Aflatoxin B1 Disrupts Gut-Microbial Metabolisms of Short-Chain Fatty Acids, Long-Chain Fatty Acids, and Bile Acids in Male F344 Rats

Aflatoxin B1 Disrupts Gut-Microbial Metabolisms of Short-Chain Fatty Acids, Long-Chain Fatty... Abstract In this study, male F344 rats were orally exposed to aflatoxin B1 (AFB1) at 0, 5, 25, and 75 μg/kg for 4 weeks. Rat feces were collected from 2 to 4 weeks following exposure and were assessed for gut-microbiota-dependent metabolites. Gut-microbiota-related organic acids were quantitated in the feces using 2-nitrophenylhydrazine derivatization coupled HPLC-profiling method which was validated and showed good reliability, accuracy and sensitivity. After 2-week exposure, AFB1 significantly reduced the levels of fecal short-chain fatty acids (SCFAs) with an over 70% reduction in the high-dose group (75 μg/kg). Mixed-effects model revealed an inverse correlation between AFB1 dose and fecal levels of SCFAs, but no significant time effect was found. When compared with the control, oral exposure to middle-dose AFB1 (25 μg/kg) resulted in remarkable elevations of fecal cholic acid (2.18-fold), linoleic acid (cis-9, cis-12–18:2) (11.3-fold), pentadecanoic acid (15: 0) (3.68-fold), pyruvic acid (4.56-fold), and 3-phenyllactic acid (3.74-fold), but deoxycholic acid level was reduced by 41% in the low-dose group (5 μg/kg). These results demonstrated the disruptions of several important gut-microbiota metabolic pathways, including the synthesis of SCFAs, pyruvic acid related pathways, metabolisms of amino acids, bile acids and long-chain fatty acids, which may further affect host digestive efficiency, energy supply, intestinal immunity, production of neurotransmitters, and enterohepatic cross-talk. Our study suggests that the impairment of gut-microbiota-dependent metabolism may contribute to pathological mechanisms of AFB1-induced adverse health effects. Aflatoxin B1, gut-microbiota, long-chain fatty acids, microbial metabolism, short-chain fatty acids Aflatoxins (AFs) are a class of food-borne mycotoxins mainly produced by Aspergillus flavus and Aspergillus parasiticus (Kumar et al., 2016). These toxigenic fungi commonly contaminate soil, and colonize on the surface of cereals, especially for maize, and groundnuts, once humidity (>17.5%) and temperature (>24°C) meet their growth needs (Trenk and Hartman, 1970). Such environmental conditions have made the tropical area more susceptible for the food contamination and human exposure to AFs, especially in the low- and middle-income developing nations (Qian et al., 2013a). Aflatoxin B1 (AFB1) is widely recognized as the most harmful AF, due to its potent toxicity, genotoxicity, and carcinogenicity as well as acute aflatoxicosis in animals and human populations (Kew, 2013; Qian et al., 2013b; Wang and Groopman, 1999). Accordingly, the detection and assessment of AFB1 contamination in human food and animal feed has been a global concern for food safety and public health (Henry et al., 1999; Torres et al., 2015). On the other hand, remarkable efforts have been made to develop novel prevention/intervention strategies against AFB1-induced adverse health effects, including liver cancer risks and growth/developmental disorders in high-risk and vulnerable populations (Mitchell et al., 2014; Xue et al., 2016). Human gastrointestinal tract harbors a complex microbiota that contains more than 100 trillion microbes with over 400 species and carries 150 times more genes than the human genome (Qin et al., 2010). The gut-microbiota constantly provides host with hundreds of micronutrients and functional metabolites, which actively participate into the host enterohepatic cross-talk, as well as the physiological regulations of many organs and systems (Ursell et al., 2014). In recent years, next-generation sequencing technologies have uncovered all kinds of intricate connections among gut-microbiota, dietary composition and host health (Chakraborty et al., 2010; Holmes et al., 2012). In this 3-way relationship, oral exposure to xenobiotics or dietary composition could lead to the alteration of gut-microbiota, and the changes of gut-microbiota may further influence host health in a significant way (Brown and Hazen, 2015). Emerging evidences have demonstrated the causative links between gut-microbial microbiome/metabolome and a series of health problems in host, eg, obesity, metabolic syndrome, nonalcoholic fatty liver disease (NAFLD), colon cancer, inflammatory bowel disease (IBD), and cardiovascular disease (Flint et al., 2012; Holmes et al., 2012; Lee and Hase, 2014; Louis et al., 2014; Ursell et al., 2014). Therapeutic manipulation of gut-microbiota has also exhibited the potential to mitigate a number of metabolic diseases such as obesity, type-2 diabetes mellitus, IBD and NAFLD, most probably by modifying gut-microbiota-dependent metabolites, which are either derived from food by gut-microbiota, or the endogenous metabolites of gut-microbes (Kootte et al., 2012; Schulberg and De Cruz, 2016). We have previously performed 16S rRNA analysis and found the compositional change of fecal microbiome in F344 rats following repeated oral exposure to AFB1 (Wang et al., 2016). Through 16S rRNA sequencing technique, notable enrichment of Clostridiales spp. and depletion of Lactobacillales spp. were found in the rat feces. In the work presented here, the potential impact of such compositional changes on host health at metabolic level was further explored by examining a group of fecal organic acids that are highly associated with gut-microbiota. The studied metabolites include acetic acid, lactic acid, propionic acid, butyric acid, valeric acid, hexanoic acid, cholic acid, deoxycholic acid, pentadecanoic acid (15: 0), 3-phenyllactic acid, pyruvic acid, and linoleic acid (cis-9, cis-12-18:2). The metabolism of these organic acids heavily depends on the metabolic pathways and community structure of gut-microbiota, and also play important roles in host physiology and global metabolic pathways. MATERIALS AND METHODS Chemicals and reagents Pyridine, 2-nitrophenylhydrazine (2-NPH), N-(3-dimethylaminopropyl)-N’-ethylcarbodiimide hydrochloride (EDC), 2-ethylbutyric acid, acetic acid, propionic acid, butyric acid, valeric acid, hexanoic acid, cholic acid, pentadecanoic acid, 3-phenyllactic acid, pyruvic acid, linoleic acid, deoxycholic acid, bisphenol A, hippuric acid, heptadecanoic acid, AFB1, and dimethyl sulfoxide (DMSO) were all purchased from Sigma-Aldrich Inc. (St Louis, Missouri). AFB1 stock solution (25 mg/ml) was prepared in DMSO and diluted to appropriate treatment concentrations upon using. All other reagents and analytical solvents, methanol, acetonitrile, and water were purchased at the highest grade commercially available from Honeywell (Morris Plains, New Jersey). Animal treatment Male Fischer 344 rats (100–120 g) were purchased from Harlan Laboratory (Indianapolis, IN, USA). The animal housing environment was under controlled light/dark cycle (12:12 h) with a temperature of 22°C ± 2°C and relative humidity of 50%–70%. Purified AIN 76A diet and tap water were maintained every day. Upon arrival, animals were allowed for one week of environmental acclimation. One hundred male F344 rats were divided into 4 groups and were gavaged with 0, 5, 25, and 75 μg AFB1/kg body weight (b.w.) per day, respectively. DMSO was used as vehicle solvent. The details of animal protocol were reported in earlier publications, together with body indexes, histopathological assessment and AFB1-Lys pharmacokinetic data (Mohammadagheri et al., 2016; Qian et al., 2013b, 2014, 2016). Briefly, animals were daily administered with AFB1 by gavage for 4 weeks. From the second week to the fourth week, rat feces were daily collected, and weekly pooled for each group. All fecal samples were stored in −80°C freezer. Animal husbandry and care, AFB1 dosing protocol, and sample collection were approved and in strict accordance with the requirements and regulations of the Institutional Animal Care and Use Committee at the University of Georgia. Sample quenching and extraction Sample extraction procedure was similar to what previously published with modifications (de Jonge et al., 2012; Hernández Bort et al., 2014). Cold methanol (−80°C)-based quenching and extraction were applied to the fecal samples for sample pretreatment. The purpose of using cold methanol was to avoid the loss of volatile compounds, and also because methanol is a solvent chemically appropriate for the reaction of 2-NPH derivatization (Peters et al., 2004; Torii et al., 2010; Winder et al., 2008). To perform sample extraction 200 mg rat feces was transferred to the Mobio PowerLyzer tube with preloaded glass beads (0.1 mm i.d.). One milliliter of cold methanol was immediately added into the tube, and fecal pellet was gently crushed using a glass pestle. After grinding, 0.5 ml cold methanol was slowly added to wash the pestle. Then the tube was capped tightly and fastened on a rotary vortex to undergo 20 min vortex at maximum level using a Vortex-Genie 2 Mixer (Scientific Industries). During vortex, sample tube was put back on ice for 2 min in every 5 min, and finally underwent centrifugation at 12 000 rpm for 10 min to spin down cellular debris. A volume of 100 µl supernatant was transferred to an Eppendorf tube, and 50 µl internal standard (2-ethylbutyric acid) stock solution was spiked into the supernatant to achieve a concentration of 1 µg/µl, which was used to compensate technical variabilities. 2-NPH derivatization To perform derivatization, 150 µl sample extract (with internal standard added) was mixed with 45 µl derivatization solution which was freshly prepared by mixing 15 µl EDC solution (0.05 g/mL H2O), 15 µl 2-NPH solution (12.5 mg/ml methanol) and 15 µl 3% pyridine in methanol (v/v). After mild vortex, the tubes were transferred to water bath at 60°C for 60 min. The tubes then were allowed to stay in room temperature for 5 min and went through brief centrifugation in order to collect the liquid left on the tube wall. All sample vials were kept in 4°C sample cooling tray and the analysis was finished within 24 h. High-performance liquid chromatography analysis An Agilent 1200 High-performance liquid chromatography (HPLC) system, consisted of a degasser, a quarterly pump, an autosampler, a diode-array detector, and a fluorescence detector, was used to perform HPLC-profiling analysis. The chromatographic separation was conducted in a Nucleosil C18 reversed-phase column (250 × 4 mm i.d.; ES industries, New Jersey) with particle size of 5 μm and pore diameter of 120 Å. The injection volume was 100 µl and flow rate was kept at 1 ml/min. Column oven temperature was set as 40°C. Mobile phase A was pH 4.5 acidified water adjusted by hydrochloric acid. Mobile phase B was acetonitrile. The gradient eluting condition was: 90% A to 80% A in 0–12 min; 80% A to 70% A in 12–20 min; 70% A to 60% A in 20–30 min; 60% A to 45% A in 30–41 min; 45% A to 10% A in 41–43 min; then keeping at 10% A in 43–58 min; finally, from 10% A to 90% A in 58–61 min for re-balance. The detection channel is 400 nm by DAD, with reference wavelength at 510 ± 60 nm. The representative chromatogram is shown in Figure 1. Lower limit of detection (LLOD), regression standard curves, as well as the other necessary quantitative parameters used for HPLC-profiling analysis are listed in Table 1. Short-chain fatty acids (SCFAs) were recovered using the recovery rates averaged from the feces spiked with SCFA standards of approximately 50%, 100%, and 200% of their levels in control group (Supplementray Table 1). The concentrations of other interested analytes were determined using the recovery rates of structurally close standards which have similar or close structure to the analytes. Specifically, the recovery rate of hippuric acid was used to recover phenyl acids (PAs); heptadecanoic acid was used to recover long-chain fatty acids (LCFAs), and bisphenol A was used to recover bile acids. Further, 2-ethylbutyric acid was used to eliminate the technical variabilities, since it has similar structure with SCFAs. Bisphenol A was used as standard to calculate recoveries for bile acid and derivatives because it is considered to have close structure with estradiol, which was used as internal standard for quantitative analysis of bile acid (Junichi et al., 1978; Rubin, 2011). And no other commercially available compound can be chromatographically separated with bile constituents for the calibration of recovery using current method. Table 1. Analytical Parameters of HPLC-Profiling Analysis Used for the Measurement of Interested Fecal Metabolites Component  Category  RT*  Detection Channel  Regression (X, AUC; Y, ng/μl)  R2  Linear Range ng/μl  LLOD ng/μl  Acetic acid  SCFA  14.9  400 nm  y = 0.0063x − 0.3726  0.9993  0.016–64.8  0.008  Propionic acid  SCFA  19.6  400 nm  y = 0.021x − 0.9273  0.999  0.07–143  0.03  Butyric acid  SCFA  25.1  400 nm  y = 0.0256x − 0.4994  0.9991  0.078–79.5  0.04  Valeric acid  SCFA  31.5  400 nm  y = 0.0208x − 0.4974  0.999  0.054–56.1  0.03  Hexanoic acid  SCFA  37.6  400 nm  y = 0.0309x − 0.356  0.9994  0.074–75.6  0.04  Lactic acid  SCFA  13.8  400 nm  y = 0.0244x − 0.2351  0.9991  0.11–14.33  0.05  Pyruvic acid  Alpha-keto acid  41.3  400 nm  y = 0.0166x + 0.7136  0.9997  6.2–500  0.19  2-Ethylbutyric acid  IS for SCFA  34.2  400 nm  y = 0.1662x - 0.4527  0.9991  0.56–1138  0.28  Niacin  PA  22.1  210 nm  y = 0.0313x − 6.1768  0.9954  1–430  0.25  3-Phenyllactic acid  PA  31.2  400 nm  y = 0.1003x − 0.7134  0.9994  4.7–300  0.58  Hippuric acid  IS for PA  26.3  400 nm  y = 0.6161x + 2.8275  0.9996  4.45–570  2.25  Cholic acid  SA  45.1  400 nm  y = 0.1219x − 6.6046  0.9930  3.9–250  0.49  Deoxycholic acid  SA  47.1  400 nm  y = 0.0371x − 4.8658  0.9930  2.5–330  0.64  Cholesterol  Sterol  47.4  400 nm  y = 0.0686x − 2.4795  0.9900  1.95–125  0.98  Bisphenol A  IS for SA  35.0  210 nm  y = 0.0148x − 6.9647  0.992  0.33–685  0.17  Linoleic acid  LCFA  50.9  400 nm  y = 0.3705x − 31.314  0.9948  3.9–1000  3.9  Pentadecanoic acid  LCFA  51.2  400 nm  y = 0.0636x − 0.3641  0.9990  1.95–500  0.5  Heptadecanoic acid  IS for LCFA  54.5  400 nm  y = 0.1436x − 4.5852  0.9952  2.15–275  1.07  Component  Category  RT*  Detection Channel  Regression (X, AUC; Y, ng/μl)  R2  Linear Range ng/μl  LLOD ng/μl  Acetic acid  SCFA  14.9  400 nm  y = 0.0063x − 0.3726  0.9993  0.016–64.8  0.008  Propionic acid  SCFA  19.6  400 nm  y = 0.021x − 0.9273  0.999  0.07–143  0.03  Butyric acid  SCFA  25.1  400 nm  y = 0.0256x − 0.4994  0.9991  0.078–79.5  0.04  Valeric acid  SCFA  31.5  400 nm  y = 0.0208x − 0.4974  0.999  0.054–56.1  0.03  Hexanoic acid  SCFA  37.6  400 nm  y = 0.0309x − 0.356  0.9994  0.074–75.6  0.04  Lactic acid  SCFA  13.8  400 nm  y = 0.0244x − 0.2351  0.9991  0.11–14.33  0.05  Pyruvic acid  Alpha-keto acid  41.3  400 nm  y = 0.0166x + 0.7136  0.9997  6.2–500  0.19  2-Ethylbutyric acid  IS for SCFA  34.2  400 nm  y = 0.1662x - 0.4527  0.9991  0.56–1138  0.28  Niacin  PA  22.1  210 nm  y = 0.0313x − 6.1768  0.9954  1–430  0.25  3-Phenyllactic acid  PA  31.2  400 nm  y = 0.1003x − 0.7134  0.9994  4.7–300  0.58  Hippuric acid  IS for PA  26.3  400 nm  y = 0.6161x + 2.8275  0.9996  4.45–570  2.25  Cholic acid  SA  45.1  400 nm  y = 0.1219x − 6.6046  0.9930  3.9–250  0.49  Deoxycholic acid  SA  47.1  400 nm  y = 0.0371x − 4.8658  0.9930  2.5–330  0.64  Cholesterol  Sterol  47.4  400 nm  y = 0.0686x − 2.4795  0.9900  1.95–125  0.98  Bisphenol A  IS for SA  35.0  210 nm  y = 0.0148x − 6.9647  0.992  0.33–685  0.17  Linoleic acid  LCFA  50.9  400 nm  y = 0.3705x − 31.314  0.9948  3.9–1000  3.9  Pentadecanoic acid  LCFA  51.2  400 nm  y = 0.0636x − 0.3641  0.9990  1.95–500  0.5  Heptadecanoic acid  IS for LCFA  54.5  400 nm  y = 0.1436x − 4.5852  0.9952  2.15–275  1.07  The minimum data point in the linear regression range (R2 > 0.999) was noted as LOQ. Abbreviations: IS, internal standard for quality control; R2, regression coefficient; LLOD, lower limit of detection; LCFA, long-chain fatty acid; PA, phenyl acid; RT, retention time (min) in chromatogram; SA, steroid acid; SCFA, short-chain fatty acid. The analyte level which generated a signal-to-noise (S/N) ratio of 3 was noted as the LLOD for that analyte. Niacin and cholesterol were not detected in most sample extracts. Table 1. Analytical Parameters of HPLC-Profiling Analysis Used for the Measurement of Interested Fecal Metabolites Component  Category  RT*  Detection Channel  Regression (X, AUC; Y, ng/μl)  R2  Linear Range ng/μl  LLOD ng/μl  Acetic acid  SCFA  14.9  400 nm  y = 0.0063x − 0.3726  0.9993  0.016–64.8  0.008  Propionic acid  SCFA  19.6  400 nm  y = 0.021x − 0.9273  0.999  0.07–143  0.03  Butyric acid  SCFA  25.1  400 nm  y = 0.0256x − 0.4994  0.9991  0.078–79.5  0.04  Valeric acid  SCFA  31.5  400 nm  y = 0.0208x − 0.4974  0.999  0.054–56.1  0.03  Hexanoic acid  SCFA  37.6  400 nm  y = 0.0309x − 0.356  0.9994  0.074–75.6  0.04  Lactic acid  SCFA  13.8  400 nm  y = 0.0244x − 0.2351  0.9991  0.11–14.33  0.05  Pyruvic acid  Alpha-keto acid  41.3  400 nm  y = 0.0166x + 0.7136  0.9997  6.2–500  0.19  2-Ethylbutyric acid  IS for SCFA  34.2  400 nm  y = 0.1662x - 0.4527  0.9991  0.56–1138  0.28  Niacin  PA  22.1  210 nm  y = 0.0313x − 6.1768  0.9954  1–430  0.25  3-Phenyllactic acid  PA  31.2  400 nm  y = 0.1003x − 0.7134  0.9994  4.7–300  0.58  Hippuric acid  IS for PA  26.3  400 nm  y = 0.6161x + 2.8275  0.9996  4.45–570  2.25  Cholic acid  SA  45.1  400 nm  y = 0.1219x − 6.6046  0.9930  3.9–250  0.49  Deoxycholic acid  SA  47.1  400 nm  y = 0.0371x − 4.8658  0.9930  2.5–330  0.64  Cholesterol  Sterol  47.4  400 nm  y = 0.0686x − 2.4795  0.9900  1.95–125  0.98  Bisphenol A  IS for SA  35.0  210 nm  y = 0.0148x − 6.9647  0.992  0.33–685  0.17  Linoleic acid  LCFA  50.9  400 nm  y = 0.3705x − 31.314  0.9948  3.9–1000  3.9  Pentadecanoic acid  LCFA  51.2  400 nm  y = 0.0636x − 0.3641  0.9990  1.95–500  0.5  Heptadecanoic acid  IS for LCFA  54.5  400 nm  y = 0.1436x − 4.5852  0.9952  2.15–275  1.07  Component  Category  RT*  Detection Channel  Regression (X, AUC; Y, ng/μl)  R2  Linear Range ng/μl  LLOD ng/μl  Acetic acid  SCFA  14.9  400 nm  y = 0.0063x − 0.3726  0.9993  0.016–64.8  0.008  Propionic acid  SCFA  19.6  400 nm  y = 0.021x − 0.9273  0.999  0.07–143  0.03  Butyric acid  SCFA  25.1  400 nm  y = 0.0256x − 0.4994  0.9991  0.078–79.5  0.04  Valeric acid  SCFA  31.5  400 nm  y = 0.0208x − 0.4974  0.999  0.054–56.1  0.03  Hexanoic acid  SCFA  37.6  400 nm  y = 0.0309x − 0.356  0.9994  0.074–75.6  0.04  Lactic acid  SCFA  13.8  400 nm  y = 0.0244x − 0.2351  0.9991  0.11–14.33  0.05  Pyruvic acid  Alpha-keto acid  41.3  400 nm  y = 0.0166x + 0.7136  0.9997  6.2–500  0.19  2-Ethylbutyric acid  IS for SCFA  34.2  400 nm  y = 0.1662x - 0.4527  0.9991  0.56–1138  0.28  Niacin  PA  22.1  210 nm  y = 0.0313x − 6.1768  0.9954  1–430  0.25  3-Phenyllactic acid  PA  31.2  400 nm  y = 0.1003x − 0.7134  0.9994  4.7–300  0.58  Hippuric acid  IS for PA  26.3  400 nm  y = 0.6161x + 2.8275  0.9996  4.45–570  2.25  Cholic acid  SA  45.1  400 nm  y = 0.1219x − 6.6046  0.9930  3.9–250  0.49  Deoxycholic acid  SA  47.1  400 nm  y = 0.0371x − 4.8658  0.9930  2.5–330  0.64  Cholesterol  Sterol  47.4  400 nm  y = 0.0686x − 2.4795  0.9900  1.95–125  0.98  Bisphenol A  IS for SA  35.0  210 nm  y = 0.0148x − 6.9647  0.992  0.33–685  0.17  Linoleic acid  LCFA  50.9  400 nm  y = 0.3705x − 31.314  0.9948  3.9–1000  3.9  Pentadecanoic acid  LCFA  51.2  400 nm  y = 0.0636x − 0.3641  0.9990  1.95–500  0.5  Heptadecanoic acid  IS for LCFA  54.5  400 nm  y = 0.1436x − 4.5852  0.9952  2.15–275  1.07  The minimum data point in the linear regression range (R2 > 0.999) was noted as LOQ. Abbreviations: IS, internal standard for quality control; R2, regression coefficient; LLOD, lower limit of detection; LCFA, long-chain fatty acid; PA, phenyl acid; RT, retention time (min) in chromatogram; SA, steroid acid; SCFA, short-chain fatty acid. The analyte level which generated a signal-to-noise (S/N) ratio of 3 was noted as the LLOD for that analyte. Niacin and cholesterol were not detected in most sample extracts. Figure 1. View largeDownload slide HPLC-profiling chromatograms of fecal extracts from control (upper) and exposure group (lower) after 2-NPH derivatization. 2-ethyl butyric acid was used as internal standard (IS). The detection channel of DAD is 400 nm with a reference channel as 510 ± 60 nm. Down-regulated organic acids are labeled on the upper panel, whereas up-regulated organic acids are labeled on the lower panel. Specific retention time and relevant information are available in Table 1. Figure 1. View largeDownload slide HPLC-profiling chromatograms of fecal extracts from control (upper) and exposure group (lower) after 2-NPH derivatization. 2-ethyl butyric acid was used as internal standard (IS). The detection channel of DAD is 400 nm with a reference channel as 510 ± 60 nm. Down-regulated organic acids are labeled on the upper panel, whereas up-regulated organic acids are labeled on the lower panel. Specific retention time and relevant information are available in Table 1. Method validation and optimization Methanol blanks were spiked with SCFA standards to generate test solutions with concentrations of approximately 50%, 100%, and 200% of the actual SCFA amounts measured in the sample extracts. The test solutions were derivatized using 2-NPH and EDC and were immediately used for HPLC-profiling analysis. The analytical precision of the method was validated based on: (1) interday coefficient of variation (CV) of the peak intensities of SCFAs at 3 spike levels in 3 consecutive days, with one bunch performed per day; (2) inter-assay CV of the peak intensities of SCFAs at 3 spike levels in 7 consecutive assays; (3) intra-assay CV of the peak intensities of SCFAs at 3 spike levels, with 4 repeats conducted at each level. Analytical accuracy was examined using recoveries with CV, and the formula to calculate recovery rate is: recovery % = (analyte amount measured in the extract of standard-spiked feces – analyte amount measured in the extract of nonspiked feces) × 100/(amount of spiked analyte) (Han et al., 2013b). 16S rRNA analysis Briefly, total fecal genomic DNA which contains 16S rRNA was extracted using QIAamp DNA stool mini kits (QIAGEN, Valencia, California). A 2-step Quadruple-index PCR method was used to prepare the 16S rRNA gene libraries according to Klindworth et al. (2013). Sequencing of these 16S rRNA fragment libraries was performed in the Georgia Genomic Facility (University of Georgia, Athens, Georgia) using the Illumina MiSeq with v2 500 cycle chemistry, resulting in paired-end 250 base reads to obtain approximately 30 000 reads per sample. The 16S rRNA fragment amplified in this study is from site 358 to 784 under Escherichia coli system of nomenclature (Klindworth et al., 2013). The raw paired-end, demultiplexed sequence read was merged using FLASH 1.2.9 in Geneious 8.1 software (Biomatters Inc, San Francisco, California). All internal tags, base spacers, and locus-specific primers of merged sequences were trimmed and sequences outranged 400–450 base-pairs were discarded. Outputs from Geneious 8.1 were quality filtered using QIIME pipeline (Quantitative Insights Into Microbial Ecology) (Caporaso et al., 2010). Representative sequences for each operational taxonomic units (OTUs) were compared with the Greengene 16S rRNA gene database 13-8 release (DeSantis et al., 2006) using uclust algorithm with the similarity threshold of 90%. The top 3 database hits that matched the above representative sequences for each OTU were selected. Statistics and software Data normality examination, homogeneity test, 1-way ANOVA, and Welch’s t test, were all performed using SPSS 22. Levene statistic was used to test homogeneity of variances and Welch-Brown-Forsythe statistic was used to test the equality of means. Tukey’s test was used for post hoc analysis in ANOVA. When data failed to follow normality of distribution, Kruskal-Wallis H test was applied to replace 1-way ANOVA. Mixed-effects model regression was performed using STATA 14.1. Pearson’s correlation analysis, and construction of heat map and hierarchical tree were performed using R. Mann-Whitney U test was used to compare the differences of fecal organic acids (except for SCFAs) when dose effect was the only factor being analyzed, with p value < .05 considered to be statistically significant. RESULTS Validation and Optimization of HPLC-Profiling Method Our initial effort was to optimize conditions for fecal sample extraction and metabolites enrichment. However, centrifugal evaporation resulted in significant loss of SCFAs (20%–50%) in the sample extracts, as found by HPLC analysis (data not shown). For this reason, sample enrichment was avoided during sample preparation. Nonetheless, interested analytes are still detected from fecal samples. In terms of precolumn derivatization and HPLC-profiling analysis, the validation work included intraassay precision, interassay precision, interday precision, and accuracy. Shown in supplementary Table 1, most values of measured metabolites showed CV < 8%. The sensitivity and LLOD were determined for all analyzed metabolites, as shown in the Table 1. Internal standards were used to confirm the precision and accuracy, and recovery rate was ranged from 33% to 74% for the all SCFA standards spiked into fecal samples of the control and AFB1-dosed rats, with CV < 5%. Using this validated method, the peak identity and concentrations of interested metabolites were further determined from the chromatogram of fecal extracts, as shown in Figure 1. Four categories of metabolites were measured in the study: SCFAs, including acetic acid, butyric acid, hexanoic acid, lactic acid, propionic acid, and valeric acid; LCFAs, including linoleic acid (cis-9, cis-12-18:2) and pentadecanoic acid (15:0), bile acids, including cholic acid and deoxycholic acid, and other metabolites, including 3-phenyllactic acid and pyruvic acid (Table 1). AFB1 Exposure Affects SCFA Production of Gut-Microbiota Rats were exposed to AFB1 at doses of 0, 5, 25, and 75 μg/kg b.w., which are noted as control, low-, middle- and high-dose groups in the study. As shown in Figure 2 and Supplementary Table 2, significant change of fecal SCFA levels was found in AFB1-exposed groups. The measured levels of SCFAs in the untreated control group were comparable over the time course from 2- to 4-week, but notable reduction of acetic acid, propionic acid, butyric acid, hexanoic acid, and lactic acid were detected in the rat feces of AFB1-exposed groups. In the low-dose group, fecal SCFA levels seemed to be affected by the time of exposure. The fecal levels of acetic acid, propionic acid, butyric acid, lactic acid, valeric acid and hexanoic acid were 46.6%, 39.9%, 68.4%, 79.9%, 95.3%, and 63% of the control after 2 weeks of exposure, but the percentages went to 70.7%, 77.6%, 35.1%, 34.4%, 75.6%, and 86.7% of the control after 4 weeks of exposure, indicating the time-effect of AFB1-exposure on SCFA levels in the low-dose group. The fecal levels of SCFAs in the middle- and high-dose groups were generally not affected by the exposure time, except for propionic acid in middle-dose group at 2-week, and lactic acid and hexanoic acid in high-dose group at 4-week, which showed about 50% changes of fecal levels compared with control. As shown in Supplementary Table 2, the fecal levels of 6 SCFAs in the middle-dose group, were 17.7%, 31.1%, 26.1%, 20.1%, 90.7%, and 19.9% of the control in 2-week, and were 21.6%, 15.3%, 24.6%, 17.2%, 88.9%, and 27.3% of the control in 4-week. Similarly, In the high-dose group the 6 SCFAs were 22%, 22.2%, 21.9%, 12.1%, 44.2%, and 20.5% of the control in 2-week, and were 25%, 34.4%, 17.2%, 6.8%, 42.2%, and 52% of the control in 4-week. Remarkable changes were found for fecal propionic acid level in middle-dose group, which was reduced from 31.1% of control to 15.3% of control from 2- to 4-week; valeric acid in high-dose group, which was reduced from 12.1% of control to 6.8% of control; and lactic acid in high-dose group, which was elevated from 20.5% of control to 52% of control from 2- to 4-week. Figure 2. View largeDownload slide Fecal SCFA levels of rats treated with 0, 5, 25, and 75 μg AFB1/kg b.w. X-axis indicates duration of treatment. Significance of one-way ANOVA or Kruskal-Wallis H Test is indicated by string labels: same string indicating p > .05; string with partly overlapped character(s) indicating p < .05; totally different string indicating p < .01. Error bar indicates standard deviation (n = 5). Specific data are available in Supplementary Table 2. Figure 2. View largeDownload slide Fecal SCFA levels of rats treated with 0, 5, 25, and 75 μg AFB1/kg b.w. X-axis indicates duration of treatment. Significance of one-way ANOVA or Kruskal-Wallis H Test is indicated by string labels: same string indicating p > .05; string with partly overlapped character(s) indicating p < .05; totally different string indicating p < .01. Error bar indicates standard deviation (n = 5). Specific data are available in Supplementary Table 2. To examine the AFB1 dose-, time-, and time × dose interaction effects on fecal SCFA levels, mixed-effects regression model was applied to analyze the linear correlation between the AFB1-dose/time and SCFA levels. As shown in Table 2, significant dose effect and dose × time interaction effect were found. Further, Pearson’s correlation analysis was performed to examine the possible link between the changes of SCFA levels and the community structure of gut-microbiota. The correlation results were shown in the hierarchical tree and heat map in Figure 3. Briefly, strains belonging to Firmicutes Clostridiales order were highly clustered and showed inverse correlation with the fecal levels of SCFAs following AFB1-exposure, while Lactobacillales Streptococcus and Clostridiales Roseburia, 2 SCFA-producing strains, were depleted in the feces. All 6 SCFAs are correlated in the same cluster of Pearson’s r distance. Table 2. Mixed-Effects Model Analysis Between AFB1-Treatment (Dose, Time and Interaction) and Fecal Levels of SCFAs   Fixed Effect   Random Effect   SCFAs  Dose  SE  p  Time  SE  p  Interactiona  SE  p  Estimate  SE  Acetic acid  −0.1103  0.025  <.001  0.1613  0.133  0.227  10.361  2.149  <.001  31.288  6.021  Propionic acid  −0.0617  0.015  <.001  0.054  0.082  0.509  6.718  1.301  <.001  11.798  2.270  Butyric acid  −0.1442  0.015  <.001  −0.0202  0.086  0.921  13.226  1.663  <.001  18.316  3.525  Valeric acid  −0.0028  0.001  <.001  −0.0007  0.004  0.863  7.900  0.977  <.001  0.026  0.005  Hexanoic acid  −0.0179  0.004  <.001  −0.0074  0.022  0.740  2.113  0.365  <.001  0.881  0.170  Lactic acid  −0.0207  0.003  <.001  0.0228  0.018  0.210  1.280  0.298  <.001  0.586  0.112    Fixed Effect   Random Effect   SCFAs  Dose  SE  p  Time  SE  p  Interactiona  SE  p  Estimate  SE  Acetic acid  −0.1103  0.025  <.001  0.1613  0.133  0.227  10.361  2.149  <.001  31.288  6.021  Propionic acid  −0.0617  0.015  <.001  0.054  0.082  0.509  6.718  1.301  <.001  11.798  2.270  Butyric acid  −0.1442  0.015  <.001  −0.0202  0.086  0.921  13.226  1.663  <.001  18.316  3.525  Valeric acid  −0.0028  0.001  <.001  −0.0007  0.004  0.863  7.900  0.977  <.001  0.026  0.005  Hexanoic acid  −0.0179  0.004  <.001  −0.0074  0.022  0.740  2.113  0.365  <.001  0.881  0.170  Lactic acid  −0.0207  0.003  <.001  0.0228  0.018  0.210  1.280  0.298  <.001  0.586  0.112  a Estimate of interaction effect resulted by both dose and treatment time on fecal SCFA levels. Table 2. Mixed-Effects Model Analysis Between AFB1-Treatment (Dose, Time and Interaction) and Fecal Levels of SCFAs   Fixed Effect   Random Effect   SCFAs  Dose  SE  p  Time  SE  p  Interactiona  SE  p  Estimate  SE  Acetic acid  −0.1103  0.025  <.001  0.1613  0.133  0.227  10.361  2.149  <.001  31.288  6.021  Propionic acid  −0.0617  0.015  <.001  0.054  0.082  0.509  6.718  1.301  <.001  11.798  2.270  Butyric acid  −0.1442  0.015  <.001  −0.0202  0.086  0.921  13.226  1.663  <.001  18.316  3.525  Valeric acid  −0.0028  0.001  <.001  −0.0007  0.004  0.863  7.900  0.977  <.001  0.026  0.005  Hexanoic acid  −0.0179  0.004  <.001  −0.0074  0.022  0.740  2.113  0.365  <.001  0.881  0.170  Lactic acid  −0.0207  0.003  <.001  0.0228  0.018  0.210  1.280  0.298  <.001  0.586  0.112    Fixed Effect   Random Effect   SCFAs  Dose  SE  p  Time  SE  p  Interactiona  SE  p  Estimate  SE  Acetic acid  −0.1103  0.025  <.001  0.1613  0.133  0.227  10.361  2.149  <.001  31.288  6.021  Propionic acid  −0.0617  0.015  <.001  0.054  0.082  0.509  6.718  1.301  <.001  11.798  2.270  Butyric acid  −0.1442  0.015  <.001  −0.0202  0.086  0.921  13.226  1.663  <.001  18.316  3.525  Valeric acid  −0.0028  0.001  <.001  −0.0007  0.004  0.863  7.900  0.977  <.001  0.026  0.005  Hexanoic acid  −0.0179  0.004  <.001  −0.0074  0.022  0.740  2.113  0.365  <.001  0.881  0.170  Lactic acid  −0.0207  0.003  <.001  0.0228  0.018  0.210  1.280  0.298  <.001  0.586  0.112  a Estimate of interaction effect resulted by both dose and treatment time on fecal SCFA levels. Figure 3. View largeDownload slide Hierarchical cluster tree and heat map to show cross correlations for SCFAs and top 18 significantly altered gut-microbial strains discovered by previous 16s rRNA data. Data were transferred to fold change of exposure group versus control. Hierarchical clusters are constructed based on Pearson’s r distance. Red-blue color bar indicates Pearson’s correlation coefficient between 2 correlated components. SCFAs are negatively correlated with the Clostridial Ruminococcaceae strains that are frequently seen in the stools from patients with Crohn’s disease and obesity. The suppressed strains belong to Lactobacillales and Clostridial Roseburia. Phylogenetic taxa information can be accessed in reference (Wang et al., 2016). (For interpretation of the reference to color in this figure legend, the reader is referred to the web version of this article.) Figure 3. View largeDownload slide Hierarchical cluster tree and heat map to show cross correlations for SCFAs and top 18 significantly altered gut-microbial strains discovered by previous 16s rRNA data. Data were transferred to fold change of exposure group versus control. Hierarchical clusters are constructed based on Pearson’s r distance. Red-blue color bar indicates Pearson’s correlation coefficient between 2 correlated components. SCFAs are negatively correlated with the Clostridial Ruminococcaceae strains that are frequently seen in the stools from patients with Crohn’s disease and obesity. The suppressed strains belong to Lactobacillales and Clostridial Roseburia. Phylogenetic taxa information can be accessed in reference (Wang et al., 2016). (For interpretation of the reference to color in this figure legend, the reader is referred to the web version of this article.) AFB1 Exposure Affects Metabolism of Other Gut-Microbiota Dependent Organic Acids We next examined the impacts of AFB1 treatment on a set of key organic acids after 4 weeks of AFB1 exposure, including cholic acid, deoxycholic acid, 3-phenyllactic acid, pyruvic acid, pentadecanoic acid (15:0), and linoleic acid (cis-9, cis-12-18:2). Oral AFB1 exposure significantly elevated fecal LCFAs (linoleic acid and pentadecanoic acid). Specifically, the level of linoleic acid was 95.51 ± 24.18 ng/mg in the control group, and increased to 1274.82 ± 363.02 ng/mg in the low-dose group and 1079.18 ± 760.29 ng/mg in the middle-dose group; the level of pentadecanoic acid in the control group was 20.26 ± 21.99 ng/mg, and increased to 64.76 ± 36.57 ng/mg in the low-dose group and 74.60 ± 53.35 ng/mg in the middle-dose group; the most significantly altered organic acid was linoleic acid, with over 10-fold increase found in low- and middle-dose groups (Figure 4). Figure 4. View largeDownload slide Fecal concentrations of cholic acid, deoxycholic acid, linoleic acid, pentadecanoic acid, pyruvic acid, and 3-phenyllactic acid measured from the experimental groups treated with 0, 5, and 25 μg AFB1/kg b.w. via HPLC-profiling analysis. Nonparametric Mann-Whitney U test was applied for all comparisons (n = 10). Box with middle vertical line represents 25%, 50%, and 75% percentile of data. Vertical lines of box plots indicate SD, multiplied with 1.5-fold coefficient in order to stretch out from box. Figure 4. View largeDownload slide Fecal concentrations of cholic acid, deoxycholic acid, linoleic acid, pentadecanoic acid, pyruvic acid, and 3-phenyllactic acid measured from the experimental groups treated with 0, 5, and 25 μg AFB1/kg b.w. via HPLC-profiling analysis. Nonparametric Mann-Whitney U test was applied for all comparisons (n = 10). Box with middle vertical line represents 25%, 50%, and 75% percentile of data. Vertical lines of box plots indicate SD, multiplied with 1.5-fold coefficient in order to stretch out from box. Oral AFB1 exposure also significantly elevated fecal levels of cholic acid, pyruvic acid, and 3-phenyllactic acid. The level of cholic acid in the control group was 56.15 ± 27.15 ng/mg, and increased to 128.46 ± 15.35 ng/mg in the low-dose group and 122.60 ± 7.32 ng/mg in the middle-dose group; the level of pyruvic acid in the control group was 38.46 ± 26.92 ng/mg, and increased to 75.57 ± 22.18 ng/mg in the low-dose group and 175.23 ± 74.98 ng/mg in the middle-dose group, and the level of 3-phenyllactic acid in the control group was 28.82 ± 9.04 ng/mg, and increased to 83.89 ± 18.10 ng/mg in the low-dose group and 107.84 ± 74.9 ng/mg in the middle-dose group, respectively. On the other hand, the level of deoxycholic acid was significantly reduced, to about the half level (5.13 ± 5.09 ng/mg) in the low-dose group from 10.18 ± 8.69 ng/mg in the control group, and completely dropped to undetectable level in the middle-dose group. DISCUSSION Results of this study clearly demonstrated that up to 2-week oral AFB1 exposure disrupted metabolism of gut microbiota-dependent organic acids, as evidenced by significant reduction in fecal level of SCFAs and deoxycholic acid, and significant increases in LCFAs and other organic acids such as pyruvic acid, 3-phenyllactic acid, and cholic acid. All these microbial metabolites play key roles in the metabolism of gut-microbiota and the maintenance of host nutrition and health. The detection of trace amounts of SCFAs in complex media, eg, bio-fluids and fecal extracts, has been reported by several studies using HPLC-profiling combined with precolumn derivatization with 2-NPH (Han et al., 2013a; Miwa et al., 1985; Peters et al., 2004), but the application of this method has not yet reported in AFB1-exposed rat models. The chemical derivatization is usually performed in mild aqueous or alcohol environment, in which carbonyl compounds (carboxylic acid, aldehyde, and ketone) bonded to 2-NPH and form hydrazides. The reaction is activated by water-soluble EDC which serves as carbodiimide crosslinker. Before in-lab analysis, method validation was conducted to confirm whether the analytical procedure is suitable and reliable for a specific analytical task (VanHook, 2016). The accuracy and reliability of analytical method were further carefully validated (Supplementary Table 1). The measured values and interclass ratio of SCFAs in our study are comparable with several other publications (Cummings et al., 1987; Torii et al., 2010; Zhao et al., 2006). In this study we found significant inhibitory effects of AFB1-exposure on synthesis of SCFAs, which has not previously reported. The decrease in SCFAs was consistent with the depletion of SCFA-producing strains such as Lactobacillales Streptococcus and Clostridiales Roseburia (Duncan et al., 2002; Kleessen et al., 1997). SCFAs are a group of beneficial aliphatic acids that are mainly produced by the anaerobic bacterial fermentation of resistant starches and insoluble fibers in the gastrointestinal tract of human and other mammals (Brockman, 2005). They are structurally constructed by 1–6 carbon atom(s), including formic acid (C1), acetic acid (C2), propionic acid (C3), butyric acid (C4), valeric acid (C5), hexanoic acid (C6), and a variety of branched-chain isomers of these acids. A variety of nutritional and physiological associations of SCFAs with liver diseases, general immunity, IBD, cardiovascular disease, and diabetes were found in many epidemiological studies and in various in vivo and in vitro models (Corrêa-Oliveira et al., 2016; Galisteo et al., 2008; Morrison and Preston, 2016; Wong et al., 2006; Zhao et al., 2006). Acetic acid, butyric acid and propionic acid can be produced by gut-microbiota via fermentation of insoluble fibers (Corrêa-Oliveira et al., 2016; Morrison and Preston, 2016; Torii et al., 2010). SCFAs were mainly produced from the fermentation process of certain strains such as Lactobacillales Streptococcus. The aflatoxin-caused reduction in these microbial strains (Wang et al., 2015) could eventually affect the fermentation process and cause reduction of SCFAs. Mixed-effects model analysis showed that—3 major SCFAs, ie, acetic acid, butyric acid and propionic acid were the most significantly affected by AFB1-dose and dose × time interaction, but not time of treatment (Table 2). It was demonstrated in our earlier 16S rRNA analysis, that the adaption of gut-microbiota community structure was featured by the elevation of relative abundances of Clostridiales spp., but decrease of Lactobacillales Streptococcus and Clostridiales Roseburia (Wang et al., 2016). Given that dose-response was also found for specific gut-microbial strains, Pearson’s correlation analysis between fecal SCFA levels and gut-microbial strains was performed to show their correlation. We found that strains from Firmicutes Clostridiales, an order associated with diarrhea in human and other mammals (Suchodolski et al., 2015), were highly clustered, and exhibited inverse correlation with SCFAs. By contrast, the relative abundances of Lactobacillales Streptococcus and Clostridiales Roseburia were positively correlated with fecal SCFAs. Both of these microbes are SCFA-producing strains (Duncan et al., 2002; Kleessen et al., 1997). The depletion of SCFAs in feces reflected the suppression of microbial fermentation on resistant starches and insoluble fibers. This may further result in a wide range of adverse consequences, because the receptors of SCFAs such as GPR43, GPR41, OLFR78, and GPR109A, are extensively distributed in different organs and systems, and are involved in a myriad of regulatory axis and pathways, such as mobility of gut epithelium, liver lipogenesis, global immunity, cell cycle, oncogenesis, apoptosis and proliferation (Brown et al., 2003; Natarajan and Pluznick, 2014; Smith et al., 2013). Moreover, dietary supply of SCFAs has recently been found to be able to protect against type-I diabetes in mice model (Wen and Wong, 2017). In addition to SCFAs, there are a great number of organic acids present in gut and feces that play important physiological roles. They are either food-derived nutrients or the metabolic products generated in gut-microbiota and host metabolisms. Interested organic acids in our study included fecal linoleic acid (cis-9, cis-12–18:2), pentadecanoic acid (15:0), pyruvic acid, 3-phenyllactic acid, cholic acid, and deoxycholic acid, which were remarkably altered in the feces following AFB1 exposure (Figure 4). Linoleic acid is an omega-6 polyunsaturated fatty acid known as an essential dietary nutrient that cannot be de novo synthesized by human body. The unsaturated fatty acids are known to carry with various health-promoting functions, such as antioxidant defense, suppression of blood levels of triglycerides and cholesterol, maintenance of glucose tolerance, and mitigation of hyperinsulinemia (Whelan and Fritsche, 2013). Most of these beneficial functions has been identified in conjugated linoleic acids, mainly as cis-9, trans-11 C18:2, trans-9, trans-11 C18:2, and trans-10, cis-12 C18:2 (Worley and Powers, 2016; Yatsunenko et al., 2012). Pentadecanoic acid is known to carry a variety of regulatory functions in cell signaling, glucose utilization, and the maintenance of the integrity and stability of gut epithelium (Santaren et al., 2014). The abnormal accumulation of linoleic acid and pentadecanoic acid in rat feces suggested a suppressed intestinal absorption of LCFAs, which is disadvantageous for host health. The deficient bioavailability may be caused by several conditions. First, the decrease of SCFAs may affect the epithelial delivery of nutrients to hepatic portal vein, since SCFAs are well-known nutrients that are able to enhance colonic blood flow and epithelial motility by providing energy and activating G-protein receptors (Scheppach, 1994). Second, certain gut-microbial strains are capable of transferring LCFAs into their conjugated forms which are easier to be absorbed (Druart et al., 2014). For example, Lactobacillus, Propionibacterium, and Bifidobacterium species can produce conjugated linoleic acid from dietary linoleic acid by using microbial lipoxygenases and cyclooxygenases—a process known to facilitate the absorption of LCFAs (Yatsunenko et al., 2012). Our previous 16S rRNA analysis demonstrated that these strains were suppressed by AFB1, which could affect the uptake and reduce bioavailability of LCFAs (Wang et al., 2016). Bile acids are endogenous steroid acids synthesized from cholesterol by liver cells of most vertebrates. Different species have distinct molecular forms of bile acids generated, but some major types of bile acids are shared by different species, eg, cholic acid and chenodeoxycholic acid in human and rat (Whittaker and Chipley, 1986). In human, bile acids are stored in the gallbladder, and are released into duodenum with bile juice under the dietary stimulation. Upon arriving small intestine, bile acids participate in the digestion and absorption of fats and fat-soluble vitamins and can be further metabolized into a variety of secondary metabolites by gut-microbiota. In this work, cholic acid and deoxycholic acid were selected as representative primary and secondary bile acids to probe the microbial metabolism of bile acids, since they are found in both human and rat feces at comparatively high levels. We found a remarkable elevation of cholic acid level with a significant reduction of deoxycholic acid level in AFB1 exposed rat feces. The significant elevation of cholic acid is generally considered to be harmful to host health. Abnormal increase of cholic acid is associated with liver pathogenesis such as cirrhosis and steatosis (Mouzaki et al., 2016), and is also known as a risk factor for intestinal inflammation (Camilleri, 2011). Besides, extra cholic acid in gut may partially contribute to the incidence of colon cancer by stimulating the growth of a small-size benign adenoma to larger size (Rowland, 2012). In correspondence with the increase of cholic acid, we found severe liver damages and pathogenesis in the AFB1-treated rats (Qian et al., 2013b, 2016). The abnormal reduction of deoxycholic acid can be attributed to the relative abundances of the deoxycholic acid-producing microbes, such as Lachospiraceae, Clostridiaceae, and Ruminococcaceae, were all decreased by AFB1 exposure (Wang et al., 2016). In both human studies and rodent models these strains can metabolize primary bile acids into secondary bile acids (Labbé et al., 2014). There are also interactions among primary bile acids, secondary bile acids, and SCFAs in regulating host health, and the elevation of intestinal primary bile acids with decreased secondary bile acid was associated with the incidences of dysbiosis and IBD in humans (Lefebvre et al., 2009). The increase of fecal cholic acid in combination with the decrease of SCFAs were previously observed in the patients with colon cancer (Weir et al., 2013). Pyruvic acid is a well-known energetic α-keto acid that is involved in a number of important metabolic pathways of both gut-microbiota and host. It serves as energy supply to cells through Krebs cycle, and can be transferred to SCFAs by Lactobacilli strains through glycolytic pathway (Pessione, 2012). Pyruvic acid can be transferred to carbohydrates via gluconeogenesis, or participate in the biosynthesis of fatty acids after binding with acetyl-CoA (Kim et al., 2016). Since pyruvic acid takes such a central role in the catabolism of carbohydrates, its unusual accumulation in rat feces reflected a suppressed energy utilization and disruption of glycolysis of gut-microbiota. This may also result in the decrease of microbial synthesis of SCFAs (VanHook, 2016). It seems that the reduction of SCFAs is not only caused by alteration of community structure of gut-microbiota, but also related with the specific metabolic pathway. Last, 3-phenyllactic acid, a central intermediate product in the upstream of phenylalanine catabolism (Stark et al., 1979), was accumulated in the rat feces following exposure to AFB1. The abnormal accumulation of 3-phenyllactic acid suggested the disruption of gut-microbial phenylalanine pathway (Camilleri, 2011). The phenylalanine pathway is known to generate L-3, 4-dihydroxyphenylalanine (L-DOPA) and tyrosine. L-DOPA is the precursor to a number of important neurotransmitters such as dopamine, norepinephrine, and epinephrine. In addition, L-DOPA itself also mediates neurotrophic factor release by the brain and central neuro system (CNS) (Lopez et al., 2008). For these reasons, the down-regulation of phenylalanine pathway may interfere with host CNS function and cause-related health problems. Dietary AFB1 exposure and AFB1-induced adverse health effects remain a major public health problem in many tropical developing nations. The range of dosage used in this study (5–75 µg/kg b.w.) was relevant to human exposure, based on 300 g corn consumption per day (Gwirtz and Garcia-Casal, 2014) and oral exposure levels ranged from 100 to 1000 μg/kg corn for high-risk human populations in Kenya, Ghana, and Guangxi area of China (Azziz-Baumgartner et al., 2005; Groopman et al., 1992; Tang et al., 2009). The dose was multiplied by an adjusting factor of 6.2 in order to transfer human exposure to that in rats (Nair and Jacob, 2016). Regarding the mechanisms behind the metabolite alterations found in this study, there are several mechanisms involved: (1) AFB1, as a natural antimicrobial agent, can selectively inhibit certain bacterial strains and influence on the growth of other strains (Arai et al., 1967; Haskard et al., 2001), as shown in the compositional changes of gut-microbiota revealed by 16 s rRNA analysis; (2) AFB1, as a potent hepatic toxin, can damage liver—the major metabolic organ and in turn induce the metabolic changes for the supply of nutrients and metabolites to host cells and tissues, including gut cells, which may play an important role in the metabolism of gut-microbiota (Atroshi et al., 1998). However, the more specific mechanism related to how AFB1 induces changes of gut-microbiota community structure and the dependent metabolites still need to be clarified in future study. Taken together, as summarized in Figure 5 based on our previous studies (Qian et al., 2013a,b,, 2014; Wang et al., 2016), oral exposure to AFB1 in rat results in significant toxic effects, biochemical alterations, and induction of preneoplastic GST-P positive liver foci. With same study design, here we show that AFB1 can induce the adverse change of community structure of gut-microbiota and significant disruption of multiple metabolic pathways, such as production of SCFAs, secretion, and metabolism of bile acids, absorption of LCFAs, catabolism of phenylalanine, and metabolism of pyruvic acid. These pathways take central and key positions in the global metabolism of gut-microbiota and maintenance of host health, for examples, energy-delivery pathways related with pyruvic acid, including gluconeogenesis, fatty acid synthesis, Krebs cycle and production of lactic acid. Therefore, our data suggest that gut-microbiota may partially be involved in the pathological mechanism and progressions of AFB1-exposure induced adverse health outcomes in F344 rat model, and presumably also in humans. Figure 5. View largeDownload slide Summary of the adverse health outcomes associated with dietary exposure to AFB1 in F344 rat model. Gray arrow indicates the changing trends of microbial taxa, biomarkers, phenotypes, and metabolites induced by AFB1-treatment. The establishment of rat model for AFB1 oral exposure, as well as the 16s rRNA analysis have been published already (Qian et al., 2013a,b, 2014; Wang et al., 2016). Briefly, male F344 rats were gavaged with AFB1 at doses of 0, 5, 10, 25, 50, and 75 μg/kg b.w. per day. The major pathological changes are summarized on the left panel. After 3 weeks of exposure to 75 μg AFB1/kg b.w., bile duct proliferation, liver GST-P+ foci co-occurred, followed by proliferation foci formation after 4 weeks and dramatic alanine transaminase, aspartate transaminase and creatine kinase elevations after 5 weeks of treatment. Figure 5. View largeDownload slide Summary of the adverse health outcomes associated with dietary exposure to AFB1 in F344 rat model. Gray arrow indicates the changing trends of microbial taxa, biomarkers, phenotypes, and metabolites induced by AFB1-treatment. The establishment of rat model for AFB1 oral exposure, as well as the 16s rRNA analysis have been published already (Qian et al., 2013a,b, 2014; Wang et al., 2016). Briefly, male F344 rats were gavaged with AFB1 at doses of 0, 5, 10, 25, 50, and 75 μg/kg b.w. per day. The major pathological changes are summarized on the left panel. After 3 weeks of exposure to 75 μg AFB1/kg b.w., bile duct proliferation, liver GST-P+ foci co-occurred, followed by proliferation foci formation after 4 weeks and dramatic alanine transaminase, aspartate transaminase and creatine kinase elevations after 5 weeks of treatment. SUPPLEMENTARY DATA Supplementary data are available at Toxicological Sciences online. ACKNOWLEDGMENTS Authors thank Dr Guoqing Qian and Dr Kathy Xue for their assistance in animal experiments. Interdisciplinary Toxicology Program at the University of Georgia Graduate School provided stipend supports. FUNDING This work was supported partially by the research contract (ECG-A-00-07-00001-00), from the United States Agency for International Development via Peanut CRSP and the Center for Mycotoxin Research at the College of Public Health, University of Georgia. REFERENCES Arai T., Ito T., Koyama Y. ( 1967). Antimicrobial activity of aflatoxins. J. Bacteriol . 93, 59– 64. Google Scholar PubMed  Atroshi F., Rizzo A., Westermarck T., Ali-Vehmas T. ( 1998). Effects of tamoxifen, melatonin, coenzyme q10, andl-carnitine supplementation on bacterial growth in the presence of mycotoxins. Pharmacol. Res . 38, 289– 295. Google Scholar CrossRef Search ADS PubMed  Azziz-Baumgartner E., Lindblade K., Gieseker K., Rogers H. S., Kieszak S., Njapau H., Schleicher R., McCoy L. F., Misore A., DeCock K.et al.  , . ( 2005). Case-control study of an acute aflatoxicosis outbreak, Kenya, 2004. Environ. Health Perspect . 113, 1779– 1783. Google Scholar CrossRef Search ADS PubMed  Brockman R. ( 2005). Glucose and short-chain fatty acid metabolism. In Quantitative Aspects of Ruminant Digestion and Metabolism , pp. 291– 310 ( Dijkstra J., Forbes J. M., France J., Eds.), CAB International, Wallingford, UK. Google Scholar CrossRef Search ADS   Brown A. J., Goldsworthy S. M., Barnes A. A., Eilert M. M., Tcheang L., Daniels D., Muir A. I., Wigglesworth M. J., Kinghorn I., Fraser N. J.et al.  , . ( 2003). The Orphan G protein-coupled receptors GPR41 and GPR43 are activated by propionate and other short chain carboxylic acids. J. Biol. Chem . 278, 11312– 11319. Google Scholar CrossRef Search ADS PubMed  Brown J. M., Hazen S. L. ( 2015). The gut microbial endocrine organ: Bacterially-derived signals driving cardiometabolic diseases. Annu. Rev. Med . 66, 343. Google Scholar CrossRef Search ADS PubMed  Camilleri M. ( 2011). Managing symptoms of irritable bowel syndrome in patients with inflammatory bowel disease. Gut  60, 425– 428. Google Scholar CrossRef Search ADS PubMed  Caporaso J. G., Kuczynski J., Stombaugh J., Bittinger K., Bushman F. D., Costello E. K., Fierer N., Peña A. G., Goodrich J. K., Gordon J. I.et al.  , . ( 2010). QIIME allows analysis of high-throughput community sequencing data. Nat. Methods  7, 335. Google Scholar CrossRef Search ADS PubMed  Chakraborty S., Zawieja S., Wang W., Zawieja D. C., Muthuchamy M. ( 2010). Lymphatic system: A vital link between metabolic syndrome and inflammation. Ann. N.Y. Acad. Sci . 1207, E94– 102. Google Scholar CrossRef Search ADS PubMed  Corrêa-Oliveira R., Fachi J. L., Vieira A., Sato F. T., Vinolo M. A. R. ( 2016). Regulation of immune cell function by short-chain fatty acids. Clin. Transl. Immunol.  5, e73– e80. Google Scholar CrossRef Search ADS   Cummings J., Pomare E., Branch W., Naylor C., Macfarlane G. ( 1987). Short chain fatty acids in human large intestine, portal, hepatic and venous blood. Gut  28, 1221– 1227. Google Scholar CrossRef Search ADS PubMed  DeSantis T., Hugenholtz P., Keller K., Brodie E., Larsen N., Piceno Y., Phan R., Andersen G. L. ( 2006). NAST: A multiple sequence alignment server for comparative analysis of 16S rRNA genes. Nucleic Acids Res . 34, W394– W399. Google Scholar CrossRef Search ADS PubMed  de Jonge L. P., Douma R. D., Heijnen J. J., van Gulik W. M. ( 2012). Optimization of cold methanol quenching for quantitative metabolomics of Penicillium chrysogenum. Metabolomics  8, 727– 735. Google Scholar CrossRef Search ADS PubMed  Druart C., Neyrinck A. M., Vlaeminck B., Fievez V., Cani P. D., Delzenne N. M. ( 2014). Role of the lower and upper intestine in the production and absorption of gut microbiota-derived PUFA metabolites. PLoS One  9, e87560. Google Scholar CrossRef Search ADS PubMed  Duncan S. H., Hold G. L., Barcenilla A., Stewart C. S., Flint H. J. ( 2002). Roseburia intestinalis sp. nov., a novel saccharolytic, butyrate-producing bacterium from human faeces. Int. J. Syst. Evol. Microbiol . 52, 1615– 1620. Google Scholar PubMed  Flint H. J., Scott K. P., Louis P., Duncan S. H. ( 2012). The role of the gut microbiota in nutrition and health. Nat. Rev. Gastroenterol. Hepatol . 9, 577– 589. Google Scholar CrossRef Search ADS PubMed  Galisteo M., Duarte J., Zarzuelo A. ( 2008). Effects of dietary fibers on disturbances clustered in the metabolic syndrome. J. Nutr. Biochem . 19, 71– 84. Google Scholar CrossRef Search ADS PubMed  Groopman J. D., Jiaqi Z., Donahue P. R., Pikul A., Lisheng Z., Jun-shi C., Wogan G. N. ( 1992). Molecular dosimetry of urinary aflatoxin-DNA adducts in people living in Guangxi Autonomous Region, People's Republic of China. Cancer Res . 52, 45– 52. Google Scholar PubMed  Gwirtz J. A., Garcia-Casal M. N. ( 2014). Processing maize flour and corn meal food products. Ann. N. Y. Acad. Sci . 1312, 66– 75. Google Scholar CrossRef Search ADS PubMed  Haskard C. A., El-Nezami H. S., Kankaanpaa P. E., Salminen S., Ahokas J. T. ( 2001). Surface binding of aflatoxin B1 by lactic acid bacteria. Appl. Environ. Microbiol . 67, 3086– 3091. Google Scholar CrossRef Search ADS PubMed  Han J., Gagnon S., Eckle T., Borchers C. H. ( 2013a). Metabolomic analysis of key central carbon metabolism carboxylic acids as their 3-nitrophenylhydrazones by UPLC/ESI-MS. Electrophoresis  34, 2891– 2900. Han J., Gagnon S., Eckle T., Borchers C. H. ( 2013b). Metabolomic analysis of key central carbon metabolism carboxylic acids as their 3‐nitrophenylhydrazones by UPLC/ESI‐MS. Electrophoresis  34, 2891– 2900. Henry S. H., Bosch F. X., Troxell T. C., Bolger P. M. ( 1999). Reducing liver cancer–global control of aflatoxin. Science  286, 2453– 2454. Google Scholar CrossRef Search ADS PubMed  Hernández Bort J. A., Shanmukam V., Pabst M., Windwarder M., Neumann L., Alchalabi A., Krebiehl G., Koellensperger G., Hann S., Sonntag D.et al.  , . ( 2014). Reduced quenching and extraction time for mammalian cells using filtration and syringe extraction. J. Biotechnol . 182-183, 97– 103. Google Scholar CrossRef Search ADS PubMed  Holmes E., Li J. V., Marchesi J. R., Nicholson J. K. ( 2012). Gut microbiota composition and activity in relation to host metabolic phenotype and disease risk. Cell Metab . 16, 559– 564. Google Scholar CrossRef Search ADS PubMed  Junichi G., Masatoshi H., Hiroaki K., Toshio N. ( 1978). A new method for simultaneous determination of bile acids in human bile without hydrolysis. Clin. Chim. Acta  87, 141– 147. Google Scholar CrossRef Search ADS PubMed  Kew M. C. ( 2013). Aflatoxins as a cause of hepatocellular carcinoma. J. Gastrointestin. Liver Dis . 22, 305– 310. Google Scholar PubMed  Kim M., Qie Y., Park J., Kim C. H. ( 2016). Gut microbial metabolites fuel host antibody responses. Cell Host Microbe  20, 202– 214. Google Scholar CrossRef Search ADS PubMed  Kleessen B., Stoof G., Proll J., Schmiedl D., Noack J., Blaut M. ( 1997). Feeding resistant starch affects fecal and cecal microflora and short-chain fatty acids in rats. J. Anim. Sci . 75, 2453– 2462. Google Scholar CrossRef Search ADS PubMed  Klindworth A., Pruesse E., Schweer T., Peplies J., Quast C., Horn M., Glöckner F. O. ( 2013). Evaluation of general 16S ribosomal RNA gene PCR primers for classical and next-generation sequencing-based diversity studies. Nucleic Acids Res . 41, e1– e1. Google Scholar CrossRef Search ADS PubMed  Kootte R. S., Vrieze A., Holleman F., Dallinga-Thie G. M., Zoetendal E. G., de Vos W. M., Groen A. K., Hoekstra J. B., Stroes E. S., Nieuwdorp M. ( 2012). The therapeutic potential of manipulating gut microbiota in obesity and type 2 diabetes mellitus. Diabetes Obes. Metab . 14, 112– 120. Google Scholar CrossRef Search ADS PubMed  Kumar P., Mahato D. K., Kamle M., Mohanta T. K., Kang S. G. ( 2016). Aflatoxins: A Global Concern for Food Safety, Human Health and Their Management. Front. Microbiol . 7, 2170. Google Scholar PubMed  Labbé A., Ganopolsky J. G., Martoni C. J., Prakash S., Jones M. L. ( 2014). Bacterial bile metabolising gene abundance in Crohn’s, ulcerative colitis and type 2 diabetes metagenomes. PLoS One  9, e115175. Google Scholar CrossRef Search ADS PubMed  Lee W.-J., Hase K. ( 2014). Gut microbiota-generated metabolites in animal health and disease. Nat. Chem. Biol . 10, 416– 424. Google Scholar CrossRef Search ADS PubMed  Lefebvre P., Cariou B., Lien F., Kuipers F., Staels B. ( 2009). Role of bile acids and bile acid receptors in metabolic regulation. Physiol. Rev . 89, 147– 191. Google Scholar CrossRef Search ADS PubMed  Lopez V. M., Decatur C. L., Stamer W. D., Lynch R. M., McKay B. S. ( 2008). L-DOPA Is an Endogenous Ligand for OA1. PLoS Biol . 6, e236. Google Scholar CrossRef Search ADS PubMed  Louis P., Hold G. L., Flint H. J. ( 2014). The gut microbiota, bacterial metabolites and colorectal cancer. Nat. Rev. Microbiol . 12, 661– 672. Google Scholar CrossRef Search ADS PubMed  Mitchell N. J., Xue K. S., Lin S., Marroquin-Cardona A., Brown K. A., Elmore S. E., Tang L., Romoser A., Gelderblom W. C. A., Wang J.-S.et al.  , . ( 2014). Calcium montmorillonite clay reduces AFB1 and FB1 biomarkers in rats exposed to single and co‐exposures of aflatoxin and fumonisin. J. Appl. Toxicol . 34, 795– 804. Google Scholar CrossRef Search ADS PubMed  Miwa H., Hiyama C., Yamamoto M. ( 1985). High-performance liquid chromatography of short-and long-chain fatty acids as 2-nitrophenylhydrazides. J. Chromatogr. A  321, 165– 174. Google Scholar CrossRef Search ADS   Mohammadagheri N., Najafi R., Najafi G. ( 2016). Effects of dietary supplementation of organic acids and phytase on performance and intestinal histomorphology of broilers. Vet. Res. Forum  7, 189– 195. Google Scholar PubMed  Morrison D. J., Preston T. ( 2016). Formation of short chain fatty acids by the gut microbiota and their impact on human metabolism. Gut Microbes  7, 189– 200. Google Scholar CrossRef Search ADS PubMed  Mouzaki M., Wang A. Y., Bandsma R., Comelli E. M., Arendt B. M., Zhang L., Fung S., Fischer S. E., McGilvray I. G., Allard J. P. ( 2016). Bile acids and dysbiosis in non-alcoholic fatty liver disease. PLoS One  11, e0151829. Google Scholar CrossRef Search ADS PubMed  Nair A. B., Jacob S. ( 2016). A simple practice guide for dose conversion between animals and human. J. Basic. Clin. Pharm . 7, 27– 31. Google Scholar CrossRef Search ADS PubMed  Natarajan N., Pluznick J. L. ( 2014). From microbe to man: The role of microbial short chain fatty acid metabolites in host cell biology. Am. J. Physiol. Cell. Physiol . 307, C979– C985. Google Scholar CrossRef Search ADS PubMed  Pessione E. ( 2012). Lactic acid bacteria contribution to gut microbiota complexity: Lights and shadows. Front. Cell. Infect. Microbiol . 2, 86– 100. Google Scholar CrossRef Search ADS PubMed  Peters R., Hellenbrand J., Mengerink Y., Van der Wal S. ( 2004). On-line determination of carboxylic acids, aldehydes and ketones by high-performance liquid chromatography-diode array detection-atmospheric pressure chemical ionisation mass spectrometry after derivatization with 2-nitrophenylhydrazine. J. Chromatogr. A  1031, 35– 50. Google Scholar CrossRef Search ADS PubMed  Qian G., Tang L., Guo X., Wang F., Massey M. E., Su J., Guo T. L., Williams J. H., Phillips T. D., Wang J. S. ( 2014). Aflatoxin B1 modulates the expression of phenotypic markers and cytokines by splenic lymphocytes of male F344 rats. J. Appl. Toxicol . 34, 241– 249. Google Scholar CrossRef Search ADS PubMed  Qian G., Tang L., Lin S., Xue K. S., Mitchell N. J., Su J., Gelderblom W. C., Riley R. T., Phillips T. D., Wang J.-S. ( 2016). Sequential dietary exposure to Aflatoxin B1 and Fumonisin B1 in F344 rats increases liver preneoplastic changes indicative of a synergistic interaction. Food Chem. Toxicol . 95, 188– 195. Google Scholar CrossRef Search ADS PubMed  Qian G., Tang L., Wang F., Guo X., Massey M. E., Williams J. H., Phillips T. D., Wang J.-S. ( 2013a). Physiologically based toxicokinetics of serum Aflatoxin B1-lysine adduct in F344 rats. Toxicology  303, 147– 151. Google Scholar CrossRef Search ADS   Qian G., Wang F., Tang L., Massey M. E., Mitchell N. J., Su J., Williams J. H., Phillips T. D., Wang J.-S. ( 2013b). Integrative toxicopathological evaluation of Aflatoxin B1 exposure in F344 rats. Toxicol. Pathol . 41, 1093– 1105. Google Scholar CrossRef Search ADS   Qin J., Li R., Raes J., Arumugam M., Burgdorf K. S., Manichanh C., Nielsen T., Pons N., Levenez F., Yamada T.et al.  , . ( 2010). A human gut microbial gene catalogue established by metagenomic sequencing. Nature  464, 59– 65. Google Scholar CrossRef Search ADS PubMed  Rowland I. ( 2012). Metabolism of Fats Bile Acids and Steroids. Role of the Gut Flora in Toxicity and Cancer , pp. 265. Elsevier, Amsterdam, Netherlands. Rubin B. S. ( 2011). Bisphenol A: An endocrine disruptor with widespread exposure and multiple effects. J. Steroid Biochem. Mol. Biol.  127, 27– 34. Google Scholar CrossRef Search ADS PubMed  Santaren I. D., Watkins S. M., Liese A. D., Wagenknecht L. E., Rewers M. J., Haffner S. M., Lorenzo C., Hanley A. J. ( 2014). Serum pentadecanoic acid (15: 0), a short-term marker of dairy food intake, is inversely associated with incident type 2 diabetes and its underlying disorders. Am. J. Clin. Nutr . 100, 1532– 1540. Google Scholar CrossRef Search ADS PubMed  Scheppach W. ( 1994). Effects of short chain fatty acids on gut morphology and function. Gut  35, S35– S38. Google Scholar CrossRef Search ADS PubMed  Schulberg J., De Cruz P. ( 2016). Characterisation and therapeutic manipulation of the gut microbiome in inflammatory bowel disease. Intern. Med. J  46, 266– 273. Google Scholar CrossRef Search ADS PubMed  Smith P. M., Howitt M. R., Panikov N., Michaud M., Gallini C. A., Bohlooly-y M., Glickman J. N., Garrett W. S. ( 2013). The microbial metabolites, short-chain fatty acids, regulate colonic Treg cell homeostasis. Science  341, 569– 573. Google Scholar CrossRef Search ADS PubMed  Stark A., Essigmann J., Demain A., Skopek T., Wogan G. ( 1979). Aflatoxin B1 mutagenesis, DNA binding, and adduct formation in Salmonella typhimurium. Proc. Natl. Acad. Sci. U.S.A . 76, 1343– 1347. Google Scholar CrossRef Search ADS PubMed  Suchodolski J. S., Foster M. L., Sohail M. U., Leutenegger C., Queen E. V., Steiner J. M., Marks S. L. ( 2015). The fecal microbiome in cats with diarrhea. PLoS One  10, e0127378. Google Scholar CrossRef Search ADS PubMed  Tang L., Xu L., Afriyie-Gyawu E., Liu W., Wang P., Tang Y., Wang Z., Huebner H. J., Ankrah N.-A., Ofori-Adjei D.et al.  , . ( 2009). Aflatoxin–albumin adducts and correlation with decreased serum levels of vitamins A and E in an adult Ghanaian population. Food Addit. Contam . 26, 108– 118. Google Scholar CrossRef Search ADS   Torii T., Kanemitsu K., Wada T., Itoh S., Kinugawa K., Hagiwara A. ( 2010). Measurement of short-chain fatty acids in human faeces using high-performance liquid chromatography: Specimen stability. Ann. Clin. Biochem . 47, 447– 452. Google Scholar CrossRef Search ADS PubMed  Torres O., Matute J., Gelineau-Van Waes J., Maddox J., Gregory S., Ashley-Koch A., Showker J., Voss K., Riley R. ( 2015). Human health implications from co-exposure to aflatoxins and fumonisins in maize-based foods in Latin America: Guatemala as a case study. World Mycotoxin J . 8, 143– 159. Google Scholar CrossRef Search ADS   Trenk H. L., Hartman P. A. ( 1970). Effects of moisture content and temperature on aflatoxin production in corn. Appl. Microbiol . 19, 781– 784. Google Scholar PubMed  Ursell L. K., Haiser H. J., Van Treuren W., Garg N., Reddivari L., Vanamala J., Dorrestein P. C., Turnbaugh P. J., Knight R. ( 2014). the intestinal metabolome—an intersection between microbiota and host. Gastroenterology  146, 1470– 1476. Google Scholar CrossRef Search ADS PubMed  VanHook A. M. ( 2016). Another reason to eat more fiber. Sci. Signal . 9, ec184– ec184. Google Scholar CrossRef Search ADS   Wang J.-S., Groopman J. D. ( 1999). DNA damage by mycotoxins. Mutat. Res . 424, 167– 181. Google Scholar CrossRef Search ADS PubMed  Wang J., Tang L., Glenn T. C., Wang J.-S. ( 2016). Aflatoxin B1 induced compositional changes in gut microbial communities of male F344 rats. Toxicol. Sci . 150, 54– 63. Google Scholar CrossRef Search ADS PubMed  Weir T. L., Manter D. K., Sheflin A. M., Barnett B. A., Heuberger A. L., Ryan E. P. ( 2013). Stool microbiome and metabolome differences between colorectal cancer patients and healthy adults. PLoS One  8, e70803. Google Scholar CrossRef Search ADS PubMed  Wen L., Wong F. S. ( 2017). Dietary short-chain fatty acids protect against type 1 diabetes. Nat. Immunol . 18, 484– 486. Google Scholar CrossRef Search ADS PubMed  Whelan J., Fritsche K. ( 2013). Linoleic Acid. Adv. Nutr . 4, 311– 312. Google Scholar CrossRef Search ADS PubMed  Whittaker B. L., Chipley J. R. ( 1986). Interaction of aflatoxin B1 with bacterial DNA: Possible relationship to bacteriophage induction in lysogenic Bacillus megaterium. J. Environ. Pathol. Toxicol. Oncol . 6, 295– 304. Google Scholar PubMed  Winder C. L., Dunn W. B., Schuler S., Broadhurst D., Jarvis R., Stephens G. M., Goodacre R. ( 2008). Global metabolic profiling of Escherichia coli cultures: An evaluation of methods for quenching and extraction of intracellular metabolites. Anal. Chem . 80, 2939– 2948. Google Scholar CrossRef Search ADS PubMed  Wong J. M., De Souza R., Kendall C. W., Emam A., Jenkins D. J. ( 2006). Colonic health: Fermentation and short chain fatty acids. J. Clin. Gastroenterol . 40, 235– 243. Google Scholar CrossRef Search ADS PubMed  Worley B., Powers R. ( 2016). PCA as a practical indicator of OPLS-DA model reliability. Curr. Metabolomics  4, 97– 103. Google Scholar CrossRef Search ADS PubMed  Xue K. S., Cai W., Tang L., Wang J.-S. ( 2016). Aflatoxin B 1-lysine adduct in dried blood spot samples of animals and humans. Food Chem. Toxicol . 98, 210– 219. Google Scholar CrossRef Search ADS PubMed  Yatsunenko T., Rey F. E., Manary M. J., Trehan I., Dominguez-Bello M. G., Contreras M., Magris M., Hidalgo G., Baldassano R. N., Anokhin A. P.et al.  , . ( 2012). Human gut microbiome viewed across age and geography. Nature  486, 222– 227. Google Scholar CrossRef Search ADS PubMed  Zhao G., Nyman M., Åke Jönsson J. ( 2006). Rapid determination of short‐chain fatty acids in colonic contents and faeces of humans and rats by acidified water‐extraction and direct‐injection gas chromatography. Biomed. Chromatogr . 20, 674– 682. Google Scholar CrossRef Search ADS PubMed  © The Author(s) 2018. Published by Oxford University Press on behalf of the Society of Toxicology. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Toxicological Sciences Oxford University Press

Aflatoxin B1 Disrupts Gut-Microbial Metabolisms of Short-Chain Fatty Acids, Long-Chain Fatty Acids, and Bile Acids in Male F344 Rats

Loading next page...
 
/lp/ou_press/aflatoxin-b1-disrupts-gut-microbial-metabolisms-of-short-chain-fatty-e5JVa4Q004
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of the Society of Toxicology. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com
ISSN
1096-6080
eISSN
1096-0929
D.O.I.
10.1093/toxsci/kfy102
Publisher site
See Article on Publisher Site

Abstract

Abstract In this study, male F344 rats were orally exposed to aflatoxin B1 (AFB1) at 0, 5, 25, and 75 μg/kg for 4 weeks. Rat feces were collected from 2 to 4 weeks following exposure and were assessed for gut-microbiota-dependent metabolites. Gut-microbiota-related organic acids were quantitated in the feces using 2-nitrophenylhydrazine derivatization coupled HPLC-profiling method which was validated and showed good reliability, accuracy and sensitivity. After 2-week exposure, AFB1 significantly reduced the levels of fecal short-chain fatty acids (SCFAs) with an over 70% reduction in the high-dose group (75 μg/kg). Mixed-effects model revealed an inverse correlation between AFB1 dose and fecal levels of SCFAs, but no significant time effect was found. When compared with the control, oral exposure to middle-dose AFB1 (25 μg/kg) resulted in remarkable elevations of fecal cholic acid (2.18-fold), linoleic acid (cis-9, cis-12–18:2) (11.3-fold), pentadecanoic acid (15: 0) (3.68-fold), pyruvic acid (4.56-fold), and 3-phenyllactic acid (3.74-fold), but deoxycholic acid level was reduced by 41% in the low-dose group (5 μg/kg). These results demonstrated the disruptions of several important gut-microbiota metabolic pathways, including the synthesis of SCFAs, pyruvic acid related pathways, metabolisms of amino acids, bile acids and long-chain fatty acids, which may further affect host digestive efficiency, energy supply, intestinal immunity, production of neurotransmitters, and enterohepatic cross-talk. Our study suggests that the impairment of gut-microbiota-dependent metabolism may contribute to pathological mechanisms of AFB1-induced adverse health effects. Aflatoxin B1, gut-microbiota, long-chain fatty acids, microbial metabolism, short-chain fatty acids Aflatoxins (AFs) are a class of food-borne mycotoxins mainly produced by Aspergillus flavus and Aspergillus parasiticus (Kumar et al., 2016). These toxigenic fungi commonly contaminate soil, and colonize on the surface of cereals, especially for maize, and groundnuts, once humidity (>17.5%) and temperature (>24°C) meet their growth needs (Trenk and Hartman, 1970). Such environmental conditions have made the tropical area more susceptible for the food contamination and human exposure to AFs, especially in the low- and middle-income developing nations (Qian et al., 2013a). Aflatoxin B1 (AFB1) is widely recognized as the most harmful AF, due to its potent toxicity, genotoxicity, and carcinogenicity as well as acute aflatoxicosis in animals and human populations (Kew, 2013; Qian et al., 2013b; Wang and Groopman, 1999). Accordingly, the detection and assessment of AFB1 contamination in human food and animal feed has been a global concern for food safety and public health (Henry et al., 1999; Torres et al., 2015). On the other hand, remarkable efforts have been made to develop novel prevention/intervention strategies against AFB1-induced adverse health effects, including liver cancer risks and growth/developmental disorders in high-risk and vulnerable populations (Mitchell et al., 2014; Xue et al., 2016). Human gastrointestinal tract harbors a complex microbiota that contains more than 100 trillion microbes with over 400 species and carries 150 times more genes than the human genome (Qin et al., 2010). The gut-microbiota constantly provides host with hundreds of micronutrients and functional metabolites, which actively participate into the host enterohepatic cross-talk, as well as the physiological regulations of many organs and systems (Ursell et al., 2014). In recent years, next-generation sequencing technologies have uncovered all kinds of intricate connections among gut-microbiota, dietary composition and host health (Chakraborty et al., 2010; Holmes et al., 2012). In this 3-way relationship, oral exposure to xenobiotics or dietary composition could lead to the alteration of gut-microbiota, and the changes of gut-microbiota may further influence host health in a significant way (Brown and Hazen, 2015). Emerging evidences have demonstrated the causative links between gut-microbial microbiome/metabolome and a series of health problems in host, eg, obesity, metabolic syndrome, nonalcoholic fatty liver disease (NAFLD), colon cancer, inflammatory bowel disease (IBD), and cardiovascular disease (Flint et al., 2012; Holmes et al., 2012; Lee and Hase, 2014; Louis et al., 2014; Ursell et al., 2014). Therapeutic manipulation of gut-microbiota has also exhibited the potential to mitigate a number of metabolic diseases such as obesity, type-2 diabetes mellitus, IBD and NAFLD, most probably by modifying gut-microbiota-dependent metabolites, which are either derived from food by gut-microbiota, or the endogenous metabolites of gut-microbes (Kootte et al., 2012; Schulberg and De Cruz, 2016). We have previously performed 16S rRNA analysis and found the compositional change of fecal microbiome in F344 rats following repeated oral exposure to AFB1 (Wang et al., 2016). Through 16S rRNA sequencing technique, notable enrichment of Clostridiales spp. and depletion of Lactobacillales spp. were found in the rat feces. In the work presented here, the potential impact of such compositional changes on host health at metabolic level was further explored by examining a group of fecal organic acids that are highly associated with gut-microbiota. The studied metabolites include acetic acid, lactic acid, propionic acid, butyric acid, valeric acid, hexanoic acid, cholic acid, deoxycholic acid, pentadecanoic acid (15: 0), 3-phenyllactic acid, pyruvic acid, and linoleic acid (cis-9, cis-12-18:2). The metabolism of these organic acids heavily depends on the metabolic pathways and community structure of gut-microbiota, and also play important roles in host physiology and global metabolic pathways. MATERIALS AND METHODS Chemicals and reagents Pyridine, 2-nitrophenylhydrazine (2-NPH), N-(3-dimethylaminopropyl)-N’-ethylcarbodiimide hydrochloride (EDC), 2-ethylbutyric acid, acetic acid, propionic acid, butyric acid, valeric acid, hexanoic acid, cholic acid, pentadecanoic acid, 3-phenyllactic acid, pyruvic acid, linoleic acid, deoxycholic acid, bisphenol A, hippuric acid, heptadecanoic acid, AFB1, and dimethyl sulfoxide (DMSO) were all purchased from Sigma-Aldrich Inc. (St Louis, Missouri). AFB1 stock solution (25 mg/ml) was prepared in DMSO and diluted to appropriate treatment concentrations upon using. All other reagents and analytical solvents, methanol, acetonitrile, and water were purchased at the highest grade commercially available from Honeywell (Morris Plains, New Jersey). Animal treatment Male Fischer 344 rats (100–120 g) were purchased from Harlan Laboratory (Indianapolis, IN, USA). The animal housing environment was under controlled light/dark cycle (12:12 h) with a temperature of 22°C ± 2°C and relative humidity of 50%–70%. Purified AIN 76A diet and tap water were maintained every day. Upon arrival, animals were allowed for one week of environmental acclimation. One hundred male F344 rats were divided into 4 groups and were gavaged with 0, 5, 25, and 75 μg AFB1/kg body weight (b.w.) per day, respectively. DMSO was used as vehicle solvent. The details of animal protocol were reported in earlier publications, together with body indexes, histopathological assessment and AFB1-Lys pharmacokinetic data (Mohammadagheri et al., 2016; Qian et al., 2013b, 2014, 2016). Briefly, animals were daily administered with AFB1 by gavage for 4 weeks. From the second week to the fourth week, rat feces were daily collected, and weekly pooled for each group. All fecal samples were stored in −80°C freezer. Animal husbandry and care, AFB1 dosing protocol, and sample collection were approved and in strict accordance with the requirements and regulations of the Institutional Animal Care and Use Committee at the University of Georgia. Sample quenching and extraction Sample extraction procedure was similar to what previously published with modifications (de Jonge et al., 2012; Hernández Bort et al., 2014). Cold methanol (−80°C)-based quenching and extraction were applied to the fecal samples for sample pretreatment. The purpose of using cold methanol was to avoid the loss of volatile compounds, and also because methanol is a solvent chemically appropriate for the reaction of 2-NPH derivatization (Peters et al., 2004; Torii et al., 2010; Winder et al., 2008). To perform sample extraction 200 mg rat feces was transferred to the Mobio PowerLyzer tube with preloaded glass beads (0.1 mm i.d.). One milliliter of cold methanol was immediately added into the tube, and fecal pellet was gently crushed using a glass pestle. After grinding, 0.5 ml cold methanol was slowly added to wash the pestle. Then the tube was capped tightly and fastened on a rotary vortex to undergo 20 min vortex at maximum level using a Vortex-Genie 2 Mixer (Scientific Industries). During vortex, sample tube was put back on ice for 2 min in every 5 min, and finally underwent centrifugation at 12 000 rpm for 10 min to spin down cellular debris. A volume of 100 µl supernatant was transferred to an Eppendorf tube, and 50 µl internal standard (2-ethylbutyric acid) stock solution was spiked into the supernatant to achieve a concentration of 1 µg/µl, which was used to compensate technical variabilities. 2-NPH derivatization To perform derivatization, 150 µl sample extract (with internal standard added) was mixed with 45 µl derivatization solution which was freshly prepared by mixing 15 µl EDC solution (0.05 g/mL H2O), 15 µl 2-NPH solution (12.5 mg/ml methanol) and 15 µl 3% pyridine in methanol (v/v). After mild vortex, the tubes were transferred to water bath at 60°C for 60 min. The tubes then were allowed to stay in room temperature for 5 min and went through brief centrifugation in order to collect the liquid left on the tube wall. All sample vials were kept in 4°C sample cooling tray and the analysis was finished within 24 h. High-performance liquid chromatography analysis An Agilent 1200 High-performance liquid chromatography (HPLC) system, consisted of a degasser, a quarterly pump, an autosampler, a diode-array detector, and a fluorescence detector, was used to perform HPLC-profiling analysis. The chromatographic separation was conducted in a Nucleosil C18 reversed-phase column (250 × 4 mm i.d.; ES industries, New Jersey) with particle size of 5 μm and pore diameter of 120 Å. The injection volume was 100 µl and flow rate was kept at 1 ml/min. Column oven temperature was set as 40°C. Mobile phase A was pH 4.5 acidified water adjusted by hydrochloric acid. Mobile phase B was acetonitrile. The gradient eluting condition was: 90% A to 80% A in 0–12 min; 80% A to 70% A in 12–20 min; 70% A to 60% A in 20–30 min; 60% A to 45% A in 30–41 min; 45% A to 10% A in 41–43 min; then keeping at 10% A in 43–58 min; finally, from 10% A to 90% A in 58–61 min for re-balance. The detection channel is 400 nm by DAD, with reference wavelength at 510 ± 60 nm. The representative chromatogram is shown in Figure 1. Lower limit of detection (LLOD), regression standard curves, as well as the other necessary quantitative parameters used for HPLC-profiling analysis are listed in Table 1. Short-chain fatty acids (SCFAs) were recovered using the recovery rates averaged from the feces spiked with SCFA standards of approximately 50%, 100%, and 200% of their levels in control group (Supplementray Table 1). The concentrations of other interested analytes were determined using the recovery rates of structurally close standards which have similar or close structure to the analytes. Specifically, the recovery rate of hippuric acid was used to recover phenyl acids (PAs); heptadecanoic acid was used to recover long-chain fatty acids (LCFAs), and bisphenol A was used to recover bile acids. Further, 2-ethylbutyric acid was used to eliminate the technical variabilities, since it has similar structure with SCFAs. Bisphenol A was used as standard to calculate recoveries for bile acid and derivatives because it is considered to have close structure with estradiol, which was used as internal standard for quantitative analysis of bile acid (Junichi et al., 1978; Rubin, 2011). And no other commercially available compound can be chromatographically separated with bile constituents for the calibration of recovery using current method. Table 1. Analytical Parameters of HPLC-Profiling Analysis Used for the Measurement of Interested Fecal Metabolites Component  Category  RT*  Detection Channel  Regression (X, AUC; Y, ng/μl)  R2  Linear Range ng/μl  LLOD ng/μl  Acetic acid  SCFA  14.9  400 nm  y = 0.0063x − 0.3726  0.9993  0.016–64.8  0.008  Propionic acid  SCFA  19.6  400 nm  y = 0.021x − 0.9273  0.999  0.07–143  0.03  Butyric acid  SCFA  25.1  400 nm  y = 0.0256x − 0.4994  0.9991  0.078–79.5  0.04  Valeric acid  SCFA  31.5  400 nm  y = 0.0208x − 0.4974  0.999  0.054–56.1  0.03  Hexanoic acid  SCFA  37.6  400 nm  y = 0.0309x − 0.356  0.9994  0.074–75.6  0.04  Lactic acid  SCFA  13.8  400 nm  y = 0.0244x − 0.2351  0.9991  0.11–14.33  0.05  Pyruvic acid  Alpha-keto acid  41.3  400 nm  y = 0.0166x + 0.7136  0.9997  6.2–500  0.19  2-Ethylbutyric acid  IS for SCFA  34.2  400 nm  y = 0.1662x - 0.4527  0.9991  0.56–1138  0.28  Niacin  PA  22.1  210 nm  y = 0.0313x − 6.1768  0.9954  1–430  0.25  3-Phenyllactic acid  PA  31.2  400 nm  y = 0.1003x − 0.7134  0.9994  4.7–300  0.58  Hippuric acid  IS for PA  26.3  400 nm  y = 0.6161x + 2.8275  0.9996  4.45–570  2.25  Cholic acid  SA  45.1  400 nm  y = 0.1219x − 6.6046  0.9930  3.9–250  0.49  Deoxycholic acid  SA  47.1  400 nm  y = 0.0371x − 4.8658  0.9930  2.5–330  0.64  Cholesterol  Sterol  47.4  400 nm  y = 0.0686x − 2.4795  0.9900  1.95–125  0.98  Bisphenol A  IS for SA  35.0  210 nm  y = 0.0148x − 6.9647  0.992  0.33–685  0.17  Linoleic acid  LCFA  50.9  400 nm  y = 0.3705x − 31.314  0.9948  3.9–1000  3.9  Pentadecanoic acid  LCFA  51.2  400 nm  y = 0.0636x − 0.3641  0.9990  1.95–500  0.5  Heptadecanoic acid  IS for LCFA  54.5  400 nm  y = 0.1436x − 4.5852  0.9952  2.15–275  1.07  Component  Category  RT*  Detection Channel  Regression (X, AUC; Y, ng/μl)  R2  Linear Range ng/μl  LLOD ng/μl  Acetic acid  SCFA  14.9  400 nm  y = 0.0063x − 0.3726  0.9993  0.016–64.8  0.008  Propionic acid  SCFA  19.6  400 nm  y = 0.021x − 0.9273  0.999  0.07–143  0.03  Butyric acid  SCFA  25.1  400 nm  y = 0.0256x − 0.4994  0.9991  0.078–79.5  0.04  Valeric acid  SCFA  31.5  400 nm  y = 0.0208x − 0.4974  0.999  0.054–56.1  0.03  Hexanoic acid  SCFA  37.6  400 nm  y = 0.0309x − 0.356  0.9994  0.074–75.6  0.04  Lactic acid  SCFA  13.8  400 nm  y = 0.0244x − 0.2351  0.9991  0.11–14.33  0.05  Pyruvic acid  Alpha-keto acid  41.3  400 nm  y = 0.0166x + 0.7136  0.9997  6.2–500  0.19  2-Ethylbutyric acid  IS for SCFA  34.2  400 nm  y = 0.1662x - 0.4527  0.9991  0.56–1138  0.28  Niacin  PA  22.1  210 nm  y = 0.0313x − 6.1768  0.9954  1–430  0.25  3-Phenyllactic acid  PA  31.2  400 nm  y = 0.1003x − 0.7134  0.9994  4.7–300  0.58  Hippuric acid  IS for PA  26.3  400 nm  y = 0.6161x + 2.8275  0.9996  4.45–570  2.25  Cholic acid  SA  45.1  400 nm  y = 0.1219x − 6.6046  0.9930  3.9–250  0.49  Deoxycholic acid  SA  47.1  400 nm  y = 0.0371x − 4.8658  0.9930  2.5–330  0.64  Cholesterol  Sterol  47.4  400 nm  y = 0.0686x − 2.4795  0.9900  1.95–125  0.98  Bisphenol A  IS for SA  35.0  210 nm  y = 0.0148x − 6.9647  0.992  0.33–685  0.17  Linoleic acid  LCFA  50.9  400 nm  y = 0.3705x − 31.314  0.9948  3.9–1000  3.9  Pentadecanoic acid  LCFA  51.2  400 nm  y = 0.0636x − 0.3641  0.9990  1.95–500  0.5  Heptadecanoic acid  IS for LCFA  54.5  400 nm  y = 0.1436x − 4.5852  0.9952  2.15–275  1.07  The minimum data point in the linear regression range (R2 > 0.999) was noted as LOQ. Abbreviations: IS, internal standard for quality control; R2, regression coefficient; LLOD, lower limit of detection; LCFA, long-chain fatty acid; PA, phenyl acid; RT, retention time (min) in chromatogram; SA, steroid acid; SCFA, short-chain fatty acid. The analyte level which generated a signal-to-noise (S/N) ratio of 3 was noted as the LLOD for that analyte. Niacin and cholesterol were not detected in most sample extracts. Table 1. Analytical Parameters of HPLC-Profiling Analysis Used for the Measurement of Interested Fecal Metabolites Component  Category  RT*  Detection Channel  Regression (X, AUC; Y, ng/μl)  R2  Linear Range ng/μl  LLOD ng/μl  Acetic acid  SCFA  14.9  400 nm  y = 0.0063x − 0.3726  0.9993  0.016–64.8  0.008  Propionic acid  SCFA  19.6  400 nm  y = 0.021x − 0.9273  0.999  0.07–143  0.03  Butyric acid  SCFA  25.1  400 nm  y = 0.0256x − 0.4994  0.9991  0.078–79.5  0.04  Valeric acid  SCFA  31.5  400 nm  y = 0.0208x − 0.4974  0.999  0.054–56.1  0.03  Hexanoic acid  SCFA  37.6  400 nm  y = 0.0309x − 0.356  0.9994  0.074–75.6  0.04  Lactic acid  SCFA  13.8  400 nm  y = 0.0244x − 0.2351  0.9991  0.11–14.33  0.05  Pyruvic acid  Alpha-keto acid  41.3  400 nm  y = 0.0166x + 0.7136  0.9997  6.2–500  0.19  2-Ethylbutyric acid  IS for SCFA  34.2  400 nm  y = 0.1662x - 0.4527  0.9991  0.56–1138  0.28  Niacin  PA  22.1  210 nm  y = 0.0313x − 6.1768  0.9954  1–430  0.25  3-Phenyllactic acid  PA  31.2  400 nm  y = 0.1003x − 0.7134  0.9994  4.7–300  0.58  Hippuric acid  IS for PA  26.3  400 nm  y = 0.6161x + 2.8275  0.9996  4.45–570  2.25  Cholic acid  SA  45.1  400 nm  y = 0.1219x − 6.6046  0.9930  3.9–250  0.49  Deoxycholic acid  SA  47.1  400 nm  y = 0.0371x − 4.8658  0.9930  2.5–330  0.64  Cholesterol  Sterol  47.4  400 nm  y = 0.0686x − 2.4795  0.9900  1.95–125  0.98  Bisphenol A  IS for SA  35.0  210 nm  y = 0.0148x − 6.9647  0.992  0.33–685  0.17  Linoleic acid  LCFA  50.9  400 nm  y = 0.3705x − 31.314  0.9948  3.9–1000  3.9  Pentadecanoic acid  LCFA  51.2  400 nm  y = 0.0636x − 0.3641  0.9990  1.95–500  0.5  Heptadecanoic acid  IS for LCFA  54.5  400 nm  y = 0.1436x − 4.5852  0.9952  2.15–275  1.07  Component  Category  RT*  Detection Channel  Regression (X, AUC; Y, ng/μl)  R2  Linear Range ng/μl  LLOD ng/μl  Acetic acid  SCFA  14.9  400 nm  y = 0.0063x − 0.3726  0.9993  0.016–64.8  0.008  Propionic acid  SCFA  19.6  400 nm  y = 0.021x − 0.9273  0.999  0.07–143  0.03  Butyric acid  SCFA  25.1  400 nm  y = 0.0256x − 0.4994  0.9991  0.078–79.5  0.04  Valeric acid  SCFA  31.5  400 nm  y = 0.0208x − 0.4974  0.999  0.054–56.1  0.03  Hexanoic acid  SCFA  37.6  400 nm  y = 0.0309x − 0.356  0.9994  0.074–75.6  0.04  Lactic acid  SCFA  13.8  400 nm  y = 0.0244x − 0.2351  0.9991  0.11–14.33  0.05  Pyruvic acid  Alpha-keto acid  41.3  400 nm  y = 0.0166x + 0.7136  0.9997  6.2–500  0.19  2-Ethylbutyric acid  IS for SCFA  34.2  400 nm  y = 0.1662x - 0.4527  0.9991  0.56–1138  0.28  Niacin  PA  22.1  210 nm  y = 0.0313x − 6.1768  0.9954  1–430  0.25  3-Phenyllactic acid  PA  31.2  400 nm  y = 0.1003x − 0.7134  0.9994  4.7–300  0.58  Hippuric acid  IS for PA  26.3  400 nm  y = 0.6161x + 2.8275  0.9996  4.45–570  2.25  Cholic acid  SA  45.1  400 nm  y = 0.1219x − 6.6046  0.9930  3.9–250  0.49  Deoxycholic acid  SA  47.1  400 nm  y = 0.0371x − 4.8658  0.9930  2.5–330  0.64  Cholesterol  Sterol  47.4  400 nm  y = 0.0686x − 2.4795  0.9900  1.95–125  0.98  Bisphenol A  IS for SA  35.0  210 nm  y = 0.0148x − 6.9647  0.992  0.33–685  0.17  Linoleic acid  LCFA  50.9  400 nm  y = 0.3705x − 31.314  0.9948  3.9–1000  3.9  Pentadecanoic acid  LCFA  51.2  400 nm  y = 0.0636x − 0.3641  0.9990  1.95–500  0.5  Heptadecanoic acid  IS for LCFA  54.5  400 nm  y = 0.1436x − 4.5852  0.9952  2.15–275  1.07  The minimum data point in the linear regression range (R2 > 0.999) was noted as LOQ. Abbreviations: IS, internal standard for quality control; R2, regression coefficient; LLOD, lower limit of detection; LCFA, long-chain fatty acid; PA, phenyl acid; RT, retention time (min) in chromatogram; SA, steroid acid; SCFA, short-chain fatty acid. The analyte level which generated a signal-to-noise (S/N) ratio of 3 was noted as the LLOD for that analyte. Niacin and cholesterol were not detected in most sample extracts. Figure 1. View largeDownload slide HPLC-profiling chromatograms of fecal extracts from control (upper) and exposure group (lower) after 2-NPH derivatization. 2-ethyl butyric acid was used as internal standard (IS). The detection channel of DAD is 400 nm with a reference channel as 510 ± 60 nm. Down-regulated organic acids are labeled on the upper panel, whereas up-regulated organic acids are labeled on the lower panel. Specific retention time and relevant information are available in Table 1. Figure 1. View largeDownload slide HPLC-profiling chromatograms of fecal extracts from control (upper) and exposure group (lower) after 2-NPH derivatization. 2-ethyl butyric acid was used as internal standard (IS). The detection channel of DAD is 400 nm with a reference channel as 510 ± 60 nm. Down-regulated organic acids are labeled on the upper panel, whereas up-regulated organic acids are labeled on the lower panel. Specific retention time and relevant information are available in Table 1. Method validation and optimization Methanol blanks were spiked with SCFA standards to generate test solutions with concentrations of approximately 50%, 100%, and 200% of the actual SCFA amounts measured in the sample extracts. The test solutions were derivatized using 2-NPH and EDC and were immediately used for HPLC-profiling analysis. The analytical precision of the method was validated based on: (1) interday coefficient of variation (CV) of the peak intensities of SCFAs at 3 spike levels in 3 consecutive days, with one bunch performed per day; (2) inter-assay CV of the peak intensities of SCFAs at 3 spike levels in 7 consecutive assays; (3) intra-assay CV of the peak intensities of SCFAs at 3 spike levels, with 4 repeats conducted at each level. Analytical accuracy was examined using recoveries with CV, and the formula to calculate recovery rate is: recovery % = (analyte amount measured in the extract of standard-spiked feces – analyte amount measured in the extract of nonspiked feces) × 100/(amount of spiked analyte) (Han et al., 2013b). 16S rRNA analysis Briefly, total fecal genomic DNA which contains 16S rRNA was extracted using QIAamp DNA stool mini kits (QIAGEN, Valencia, California). A 2-step Quadruple-index PCR method was used to prepare the 16S rRNA gene libraries according to Klindworth et al. (2013). Sequencing of these 16S rRNA fragment libraries was performed in the Georgia Genomic Facility (University of Georgia, Athens, Georgia) using the Illumina MiSeq with v2 500 cycle chemistry, resulting in paired-end 250 base reads to obtain approximately 30 000 reads per sample. The 16S rRNA fragment amplified in this study is from site 358 to 784 under Escherichia coli system of nomenclature (Klindworth et al., 2013). The raw paired-end, demultiplexed sequence read was merged using FLASH 1.2.9 in Geneious 8.1 software (Biomatters Inc, San Francisco, California). All internal tags, base spacers, and locus-specific primers of merged sequences were trimmed and sequences outranged 400–450 base-pairs were discarded. Outputs from Geneious 8.1 were quality filtered using QIIME pipeline (Quantitative Insights Into Microbial Ecology) (Caporaso et al., 2010). Representative sequences for each operational taxonomic units (OTUs) were compared with the Greengene 16S rRNA gene database 13-8 release (DeSantis et al., 2006) using uclust algorithm with the similarity threshold of 90%. The top 3 database hits that matched the above representative sequences for each OTU were selected. Statistics and software Data normality examination, homogeneity test, 1-way ANOVA, and Welch’s t test, were all performed using SPSS 22. Levene statistic was used to test homogeneity of variances and Welch-Brown-Forsythe statistic was used to test the equality of means. Tukey’s test was used for post hoc analysis in ANOVA. When data failed to follow normality of distribution, Kruskal-Wallis H test was applied to replace 1-way ANOVA. Mixed-effects model regression was performed using STATA 14.1. Pearson’s correlation analysis, and construction of heat map and hierarchical tree were performed using R. Mann-Whitney U test was used to compare the differences of fecal organic acids (except for SCFAs) when dose effect was the only factor being analyzed, with p value < .05 considered to be statistically significant. RESULTS Validation and Optimization of HPLC-Profiling Method Our initial effort was to optimize conditions for fecal sample extraction and metabolites enrichment. However, centrifugal evaporation resulted in significant loss of SCFAs (20%–50%) in the sample extracts, as found by HPLC analysis (data not shown). For this reason, sample enrichment was avoided during sample preparation. Nonetheless, interested analytes are still detected from fecal samples. In terms of precolumn derivatization and HPLC-profiling analysis, the validation work included intraassay precision, interassay precision, interday precision, and accuracy. Shown in supplementary Table 1, most values of measured metabolites showed CV < 8%. The sensitivity and LLOD were determined for all analyzed metabolites, as shown in the Table 1. Internal standards were used to confirm the precision and accuracy, and recovery rate was ranged from 33% to 74% for the all SCFA standards spiked into fecal samples of the control and AFB1-dosed rats, with CV < 5%. Using this validated method, the peak identity and concentrations of interested metabolites were further determined from the chromatogram of fecal extracts, as shown in Figure 1. Four categories of metabolites were measured in the study: SCFAs, including acetic acid, butyric acid, hexanoic acid, lactic acid, propionic acid, and valeric acid; LCFAs, including linoleic acid (cis-9, cis-12-18:2) and pentadecanoic acid (15:0), bile acids, including cholic acid and deoxycholic acid, and other metabolites, including 3-phenyllactic acid and pyruvic acid (Table 1). AFB1 Exposure Affects SCFA Production of Gut-Microbiota Rats were exposed to AFB1 at doses of 0, 5, 25, and 75 μg/kg b.w., which are noted as control, low-, middle- and high-dose groups in the study. As shown in Figure 2 and Supplementary Table 2, significant change of fecal SCFA levels was found in AFB1-exposed groups. The measured levels of SCFAs in the untreated control group were comparable over the time course from 2- to 4-week, but notable reduction of acetic acid, propionic acid, butyric acid, hexanoic acid, and lactic acid were detected in the rat feces of AFB1-exposed groups. In the low-dose group, fecal SCFA levels seemed to be affected by the time of exposure. The fecal levels of acetic acid, propionic acid, butyric acid, lactic acid, valeric acid and hexanoic acid were 46.6%, 39.9%, 68.4%, 79.9%, 95.3%, and 63% of the control after 2 weeks of exposure, but the percentages went to 70.7%, 77.6%, 35.1%, 34.4%, 75.6%, and 86.7% of the control after 4 weeks of exposure, indicating the time-effect of AFB1-exposure on SCFA levels in the low-dose group. The fecal levels of SCFAs in the middle- and high-dose groups were generally not affected by the exposure time, except for propionic acid in middle-dose group at 2-week, and lactic acid and hexanoic acid in high-dose group at 4-week, which showed about 50% changes of fecal levels compared with control. As shown in Supplementary Table 2, the fecal levels of 6 SCFAs in the middle-dose group, were 17.7%, 31.1%, 26.1%, 20.1%, 90.7%, and 19.9% of the control in 2-week, and were 21.6%, 15.3%, 24.6%, 17.2%, 88.9%, and 27.3% of the control in 4-week. Similarly, In the high-dose group the 6 SCFAs were 22%, 22.2%, 21.9%, 12.1%, 44.2%, and 20.5% of the control in 2-week, and were 25%, 34.4%, 17.2%, 6.8%, 42.2%, and 52% of the control in 4-week. Remarkable changes were found for fecal propionic acid level in middle-dose group, which was reduced from 31.1% of control to 15.3% of control from 2- to 4-week; valeric acid in high-dose group, which was reduced from 12.1% of control to 6.8% of control; and lactic acid in high-dose group, which was elevated from 20.5% of control to 52% of control from 2- to 4-week. Figure 2. View largeDownload slide Fecal SCFA levels of rats treated with 0, 5, 25, and 75 μg AFB1/kg b.w. X-axis indicates duration of treatment. Significance of one-way ANOVA or Kruskal-Wallis H Test is indicated by string labels: same string indicating p > .05; string with partly overlapped character(s) indicating p < .05; totally different string indicating p < .01. Error bar indicates standard deviation (n = 5). Specific data are available in Supplementary Table 2. Figure 2. View largeDownload slide Fecal SCFA levels of rats treated with 0, 5, 25, and 75 μg AFB1/kg b.w. X-axis indicates duration of treatment. Significance of one-way ANOVA or Kruskal-Wallis H Test is indicated by string labels: same string indicating p > .05; string with partly overlapped character(s) indicating p < .05; totally different string indicating p < .01. Error bar indicates standard deviation (n = 5). Specific data are available in Supplementary Table 2. To examine the AFB1 dose-, time-, and time × dose interaction effects on fecal SCFA levels, mixed-effects regression model was applied to analyze the linear correlation between the AFB1-dose/time and SCFA levels. As shown in Table 2, significant dose effect and dose × time interaction effect were found. Further, Pearson’s correlation analysis was performed to examine the possible link between the changes of SCFA levels and the community structure of gut-microbiota. The correlation results were shown in the hierarchical tree and heat map in Figure 3. Briefly, strains belonging to Firmicutes Clostridiales order were highly clustered and showed inverse correlation with the fecal levels of SCFAs following AFB1-exposure, while Lactobacillales Streptococcus and Clostridiales Roseburia, 2 SCFA-producing strains, were depleted in the feces. All 6 SCFAs are correlated in the same cluster of Pearson’s r distance. Table 2. Mixed-Effects Model Analysis Between AFB1-Treatment (Dose, Time and Interaction) and Fecal Levels of SCFAs   Fixed Effect   Random Effect   SCFAs  Dose  SE  p  Time  SE  p  Interactiona  SE  p  Estimate  SE  Acetic acid  −0.1103  0.025  <.001  0.1613  0.133  0.227  10.361  2.149  <.001  31.288  6.021  Propionic acid  −0.0617  0.015  <.001  0.054  0.082  0.509  6.718  1.301  <.001  11.798  2.270  Butyric acid  −0.1442  0.015  <.001  −0.0202  0.086  0.921  13.226  1.663  <.001  18.316  3.525  Valeric acid  −0.0028  0.001  <.001  −0.0007  0.004  0.863  7.900  0.977  <.001  0.026  0.005  Hexanoic acid  −0.0179  0.004  <.001  −0.0074  0.022  0.740  2.113  0.365  <.001  0.881  0.170  Lactic acid  −0.0207  0.003  <.001  0.0228  0.018  0.210  1.280  0.298  <.001  0.586  0.112    Fixed Effect   Random Effect   SCFAs  Dose  SE  p  Time  SE  p  Interactiona  SE  p  Estimate  SE  Acetic acid  −0.1103  0.025  <.001  0.1613  0.133  0.227  10.361  2.149  <.001  31.288  6.021  Propionic acid  −0.0617  0.015  <.001  0.054  0.082  0.509  6.718  1.301  <.001  11.798  2.270  Butyric acid  −0.1442  0.015  <.001  −0.0202  0.086  0.921  13.226  1.663  <.001  18.316  3.525  Valeric acid  −0.0028  0.001  <.001  −0.0007  0.004  0.863  7.900  0.977  <.001  0.026  0.005  Hexanoic acid  −0.0179  0.004  <.001  −0.0074  0.022  0.740  2.113  0.365  <.001  0.881  0.170  Lactic acid  −0.0207  0.003  <.001  0.0228  0.018  0.210  1.280  0.298  <.001  0.586  0.112  a Estimate of interaction effect resulted by both dose and treatment time on fecal SCFA levels. Table 2. Mixed-Effects Model Analysis Between AFB1-Treatment (Dose, Time and Interaction) and Fecal Levels of SCFAs   Fixed Effect   Random Effect   SCFAs  Dose  SE  p  Time  SE  p  Interactiona  SE  p  Estimate  SE  Acetic acid  −0.1103  0.025  <.001  0.1613  0.133  0.227  10.361  2.149  <.001  31.288  6.021  Propionic acid  −0.0617  0.015  <.001  0.054  0.082  0.509  6.718  1.301  <.001  11.798  2.270  Butyric acid  −0.1442  0.015  <.001  −0.0202  0.086  0.921  13.226  1.663  <.001  18.316  3.525  Valeric acid  −0.0028  0.001  <.001  −0.0007  0.004  0.863  7.900  0.977  <.001  0.026  0.005  Hexanoic acid  −0.0179  0.004  <.001  −0.0074  0.022  0.740  2.113  0.365  <.001  0.881  0.170  Lactic acid  −0.0207  0.003  <.001  0.0228  0.018  0.210  1.280  0.298  <.001  0.586  0.112    Fixed Effect   Random Effect   SCFAs  Dose  SE  p  Time  SE  p  Interactiona  SE  p  Estimate  SE  Acetic acid  −0.1103  0.025  <.001  0.1613  0.133  0.227  10.361  2.149  <.001  31.288  6.021  Propionic acid  −0.0617  0.015  <.001  0.054  0.082  0.509  6.718  1.301  <.001  11.798  2.270  Butyric acid  −0.1442  0.015  <.001  −0.0202  0.086  0.921  13.226  1.663  <.001  18.316  3.525  Valeric acid  −0.0028  0.001  <.001  −0.0007  0.004  0.863  7.900  0.977  <.001  0.026  0.005  Hexanoic acid  −0.0179  0.004  <.001  −0.0074  0.022  0.740  2.113  0.365  <.001  0.881  0.170  Lactic acid  −0.0207  0.003  <.001  0.0228  0.018  0.210  1.280  0.298  <.001  0.586  0.112  a Estimate of interaction effect resulted by both dose and treatment time on fecal SCFA levels. Figure 3. View largeDownload slide Hierarchical cluster tree and heat map to show cross correlations for SCFAs and top 18 significantly altered gut-microbial strains discovered by previous 16s rRNA data. Data were transferred to fold change of exposure group versus control. Hierarchical clusters are constructed based on Pearson’s r distance. Red-blue color bar indicates Pearson’s correlation coefficient between 2 correlated components. SCFAs are negatively correlated with the Clostridial Ruminococcaceae strains that are frequently seen in the stools from patients with Crohn’s disease and obesity. The suppressed strains belong to Lactobacillales and Clostridial Roseburia. Phylogenetic taxa information can be accessed in reference (Wang et al., 2016). (For interpretation of the reference to color in this figure legend, the reader is referred to the web version of this article.) Figure 3. View largeDownload slide Hierarchical cluster tree and heat map to show cross correlations for SCFAs and top 18 significantly altered gut-microbial strains discovered by previous 16s rRNA data. Data were transferred to fold change of exposure group versus control. Hierarchical clusters are constructed based on Pearson’s r distance. Red-blue color bar indicates Pearson’s correlation coefficient between 2 correlated components. SCFAs are negatively correlated with the Clostridial Ruminococcaceae strains that are frequently seen in the stools from patients with Crohn’s disease and obesity. The suppressed strains belong to Lactobacillales and Clostridial Roseburia. Phylogenetic taxa information can be accessed in reference (Wang et al., 2016). (For interpretation of the reference to color in this figure legend, the reader is referred to the web version of this article.) AFB1 Exposure Affects Metabolism of Other Gut-Microbiota Dependent Organic Acids We next examined the impacts of AFB1 treatment on a set of key organic acids after 4 weeks of AFB1 exposure, including cholic acid, deoxycholic acid, 3-phenyllactic acid, pyruvic acid, pentadecanoic acid (15:0), and linoleic acid (cis-9, cis-12-18:2). Oral AFB1 exposure significantly elevated fecal LCFAs (linoleic acid and pentadecanoic acid). Specifically, the level of linoleic acid was 95.51 ± 24.18 ng/mg in the control group, and increased to 1274.82 ± 363.02 ng/mg in the low-dose group and 1079.18 ± 760.29 ng/mg in the middle-dose group; the level of pentadecanoic acid in the control group was 20.26 ± 21.99 ng/mg, and increased to 64.76 ± 36.57 ng/mg in the low-dose group and 74.60 ± 53.35 ng/mg in the middle-dose group; the most significantly altered organic acid was linoleic acid, with over 10-fold increase found in low- and middle-dose groups (Figure 4). Figure 4. View largeDownload slide Fecal concentrations of cholic acid, deoxycholic acid, linoleic acid, pentadecanoic acid, pyruvic acid, and 3-phenyllactic acid measured from the experimental groups treated with 0, 5, and 25 μg AFB1/kg b.w. via HPLC-profiling analysis. Nonparametric Mann-Whitney U test was applied for all comparisons (n = 10). Box with middle vertical line represents 25%, 50%, and 75% percentile of data. Vertical lines of box plots indicate SD, multiplied with 1.5-fold coefficient in order to stretch out from box. Figure 4. View largeDownload slide Fecal concentrations of cholic acid, deoxycholic acid, linoleic acid, pentadecanoic acid, pyruvic acid, and 3-phenyllactic acid measured from the experimental groups treated with 0, 5, and 25 μg AFB1/kg b.w. via HPLC-profiling analysis. Nonparametric Mann-Whitney U test was applied for all comparisons (n = 10). Box with middle vertical line represents 25%, 50%, and 75% percentile of data. Vertical lines of box plots indicate SD, multiplied with 1.5-fold coefficient in order to stretch out from box. Oral AFB1 exposure also significantly elevated fecal levels of cholic acid, pyruvic acid, and 3-phenyllactic acid. The level of cholic acid in the control group was 56.15 ± 27.15 ng/mg, and increased to 128.46 ± 15.35 ng/mg in the low-dose group and 122.60 ± 7.32 ng/mg in the middle-dose group; the level of pyruvic acid in the control group was 38.46 ± 26.92 ng/mg, and increased to 75.57 ± 22.18 ng/mg in the low-dose group and 175.23 ± 74.98 ng/mg in the middle-dose group, and the level of 3-phenyllactic acid in the control group was 28.82 ± 9.04 ng/mg, and increased to 83.89 ± 18.10 ng/mg in the low-dose group and 107.84 ± 74.9 ng/mg in the middle-dose group, respectively. On the other hand, the level of deoxycholic acid was significantly reduced, to about the half level (5.13 ± 5.09 ng/mg) in the low-dose group from 10.18 ± 8.69 ng/mg in the control group, and completely dropped to undetectable level in the middle-dose group. DISCUSSION Results of this study clearly demonstrated that up to 2-week oral AFB1 exposure disrupted metabolism of gut microbiota-dependent organic acids, as evidenced by significant reduction in fecal level of SCFAs and deoxycholic acid, and significant increases in LCFAs and other organic acids such as pyruvic acid, 3-phenyllactic acid, and cholic acid. All these microbial metabolites play key roles in the metabolism of gut-microbiota and the maintenance of host nutrition and health. The detection of trace amounts of SCFAs in complex media, eg, bio-fluids and fecal extracts, has been reported by several studies using HPLC-profiling combined with precolumn derivatization with 2-NPH (Han et al., 2013a; Miwa et al., 1985; Peters et al., 2004), but the application of this method has not yet reported in AFB1-exposed rat models. The chemical derivatization is usually performed in mild aqueous or alcohol environment, in which carbonyl compounds (carboxylic acid, aldehyde, and ketone) bonded to 2-NPH and form hydrazides. The reaction is activated by water-soluble EDC which serves as carbodiimide crosslinker. Before in-lab analysis, method validation was conducted to confirm whether the analytical procedure is suitable and reliable for a specific analytical task (VanHook, 2016). The accuracy and reliability of analytical method were further carefully validated (Supplementary Table 1). The measured values and interclass ratio of SCFAs in our study are comparable with several other publications (Cummings et al., 1987; Torii et al., 2010; Zhao et al., 2006). In this study we found significant inhibitory effects of AFB1-exposure on synthesis of SCFAs, which has not previously reported. The decrease in SCFAs was consistent with the depletion of SCFA-producing strains such as Lactobacillales Streptococcus and Clostridiales Roseburia (Duncan et al., 2002; Kleessen et al., 1997). SCFAs are a group of beneficial aliphatic acids that are mainly produced by the anaerobic bacterial fermentation of resistant starches and insoluble fibers in the gastrointestinal tract of human and other mammals (Brockman, 2005). They are structurally constructed by 1–6 carbon atom(s), including formic acid (C1), acetic acid (C2), propionic acid (C3), butyric acid (C4), valeric acid (C5), hexanoic acid (C6), and a variety of branched-chain isomers of these acids. A variety of nutritional and physiological associations of SCFAs with liver diseases, general immunity, IBD, cardiovascular disease, and diabetes were found in many epidemiological studies and in various in vivo and in vitro models (Corrêa-Oliveira et al., 2016; Galisteo et al., 2008; Morrison and Preston, 2016; Wong et al., 2006; Zhao et al., 2006). Acetic acid, butyric acid and propionic acid can be produced by gut-microbiota via fermentation of insoluble fibers (Corrêa-Oliveira et al., 2016; Morrison and Preston, 2016; Torii et al., 2010). SCFAs were mainly produced from the fermentation process of certain strains such as Lactobacillales Streptococcus. The aflatoxin-caused reduction in these microbial strains (Wang et al., 2015) could eventually affect the fermentation process and cause reduction of SCFAs. Mixed-effects model analysis showed that—3 major SCFAs, ie, acetic acid, butyric acid and propionic acid were the most significantly affected by AFB1-dose and dose × time interaction, but not time of treatment (Table 2). It was demonstrated in our earlier 16S rRNA analysis, that the adaption of gut-microbiota community structure was featured by the elevation of relative abundances of Clostridiales spp., but decrease of Lactobacillales Streptococcus and Clostridiales Roseburia (Wang et al., 2016). Given that dose-response was also found for specific gut-microbial strains, Pearson’s correlation analysis between fecal SCFA levels and gut-microbial strains was performed to show their correlation. We found that strains from Firmicutes Clostridiales, an order associated with diarrhea in human and other mammals (Suchodolski et al., 2015), were highly clustered, and exhibited inverse correlation with SCFAs. By contrast, the relative abundances of Lactobacillales Streptococcus and Clostridiales Roseburia were positively correlated with fecal SCFAs. Both of these microbes are SCFA-producing strains (Duncan et al., 2002; Kleessen et al., 1997). The depletion of SCFAs in feces reflected the suppression of microbial fermentation on resistant starches and insoluble fibers. This may further result in a wide range of adverse consequences, because the receptors of SCFAs such as GPR43, GPR41, OLFR78, and GPR109A, are extensively distributed in different organs and systems, and are involved in a myriad of regulatory axis and pathways, such as mobility of gut epithelium, liver lipogenesis, global immunity, cell cycle, oncogenesis, apoptosis and proliferation (Brown et al., 2003; Natarajan and Pluznick, 2014; Smith et al., 2013). Moreover, dietary supply of SCFAs has recently been found to be able to protect against type-I diabetes in mice model (Wen and Wong, 2017). In addition to SCFAs, there are a great number of organic acids present in gut and feces that play important physiological roles. They are either food-derived nutrients or the metabolic products generated in gut-microbiota and host metabolisms. Interested organic acids in our study included fecal linoleic acid (cis-9, cis-12–18:2), pentadecanoic acid (15:0), pyruvic acid, 3-phenyllactic acid, cholic acid, and deoxycholic acid, which were remarkably altered in the feces following AFB1 exposure (Figure 4). Linoleic acid is an omega-6 polyunsaturated fatty acid known as an essential dietary nutrient that cannot be de novo synthesized by human body. The unsaturated fatty acids are known to carry with various health-promoting functions, such as antioxidant defense, suppression of blood levels of triglycerides and cholesterol, maintenance of glucose tolerance, and mitigation of hyperinsulinemia (Whelan and Fritsche, 2013). Most of these beneficial functions has been identified in conjugated linoleic acids, mainly as cis-9, trans-11 C18:2, trans-9, trans-11 C18:2, and trans-10, cis-12 C18:2 (Worley and Powers, 2016; Yatsunenko et al., 2012). Pentadecanoic acid is known to carry a variety of regulatory functions in cell signaling, glucose utilization, and the maintenance of the integrity and stability of gut epithelium (Santaren et al., 2014). The abnormal accumulation of linoleic acid and pentadecanoic acid in rat feces suggested a suppressed intestinal absorption of LCFAs, which is disadvantageous for host health. The deficient bioavailability may be caused by several conditions. First, the decrease of SCFAs may affect the epithelial delivery of nutrients to hepatic portal vein, since SCFAs are well-known nutrients that are able to enhance colonic blood flow and epithelial motility by providing energy and activating G-protein receptors (Scheppach, 1994). Second, certain gut-microbial strains are capable of transferring LCFAs into their conjugated forms which are easier to be absorbed (Druart et al., 2014). For example, Lactobacillus, Propionibacterium, and Bifidobacterium species can produce conjugated linoleic acid from dietary linoleic acid by using microbial lipoxygenases and cyclooxygenases—a process known to facilitate the absorption of LCFAs (Yatsunenko et al., 2012). Our previous 16S rRNA analysis demonstrated that these strains were suppressed by AFB1, which could affect the uptake and reduce bioavailability of LCFAs (Wang et al., 2016). Bile acids are endogenous steroid acids synthesized from cholesterol by liver cells of most vertebrates. Different species have distinct molecular forms of bile acids generated, but some major types of bile acids are shared by different species, eg, cholic acid and chenodeoxycholic acid in human and rat (Whittaker and Chipley, 1986). In human, bile acids are stored in the gallbladder, and are released into duodenum with bile juice under the dietary stimulation. Upon arriving small intestine, bile acids participate in the digestion and absorption of fats and fat-soluble vitamins and can be further metabolized into a variety of secondary metabolites by gut-microbiota. In this work, cholic acid and deoxycholic acid were selected as representative primary and secondary bile acids to probe the microbial metabolism of bile acids, since they are found in both human and rat feces at comparatively high levels. We found a remarkable elevation of cholic acid level with a significant reduction of deoxycholic acid level in AFB1 exposed rat feces. The significant elevation of cholic acid is generally considered to be harmful to host health. Abnormal increase of cholic acid is associated with liver pathogenesis such as cirrhosis and steatosis (Mouzaki et al., 2016), and is also known as a risk factor for intestinal inflammation (Camilleri, 2011). Besides, extra cholic acid in gut may partially contribute to the incidence of colon cancer by stimulating the growth of a small-size benign adenoma to larger size (Rowland, 2012). In correspondence with the increase of cholic acid, we found severe liver damages and pathogenesis in the AFB1-treated rats (Qian et al., 2013b, 2016). The abnormal reduction of deoxycholic acid can be attributed to the relative abundances of the deoxycholic acid-producing microbes, such as Lachospiraceae, Clostridiaceae, and Ruminococcaceae, were all decreased by AFB1 exposure (Wang et al., 2016). In both human studies and rodent models these strains can metabolize primary bile acids into secondary bile acids (Labbé et al., 2014). There are also interactions among primary bile acids, secondary bile acids, and SCFAs in regulating host health, and the elevation of intestinal primary bile acids with decreased secondary bile acid was associated with the incidences of dysbiosis and IBD in humans (Lefebvre et al., 2009). The increase of fecal cholic acid in combination with the decrease of SCFAs were previously observed in the patients with colon cancer (Weir et al., 2013). Pyruvic acid is a well-known energetic α-keto acid that is involved in a number of important metabolic pathways of both gut-microbiota and host. It serves as energy supply to cells through Krebs cycle, and can be transferred to SCFAs by Lactobacilli strains through glycolytic pathway (Pessione, 2012). Pyruvic acid can be transferred to carbohydrates via gluconeogenesis, or participate in the biosynthesis of fatty acids after binding with acetyl-CoA (Kim et al., 2016). Since pyruvic acid takes such a central role in the catabolism of carbohydrates, its unusual accumulation in rat feces reflected a suppressed energy utilization and disruption of glycolysis of gut-microbiota. This may also result in the decrease of microbial synthesis of SCFAs (VanHook, 2016). It seems that the reduction of SCFAs is not only caused by alteration of community structure of gut-microbiota, but also related with the specific metabolic pathway. Last, 3-phenyllactic acid, a central intermediate product in the upstream of phenylalanine catabolism (Stark et al., 1979), was accumulated in the rat feces following exposure to AFB1. The abnormal accumulation of 3-phenyllactic acid suggested the disruption of gut-microbial phenylalanine pathway (Camilleri, 2011). The phenylalanine pathway is known to generate L-3, 4-dihydroxyphenylalanine (L-DOPA) and tyrosine. L-DOPA is the precursor to a number of important neurotransmitters such as dopamine, norepinephrine, and epinephrine. In addition, L-DOPA itself also mediates neurotrophic factor release by the brain and central neuro system (CNS) (Lopez et al., 2008). For these reasons, the down-regulation of phenylalanine pathway may interfere with host CNS function and cause-related health problems. Dietary AFB1 exposure and AFB1-induced adverse health effects remain a major public health problem in many tropical developing nations. The range of dosage used in this study (5–75 µg/kg b.w.) was relevant to human exposure, based on 300 g corn consumption per day (Gwirtz and Garcia-Casal, 2014) and oral exposure levels ranged from 100 to 1000 μg/kg corn for high-risk human populations in Kenya, Ghana, and Guangxi area of China (Azziz-Baumgartner et al., 2005; Groopman et al., 1992; Tang et al., 2009). The dose was multiplied by an adjusting factor of 6.2 in order to transfer human exposure to that in rats (Nair and Jacob, 2016). Regarding the mechanisms behind the metabolite alterations found in this study, there are several mechanisms involved: (1) AFB1, as a natural antimicrobial agent, can selectively inhibit certain bacterial strains and influence on the growth of other strains (Arai et al., 1967; Haskard et al., 2001), as shown in the compositional changes of gut-microbiota revealed by 16 s rRNA analysis; (2) AFB1, as a potent hepatic toxin, can damage liver—the major metabolic organ and in turn induce the metabolic changes for the supply of nutrients and metabolites to host cells and tissues, including gut cells, which may play an important role in the metabolism of gut-microbiota (Atroshi et al., 1998). However, the more specific mechanism related to how AFB1 induces changes of gut-microbiota community structure and the dependent metabolites still need to be clarified in future study. Taken together, as summarized in Figure 5 based on our previous studies (Qian et al., 2013a,b,, 2014; Wang et al., 2016), oral exposure to AFB1 in rat results in significant toxic effects, biochemical alterations, and induction of preneoplastic GST-P positive liver foci. With same study design, here we show that AFB1 can induce the adverse change of community structure of gut-microbiota and significant disruption of multiple metabolic pathways, such as production of SCFAs, secretion, and metabolism of bile acids, absorption of LCFAs, catabolism of phenylalanine, and metabolism of pyruvic acid. These pathways take central and key positions in the global metabolism of gut-microbiota and maintenance of host health, for examples, energy-delivery pathways related with pyruvic acid, including gluconeogenesis, fatty acid synthesis, Krebs cycle and production of lactic acid. Therefore, our data suggest that gut-microbiota may partially be involved in the pathological mechanism and progressions of AFB1-exposure induced adverse health outcomes in F344 rat model, and presumably also in humans. Figure 5. View largeDownload slide Summary of the adverse health outcomes associated with dietary exposure to AFB1 in F344 rat model. Gray arrow indicates the changing trends of microbial taxa, biomarkers, phenotypes, and metabolites induced by AFB1-treatment. The establishment of rat model for AFB1 oral exposure, as well as the 16s rRNA analysis have been published already (Qian et al., 2013a,b, 2014; Wang et al., 2016). Briefly, male F344 rats were gavaged with AFB1 at doses of 0, 5, 10, 25, 50, and 75 μg/kg b.w. per day. The major pathological changes are summarized on the left panel. After 3 weeks of exposure to 75 μg AFB1/kg b.w., bile duct proliferation, liver GST-P+ foci co-occurred, followed by proliferation foci formation after 4 weeks and dramatic alanine transaminase, aspartate transaminase and creatine kinase elevations after 5 weeks of treatment. Figure 5. View largeDownload slide Summary of the adverse health outcomes associated with dietary exposure to AFB1 in F344 rat model. Gray arrow indicates the changing trends of microbial taxa, biomarkers, phenotypes, and metabolites induced by AFB1-treatment. The establishment of rat model for AFB1 oral exposure, as well as the 16s rRNA analysis have been published already (Qian et al., 2013a,b, 2014; Wang et al., 2016). Briefly, male F344 rats were gavaged with AFB1 at doses of 0, 5, 10, 25, 50, and 75 μg/kg b.w. per day. The major pathological changes are summarized on the left panel. After 3 weeks of exposure to 75 μg AFB1/kg b.w., bile duct proliferation, liver GST-P+ foci co-occurred, followed by proliferation foci formation after 4 weeks and dramatic alanine transaminase, aspartate transaminase and creatine kinase elevations after 5 weeks of treatment. SUPPLEMENTARY DATA Supplementary data are available at Toxicological Sciences online. ACKNOWLEDGMENTS Authors thank Dr Guoqing Qian and Dr Kathy Xue for their assistance in animal experiments. Interdisciplinary Toxicology Program at the University of Georgia Graduate School provided stipend supports. FUNDING This work was supported partially by the research contract (ECG-A-00-07-00001-00), from the United States Agency for International Development via Peanut CRSP and the Center for Mycotoxin Research at the College of Public Health, University of Georgia. REFERENCES Arai T., Ito T., Koyama Y. ( 1967). Antimicrobial activity of aflatoxins. J. Bacteriol . 93, 59– 64. Google Scholar PubMed  Atroshi F., Rizzo A., Westermarck T., Ali-Vehmas T. ( 1998). Effects of tamoxifen, melatonin, coenzyme q10, andl-carnitine supplementation on bacterial growth in the presence of mycotoxins. Pharmacol. Res . 38, 289– 295. Google Scholar CrossRef Search ADS PubMed  Azziz-Baumgartner E., Lindblade K., Gieseker K., Rogers H. S., Kieszak S., Njapau H., Schleicher R., McCoy L. F., Misore A., DeCock K.et al.  , . ( 2005). Case-control study of an acute aflatoxicosis outbreak, Kenya, 2004. Environ. Health Perspect . 113, 1779– 1783. Google Scholar CrossRef Search ADS PubMed  Brockman R. ( 2005). Glucose and short-chain fatty acid metabolism. In Quantitative Aspects of Ruminant Digestion and Metabolism , pp. 291– 310 ( Dijkstra J., Forbes J. M., France J., Eds.), CAB International, Wallingford, UK. Google Scholar CrossRef Search ADS   Brown A. J., Goldsworthy S. M., Barnes A. A., Eilert M. M., Tcheang L., Daniels D., Muir A. I., Wigglesworth M. J., Kinghorn I., Fraser N. J.et al.  , . ( 2003). The Orphan G protein-coupled receptors GPR41 and GPR43 are activated by propionate and other short chain carboxylic acids. J. Biol. Chem . 278, 11312– 11319. Google Scholar CrossRef Search ADS PubMed  Brown J. M., Hazen S. L. ( 2015). The gut microbial endocrine organ: Bacterially-derived signals driving cardiometabolic diseases. Annu. Rev. Med . 66, 343. Google Scholar CrossRef Search ADS PubMed  Camilleri M. ( 2011). Managing symptoms of irritable bowel syndrome in patients with inflammatory bowel disease. Gut  60, 425– 428. Google Scholar CrossRef Search ADS PubMed  Caporaso J. G., Kuczynski J., Stombaugh J., Bittinger K., Bushman F. D., Costello E. K., Fierer N., Peña A. G., Goodrich J. K., Gordon J. I.et al.  , . ( 2010). QIIME allows analysis of high-throughput community sequencing data. Nat. Methods  7, 335. Google Scholar CrossRef Search ADS PubMed  Chakraborty S., Zawieja S., Wang W., Zawieja D. C., Muthuchamy M. ( 2010). Lymphatic system: A vital link between metabolic syndrome and inflammation. Ann. N.Y. Acad. Sci . 1207, E94– 102. Google Scholar CrossRef Search ADS PubMed  Corrêa-Oliveira R., Fachi J. L., Vieira A., Sato F. T., Vinolo M. A. R. ( 2016). Regulation of immune cell function by short-chain fatty acids. Clin. Transl. Immunol.  5, e73– e80. Google Scholar CrossRef Search ADS   Cummings J., Pomare E., Branch W., Naylor C., Macfarlane G. ( 1987). Short chain fatty acids in human large intestine, portal, hepatic and venous blood. Gut  28, 1221– 1227. Google Scholar CrossRef Search ADS PubMed  DeSantis T., Hugenholtz P., Keller K., Brodie E., Larsen N., Piceno Y., Phan R., Andersen G. L. ( 2006). NAST: A multiple sequence alignment server for comparative analysis of 16S rRNA genes. Nucleic Acids Res . 34, W394– W399. Google Scholar CrossRef Search ADS PubMed  de Jonge L. P., Douma R. D., Heijnen J. J., van Gulik W. M. ( 2012). Optimization of cold methanol quenching for quantitative metabolomics of Penicillium chrysogenum. Metabolomics  8, 727– 735. Google Scholar CrossRef Search ADS PubMed  Druart C., Neyrinck A. M., Vlaeminck B., Fievez V., Cani P. D., Delzenne N. M. ( 2014). Role of the lower and upper intestine in the production and absorption of gut microbiota-derived PUFA metabolites. PLoS One  9, e87560. Google Scholar CrossRef Search ADS PubMed  Duncan S. H., Hold G. L., Barcenilla A., Stewart C. S., Flint H. J. ( 2002). Roseburia intestinalis sp. nov., a novel saccharolytic, butyrate-producing bacterium from human faeces. Int. J. Syst. Evol. Microbiol . 52, 1615– 1620. Google Scholar PubMed  Flint H. J., Scott K. P., Louis P., Duncan S. H. ( 2012). The role of the gut microbiota in nutrition and health. Nat. Rev. Gastroenterol. Hepatol . 9, 577– 589. Google Scholar CrossRef Search ADS PubMed  Galisteo M., Duarte J., Zarzuelo A. ( 2008). Effects of dietary fibers on disturbances clustered in the metabolic syndrome. J. Nutr. Biochem . 19, 71– 84. Google Scholar CrossRef Search ADS PubMed  Groopman J. D., Jiaqi Z., Donahue P. R., Pikul A., Lisheng Z., Jun-shi C., Wogan G. N. ( 1992). Molecular dosimetry of urinary aflatoxin-DNA adducts in people living in Guangxi Autonomous Region, People's Republic of China. Cancer Res . 52, 45– 52. Google Scholar PubMed  Gwirtz J. A., Garcia-Casal M. N. ( 2014). Processing maize flour and corn meal food products. Ann. N. Y. Acad. Sci . 1312, 66– 75. Google Scholar CrossRef Search ADS PubMed  Haskard C. A., El-Nezami H. S., Kankaanpaa P. E., Salminen S., Ahokas J. T. ( 2001). Surface binding of aflatoxin B1 by lactic acid bacteria. Appl. Environ. Microbiol . 67, 3086– 3091. Google Scholar CrossRef Search ADS PubMed  Han J., Gagnon S., Eckle T., Borchers C. H. ( 2013a). Metabolomic analysis of key central carbon metabolism carboxylic acids as their 3-nitrophenylhydrazones by UPLC/ESI-MS. Electrophoresis  34, 2891– 2900. Han J., Gagnon S., Eckle T., Borchers C. H. ( 2013b). Metabolomic analysis of key central carbon metabolism carboxylic acids as their 3‐nitrophenylhydrazones by UPLC/ESI‐MS. Electrophoresis  34, 2891– 2900. Henry S. H., Bosch F. X., Troxell T. C., Bolger P. M. ( 1999). Reducing liver cancer–global control of aflatoxin. Science  286, 2453– 2454. Google Scholar CrossRef Search ADS PubMed  Hernández Bort J. A., Shanmukam V., Pabst M., Windwarder M., Neumann L., Alchalabi A., Krebiehl G., Koellensperger G., Hann S., Sonntag D.et al.  , . ( 2014). Reduced quenching and extraction time for mammalian cells using filtration and syringe extraction. J. Biotechnol . 182-183, 97– 103. Google Scholar CrossRef Search ADS PubMed  Holmes E., Li J. V., Marchesi J. R., Nicholson J. K. ( 2012). Gut microbiota composition and activity in relation to host metabolic phenotype and disease risk. Cell Metab . 16, 559– 564. Google Scholar CrossRef Search ADS PubMed  Junichi G., Masatoshi H., Hiroaki K., Toshio N. ( 1978). A new method for simultaneous determination of bile acids in human bile without hydrolysis. Clin. Chim. Acta  87, 141– 147. Google Scholar CrossRef Search ADS PubMed  Kew M. C. ( 2013). Aflatoxins as a cause of hepatocellular carcinoma. J. Gastrointestin. Liver Dis . 22, 305– 310. Google Scholar PubMed  Kim M., Qie Y., Park J., Kim C. H. ( 2016). Gut microbial metabolites fuel host antibody responses. Cell Host Microbe  20, 202– 214. Google Scholar CrossRef Search ADS PubMed  Kleessen B., Stoof G., Proll J., Schmiedl D., Noack J., Blaut M. ( 1997). Feeding resistant starch affects fecal and cecal microflora and short-chain fatty acids in rats. J. Anim. Sci . 75, 2453– 2462. Google Scholar CrossRef Search ADS PubMed  Klindworth A., Pruesse E., Schweer T., Peplies J., Quast C., Horn M., Glöckner F. O. ( 2013). Evaluation of general 16S ribosomal RNA gene PCR primers for classical and next-generation sequencing-based diversity studies. Nucleic Acids Res . 41, e1– e1. Google Scholar CrossRef Search ADS PubMed  Kootte R. S., Vrieze A., Holleman F., Dallinga-Thie G. M., Zoetendal E. G., de Vos W. M., Groen A. K., Hoekstra J. B., Stroes E. S., Nieuwdorp M. ( 2012). The therapeutic potential of manipulating gut microbiota in obesity and type 2 diabetes mellitus. Diabetes Obes. Metab . 14, 112– 120. Google Scholar CrossRef Search ADS PubMed  Kumar P., Mahato D. K., Kamle M., Mohanta T. K., Kang S. G. ( 2016). Aflatoxins: A Global Concern for Food Safety, Human Health and Their Management. Front. Microbiol . 7, 2170. Google Scholar PubMed  Labbé A., Ganopolsky J. G., Martoni C. J., Prakash S., Jones M. L. ( 2014). Bacterial bile metabolising gene abundance in Crohn’s, ulcerative colitis and type 2 diabetes metagenomes. PLoS One  9, e115175. Google Scholar CrossRef Search ADS PubMed  Lee W.-J., Hase K. ( 2014). Gut microbiota-generated metabolites in animal health and disease. Nat. Chem. Biol . 10, 416– 424. Google Scholar CrossRef Search ADS PubMed  Lefebvre P., Cariou B., Lien F., Kuipers F., Staels B. ( 2009). Role of bile acids and bile acid receptors in metabolic regulation. Physiol. Rev . 89, 147– 191. Google Scholar CrossRef Search ADS PubMed  Lopez V. M., Decatur C. L., Stamer W. D., Lynch R. M., McKay B. S. ( 2008). L-DOPA Is an Endogenous Ligand for OA1. PLoS Biol . 6, e236. Google Scholar CrossRef Search ADS PubMed  Louis P., Hold G. L., Flint H. J. ( 2014). The gut microbiota, bacterial metabolites and colorectal cancer. Nat. Rev. Microbiol . 12, 661– 672. Google Scholar CrossRef Search ADS PubMed  Mitchell N. J., Xue K. S., Lin S., Marroquin-Cardona A., Brown K. A., Elmore S. E., Tang L., Romoser A., Gelderblom W. C. A., Wang J.-S.et al.  , . ( 2014). Calcium montmorillonite clay reduces AFB1 and FB1 biomarkers in rats exposed to single and co‐exposures of aflatoxin and fumonisin. J. Appl. Toxicol . 34, 795– 804. Google Scholar CrossRef Search ADS PubMed  Miwa H., Hiyama C., Yamamoto M. ( 1985). High-performance liquid chromatography of short-and long-chain fatty acids as 2-nitrophenylhydrazides. J. Chromatogr. A  321, 165– 174. Google Scholar CrossRef Search ADS   Mohammadagheri N., Najafi R., Najafi G. ( 2016). Effects of dietary supplementation of organic acids and phytase on performance and intestinal histomorphology of broilers. Vet. Res. Forum  7, 189– 195. Google Scholar PubMed  Morrison D. J., Preston T. ( 2016). Formation of short chain fatty acids by the gut microbiota and their impact on human metabolism. Gut Microbes  7, 189– 200. Google Scholar CrossRef Search ADS PubMed  Mouzaki M., Wang A. Y., Bandsma R., Comelli E. M., Arendt B. M., Zhang L., Fung S., Fischer S. E., McGilvray I. G., Allard J. P. ( 2016). Bile acids and dysbiosis in non-alcoholic fatty liver disease. PLoS One  11, e0151829. Google Scholar CrossRef Search ADS PubMed  Nair A. B., Jacob S. ( 2016). A simple practice guide for dose conversion between animals and human. J. Basic. Clin. Pharm . 7, 27– 31. Google Scholar CrossRef Search ADS PubMed  Natarajan N., Pluznick J. L. ( 2014). From microbe to man: The role of microbial short chain fatty acid metabolites in host cell biology. Am. J. Physiol. Cell. Physiol . 307, C979– C985. Google Scholar CrossRef Search ADS PubMed  Pessione E. ( 2012). Lactic acid bacteria contribution to gut microbiota complexity: Lights and shadows. Front. Cell. Infect. Microbiol . 2, 86– 100. Google Scholar CrossRef Search ADS PubMed  Peters R., Hellenbrand J., Mengerink Y., Van der Wal S. ( 2004). On-line determination of carboxylic acids, aldehydes and ketones by high-performance liquid chromatography-diode array detection-atmospheric pressure chemical ionisation mass spectrometry after derivatization with 2-nitrophenylhydrazine. J. Chromatogr. A  1031, 35– 50. Google Scholar CrossRef Search ADS PubMed  Qian G., Tang L., Guo X., Wang F., Massey M. E., Su J., Guo T. L., Williams J. H., Phillips T. D., Wang J. S. ( 2014). Aflatoxin B1 modulates the expression of phenotypic markers and cytokines by splenic lymphocytes of male F344 rats. J. Appl. Toxicol . 34, 241– 249. Google Scholar CrossRef Search ADS PubMed  Qian G., Tang L., Lin S., Xue K. S., Mitchell N. J., Su J., Gelderblom W. C., Riley R. T., Phillips T. D., Wang J.-S. ( 2016). Sequential dietary exposure to Aflatoxin B1 and Fumonisin B1 in F344 rats increases liver preneoplastic changes indicative of a synergistic interaction. Food Chem. Toxicol . 95, 188– 195. Google Scholar CrossRef Search ADS PubMed  Qian G., Tang L., Wang F., Guo X., Massey M. E., Williams J. H., Phillips T. D., Wang J.-S. ( 2013a). Physiologically based toxicokinetics of serum Aflatoxin B1-lysine adduct in F344 rats. Toxicology  303, 147– 151. Google Scholar CrossRef Search ADS   Qian G., Wang F., Tang L., Massey M. E., Mitchell N. J., Su J., Williams J. H., Phillips T. D., Wang J.-S. ( 2013b). Integrative toxicopathological evaluation of Aflatoxin B1 exposure in F344 rats. Toxicol. Pathol . 41, 1093– 1105. Google Scholar CrossRef Search ADS   Qin J., Li R., Raes J., Arumugam M., Burgdorf K. S., Manichanh C., Nielsen T., Pons N., Levenez F., Yamada T.et al.  , . ( 2010). A human gut microbial gene catalogue established by metagenomic sequencing. Nature  464, 59– 65. Google Scholar CrossRef Search ADS PubMed  Rowland I. ( 2012). Metabolism of Fats Bile Acids and Steroids. Role of the Gut Flora in Toxicity and Cancer , pp. 265. Elsevier, Amsterdam, Netherlands. Rubin B. S. ( 2011). Bisphenol A: An endocrine disruptor with widespread exposure and multiple effects. J. Steroid Biochem. Mol. Biol.  127, 27– 34. Google Scholar CrossRef Search ADS PubMed  Santaren I. D., Watkins S. M., Liese A. D., Wagenknecht L. E., Rewers M. J., Haffner S. M., Lorenzo C., Hanley A. J. ( 2014). Serum pentadecanoic acid (15: 0), a short-term marker of dairy food intake, is inversely associated with incident type 2 diabetes and its underlying disorders. Am. J. Clin. Nutr . 100, 1532– 1540. Google Scholar CrossRef Search ADS PubMed  Scheppach W. ( 1994). Effects of short chain fatty acids on gut morphology and function. Gut  35, S35– S38. Google Scholar CrossRef Search ADS PubMed  Schulberg J., De Cruz P. ( 2016). Characterisation and therapeutic manipulation of the gut microbiome in inflammatory bowel disease. Intern. Med. J  46, 266– 273. Google Scholar CrossRef Search ADS PubMed  Smith P. M., Howitt M. R., Panikov N., Michaud M., Gallini C. A., Bohlooly-y M., Glickman J. N., Garrett W. S. ( 2013). The microbial metabolites, short-chain fatty acids, regulate colonic Treg cell homeostasis. Science  341, 569– 573. Google Scholar CrossRef Search ADS PubMed  Stark A., Essigmann J., Demain A., Skopek T., Wogan G. ( 1979). Aflatoxin B1 mutagenesis, DNA binding, and adduct formation in Salmonella typhimurium. Proc. Natl. Acad. Sci. U.S.A . 76, 1343– 1347. Google Scholar CrossRef Search ADS PubMed  Suchodolski J. S., Foster M. L., Sohail M. U., Leutenegger C., Queen E. V., Steiner J. M., Marks S. L. ( 2015). The fecal microbiome in cats with diarrhea. PLoS One  10, e0127378. Google Scholar CrossRef Search ADS PubMed  Tang L., Xu L., Afriyie-Gyawu E., Liu W., Wang P., Tang Y., Wang Z., Huebner H. J., Ankrah N.-A., Ofori-Adjei D.et al.  , . ( 2009). Aflatoxin–albumin adducts and correlation with decreased serum levels of vitamins A and E in an adult Ghanaian population. Food Addit. Contam . 26, 108– 118. Google Scholar CrossRef Search ADS   Torii T., Kanemitsu K., Wada T., Itoh S., Kinugawa K., Hagiwara A. ( 2010). Measurement of short-chain fatty acids in human faeces using high-performance liquid chromatography: Specimen stability. Ann. Clin. Biochem . 47, 447– 452. Google Scholar CrossRef Search ADS PubMed  Torres O., Matute J., Gelineau-Van Waes J., Maddox J., Gregory S., Ashley-Koch A., Showker J., Voss K., Riley R. ( 2015). Human health implications from co-exposure to aflatoxins and fumonisins in maize-based foods in Latin America: Guatemala as a case study. World Mycotoxin J . 8, 143– 159. Google Scholar CrossRef Search ADS   Trenk H. L., Hartman P. A. ( 1970). Effects of moisture content and temperature on aflatoxin production in corn. Appl. Microbiol . 19, 781– 784. Google Scholar PubMed  Ursell L. K., Haiser H. J., Van Treuren W., Garg N., Reddivari L., Vanamala J., Dorrestein P. C., Turnbaugh P. J., Knight R. ( 2014). the intestinal metabolome—an intersection between microbiota and host. Gastroenterology  146, 1470– 1476. Google Scholar CrossRef Search ADS PubMed  VanHook A. M. ( 2016). Another reason to eat more fiber. Sci. Signal . 9, ec184– ec184. Google Scholar CrossRef Search ADS   Wang J.-S., Groopman J. D. ( 1999). DNA damage by mycotoxins. Mutat. Res . 424, 167– 181. Google Scholar CrossRef Search ADS PubMed  Wang J., Tang L., Glenn T. C., Wang J.-S. ( 2016). Aflatoxin B1 induced compositional changes in gut microbial communities of male F344 rats. Toxicol. Sci . 150, 54– 63. Google Scholar CrossRef Search ADS PubMed  Weir T. L., Manter D. K., Sheflin A. M., Barnett B. A., Heuberger A. L., Ryan E. P. ( 2013). Stool microbiome and metabolome differences between colorectal cancer patients and healthy adults. PLoS One  8, e70803. Google Scholar CrossRef Search ADS PubMed  Wen L., Wong F. S. ( 2017). Dietary short-chain fatty acids protect against type 1 diabetes. Nat. Immunol . 18, 484– 486. Google Scholar CrossRef Search ADS PubMed  Whelan J., Fritsche K. ( 2013). Linoleic Acid. Adv. Nutr . 4, 311– 312. Google Scholar CrossRef Search ADS PubMed  Whittaker B. L., Chipley J. R. ( 1986). Interaction of aflatoxin B1 with bacterial DNA: Possible relationship to bacteriophage induction in lysogenic Bacillus megaterium. J. Environ. Pathol. Toxicol. Oncol . 6, 295– 304. Google Scholar PubMed  Winder C. L., Dunn W. B., Schuler S., Broadhurst D., Jarvis R., Stephens G. M., Goodacre R. ( 2008). Global metabolic profiling of Escherichia coli cultures: An evaluation of methods for quenching and extraction of intracellular metabolites. Anal. Chem . 80, 2939– 2948. Google Scholar CrossRef Search ADS PubMed  Wong J. M., De Souza R., Kendall C. W., Emam A., Jenkins D. J. ( 2006). Colonic health: Fermentation and short chain fatty acids. J. Clin. Gastroenterol . 40, 235– 243. Google Scholar CrossRef Search ADS PubMed  Worley B., Powers R. ( 2016). PCA as a practical indicator of OPLS-DA model reliability. Curr. Metabolomics  4, 97– 103. Google Scholar CrossRef Search ADS PubMed  Xue K. S., Cai W., Tang L., Wang J.-S. ( 2016). Aflatoxin B 1-lysine adduct in dried blood spot samples of animals and humans. Food Chem. Toxicol . 98, 210– 219. Google Scholar CrossRef Search ADS PubMed  Yatsunenko T., Rey F. E., Manary M. J., Trehan I., Dominguez-Bello M. G., Contreras M., Magris M., Hidalgo G., Baldassano R. N., Anokhin A. P.et al.  , . ( 2012). Human gut microbiome viewed across age and geography. Nature  486, 222– 227. Google Scholar CrossRef Search ADS PubMed  Zhao G., Nyman M., Åke Jönsson J. ( 2006). Rapid determination of short‐chain fatty acids in colonic contents and faeces of humans and rats by acidified water‐extraction and direct‐injection gas chromatography. Biomed. Chromatogr . 20, 674– 682. Google Scholar CrossRef Search ADS PubMed  © The Author(s) 2018. Published by Oxford University Press on behalf of the Society of Toxicology. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

Toxicological SciencesOxford University Press

Published: Apr 20, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off