective.

PROJECT SUMMARY The rate of early onset colorectal cancer (EOCRC; <50 years of age at diagnosis) continues to increase, even as CRC rates for individuals over 50 have been declining, largely as a result of prevention by enhanced colonoscopic screening. Cancers diagnosed in younger patients tend to be more distal/rectal, mucinous, poorly differentiated and diagnosed at an advanced stage, suggesting a rapid disease progression. Although sporadic EOCRCs are less likely to show aneuploidy, BRAF mutations or CIMP, they are otherwise similar at a molecular level to cancers in individuals greater than 50 years of age. Given the early formation and rapid progression of EOCRC, it is likely that strong promotional factors are at play. Fibroblasts are a key cell type in establishing a microenvironment conducive to cancer progression as they coordinate the activities of epithelial, endothelial and immune cells in the tissue. Fibroblasts can exist in a number of distinct states with dramatically different activities. Resident fibroblast in healthy tissue are non-dividing cells that help establish tissue architecture and crypt cell dynamics. However, when adjacent to cancerous cells, fibroblasts can become persistently activated as cancer-associated fibroblasts (CAFs) that promote tumor growth and angiogenesis, while suppressing immune responses. CAFs can also become senescent and acquire an irreversible senescence associated secretory phenotype (SASP) that establishes a ?permanent? cancer promoting microenvironment. We hypothesize that the underlying stroma advances rapidly in EOCRC to drive early disease pathogenesis. Specifically, we propose that environmental and/or life-style factors cause aberrant fibroblast activation that negatively impacts the normal function of immunoregulatory cells within the stroma, while promoting epithelial cell division. The exploratory experiments proposed here will assess fibroblast proliferation, activation and senescence at different stages of cancer development in young patients. Understanding fibroblast dysregulation in individuals at risk of EOCRC could provide important information for understanding the factors responsible for the increasing incidence of EOCRC and ultimately point to approaches that reduce this risk. We will study fibroblast populations in colonic lesions from patients under 45 and over 60 years old. Using a combination of laser-capture microdissection combined with targeted RNA expression analysis and Imaging Mass CytometryTM, we will define the distinguishing set of molecular alterations characterize define EOCRC cases. Our study population will include normal mucosa, preneoplastic tissue with activated fibroblasts, advanced adenomas and CRCs. Overall, these studies will determine how a hyper-responsive ?hot? stromal microenvironment established by activated and/or senescent fibroblasts relates to other stromal events that contribute to the rapid advancement of EOCRC. Once details of the activated fibroblast populations present in these tissues is determined, long term studies would aim to determine how personal factors relate to their appearance and how their activity might be mitigated to suppress CRC risk.

Project Summary The human diet can positively or negatively impact cancer incidence, with plant-derived compounds ? such as polyphenols ? often exhibiting antioxidant cancer-preventive properties. Walnuts are an exceptional source of polyphenolic ellagitannins (ETs) that are converted to ellagic acid and various urolithins by gut microbiota in the colon. Urolithin A (UroA) is of particular interest for its potent anti-cancer, anti-inflammatory, and prebiotic activities. However, UroA production in individuals can vary significantly, likely based on differences in gut microbiota. We will substantiate the anti-cancer benefits of a prebiotic/probiotic complex derived from consuming walnuts and determine the basis of human inter-individual variability in UroA formation. Our overall hypothesis is that walnut supplementation improves colonic health and lowers colorectal cancer (CRC) risk through UroA formation. This leads to several working hypotheses guiding our Aims: Working Hypothesis 1. UroA producers are at a lower risk of having an advanced colonic lesion; Working Hypothesis 2. Walnut supplementation will increase urinary UroA levels; and Working Hypothesis 3. CRC prevention by walnuts will be greater in UroA-producers than in non-producers. In Aim 1, we propose a randomized, controlled crossover trial in 69 patients (45-75 y) to examine walnut effects on CRC risk factors. We will associate an individual's ability to produce UroA with biomarkers of inflammation and CRC risk, and identify the bacterial species responsible for urolithin metabolism. In Aim 2, we will investigate prebiotic effects of ET-containing walnuts in two conditional mouse CRC models, focusing on important processes in CRC and inflammation, including bile acid metabolism, inflammation, and short chain fatty acid production. In Aim 3, we will test the probiotic effects of human UroA-producing microbiota in a mouse fecal microbiota transplant (MT) study and demonstrate a causal role for specific microbes in UroA formation. This will enable us to validate the concept that important protective effects of walnuts and other ET-rich foods occur through specific microbiota-derived metabolites. This will also define biomarkers and probiotics that highlight the benefits of these foods. Our approach incorporates personalized nutrition with a focus on UroA producers and non-producers in colonic health. Ultimately, our human and pre-clinical mouse studies may lead to prebiotics and probiotics that increase protective urolithins for CRC prevention. These highly significant studies will test the ability of the microbiota to generate colonic mucosa-protective agents (e.g., UroA). It is possible that high-risk patients can be efficiently converted to a protective state by taking probiotics to realize the full benefits of ET-rich foods.