Prevotella copri promotes vascular calcification via lipopolysaccharide through activation of NF-κB signaling pathway

ABSTRACT

Emerging evidence indicates that alteration of gut microbiota plays an important role in chronic kidney disease (CKD)-related vascular calcification (VC). We aimed to investigate the specific gut microbiota and the underlying mechanism involved in CKD-VC. We identified an increased abundance of Prevotella copri (P. copri) in the feces of CKD rats (induced by using 5/6 nephrectomy followed by a high calcium and phosphate diet) with aortic calcification via amplicon sequencing of 16S rRNA genes. In patients with CKD, we further confirmed a positive correlation between abundance of P. copri and aortic calcification scores. Moreover, oral administration of live P. copri aggravated CKD-related VC and osteogenic differentiation of vascular smooth muscle cells in vivo, accompanied by intestinal destruction, enhanced expression of Toll-like receptor-4 (TLR4), and elevated lipopolysaccharide (LPS) levels. In vitro and ex vivo experiments consistently demonstrated that P. copri-derived LPS (Pc-LPS) accelerated high phosphate-induced VC and VSMC osteogenic differentiation. Mechanistically, Pc-LPS bound to TLR4, then activated the nuclear factor κB (NF-κB) and nucleotide-binding domain, leucine-rich–containing family, pyrin domain–containing-3 (NLRP3) inflammasome signals during VC. Inhibition of NF-κB reduced NLRP3 inflammasome and attenuated Pc-LPS-induced VSMC calcification. Our study clarifies a novel role of P. copri in CKD-related VC, by the mechanisms involving increased inflammation-regulating metabolites including Pc-LPS, and activation of the NF-κB/NLRP3 signaling pathway. These findings highlight P. copri and its-derived LPS as potential therapeutic targets for VC in CKD.

KEYWORDS:

• Vascular calcification
• prevotella copri
• gut microbiota
• chronic kidney disease
• lipopolysaccharides

1. Introduction

Vascular calcification (VC) is prevalent in patients with chronic kidney disease (CKD) and is a significant contributor to subsequent cardiovascular morbidity and mortality.^1–3 VC is an actively regulated pathophysiological process, analogous to osteogenesis, which involves the phenotypic transformation of vascular smooth muscle cells (VSMCs),^4–6 characterized by the upregulation of osteogenic factors such as alkaline phosphatase (ALP), runt-related transcription factor 2 (RUNX2), and bone morphogenetic protein-2 (BMP2), and downregulation of the contractile phenotype marker α-smooth muscle actin (α-SMA) ultimately leading to VC,^7–9 Emerging evidence implies that gut microbiota plays a central role in the pathogenesis of cardiovascular complications in CKD.^10,11 The microbial population within the intestine surpasses the total count of human cells, forms an intricate ecosystem which is equated to a human organ.¹² Gut dysbiosis is also involved in the development of hypertension, atherosclerosis and thrombosis.^13,14 However, whether the gut microbiota is involved in the development of VC in CKD needs to be unraveled.

Both experimental and clinical studies have identified a distinct gut microflora pattern associated with CKD. Particularly in advanced CKD stages, microbial families tied to uremic toxin production become predominant.^11,15 The uremic milieu further aggravates gut dysbiosis, in turn undermines the integrity of the gut barrier. As a result, it facilitates translocation of live bacteria, endotoxins like lipopolysaccharides (LPS), and gut-derived uremic toxins into systemic circulation, which can instigate systemic inflammation and increase the risk of cardiovascular complications.¹⁶ Others’ and our teams have demonstrated the associations between elements such as uremic toxins, trimethylamine N-oxide (TMAO), vitamin K, and short-chain fatty acids with cardiovascular outcomes in CKD.^8,17 These insights hint that a thorough exploration of gut microbiota might unravel the intricacies of VC in CKD. Nevertheless, pinpointing the specific pathogenic bacteria and understanding their mechanisms in driving VC pathogenesis remains a challenge.

To probe deeper into the gut microbiota’s influence on VC in the context of CKD, we leveraged 16S rRNA gene sequencing of fecal samples from both humans and animals, coupled with in vitro, ex vivo, and in vivo models, to explore the potential role of gut microbiota in regulating VC within CKD.

2. Methods

2.1. Clinical observational cohort study and analysis

Individuals diagnosed with CKD and admitted to Zhujiang Hospital between February and June 2023 were incorporated into this investigation (ClinicalTrials.gov ID: ChiCTR2300074963). Calcium scores of aortas were calculated for all enrolled patients, in addition to the collection of data on demography (sex, age, body mass index), cardiovascular indicators (systolic blood pressures, diastolic blood pressures, heart rate), blood biochemical profiles (serum creatinine [SCr], blood urea nitrogen [BUN], estimated glomerular filtration rate [eGFR], calcium, phosphorus, parathyroid hormone [PTH]), and urinary biochemical markers (microalbuminuria, 24-hour urinary protein excretion, urinary protein/creatinine), the modality of dialysis therapy (peritoneal dialysis and hemodialysis) and medication use (statin and sodium-glucose cotransporter 2). At the admission, the clinicians also recorded the diagnosis of CKD. This diagnosis could be based on previous kidney biopsy where this information was available, but a biopsy-confirmed diagnosis was not mandatory. The following prespecified categories of causes of CKD were defined: diabetic nephropathy, hypertensive nephropathy, glomerulonephritis and other or unknown causes of CKD.¹⁸ Specifically, the eGFR was calculated using the Chronic Kidney Disease Epidemiology Collaboration equation.¹⁹ Additionally, fecal samples from the participants were secured for analysis. (1) Inclusion criteria: participants (aged 18–80 years) met the diagnostic criteria of chronic kidney disease stage 5 (uremia stage) (KDIGO 2012 Clinical Practice Guideline for the Evaluation and Management of Chronic Kidney Disease) with available computed tomography (CT) scans. (2) Exclusion criteria: pregnant and breastfeeding women; patients with severe infection; patients with other important organ dysfunction; patients with blood and immune system diseases. Regarding the potential influence of dietary intake on gut microbiota composition, we further excluded those who had consumed antibiotics, probiotics, prebiotics, synbiotics, or laxatives in the 4 weeks preceding the study (Figure S1). All consenting participants provided their affirmation through written informed consent. The research protocol was meticulously reviewed and subsequently endorsed by the Ethics Committee of Zhujiang Hospital (ID: 2022-KY-287-01).

All the images were analyzed by 2 different experienced radiologists by using the software (Inobitec PRO 2.9.0), who were blind to the study design. Depending on Agatston scores,^20,21 we calculated the total calcium scores from the ascending aorta, aortic arch, and descending thoracic aorta to represent the aorta’s total calcification score. In brief, the criteria for calcification included CT values ≥ 130 Hounsfield units (HU) at the lesion with a calcified area ≥1 mm². The calcification score was derived by multiplying the calcified area by its respective coefficient. Each CT slice was analyzed independently, and the summative score across all slices represented the patient’s aortic calcification score. The use of chest CT in CKD patients was driven by specific clinical indications: 1) the enhanced risk of infections due to altered immune responses, making CKD patients particularly vulnerable to pneumonia; 2) the need to assess pulmonary complications such as edema and pleural effusions, which often accompany cardiovascular issues and significantly affect outcomes; and 3) the ability to exclude other thoracic conditions like nodules, masses, or lymphadenopathy that could impact patient management. These scans were performed without contrast agents to avoid potential renal damage. By distinguishing patients with or without aortic calcification, we aimed to explore the link between VC and gut microbiome changes in CKD progression.

2.2. Animal studies

The study obtained approval for all animal experiments from the Institutional Animal Care Committee at Zhujiang Hospital, Southern Medical University, China (LAEC-2023-088). It adhered to the guidelines set forth by the US National Institutes of Health Guide for the Care and Use of Laboratory Animals (8th Edition, 2011). Male Sprague–Dawley rats weighing 220–250 g were procured from the Central Animal Care Facility of Southern Medical University, China. All animals were kept under standard laboratory conditions following a 12-h light/12-h dark cycle, with unrestricted access to tap water and food. For the study, rats were randomly assigned to two groups: the CKD group (CKD, n = 5) and the VC group (Model, n = 5). The VC model was performed as described previously.^7,8,22 In brief, rats were anaesthetized with sodium pentobarbital (50 mg/kg) via intraperitoneal injection. Subsequently, a 2/3 right nephrectomy was performed, followed by the complete removal of the left kidney a week later to induce CKD. SCr levels were assessed 2 weeks post-surgery. Subsequently, the CKD rats were fed a high-calcium and high-phosphorus diet (4% calcium and 1.8% phosphate, Guangdong Medical Laboratory Animal Center, China), along with a daily oral gavage of 1 µg/kg calcitriol. Following a 4-week period, the CKD and VC rats were euthanized, and feces samples were collected for 16S rRNA analysis. To explore the potential role of Prevotella copri (P. copri, Guangdong Microbial Culture Collection Center, China) in promoting VC, the rats were grouped into 3 groups: Sham (n = 5), Model + Vehicle (n = 5, VC rats received phosphate buffered saline for 3 weeks as control), and Model + P. copri (n = 5, VC rats were orally gavaged with a P. copri bacterial solution [1 × 10⁹ CFU/200 µL phosphate buffered saline] for 3 weeks). Further experiments delved into the impact of P. copri removal on VC in CKD, rats were randomized into 3 groups: Sham (n = 5), Model+ Vehicle (n = 5, CKD rats were oral gavage with a comparable volume of 0.9% saline as vehicle control for 1 week, followed by 4 weeks of a high calcium and high phosphate diet with calcitriol [1 µg/kg]), and Model + Met (n = 5, CKD rats were orally gavaged with a metronidazole [200 mg/kg/d] for 1 week to clear P. copri,^23,24 followed by the same 4-week treatment as Model+ Vehicle group). At the end of the experiment, aortic tissues were collected from the euthanized rats for further analysis.

2.3. Cell culture

Primary rat VSMCs were obtained from the thoracic aortas of 2-month-old male Sprague-Dawley rats using the explant method detailed in our prior studies.^7,8,22 In brief, the rats were euthanized intraperitoneally with sodium pentobarbital (150 mg/kg) for aortic isolation. The aortas were dissected, longitudinally opened with sterile scissors, and then segmented into small pieces under sterile conditions. These aortic segments were cultivated in Dulbecco’s Modified Eagle’s Medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 100 U/ml penicillin, and 100 mg/ml streptomycin. The resulting VSMCs migrated from the aortic explants were sustained in the growth medium at 37°C within a humidified incubator with a 5% CO₂ atmosphere. Upon reaching 80% confluence, subculturing of VSMCs was carried out. Cells from passages 3 to 8 were employed in the experiments. For calcification induction, VSMCs were treated with a calcifying medium composed of DMEM supplemented with 10 mM β-glycerophosphate (BGP) and 3 mM calcium chloride (CaCl₂). Pc-LPS was extracted from P. copri using an LPS Extraction Kit (Abcam, ab239718). In some experiments, VSMCs were treated with variable concentrations of Pc-LPS (0.2, 0.4, and 1.2 ng/mL) in conjunction with the calcifying medium over a period of 3 to 8 days. Pyrrolidinedithiocarbamate ammonium (PDTC, 1 μmol/L; Selleck), a potent inhibitor of NF-κB was used to treat VSMCs in some experiments.

2.4. Arterial ring organ culture

Thoracic aortas were dissected from 2-month-old male Sprague–Dawley rats that had been euthanized via intraperitoneal injection of sodium pentobarbital (150 mg/kg) as described previously.^7,8,22 The aortas were carefully segmented into 2–5 mm rings and then incubated in the growth medium, calcifying medium or calcifying medium with Pc-LPS (0.2 ng/mL) for 6 to 8 days, with fresh medium changes taking place every two days. The aortic rings were collected for subsequent evaluation and analysis at the predetermined time points.

2.5. ALP activity assay

The activity of ALP was ascertained utilizing an ALP Assay Kit (Beyotime, China), adhering to the guidelines provided by the manufacturer. Specifically, proteins were extracted from the cellular samples and arterial tissues, following which the concentration of the obtained proteins was quantified via a BCA Protein Assay. These protein samples were subsequently mixed with a para-nitrophenyl phosphate substrate and subjected to a 10-minute incubation period at 37°C. Post-incubation, 3 M NaOH was added to terminate the reaction. Absorbance was then measured at 405 nm utilizing a microplate reader. Data were ultimately represented as units of ALP activity per milligram of protein.

2.6. Alizarin Red staining and quantification of calcium content

Alizarin Red staining was utilized to evaluate the presence and extent of calcification in VSMCs and arterial tissues, as described in our previous studies.^7,8,22 For cell staining, VSMCs were fixed with 4% paraformaldehyde for 10 minutes, followed by washing in PBS and exposure to 2% Alizarin Red solution (pH 4.2, Solarbio, China) at room temperature for 5 minutes. Excess dye was removed with deionized water, and the resulting calcium deposition in VSMCs was visualized and imaged using a Leica Microsystems microscope. Subsequently, the dye was eluted with 10% formic acid, and the absorbance was measured at 405 nm using a spectrophotometer or a microplate reader to quantify the extent of calcification.²⁵ Similarly, tissue staining involved fixing rat aortic segments in 4% paraformaldehyde, embedding them in paraffin, and sectioning them to a 5–6 μm thickness. The sections were then stained with Alizarin Red solution for 5 minutes, followed by washing to remove excess dye, dehydration in acetone, and mounting with mounting medium. The stained sections were visualized and imaged using an inverted microscope, and positive red staining areas in aortic rings were analyzed with ImageJ software. Whole-mount staining of aortic tissues involved fixing them in 95% ethanol for 24 h, staining with 0.003% Alizarin Red solution in 1% potassium hydroxide overnight, rinsing in 2% potassium hydroxide, and photographing them with an inverted microscope. Calcium content was determined using a commercial calcium assay kit (Leagene Biotechnology, China) for arterial calcium content detection. Concisely, both VSMCs and aortic tissues underwent homogenization, after which the supernatant was isolated via centrifugation. This supernatant was subsequently mixed with a Methyl thymol blue solution and incubated. The absorbance was subsequently assessed at 610 nm utilizing a microplate reader, with the relative calcium content being adjusted to align with the protein content.

2.7. Micro-CT imaging of aortic calcification

The extent of aortic calcification was evaluated using Micro-CT as previously described.^7,8,25 Initially, rats were euthanized by an intraperitoneal injection of sodium pentobarbital (150 mg/kg), and their aortas were harvested and promptly fixed with 4% formaldehyde. These aortic samples were scanned with a Micro-CT scanner, from Siemens Inveon, at a resolution of 0.079 mm. Following scanning, Micro-CT images were analyzed using the Siemens Inveon research workplace software.

2.8. Measurements of serum biochemistries and cytokine levels

Two weeks after 5/6 nephrectomy surgery and after the experiment, blood samples were collected from the ophthalmic veins of rats. The blood was then centrifuged to separate the serum. SCr levels in the rats were measured using a colorimetric method provided by the BioAssay Systems Creatinine Assay Kit (BioAssay Systems, USA). Furthermore, serum levels of LPS, interleukin-6 (IL-6), tumor necrosis factor-alpha (TNF-α), and interleukin-1 beta (IL-1β) were determined by rat ELISA kits (Shanghai FanKe biological technology co., LTD, China), adhering to the manufacturer’s guidelines. Serum calcium and phosphate levels were analyzed with a semiautomatic biochemical analyzer (ECA-2000A, Jilin, China). PTH and fibroblast growth factor 23 (FGF23) were also assessed using rat-specific ELISA kits (Shanghai FanKe Biological Technology Co., Ltd., China).

2.9. Western blot analysis

Protein extraction from VSMCs and arterial tissues was performed using RIPA buffer containing proteinase and phosphatase inhibitors. According to the manufacturer’s instructions, the protein concentrations were quantified spectrophotometrically using a Pierce™ BCA Protein Assay Kit (Thermo Fisher Scientific, USA). Subsequently, proteins were separated by SDS-PAGE and transferred to PVDF membranes (Millipore, USA). Membranes were blocked with 5% skim milk for 2 hours, and then incubated overnight at 4°C with specific primary antibodies, including BMP2 (1:1000, Abcam, USA, ab214821), RUNX2 (1:1000, Cell Signaling Technology, USA, #12556), α-SMA (1:1000, Cell Signaling Technology, USA, 19245S), Occludin (1:1000, Proteintech, China 27,260–1-AP), ZO-1 (1:1000, Proteintech, China 21,773–1-AP), TLR4 (1:1000, Signalway Antibody, USA, #35463), Phospho-NF-κB antibody (1:1000, Cell Signaling Technology, USA, #3033), NF-κB antibody (1:1000, Cell Signaling Technology, USA, #8242), leucine-rich–containing family, pyrin domain– containing-3 (NLRP3, 1:1000, Abcam, USA, ab263899) and GAPDH (1:8000, Bioworld, China, AP006). After the incubation with the primary antibodies, the membranes were washed three times. Subsequently, they were incubated with horseradish peroxidase-conjugated secondary antibodies for 2 hours at room temperature. Visualization of blots was achieved with the Amersham Imager 600 imaging system. Densitometry analysis of the blot bands was performed using Image J software, with relative protein expression levels normalized against GAPDH expression.

2.10. Quantitative real-time polymerase chain reaction (qRT-PCR)

Total ribonucleic acid (RNA) was isolated from cultured VSMCs using TRIzol Reagent (Thermo Fisher, USA) and reverse-transcribed into complementary DNA (cDNA) using the PrimeScript RT reagent kit (TaKaRa, Japan), in accordance with the respective manufacturer’s instructions. qRT-PCR was then conducted on the cDNA samples using a 7500 FAST Real-Time PCR System (Applied Biosystems, USA) with the SYBR Green mixture (TaKaRa, Japan). The relative mRNA expression levels were determined using the comparative Ct (ΔΔCt) method, which allowed for calculating the fold change in gene expression relative to controls. GAPDH served as the internal reference for normalization. Primer sets for genes are presented in Table S1.

2.11. Histological analysis

The ileum from each rat, located 1 cm from the cecum, was carefully harvested and fixed with 4% paraformaldehyde for 24 hours. These tissue samples were then embedded in paraffin and processed to generate 5 µm-thick sections. These sections were subsequently dewaxed and stained with hematoxylin and eosin using standard protocols to analyze the organizational morphology. After staining, sections were inspected under a Leica Microsystems microscope, and measurements were quantified using Image Pro Plus 6.0 software. Crypt depth was measured from the bottom of the crypt to the crypt-villus junction, and villus length was gauged from the crypt-villus junction to the tip of the villus.

2.12. Statistical analysis

All data were expressed as mean±SD. Normality of distribution for the investigated parameters was evaluated using the Shapiro-Wilk test. Statistical differences between 2 groups were analyzed using Student’s t-test or Pearson’s Chi-squared test where appropriate, while differences among more than 2 groups were compared using one-way analysis of variance (ANOVA) followed by Bonferroni post hoc test. All statistical analyses were performed using GraphPad Prism statistical software (version 9.0), and p < 0.05 was accepted as statistically significant.

3. Results

3.1. The abundance of P. copri was positively associated with aortic calcification

To examine the relationship between the gut microbiome and VC in CKD, amplicon sequencing of 16S rRNA genes was performed between CKD rats and Model rats. The α-diversity, analyzed via Chao1, Observed species, PD whole tree, and Shannon indices, delineated significant disparities between CKD groups and Model group (Figure 1a). These variances in microbiota composition were further substantiated by Principal component analysis (Figure 1b). Utilizing linear discriminant analysis effect size, we pinpointed Prevotella as a prominent taxonomic marker differentiating Model rats from CKD rats (Figure 1c). Of particular interest, the abundance of P. copri species was markedly augmented in the gut microbiota of CKD rats with aortic calcification (Figure 1d).

Figure 1. Gut microbiota composition was changed during vascular calcification in chronic kidney disease (CKD) rats, presenting higher prevotella copri (P. copri) abundance. (a) Gut microbial α diversity (Chao 1, observed species, PD whole tree and Shannon index) among CKD and Model groups. (b) Principal component analysis plot of unweighted UniFrac distances between 2 groups. (c) Linear discriminant analysis effect size showed bacterial taxa with significantly different abundances between 2 groups. (d) Relative abundances of P. copri in 2 groups. Data are expressed as mean±SD, CKD = 5, Model = 5. **p < .01 and ***p < .001 by Student’s t-test.

As shown in Figure S1, a total of 149 CKD patients were initially recruited. Of these patients, 113 patients were subsequently excluded due to either missing fecal samples or other exclusion criteria. Finally, our study cohort comprised 36 CKD patients: 18 CKD patients with aortic calcification (average age 57.28 ± 7.60 years; 55.56% of them were male) and 18 matched CKD controls without aortic calcification (average age 54.44 ± 8.08 years; 61.11% of them were male). As expected, there were no significant differences of various baseline parameters between the 2 groups (Table 1). Notably, the fecal samples from the CKD patients with aortic calcification demonstrated a markedly increased presence of P. copri compared to those without (Figure 2a,b), with elevated serum LPS levels (Figure S2). Moreover, Spearman’s correlation analysis further reinforced this observation (Figure 2c). These findings revealed a significant positive correlation between the abundance of P. copri and the aortic calcium scores.

Figure 2. Correlation analysis of the prevotella copri (P. copri) abundance with calcification scores in chronic kidney disease (CKD) patients. (a) Aortic calcification in CKD patients was detected by computed tomography scan. (b) Relative abundances of P. copri in faeces. (c) Spearman’s correlation analyses of the association between P. copri in feces with calcification scores. Data are expressed as mean±SD, n = 18 per group. ***p < .001 by Student’s t-test.

Table 1. Baseline characteristics of enrolled CKD patients.

Characteristics	Aortic without calcification (n = 18)	Aortic with calcification (n = 18)	P value
Demographic characteristics
Male (n, %)	11 (61.11%)	10 (55.56%)	0.735
Age (years)	54.44 ± 8.08	57.28 ± 7.60	0.286
BMI (kg/m^Citation2)	23.07 ± 2.53	23.88 ± 3.34	0.417
SBP (mmHg)	134.67 ± 25.11	139.06 ± 25.72	0.608
DBP (mmHg)	82.61 ± 14.21	78.78 ± 12.80	0.401
Heart rate (bpm)	91.67 ± 8.82	86.67 ± 9.52	0.111
Calcification score	0	1200.09 [245.84, 8050.23]	0.004
Blood biochemical characteristics
SCr (μmol/L)	824.44 ± 368.80	872.39 ± 324.11	0.681
BUN (mmol/L)	24.24 ± 7.64	20.93 ± 11.15	0.306
eGFR (ml/min per 1.73 m^2)	6.69 ± 3.20	5.44 ± 2.45	0.197
Ca (mmol/L)	1.99 ± 0.21	2.01 ± 0.15	0.795
Pi (mmol/L)	1.68 ± 0.51	1.63 ± 0.46	0.761
PTH (pg/mL)	270.00 [197.50, 441.00]	221.50 [84.50, 309.25]	0.628
Urinary biochemical characteristics
Microalbuminuria (n, %)	4 (22.22%)	6 (33.33%)	0.457
24-hour urinary protein (mg/24 h)	1768.85 [694.25, 3568.5]	1202.80 [795.30, 1630.05]	0.609
Urinary protein/creatinine (mg/mmol)	459.50 [186.75, 754.5]	425.00 [168.75, 567.00]	0.380
Cause of CKD			0.895
Diabetic nephropathy (n, %)	5 (27.78%)	7 (38.89%)
Hypertensive nephropathy (n, %)	4 (22.22%)	5 (27.78%)
Glomerulonephritis (n, %)	3 (16.67%)	2 (11.11%)
Other or unknown causes (n, %)	6 (33.33%)	4 (22.22%)
Type of dialysis			0.630
Peritoneal dialysis (n, %)	2 (11.11%)	3 (16.67%)
Hemodialysis (n, %)	16 (88.89%)	15 (83.33%)
Medication
Statin (n, %)	12 (66.67%)	9 (50.00%)	0.311
SGLT2 Inhibitor (n, %)	1 (5.56%)	2 (11.11%)	0.547

Abbreviations: BMl, body mass index; BUN, blood urea nitrogen; Ca, calcium; CKD, chronic kidney disease; SCr, serum creatinine; DBP, diastolic blood pressure; eGFR, estimated glomerular filtration rate; Pi, phosphorus; P. copri, Prevotella copri; PTH, parathyroid hormone; SBP, systolic blood pressure; SGLT2, sodium-glucose cotransporter 2.

3.2. P. copri promoted aortic calcification in CKD rats

To further investigate the regulatory role of P. copri in VC in vivo, mono-colonization of live P. copri was intragastric administered to Model rats (Figure 3a). As shown in Figure 3b, rats in the Model group exhibited significant weight loss compared to those in the Sham group, which was even more pronounced in P. copri-treated rats. SCr levels revealed no significant difference between the 2 Model groups prior to P. copri colonization (Figure 3c). After the experiment, qRT-PCR confirmed successful colonization in P. copri group (Figure 3d). After colonization with P. copri, there were no significant differences observed in SCr, calcium, phosphate, PTH, and FGF23 levels between the 2 Model groups (Table S2). Micro-CT analysis, Alizarin Red staining, and calcium content quantification collectively confirmed the increased aortic calcification in P. copri-treated rats compared to saline-treated ones (Figure 3e–i). As shown in Figure 3j, compared to Sham group, ALP activity was significantly increased in Model group, and was further enhanced in P. copri group. Moreover, after P. copri treatment, the expression of contractile marker, α-SMA was downregulated, while the expression of osteogenic markers, RUNX2 and BMP2 was upregulated in CKD rat aortas (Figure 3k–n).

Figure 3. Prevotella copri (P. copri) promoted aortic calcification in chronic kidney disease (CKD) rats. (a) Experimental design. (b) The bodyweight of rats was measured weekly. (c) Serum creatinine (SCr) levels were measured 2 weeks after surgery. (d) Relative abundances of P. copri in faeces were measured at the end of the experiment. (e) Micro-CT analysis of rat aortic calcification. Scale bar: 10 mm. (f) Representative image of Alizarin Red-stained aortas. Scale bar: 10 mm. (g) Representative image of Alizarin Red-stained aortic sections. Scale bar: 500 μm (upper) and 250 μm (lower). (h) Alizarin Red positive area was quantified by image J software. (i) Quantification of calcium content in aortas using a Ca assay kit. (j) Alkaline phosphatase (ALP) activity was assessed using an ALP activity assay kit. (k) Representative western blots of runt-related transcription factor 2 (RUNX2), bone morphogenetic protein-2 (BMP2), and α-smooth muscle actin (α-SMA). (l–n) Quantification of RUNX2, BMP2 and α-SMA protein expression by densitometry. Data are expressed as mean±SD, n = 5 per group. Not significant (ns), *p < .05, **p < .01 and ***p < .001 by one-way ANOVA with Bonferroni post hoc test.

In light of previous findings indicating metronidazole’s ability to substantially decrease a broad spectrum of bacterial populations, including P. copri,^24,26 we embarked on the effects of clearing P. copri by metronidazole on VC in rats (Figure 4a). Model group exhibited significant weight loss compared to Sham group, however, the administration of metronidazole successfully mitigated this effect (Figure 4b). SCr levels revealed no significant difference between the 2 Model groups before metronidazole treatment (Figure 4c). As expected, we observed a significant reduction in the abundance of P. copri in the metronidazole-treated rats compared to their untreated counterparts (Figure 4d). However, metronidazole treatment did not significantly influence SCr, and calcium, phosphate, PTH, and FGF23 levels (Table S3). Intriguingly, Alizarin Red staining, calcium content and ALP activity measurements revealed a reduction in aortic calcification following metronidazole treatment (Figure 4e–i). Similarly, metronidazole treatment downregulated RUNX2 protein levels, accompanied with upregulation of α-SMA protein levels (Figure 4j-l).

Figure 4. Effect of prevotella copri (P. copri) clearance by metronidazole on aortic calcification in chronic kidney disease (CKD) rats. (a) Experimental design. (b) The bodyweight of rats was measured weekly. (c) Serum creatinine (SCr) levels were measured 2 weeks after surgery. (d) Relative abundances of P. copri in faeces were measured at the end of the experiment. (e) Representative image of Alizarin Red-stained aortas. Calcium deposition in the aortas was stained with Alizarin Red solution. Scale bar: 10 mm. (f) Representative image of Alizarin Red-stained aortic sections. Scale bar: 500 μm (upper) and 250 μm (lower). (g) Alizarin Red positive area was quantified by image J software. (h) Quantification of calcium content in aortas using a Ca assay kit. (i) Alkaline phosphatase (ALP) activity was assessed using an ALP activity assay kit. (j) Representative western blots of RUNX2 (runt-related transcription factor 2) and α-smooth muscle actin (α-SMA). (k–l) quantification of RUNX2 and α-SMA protein expression by densitometry. Data are expressed as mean±SD, n = 5 per group. Not significant (ns), *p<.05, **p < .01 and ***p < .001 by one-way ANOVA with Bonferroni post hoc test.

3.3. P. copri aggravated gut barrier disruption and increased serum LPS levels in CKD rats

Growing evidence underscores the role of heightened gut barrier permeability, facilitating the translocation of harmful metabolites or bacterial components from the gut to the bloodstream, in eliciting systemic inflammatory responses stimulated by gut dysbiosis.^27,28 In order to demystify the underpinnings of P. copri-induced VC, we embarked on a comprehensive examination of intestinal morphology. As shown in Figure 5a-c, the rats colonized with P. copri exhibited observable epithelial disruption, manifested by a reduction in villus height and crypt depth. Concurrently, we detected a significant decrease in the mRNA and protein expression levels of crucial tight-junction proteins, specifically ZO-1 and Occludin, in P. copri group (Figure 5d–h). In addition, we also found significantly higher serum LPS levels (Figure 5i) and intestinal expression levels of TLR-4 (Figure S3), a receptor sensitive to LPS, in Model group than Sham group, which was further enhanced in P. copri group. These results indicate that P. copri supplementation intensifies the inherent gut barrier dysfunction associated with VC, leading to increased LPS leakage. Additionally, the detection of TLR4 expression in both VSMCs and aortic tissues was consistent with those observed in intestine tissues, suggesting its importance in Pc-LPS-induced VC (Figure 5j–m).

Figure 5. Prevotella copri (P. copri) damaged the gut barrier and promoted the translocation of bacterial lipopolysaccharide (LPS) in chronic kidney disease (CKD) rats with aortic calcification. (a–c) morphology of the ileum was assessed using hematoxylin and eosin staining (scale bar = 100 μm) and the average villus height and crypt depth was analyzed. (d–e) mRNA expression levels of intestinal tight junction proteins zonula occludens-1 (ZO-1) and Occludin in the three groups. (f) Representative western blots of ZO-1 and Occludin. (g–h) quantification of ZO-1 and Occludin protein expression by densitometry. (i) Levels of LPS in the peripheral blood. (j) Representative western blots of Toll-like receptor-4 (TLR4) in aortic tissues of rats. (k) Quantification of TLR4 protein expression by densitometry. (l) Representative western blots of TLR4 in vascular smooth muscle cells (VSMCs). (m) Quantification of TLR4 protein expression by densitometry. Data are expressed as mean±SD, n = 5 per group. *p < .05, **p < .01 and ***p < .001 by one-way ANOVA with Bonferroni post hoc test.

3.4. Pc-LPS promoted VC in vitro and ex vivo

The chemical structures and biological properties of LPS vary considerably across different organisms, displaying characteristics that distinctly diverge from those found in enterobacteria such as Escherichia coli.²⁷ In order to explore the effect of Pc-LPS on VC, rat aortic rings and VSMCs were treated with Pc-LPS in the calcifying medium for 1 week (Figure 6a). Alizarin Red staining revealed that Pc-LPS significantly promoted VSMC calcification in a dose-dependent manner (Figure 6b,c). Consequently, for subsequent experiments in vitro and ex vivo, we utilized the lowest concentration of Pc-LPS that still induced VC. As shown in Figure 6d–j, addition of Pc-LPS augmented the ALP activity in VSMCs, accompanied by elevated RUNX2 mRNA and protein levels, and decreased α-SMA mRNA and protein levels. Alizarin Red staining, calcium content, and ALP activity assays further corroborated these results using rat aortic rings (Figure 6k–n). Additionally, Pc-LPS notably increased RUNX2 expression and reduced the expression of α-SMA at protein levels in the rat aortic ring (Figure 6o–q).

Figure 6. Lipopolysaccharide from Prevotella copri (pc-LPS) promoted vascular calcification in vitro and ex vivo. (a) Experimental design. (b) Representative images of Alizarin Red-stained vascular smooth muscle cells (VSMCs). Scale bar: 500 μm. (c) Quantitative analysis of alizarin red dye was performed by a microplate reader. (d) Quantification of calcium content using a Ca assay kit. (e) Alkaline phosphatase (ALP) activity was assessed using an ALP activity assay kit. (f–g) mRNA expression of runt-related transcription factor 2 (RUNX2) and α-smooth muscle actin (α-SMA). (h) Representative western blots of RUNX2 and α-SMA. (i-j) quantification of RUNX2 and α-SMA protein expression by densitometry. (k) Representative image of Alizarin Red-stained aortic sections. Scale bar: 500 μm. (l) Alizarin Red positive area was quantified by image J software. (m) Quantification of calcium content in aortas using a Ca assay kit. (n) ALP activity was assessed using an ALP activity assay kit. (o) Representative western blots for RUNX2 and α-SMA. (p-q) quantification of RUNX2 and α-SMA protein expression levels by densitometry. Data are expressed as mean±SD, n = 5 per group. *p < .05, **p < .01 and ***p < .001 by one-way ANOVA with Bonferroni post hoc test.

3.5. Pc-LPS activated NF-κB signaling pathway activation during VC

We next investigated the molecular mechanism by which Pc-LPS regulated VC. Previous studies showed that LPS promoted vascular inflammation by activating the NF-κB signaling pathway, which was involved in the pathogenesis of VC.^8,29 Thus, we examined whether Pc-LPS had an effect on the activation of NF-κB during VC. As shown in Figure 7a–c, treatment with P. copri. significantly increased the levels of phosphorylated NF-κB and NLRP3 protein expression in rat aorta. In addition, P. copri increased the levels of inflammatory cytokines, TNF-α, IL-6, and IL-1β, both in aortic arteries (Figure 7d). In line with these in vivo results, the expression levels of phosphorylated NF-κB, NLRP3, TNF-α, IL-6, and IL-1β in VSMCs were significantly higher in Pc-LPS group than in Model group (Figure 7e–h). In contrast, inhibition of NF-κB by PDTC effectively attenuated Pc-LPS-induced VSMC calcification (Figure 7i–l). Addition of PDTC also significantly decreased RUNX2, but increased α-SMA mRNA and protein levels in the Pc-LPS-treated VSMCs (Figure 7m–q).

Figure 7. Activation of nuclear factor κB (NF-κB) was required for lipopolysaccharide from Prevotella copri (pc-LPS)-induced calcification of vascular smooth muscle cells (VSMCs). (a) Representative western blots of phosphorylated NF-κB (pNF-κB) and nucleotide-binding oligomerization domain, leucine-rich repeat and pyrin domain-containing 3 (NLRP3) of rat aorta. (b–c) quantification of pNF-κB and NLRP3 protein expression by densitometry. (d) The concentration of tumor necrosis factor-alpha (TNF-α), interleukin-6 (IL)-6 and IL-18 in rat aorta was determined by ELISA assay. (e) Representative western blots of pNF-κB and NLRP3 of VSMCs. (f–g) quantification of pNF-κB and NLRP3 protein expression by densitometry. (h) The concentration of TNF-α, IL-6 and IL-18 in VSMCs was determined by qRT-PCR. (i–q) VSMCs were incubated with pc-LPS in the presence of calcifying medium with or without NF-κB inhibitor pyrrolidinedithiocarbamate ammonium (PDTC). (l) Representative images of alizarin red-stained VSMCs. Scale bar: 500 μm. (j) Quantitative analysis of alizarin red dye was performed by a microplate reader. (k) Quantification of calcium content using a Ca assay kit. (l) ALP activity was assessed using an ALP activity assay kit. (m-n) mRNA expression of RUNX2 and α-SMA. (o) Representative western blots for RUNX2 and α-SMA. (p-q) Quantification of RUNX2 and α-SMA protein expression levels by densitometry. Data are expressed as mean±SD, n = 5 per group. *p < .05, **p < .01 and ***p < .001 by one-way ANOVA with Bonferroni post hoc test.

4. Discussion

The present study demonstrates several novel findings: (1) both CKD patients and rats with VC exhibit distinct gut microbiota dysbiosis, with an increased abundance of P. copri; (2) P. copri colonization deteriorates the gut barrier, elevated serum LPS levels, and in turn aggravates VC in CKD rats; (4) LPS from P. copri significantly induces osteogenic differentiation and calcification of VSMCs via the activation of the NF-κB signaling pathway. To our best knowledge, this is the first study to reveal not only the etiological role of gut dysbiosis in VC but also the typical pathogenic bacteria of this dysbiosis in the setting of CKD (Figure 8).

Figure 8. A schematic diagram of the mechanism of action. In rats with chronic kidney disease (CKD)-induced vascular calcification (VC), gut microbiota homeostasis is disrupted, evidenced by heightened prevotella copri (P. copri) levels and their increased lipopolysaccharide (LPS) content. Concurrently, there is an augmented expression of Toll-like receptor-4 (TLR4), a receptor specific to LPS. This imbalance leads to mucosal barrier disruption and subsequent gut ‘leakage,’ which activates nuclear factor κB (NF-κB) and nucleotide-binding domain, leucine-rich-containing family, pyrin domain-containing-3 (NLRP3) inflammasome signals and then instigates systemic inflammatory responses, further exacerbating VC.

Although the link between gut microbiota and cardiovascular disease has been previously reported,^30–32 whether gut bacteria mediated the effect of VC are still unclear. Our prior research showed that treating CKD rats with antibiotics to suppress gut microbiota, can reduce aortic calcification, suggesting the potential of reshaping gut microbiota as a therapeutic strategy for VC.⁸ Consequently, identifying specific gut bacteria associated with VC development becomes crucial for the therapeutic strategy. Notably, in the present study, a marked difference in gut microbial compositions was observed between Sham and Model groups. Specifically, CKD rats with aortic calcification had an increased abundance of Prevotella, which aligned with an earlier finding in cardiac valve calcification.²⁴ Subsequent analyses pinpointed P. copri as the dominant species in the gut of both CKD patients and rats exhibiting VC. P. copri is a species of bacteria predominantly found in the human gut, especially in non-Westernized populations.³³ It is responsible for the digestion of complex carbohydrates. While P. copri is a natural component of the human microbiome, its overrepresentation has been associated with both ankylosing spondylitis and rheumatoid arthritis,^33–36 Concurrently, P. copri was linked to steatohepatitis and moderate to severe fibrosis in nonalcoholic fatty liver disease.³⁷ Furthermore, P. copri was reported to exacerbate chronic inflammation in pigs by producing metabolites that significantly increase host fat deposition.^38,39 Thus, due to its roles in inflammation, P. copri is a possible critical pathogen involved in cardiovascular disease. We had conducted a preliminary experiment where Sham rats were colonized with P. copri. Interestingly, P. copri colonization alone did not induce aortic calcification in the absence of CKD, suggesting that the elevated P. copri observed in human feces may be a promoter rather than an initiator of CKD-related VC. In this study, we for the first time demonstrated that P. copri colonization exacerbated aortic calcification in CKD rats, which was attenuated by clearance of P. copri. These findings suggest that targeting P. copri is a potential therapeutic strategy for retarding VC development in CKD.

The elevated P. copri in CKD human with VC, though not explicitly delineated in the current study, may arise from the intricate interplay of CKD effects and gut microbiota dynamics. Specifically, CKD-induced gut environmental shifts, including altered pH, augmented urea, and modified nutrient profiles, favors P. copri proliferation. Additionally, dietary constraints in CKD may further modulate this microbial balance. Importantly, the inflammatory cascade in VC with CKD potentially amplifies P. copri growth, perpetuating inflammation and compromising intestinal barriers. This compromised state facilitates bacterial translocation, eliciting systemic inflammation and perhaps further boosting P. copri dominance.^11,15,16 To conclude why P. copri is elevated in CKD patients with VC, rigorous experimental studies exploring each potential mechanism are needed. Our focus was to elucidate the signaling pathways bridging P. copri abundance with VC. LPS, an inflammatory stimulant found in the outer membrane of Gram-negative bacteria, emerged as a pivotal actor. Although prior studies highlighted the role of LPS derived from Escherichia coli in VC,^40,41 our model showed no differential changes in Escherichia coli levels. Importantly, Escherichia coli is not the exclusive source of LPS. Gram-negative bacterial species possess diverse LPS structures.⁴² Significantly, our findings showed that CKD rats colonized with P. copri had pronounced intestinal disruption, resulting in upregulated TLR4 expression—a receptor attuned to LPS.⁴³ This led to a substantial increase in LPS levels. We further demonstrated that LPS specifically derived from P. copri (Pc-LPS) not only promoted VC but also instigated a phenotypical shift of VSMCs from their contractile to osteogenic forms, positioning Pc-LPS as the primary effector in P. copri-driven VC promotion.

The connection between inflammation and VC is well-established.^44,45 Both our prior and current research highlight that the transcription factor NF-κB, a major regulator of vascular inflammation, advances VC in CKD rat models.^8,46 Significantly, LPS binding to TLR4 activates the NF-κB, triggering the transcription of multiple inflammatory factors, including IL-6, IL-1β, and TNF-α.^47,48 Given these insights, we investigated whether NF-κB signaling was implicated in Pc-LPS-induced VC. Our data indicate that Pc-LPS bind to TLR4, activates the NF-κB and NLRP3 inflammasome signals in VSMCs, accompanied with increased levels of TNF-α, IL-1β, and IL-6. When NF-κB was inhibited using PDTC, there was a marked reduction in Pc-LPS-driven osteogenic differentiation and VSMC calcification. These results underscore the role of NF-κB signaling in modulating Pc-LPS-induced VSMC calcification.

Our study acknowledges several limitations. First, microbial composition can significantly differ throughout the gastrointestinal tract. Since fecal samples mainly represent microbial populations from the distal colon, rather than from other locations such as the small intestine, we might have missed associations of species not abundantly present in fecal samples. Second, our study didn’t delve into the intricate interactions among bacterial species, so we cannot fully exclude the effects of other differentially abundant bacterial species on VC. Third, the model we employed, consisting of a 5/6 nephrectomy accompanied by calcitriol gavage and a diet enriched in calcium and phosphorus, cannot completely replicate the intricacies of a typical clinical setting. Further studies are warranted to validate the usefulness of leveraging P. copri as potential biomarkers or therapeutic targets. Furthermore, while our focus was specifically on the clearance of P. copri, we acknowledged that metronidazole potentially affected the overall composition of the gut microbiota. Recognizing this, future studies should be designed to more precisely evaluate these broader impacts. A more detailed understanding of how broad-spectrum antibiotics like metronidazole affect the gut microbiome, especially in the context of CKD and its association with VC, requires further studies.

5. Conclusion

In summary, we identify the crucial roles of P. copri and its derived Pc-LPS in CKD-related VC, by the mechanism involving inhibition of the NF-κB/NLRP3 signaling pathway. Targeting specific intestinal bacterial communities, particularly P. copri, represent a promising therapeutic strategy for VC in CKD. Further emphasis on monitoring early bacterial markers may provide more reliable guidance for VC.

Author contributions

QY Hao performed experiments. J Yan analyzed the data, and wrote the draft. JT Wei performed experiments and analyzed the data. YH Zeng, LY Feng, DD Que, SC Li, JB Guo, Y Fan, YF Ding and XL Zhang performed experiments. PZ Yang conceived and designed the experiments. JW Gao conceived and designed the study and analyzed the data. ZH Li conceived and designed the experiments, analyzed the data, and revised the manuscript.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/19490976.2024.2351532

Additional information

Funding

This work was supported by grants from the National Natural Science Foundation of China (82200442, 82070247, 82370237 and 82000460) and the Guangdong Basic and Applied Basic Research Foundation (2022A1515012263).