Molecular Medicine. Gut Microbiota Metabolism of Anthocyanin Promotes Reverse Cholesterol Transport in Mice Via Repressing mirna-10b

Transcription

1 Molecular Medicine Gut Microbiota Metabolism of Anthocyanin Promotes Reverse Cholesterol Transport in Mice Via Repressing mirna-10b Dongliang Wang, Min Xia, Xiao Yan, Dan Li, Lei Wang, Yuxuan Xu, Tianru Jin, Wenhua Ling Downloaded from by guest on May 8, 2018 Rationale: We and others have demonstrated that anthocyanins have antiatherogenic capability. Because intact anthocyanins are absorbed very poorly, the low level of circulating parent anthocyanins may not fully account for their beneficial effect. We found recently that protocatechuic acid (PCA), a metabolite of cyanidin-3 to 0-β-glucoside (Cy-3-G), has a remarkable antiatherogenic effect. Objective: To investigate whether mouse gut microbiota metabolizes Cy-3-G into PCA and to determine whether and how PCA contributes to the antiatherogenic potency of its precursor, Cy-3-G. Methods and Results: PCA was determined as a gut microbiota metabolite of Cy-3-G in ApoE / mice, verified by the utilization of antibiotics to eliminate gut microbiota and further microbiota acquisition. PCA but not Cy-3-G at physiologically reachable concentrations promoted cholesterol efflux from macrophages and macrophage ABCA1 and ABCG1 expression. By conducting a mirna microarray screening, we revealed that expression of mirna-10b in macrophages can be reduced by PCA. Functional analyses demonstrated that mirna-10b directly represses ABCA1 and ABCG1 and negatively regulates cholesterol efflux from murine- and human-derived macrophages. Further in vitro and ex vivo analyses verified that PCA accelerates macrophage cholesterol efflux, correlating with the regulation of mirna-10b-abca1/abcg1 cascade, whereas Cy-3-G consumption promoted macrophage RCT and regressed atherosclerotic lesion in a gut microbiotaendependent manner. Conclusions: PCA, as the gut microbiota metabolite of Cy-3-G, exerts the antiatherogenic effect partially through this newly defined mirna-10b-abca1/abcg1-cholesterol efflux signaling cascade. Thus, gut microbiota is a potential novel target for atherosclerosis prevention and treatment. (Circ Res. 2012;111: ) Key Words: anthocyanin protocatechuic acid gut microbiota reverse cholesterol transport micro RNA atherosclerosis monocytes cholesterol homeostasis vascular remodeling Anthocyanins, as a group of flavonoids, are most abundant in various colorful fruits, vegetables, red wine, and grains. 1 Epidemiological and animal investigations have shown that anthocyanins reduce the severity or risk of atherosclerosis. 2 5 Many in vitro studies have revealed that anthocyanins have antioxidant and anti-inflammatory properties as well as capability in promoting cholesterol efflux from macrophages. 1,6,7 However, the effective levels were determined at a range between 10 to 200 μmol/l. In vivo intervention in human subjects and most animal test systems have shown that only approximately 0.1% of the intact anthocyanins are typically absorbed, and the maximum plasma concentration (Cmax) of anthocyanins is usually less than 0.1 μmol/l Recent studies have suggested that after the ingestion, anthocyanins are rapidly biotransformed into various metabolites, 11,12 raising the possibility that those metabolites, either working on their own or in combination with the parent anthocyanins, in protecting the subjects from developing atherosclerosis. In This Issue, see p 947 Editorial, see p 948 Very recently, we found that protocatechuic acid (PCA), a metabolite of cyanidin-3-o-β-glucoside (Cy-3-G), 13 could greatly inhibit the development of early atherosclerosis in the apolipoprotein Eendeficient (ApoE / ) mice. 14 However, it remains unknown how the conversion of Cy-3-G into PCA takes place in vivo, although a few in vitro studies have shown that PCA could be derived from human gut microbiota Original received February 4, 2012; revision received July 8, 2012; accepted July 19, In June 2012, the average time from submission to first decision for all original research papers submitted to Circulation Research was days. From the Department of Nutrition, School of Public Health, Sun Yat-Sen University, Guangzhou, PR China (D.W., M.X., X.Y., D.L., L.W., Y.X., T.J., W.L.); the Department of Physiology, University of Toronto, Toronto, Canada (T.J.); and Guangdong Provincial Key Laboratory of Food, Nutrition, and Health, Guangzhou, PR China (W.L.). The online-only Data Supplement is available with this article at /-/DC1. Correspondence to Wenhua Ling, PhD, Department of Nutrition, School of Public Health, Sun Yat-Sen University, Guangzhou , PR China. lingwh@mail.sysu.edu.cn 2012 American Heart Association, Inc. Circulation Research is available at DOI: /CIRCRESAHA

2 968 Circulation Research September 28, 2012 Downloaded from by guest on May 8, 2018 Non-standard Abbreviations and Acronyms ABCA1 ATP-binding cassette transporter A1 ABCG1 ATP-binding cassette transporter G1 Abx antibiotics AcLDL acetylated low-density lipoprotein ActD actinomycin D Apo apolipoprotein ApoE / apolipoprotein Eendeficient BW body weight Cmax maximum plasma concentration Cy-3-G cyanidin-3-o-β-glucoside DMSO dimethyl sulfoxide FBS fetal bovine serum FPLC fast protein liquid chromatography HDL-C high-density lipoprotein cholesterol HOXD10 homeobox D10 HPLC/MS/MS high-performance liquid chromatography-tandem mass spectrometry J774 J774A.1 KLF-4 Kruppel-like factor 4 LDL-C low-density lipoprotein cholesterol LXRα liver X receptor-α mir-10b mirna-10b mirnas micrornas MPMs mouse peritoneal macrophages PCA protocatechuic acid qrt-pcr quantitative real-time polymerase chain reaction RCT reverse cholesterol transport shrna short-hairpin RNA SR-BI scavenger receptor class B type I TC total cholesterol TG total triglyceride TNF-α tumor necrosis factor-α 3 -UTR 3 -untranslated region metabolism of Cy-3-G or produced by the spontaneous degradation of Cy-3-G under physiological conditions at neutral ph Moreover, if PCA is among the primary contributors in attenuating atherosclerotic lesions, mechanisms underlying the function of PCA remain to be explored. Plasma high-density lipoprotein cholesterol (HDL-C) levels inversely correlate with the development of atherosclerotic diseases. 18 In addition to its antioxidative and anti-inflammatory properties, HDL may attenuate atherosclerosis via stimulating reverse cholesterol transport (RCT), a process that entails the efflux of excess cholesterol from macrophages into the liver followed by its excretion into the bile and then the feces. 19 The promotion of macrophage RCT is a potential novel approach for atherosclerosis prevention and treatment, which can be achieved by upregulating the expression of ATP-binding cassette transporter A1 (ABCA1) and G1 (ABCG1) in macrophages. 20 MicroRNA (mirna) regulates gene expression through recognizing the 3 -untranslated region (3 -UTR) of its target mrna to repress its translation or facilitate its cleavage. 21 Recently, a few mirnas, including mirna-33 (mir-33) 22 and mir-758, 23 were shown to inhibit macrophage cholesterol efflux by downregulating ABCA1 or ABCG1 expression. Whether there are other mirnas in macrophages that modulate ABCA1 and ABCG1 expression remain unknown. In the present study, we investigated the contribution of PCA in the antiatherosclerotic effects of the parent Cy- 3-G, determined its effect in promoting macrophage RCT, and revealed the existence of a novel mechanism by which PCA represses mir-10b expression, increases ABCA1 and ABCG1 expression, and accelerates macrophage RCT in the ApoE / mice. Methods Gut Microbiota Metabolism of Cy-3-G The metabolism of Cy-3-G (Polyphenols AS, Sandnes, Norway) was investigated in both ApoE / mice (Jackson Laboratories, Sacramento, CA) and germ-free Kunming mice (Third Military Medical University, Chongqing, China). All of the animal procedures were approved by the Animal Care and User Committee of Sun Yat-Sen University. The expanded Methods are available in the Online Data Supplement. Cell Culture Mouse peritoneal macrophages (MPMs) from adult C57BL/6J mice (Jackson Laboratories) were obtained by peritoneal lavage 4 days after the intraperitoneal injection of a 1.0-mL volume of 4% thioglycollate. 13 The thioglycollate-elicited MPMs were maintained in RPMI 1640 medium containing 10% fetal bovine serum (FBS) (Gibco BRL, Grand Island, NY). Unless otherwise stated, references to MPMs concern the thioglycollate-elicited MPMs. J774.A1 (J774) murine macrophages, human THP-1 monocytes, and HEK293 cells were obtained from the American Type Culture Collection (Manassas, VA) and cultured per manufacturer s instructions. The differentiation of THP-1 monocytes into macrophages was induced in the presence of 167 nmol/l phorbol 12-myristate 13-acetate (Wako, Tokyo, Japan) for 48 hours. 24 In Vitro Cholesterol Efflux Assays Assays of cellular cholesterol efflux were performed as described previously. 7 Details are provided in the Online Data Supplement. Quantitative Real-Time RT-PCR Total RNA from cells was isolated using the mirneasy isolation kit (Qiagen, Hilden, Germany). Quantitative real-time polymerase chain reaction (qrt-pcr) was performed using iq SYBR Green Supermix (Bio-Rad, Hercules, CA) on the Applied Biosystems 7000 sequence detection system. GAPDH mrna was used as the internal control. Primer sequences for all the genes examined in this study are shown in Online Table I. For mirna quantification, total RNA was reverse-transcribed using the RT 2 mirna First Strand kit (SABiosciences, San Diego, CA). Primers specific for mouse mir-10b, mir-300, mir-370, and mir-493 (SABiosciences) were used and the values were normalized to U6 as a housekeeping gene. ABCA1 and ABCG1 mrna Stability Assay Actinomycin D (ActD) (Sigma-Aldrich, St Louis, MO), a transcriptional inhibitor, 25 was used to examine the effect of Cy-3-G or PCA (Sigma-Aldrich) on ABCA1 and ABCG1 mrna stability. The expanded Methods are available in the Online Data Supplement. mirna Microarray Analysis MPMs preloaded with acetylated low-density lipoprotein (AcLDL) (25 μg/ml) for 24 hours were incubated in RPMI 1640 medium containing 0.1% bovine serum albumin and PCA (1 μmol/l) or the vehicle (DMSO) for another 24 hours. Total RNA was extracted

3 Wang et al Cy-3-G Needs Gut Flora to Promote RCT Via MiR-10b 969 Downloaded from by guest on May 8, 2018 using Trizol and the mirneasy Mini Kit (Qiagen). The RNA quality and quantity were determined with a Nanodrop spectrophotometer (ND-1000, Nanodrop Technologies), and the RNA integrity was determined by gel electrophoresis. The labeling, hybridization, and scanning of the isolated RNA were performed by Exiqon (Denmark) using an LNA mircury array (v.16.0). The scanned images were then imported into the GenePix Pro 6.0 software (Axon) for grid alignment and data extraction. The mirna expression levels that were modulated with Fold Change Filtering (treated/control) of 2.0 or more were considered to be differentially expressed. mir-10b and Anti mir-10b Transfection MPMs and THP-1 macrophages preloaded with or without AcLDL and HEK293 were transfected with miridian mirna mimics (mir-10b) or with miridian mirna inhibitors (antienmir-10b) (Dharmacon, Lafayette, CO) at the concentrations indicated in the corresponding figure legends, using Oligofectamine (Invitrogen) for 48 hours. All the control samples were treated with an equal concentration of a nontargeting negative control mimic sequence (Con mir) or an inhibitor negative control sequence (Con Inh) (Dharmacon) for use as controls for nonensequence-specific effects in the mirna experiments. 3 -UTR Vector Construction and Luciferase Reporter Assays The entire 3 -UTRs sequences of mouse ABCA1 (mabca1) (3153 nt, NM_013454), mabcg1 (3696 nt, NM_009593), habca1 (3309 nt, NM_005502), and habcg1 (850 nt, NM_004915) were cloned into the EcoRV and Xba1 sites of pgl3 cm created based on the vector pgl3-control (Promega), generating the pgl3-m/hab- CA1-WT and pgl3-m/habcg1-wt constructs. Point mutations in the seed regions of predicted mir-10b sites within the 3 -UTRs of m/h ABCA1 and ABCG1 were generated using QuikChange Multi Site-Directed Mutagenesis (Stratagene) according to the manufacturer s protocol. All the constructs were confirmed by sequencing. For the luciferase reporter assay, HEK293 cells were seeded at cells/well in a 96-well plate 24 hours before transfection. The cells were cotransfected with a series of concentrations of mir-10b mimics or negative control mimics (Dharmacon) (60 nmol/l), either 50 ng of pgl3-wt or pgl3-mutation and 50 ng of prl-tk (Promega) by Lipofectamine 2000 (Invitrogen) for 48 hours. Next, the luciferase activity was measured using the Dual-Luciferase Reporter Assay System (Promega), and prl-tk was used as an internal control. The experiments were performed in triplicate wells of a 96-well plate and repeated at least 3 times. In determining the effect of Cy-3-G or PCA, MPMs, and THP- 1 macrophages preloaded with AcLDL (25 μg/ml) for 24 hours were cotransfected with 50 ng of the corresponding pgl3-wt and prl-tk in the presence or absence of Cy-3-G or PCA (1 μmol/l) for 24 hours. The luciferase activity was then determined as described above. Construction of Short-Hairpin RNA Plasmids The short-hairpin RNA (shrna) sequences for murine ABCA1 and ABCG1 were designed using Invitrogen s BLOCK-iT RNAi Designer. Negative control shrna sequences for LacZ were provided by Invitrogen. cdna oligonucleotides were annealed and cloned into pentr/u6 (Invitrogen). The expression constructs were then generated by an LR recombination reaction between the pentr/u6 entry construct and the plenti6/block-it-dest expression construct (Invitrogen) according to the manufacturer s instructions. The final constructs were confirmed by DNA sequencing. The murine ABCA1 and ABCG1 RNAi sequences were 5 -TCCCTGGGTGTCAGTAATTCT-3 and 5 -GGAAGTCCACACTCATGAATA-3, respectively. Lentiviral Vector Transduction Lentiviral stocks were produced by cotransfecting the plenti6/ BLOCK-iT-DEST constructs expressing shrna targeting ABCA1 and ABCG1 or LacZ with ViraPower Packaging Mix (Invitrogen) into the 293FT Producer Cell Line (Invitrogen). These lentiviral stocks were then used to transduce J774 cells. Stably transduced cell lines were selected in DMEM supplemented with 10% FBS and 5 μg/ml blasticidin. ABCA1 or ABCG1 mrna and protein levels in each cell line were determined by qrt-pcr or Western blotting, respectively. Mouse Studies Male ApoE / mice were weaned at 4 weeks of age and fed with an AIN-93G diet for another 26 weeks, at which point the mice were either euthanized (baseline, n=12) or subjected to the following experiments. Experiment 1 The ApoE / mice fed with the AIN-93G diet in the absence and presence of an antibiotic cocktail (0.5 g/l vancomycin, 1 g/l neomycin sulfate, 1 g/l metronidazole, 1 g/l ampicillin) were orally gavaged with Cy-3-G (50 mg/kg body weight BW), PCA (5 mg/kg BW), or Cy-3-G (50 mg/kg BW) plus PCA (5 mg/kg BW) once daily for 14 days. On day 10, half of the mice (n=24) in each group were intraperitoneally injected with thioglycollate to elicit peritoneal macrophages. The remaining mice were individually housed in metabolic cages during the last 48 hours, and the efficiency of macrophage RCT in vivo was then quantified (described below). Experiment 2 The ApoE / mice fed with the AIN-93G diet off and on an antibiotic cocktail were orally gavaged with Cy-3-G (50 mg/kg BW), PCA (5 mg/kg BW), or Cy-3-G (50 mg/kg BW) plus PCA (5 mg/kg BW) once daily for 4 weeks. At the end of the experimental period, overnight-fasted mice were anesthetized with isoflurane and exsanguinated by cardiac puncture. The MPMs, hearts, and thoracic and abdominal aortas were then collected for further analysis. In Vivo Macrophage RCT Assay RCT studies were performed as previously described. 26 Briefly, 3 H-cholesterol-labeled and AcLDL-loaded macrophages were resuspended in RPMI 1640 medium followed by intraperitoneal injection into the ApoE-deficient mice. 3 H-tracer recovery in blood, liver, and feces was then monitored and converted into RCT. For more details, see the Online Data Supplement. Other Laboratory Procedures Methods for the quantification of cellular protein expression by Western blotting, 27 plasma lipoprotein preparation and cellular lipid content, 7 plasma lipids and apolipoprotein A-I (ApoA-I) using commercial kits, cholesterol distribution in plasma lipoproteins separated by fast protein liquid chromatography (FPLC) gel filtration, 25 atherosclerosis at the aortic sinus, 14 and aortic cholesterol content 7 are shown in the Online Data Supplement. Statistical Analysis Results are presented as mean±sem. The data were analyzed by 2-way repeated-measures ANOVA coupled with the post hoc Bonferroni-corrected t test, 1-way ANOVA coupled with the Student- Newman-Keuls multiple comparison test, or full factorial ANOVA (univariate general linear model) coupled with the independent Student t test where appropriate with the use of SPSS for Windows software (version 13.0; SPSS Inc, Chicago, IL). Statistical significance was accepted for P<0.05. Results PCA Is a Gut Microbiota Metabolite of Cy-3-G ApoE / mice were orally gavaged with Cy-3-G at 5 or 50 mg/kg BW that can be achieved in the human diet. 8 Cy-3-G feeding at 50 mg/kg BW resulted in a plasma Cmax of Cy- 3-G and its metabolite PCA at 295 nmol/l or 979 nmol/l (Figure 1A), respectively. To determine the role of gut

4 970 Circulation Research September 28, 2012 Downloaded from by guest on May 8, 2018 Figure 1. PCA is a gut microbiota metabolite of Cy-3-G. Plasma PCA and Cy-3-G concentrations were measured before, 0.5 hours, and 2 hours after oral Cy-3-G administration in the ApoE / mice (A). The metabolic assays were then made after the suppression of the gut microbiota with an antibiotic cocktail (3 weeks) (B) and after the placement of the mice into conventional cages (4 weeks) with nonsterile mice (conventionalized) (C). The top and bottom (A, B, and C) indicate mice orally gavaged once with 5 mg/kg BW and 50 mg/ kg BW of Cy-3-G, respectively. Results are mean±sem, n=12. microbiota in the conversion of Cy-3-G to PCA, we applied 2 strategies. First, ApoE / mice were treated with antibiotics to eliminate the gut microbiota, resulting in the lack of detection of PCA in the plasma (Figure 1B). When the antibioticstreated mice were recolonized with normal gut microbiota, plasma PCA reappeared (Figure 1C). Second, germ-free mice orally fed Cy-3-G (50 mg/kg BW) did not produce PCA (Online Figure IA), whereas these mice colonized with normal gut microbiota generated PCA after Cy-3-G oral administration (Online Figure IB). Collectively, these data demonstrated that PCA is a gut microbiota metabolite of Cy-3-G. PCA But Not Cy-3-G at Physiologically Reachable Concentrations Promotes Cholesterol Efflux From Macrophages In Vitro We demonstrated previously that Cy-3-G at supraphysiological concentrations ( μmol/l) promotes cholesterol efflux from macrophages to ApoA-I. 7 We showed that Cy-3-G at physiologically reachable concentrations ( μmol/l) had no effect on cholesterol efflux from AcLDLloaded MPMs or THP-1 macrophages to ApoA-I or HDL (Figure 2A and 2B). However, PCA at physiologically reachable concentrations ( μmol/l) greatly increased cholesterol efflux from the 2 AcLDL-loaded macrophages (Figure 2A and 2B). Additionally, there was no additive effect on cholesterol efflux from either AcLDL-loaded MPMs or THP-1 macrophages with the combination of PCA (1 μmol/l) and its precursor Cy-3-G (1 μmol/l) (data not shown). Because Cy-3-G at the concentration lower than 1 μmol/l could not affect macrophage cholesterol efflux, it was used as the negative control for further PCA studies unless otherwise indicated. PCA at Physiologically Reachable Concentrations Induces ABCA1 and ABCG1 Expression Via Downregulating mir-10b Expression in Macrophages ABCA1 is known to mediate the cholesterol efflux from macrophages to ApoA-I, whereas ABCG1 and scavenger receptor class B type I (SR-BI) facilitate the cholesterol efflux from macrophages to HDL. 20 We found by qrt-pcr and Western blotting that PCA but not Cy-3-G increased ABCA1 and ABCG1 but not SR-BI expression in AcLDLloaded MPMs (Figure 2C and 2D) and THP-1 macrophages (Figure 2E and 2F).

5 Wang et al Cy-3-G Needs Gut Flora to Promote RCT Via MiR-10b 971 Downloaded from by guest on May 8, 2018 Figure 2. PCA but not Cy-3-G promotes cholesterol efflux from murine- and human-derived macrophages. A and B, Cholesterol efflux to ApoA-I or HDL in AcLDL-loaded MPMs (A) and THP-1 macrophages (B) treated with Cy-3-G, PCA or the vehicle (DMSO) for 24 hours. The mrna and protein expression of ABCA1, ABCG1, and SR-BI in AcLDL-loaded MPMs (C and D) and THP-1 macrophages (E and F) were treated with Cy-3-G, PCA, or DMSO for 24 hours. The half-lives of ABCA1 and ABCG1 mrna in AcLDLloaded MPMs (G) and THP-1 macrophages (H) pretreated with Cy-3-G (1 μmol/l), PCA (1 μmol/l), or DMSO for 24 hours were determined with the addition of actinomycin D. PCA but not Cy-3-G (1 μmol/l) treatment restores the attenuated luciferase activity induced by the presence of mouse and human ABCA1/ABCG1 3 -UTR in AcLDL-loaded MPMs (I) and THP-1 macrophages (J), respectively. D and F are representative images of 3 independent experiments. For the other panels, results are mean±sem, n=3 to 6. *P<0.05 vs. control. #P<0.05. To investigate how PCA stimulates ABCA1 and ABCG1 expression, we applied 2 strategies. As ABCA1 and ABCG1 are transcriptionally regulated by liver X receptor alpha (LXRα), 28 we have first assessed the effect of PCA on the ABCA1 or ABCG1 promoter activities using human ABCA1 or ABCG1 promoter luciferase fusion gene constructs. Both PCA and Cy-3-G ( μmol/l) showed no appreciable effect on human ABCA1 or ABCG1 promoter activities in AcLDL-loaded MPMs or THP-1 macrophages (Online Figure II A and B). Moreover, PCA or Cy-3-G ( μmol/l) did not affect LXRα mrna and protein levels in the 2 macrophages (Online Figure III A through D). Furthermore, PCA or Cy-3-G had no effect on LXRα activation by assessing liver X receptor element (LXRE)-mediated luciferase activity (Online Figure III E and F). Next, ActD was used to assess the mrna stability of ABCA1 and ABCG1. We found that PCA but not Cy-3-G (1 μmol/l) significantly prolonged the half-lives of ABCA1 and ABCG1 mrna in AcLDL-loaded MPMs and THP-1 macrophages (Figure 2G and 2H). We have also investigated whether PCA modulates 3 -UTRs, a mechanism that destabilizes mrna. As shown in Figure 2I and 2J, PCA but not Cy-3-G (1 μmol/l) restored the attenuated luciferase reporter activities induced by the presence of mouse and human ABCA1 or ABCG1 3 -UTR.

6 972 Circulation Research September 28, 2012 Downloaded from by guest on May 8, 2018 We hence suggest that PCA posttranscriptionally regulates ABCA1 and ABCG1 mrna expression, at least in part, through enhancing the stability of those mrnas mediated by their 3 -UTR sequences. mirna plays critical roles in mrna stability via interacting with 3 -UTRs of their target gene. 29 Several flavonoids were shown to affect mirna expression. 30 We then performed unbiased mirna microarray analysis with the chip containing 1055 mirna probes. Online Table II shows that the expression of four mirnas (mir-10b, 300, 370, and 493) was at least 2-fold different in the PCA-treated MPMs compared with vehicle-treated cells. qrt-pcr further confirmed that PCA but not Cy-3-G (1 μmol/l) induced these 4 mir- NAs changes in AcLDL-loaded MPMs (Figure 3A) and J774 murine macrophages (Online Figure IV). In AcLDL-loaded THP-1 macrophages, except for mir-493, the other 3 mir- NAs were also shown to be regulated by PCA (Figure 3B). Using several mirna target prediction databases, 31 we identified putative binding sites for mir-10b in the 3 -UTR of mouse and human ABCA1 and ABCG1 genes (show in the bottom panels of Figure 3E and 3F). To evaluate the effects of mir-10b on the 3 -UTR of mouse and human ABCA1 and ABCG1, we used luciferase reporter constructs. We found that mir-10b dose-dependently repressed ABCA1 or ABCG1 3 -UTR activities of either mouse or human origins (Figure 3C and 3D). To ensure the specificity of these interactions, mutants of ABCA1 and ABCG1 3 - UTR of either mouse or human origins that contained a 4-nt change in the seed region of the target sequence were constructed (Figure 3E and 3F, bottom panels). The mutation of the mir-10b target sites relieved mir-10b repression of mouse and human ABCA1 and ABCG1 3 -UTR activities (Figure 3E and 3F, top panels), consistent with a direct interaction of mir- 10b with these sites. Furthermore, mir-10b strongly decreased both ABCA1 and ABCG1 mrna and protein expression in AcLDL-loaded MPMs (Figure 3G and 3H) and THP-1 macrophages (Figure 3I and 3J), whereas antagonism of endogenous mir-10b increased their expression in both cells (Figure 3K through 3N). Together, these observations suggest that ABCA1 and ABCG1 are functional targets of mir-10b. Next, we tested whether mir-10b affects the ability of macrophages to efflux free cholesterol. As shown in Figure 4A and 4B, mir-10b attenuated cholesterol efflux from both AcLDLloaded MPMs and THP-1 macrophages to ApoA-I and HDL, whereas antagonism of endogenous mir-10b increased the cholesterol efflux (Figure 4C and 4D). To further explore the role of mir-10b in PCA-induced ABCA1 and ABCG1 expression in macrophages and its induced cholesterol efflux function, we used gain-of-function (overexpressed mir-10b) and loss-of-function (down-expressed mir-10b) assays in AcLDL-loaded MPMs. The results of the gain-of-function experiment revealed that PCA treatment reversed the inhibitory effects of mir-10b on the expression of ABCA1 and ABCG1 at both protein and mrna levels (Figure 5A and 5B), as well as cholesterol efflux to ApoA-I and HDL (Figure 5C), whereas PCA-induced functions were not observed in AcLDL-loaded MPMs transfected with antienmir-10b (Figure, 5A, 5D, and 5E). Additionally, Cy-3-G did not show any significant effect in these testing systems as PCA did (Figure 5A through 5E). Because PCA also markedly reduced mir-300 and increased the expression of mir-370 and mir-493 (Figure 3A), we further evaluated the potential effects of these mirnas on cholesterol efflux from AcLDL-loaded MPMs. Antagonism of endogenous mir-300 (Online Figure VA) or overexpression of mir-370 or 493 (Online Figure VB and C) had no effect on cholesterol efflux from the cells. Collectively, the observations imply that the stimulatory effects of PCA on cholesterol efflux from AcLDL-loaded MPMs via upregulating ABCA1 and ABCG1 expression involve the downregulation of mir-10b. PCA Is Responsible for the Stimulatory Efficacy of Cy-3-G on Macrophage RCT In Vivo To test whether the in vitro findings on cholesterol efflux from macrophages by PCA could be extended into the in vivo circumstances, 30-week-old male ApoE / mice were orally gavaged once daily with PCA (5 mg/kg BW) for 14 days. PCA feeding resulted in approximately 0.8 μmol/l PCA in the circulatory system (Online Figure VI), which is effective in driving cholesterol efflux from macrophages in vitro (Figure 2A and 2B). MPMs isolated from mice treated with PCA for 14 days had lower levels of cellular mir-10b (Figure 6A), associated with increased ABCA1 and ABCG1 expression at both mrna and protein levels (Figure 6B through 6D), as well as enhanced cholesterol efflux ability to ApoA-I or HDL (Figure 6E and 6F). Because feeding with Cy-3-G (50 mg/kg BW) produced similar levels of plasma PCA (Figure 1 and Online Figure VI), it is assumed that Cy-3-G at a higher dose may have an analogous effect with PCA. As result, orally feeding Cy-3-G (50 mg/kg BW) in 30-weekold male ApoE / mice for 14 days improved the cholesterol efflux, similar to the effect by a lower-dose PCA feeding (Figure 6A through 6F). To further verify that Cy-3-G-induced beneficial effects (Figure 6A through 6F) depend on its gut microbiota metabolite PCA, mice that received Cy-3-G or PCA were concomitantly administered antibiotics. Notably, antibiotics administration completely abolished the Cy-3-Geninduced effects, including the repression of mir-10b, the activation of ABCA1 and ABCG1, as well as the stimulation of cholesterol efflux in isolated MPMs (Figure 6A through 6F), whereas antibiotics had no effect on PCA-induced changes (Figure 6A through 6F). To further verify whether PCA is required for the effects induced by its parent Cy-3-G, we examined the impact of Cy-3-G (50 mg/kg BW) plus PCA (5 mg/kg BW) with or without antibiotics treatment. Evidently, Cy-3-G plus PCA reversed the inhibitory effect of antibiotics on Cy-3-Geninduced effects (Figure 6A through 6F). Moreover, there was an additive effect between Cy-3-G and PCA without antibiotics treatment (Figure 6A through 6F). To determine whether the changes of ABCA1 and ABCG1 expression are regulated by mir-10b, we examined a time course of their expression in isolated MPMs related to the pharmacokinetics of PCA in ApoE-deficient mice. The mice were treated orally with PCA (5 mg/kg BW) for 0 to 16 hours. The variation tendency of PCA concentrations in plasma, mir-10b, and ABCA1/ABCG1 mrna expression in isolated MPMs were showed in Online Figure VII. Plasma PCA

7 Wang et al Cy-3-G Needs Gut Flora to Promote RCT Via MiR-10b 973 Downloaded from by guest on May 8, 2018 Figure 3. ABCA1 and ABCG1 genes are direct targets of mir-10b. A and B, Expression of 4 mirnas in AcLDL-loaded MPMs (A) and THP-1 macrophages (B) treated with Cy-3-G (1 μmol/l), PCA (1 μmol/l), or the vehicle (DMSO) for 24 hours. C and D, The activity of luciferase reporter constructs fused to the 3 -UTRs of mouse (C) and human (D) ABCA1/ABCG1 in HEK293 cells transfected with mir- 10b or control mir (Con mir). E and F, The mir-10b-binding sites in mouse and human ABCA1/ABCG1 are shown in bottom panels of E and F. Mutants were generated in the mouse and human ABCA1/ABCG1 3 -UTR seed regions, as indicated. Luciferase reporter activity in HEK293 cells transfected with Con mir (60 nmol/l) or mir-10b (60 nmol/l) of the mouse (E) or human (F) ABCA1/ABCG1 3 -UTRs containing the indicated point mutations (PM) in the mir-10b target sites. G to J, Quantitative RT-PCR (G and I) and Western blotting (H and J) analyses of mouse and human ABCA1/ABCG1 in AcLDL-loaded MPMs and THP-1 macrophages transfected with the indicated concentrations of Con mir or mir-10b. K to N, Quantitative RT-PCR (K and M) and Western blotting (L and N) analyses of mouse and human ABCA1/ABCG1 in AcLDL-loaded MPMs and THP-1 macrophages transfected with Con mir (150 nmol/l) or mir-10b (150 nmol/l) in the presence or absence of a control inhibitor (Con Inh) (150 nmol/l) or antienmir-10b (150 nmol/l). H, J, L, and N are representative images of 3 independent assays. C to F, Data are expressed as the mean percentage of the 3 -UTR activity of Con mir±sem, n=3. G, I, K, and M, Data are expressed as meansem of fold of either Con mir (G and I) or Con mir and Con Inh (K and M). n=3. *P<0.05 vs. Con mir or Con mir and Con Inh. concentration reached the peak at 0.5 hours, whereas the level of macrophage mir-10b began to decrease at 1 hour, reached the lowest at 4 hours, then maintained lower level till 16 hours. ABCA1 and ABCG1 mrna expression started to increase at 2 hours, reached the highest at 8 hours, then remained at higher level until 16 hours. The lag response of ABCA1 and ABCG1 to mir-10b suggests that PCA may first reduce mir-10b levels and then in turn induce ABCA1 and ABCG1 expression. Consistent with the results from the ex vivo studies, a 14- day intervention with Cy-3-G (50 mg/kg BW) or PCA (5 mg/kg BW) significantly increased the cholesterol transport from macrophages to feces in the ApoE / mice (Figure 6G). Moreover, antibiotics treatment blocked the positive effect of

8 974 Circulation Research September 28, 2012 Downloaded from by guest on May 8, 2018 Figure 4. mir-10b regulates the ability of macrophages to efflux free cholesterol. A and B, MPMs and THP-1 macrophages preloaded with 3H-cholesterol and AcLDL, as described in Methods, were transfected with control mir (Con mir) or mir-10b for 48 hours. Cholesterol efflux was then initiated to ApoA-I or HDL from the 3H-cholesterolenlabeled and AcLDL-loaded MPMs (A) and THP-1 cells (B) for 24 hours. C and D, Cholesterol efflux to ApoA-I or HDL from the 3H-cholesterolenlabeled and AcLDL-loaded MPMs (C) and THP-1 macrophages (D) transfected with control inhibitor (Con Inh) or antienmir-10b. Results are mean±sem, n=3 to 6. *P<0.05 vs. Con mir (0 nmol/l) or Con Inh (0 nmol/l). Cy-3-G but not PCA on macrophage RCT (Figure 6G), whereas the inhibitory effect of antibiotics on Cy-3- Geninduced macrophage RCT was reversed by cotreatment with PCA (Figure 6G). In contrast, both Cy-3-G and PCA had no significant effect on cholesterol tracer recovery in the liver at 48 hours (Figure 6H) or in plasma at 6, 24, and 48 hours (Figure 6I). Taken together, these data suggest that the Cy-3-Genmediated macrophage RCT is at least mainly, if not completely, dependent on its gut microbiota metabolite PCA, involving the repression of mir-10b and the activation of ABCA1 and ABCG1-cholesterol efflux pathway in macrophages. Macrophage ABCA1 and ABCG1 Are Partially Responsible for the Stimulatory Effect of PCA on Macrophage RCT In Vivo Because PCA markedly increased 3 H-tracer recovery in feces but not in plasma and liver, we then performed a proof-of-principle study to investigate how importantly PCA-induced macrophage cholesterol efflux does contribute to PCA-enhanced RCT in ApoE-deficient mice. We first constructed lentiviral vectors that drive the expression of a shrna targeting murine ABCA1 or ABCG1 and used both of them to transduce J774 macrophages to obtain cell lines in which the expression of both ABCA1 and ABCG1 has been knocked down (J774-DKD). J774 macrophages transduced with a lentiviral vector encoding LacZ shrna (J774-Con) and nontransfected cells served as the control cells. ABCA1 and ABCG1 mrna (Online Figure VIIIA and B) as well as their protein expression (Figure 6J) were significantly lower in J774-DKD macrophages than J774-Con cells in either the absence or presence of AcLDL. As expected, a substantial reduction in cholesterol efflux to ApoA-I or HDL from the AcLDL-loaded J774-DKD cells was observed compared with the AcLDLloaded J774-Con cells (Online Figure VIIIC and D). Moreover, in agreement with a previous study, 20 ApoE-deficient mice injected with 3 H-cholesterol and AcLDL-loaded J774-DKD macrophages had a lower ability to transport cholesterol from macrophages to feces than the control mice injected with AcLDL-loaded J774-Con cells (Figure 6K). Thus, knockdown of ABCA1 and ABCG1 in J774-DKD cells effectively impaired macrophage cholesterol efflux in vitro and in vivo. Next, we performed a 2 2 factorial design of RCT study with 2 factors (named A and B). Factor A is 3 H-cholesterol-labeled J774-Con macrophages versus 3 H-cholesterol-labeled J774- DKD cells, and factor B is PCA or Cy-3-G treatment versus its vehicle (normal saline) treatment. The full factorial ANOVA analysis showed a significant interaction between macrophage ABCA1 and ABCG1 expression and PCA or Cy-3-G treatment (PCA: F=12.7, P=0.001; Cy-3-G: F=10.5, P=0.002). ABCA1 and ABCG1 in macrophages contributed to the promotion of macrophage RCT induced by a 14-day intervention with PCA (5 mg/kg BW) or Cy-3-G (50 mg/kg BW) by 55.4% (95% confidence interval, 40.4% to 70.4%) and 57.9% (95% confidence interval, 42.1% to 73.7%) (Figure 6L and 6M), respectively. Collectively, these data firmly demonstrated that PCA or Cy- 3-G promotes macrophage RCT in vivo partially via upregulating ABCA1 and ABCG1 expression in macrophages.

9 Wang et al Cy-3-G Needs Gut Flora to Promote RCT Via MiR-10b 975 Downloaded from by guest on May 8, 2018 Figure 5. mir-10b regulates the stimulatory effect of PCA on cholesterol efflux in macrophages. AcLDL-loaded MPMs were transfected with either a control mir (Con mir) (100 nmol/l) or mir-10b (100 nmol/l) or a control inhibitor (Con Inh) (100 nmol/l) or antienmir-10b (100 nmol/l) for 48 hours, after treatment with Cy-3-G (1 μmol/l), PCA (1 μ mol/l), or the vehicle (DMSO) for 24 hours. ABCA1 and ABCG1 protein and mrna expression were then determined by Western blotting (A) and qrt-pcr (B and D), respectively. C and E, The 3H-cholesterolenlabeled and AcLDL-loaded MPMs were transfected with either Con mir (100 nmol/l) or mir-10b (100 nmol/l) (C) or Con Inh (100 nmol/l) or antienmir- 10b (100 nmol/l) (E) for 48 hours, after treatment with Cy-3-G (1 μmol/l), PCA (1 μmol/l), or DMSO for 24 hours. Cholesterol efflux to ApoA-I and HDL was then determined in the 3H-cholesterolenlabeled and AcLDL-loaded MPMs transfected with either Con mir or mir-10b (C) or Con Inh or antienmir-10b (E). A, Representative image of 3 independent experiments. Results in B and D are expressed as mean±sem of fold of either Con mir (B) or Con Inh (D). Results in C and E are expressed as mean±sem, n=4 to 6. #P<0.05. NS indicates not significant. PCA Masters the Regressive Potential of Its Precursor Cy-3-G on Established Atherosclerotic Lesions Macrophage RCT plays a key role in the formation and/or regression of cholesterol-loaded macrophage foam cells. 32 Isolated MPMs from the ApoE / mice received Cy-3-G (50 mg/kg BW), PCA (5 mg/kg BW), or Cy-3-G (50 mg/kg BW) plus PCA (5 mg/kg BW) treatment for 4 weeks had a lower level of endogenous macrophage foam cell formation, indicated by comparing total cellular lipid (Figure 7A) and cholesterol (Figure 7B) levels in those mice with that of the control mice or the testing mice before PCA or Cy-3-G intervention. This observation indicates that Cy-3-G, or PCA, or Cy-3-G plus PCA intervention regresses the established endogenous macrophage foam cells. Antibiotics treatment abolished the regressive effect of Cy-3-G but not PCA treatment on the established endogenous macrophage foam cells (Figure 7A and 7B), whereas the inhibitory effect of antibiotics on the Cy-3-Geninduced regressive effect was restored by cotreatment with PCA (Figure 7A and 7B).

10 976 Circulation Research September 28, 2012 Downloaded from by guest on May 8, 2018 Figure 6. Cy-3-G accelerates cholesterol efflux from macrophages ex vivo and in vivo through its metabolite PCA. Male 30-weekold ApoE / mice with or without the antibiotics were orally gavaged with Cy-3-G (50 mg/kg BW), PCA (5 mg/kg BW), Cy-3-G (50 mg/ kg BW) plus PCA (5 mg/kg BW), or the vehicle (normal saline) once daily for 14 days. Four days before euthanasia, mice (n=12 per group) were intraperitoneally injected with thioglycollate to obtain thioglycollate-elicited MPMs for the analyses of cellular mir-10b content (A), ABCA1 and ABCG1 mrna (B and C) and protein (D) expression levels, and cholesterol efflux to ApoA-I (E) and HDL (F). The isolated MPMs were pooled with 4 mice in each group. D, Representative image of 3 independent experiments. The results in A, B, and C are mean±sem of fold of the control Abx ( ).The results in E and F are mean±sem, n=3. *P<0.05 vs. the control Abx ( ); #P<0.05; NS indicates not significant; Abx, antibiotics. G to I, Thirty-week-old male ApoE / mice with or without the antibiotics were orally gavaged with Cy-3-G (50 mg/kg BW), PCA (5 mg/kg BW), Cy-3-G (50 mg/kg BW) plus PCA (5 mg/kg BW), or the vehicle once daily for 14 days. On day 12, 3H-cholesterolenlabeled and AcLDL-loaded MPMs ( cells at cpm per mouse) were intraperitoneally injected into the mice (n=12 per group). After cell injection, blood at 6, 24, and 48 hours, liver at 48 hours (after the animals were euthanized), and feces within the 48-hour experimental period were collected for the analysis of macrophage RCT. 3H-cholesterol recovery in the feces (G), liver (H), and plasma (I). Results in G, H, and I are mean±sem, n=12. *P<0.05 vs. control Abx ( ); #P<0.05. J, Reduction in ABCA1 and ABCG1 expression on the protein level was determined in J774 macrophages stably transfected with vectors encoding for shrna against mouse ABCA1 (shabca1) and ABCG1 (shabcg1) that were loaded with or without AcLDL (25 μg/ml) for 24 hours. As control, cells transfected with vector encoding for shrna against LacZ (J774-Con) or nontransfected cells (NT) were used. J, Representative image of 3 independent experiments. K, Double knockdown of ABCA1 and ABCG1 in J774 macrophages (J774-DKD) impairs cholesterol efflux in vivo. Thirty-week-old male ApoE / mice were intraperitoneally injected with 3H-cholesterolenlabeled and AcLDL-loaded J774- DKD macrophages ( cells at cpm per mouse) or J774-Con cells ( cells at cpm per mouse). After cell injection, feces within 48 hours were obtained and 3H-tracer levels were measured. Results are mean±sem, n=8. #P<0.05. L and M, Thirty-week-old male ApoE / mice were orally gavaged with PCA (5 mg/kg BW) (L), Cy-3-G (50 mg/kg BW) (M), or vehicle once daily for 14 days. On day 12, 3H-cholesterolenlabeled and AcLDL-loaded J774-DKD ( cells at cpm per mouse) and J774- Con ( cells at cpm per mouse) were intraperitoneally injected into the mice. Feces within 48 hours were obtained and 3H-tracer levels were measured. Results are mean±sem, n=12. #P<0.05.

11 Wang et al Cy-3-G Needs Gut Flora to Promote RCT Via MiR-10b 977 Downloaded from by guest on May 8, 2018 Figure 7. Cy-3-G regresses endogenous foam cell formation and establishes atherosclerosis through its metabolite PCA in vivo. Male 30-week-old ApoE / mice with or without antibiotics were orally gavaged with Cy-3-G (50 mg/kg BW), PCA (5 mg/kg BW), Cy-3-G (50 mg/kg BW) plus PCA (5 mg/kg BW), or the vehicle (normal saline) once daily for 4 weeks. Four days before euthanasia, mice (n=12 per group) were intraperitoneally injected with thioglycollate to obtain MPMs for the analyses of total lipid content (A) and cholesterol content (B). C, Representative Oil-red-O staining of aortic sinus lesion and quantification of the aortic sinus lesion area. Area of green circles indicates atherosclerotic lesion. D, Total cholesterol content in the thoracic and abdominal aorta. Left in C is representative images of 12 independent experiments (n=12). The results in A, B, right of C, and D are mean±sem, n=12. *P<0.05 vs. baseline; #P<0.05. NS indicates not significant; Abx, antibiotics. E, Model depicting the gut microbiotaendependent mechanism of anthocyanin in the promotion of macrophage RCT and atherosclerosis regression in the ApoE-deficient mice. Finally, we determined whether the changes in RCT and macrophage foam cells induced by Cy-3-G, or PCA, or Cy- 3-G plus PCA are associated with atherosclerosis regression in ApoE / mice. As shown in Figure 7C and 7D, both the atherosclerotic lesion area in the aortic sinus and the cholesterol content in the whole aorta, as the indexes for the severity of atherosclerosis, were lower in Cy-3-Gentreated, PCA-treated, or Cy-3-G plus PCAentreated mice. However, antibiotics abolished the regressive effect of Cy-3-G but not PCA on established atherosclerosis (Figure 7C and 7D). Moreover, cotreatment with PCA reversed the suppressive effect of antibiotics on the Cy-3-Geninduced regressive effect (Figure 7C and 7D). Collectively, these data implied that the effect of Cy-3-G on the regression of atherosclerosis depends on its gut microbiota metabolite PCA. Important drivers of atherosclerotic lesion regression are plasma lipids and ApoA-I. The Table shows that Cy-3-G but not PCA reduced plasma total cholesterol (TC) and total triglyceride (TG) and increased HDL-C and ApoA-I levels. Evidently, antibiotics selectively abrogated the incremental effect of Cy-3-G on the plasma HDL-C and ApoA-I levels (Table). FPLC analysis of pooled plasma samples further showed that Cy-3-G especially decreased cholesterol contents in the fraction of VLDL (Online Figure IX). These data suggest that Cy-3-G alone is sufficient to decrease plasma TC and TG content, whereas PCA is a perquisite for its parent Cy-3-G to elevate HDL-C and ApoA-I levels.

12 978 Circulation Research September 28, 2012 Downloaded from by guest on May 8, 2018 Table 1. Plasma lipid profiles and ApoA-I levels in ApoE -/- mice from different experimental groups Groups n TC (mg/dl) TG (mg/dl) HDL-C (mg/dl) ApoA-I (mg/dl) Control ±43 117± ± ±0.6 Control+Abx ±55 129± ± ±0.8 Cy-3-G ±29* 69.6±3.1* 51.6±5.3* 21.8±1.1* Cy-3-G+Abx ±32* 74.4±4.2* 28.7± ±0.8 PCA ±44 125± ± ±0.9 PCA+Abx ±52 121± ± ±1.2 Cy-3-G+PCA ±37* 73.3±3.6* 70.2±6.2* 28.6±0.9* Cy-3-G+PCA+Abx ±43* 78.2±3.4* 55.7±3.5* 22.5±0.7* The data were the means±sem from the indicated numbers of male ApoE -/- mice in each group. *The means in the same column as an asterisk differ from the control group, P<0.05. Discussion The current study demonstrated that pure anthocyanin Cy-3-G accelerates macrophage RCT and promotes the regression of established atherosclerosis in the ApoE / mouse model. Utilizing antibiotics-treated and germ-free mice, we revealed that Cy-3-G is metabolized into PCA via the gut microbiota. Of note, we observed that these beneficial effects of Cy-3-G are at least mainly, if not completely, dependent on its gut microbiota metabolite PCA. We showed for the first time that mir-10b negatively regulates cholesterol efflux from murineand human-derived macrophages by repressing the expression of ABCA1 and ABCG1, whereas PCA reduces mir-10b levels, leading to enhanced cholesterol efflux from macrophages. Taken together, our data provided the unequivocal evidence that gut microbiota modulates the potency of Cy-3-G in promoting macrophage RCT and regressing atherosclerosis in vivo. At the same time, this finding also provided new insight that the gut microbiota is an attractive therapeutic target for regressing established atherosclerosis. Because anthocyanins are poorly absorbed in the gut, it has been doubtful that relatively low concentrations of intact anthocyanins in vivo could account for these notable beneficial effects. Recently, it has been reported that substantial amounts of unabsorbed anthocyanin undergo extensive metabolic transformation in the colon via the gut microbiota, producing a wide range of aromatic and phenolic compounds. 11 Several in vivo studies showed that one of these compounds, PCA, can be detected in the systemic circulation of rats and humans after the consumption of the parent compounds. 12,33 In addition to the gut bacterial transformation, several studies showed that anthocyanin could be spontaneously degraded into to PCA in vitro under physiological conditions at neutral ph such as intestinal content and cell culture mediums. 17 In the current study, we provided the evidence that PCA is a gut microbiota metabolite of Cy-3-G. PCA could not be detected in plasma in germ-free Kunming mice fed with Cy-3-G, or in Cy-3-Genfed ApoE / mice treated with antibiotics. These findings suggest that spontaneous degradation of anthocyanin into PCA has limited physiological relevance during Cy-3-G intervention. In addition to anthocyanin, many other flavonoids are also poorly absorbed and undergo extensive gut microbiota metabolism It is reasonable to expand our study into other flavonoids and to hypothesize that other gut microbiota metabolites of the parent flavonoids may also play important biological roles. We found that PCA but not Cy-3-G selectively induces ABCA1 and ABCG1 expression in an LXRα-independent manner. Instead, PCA regulates the cholesterol efflux from murine- and human-derived macrophages by posttranscriptional regulation of the 3 -UTR of ABCA1 or ABCG1 transcript through downregulating mir-10b level. We hence conclude that in addition to mir and mir-758, 23 mir- 10b is the third mirna that is involved in functionally regulating macrophage cholesterol efflux. It is worth noting that a recent study shows that mir-10b level is significantly increased in the atherosclerotic plaque tissue of human subjects, 37 suggesting that mir-10b is a novel therapeutic target for preventing atherosclerotic development. The observation that antagonism of mir-33 in LDL receptorendeficient mice markedly promotes macrophage RCT partially via increasing macrophage ABCA1 and ABCG1 expression 38 prompted us to hypothesize that mir-10b may function in a similar way as mir-33. Supporting our hypothesis, we found that PCA could promote macrophage RCT in mice and also alter expression of mir-10b and ABCA1/ABCG1 in isolated MPMs. Because the cause-effect relationship between mir-10b and ABCA1/ ABCG1 mediated by PCA is ambiguous, we performed a time-course experiment. The results revealed sequential changes: first, appearance of plasma PCA, then macrophage mir-10b level decreased; last, ABCA1/ABCG1 expression increased. These findings strongly imply that PCA induces macrophage ABCA1 and ABCG1 expression in vivo, at least in part, through downregulation of mir-10b. Interestingly, plasma PCA is rapidly eliminated from circulation, but its downstream effects of mir-10b and ABCA1/ABCG1 last for a much longer duration. This phenomenon may be explained by longer stability of mir-10b, which is consistent with the concept that the half-lives of mirnas usually ranged from 28 to 220 hours. 39 Furthermore, we demonstrated that the potency of PCA in accelerating in vivo macrophage RCT is reduced by 55.4% when macrophage ABCA1 and ABCG1 were knocked down. Because ABCA1 and ABCG1 in J774-DKD macrophages were not completely knocked down in our cell models, we speculate that the contribution of macrophage cholesterol efflux mediated by ABCA1 and ABCG1 to the PCA-induced macrophage RCT may have been underestimated. How PCA affects mir-10b expression in macrophages is currently unknown. Several groups have independently revealed that expression of mir-10b can be induced by the transcription factor Twist that directly binds to the putative promoter of mir-10b. 40 Twist is positively regulated by CD44, 41 which could be induced by tumor necrosis factor-α (TNF-α). 42 Given that the macrophages express CD44 and Twist, 42,43 and PCA can inhibit TNF-α production in lipopolysaccharide-induced macrophages, 44 we wonder whether PCA decreases mir-10b expression through the inhibition of the CD44-activated Twist pathway via decreasing TNF-α production. This is worth further examination.

13 Wang et al Cy-3-G Needs Gut Flora to Promote RCT Via MiR-10b 979 Downloaded from by guest on May 8, 2018 Although numerous flavonoid-rich foods, extracts rich in flavonoids, and individual flavonoids are able to inhibit the early stages of atherosclerosis in animal models, 36 no specific flavonoid has been shown to regress the advanced stages of atherosclerosis. We have found previously that anthocyanin extract from black rice (rich in Cy-3-G) significantly attenuated the progression of advanced plaque lesions in the ApoE / mouse model. 3 The current work revealed for the first time that an individual purified anthocyanin Cy-3-G promotes the regression of advanced plaque lesions. Mechanistically, we proposed that this effect is probably achieved with the aid of its gut microbiota metabolite PCA. It has been shown that Cy-3-G inhibits the formation of early atherosclerosis through several potential mechanisms: antioxidation, 6 antimonocyte infiltration, 13 and elevating plasma HDL-C levels. 3 However, it is very unlikely that those effects largely contribute to the regressive effect of Cy-3-G on established atherosclerosis. First, antioxidants therapies, such as vitamin E, vitamin C, or β-carotene supplementation, have no regressive effect on already established atherosclerosis in numerous preclinical and clinical studies. 45 Second, it is generally accepted that the antimonocyte infiltration by downregulation of CC chemokine receptor 2 expression could not inhibit the progression of established atherosclerosis in the same ApoE / mouse model. 46 Finally, in contrast to the recent finding that HDL could promote atherosclerosis regression partially by its antiinflammatory property, 47 the elevated plasma HDL-C levels by Cy-3-G treatment did not cause a more favorable effect on atherosclerosis regression than PCA treatment in present study. This paradoxical phenomenon indicated that HDL in the Cy-3-Gentreated mice may not be functional in regressing atherosclerotic lesion. Supporting this speculation, a previous study has demonstrated that HDL derived from ApoE / mice is proinflammatory. 48 Thus, the present findings may further support the concept that functional properties of HDL rather than mere HDL concentration faithfully reflect the atheroprotective potency of HDL. 19 A single mirna may target more than 100 mrnas, whereas a single mrna could be regulated by different mirnas. In addition to the promotion of macrophage RCT through posttranscriptional regulation of ABCA1 and ABCG1, PCA may promote lesion regression through other targets of mir-10b, such as Kruppel-like factor 4 (KLF-4) 49 and homeobox D10 (HOXD10). 50 KLF-4 was shown to play a critical role in regulating macrophage polarization from a proinflammatory to an anti-inflammatory functional phenotype. 51 HOXD10 represses several genes involved in cell migration and extracellular matrix remodeling. 50 If mir-10b functionally targets KLF4 and HOXD10 in macrophages or vascular cells such as endothelial cells, PCA may reduce atherosclerosis partially via inhibiting the mir-10b-inflammatory pathway and improving extracellular matrix remodeling in the atherosclerotic plaques besides its promotion of RCT. The gut microbiota forms an enormous and diverse ecosystem known to function in nutrient absorption, vitamin production, gut epithelial cell health, and innate immunity. 52 The gut microbiota has also been implicated in the development of several metabolic phenotypes, such as obesity, insulin resistance, and atherosclerosis Alterations in the composition of the intestinal ecosystem can influence the development of chronic disease by affecting the dominant bacteria and the production and absorption of endogenous endotoxins, which can trigger inflammation in obesity and type 2 diabetes. 57,58 In addition, nutritional substances undergo metabolism by the gut microbiota, and their metabolites subsequently induce various functions. 59 A recent study revealed the direct link between the gut microbiota, dietary phosphatidylcholine, and cardiovascular disease risk. 60 Furthermore, it was proposed that a gut microbiota metabolite of isoflavone, equol, is a critical factor contributing to the beneficial effects of its parent, isoflavone. 61 The present study revealed the importance of PCA as a gut microbiota metabolite of the parent anthocyanin, being substantially responsible for the antiatherosclerotic effects of dietary anthocyanin. These observations collectively indicate that intestinal microbiota plays a critical role in the development of chronic disease. Additionally, the intestinal microbiota ecosystem is a potential target for the prevention and treatment of chronic diseases. Cholesterol effluxed from macrophages is carried by plasma HDL in the removal of cellular cholesterol excess. The plasma concentration of effluxed cholesterol from macrophages is affected by many components of RCT, including the clearance rate by the liver. 62 A few recent studies evidenced that HDL also deliver cholesterol to intestinal tissue and that the cholesterol is then directly transported into intestinal lumen without metabolism in the liver. 63 Therefore, the plasma 3 H-tracer levels could not faithfully reflect the cholesterol efflux from macrophage. The present results showed that plasma and hepatic 3 H-tracer levels did not differ from the control group after PCA administration. It is possible that PCA promotes cholesterol transportation of HDL from intestinal tissue into gut lumen or reduces intestinal cholesterol and bile acid reabsorption, thus partially contributing to PCA-induced macrophage RCT. Nevertheless, the effect of PCA on cholesterol transport in intestinal tissue requires further investigation. In summary, we present the following model for the role of Cy-3-G in promoting RCT and regressing established atherosclerosis in the ApoE / mouse model (Figure 7E): dietary Cy-3-G is intensively metabolized into PCA by the gut microbiota, and the latter induces ABCA1 and ABCG1 expression in macrophages by decreasing the expression of mir-10b, which contributes to the accelerated macrophage RCT. As a consequence, enhanced macrophage RCT can, at least in part, result in the regression of established atherosclerosis. Importantly, the observations made in this study allow us to propose that gut microbiota is a potential and promising target for the promotion of macrophage RCT and hence atherosclerosis regression in humans. Acknowledgments We thank Benhua Zeng (Third Military Medical University, Chongqing, China) for handling germ-free animals. mirna microarray experiments were performed by KangChen Bio-tech, Shanghai, China. Sources of Funding This work was supported by grants from the National Natural Science Foundation of China ( ), the Joint project of NSFC, China- CIHR, Canada ( ), and the Medical Scientific Research Foundation of Guangdong Province (B ).

14 980 Circulation Research September 28, 2012 Downloaded from by guest on May 8, 2018 None. Disclosures References 1. Prior RL, Wu X. Anthocyanins: structural characteristics that result in unique metabolic patterns and biological activities. Free Radic Res. 2006;40: de Pascual-Teresa S, Moreno DA, Garcia-Viguera C. Flavanols and anthocyanins in cardiovascular health: a review of current evidence. Int J Mol Sci. 2010;11: Xia X, Ling W, Ma J, Xia M, Hou M, Wang Q, Zhu H, Tang Z. An anthocyanin-rich extract from black rice enhances atherosclerotic plaque stabilization in apolipoprotein E-deficient mice. J Nutr. 2006;136: Qin Y, Xia M, Ma J, Hao Y, Liu J, Mou H, Cao L, Ling W. Anthocyanin supplementation improves serum LDL- and HDL-cholesterol concentrations associated with the inhibition of cholesteryl ester transfer protein in dyslipidemic subjects. Am J Clin Nutr. 2009;90: Galvano F, Vitaglione P, Li Volti G, Di Giacomo C, Gazzolo D, Vanella L, La Fauci L, Fogliano V. Protocatechuic acid: the missing human cyanidins metabolite. Mol Nutr Food Res. 2008;52: Tsuda T. Dietary anthocyanin-rich plants: biochemical basis and recent progress in health benefits studies. Mol Nutr Food Res. 2012; 56: Xia M, Hou M, Zhu H, Ma J, Tang Z, Wang Q, Li Y, Chi D, Yu X, Zhao T, Han P, Xia X, Ling W. Anthocyanins induce cholesterol efflux from mouse peritoneal macrophages: the role of the peroxisome proliferatoractivated receptor γ-liver X receptor α-abca1 pathway. J Biol Chem. 2005;280: McGhie TK, Walton MC. The bioavailability and absorption of anthocyanins: towards a better understanding. Mol Nutr Food Res. 2007;51: Scalbert A, Williamson G. Dietary intake and bioavailability of polyphenols. J Nutr. 2000;130:2073S 2085S. 10. Kay CD. Aspects of anthocyanin absorption, metabolism and pharmacokinetics in humans. Nutr Res Rev. 2006;19: Williamson G, Clifford MN. Colonic metabolites of berry polyphenols: the missing link to biological activity? Br J Nutr. 2010;104:S48--S Vitaglione P, Donnarumma G, Napolitano A, Galvano F, Gallo A, Scalfi L, Fogliano V. Protocatechuic acid is the major human metabolite of cyanidin-glucosides. J Nutr. 2007;137: Wang D, Zou T, Yang Y, Yan X, Ling W. Cyanidin-3-O-beta-glucoside with the aid of its metabolite protocatechuic acid, reduces monocyte infiltration in apolipoprotein E-deficient mice. Biochem Pharmacol. 2011;82: Wang D, Wei X, Yan X, Jin T, Ling W. Protocatechuic acid, a metabolite of anthocyanins, inhibits monocyte adhesion and reduces atherosclerosis in apolipoprotein E-deficient mice. J Agric Food Chem. 2010;58: Keppler K, Humpf HU. Metabolism of anthocyanins and their phenolic degradation products by the intestinal microflora. Bioorg Med Chem. 2005;13: Aura AM, Martin-Lopez P, O Leary KA, Williamson G, Oksman- Caldentey KM, Poutanen K, Santos-Buelga C. In vitro metabolism of anthocyanins by human gut microflora. Eur J Nutr. 2005;44: Kay CD, Kroon PA, Cassidy A. The bioactivity of dietary anthocyanins is likely to be mediated by their degradation products. Mol Nutr Food Res. 2009;53:S92--S Rader DJ. Molecular regulation of HDL metabolism and function: implications for novel therapies. J Clin Invest. 2006;116: degoma EM, degoma RL, Rader DJ. Beyond high-density lipoprotein cholesterol levels evaluating high-density lipoprotein function as influenced by novel therapeutic approaches. J Am Coll Cardiol. 2008;51: Wang X, Collins HL, Ranalletta M, Fuki IV, Billheimer JT, Rothblat GH, Tall AR, Rader DJ. Macrophage ABCA1 and ABCG1, but not SR-BI, promote macrophage reverse cholesterol transport in vivo. J Clin Invest. 2007;117: Bartel DP. MicroRNAs: genomics, biogenesis, mechanism, and function. Cell. 2004;116: Rayner KJ, Suarez Y, Davalos A, Parathath S, Fitzgerald ML, Tamehiro N, Fisher EA, Moore KJ, Fernandez-Hernando C. MiR-33 contributes to the regulation of cholesterol homeostasis. Science. 2010; 328: Ramirez CM, Davalos A, Goedeke L, Salerno AG, Warrier N, Cirera- Salinas D, Suarez Y, Fernandez-Hernando C. MicroRNA-758 regulates cholesterol efflux through posttranscriptional repression of ATP-binding cassette transporter A1. Arterioscler Thromb Vasc Biol. 2011;31: Wang Q, Xia M, Liu C, Guo H, Ye Q, Hu Y, Zhang Y, Hou M, Zhu H, Ma J, Ling W. Cyanidin-3-O-beta-glucoside inhibits inos and COX-2 expression by inducing liver X receptor alpha activation in THP-1 macrophages. Life Sci. 2008;83: Li D, Wang D, Wang Y, Ling W, Feng X, Xia M. Adenosine monophosphate activated protein kinase induces cholesterol efflux from macrophage-derived foam cells and alleviates atherosclerosis in apolipoprotein E-deficient mice. J Biol Chem. 2010;285: Zhang Y, Zanotti I, Reilly MP, Glick JM, Rothblat GH, Rader DJ. Overexpression of apolipoprotein A-I promotes reverse transport of cholesterol from macrophages to feces in vivo. Circulation. 2003;108: Guo H, Ling W, Wang Q, Liu C, Hu Y, Xia M. Cyanidin 3-glucoside protects 3T3-L1 adipocytes against H2O2- or TNF-alpha-induced insulin resistance by inhibiting c-jun NH2-terminal kinase activation. Biochem Pharmacol. 2008;75: Tontonoz P, Mangelsdorf DJ. Liver X receptor signaling pathways in cardiovascular disease. Mol Endocrinol. 2003;17: Bartel DP. MicroRNAs: target recognition and regulatory functions. Cell. 2009;136: Li Y, Kong D, Wang Z, Sarkar FH. Regulation of micrornas by natural agents: an emerging field in chemoprevention and chemotherapy research. Pharm Res. 2010;27: Marin RM, Vanicek J. Efficient use of accessibility in microrna target prediction. Nucleic Acids Res. 2011;39: Cuchel M, Rader DJ. Macrophage reverse cholesterol transport: key to the regression of atherosclerosis? Circulation. 2006;113: Tsuda T, Horio F, Osawa T. Absorption and metabolism of cyanidin 3-O-beta-D-glucoside in rats. FEBS Lett. 1999;449: Selma MV, Espin JC, Tomas-Barberan FA. Interaction between phenolics and gut microbiota: role in human health. J Agric Food Chem. 2009;57: van Duynhoven J, Vaughan EE, Jacobs DM, Kemperman RA, van Velzen EJ, Gross G, Roger LC, Possemiers S, Smilde AK, Dore J, Westerhuis JA, Van de Wiele T. Metabolic fate of polyphenols in the human superorganism. Proc Natl Acad Sci U S A. 2011;108: Manach C, Mazur A, Scalbert A. Polyphenols and prevention of cardiovascular diseases. Curr Opin Lipidol. 2005;16: Bidzhekov K, Gan L, Denecke B, Rostalsky A, Hristov M, Koeppel TA, Zernecke A, Weber C. microrna expression signatures and parallels between monocyte subsets and atherosclerotic plaque in humans. Thromb Haemost. 2012;107: Rayner KJ, Sheedy FJ, Esau CC, Hussain FN, Temel RE, Parathath S, van Gils JM, Rayner AJ, Chang AN, Suarez Y, Fernandez-Hernando C, Fisher EA, Moore KJ. Antagonism of mir-33 in mice promotes reverse cholesterol transport and regression of atherosclerosis. J Clin Invest. 2011;121: Gantier MP, McCoy CE, Rusinova I, Saulep D, Wang D, Xu D, Irving AT, Behlke MA, Hertzog PJ, Mackay F, Williams BR. Analysis of microrna turnover in mammalian cells following Dicer1 ablation. Nucleic Acids Res. 2011;39: Ma L. Role of mir-10b in breast cancer metastasis. Breast Cancer Res. 2010;12: Bourguignon LY, Wong G, Earle C, Krueger K, Spevak CC. Hyaluronan- CD44 interaction promotes c-src-mediated twist signaling, microrna-10b expression, and RhoA/RhoC upregulation, leading to Rho-kinase-associated cytoskeleton activation and breast tumor cell invasion. J Biol Chem. 2010;285: Maiti A, Maki G, Johnson P. TNF-alpha induction of CD44-mediated leukocyte adhesion by sulfation. Science. 1998;282: Pozharskaya V, Torres-Gonzalez E, Rojas M, Gal A, Amin M, Dollard S, Roman J, Stecenko AA, Mora AL. Twist: a regulator of epithelialmesenchymal transition in lung fibrosis. PLoS One. 2009;4:e Min SW, Ryu SN, Kim DH. Anti-inflammatory effects of black rice, cyanidin-3-o-beta-d-glycoside, and its metabolites, cyanidin and protocatechuic acid. Int Immunopharmacol. 2010;10: Brown BG, Cheung MC, Lee AC, Zhao XQ, Chait A. Antioxidant vitamins and lipid therapy: end of a long romance? Arterioscler Thromb Vasc Biol. 2002;22:

15 Wang et al Cy-3-G Needs Gut Flora to Promote RCT Via MiR-10b 981 Downloaded from by guest on May 8, Guo J, de Waard V, Van Eck M, Hildebrand RB, van Wanrooij EJ, Kuiper J, Maeda N, Benson GM, Groot PH, Van Berkel TJ. Repopulation of apolipoprotein E knockout mice with CCR2-deficient bone marrow progenitor cells does not inhibit ongoing atherosclerotic lesion development. Arterioscler Thromb Vasc Biol. 2005;25: Feig JE, Rong JX, Shamir R, Sanson M, Vengrenyuk Y, Liu J, Rayner K, Moore K, Garabedian M, Fisher EA. HDL promotes rapid atherosclerosis regression in mice and alters inflammatory properties of plaque monocyte-derived cells. Proc Natl Acad Sci U S A. 2011;108: Navab M, Anantharamaiah GM, Reddy ST, Hama S, Hough G, Frank JS, Grijalva VR, Ganesh VK, Mishra VK, Palgunachari MN, Fogelman AM. Oral small peptides render HDL antiinflammatory in mice and monkeys and reduce atherosclerosis in ApoE null mice. Circ Res. 2005;97: Tian Y, Luo A, Cai Y, Su Q, Ding F, Chen H, Liu Z. MicroRNA-10b promotes migration and invasion through KLF4 in human esophageal cancer cell lines. J Biol Chem. 2010;285: Ma L, Teruya-Feldstein J, Weinberg RA. Tumour invasion and metastasis initiated by microrna-10b in breast cancer. Nature. 2007;449: Liao X, Sharma N, Kapadia F, et al. Kruppel-like factor 4 regulates macrophage polarization. J Clin Invest. 2011;121: Kau AL, Ahern PP, Griffin NW, Goodman AL, Gordon JI. Human nutrition, the gut microbiome and the immune system. Nature. 2011;474: Backhed F, Ding H, Wang T, Hooper LV, Koh GY, Nagy A, Semenkovich CF, Gordon JI. The gut microbiota as an environmental factor that regulates fat storage. Proc Natl Acad Sci U S A. 2004;101: Turnbaugh PJ, Ley RE, Mahowald MA, Magrini V, Mardis ER, Gordon JI. An obesity-associated gut microbiome with increased capacity for energy harvest. Nature. 2006;444: What Is Known? Anthocyanins, similar to other flavonoids, have been generally recognized as cardioprotective compounds despite their very limited bioavailability. Protocatechuic acid (PCA), one metabolite of cyanidin-3-o-βglucoside (Cy-3-G), has been shown to inhibit the development of early atherosclerosis. mirnas are known to regulate macrophage cholesterol efflux via altering expression of their target genes ATP-binding cassette transporter A1 (ABCA1) and G1 (ABCG1), both of which facilitate cellular cholesterol efflux. What New Information Does This Article Contribute? PCA, a gut microbiota metabolite of Cy-3-G, mediates the stimulatory effects of its precursor Cy-3-G on macrophage reverse cholesterol transport (RCT) and established atherosclerosis regression. mirna-10b negatively regulates macrophage cholesterol efflux via altering its target genes ABCA1 and ABCG1 expression, whereas PCA reduces mirna-10b levels, the net effect being enhanced macrophage cholesterol efflux and atherosclerosis regression. Novelty and Significance 55. Cani PD, Amar J, Iglesias MA, et al. Metabolic endotoxemia initiates obesity and insulin resistance. Diabetes. 2007;56: Caesar R, Fak F, Backhed F. Effects of gut microbiota on obesity and atherosclerosis via modulation of inflammation and lipid metabolism. J Intern Med. 2010;268: Burcelin R, Luche E, Serino M, Amar J. The gut microbiota ecology: a new opportunity for the treatment of metabolic diseases? Front Biosci. 2009;14: Geurts L, Lazarevic V, Derrien M, Everard A, Van Roye M, Knauf C, Valet P, Girard M, Muccioli GG, Francois P, de Vos WM, Schrenzel J, Delzenne NM, Cani PD. Altered gut microbiota and endocannabinoid system tone in obese and diabetic leptin-resistant mice: impact on apelin regulation in adipose tissue. Front Microbiol. 2011;2: Blaut M, Clavel T. Metabolic diversity of the intestinal microbiota: implications for health and disease. J Nutr. 2007;137:751S--755S. 60. Wang Z, Klipfell E, Bennett BJ, et al. Gut flora metabolism of phosphatidylcholine promotes cardiovascular disease. Nature. 2011;472: Setchell KD, Brown NM, Lydeking-Olsen E. The clinical importance of the metabolite equol-a clue to the effectiveness of soy and its isoflavones. J Nutr. 2002;132: McGillicuddy FC, de la Llera Moya M, Hinkle CC, Joshi MR, Chiquoine EH, Billheimer JT, Rothblat GH, Reilly MP. Inflammation impairs reverse cholesterol transport in vivo. Circulation. 2009;119: Temel RE, Sawyer JK, Yu L, Lord C, Degirolamo C, McDaniel A, Marshall S, Wang N, Shah R, Rudel LL, Brown JM. Biliary sterol secretion is not required for macrophage reverse cholesterol transport. Cell Metab. 2010;12: The relatively low concentrations of intact anthocyanins in vivo have cast doubt about their antiatherogenic effect. Because anthocyanins, similar to other flavonoids, are intensively metabolized by gut microbiota, we tested the protective effects of PCA on atherosclerosis. We showed that the stimulatory effects of anthocyanin Cy-3-G on macrophage RCT and established atherosclerosis regression were mediated by PCA. We also showed that mirna-10b negatively regulates cholesterol efflux from murine- and human-derived macrophages by repressing expression of ABCA1 and ABCG1, whereas PCA reduces mir-10b levels, which together contribute to accelerated cholesterol efflux from macrophages and established atherosclerosis regression. The discovery of a relationship between gut microbiotaendependent metabolism of dietary anthocyanin and regression of atherosclerosis provides for the opportunity for the utilization of gut microbiota in the treatment of atherosclerosis. Moreover, our novel findings also support the notion that mirna-10b might be an attractive candidate for promotion macrophage RCT and established atherosclerosis regression.

16 Downloaded from by guest on May 8, 2018 Gut Microbiota Metabolism of Anthocyanin Promotes Reverse Cholesterol Transport in Mice Via Repressing mirna-10b Dongliang Wang, Min Xia, Xiao Yan, Dan Li, Lei Wang, Yuxuan Xu, Tianru Jin and Wenhua Ling Circ Res. 2012;111: ; originally published online July 19, 2012; doi: /CIRCRESAHA Circulation Research is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX Copyright 2012 American Heart Association, Inc. All rights reserved. Print ISSN: Online ISSN: The online version of this article, along with updated information and services, is located on the World Wide Web at: Data Supplement (unedited) at: Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published in Circulation Research can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office. Once the online version of the published article for which permission is being requested is located, click Request Permissions in the middle column of the Web page under Services. Further information about this process is available in the Permissions and Rights Question and Answer document. Reprints: Information about reprints can be found online at: Subscriptions: Information about subscribing to Circulation Research is online at:

17 1 Supplemental Material Gut microbiota metabolism of anthocyanin promotes reverse cholesterol transport in mice via repressing mirna-10b Dongliang Wang 1, Min Xia 1, Xiao Yan 1, Dan Li 1, Lei Wang 1, Yuxuan Xu 1, Tianru Jin 1,2, and Wenhua Ling 1,3* SUPPLEMENTAL MATERIALS AND METHODS Gut Microbiota Metabolism of Cy-3-G 1) Ten-week-old male apolipoprotein E-deficient (ApoE -/- ) mice in the C57BL/6 genetic background (Jackson Laboratories, Sacramento, CA) that had been fasted overnight were orally gavaged once with cyanidin-3-o- -glucoside (Cy-3-G) (5 or 50 mg/kg BW) or the vehicle (normal saline). Blood samples were obtained from the retro-orbital venous plexus before as well as 0.5 and 2 h after Cy-3-G admission. Next, an antibiotic cocktail (0.5 g/l vancomycin, 1 g/l neomycin sulfate, 1 g/l metronidazole, 1 g/l ampicillin) was administered via the drinking water to deplete commensal microbiota. 1 Three weeks later, the metabolic experiment with Cy-3-G was performed again. The mice were then placed in conventional cages with non-sterile ApoE -/- female mice to facilitate the transfer of commensal organisms. Four weeks later, the conventionalized mice received Cy-3-G gavage again, followed by blood sample collection. 2) Ten-week-old male germ-free Kunming mice (Third Military Medical University, Chongqing, China) were gavaged once with Cy-3-G (50 mg/kg BW) or the vehicle. Blood samples were then collected successively, as indicated in the figure legends. Next, the mice were placed in conventional cages with non-sterile Kunming female mice to facilitate the transfer of commensal organisms. Four weeks later, the conventionalized mice underwent another round of Cy-3-G administration, followed by blood sample collection. The concentrations of Cy-3-G and protocatechuic acid (PCA) in plasma were quantified by high performance liquid chromatography-tandem mass spectrometry (HPLC/MS/MS) as previously described. 2 Briefly, an Agilent 1200 series high-performance liquid chromatography coupled to an Agilent 6410 triple quadrupole mass spectrometer and an Agilent Zorbax SB-C18 column (2.1 mm 50 mm, 1.8 m) were used to quantify the levels of plasma Cy-3-G and PCA. The solvents were A (5% formic acid in water, v/v) and solvent B (acetonitrile), the gradient elution program was performed as follows: a linear gradient of 2-8% B from 1 to 3 min, followed by a 7 min linear gradient to 10% B, a 5 min linear gradient to the initial condition, and holding for 5 min. Flow rate was 0.2 ml/min, and injection volume was 5 l. The MS/MS detection was performed by acquiring data in positive ion mode for Cy-3-G and in negative ion mode for PCA. Electrospray ionisation (ESI) was performed with the following spray chamber conditions: drying gas flow of 8 L/min, nebulizer pressure of 40 psi, and drying gas temperature of 350 C, applying a voltage of 5000 V. Plasma Cy-3-G and PCA levels were quantified by calibration curves obtained with the corresponding standard substances. Plasma Lipoprotein Preparation Human low-density lipoprotein (LDL) (density g/ml) and high-density lipoprotein (HDL) (density g/ml) were isolated by preparative ultracentrifugation

18 2 from normolipidemic human plasma and stored in PBS containing 1 mmol/l EDTA. Acetylation of LDL was performed by the addition of 4 aliquots each of 1 l of acetic anhydride at 10-min intervals to 2 mg of LDL in 600 l of ice-cold 50% saturated sodium acetate. LDL was aggregated by vortexing a 1 mg/ml solution at low speed for 15 s. The acetylated LDL (AcLDL) was dialyzed against PBS overnight, and the protein concentration was quantified using a BCA protein assay kit (Pierce, Rockford, IL). Western Blotting Analysis Mouse peritoneal macrophages (MPMs) were rinsed once with PBS and lysed with ice cold whole-cell lysis buffer [50 mmol/l Tris-HCl (ph 7.4), 150 mmol/l NaCl, 0.25% sodium deoxycholate, 1% NP-40, 1 mmol/l EGTA, 1 mmol/l Na 3 VO 4, 1 mmol/l NaF, 1 mmol/l PMSF, 1 mmol/l DTT, 1 g/ml leupeptin, 1 g/ml aprotinin and 1 g/ml pepstatin] for 30 min. The collections were centrifuged at 10,000 g for 15 min at 4 C and supernatants were retained as total cell lysates. Nuclear extracts were prepared with both a hypotonic buffer and high salt buffer. The hypotonic buffer contained 20 mmol/l Hepes (ph 7.9), 20 mmol/l NaF, 1 mmol/l Na 4 P 2 O 7, 1 mmol/l EDTA, 1 mmol/l Na 3 VO 4, 1 mmol/l DTT, 0.5 mmol/l PMSF, 1 g/ml leupeptin, 1 g/ml aprotinin and 1 g/ml pepstatin. The high salt buffer contained 20 mmol/l Hepes (ph 7.9), 20 mmol/l NaF, 1 mmol/l Na 4 P 2 O 7, 1 mmol/l EDTA, 420 mmol/l NaCl, 10% glycerol, 1 mmol/l Na 3 VO 4, 1 mmol/l DTT, 0.5 mmol/l PMSF, 1 g/ml leupeptin, 1 g/ml aprotinin and 1 g/ml pepstatin. Briefly, cells were rinsed once with PBS, scraped into an Eppendorf tube and spun for 15 s at 4 C. The cell pellets were dispersed into five volumes of hypotonic buffer to swell for 10 min on ice. Nuclei were precipitated by centrifugation at 12,000 g for 20 s at 4 C. The precipitations were resuspended in the high salt buffer, and incubated on ice for 30 min. The supernatants were retained as nuclear protein extraction after another centrifugation at 10,000 g for 15 min at 4 C. Protein concentration was determined using a BCA protein assay kit. Protein (~30 g total protein) for each sample was subjected to SDS-PAGE, followed by transfer to polyvinylidene difluoride membranes (Millipore, Bedford, MA). Membranes were immunoblotted with primary antibody and the appropriate horseradish peroxidase-conjugated secondary antibody (Santa Cruz Biotechnology, Santa Cruz, CA). Immunoblots blots were visualized by the enhanced chemiluminescence technique (PerkinElmer, Wellesley, MA). Primary antibodies include ATP-binding cassette transporter A1 (ABCA1) (cat# sc-53482, Santa Cruz Biotechnology), ATP-binding cassette transporter G1 (ABCG1) (cat# ab36969, Abcam, Cambridge, UK), scavenger receptor class B type I (SR-BI) (cat# NB , Novus Biologicals, Littleton, CO), liver X receptor alpha (LXR ) (cat# sc-1000, Santa Cruz Biotechnology), histone H1 (cat# sc-8615, Santa Cruz Biotechnology) and -actin (cat# 4967, Cell Signaling Technology, Danvers, MA). Construction of Promoter-Luciferase Vector and Transfection The human (h) ABCA1 or ABCG1 promoter was generated by PCR against human genomic DNA with the following primers. For habca1 promoter (from -940 to +110): forward, 5'-CCCAAGCTTAAGTTGGAGGTCTGGAGTGG-3'; backward, 5'-CCCAAGCTTACCGGCTCTGTTGGTGCGCGG-3'. For habcg1 promoter (from to + 37): forward, 5'-AAGGGGTACCATGAATGAAAGAAGCCAGACACAAA-3'; backward, 5'-CCTATCCGCTCGAGCACAAACATAGGTAGTCCAGCTGC-3'. After verification by DNA sequencing, the PCR product was digested by XhoI and HindIII

19 3 followed by ligation with pgl4 luciferase reporter vector (Promega, Madison, WI), transformed, and amplified. For promoter activity assay, AcLDL-loaded MPMs or THP-1 macrophages in 24-well plates were co-transfected with habca1 or habcg1 promoter and Renilla (for internal normalization) using Lipofectamine 2000 (Invitrogen, Carlsbad, CA) for 6 h. Cells were then treated with Cy-3-G or PCA at the indicated concentrations, or its vehicle, dimethyl sulfoxide (DMSO) for 12 h. The cells were harvested for measuring firefly and Renilla luciferases using the Dual-Luciferase Reporter Assay System (Promega). In Vitro Cholesterol Efflux Assays The method for in vitro assaying cholesterol efflux from macrophages has been described previously. 3 Briefly, macrophages were labeled with 3 H-cholesterol (PerkinElmer, Boston, MA) (1.0 Ci/mL) in RPMI 1640 medium supplemented with or without 25 g/ml AcLDL for 24 h. The cells were then incubated in medium containing 0.1% bovine serum albumin (BSA) in the presence or absence of Cy-3-G or PCA at the indicated concentrations for 24 h. Next, the cells were washed twice with PBS, and cholesterol efflux was initiated by medium containing 0.1% BSA in the presence or absence of ApoA-I (Sigma-Aldrich, St Louis, MO) (10 g/ml) or HDL (100 g/ml). The supernatants were collected after 24 h and expressed as a percentage of total cell 3 H-cholesterol content (total effluxed 3 H-cholesterol + cell-associated 3 H-cholesterol). ABCA1 and ABCG1 mrna Stability Assay For mrna stability measurement, actinomycin D (ActD) (Sigma-Aldrich) was used to block gene transcription. MPMs and THP-1 macrophages preloaded with AcLDL (25 g/ml) for 24 h were treated with Cy-3-G or PCA (1 µmol/l), or the vehicle (DMSO) for another 24 h. The cells were then treated with ActD (5 µg/ml), and the total RNA was harvested 0, 0.5, 1.0, 1.5, 2.0, 2.5, and 3.0 h after ActD treatment. ABCA1 and ABCG1 mrna levels at each time point were normalized to the corresponding level of the GAPDH transcript. The remaining mrna was determined by comparison with the expression level of the relevant gene at the zero time point (designated 100%) when ActD was added. GraphPad Prism software version 5.0 was used to calculate the transcript half-lives based on a one-phase exponential decay model. In Vivo Macrophage RCT Assay Macrophage RCT in vivo was measured as previously described, 4 using MPMs from adult C57BL/6J mice or J774 macrophages in the presence or absence of ABCA1 and ABCG1 as cholesterol donors. Briefly, macrophages were incubated with 5 Ci/mL 3 H-cholesterol and 25 g/ml AcLDL for 48 h. After a 6-h equilibration period in a medium containing 0.2% BSA, the 3 H-cholesterol-labeled MPMs were spun down and resuspended in ice-cold medium following intraperitoneal injection into the ApoE -/- mice. Blood was collected at 6, 24, and 48 h, and an aliquot of heparinized plasma was counted in a liquid scintillation counter (LSC) (Packard, Ramsey, MN). At 48 h after injection, mice were exsanguinated and perfused with ice-cold PBS. Liver lipids were extracted according to the method of Bligh and Dyer. 5 Briefly, a 50-mg piece of tissue was homogenized in water, and then lipids were extracted with a mixture of chloroform/methanol 2:1 (vol/vol). The lipid layer was collected, evaporated, resuspended in toluene, and counted in a LSC. Total feces collected from 0 to 48 h were vacuum dried and powdered, and fecal total sterols including neutral sterols and bile acids were then extracted as described by Batta et al. 6 Briefly, 15 mg fecal samples were digested

20 4 with 1 ml of 1 N sodium hydroxide for 1 h at 90 C. After cooling, the product was diluted with water (5 ml), and neutral sterols were repeatedly extracted with n-hexane (4 3 ml). The extracts were pooled, evaporated, re-suspended in toluene, and then counted in a LSC. To extract bile acids, the remaining aqueous layer after extraction of neutral sterols was acidified with concentrated HCl followed by extraction with ethyl acetate (4 3 ml). The extracts were pooled, evaporated, re-suspended in ethyl acetate, and counted in a LSC. The amounts of radioactivity per milliliter of plasma, in the lipids of the whole liver and in the total fecal sterols were expressed as a percentage of total injected 3 H-cholesterol. 4 Cellular Lipid Measurement MPMs were fixed with 4% paraformaldehyde and then stained by 0.5% Oil-red-O. The density of the total cellular lipid content was then measured with a microplate reader (absorbance at 540 nm, BioTek Instruments, Inc.) after alcohol extraction. 7 Total lipid in the isolated MPMs was extracted by hexane/isopropanol (3/2, vol/vol). Lipid extracts was dried under nitrogen flush, resuspended in a solution with 90% isopropanol and 10% Triton X-100. Total cholesterol content was measured enzymatically. 3 Plasma Lipid Profiles and ApoA-I Analysis Plasma total triglyceride (TG) and cholesterol (TC), and HDL-C were measured using commercial kits (BioSino Biotechnology Company, Ltd., Beijing, China). Plasma apolipoprotein A-I (ApoA-I) levels were quantified by a mouse ApoA-I enzyme-linked immunosorbent kit (Cusabio Biotech Co., Ltd, Wuhan, China). Pooled plasma samples were subjected to fast protein liquid chromatography (FPLC) gel filtration employing a Superose 6B column (Pharmacia Biotech) as previously described. 8 In brief, plasma from mice was subjected to FPLC analysis using a Superose 6 column on an HPLC system model 600 (Waters Chromatography). A 100 l aliquot of plasma was injected onto the column and separated with a buffer containing 0.15 mol/l NaCl, 0.01 mol/l Na 2 HPO 4, 0.1 mmol/l EDTA (ph 7.4), at a flow rate of 0.5 ml/min. Forty 0.5 ml fractions were collected, and tubes were analysed for cholesterol as described above. Fractions contain VLDL and chylomicra; fractions contain LDL and IDL; fractions contain HDL; and fractions contain non-lipoprotein-associated proteins. Quantification of Atherosclerosis The method for measuring atherosclerotic lesions at the aortic sinus was described previously. 9 Briefly, upper sections of the hearts were embedded in OCT compound (Sigma-Aldrich) and frozen at -20 C. Sections (10 m) were collected, beginning at the aortic root and extending for 400 m, according to the method of Paigen et al. 10 Lesions from 15 alternating sections were stained with Oil-Red-O and quantified using the Optimas Image Analysis software package (Bioscan Inc., Edmonds, WA). Total cholesterol contents in the mouse aortas were determined by enzymatic assays. 3

21 SUPPLEMENTAL REFERENCES 1. Rakoff-Nahoum S, Paglino J, Eslami-Varzaneh F, Edberg S, Medzhitov R. Recognition of commensal microflora by toll-like receptors is required for intestinal homeostasis. Cell. 2004;118: Wang D, Zou T, Yang Y, Yan X, Ling W. Cyanidin-3-O-beta-glucoside with the aid of its metabolite protocatechuic acid, reduces monocyte infiltration in apolipoprotein E-deficient mice. Biochem Pharmacol. 2011;82: Xia M, Hou M, Zhu H, Ma J, Tang Z, Wang Q, Li Y, Chi D, Yu X, Zhao T, Han P, Xia X, Ling W. Anthocyanins induce cholesterol efflux from mouse peritoneal macrophages: the role of the peroxisome proliferator-activated receptor {gamma}-liver X receptor {alpha}-abca1 pathway. J Biol Chem. 2005;280: Zhang Y, Zanotti I, Reilly MP, Glick JM, Rothblat GH, Rader DJ. Overexpression of apolipoprotein A-I promotes reverse transport of cholesterol from macrophages to feces in vivo. Circulation. 2003;108: Iverson SJ, Lang SL, Cooper MH. Comparison of the Bligh and Dyer and Folch methods for total lipid determination in a broad range of marine tissue. Lipids. 2001;36: Batta AK, Salen G, Rapole KR, Batta M, Batta P, Alberts D, Earnest D. Highly simplified method for gas-liquid chromatographic quantitation of bile acids and sterols in human stool. J Lipid Res. 1999;40: Lu KY, Ching LC, Su KH, Yu YB, Kou YR, Hsiao SH, Huang YC, Chen CY, Cheng LC, Pan CC, Lee TS. Erythropoietin suppresses the formation of macrophage foam cells: role of liver X receptor alpha. Circulation. 2010;121: Li D, Wang D, Wang Y, Ling W, Feng X, Xia M. Adenosine Monophosphate Activated Protein Kinase Induces Cholesterol Efflux from Macrophage-Derived Foam Cells and Alleviates Atherosclerosis in Apolipoprotein E-Deficient Mice. J Biol Chem. 2010;285: Wang D, Wei X, Yan X, Jin T, Ling W. Protocatechuic acid, a metabolite of anthocyanins, inhibits monocyte adhesion and reduces atherosclerosis in apolipoprotein E-deficient mice. J Agric Food Chem. 2010;58: Paigen B, Morrow A, Holmes PA, Mitchell D, Williams RA. Quantitative assessment of atherosclerotic lesions in mice. Atherosclerosis. 1987;68:

22 6 SUPPLEMENTAL TABLE Online Table I. Sequences of primers for real-time quantitative PCR Gene Species Forward primer (5' to 3' ) Reverse primer (5' to 3' ) ABCA1 mouse GGACTTGCCTTGTTCCGAGAG GCTGCCACATAACTGATAGCGA ABCA1 human ACCCACCCTATGAACAACATGA GAGTCGGGTAACGGAAACAGG ABCG1 mouse CTCCATCGTCTGTACCATCC CTCCATCGTCTGTACCATCC ABCG1 human ATTCAGGGACCTTTCCTATTCGG CTCACCACTATTGAACTTCCCG SR-BI mouse GGCTGCTGTTTGCTGCG GCTGCTTGATGAGGGAGGG SR-BI human ACTTCTGGCATTCCGATCAGT ACGAAGCGATAGGTGGGGAT LXR mouse GAGCCGACAGAGCTTCGTC GCGTGCTCCCTTGATGACA LXR human ACACCTACATGCGTCGCAAG GACGAGCTTCTCGATCATGCC GAPDH mouse GGCAGCTTCGGCACATATTTC CCAGGGTTATAGTCCTTCTCGT GAPDH human CTCACCGGATGCACCAATGTT CGCGTTGCTCACAATGTTCAT

23 7 Online Table II. Relative mirna expression in PCA treated vs. control macrophages mirna ID Fold change mmu-mir-10b mmu-mir mmu-mir mmu-mir mmu-mir mmu-mir MPMs preloaded with AcLDL were treated with PCA (1 mol/l) or the vehicle (DMSO), and the relative mirna expression was analyzed by qrt-pcr mirna PCR Array. The data are the means of three independent experiments.

24 SUPPLEMENTAL FIGURES AND FIGURE LEGENDS Online Figure I Mouse gut microbiota metabolizes Cy-3-G into PCA. (A) Plasma PCA and Cy-3-G concentrations were measured before, 0.5 and 2 h after oral Cy-3-G administration (50 mg/kg BW) in the germ-free Kunming mice. (B) The measurements were then made following the acquisition of normal gut microbiota via the placement of the mice into conventional cages (4 weeks) with non-sterile mice. The results are the means ± SEM. n = 10.

25 Online Figure II Cy-3-G and PCA cannot alter the promoter activity of human ABCA1 and ABCG1 in macrophages. AcLDL-loaded MPMs (A) and THP-1 macrophages (B) co-transfected with human ABCA1 or ABCG1 promoter and Renilla (for internal normalization) were treated with Cy-3-G, PCA, or the vehicle (DMSO) for 12 h. Cellular lysate was then used to determine the activity of firefly and Renilla luciferases. The results are the means ± SEM of fold of the Control from three independent experiments. n = 3. 9

26 10 Online Figure III Cy-3-G and PCA cannot alter LXR expression and activation in macrophages. AcLDL-loaded MPMs and THP-1 macrophages were treated with Cy-3-G, PCA, or the vehicle (DMSO) for 24 h. (A to D) LXR mrna and protein expression in nuclear extract in AcLDL-loaded MPMs (A and C) and THP-1 macrophages (B and D) were assessed by qrt-pcr and Western blotting, respectively. Panel C and D are representative images of three independent experiments. For panels A and B, results are the means ± SEM of fold of the Control (n = 6 per group). (E and F) AcLDL-loaded MPMs (E) and THP-1 macrophages (F) were transfected with a luciferase reporter plasmid containing liver X receptor element upstream of the thymidine kinase promoter (LXRE-tk-Luc) (kindly provided by David J. Mangelsdorf, University of Texas Southwestern Medical Center) in the presence of -galactosidase ( -gal) as a reference plasmid. Twelve hour after transfection, cells were treated with Cy-3-G, PCA, or DMSO for 24 h. Luciferase and -gal activities were then determined in the cell lysates. The values are the means ± SEM of fold of the Control. n = 6.

27 11 Online Figure IV PCA but not Cy-3-G reduces mir-10b expression in macrophages. AcLDL-loaded J774 murine macrophages were treated with Cy-3-G (1 mol/l), PCA (1 mol/l), or the vehicle (DMSO) for 24 h. Total RNA was then isolated and mirna expression was determined by qrt-pcr. The results are the means ± SEM of fold of the Control. n = 3-5. *, P < 0.05 versus Control.

28 Online Figure V Effect of mir-300, 370 or 493 on cholesterol efflux from MPMs. MPMs preloaded with 3 H-cholesterol and AcLDL, as described in the Methods, were transfected with either (A) control inhibitor (Con Inh) or anti-mir-300, or (B and C) control mir (Con mir) or mir-370 (B), or 493 (C) for 48 h. Cholesterol efflux was then initiated to ApoA-I or HDL for 24 h. The results are the means ± SEM. n =

29 Online Figure VI Plasma PCA concentrations after oral gavage of PCA in mice. Twelve 10-week-old ApoE -/- mice fasted overnight were once orally gavaged with PCA (2.5, 5, or 10 mg/kg BW), and blood samples were then collected successively from the retro-orbital venous plexus as indicated. Plasma PCA levels were determined as described in Methods. The results are the means ± SEM. n =

30 Online Figure VII Time course of PCA-induced changes in mir-10b and ABCA1 and ABCG1 gene expression in vivo. Male 30-week-old ApoE -/- mice were intraperitoneally injected with thioglycollate. After four days, four-hour fasted mice were orally treated with PCA (5 mg/kg BW) for 0-16 h. Plasma PCA concentrations (A), cellular mir-10b (B), and ABCA1 and ABCG1 mrna (C) expression levels in isolated MPMs were quantified. The results in A are the means ± SEM. The results in B and C are the means ± SEM of fold of the Control (untreated mice served as controls), which was set to 1. n = 6. *, P < 0.05 versus Control. 14

31 15 Online Figure VIII Double knockdown of ABCA1 and ABCG1 in J774 macrophages reduces cholesterol efflux potential in vitro. J774 macrophages stably transfected with vectors encoding for shrna against mouse ABCA1 (shabca1) and ABCG1 (shabcg1) were loaded with AcLDL (25 g/ml) for 24 h. As control, cells transfected with the vector encoding for shrna against LacZ or not transfected cells (NT) were used. ABCA1 and ABCG1 expression levels were then assessed by qrt-pcr (A and B), and cholesterol efflux to (C) ApoA-I (10 g/ml) or (D) HDL (100 g/ml) was measured as described in Method. The results in A and B are the means ± SEM of fold of the Control cells (not transfected) from three independent experiments. The results in C and D are the means ± SEM from three independent experiments. #, P < NT, not transfected.

32 Online Figure IX Distribution of cholesterol in plasma lipoproteins. Pooled plasma samples (n=4 mice of each group) were separated by fast protein liquid chromatography. Individual fractions were assayed for cholesterol content. Fractions contain VLDL and chylomicra; fractions contain IDL and LDL; fractions contain HDL; and fractions contain non-lipoprotein-associated proteins. Data are expressed as the means. VLDL, very low density lipoprotein; IDL, intermediate density lipoprotein; LDL, low density lipoprotein; HDL, high density lipoprotein. 16