TMAO-Activated Hepatocyte-Derived Exosomes Are Widely Distributed in Mice with Different Patterns and Promote Vascular Inflammation

Background Trimethylamine-N-oxide (TMAO) has been shown to be an important player in cardiovascular disease (CVD) by promoting vascular inflammation and endothelial dysfunction. We recently found that exosomes (Exos) released from TMAO-activated hepatocytes (TMAO-Exos) could significantly induce inflammation and endothelial dysfunction. However, understandings of how are the Exos secreted by hepatocytes, where are they distributed in vivo, and what effects will they have on vascular inflammation remain limited. The present study aimed to explore the hub genes involved in the production of TMAO-Exos and their distributions in vivo and effects on inflammation. Methods The transcriptome profiles of primary rat hepatocytes stimulated with TMAO were obtained from the GSE135856 dataset in the Gene Expression Omnibus repository, and the hub genes associated with Exos were screened and verified by qPCR. Next, Exos derived from TMAO-treated hepatocytes were isolated using differential centrifugation and given intravenously to mice. After 24 h, the distributions of DiI-labelled Exos were visualized with a fluorescence microscope, and the levels of proinflammatory genes in the aorta were detected by qPCR. Results Phgdh, Mdh2, Echs1, Rap2a, Gpd1l, and Slc3a2 were identified as hub genes that may be involved in the production of TMAO-Exos. And TMAO-Exos were found to be efficiently taken up by cardiomyocytes, hepatocytes, and endothelial cells in the aorta and gastrocnemius muscle. Furthermore, TMAO-Exos, but not control-Exos, could significantly promote the mRNA expressions of Tnf, Icam1, Sele, and Cox-2 in the aorta. Conclusions We provided clues about how TMAO may stimulate hepatocytes to produce Exos and further offered evidence that Exos secreted by TMAO-treated hepatocytes could be widely distributed in vivo and promote vascular inflammation.


Introduction
ere is increasing evidence that gut microbiota and its metabolites play a key role in the pathogenesis and development of cardiovascular disease (CVD) [1]. ereinto, trimethylamine-N-oxide (TMAO) has been found to be an independent risk factor for adverse cardiovascular events [2][3][4][5][6], which may be related to excessive vascular inflammation and endothelial dysfunction provoked [7][8][9][10]. It has been shown that dietary precursors such as choline, betaine, and L-carnitine can be metabolized into trimethylamine in the gut flora and further catalyzed into trimethylamine-Noxide (TMAO) in the liver [2,5,11]. However, the mechanisms underlying TMAO are still not completely understood.
Exosomes (Exos) have gained a growing concern for the pivotal roles in cardiovascular physiology [12]. Exos are nanosized membrane vesicles produced by nearly all types of cells, ranging from approximately 30 to 100 nm in diameter and containing a variety of bioactive molecules [12][13][14]. Recent research has found that Exos derived from hepatocytes play an important role in inflammation, endothelial function, and metabolic disorders [15][16][17][18][19]. In the latest work, we found that Exos secreted by TMAO-treated hepatocytes (TMAO-Exos) contained a distinctive profile of miRNAs compared to those from the TMAO-free group (control-Exos), and furthermore, TMAO-Exos could notably promote inflammation, damage vascular endothelial cells (VECs), and impair endothelium-dependent vasodilation [20]. However, understandings of how these Exos are produced by hepatocytes, where are they distributed in vivo, and what effects will they have on vascular inflammation remain limited.
In the present study, we first obtained the transcriptome profiles of primary rat hepatocytes stimulated with or without TMAO [21] and then identified the hub genes related to Exos. Next, Exos were isolated from hepatocytes treated with or without TMAO and given intravenously to mice, and then their distributions in the heart, liver, aorta, and gastrocnemius muscle and effects on vascular inflammation were examined.

Hepatic Gene Expressions Induced by TMAO.
e transcriptome profiles of primary rat hepatocytes stimulated with or without TMAO (50 μmol/L) for 30 h were obtained from the GSE135856 dataset in the Gene Expression Omnibus repository [21,22]. e microarray experiments were performed using a single Affymetrix Clariom S Rat array (Affymetrix, Santa Clara, CA, USA) with three biological repetitions.
e open-source Bioconductor packages, affy and limma, were used to process the data [21].

Enrichment Analysis.
e differentially expressed genes (P < 0.05) related to Exos were screened and visualized on a heatmap constructed in R. DAVID (Database for Annotation, Visualization, and Integrated Discovery) was used for investigating the functional annotation of genes. GO (gene ontology) analysis was performed to elaborate the biological process, STRING (Search Tool for the Retrieval of Interacting Genes/Proteins) database (v11.0) [23] was used for analyzing the protein-protein interaction (PPI), and networks were constructed on Cytoscape platform (v3.8.2) [24]. CytoHubba plugin was used to identify hub genes with a threshold value >0.4, and the color of the nodes represented the degree of gene interaction.

Cell Culture and Exos Isolation.
e procedures were performed as described in our recent study [20]. In brief, AML12 cells (iCell Bioscience Inc., Shanghai) were cultured in DMEM/F12 (iCell Bioscience Inc., Shanghai) containing Exos-depleted serum (ViVaCell, Shanghai) and treated with TMAO (Tokyo Chemical Industry Co., Ltd.) at a physiological concentration of 50 μmol/L for 48 h (TMAO-Exos). e untreated group served as the control (control-Exos). Exos were isolated and purified from the culture supernatant using differential centrifugation and then resuspended in PBS.

Exos Identification and Labelling.
e protein levels of the Exos were measured using the BCA protein assay kit ( ermo Fisher Scientific, MA, USA). e ultrastructure of the Exos was inspected using a transmission electron microscope (JEM 1200-EX, Japan). In brief, Exos suspensions were loaded on 200-mesh formvar-coated grids and then negatively stained with phosphotungstic acid. e samples were observed under a transmission electron microscope at a voltage of 100 kV. e concentration and size distribution of the Exos were detected by nanoparticle tracking analysis (Nanosight NS300, Malvern, UK). Exosomal markers of CD9 and TSG101 and the negative marker of calnexin were detected by western blotting. In brief, the samples were separated by SDS-PAGE and transferred onto Millipore polyvinylidene difluoride membranes. e membranes were incubated overnight at 4°C with the primary antibodies of CD9 (Zenbio, Chengdu, China), TSG101 (Zenbio, Chengdu, China), and calnexin (Affinity Biosciences, Jiangsu, China) and visualized with enhanced chemiluminescence reagent (Millipore, MA, USA). Exos were labelled with DiI (Beyotime Biotechnology, Shanghai, China) for in vivo tracer experiments.

Animal Experiments.
All experiments conform to the protocols approved by the Institutional Animal Care and Use Committee of Guangdong Provincial People's Hospital. Male wild-type C57BL/6 mice (six weeks old) were purchased from the Experimental Animal Center of Sun Yat-sen University. e mice were intravenously injected with 30 μg of Exos resuspended in 100 μL of PBS. After 24 h, the mice were anesthetized by pentobarbital sodium (50 mg/kg), and then the hearts, thoracic aortas, livers, and gastrocnemius muscles were collected for subsequent study.

Fluorescence Detection.
e hearts, thoracic aortas, livers, and gastrocnemius muscles were fixed in 4% paraformaldehyde and then dehydrated and embedded in paraffin. Samples underwent dewaxing and antigen retrieval. e slides were blocked in 10% goat serum for 30 min at room temperature and then incubated with the primary antibody of CD31 (Abcam, MA, USA) overnight. Slides were incubated with Alexa Fluor® 488 donkey anti-rabbit IgG (H + L). Wheat germ agglutinin (WGA, Sigma-Aldrich, USA) staining was used to outline the cardiomyocytes. Slides were then washed and stained with DAPI (Solarbio, Beijing, China). e positive signals were detected with a fluorescence microscope (Olympus, Tokyo, Japan).

2.7.
Quantitative Polymerase Chain Reaction. Quantitative polymerase chain reaction (qPCR) was performed as described in our previous study [25]. In brief, total RNA was extracted using TRIzol reagent (Invitrogen, USA), and concentration was measured using the NanoDrop 2000 spectrophotometer ( ermo Fisher Scientific, MA, USA). en, RNA was reversely transcribed into cDNA using Color Reverse Transcription Kit (EZBioscience, CA, USA), and qPCR was performed on Bio-Rad CFX-96 (Bio-Rad, CA, USA) with Color SYBR Green qPCR Master Mix (EZBioscience, CA, USA). e expressions were normalized to Gapdh. e qPCR primers used in the study are listed in Table 1.

Statistical
Analysis. Statistical analysis was conducted using SPSS 20.0 software (SPSS Inc., Chicago, IL, USA), and the graphs were plotted by GraphPad Prism (version 9.3.0, San Diego, USA). Data were presented as mean ± standard error of the mean (SEM). For continuous variables with normal distribution, the comparisons between two groups were performed with independent t-test. For continuous variables with nonnormal distribution, the comparisons between two groups were performed with Wilcoxon ranksum test. A P value <0.05 was considered statistically significant.

Exos-Related Genes in Hepatocytes
Were Dysregulated by TMAO. Microarray assay was performed to determine the expression profiles of genes in primary rat hepatocytes stimulated with TMAO. Compared to the untreated group, a total of 101 genes related to Exos changed significantly (P < 0.05), and among these genes, 43 were upregulated and 58 were downregulated (Figure 1(a)). GO analysis was used to explore the predominant biological processes, and the results showed that the differentially expressed genes (DEGs) were significantly enriched in secretion, peptide metabolic process, export from the cell, membrane fusion, and small-molecule biosynthetic process (Figure 1(b)). Besides, the PPI networks were constructed to further elaborate the interactions among the DEGs and identified Phgdh, Mdh2, Echs1, Rap2a, Gpd1l, Slc3a2, Hyou1, Hspa9, Psat1, and Pck2 as the top 10 hub genes (Figure 1(c)).

Verification of the Hub Genes.
To verify the changes of the hub genes, AML12 cells were stimulated with or without TMAO for 48 h, and the mRNA levels of the top 10 hub genes were detected by qPCR. It was shown that TMAO remarkably enhanced the mRNA expressions of Phgdh, Mdh2, Echs1, Rap2a, and Gpd1l (Figures 2(a)-2(e)) and reduced Slc3a2 levels (Figure 2(f )). However, Hyou1, Hspa9, Psat1, and Pck2 remained unchanged (Figures 2(g)-2(j)).

Isolation and Identification of Exos.
Nanovesicles with a cup-shaped morphology and a typical size around 100 nm were isolated and purified from the cell culture supernatant (Figure 3(a)). e size distribution profiles and concentrations of the Exos showed no significant differences between the two groups ( Figure 3(b)). Exosomal markers of CD9 and TSG101 were mainly enriched in control-Exos and TMAO-Exos, and the negative marker of calnexin was detected only in the whole cell lysate (Figure 3(c)).

Exos Were Widely Distributed In Vivo with Different
Patterns. TMAO-Exos were labelled with DiI and administered intravenously. e results clearly showed that Exos could be efficiently taken up by cardiomyocytes (Figure 4(a)), hepatocytes (Figure 4(b)), and endothelial cells in the aorta (Figure 4(c)). It was worth noting that these Exos were predominantly localized in the nuclei of hepatocytes. And unlike in cardiomyocytes, these Exos appeared to keep themselves out of the skeletal muscle cells and preferentially located in the endothelial cells, as indicated by the same subcellular localizations of Exos and CD31 proteins in the gastrocnemius muscle (Figure 4(d)).

TMAO-Exos Promoted Vascular Inflammation.
e in vivo tracer experiments clearly showed that TMAO-Exos could target endothelial cells in the aorta and gastrocnemius

Discussion
In the current work, we first identified the hub genes in hepatocytes after TMAO treatment, which may be related to the production of TMAO-Exos. Next, we provided evidence that TMAO-Exos could be taken up by several tissues including the heart, liver, aorta, and gastrocnemius muscle with different distribution patterns. And finally, we found that TMAO-Exos, but not control-Exos, could significantly promote the mRNA expressions of Tnf, Icam1, Sele, and Cox-2 in the aorta. ere is a close relationship between TMAO and the liver. TMAO, at a physiological concentration of 50 μmol/L, has been shown to be able to target hepatocytes and result in the abnormal expressions of hepatic genes, thus exerting an influence on metabolic disorder [21]. Moreover, we recently found that TMAO (50 μmol/L) could stimulate hepatocytes to release functional Exos [20]. However, the molecular pathways and hub genes behind it remain unknown. It was found that the fusion of multivesicular bodies (MVBs) with the plasma membrane is critical for the extracellular secretion of Exos, and intracellular molecules and pathways such as Rab27a/b and V-ATPase-mediated acidification of MVBs are implicated [26]. An interference study showed that HGS, Alix, TSG101, and nSmase2 were necessary for the Exos production by normal or ethanol-treated hepatocytes [27]. And other studies revealed that Tsg101 and Vps4a played a pivotal role in exosomal secretion [28,29]. Based on the aforementioned study [21], a Exos-related gene set was selected to identify the predominant biological processes and hub genes which responded to TMAO treatment. e results showed that the DEGs were mainly enriched in secretion, peptide metabolic process, export from the cell, membrane fusion, and small-molecule biosynthetic process, which suggested that these DEGs may be related to the synthesis and secretion of TMAO-Exos. Furthermore, we verified that Phgdh, Mdh2, Echs1, Rap2a, and Gpd1l were upregulated, and Slc3a2 was downregulated by TMAO. Phgdh and Mdh2 are classified into the biological process of "small-molecule biosynthetic process." e protein encoded by Phgdh is essential for serine synthesis and contributes to glucose and lipid homeostasis [30][31][32]. And gene Mdh2 encodes a mitochondrial enzyme that plays a pivotal role in the malateaspartate shuttle and glucose homeostasis [33,34]. In addition, Slc3a2 encodes a membrane protein and is involved in amino acid transport and endoplasmic reticulum stress [35][36][37]. However, the results simply provided some early clues about how TMAO may stimulate hepatocytes to  produce Exos, and the exact mechanisms remain to be further elucidated by using the gain-and loss-of-function strategies.
Similarly, the distribution patterns of Exos remain poorly defined. It was suggested that the processes may depend on the distinctive factors or receptors located on the recipient cell surfaces [38]. However, Horibe et al. found that Exos could be nonselectively incorporated into recipient cells via a different mechanism, which may depend on the recipient cells rather than the donor cells [39]. In the present work, we showed that TMAO-Exos could be distributed in tissues including the heart, liver, aorta, and gastrocnemius muscle after systemic administration, suggesting that these Exos spread widely in vivo through the bloodstream. However, it seemed that the distribution of these Exos varied depending on the recipient cell type, as indicated by the observations that they could be incorporated by cardiomyocytes, but appeared to keep themselves out of the skeletal muscle cells and preferentially located in the endothelial cells. In addition, it was interesting to see that these Exos were able to be captured by hepatocytes and localized in the nuclei. ese findings were consistent with previous research, in which Exos were able to be captured by the donor cells themselves [27,39]. Furthermore, we found that TMAO-Exos could significantly promote the mRNA expressions of Tnf, Icam1, Sele, and Cox-2 in the aortas. ese proinflammatory genes have been shown to play pivotal roles in the regulation of inflammation, endothelial dysfunction, and atherosclerosis [40][41][42][43].

Conclusions
In the current study, we provided clues about how TMAO may stimulate hepatocytes to produce Exos and further offered evidence that Exos secreted by TMAO-treated hepatocytes could be widely distributed in vivo and promote vascular inflammation. ese findings may help to advance our understanding of the mechanisms by which TMAO promotes CVD.

Data Availability
e data relevant to this study are available from the corresponding author upon reasonable request.

Conflicts of Interest
e authors declare that they have no conflicts of interest regarding the publication of this paper.

Authors' Contributions
Jimei Chen contributed to the concept, designed the study, directed the research, and revised the manuscript. Xiang Liu, Jiazichao Tu, Ziqin Zhou, Bingxin Huang, and Jianrong Zhou performed the experiments and analyzed the data. Xiang Liu drafted the manuscript. All authors read and approved the final manuscript.