10601 Background: Gut bacteria with hormone-modulating potential may influence breast cancer (BC) risk: Equol-producing bacteria convert dietary isoflavones into equol, a phytoestrogen that binds estrogen receptors, while beta-glucosidating bacteria affect glycosides of dietary phytoestrogens, and beta-glucuronidating bacteria regulate the deconjugation and reabsorption of endogenous estrogens. We aimed to investigate whether gut bacterial species that have the potential to affect phytoestrogens or endogenous estrogens can be detected using 16S rDNA sequencing methods, and to what extent this detection may be influenced by the choice of 16S hypervariable region targeted for sequencing. Methods: DNA extracted from stool samples of women with BC (n = 112) and healthy controls (n = 171) was subjected to PCR amplification and high-throughput sequencing of the V1-V3 and V3-V4 16S rRNA hypervariable regions, commonly used regions in microbiome research. Data was processed using QIIME, and taxonomy assignment was performed with the SILVA database to the species level. Relative abundance and detection of taxa in cases vs controls were evaluated using Wilcoxon and McNemar’s tests in R. Taxa were classified as equol-producing based on Setchell and Clerici (2010) or beta-glucuronidating and beta-glucosidating based on Dabek et al (2008). Results: The relative abundance of equol-producing bacteria was significantly higher in the BC group vs controls in the V1–V3 (p = 0.0007); V3–V4 (p = 0.0089) regions, and when both regions were summed (p = 0.0011). Adlercreutzia equolifaciens was detected 29.6% more often in a presence/absence analysis of the V1-V3 region (V1-V3 presence: 93.0%, V3-V4 presence: 63.4%, p < 0.001), while Finegoldia magna was detected 21.1% more often in the V3-V4 region (V1-V3 presence: 18.0%, V3-V4 presence: 39.1%, p < 0.001). The relative abundance of beta-glucuronidating and beta-glucosidating bacteria was significantly higher in the BC group vs controls in the V1-V3 region (p = 0.0094) and when both regions were summed (p = 0.0254), and approached significance in the V3-V4 region (p = 0.0534) . Presence of Bifidobacterium adolescentis was detected 21.8% more often in the V1-V3 region (V1-V3 presence: 21.8%, V3-V4 presence: 0.00%, p < 0.001), and Roseburia inulinivorans was detected 21.1% more often in the V3-V4 region (V1-V3 presence: 63.0%, V3-V4 presence: 83.1%, p < 0.001). Conclusions: Equol-producing and beta-glucuronidating and beta-glucosidating bacteria appear enriched at the species-level analysis in BC subjects, but detection of specific organisms is sensitive to the 16S region sequenced. Accurate identification of hormone-modulating taxa requires both the V1-V3 and V3-V4 regions, as either alone is insufficient to capture the full range of taxa. Additional analyses with absolute quantification such as qPCR are needed to validate the results.

e15737 Background: Colon adenocarcinoma is biologically heterogeneous and is commonly summarized using binary tumor sidedness. While clinically convenient, this framework may obscure more granular anatomic variation, particularly within the proximal colon. We examined whether key molecular phenotypes vary across specific proximal colonic segments using uniformly annotated phenotype data from The Cancer Genome Atlas (TCGA). Methods: Colon adenocarcinoma cases from the TCGA-COAD cohort with annotated anatomic tumor location were evaluated. Analyses were restricted to proximal colonic segments (cecum, ascending colon, hepatic flexure, transverse colon). To ensure uniform comparability across molecular features, analyses were limited to tumors with complete annotation for all evaluated phenotypes, including hypermutation status, mismatch repair (MMR) protein loss by immunohistochemistry, MLH1 silencing, and CpG island methylator phenotype (CIMP). Differences in age at diagnosis across segments were assessed using one-way analysis of variance (ANOVA). Segment-level differences in molecular phenotypes were assessed using chi-square testing. Age- and stage-adjusted analyses were performed using regression models to evaluate whether observed segment-level differences persisted after adjustment. Results: The final analytic cohort included 68 proximal colon tumors with complete molecular phenotype annotation. Age at diagnosis did not differ significantly across proximal colonic segments (one-way ANOVA, p = 0.15), and pathologic stage distribution was similar across segments (χ² = 5.25, p = 0.81). In unadjusted analyses, segment-level heterogeneity was observed for hypermutation (χ² = 9.34, p = 0.025) and MLH1 silencing (χ² = 8.10, p = 0.044), with prevalence increasing from the cecum to the ascending colon and peaking at the hepatic flexure, followed by lower prevalence in the transverse colon. CIMP-high status (χ² = 5.69, p = 0.13) and MMR protein loss (χ² = 2.47, p = 0.48) did not demonstrate statistically significant segment-level differences. After adjustment for age, the association between segment and hypermutation was attenuated to trend level (p = 0.059). In contrast, after adjustment for pathologic stage, segment-level differences in hypermutation (p = 0.041) and MLH1 silencing (p = 0.039) remained statistically significant. Conclusions: Colon adenocarcinoma exhibits clinically relevant molecular heterogeneity within the proximal colon that is not captured by binary sidedness classifications. Hypermutation status and MLH1 silencing vary significantly across proximal colonic segments independent of tumor stage, with enrichment at the hepatic flexure. These findings support a more granular, segment-based model of colon cancer biology and suggest that anatomic factors unique to specific proximal segments may contribute to observed molecular heterogeneity.

INTRODUCTION:Post-traumatic stress disorder (PTSD) is diagnosed when a person experiences chronic psychological distress following a traumatic event that involves a real or perceived threat to life or bodily harm. Up to one-third of inflammatory bowel disease (IBD) patients report some PTSD symptoms due to their disease (IBD-PTS); however, most data are collected in Caucasian populations. IBD-PTS is associated with poorer outcomes. To our knowledge, this is the first study to evaluate IBD-PTS in Black and Latino/a patients. METHODS:Adult IBD patients of Black and/or Latino/a race were recruited at 2 IBD clinics. Those meeting eligibility requirements completed these assessment tools online: PTSD Checklist for the Diagnostic and Statistical Manual of Mental Disorders (Fifth Edition) (PCL-5), Adverse Childhood Experience (ACE) Questionnaire, Inflammatory Bowel Disease Questionnaire (IBDQ), and Demographic and Clinical Information. White, non-Hispanic IBD patients from a previously published dataset from IBD Partners were used for age- and sex-matched controls (2:1). RESULTS:In total, the study included 180 matched controls and 90 participants, including 57.8% of Black and Latino/a patients who had experienced extreme fear or threats to life/bodily harm related to IBD; 33.3% who scored ≥31 on the PCL-5 (minimum criteria for IBD-PTS) vs 12.8% of matched controls. Black and Latino/a patients also reported more severe IBD-PTS symptoms. Those with higher PCL-5 scores had more adverse childhood experiences and were more likely to have been hospitalized, had surgery, or used the emergency room for IBD. More severe IBD-PTS was associated with poorer HRQoL. CONCLUSIONS:Black and Latino/a IBD patients may have rates of IBD-PTS 2.5 times higher than White peers, with more severe PTSD symptoms. Prior nonmedical adverse life events may be risk factors for the development of IBD-PTS. Additional research is warranted.

574 Background: Sulfidogenic bacteria produce hydrogen sulfide, the excess of which is genotoxic and proinflammatory and is absorbed through the GI tract into the systemic circulation. The role of sulfidogenic bacteria in BC has not been explored. In this study, we aimed to evaluate sulfidogenic bacterial abundance and diversity in BC. We hypothesized that there will be lower alpha and beta diversity in BC groups. We also hypothesized that there will be a higher relative abundance of sulfidogenic bacteria in BC vs. controls. Methods: We used high throughput sequencing of the dsrA gene in stool samples of female participants with BC and controls without BC (n=283). Controls were further delineated into normal mammogram, abnormal mammogram with no evidence of BC (e.g. cysts, etc), and controls without BC but with a first degree relative who has BC. To process the sequences, QIIME2 bioinformatics pipelines were used with taxonomy assigned using the SILVA database. We performed alpha diversity analyses with nonparametric statistics, and beta diversity analyses using the Bray-Curtis and weighted/unweighted UniFrac metrics. Differential abundance of sulfidogenic bacteria at the genus level was calculated in R using the DESeq2 package. In the same samples, we used sequencing of the V3-V4 region of the 16s rDNA gene to compare relative abundance of total sulfidogenic bacteria between BC samples and controls in R using the DESeq2 package. Results: Using the Chao1 metric, we found higher sulfidogenic alpha diversity in BC vs. all controls (p < 0.001), BC vs. controls who have a first degree relative with BC (p = 0.005). as well as within controls with a normal mammogram vs. abnormal mammogram without BC (p = 0.037). We did not find significant differences in beta diversity between BC vs. controls. When analyzed visually using PCoA of unweighted UniFrac distances, there were two clusters within all patient groups, indicative of two enterotypes across all subjects. We did not find a significant difference in the relative abundance of overall sulfidogenic bacteria as identified by the 16S rDNA gene between BC and controls. However, Desulfovibrio was enriched in BC compared to normal mammogram samples with a log fold change of 1.69 (p = 0.013). Conclusions: These results suggest altered sulfidogenic gut bacterial community composition in individuals with BC, particularly in organisms with lower relative abundance. Desulfovibrio appears as a predominating organism in BC. This is supported by prior research, which suggests Desulfovibrio becomes less abundant in participants exposed to agents that inhibit BC cell proliferation. We propose that there are two enterotypes of sulfidogenic gut bacteria in women.

BACKGROUND:Accumulation of hydrophobic bile acids (BAs) is linked with cancer development. However, derivatives of deoxycholic acid (DCA) and lithocholic acid (LCA) produced via bacterial metabolism may mitigate the proinflammatory and cytotoxic effects of hydrophobic BAs. The impact of diet on secondary BA derivative production has not been determined. OBJECTIVES:This study aimed to study the associations between BA-modulating nutrients and the composition of secondary BAs and their derivatives. METHODS:Stool and blood were collected from 138 participants aged 45-75 y that self-identified as Black or non-Hispanic White. BAs were extracted from stool and serum and quantified using LC/ESI-MS/MS. Energy, macronutrients, micronutrients, and specific dietary nutrients were estimated from two 24-h diet recalls. The abundance of genes for microbial BA metabolism was assessed from stool metagenomes. Kendall τ correlation and regression-based modeling were performed to determine associations between BA categories, microbial genes, and select energy-adjusted dietary variables (alcohol, calcium, coffee, fiber, fat, and protein). RESULTS:Participants had a mean age of 60 y and a mean BMI of 31 kg/m2. BA derivatives were present in all participant stools, with lagodeoxycholic acid being the most abundant derivative quantified. Analysis of stool microbial metagenomes revealed the presence of genes for secondary BA derivative production in all participants. Protein is positively associated with the accumulation of secondary BAs. monounsaturated fatty acids (MUFA)s were negatively associated with high abundant derivatives of DCA in regression models. Total fiber and coffee intake were positively correlated with increased conversion of BAs to derivatives. Race and smoking status were significant predictors of associations between dietary variables and BA derivatives. CONCLUSION:Protein, MUFAs, total fiber and coffee are significantly associated with concentrations of secondary BAs and their derivatives. Future work should account for social and structural influences on dietary intake and its relationship with BA-elicited cancer risk.

Objectives: Archaea that produce methane gas (methanogens) are the largest component of the human gut archaeome and M. smithii is believed to be the primary methanogen. There is limited fecal metagenome data examining the association between the relative abundance of archaea (i.e. as a percentage of the total metagenomic sequences) and diet. However, studies that have quantified the archaea and concomitantly assessed dietary intake are largely lacking. We hypothesize that dietary intake of specific macro- and micronutrients are associated with abundance of human gut archaea. Methods: Subjects (n=144) were selected from a pre-existing stool sample library collected prospectively among people undergoing colonoscopy to screen for colon cancer who did not have a gastrointestinal disease. Subjects completed the 2014 Block Food Frequency Questionnaire and responses were assessed with the NutritionQuest2014 FFQ Database v3 to generate calorie and macro/micronutrient intake and normalized per 1000 kcal. Fecal DNA was isolated using a Qiagen kit. General archaea, general methanogen, haloarchaea, and M. smithii abundance were measured via qPCR as gene copy number per ng of total DNA. Results: General archaea and methanogen abundance were not correlated with any of the dietary variables measured, which included carbohydrates, sugar, protein, fat, fiber, alcohol, iron, phosphorus, magnesium, potassium, sodium, zinc, copper, and selenium. M. smithii abundance was positively associated with protein (p = 0.044, ρ = 0.168), magnesium (p = 0.028, ρ = 0.184), potassium (p = 0.023, ρ= 0.191), zinc (p = 0.030, ρ = 0.182), and copper (p = 0.022, ρ = 0.192). Haloarchaea were negatively associated with intakes of total fiber (p < 0.001, ρ= -0.306), as well as soluble (p < 0.001, ρ= -0.297) and insoluble fiber (p < 0.001, ρ = -0.291), magnesium (p = 0.021, ρ = -0.196), and potassium (p = 0.007, ρ = -0.227) intake. When assessed by tertiles of intake, abundance of haloarchaea was significantly lower in the highest tertile of fiber (p = 0.002), magnesium (p = 0.032), and potassium (p = 0.047) intake (Kruskal-Wallis). Conclusions: Within the archaeome, Haloarchaea abundance followed by M. smithii abundance are most related to nutrients intake. Additional studies are needed to examine dietary intake in conjunction with clinical variables. Funding Sources: NIH NCI University of Illinois Chicago NIH T32.

ObjectivesThe primary objective was to determine differences in Social Vulnerability Index (SVI) scores among minorities (African-Americans and Hispanics) with acute pancreatitis (AP) compared with non-Hispanic whites (NHWs) with AP. The secondary objectives were to determine differences in diet, sulfidogenic bacteria gene copy numbers (gcn) and hydrogen sulfide (H2S) levels between the 2 groups.Materials and MethodsPatients with AP were enrolled during hospitalization (n = 54). Patient residential addresses were geocoded, and the Centers for Disease Control and Prevention's SVI scores were appended. Dietary intake and serum H2S levels were determined. Microbial DNAs were isolated from stool, and gcn of sulfidogenic bacteria were determined.ResultsMinorities had higher SVI scores compared with NHWs (P = 0.006). They also had lower consumption of beneficial nutrients such as omega-3 fatty acids [stearidonic (P = 0.019), and eicosapentaenoic acid (P = 0.042)], vitamin D (P = 0.025), and protein from seafood (P = 0.031). Lastly, minorities had higher pan-dissimilatory sulfite reductase A (pan-dsrA) gcn (P = 0.033) but no significant differences in H2S levels (P = 0.226).ConclusionMinorities with AP have higher SVI compared with NHWs with AP. Higher SVI scores, lower consumption of beneficial nutrients, and increased gcn of pan-dsrA in minorities with AP suggest that neighborhood vulnerability could be contributing to AP inequities.