The intestinal microbiota regulates host cholesterol homeostasis
Management of blood cholesterol is a major focus of efforts to prevent cardiovascular diseases. The objective of this study was to investigate how the gut microbiota affects host cholesterol homeostasis at the organism scale.
We depleted the intestinal microbiota of hypercholesterolemic female Apoe−/− mice using broad-spectrum antibiotics. Measurement of plasma cholesterol levels as well as cholesterol synthesis and fluxes by complementary approaches showed that the intestinal microbiota strongly regulates plasma cholesterol level, hepatic cholesterol synthesis, and enterohepatic circulation. Moreover, transplant of the microbiota from humans harboring elevated plasma cholesterol levels to recipient mice induced a phenotype of high plasma cholesterol levels in association with a low hepatic cholesterol synthesis and high intestinal absorption pattern. Recipient mice phenotypes correlated with several specific bacterial phylotypes affiliated to Betaproteobacteria, Alistipes, Bacteroides, and Barnesiella taxa.
These results indicate that the intestinal microbiota determines the circulating cholesterol level and may thus represent a novel therapeutic target in the management of dyslipidemia and cardiovascular diseases.
KeywordsIntestinal microbiota Microbiome Microbiota transfer Cholesterol metabolism Enterohepatic cycle Antibiotics Dyslipidemia
Antibiotic-induced microbiota depletion
Gas chromatography–mass spectrometry
Principal component analysis
Amplicon sequence variant
Cholesterol is an essential component of eukaryotic cell membranes and is also a precursor of bile acids and steroid hormones. Dysregulation of cholesterol metabolism has been implicated in numerous diseases, including atherosclerosis and cardiovascular diseases , neurodegenerative diseases, non-alcoholic hepatitis , and cancers [3, 4]. Cholesterol metabolism is thus tightly regulated, and complex mechanisms regulate cholesterol levels, synthesis, and trafficking.
It has long been recognized that genetic [5, 6] and environmental factors such as the composition of the diet  as well as the amount of dietary cholesterol intake  have a strong impact on circulating cholesterol levels. Nevertheless, epidemiological studies recently reported that several bacterial taxa are associated with plasma cholesterol levels [9, 10, 11]. Other studies found a positive correlation between total and low-density lipoprotein (LDL) cholesterol and the abundance in the intestine of uncharacterized microbiota members belonging to Erysipelotrichaceae and Lachnospiraceae families . Elevated plasma cholesterol and especially LDL cholesterol levels remain a major risk factor in cardiovascular diseases (CVD) [13, 14, 15]. While the contributing role of intestinal microbiota to CVD through the production of TMAO, a proatherogenic metabolite derived from dietary carnitine and phospholipids has been thoroughly demonstrated [16, 17, 18, 19], recent data proposed that intestinal microbiota also impacts CVD pathogenesis through the modulation of circulating cholesterol levels. Moreover, dietary interventions showed that an increase in microbiota richness and diversity is associated with a decrease in circulating cholesterol [20, 21].
In normolipidemic wild type mice, germ-free (GF) condition as well as microbiota depletion through the administration of antibiotics upregulates de novo cholesterol synthesis with no raise in plasma cholesterol [22, 23, 24]. Until recently, the role of the intestinal microbiota in dyslipidemic experimental models that display a plasma lipoprotein profile closer to human, such as Apoe−/− mice, has been poorly explored. Some publications have used pre- and probiotics to delineate how they can downregulates plasma cholesterol levels [25, 26]. The latest findings report that cholesterol levels and atherosclerosis lesions are higher in Apoe−/− GF mice in comparison to Apoe−/− conventionally raised mice but the mechanisms are still missing [27, 28]. Although transfer of intestinal microbiota into GF animals demonstrated that microbiota composition/activity determines recipient phenotype and susceptibility to several diseases [29, 30], this approach has not yet been applied for plasma cholesterol levels nor with human microbiota.
The objective of this study was to investigate how the gut microbiota affects host cholesterol homeostasis at the organism scale in a dyslipidemic context. First, we investigated how depleting the microbiota using antibiotics affects host cholesterol metabolism and cholesterol enterohepatic cycle. Then, using a strategy based on human to mice gut microbiota transplant, we demonstrate that specific intestinal microbiota composition regulates cholesterol absorption, biosynthesis, and circulating cholesterol levels.
Apoe−/− and LDLr−/− mice on the C57BL/6 J background were bred and kept in a conventional animal facility at the Central Animal Facility of La Pitié Salpêtrière Hospital, with temperatures maintained at 21 °C and with 12-h light and darkness cycles. Mice had free access to water and regular chow diet (RM1, Dietex) and were weaned at 22–24 days. All mice were anesthetized with isoflurane and then sacrificed by exsanguination and cervical dislocation.
Intestinal microbiota depletion
Microbiota depletion was performed immediately after weaning by daily gavage with a combination of neomycin (200 mg/kg), metronidazole (200 mg/kg), ampicillin (200 mg/kg), and vancomycin (100 mg/kg) for 4 weeks . Antibiotics were dissolved in tap water (20 mg/ml for neomycin, metronidazole, and ampicillin, and 10 mg/ml for vancomycin), filtered on 0.22 μm, aliquoted, and stored at − 20 °C until use. Control mice received water by oral gavage. All antibiotics were obtained from Sigma Aldrich.
Mice were anesthetized by an intraperitoneal injection of a mixture of ketamin and xylazine (100 mg/kg and 1 mg/kg, respectively). Mice were placed on a heating pad at 37 ± 1 °C. A 1.5-cm incision on the abdomen was performed, and the common bile duct was ligated. Then, a polyethylene tubing (0.023 in. diameter) was inserted in the gallbladder and maintained with another ligation. Bile was collected during 1 h in a 0.5-ml tube. The bile volume was assessed by pipetting.
Intestinal microbiota transplant
Fresh human stool samples were collected in an anaerobic box (GENbag Anaert; Biomérieux). After thorough homogenization with a spatula, 1 g of stool was diluted (1:10 w/vol) and homogenized in reduced sterile Ringer solution (VWR) containing 0.5 g/L L-Cysteine (Sigma). This solution was then diluted 1:2 in reduced sterile 20% skimmed milk (Merk) and stored at − 80 °C until use. Mice were treated with antibiotics as previously described for 4 weeks in order to deplete their gut microbiota. After 2 h of fasting, mice were given 500 mg of polyethylene glycol (Colopeg, Bayer) by oral gavage to flush out antibiotics of their gut . Six hours later, stool samples were thawed at 37 °C and mice were inoculated with 300 μl of the mixture. Then, mice were allowed free access to food. To ensure good colonization, mice were re-inoculated three additional times at days 1, 3, and 7. Mice were sacrificed and tissue collected 10 weeks later.
Plasma lipids and lipoprotein profile
Total cholesterol, phospholipids, and triglycerides were analyzed with an autoanalyzer (Konelab) using commercial reagents from Roche Diagnostics and Diasys.
The lipid distribution in plasma lipoprotein fractions was assessed on pooled sera (n = 6 per group) by gel filtration as previously described . Each fraction was subsequently analyzed for total cholesterol content as above.
Intestinal cholesterol and bile acid absorption
To assess cholesterol absorption, mice were fasted overnight and then gavaged with 50 μCi [3H]-cholesterol dissolved in 250 μl of olive oil. Two hours later, the plasma and liver were collected. Ten microliters of plasma and 10 mg of liver were assayed for radioactivity in triplicates.
To assess bile acid absorption, mice were fasted overnight and then gavaged with 25 μCi [3H]-taurocholic acid dissolved in 250 μl of olive oil. Two hours later, the plasma and liver were collected. Ten microliters of plasma and 10 mg of liver were assayed for radioactivity in triplicates.
Bile acid synthesis
Mice were gavaged with 50 μCi [14C]-cholesterol dissolved in olive oil. Then, feces were collected every 24 h during 72 h. Feces were dried at 60 °C during 1 h and manually grinded with a mortar. Two hundred milligrams of feces were homogenized in 1.2 ml of NaOH 0.5% in water and 1.2 ml of cyclohexane using a vortex during 2 min. Organic and aqueous phases were separated by centrifugation at 1200g during 10 min, collected, dried, and reconstituted in 200 μl of isopropanol and water, respectively. Ten microliters of each extract was assayed for radioactivity in triplicates.
Gene expression analysis by quantitative PCR
Liver, ileum, or jejunum samples were disrupted in RNA-PLUS solution (QBiogene) using lysing matrix D in 2-ml tubes (MP Biomedicals) and Precellys homogenizer (Bertin technologies). Total RNA was extracted using Macherey-Nagel RNA extraction kit. RNA concentration and purity were determined using the Nanodrop ND-1000 spectrophotometer (Thermo Fisher Scientific) at a wavelength of 260/280 nm.
Total RNA (1.5 μg per reaction) was reverse transcribed into complimentary DNA using SuperScript II Reverse Transcriptase (Invitrogen) according to the manufacturer’s instructions. PCR amplification was performed in duplicates on cDNA diluted 1/100 using SYBR Green I Master and a Roche Lightcycler 480. The relative gene expression was calculated by the 2-ΔΔCt calculation method, using 18S and hPRT as housekeeping genes and control group as reference.
Sterols quantification in the liver and bile
Bile and liver lipids were extracted in the presence of two internal standards, pregnanol and 5α-cholestan (Steraloids), according to Folch et al. methodology . Samples were homogenized in chloroform–methanol (2:1 v/v) using lysing matrix D in 2-ml tubes (MP Biomedicals) and a Precellys homogenizer (Bertin technologies). The organic extract was dried and reconstituted in methanol. Lipids were then saponified using 15% KOH (Sigma) in methanol at 60 °C during 1 h. Then, lipids were again extracted using hexan-diethyl-ether (1:1 v/v). The organic extract was subsequently dried and reconstituted in 60 μl cyclohexane, and silylation of sterols was performed with 40 μl of N,O-bis (trimethylsilyl)trifluoroacetamide-trimethylchlorosilane (99:1) at 60 °C during 1 h. Cholesterol and lathosterol were then quantified by GC-MS using a 5972 Hewlett Packard mass spectrometer and ChemStation data acquisition system. Briefly, sterols were injected in splitless mode and separated on a RTX65 column 30 m × 0.25 mm × 0.25 μm. Sterols were ionized using electronic impact and quantified in SIM mode. Ions 458.4 m/z and 255.0 m/z were used to quantify cholesterol and lathosterol, respectively.
16S rRNA gene sequencing
Feces were collected 3 to 5 days before sacrifice and immediately frozen in liquid nitrogen and then stored at − 80 °C. Fecal DNA was extracted as previously described . The V3-V4 region of the 16S rRNA gene was amplified with the universal primers F343 (CTTTCCCTACACGACGCTCTTCCGATCTACGGRAGGCAGCAG) and R784 (GGAGTTCAGACGTGTGCTCTTCCGATCTTACCAGGGTATCTAATCCT), using 30 amplification cycles with an annealing temperature of 65 °C. The resulting PCR products were purified and sequenced at the GeT-PlaGe Genotoul INRA platform (Toulouse, France) using Illumina MiSeq technology. Sequences were trimmed for adaptors and PCR primer removal and then clustered into ASV using QIIME2. We picked a reference sequence for each ASV using Deblur and assigned it at different taxonomic levels (from phylum to species) using the Greengenes database 13_8 . We used 99% sequence identity for ASV determination. The average number of sequences per sample was 5722 ± 1143 sequences per sample. Then, we normalized the dataset to the number of sequences of the sample with the lowest sequencing depth, that is to say 3619 sequences using Rhea script without random subsampling . No sample was excluded from the downstream analyses as all the samples had a similar rarefaction curve terminal slope.
Results are represented as mean ± SEM. Statistical analysis was performed by Mann–Whitney–Wilcoxon test using StatView Graphpad 6 (SAS Institute Inc., Cary, USA) for comparing two groups or by Kruskal–Wallis test followed by Dunn’s pairwise multiple comparisons procedure using R 3.3.1 program for comparing three or four groups; p or q < 0.05 was considered statistically significant. Principal component analyses (PCA) were performed using R program and ade4 package. Interclass PCA were computed and statistically assessed by a Monte Carlo rank test to observe their net effect on the scattering of the microbiota of different mice. We used R 3.3.1 and the Hmisc and corrplot packages to produce Spearman correlations matrix and the Rhea scripts pipeline to perform statistical analysis of the microbiota data . The cladogram generator GraPhIAn was used for 16S data visualization .
Microbiota depletion of conventional mice raises plasma cholesterol level
Cholesterol in the plasma exists mainly packaged in the form of lipoproteins: chylomicrons, very-low-density lipoproteins (VLDL), low-density lipoproteins (LDL), and high-density lipoproteins (HDL). Quantitative analysis of cholesterol distribution among lipoproteins separated by gel filtration revealed an increase of the abundance of VLDL and LDL subclasses (respectively + 53% and + 36%) in AB-Mdpl mice, whereas the HDL fraction was similar in Conv-R and AB-Mdpl mice (Fig. 1c).
These experiments confirm that intestinal microbiota contribute to the regulation of plasma cholesterol levels and demonstrate that microbial depletion strongly affect several lipoproteins levels, mainly VLDL and LDL.
Intestinal microbiota depletion increases intestinal cholesterol absorption with no effect on hepatic VLDL production
As the liver secretes VLDL particles, we investigated the impact of microbiota depletion on VLDL production. Likewise, as LDL particles derive from the loss of triglycerides by VLDL and intestine originating chylomicrons in the bloodstream, we investigated intestinal cholesterol absorption.
Then, we measured the appearance of labeled cholesterol in the plasma of conventionally raised and microbiota-depleted mice 2 h following gavage of [3H]-cholesterol in olive oil. The appearance of radiolabeled cholesterol in the plasma was 25% higher in antibiotic-treated mice (Fig. 1d), indicating that depleting the microbiota raises intestinal cholesterol absorption.
We next analyzed the jejunal expression of genes involved in intestinal cholesterol absorption (Npc1l1 ) and intracellular cholesterol excretion in the gut lumen (Abcg5 and 8 ). We observed that microbiota-depleted mice displayed a threefold increase in Npc1l1 expression while Abcg8 expression was moderately raised and Abcg5 expression was not affected (Fig. 1e). Moreover, the expression of several genes encoding apolipoproteins and proteins involved in chylomicron and preβ-HDL assembly and secretion were increased at least two folds in the jejunum of microbiota-depleted mice (Fig. 1f).
VLDL are assembled in the liver from triglycerides, cholesterol, and apolipoproteins (ApoB mainly) by the chaperone Mttp. Here, liver gene expression levels of ApoB and Mttp of Conv-R and AB-Mdpl mice were similar (Additional file 3: Figure S2A). This is consistent with the similar VLDL secretion rate assessed using Triton WR-1339 as an inhibitor of peripheric lipid uptake by endothelial lipoprotein lipase  (Additional file 3: Figure S2B).
This set of experiments reveals that depleting the intestinal microbiota with antibiotics raises intestinal cholesterol absorption. On the contrary, the hypothesis of elevated VLDL levels in microbiota-depleted mice being a consequence of increased hepatic VLDL synthesis and secretion is rather unlikely.
Intestinal microbiota depletion increases the hepatic clearance of plasma cholesterol through LDLr
Intestinal microbiota depletion enhances cholesterol synthesis in the liver
The gastrointestinal tract contributes to 15–35% and the liver to 20–40% of total cholesterol synthesis in rodents . The relative expression of Hmgcs1 and HmgcoAr, encoding two key enzymes in cholesterol biosynthesis pathway, was not affected following intestinal microbiota depletion in the intestine (Additional file 5: Figure S3) but significantly increased by four- to sevenfold in the liver (Fig. 2d). We next determined the liver content of cholesterol and lathosterol, a synthesis intermediate considered as a marker of cholesterol synthesis , by gas chromatography coupled to mass spectrometry (GC-MS). Cholesterol concentration was 30% higher and lathosterol concentration was doubled in the liver of AB-Mdpl compared to Conv-R mice (Fig. 2e). This indicates that intestinal microbiota regulates cholesterol biosynthesis specifically in the liver.
The intestinal microbiota influences bile acid synthesis and biliary cholesterol secretion
The drastic depletion of intestinal microbiota increases intraluminal cholesterol absorption as well as re-excretion in the bile by the liver. To investigate how intestinal microbiota depletion influences the balance between cholesterol intake and secretion, we force-fed mice with 14C-cholesterol and collected their feces every 24 h during 72 h. We separated neutral lipids containing cholesterol from water-soluble components including bile acids and measured radioactivity in each fraction (Fig. 3d). Conv-R mice excreted 70% more radioactive sterols (sum of neutral lipids and water-soluble fraction) than AB-Mdpl over 72 h (Fig. 3d), confirming that sterols accumulated in the body in the absence of microbiota. Specifically, AB-Mdpl mice excreted threefold less cholesterol and 50% more bile acids than Conv-R mice; hence, the bile acids represented a significantly higher proportion of fecal sterols in Ab-Mdpl mice (Fig. 3e). This suggests that the absence of gut microbiota leads to an accumulation of sterols in the body and that in this context bile acids constitute a significant proportion of fecal sterols.
Next, we observed that the increased fecal bile acid excretion was associated with a regulation of enzymes in the bile synthesis pathway. Expression levels of Ak1r1d1 and Cyp7a1, the rate-limiting enzyme in the bile acid synthesis pathway, were increased in the liver in AB-Mdpl mice, supporting an increased bile acid synthesis in the absence of microbiota (Fig. 3c). However, Cyp27a1 expression was similar in both groups while Cyp8b1 expression was decreased in AB-Mdpl mice (Fig. 3c). Considering that microbiota is known to induce intestinal FXR which in turn regulates hepatic Cyp7a1 through a fibroblast growth factor 15 (Fgf-15)-dependent mechanism , we determined Fgf-15 expression in the distal ileum. We found that microbiota depletion reduces Fgf-15 expression by 75% (Fig. 3f).
As microbiota depletion raises bile acid synthesis and secretion, we needed to examine whether modification of intestinal bile acid absorption can strengthen or lessen fecal loss of bile acids. Gavage with 3H-taurocholic acid showed that microbiota depletion significantly decreases taurocholic acid absorption (Fig. 3g). This is probably not related to a decrease in active transport of bile acids, as the gene expression of the two transporters Ibat and Abcc3 was not decreased by the microbiota depletion (Fig. 3h). This decrease in taurocholic acid absorption is therefore likely the consequence of a decrease in passive absorption, the major absorption pathway of microbiota-derived unconjugated bile acids .
Plasma cholesterol level is transmissible from humans to mice by microbiota transplantation
We colonized four groups of microbiota-depleted 7-week-old female Apoe−/− mice (n = 10–14 mice per group) through repeated oral gavages with fecal microbiota from respective donors (Fig. 4a). Strikingly, after 9 weeks, the mean of plasma total cholesterol levels of HiChol recipient mice was 23% higher than those of NorChol recipients (Fig. 4b). Other plasma lipids such as triglycerides and phospholipids were also dramatically increased (Fig. 4c), suggesting that as their donors, HiChol recipient mice had an overall altered plasma lipid profile.
Intestinal microbiota regulates cholesterol absorption/synthesis balance
On the contrary, genes of the cholesterol synthesis pathway (HmgcoAr and Hmgcs1) were two times less expressed in the liver of HiChol recipients than in Norchol recipients (Fig. 5b). Consistently, the concentration of lathosterol was significantly lower in the liver of the two groups of HiChol recipients than in the liver of Norchol recipients, supporting a decrease in hepatic cholesterol synthesis in HiChol recipient mice (Fig. 5c). However, hepatic cholesterol content was not affected by the donors’ status (Fig. 5c), suggesting that other cholesterol metabolism pathways in the liver were affected by the microbiota. As cholesterol, liver phospholipids were similar in the four groups while liver triglycerides were slightly raised in HiChol recipients in comparison to NorChol recipients (Fig. 5d).
Hepatic expression of LDL receptor was lower in HiChol than in NorChol recipient mice (Fig. 5e), suggesting a decreased hepatic uptake in mice colonized with the microbiota from dyslipidemic donors. Moreover, the expression of Cyp7a1 was also reduced in HiChol recipients, which likely result from the increased expression of its suppressor Fgf15 in the distal ileum (Fig. 5f, g). There was a trend towards decreased Cyp8b1 and canalicular cholesterol Abcg5/g8 and bile acid Abcb11 transporters, but this did not reach statistical significance (Additional file 9: Figure S4).
Altogether, this set of experiments suggests an elevated intestinal cholesterol absorption and a decreased hepatic uptake and synthesis in HiChol recipient mice in comparison to NorChol recipient mice. Biliary cholesterol secretion in the intestinal lumen may also be lower in HiChol than in NorChol recipient mice. This indicates more broadly that the microbiota could be a regulator of the intestinal absorption/hepatic synthesis balance.
Mice colonized with the microbiota of normocholesterolemic and dyslipidemic human donors harbor distinct intestinal microbiota composition
HiChol-associated microbiota taxa correlate with plasma cholesterol levels
To confirm whether one or several specific gut bacteria were involved in the regulation of major cholesterol metabolism pathways, we performed multiple correlation analyses between the previously identified ASVs and plasma cholesterol level as well as parameters associated with hepatic cholesterol synthesis, lipoprotein uptake by the liver, bile acid synthesis, and intestinal absorption (Fig. 6d). Six of the ten HiChol recipient-associated ASVs were significantly and positively correlated with plasma cholesterol levels. Five of these ASVs correlated negatively with markers of hepatic cholesterol synthesis such as HmgcoAr expression and lathosterol concentration in the liver. These ASVs also positively correlated with markers of intestinal absorption such as Npcl1 and Mttp expression in the jejunum. The Fgf15 expression in the ileum and the LDLr expression in the liver were also correlated with these ASVs; however, statistical significance was not reached, suggesting that these parameters of cholesterol metabolism are less tightly regulated by the microbiota than the other parameters. The sequences of seven of these ten ASVs were not assigned to the genus level by Qiime2 pipeline; however, manual BLAST against the EzBioCloud 16S data base (update 06 august 2019)  indicated that ASV 1 belongs to the Sutterellaceae family, ASV 3 and ASV 8 belong to the Turicimonas genus, and ASV 4 to the Erysipelotrichaceae family.
We found that the expression in the liver of genes involved in bile acid secretion was downregulated in the presence of a complete microbiota, in accord with previous studies . We also observed a downregulation of cholesterol transporters Abcg5/8, following antibiotic exposure, consistent with previous studies in germ-free mice  but discrepant with others . Nevertheless, our data on Abcg5/8 expression are reinforced by the observation of a decrease in bile flow and cholesterol excretion in the presence of a complex microbiota. This dissimilarity could be the consequence of differences in diets and mice strains, with Rabot et al. and Ceasar et al. studies using wild type mice and high-fat diets with different fat composition and proportions, while in the present work we used dyslipidemic mice fed a chow, low-fat diet. Taken together, these observations support the growing idea that biological processes and in particular cholesterol metabolism can be determined by microbe–nutrient interactions and not only by microbiota and/or diet taken independently [22, 24, 49].
In our model, we also observed that the intestinal microbiota lowered intestinal absorption of dietary and biliary cholesterol, which are the following steps of enterohepatic circulation (Fig. 7). Intraluminal cholesterol is transported across the apical membrane of enterocytes by Npc1l1 , which expression was downregulated by the presence of microbiota. This somewhat surprising finding is opposed to the previous study addressing the impact of microbiota on intestinal cholesterol absorption . In both cases, Npc1l1 expression in the small intestine could be related to the effective intestinal cholesterol absorption measured by radioactivity tracking techniques. Npc1l1 transcriptional regulation is not so well described yet and existing data do not converge towards the same theory. However, mice were fed a western diet rich in lipids and sucrose in Zhong et al. study. Npc1l1 intestinal expression may be modulated by glucose  and SREBP2 , which are greatly affected by western diet only in the presence of a microbiota. We speculate that, once again, nutrients regulate cholesterol homeostasis depending on the presence or absence of intestinal microbiota. Moreover, Npc1l1 expression is decreased by a Lactobacillus strain  and prebiotic fibers . This confirms the ability of specific bacteria to decrease intestinal cholesterol absorption through Npc1l1 transcriptional downregulation in the intestine.
The following step of the enterohepatic circulation of cholesterol is the recapture of the absorbed cholesterol by the liver, where endogenous synthesis also occurs (Fig. 7). Indeed, the liver plays a pivotal role in cholesterol metabolism and plasma cholesterol levels as it (i) captures most of circulating cholesterol, (ii) secretes cholesterol-containing VLDL particles in the bloodstream, (iii) synthesizes significant amounts of cholesterol, and (iv) secretes cholesterol in the bile or converts it into bile acids (Fig. 7). We observed that both cholesterol uptake by the LDL receptor and hepatic de novo synthesis are drastically downregulated by the intestinal microbiota. We observed that the lack of microbiota triggered a downregulation of Fgf15 expression, which in turn suppresses the downregulation of bile acid synthesis in the liver, which corroborates several studies [46, 51], but is discrepant with others . The disruption of this Fgf15-dependent regulation pathway by microbiota depletion mainly explains the elevation of cholesterol synthesis and cholesterol uptake by the liver. The presence of intestinal microbiota is thus essential to whole-body cholesterol homeostasis, and we observed that the dysregulation of cholesterol enterohepatic cycle caused by microbiota depletion leads to an increase in cholesterol concentration in each compartment, notably in the plasma.
This raises the important question of the impact of variations of microbiota composition on the plasma cholesterol level in humans. To explore this, we colonized microbiota-depleted mice with the microbiota of four participants whose plasma lipid profile (total cholesterol, HDL/LDL cholesterol ratio, and triglycerides levels) was either healthy or associated with CVD risk according to the European and American cardiovascular societies [13, 56]. Murine recipients from dyslipidemic donors had significantly higher plasma cholesterol levels compared with recipients from normocholesterolemic donors. Thus, not only the presence/absence of intestinal microbiota, but also the variations in intestinal microbiota composition are sufficient to influence plasma cholesterol level. Each group of recipient mice had a specific microbial community, and very few features discriminate the cholesterol metabolism patterns. This confirms the high inter-variability in human microbiota composition and suggests that the majority of intestinal bacteria have no impact on host cholesterol metabolism while a limited number of taxa have a significant impact. Among the few bacteria that were associated with HiChol status, we found one ASV belonging to the Erysipelotrichaceae family, which has already been found to be positively correlated to plasma cholesterol [9, 12]. We were also able to positively correlate plasma cholesterol level and cholesterol metabolism to other bacterial taxa such as Alistipes, Barnesiella, and Turicimonas, which to our knowledge have not yet been associated with cholesterol metabolism. Notably, HiChol-associated bacteria were more strongly correlated with cholesterol metabolism-associated parameters than NorChol-associated bacteria. Hence, this suggests that the observed phenotypes in this study were likely the consequence of the presence of some deleterious bacteria rather than the absence of beneficial bacteria.
Several bacterial taxa are believed to lower plasma cholesterol levels or to reduce atherosclerosis development through the production of beneficial metabolites such as butyrate , through bile acid metabolism, or through entrapment of cholesterol . Conversely, the deleterious effect of gut microbiota on atherosclerosis development via TMAO production has been studied , but no mechanism explaining how some bacteria can raise plasma cholesterol has been described. Our study highlight that some bacteria are probably able to exert such deleterious activity; however, the fact that most of those bacteria are not cultivable at the present time prevents the study of the involved mechanism.
Finally, an important observation made in this study is that mice colonized with the microbiota of dyslipidemic donors had markers of elevated intestinal cholesterol absorption together with a lower cholesterol synthesis, while mice colonized with the microbiota of normocholesterolemic donors exhibited an opposite pattern. It is of high clinical interest that intestinal microbiota is able to shift the cholesterol absorption/synthesis balance, since it has repeatedly been observed in human cohorts that high absorption/low synthesis pattern is associated with higher LDL cholesterol and lower HDL cholesterol level and is predictive of CVD events [44, 59, 60, 61]. Strikingly, those high CVD risk individuals are those who have the poorest response to statins, the most used cholesterol-lowering drugs which act by inhibiting cholesterol synthesis in the liver .
In this study, we unveil the influence of intestinal microbiota on cholesterol fluxes and synthesis at the whole-body scale. By combining a series of in vivo investigations based on microbiota manipulation in dyslipidemic mice models, we report how the intestinal microbiota regulates cholesterol synthesis, absorption, and trafficking. Importantly, we show that plasma cholesterol levels can be transferred from humans to mice by intestinal microbiota transplantation, demonstrating the causal role of microbiota in the regulation of plasma cholesterol levels. We also show that intestinal microbiota regulates the balance between cholesterol synthesis and absorption. Thus, our findings open new possibilities for the prevention and treatment of CVD through modulation of the microbiota composition by the use of prebiotics, probiotics, or fecal transplantation.
We are grateful to Laurence Trystram, Dieudonné Manéné, Lucie Poupel, Eric Frisdal, Emilie Tubeuf, Laura Bouchareychas, and Raphaëlle Ballaire for the technical assistance and to Ruzica Milicevic for husbandry of mice.
Conceptualization of the manuscript was performed by TLR, PL, and PGe; methodology by TLR, EL, NK, and PL; investigation by TLR, EL, BC, MR, ML, SB, FI, JHB, and PL; resources by PL, NK, KC, TH, EM, PGi, PGe, and MG; writing of the original draft by TLR; writing, review, and editing of the manuscript by EL, BC, FI, NK, MR, EM, PGe, KC, and PL; visualization by TLR and BC; supervision by PL, KC, and NK; and funding acquisition by PL, KC, and PGe. All authors read and approved the final manuscript.
This study was supported by INSERM and the Fondation de France (00029519). T.L.R. was supported by the French National Research Agency ‘Investissements d’avenir’ ANR-10-IAHU-05. E.L. was supported by foundation Lefoulon-Delalande/Institut de France and the Region Ile-de-France CORDDIM.
Ethics approval and consent to participate
All procedures involving mice were carried out according to the Guide for the Care and Use of Laboratory Animals published by the European Commission Directive 86/609/EEC. All animal studies were approved by the regional veterinary services of the Paris police headquarters (agreement no. 75-1357) and by the Biological Services Unit of Sorbonne University.
Human stool collection was conducted in accordance with the Helsinki Declaration and was part of a study registered in a public trial registry (Clinicaltrials.gov; NCT02081066). The Ethics Committee of CPP Ile-de-France VI approved the investigations for all individuals, and written informed consent was obtained from all individuals (CPP/76-13- ID RCB: 2013-A00798-37).
Consent for publication
All authors have given consent for publication.
The authors declare that they have no competing interests.
- 1.Vallejo-Vaz AJ, Robertson M, Catapano AL, Watts GF, Kastelein JJ, Packard CJ, et al. LDL-cholesterol lowering for the primary prevention of cardiovascular disease among men with primary elevations of LDL-cholesterol levels of 190 mg/dL or above: analyses from the WOSCOPS 5-year randomised trial and 20-year observational follow-up. Circulation. 2017;136(20):1878–91.PubMedCrossRefPubMedCentralGoogle Scholar
- 14.Ference BA, Ginsberg HN, Graham I, Ray KK, Packard CJ, Bruckert E, et al. Low-density lipoproteins cause atherosclerotic cardiovascular disease. 1. Evidence from genetic, epidemiologic, and clinical studies. A consensus statement from the European Atherosclerosis Society Consensus Panel. Eur Heart J. 2017;38(32):2459–72.PubMedPubMedCentralCrossRefGoogle Scholar
- 15.Stone NJ, Robinson J, Lichtenstein AH, Merz CNB, Blum CB, Eckel RH, et al. 2013 ACC/AHA guideline on the treatment of blood cholesterol to reduce atherosclerotic cardiovascular risk in adults. Circulation. 2014;129(suppl 2):S1–S45.Google Scholar
- 43.van der Wulp MYM, Verkade HJ, Groen AK. Regulation of cholesterol homeostasis. Nucl Recept Bile Acids Cholest Homeost. 2013;368(1–2):1–16.Google Scholar
- 44.Matthan NR, Zhu L, Pencina M, D’Agostino RB, Schaefer EJ, Lichtenstein AH. Sex-specific differences in the predictive value of cholesterol homeostasis markers and 10-year cardiovascular disease event rate in Framingham offspring study participants. J Am Heart Assoc Cardiovasc Cerebrovasc Dis. 2013;2(1):e005066.Google Scholar
- 45.Dikkers A, Tietge UJ. Biliary cholesterol secretion: more than a simple ABC. World J Gastroenterol. 2010;21:5936–45.Google Scholar
- 62.Nissinen MJ, Miettinen TE, Gylling H, Miettinen TA. Applicability of non-cholesterol sterols in predicting response in cholesterol metabolism to simvastatin and fluvastatin treatment among hypercholesterolemic men. Nutr Metab Cardiovasc Dis. 2010;20(5):308–16.PubMedCrossRefPubMedCentralGoogle Scholar
- 63.Le Roy T, Lécuyer E, Chassaing B, Rhimi M, Lhomme M, Boudebbouze S, et al. Impact of the microbiome on sterols. NCBI accession number PRJNA543019 [Internet]. 2019. Available from: https://www.ncbi.nlm.nih.gov/bioproject/PRJNA543019
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.