Abstract

   graphic file with name eh3c00168_0007.jpg

   Epidemiology has associated fine particulate matter (PM[2.5]) exposure
   with an increased cardiovascular risk. However, the underlying
   mechanism, particularly from the liver perspective, remains unclear.
   Here, the influence of chronic PM[2.5] exposure on cardiovascular risk
   in mice fed a high-fat and high-cholesterol diet (HFCD) was studied by
   using a real-world PM[2.5] exposure system. Results showed that PM[2.5]
   exposure elevated the serum levels of nonhigh-density lipoprotein
   cholesterol (non-HDL-C) and oxidized low-density lipoprotein (oxLDL) in
   HFCD-fed mice, demonstrating increased cardiovascular risk. To
   investigate the molecular mechanism, lipidomics and metabolomics
   analyses were conducted and revealed that PM[2.5] exposure enhanced
   lipid accumulation and disturbed purine metabolism and glutathione
   metabolism in the liver of HFCD-fed mice, contributing to the elevated
   non-HDL-C levels and intensified oxidative stress. Moreover, PM[2.5]
   exposure increased total cholesterol levels by upregulating Hmgcr
   expression and downregulating Cyp7a1 expression in the livers of
   HFCD-fed mice. The HDL-C level was reduced by inhibiting the hepatic
   Abca1 and Abcg1 expression and decreasing the levels of ApoA-I and
   LCAT. Additionally, the PM[2.5]-induced pro-oxidative environment
   impeded the oxLDL clearance and further triggered inflammation, in turn
   exacerbating oxidative stress and oxLDL production. This study
   demonstrated a synergy of PM[2.5] and HFCD on cardiovascular risk and
   illuminated the molecular mechanism in PM[2.5]-susceptible populations.

   Keywords: PM[2.5], HFCD, Metabolomics, Lipidomics, Cardiovascular risk

Introduction

   Ambient particulate matter (PM) pollution poses a substantial global
   public health threat and ranked as the fourth leading risk factor for
   death worldwide in 2019.^[42]1 Notably, fine PM with an aerodynamic
   diameter of less than 2.5 μm (PM[2.5]) was of particular concern due to
   its ability to directly infiltrate alveoli and the circulatory system.
   Epidemiological investigations have demonstrated a strong association
   between PM[2.5] exposure and the development of metabolic
   diseases.^[43]2−[44]4 Moreover, the excessive intake of fat and
   cholesterol was reported to be a primary risk factor for metabolic
   dysfunctions. It was noteworthy that metabolic dysfunctions lead to a
   cascade of health problems, ranging from obesity and diabetes to
   nonalcoholic fatty liver (NAFLD) and cardiovascular
   disease.^[45]5,[46]6 For example, short-term exposure to a high-fat and
   high-cholesterol diet (HFCD) triggered dyslipidemia, leading to serious
   liver damage in rats.^[47]7 A cohort study found the intake of HFCD
   increased cardiovascular risk and all-cause mortality in a
   dose–response manner.^[48]8 Therefore, PM[2.5] and HFCD may interact in
   the pathogenic process. PM[2.5] exposure was found to aggravate adverse
   effects on the cardiovascular system and spleen through disturbing
   energy metabolism and lipid metabolism in mice subjected to a high-fat
   diet.^[49]9,[50]10 Guan et al. observed systemic inflammation and
   vascular oxidative stress induced by PM[2.5] in mice on a
   high-cholesterol diet, exacerbating intracranial
   atherosclerosis.^[51]11 Besides, the liver is believed to be the
   primary organ in lipid and cholesterol metabolism as well as an
   essential player in detoxification. The related metabolic diseases in
   both parts, such as NAFLD and atherosclerosis, often co-occur and are
   associated with each other.^[52]12 For example, NAFLD may drive
   atherosclerotic cardiovascular disease via mixed hyperlipidemia and the
   hepatic secretion of procoagulant factors.^[53]12 However, studies
   assessing the association between hepatotoxicity and cardiovascular
   risk caused by the combination of PM[2.5] and HFCD are lacking.

   Metabolomics was an effective method to monitor the overall change of
   small molecule metabolites in response to external stimuli, which has
   recently been used to study the mechanisms of metabolic
   dysfunction.^[54]13,[55]14 As one significant branch, lipidomics is a
   powerful means for the revelation of lipid metabolite change.^[56]15
   The combined application of metabolomics and lipidomics provides a
   comprehensive and systematic insight into the underlying metabolic
   mechanisms in organisms induced by exogenous substances in toxicity
   risk assessment. For example, a metabolism analysis of human lung
   fibroblasts demonstrated fine particulate matter exposure diminished
   mitochondria-related metabolites and exacerbated the accumulation of
   lysoglycerophospholipids, thus causing apoptosis and inflammatory
   response.^[57]16 In addition, the liver plays a critical role in
   maintaining metabolic homeostasis and even serves vascular
   function.^[58]17 However, PM[2.5]-induced hepatic responses of
   endogenous metabolites, especially in combination with HFCD, have not
   been sufficiently investigated.

   In this study, a real-ambient PM[2.5] exposure system was adopted
   because it could more realistically simulate human exposure scenarios
   than the traditional intratracheal instillation method.^[59]9,[60]10
   The influence of ambient PM[2.5] on cardiovascular risk was
   investigated in the context of cross-organ communications by comparing
   normal mice and mice fed with HFCD. Biochemical analysis, metabolomics,
   and lipidomics were integrated to further elucidate the potential
   toxicity mechanisms.

Methods and Materials

Animal Experiments

   Due to the susceptibility of the females to PM[2.5]-induced
   perturbation of blood lipids and hepatic lipid
   metabolism,^[61]18,[62]19 female C57BL/6 mice aged 4 weeks were
   obtained from Beijing Vital River Laboratory Animal Technology Co.,
   Ltd. (China) and housed in a standard animal house (24 ± 2 °C, 40–60%
   humidity, and 12 h light/dark cycle) located in Shanxi, China. The
   mouse experiments were compliant with the Animal Conservation and Use
   Committee of Shanxi University. After acclimatization for 1 week, the
   mice were weighed and randomly divided into four groups (n = 6 per
   group). Two groups were fed a normal diet (ND, 10% energy from fat) and
   exposed to either filtered air (FA) or PM[2.5] through a real-ambient
   PM[2.5] exposure system. This exposure system has two chambers, one
   fitted with a high-efficiency filter to deliver FA and the other
   allowing airborne particles to pass through to deliver PM[2.5].^[63]10
   While the other two groups were fed an HFCD (34% energy from fat, 1.25%
   cholesterol, 0.5% cholic acid) and exposed to either FA or PM[2.5]. The
   diets were obtained from Trophic Animal Feed High-Tech Co., Ltd.,
   China. The four groups of mice were designated as ND-FA, ND-PM[2.5],
   HFCD-FA, and HFCD-PM[2.5] accordingly. Mice were exposed for six months
   with water supplied ad libitum and then sacrificed with blood and liver
   collected. Serum samples were collected by centrifugation at 3000 rpm
   for 10 min. The liver and serum samples were immediately stored at −80
   °C for further use. The average concentration of PM[2.5] was 64.1 ±
   42.5 μg/m^3 during the entire exposure period, and the details are
   shown in [64]Figure S1.

Measurements of Biochemistry Parameters in Serum and Liver Samples

   The levels of total cholesterol, high-density lipoprotein cholesterol
   (HDL-C), low-density lipoprotein cholesterol (LDL-C) in serum, and
   malondialdehyde (MDA) in the liver were determined by using assay kits
   (Jiancheng Bioengineering Institute, China). Nonhigh-density
   lipoprotein cholesterol (non-HDL-C) was quantified by subtracting HDL-C
   from total cholesterol. The serum levels of high-density lipoprotein
   (HDL), lecithin cholesterol acyltransferase (LCAT), oxidized
   low-density lipoprotein (oxLDL), high-sensitivity C-reactive protein
   (hs-CRP), and monocyte chemotactic protein 1 (MCP-1) and the levels of
   apolipoprotein A-I (apoA-I) and oxLDL in liver samples were measured
   using Elisa kits (Elabscience Biotechnology, China). All procedures
   were performed following the manufacturer’s protocols. The results of
   liver samples were normalized based on protein concentrations.

Histopathology Analysis of Mouse Liver

   A portion of liver and aorta samples fixed in 4% paraformaldehyde was
   embedded in paraffin, sectioned at 5 μm thick, and then stained with
   hematoxylin and eosin (H&E). The remaining liver was frozen, cut at 5
   μm, and stained with Oil Red O. The sections were observed using a
   microscope slide scanner (Axioscan 7, ZEISS), and lipid accumulation
   was quantified by ImageJ software.

Sample Preparation for Metabolomics and Lipidomics Analysis

   For endogenous metabolite extraction, the liver (10 mg) or serum (50
   μL) was homogenized with prechilled 80% methanol (400 μL). Following
   centrifugation, the supernatants were harvested, dried in an air
   vacuum, and then reconstituted with 100 μL of methanol/water (1:1, v/v)
   containing 4-chlorophenylalanine (1 μg/mL) as the internal standard.
   The quality control (QC) sample was generated by combining 40 μL of the
   solution with each sample.

   The lipid extraction procedure was conducted according to our previous
   study.^[65]7 Briefly, liver (10 mg) was homogenized in 750 μL of
   prechilled 80% methanol. Then, the solution was added with 450 μL of
   chloroform and 150 μL of water, shaken violently for 10 min, and left
   at room temperature for 10 min. After centrifugation, the lipid layer
   (bottom) was collected, vacuum-dried, and redissolved with triglyceride
   [TG (15:0/15:0/15:0)], sphingomyelin [SM (d18:1/12:0)], ceramide [Cer
   (d18:1/17:0)], phosphatidylcholine [PC (19:0/19:0)] and
   lysophosphatidylcholine [LPC (19:0)] as the internal standards for
   analysis.

Instrumental Analysis

   Nontargeted metabolomics and lipidomics analysis was performed on
   ultrahigh-performance liquid chromatography (UHPLC), equipped with a
   QExactive Focus Hybrid Quadrupole-Orbitrap Mass Spectrometer (Thermo
   Scientific, U.S.A.). Metabolites were separated on an ACQUITY UPLC HSS
   T3 column (1.8 μm, 2.1 mm × 100 mm, Waters Corporation, U.S.A.), while
   the separation of lipids depended on an ACQUITY UPLC BEH C18 column
   (1.7 μm, 2.1 mm × 100 mm, Waters Corporation, U.S.A.). Details of the
   instrument conditions are provided in [66]Table S1 and [67]Table S2.

Data Processing

   Data acquisition and preprocessing were conducted using Xcalibur v4.1
   software (Thermo Fisher Scientific, Inc., U.S.A.). For the global
   metabolomics data, the XCMS package in the R language was used in
   chromatography peak detection and alignment. A data list was output
   containing mass-to-charge ratio (m/z), peak intensity, and retention
   time. For lipidomics data, the identification and alignment of lipids
   were conducted using LipidSearch (Thermo Fisher Scientific, Inc.,
   U.S.A.). The metabolic features with relative standard deviation >30%
   in quality control samples and disturbed signals in blank samples were
   excluded to eliminate the potential interference. The peak intensities
   of the features were normalized by the internal standard method to
   remove instrumental deviation. The qualified data was analyzed with
   partial least-squares-discriminant analysis (PLS-DA) using
   MetaboAnalyst ([68]https://www.metaboanalyst.ca/). The differentiating
   metabolic features between groups were identified based on fold change
   (FC) > 1.2 (or <0.8), p < 0.05, and variable importance in the
   projection (VIP) score >1 in the PLS-DA model. Compound Discoverer
   software (Thermo Scientific) was employed to process the MS data for
   metabolite identification. The metabolite libraries used here included
   the mzCloud library, METLIN ([69]http://metlin.scripps.edu/), and Human
   Metabolome Database ([70]https://hmdb.ca/). The pivotal metabolites
   were subsequently validated by using commercial standards. Metabolic
   pathway analysis was carried out by MetaboAnalyst based on the Kyoto
   Encyclopedia of Genes and Genomes (KEGG) pathway database.

RNA Extraction and RT-qPCR

   Total RNA in liver samples was extracted with RNAiso plus (TaKaRa,
   Japan) and converted to cDNA using a PrimeScript RT Master Mix. The
   real-time qPCR was conducted using TB Green Premix Ex Taq (TaKaRa,
   Japan) on a Piko Real Real-Time PCR System (QIAquant, Germany). The
   primer sequences are shown in [71]Table S3. Each qPCR assay was
   performed in triplicate.

Statistical Analysis

   Statistical analyses were conducted with SPSS version 17.0. The one-way
   ANOVA post hoc test using Tukey’s honestly significant difference test
   was used to determine the statistical differences between groups. A
   two-way ANOVA followed by Bonferroni correction was used to examine the
   primary effects of PM[2.5] exposure and HFCD, as well as their
   interaction (i.e., PM[2.5] × diet). The results were expressed as the
   mean ± standard deviation. Statistical significance was indicated by p
   < 0.05.

Results and Discussion

Increased Non-HDL and oxLDL Levels in Serum by PM[2.5] and HFCD

   To realistically simulate human exposure, the mice fed with an ND or
   HFCD were exposed to real-world PM[2.5] in Taiyuan, Shanxi, an area of
   moderate air pollution ([72]Figure S1), for six months. Two more groups
   of mice fed with the same diets, respectively, were housed in fresh air
   as control. PM[2.5] exposure led to slightly lower weight gain in the
   early stage of the experiment. With the extension of exposure to 4 and
   6 months, the weight gain was significantly decreased in response to
   PM[2.5] exposure, especially in the mice grown with HFCD ([73]Figure
   [74]1
   A). It was reported that a high level of PM[2.5] (198.52 μg/m^3)
   induced weight loss in 40 days.^[75]20 Our results suggested that the
   accumulated adverse effect of lower ambient PM[2.5] could cause chronic
   harmful outcomes in the long term. Of note is that HFCD did not enhance
   weight gain, presumably because cholic acid in HFCD elevated energy
   expenditure by activating thermogenesis.^[76]21

Figure 1.

   [77]Figure 1
   [78]Open in a new tab

   Impacts of PM[2.5] and HFCD on physiological condition and the levels
   of cholesterol and oxLDL in mice (n = 6). (A) Body weight gain during
   exposure experiments, (B) serum cholesterol level, (C) serum oxLDL
   level, (D) serum hs-CRP level, (E) serum MCP-1 level, (F) liver/body
   weight ratio, (G) representative liver Oil Red O images, and (H)
   quantitative results of Oil Red O staining area. Letters indicate the
   results of the one-way ANOVA post hoc test at p < 0.05. Asterisks
   indicate the results of two-way ANOVA. *, p < 0.05; **, p < 0.01; ***,
   p < 0.001; and —, p ≥ 0.05.

   As shown in [79]Figure [80]1
   B, HFCD induced significant increases (p < 0.001) in the concentration
   of serum total cholesterol, HDL-C, and LDL-C, compared with that in ND
   groups. The increased HDL-C is considered a defensive response to
   HFCD.^[81]22 It is worth noting that PM[2.5] exposure further increased
   the total cholesterol level but decreased the HDL-C level when
   comparing the two HFCD groups. Moreover, a synergistic interaction
   between PM[2.5] and HFCD significantly affected the total cholesterol
   and HDL-C (p < 0.01). Similarly in a recent study performed in a
   Chinese rural population, epidemiological data showed that a 1 μg/m^3
   increment of PM[2.5] corresponded to a 0.10% (0.07–0.19%) increase in
   total cholesterol and a 0.49% (0.38%–0.60%) decrease in HDL-C.^[82]23
   HDL-C has potentially cardiovascular-protective functions mainly due to
   HDL’s reverse cholesterol transport from the peripheral tissues to the
   liver.^[83]24 Moreover, previous studies observed strong negative
   correlations between serum HDL-C concentration and the risk for
   cardiovascular diseases.^[84]25,[85]26 The results suggested that
   PM[2.5] might exacerbate dyslipidemia and cardiovascular disease risk
   in HFCD-fed mice. Non-HDL-C and circulating oxLDL were considered
   potential biomarkers of cardiovascular disease.^[86]27,[87]28 Non-HDL-C
   represents all the proatherogenic apoB-containing lipoproteins such as
   intermediate-density lipoprotein, LDL, and VLDL.^[88]28 Given the
   alterations of total cholesterol and HDL-C, a significant increase (p <
   0.01) was found in the serum level of non-HDL-C in the HFCD-PM[2.5]
   group compared to that in the HFCD-FA group, whereas that in the ND-FA
   and ND-PM[2.5] showed no difference ([89]Figure [90]1B). We then
   measured the serum levels of oxLDL in the four groups and observed the
   change trends as the same as those of non-HDL-C ([91]Figure [92]1C).
   OxLDL is produced by LDL accumulated in the vessel walls in the
   presence of excess reactive oxygen species (ROS). Subsequently, oxLDL
   was internalized by scavenger receptors on macrophages to form foam
   cells, promoting a prothrombotic state.^[93]29 Therefore, circulating
   oxLDL is involved in early stage atherosclerotic pathogenesis by
   stimulating the formation of macrophage-derived foam cells.^[94]27
   Correspondingly, slight foam cell formation was observed in the aortic
   arch of some individuals from the HFCD-PM[2.5] group, suggesting a
   higher atherosclerotic risk in this group ([95]Figure S2). The elevated
   serum levels of non-HDL-C and oxLDL confirmed the PM[2.5]-caused
   adverse cardiovascular outcomes in mice fed with HFCD. Besides
   lipoprotein-related markers, inflammation biomarkers such as hs-CRP and
   MCP-1 were also often used for cardiovascular risk prediction.^[96]30
   As shown in [97]Figure [98]1D, a significant increase (p < 0.05) was
   observed in the serum hs-CRP level in the HFCD-PM[2.5] group compared
   to that in the HFCD-FA group. The factorial analysis showed a
   synergistic effect between PM[2.5] and HFCD on non-HDL-C, oxLDL and
   hs-CRP (p < 0.05). The results further demonstrated the higher
   cardiovascular risk caused by PM[2.5] in HFCD-fed mice.

   The liver plays a pivotal role in the maintenance of cholesterol
   balance and the elimination of circulating oxLDL.^[99]31,[100]32 As
   observed, HFCD significantly increased the relative liver weight
   compared with ND feeding, which was consistent with previous
   studies.^[101]7 PM[2.5]-treated mice exhibited lower relative liver
   weight than those exposed to FA, despite the fat or cholesterol content
   of the diets ([102]Figure [103]1
   F), which was similar to the changes in body weight ([104]Figure
   [105]1A). Together with the PM[2.5] effects on lipoproteins that were
   observed only upon the HFCD background ([106]Figure [107]1B and C),
   these results indicated that the liver was more vulnerable to PM[2.5]
   toxicity. Under our experimental conditions, PM[2.5] exposure might
   pose a markable threat to the cardiovascular system only when excessive
   fat and cholesterol were presented. Morphologically, H&E staining of
   liver sections presented macrovesicular steatosis and inflammatory cell
   infiltration in both HFCD groups, but not in the ND groups ([108]Figure
   S3). Oil Red O staining revealed that PM[2.5] exposure markedly
   increased intracytoplasmic lipid accumulation, which was
   synergistically enhanced in combination with HFCD (p < 0.05,
   [109]Figure [110]1G and H). Combined with the relative liver weight
   reduced by PM[2.5], the phenotypic results suggested that coexposure to
   PM[2.5] and HFCD may induce more severe liver damage such as lipid
   accumulation and inflammation, which may be responsible for the
   increased serum levels of non-HDL-C and oxLDL. To further study the
   underlying molecular mechanism of increased cardiovascular risk from
   the liver perspective, the abnormalities of liver endogenous metabolism
   were first investigated through lipidomics and metabolomics analysis.

Effect of PM[2.5] and HFCD on Lipid Metabolism

   PLS-DA of the liver lipidomes showed a clear separation among groups,
   indicating that HFCD and PM[2.5] treatment induced significant
   perturbation of hepatic lipids ([111]Figure [112]2
   A and B). Among the nine categories of the identified lipids, the
   abundance of seven displayed remarkable upregulation, including
   cholesteryl ester (ChE), triglyceride (TG), diacylglycerol (DG),
   phosphatidylglycerol (PG), sphingomyelin (SM), ceramide (Cer), and
   hexosylceramides (Hex1Cer) after HFCD treatment ([113]Figure [114]2C
   and [115]Figure S4). PM[2.5] exposure resulted in further increases in
   ChE, TG, SM, Cer, and Hex1Cer under the HFCD feeding conditions,
   exhibiting a synergistic interaction between these two factors (p <
   0.05). Previous studies demonstrated that PM[2.5] disrupted lipid
   metabolism and induced lipid accumulation in mouse liver by influencing
   the expression of genes involved in lipid synthesis, transport, and
   lipolysis.^[116]10,[117]33 Consistent with the changes in lipoprotein
   metabolism in serum, the hepatic accumulation of ChE in both HFCD
   groups demonstrated a serious derangement of cholesterol homeostasis in
   response to the diet shift. Also, PM[2.5] exposure aggravated this
   imbalance in the HFCD-fed mice rather than the ND-fed ones. Similarly,
   Li et al. observed PM[2.5]-induced ChE accumulation in the liver of
   female mice in the lipidomics analysis.^[118]10 These alterations might
   cause oxidative damage to mitochondria and injure hepatocytes according
   to the early reports.^[119]34 As components of neutral lipids, TG was
   strongly associated with liver diseases such as NAFLD. The TG level in
   the liver was further verified by assay kits. Consistently, the TG
   accumulation was observed in response to HFCD and PM[2.5] exposure
   ([120]Figure S5). Yan et al. found that PM[2.5] exposure increased TG
   levels in mouse livers by decreasing hepatic lipid output.^[121]35
   Besides, the excess TG could be synthesized into VLDL, an important
   component of non-HDL-C, to alleviate TG accumulation during
   NAFLD.^[122]35,[123]36 Therefore, the upregulation of TG induced by the
   cotreatment of HFCD and PM[2.5] was an important signal for
   cardiovascular disease risk and NAFLD. PG is a precursor of
   cardiolipin, a signature lipid of mitochondrial membranes in
   animals.^[124]37 The high abundance of the PG content suggested that
   HFCD and PM[2.5] treatments might disturb the optimal function of
   mitochondria. SM is a category of sphingolipid in cell membranes,
   capable of undergoing hydrolysis to produce Cer. A previous study in
   apolipoprotein E-knockout mice has reported that inhibition of SM
   biosynthesis led to decreased plasma cholesterol levels, increased
   HDL-C, and prevented atherosclerotic lesions.^[125]38 The increase of
   serum cholesterol in HFCD and PM[2.5] groups might be related to the
   accumulation of SM. Cers have multiple effects associated with
   metabolic dysfunction including apoptosis, fibrosis, oxidative stress,
   and disrupting mitochondrial bioenergetics. The inhibition of Cer
   biosynthesis in mice and rats attenuated the development of
   atherosclerotic plaque formation, hypertension, and heart
   failure.^[126]39 The accumulation of Hex1Cer in the liver was
   positively related to the risks of liver diseases, such as nonalcoholic
   steatohepatitis and chronic hepatitis C.^[127]40 Thereby the enhanced
   abundance of Hex1cer by HFCD and PM[2.5] might be a sign of liver
   diseases. Moreover, the elevated sphingolipids including Cer, SM, and
   Hex1Cer, promoted the proatherogenic effects of PM[2.5] by increasing
   the level of non-HDL-C.^[128]41 Collectively, HFCD treatment triggered
   a lipid metabolism perturbation in the liver, which was exacerbated
   upon PM[2.5] exposure.

Figure 2.

   [129]Figure 2
   [130]Open in a new tab

   Lipidomics analysis of the mouse liver (n = 6). PLS-DA plots in
   positive (A) and negative (B) ionization mode, and (C) changes in the
   overall abundance of lipid subclasses. Letters indicate the results of
   the one-way ANOVA post hoc test at p < 0.05. Asterisks indicate the
   results of two-way ANOVA. *, p < 0.05; **, p < 0.01; ***, p < 0.001;
   and —, p ≥ 0.05.

Effect of PM[2.5] and HFCD on Metabolic Pathways

   A global nontarget metabolomics analysis was performed to profile the
   change of endogenous metabolites in response to HFCD or PM[2.5]
   exposure. From the liver samples, a total of 8512 and 7034 metabolite
   features were obtained in the positive and negative ionization modes,
   respectively. Similar to the results of lipidomics, the PLS-DA revealed
   distinct differences between all four groups ([131]Figure [132]3
   A and B), suggesting that HFCD and PM[2.5] exposure conspicuously
   induced perturbations in the endogenous metabolomes. A total of 40
   endogenous metabolites were identified as metabolic biomarkers based on
   tandem mass spectra, mainly including nucleosides, nucleotides, amino
   acids, and oligopeptides ([133]Table S4). Pathway enrichment analysis
   based on KEGG pathways showed that the biomarkers exhibited strong
   associations with the disruptions in purine metabolism and glutathione
   metabolism ([134]Figure [135]3C and D).

Figure 3.

   [136]Figure 3
   [137]Open in a new tab

   Effects of PM[2.5] and HFCD on the endogenous metabolome in the mouse
   liver (n = 6). PLS-DA plots in positive (A) and negative (B) ionization
   mode, (C) pathway analysis of endogenous metabolic biomarkers in the
   liver, (D) pathway of purine metabolism and GSH metabolism, (E, F)
   changes of metabolite biomarkers in the pathway of purine metabolism
   (E) and GSH metabolism (F) in response to HFCD and PM[2.5]. Compared
   with the ND-FA group, the up- and downregulated metabolites in
   HFCD-PM[2.5] are presented in purple and green, respectively. Letters
   indicate the results of the one-way ANOVA post hoc test at p < 0.05.
   Asterisks indicate the results of two-way ANOVA. *, p < 0.05; **, p <
   0.01; **, p < 0.001; and —, p ≥ 0.05.

   As detailed in [138]Figure [139]3
   E, the combined exposure of PM[2.5] and HFCD severely disrupts purine
   metabolism. Most of the metabolic biomarkers in this pathway were
   increased in the HFCD-PM[2.5] groups. The decrease of ATP and ADP
   upstream of purine metabolism may be attributed to the upregulation of
   the downstream metabolic pathways. Correspondingly, the elevated levels
   of AMP and adenosine were observed. Purine metabolism encompasses de
   novo synthesis, salvage pathways, and catabolism. The increased levels
   of hypoxanthine, xanthine, and uric acid in the HFCD-PM[2.5] groups
   suggested elevated purine catabolism. Hypoxanthine and xanthine were
   the sources of reactive oxygen species,^[140]42 and their increase
   indicated the possibility of higher oxidative stress. Uric acid, as an
   end product of purine metabolism, is considered a marker of oxidative
   stress.^[141]43 Herein, uric acid was significantly upregulated to
   1.35-fold in HFCD-FA groups and 2.09-fold in HFCD-PM[2.5] groups,
   respectively, indicating a potential risk of oxidative stress in
   response to HFCD and PM[2.5]. Furthermore, a significantly synergistic
   interaction between PM[2.5] and HFCD existed in elevating uric acid
   levels (p < 0.05). The consistent increase in both hypoxanthine and
   xanthine caused by PM[2.5] exposure in the HFCD groups suggested that
   this pathway was upregulated and produced potential oxidative stress.
   Moreover, epidemiologic studies have reported a strong association
   between high uric acid and cardiovascular disease risk.^[142]44 The
   perturbations in purine metabolism might be one of the contributors.

   In addition, HFCD and PM[2.5] exposure disordered hepatic glutathione
   (GSH) metabolism in mice ([143]Figure [144]3
   F). GSH is a physiological antioxidant deeply involved in maintaining
   cellular redox homeostasis. The pool of GSH was significantly shifted
   from the reduced form (GSH) to the oxidized form (GSSG) in response to
   HFCD exposure, especially coexposure with PM[2.5]. The elevation of
   pyroglutamate was also related to GSH depletion in the HFCD-PM[2.5]
   groups. Moreover, the ratio of GSH/GSSG, a critical redox couple, was
   dramatically decreased, indicating a potential disruption of
   intracellular redox balance. Glutamic acid not only is a main component
   for glutathione synthesis but also plays an important role in de novo
   purine synthesis.^[145]45 The significant increase in glutamic acid in
   response to HFCD and PM[2.5] exposure may contribute to the
   upregulation of purine metabolism.

   We also carried out a similar metabolomics comparison in the serum from
   the four groups ([146]Figure S6A and B). The altered metabolites were
   mainly enriched in the pentose phosphate pathway, glycerophospholipid
   metabolism, pentose and glucuronate interconversions, and purine
   metabolism after HFCD and PM[2.5] treatments ([147]Figure S6C). These
   results indicated the disorders in glucose and lipid metabolism that
   were caused by the cotreatment. Although it was only repeating the
   findings from the liver, the synergistic response to HFCD and PM[2.5]
   treatments was observed in the serum, evident with a decreased GSH/GSSG
   ratio (0.49 folds) and increased uric acid levels (2.29 folds)
   ([148]Figure S6D and E). These revealed that PM[2.5] triggered
   oxidative stress in the cotreated mice, which was consistent with the
   alterations found in the liver. As a major metabolic organ, this
   consistency suggested that the toxic changes in the liver contributed
   significantly to the alterations in the serum. It is worth noting that
   the serum level of homocysteine, a biomarker of cardiovascular disease,
   was significantly upregulated to 2.23-fold in HFCD-PM[2.5] groups
   ([149]Figure S6F), indicating the elevated cardiovascular risk in
   response to HFCD and PM[2.5] treatments.

Effect of PM[2.5] and HFCD on Cholesterol Homeostasis

   The alterations found in lipidomics analysis indicated a disrupted
   cholesterol metabolism in the mouse liver. To examine the potential
   mechanistic links between these changes in liver metabolism and the
   increased serum cholesterol caused by PM[2.5] and HFCD, we assessed the
   expressions of genes involved in cholesterol biosynthesis, transport,
   and metabolism in the liver. The expression of
   3-hydroxy-3-methylglutaryl-coenzyme A reductase (Hmgcr), the
   rate-limiting enzyme in cholesterol biosynthesis, was evidently
   suppressed (p < 0.01) to 0.23- and 0.38-fold in HFCD groups, whereas
   PM[2.5] exposure induced the upregulation of Hmgcr mRNA (p < 0.05,
   [150]Figure [151]4
   A). In vivo studies have shown that the hepatic Hmgcr level was
   downregulated as feedback regulation of cholesterol biosynthesis to the
   high-fat and high-cholesterol diets;^[152]46,[153]47 thus, the elevated
   serum cholesterol levels may be primarily attributed to lymphatic
   absorption of dietary cholesterol. The PM[2.5]-caused upregulation of
   Hmgcr expression might further break the cholesterol homeostasis in the
   HFCD mice. Cytochrome P450 7A1 (Cyp7a1) was the first and the
   rate-limiting enzyme in bile acid synthesis, which facilitates
   cholesterol elimination.^[154]48 Obviously, HFCD significantly
   inhibited the expression of Cyp7a1 (0.55-fold), especially in the
   presence of PM[2.5] (0.16-fold, p < 0.01). This synergistic interaction
   (p < 0.05) indicated that coexposure to HFCD and PM[2.5] might inhibit
   the conversion of cholesterol into bile acids by downregulating Cyp7a1
   expression, thereby enhancing the cholesterol accumulation in the
   liver. Ando et al. reported that HFCD may decrease Cyp7a1 mRNA levels
   in mouse liver via the FXR-mediated pathway.^[155]49 Both HFCD and
   PM[2.5] showed no significant effect on the gene expression of Abcg5
   and Abcg8, the heterodimers encoded by which mediated the excretion of
   hepatocyte cholesterol into bile.^[156]50 The HFCD groups markedly
   downregulated the expression of LDL receptor (Ldlr) that transferred
   LDL-C from blood to the liver,^[157]51 while no significant difference
   was observed under PM[2.5] exposure. Besides, the total cholesterol
   levels in the liver were measured, and cholesterol accumulation was
   observed in the livers of the HFCD-PM[2.5] group ([158]Figure S7).
   These results revealed that in mice fed HFCD, PM[2.5] exposure might
   increase cholesterol production via elevating Hmgcr expression and
   decreasing its consumption by inhibiting Cyp7a1 expression in the
   liver, leading to a higher level of serum cholesterol.

Figure 4.

   [159]Figure 4
   [160]Open in a new tab

   Effect of PM[2.5] and HFCD on cholesterol homeostasis in the liver (n =
   6). (A) Hepatic expression of genes involved in cholesterol synthesis
   and metabolism, (B) hepatic expression of genes involved in the HDL
   biosynthesis, (C) ApoA-I level in the liver, (D) LCAT levels in serum,
   and (E) HDL level in serum. Letters indicate the results of the one-way
   ANOVA post hoc test at p < 0.05. Asterisks indicate the results of
   two-way ANOVA. *, p < 0.05; **, p < 0.01; ***, p < 0.001; and —, p ≥
   0.05.

   HDLs are produced in the liver and are known to transport excess
   cholesterol in the vessel wall back to the liver for elimination,
   providing an atheroprotective effect.^[161]52 Therefore, the hepatic
   expressions of genes related to HDL synthesis were measured.
   ATP-binding cassette transporter A1 (Abca1) and ATP-binding cassette
   transporter G1 (Abcg1) promote cholesterol efflux from cells to nascent
   and mature HDL particles, respectively. Scavenger receptor class B type
   I (SR-B I, encoded by the Scarb1) as HDL receptor mediates the
   selective uptake of esterified cholesterol from mature HDL.^[162]53 It
   was reported that hepatic overexpression of Abca1 contributed to
   elevated HDL-C levels in plasma, while an opposite effect was observed
   for overexpression of SR-BI in Ldlr-deficient mice.^[163]54,[164]55 The
   absence of Abcg1 resulted in the decreased plasma concentration of
   HDL-C in mice fed a high-cholesterol diet.^[165]56 As expected, HFCD
   significantly stimulated the expression of Abca1 and Abcg1 by 1.85- and
   3.98-fold, respectively, although no notable difference could be read
   in the expression of Scarb1 ([166]Figure [167]4
   B). However, PM[2.5] decreased the mRNA levels of Abca1 and Abcg1 under
   the HFCD feeding conditions, indicating that PM[2.5] exposure might
   impede hepatic HDL synthesis through gene regulation. In addition, the
   mRNA levels of Abca1 and Abcg1 were synergistically modulated by HFCD
   and PM[2.5].

   ApoA-I is the major apolipoprotein of nascent HDL and the most
   effective activator of LCAT. LCAT mediates the transformation of free
   cholesterol into cholesterol esters to form mature HDL
   particles.^[168]57 Similarly, the levels of ApoA-I and LCAT were
   increased in response to HFCD but decreased with PM[2.5] exposure (p <
   0.05, [169]Figure [170]4
   C and D). Also, similar alteration trends in the effects of HFCD and
   PM[2.5] were observed for serum concentrations of HDL particles
   ([171]Figure [172]4E). Taken together, it can be speculated that in
   response to HFCD intake, HDL synthesis was improved by upregulating
   Abca1 and Abcg1 expression and stimulating ApoA-I and LCAT synthesis.
   However, the superimposed effect of PM[2.5] disrupted this process and
   hindered HDL synthesis. The impaired production of HDL particles
   reduced its capacity for cholesterol removal, resulting in lower serum
   HDL-C. Combined with the increased cholesterol production, this might
   explain the increased levels of non-HDL-C in the presence of coexposure
   to HFCD and PM[2.5].

Oxidative Stress and Inflammatory Responses Induced by PM[2.5] and HFCD

   Based on the metabolomics analysis, it was hypothesized that the
   cotreatment of HFCD and PM[2.5] might lead to oxidative stress and a
   potentially pro-oxidative environment in mouse liver. The genes related
   to oxidative stress, including erythroid 2-related factor 2 (Nrf2),
   NAD(P)H quinone oxidoreductase-1 (Nqo-1), heme oxygenase-1 (Ho-1), and
   mitochondrial superoxide dismutase (Mnsod), were determined in the
   liver. During oxidative stress, the redox-sensitive transcription
   factor Nrf2 translocates into the nucleus, which in turn upregulates
   the expression of antioxidant defense genes (Ho-1, Mnsod) and phase II
   detoxification enzymes (NQO1).^[173]58 In our results, HFCD feeding
   significantly enhanced the expressions of Nrf2, Ho-1, and Nqo-1. In the
   HFCD-fed mice, PM[2.5] exposure not only exacerbated this trend but
   also demonstrated a pronounced synergistic interaction with HFCD
   ([174]Figure [175]5
   A). These results suggested that the cotreatment of HFCD and PM[2.5]
   stimulated the antioxidant defense system as an adaptive response. In
   addition, MDA, the peroxidation product of polyunsaturated fatty acids,
   was a classical biomarker of oxidative stress.^[176]59 A significant (p
   < 0.01) elevation of hepatic MDA levels was observed in mice fed HFCD,
   especially in those exposed to both HFCD and PM[2.5] ([177]Figure
   [178]5B), confirming the more severe oxidative injury caused by PM[2.5]
   upon HFCD feeding. Similarly, Guan et al. observed that PM[2.5]
   exposure induced oxidative damage and upregulated the mRNA levels of
   Nrf2, Ho-1, and Nqo-1 in rats fed with a high-cholesterol diet.^[179]11
   Collectively, the cotreatment of HFCD and PM[2.5] perturbed purine and
   GSH metabolisms, resulting in oxidative stress in mice liver. This
   pro-oxidation environment might cause higher oxidative stress in the
   circulation and oxidize more LDL, leading to increased oxLDL
   production.^[180]27 OxLDL is a major early product of lipid
   peroxidation, and the oxLDL in circulation is primarily cleared and
   metabolized by the liver.^[181]60 As shown in [182]Figure [183]5C,
   hepatic oxLDL levels were significantly (p < 0.01) enhanced by HFCD
   feeding, which was further increased when PM[2.5] treatment was
   combined. This trend was consistent with the serum oxLDL levels,
   indicating that the accumulation of oxLDL in the liver might be
   attributed to the serum alterations. Oxidation of LDL is a complicated
   process involving the oxidative changes of both proteins and lipids
   through enzymatic and nonenzymatic pathways, as well as the formation
   of complex products.^[184]27 OxLDL was a highly immunogenic molecule
   responsible for cellular GSH depletion.^[185]32 More oxLDL that went to
   the liver possibly contributed to the above metabolomics observation of
   the decreased GSH/GSSG ratio under the combined HFCD and PM[2.5]
   treatment, worsening the oxidative stress of the whole body.
   Additionally, cholesteryl esters from oxLDL were rapidly absorbed by
   Kupffer cells and delivered to the liver parenchymal cells, which in
   turn led to further cholesterol accumulation.^[186]60 The increased
   oxLDL level in HFCD and PM[2.5] groups was consistent with the
   variation of cholesteryl esters content observed in lipidomics
   analysis.

Figure 5.

   [187]Figure 5
   [188]Open in a new tab

   Effect of PM[2.5] and HFCD on oxidative stress and pro-inflammatory
   cytokines in the liver (n = 6). (A) Hepatic expression of genes
   associated with oxidative stress, (B) relative level of MDA in the
   liver, (C) oxLDL level in the liver, and (D) gene expression of
   pro-inflammatory cytokines in the liver. Letters indicate the results
   of the one-way ANOVA post hoc test at p < 0.05. Asterisks indicate the
   results of two-way ANOVA. *, p < 0.05; **, p < 0.01; ***, p < 0.001;
   and —, p ≥ 0.05.

   Oxidative stress and inflammation are causal and mutually reinforcing
   in various pathological states, forming a vicious cycle.^[189]61
   Moreover, oxLDL is trapped in lysosomes and stimulates the generation
   of inflammatory cytokines in Kupffer cells.^[190]62 Therefore, the
   pro-inflammatory cytokines, including interleukin-6 (Il-6),
   interleukin-1β (Il-1β), and tumor necrosis factor α (Tnfα), were
   measured in mouse liver. HFCD upregulated mRNA expressions of Il-6,
   Il-1β and Tnfα, and a synergetic effect of PM[2.5] and HFCD was
   observed on the pro-inflammatory cytokines ([191]Figure [192]5
   D), indicating PM[2.5] further exacerbated liver inflammation in mice
   fed with HFCD. The liver plays a key role in managing oxLDL, which,
   when accumulated, can trigger inflammation in liver cells.^[193]63
   Excessive inflammatory cells may produce more ROS and reactive nitrogen
   species, which in turn lead to more severe oxidative stress at the
   inflammatory lesion.^[194]61 Production of oxLDL, along with increased
   oxidative stress and inflammation, tended to create harmful cycles that
   worsen each other. With PM[2.5] exposure, this situation was aggravated
   due to additional oxidative stress caused by PM[2.5] hepatotoxicity.
   Therefore, it seemed that these negative changes likely occurred
   simultaneously and intensified with each other. The altered redox
   status in the liver contributed to the oxidative modification of LDL,
   which then triggered the inflammatory response. Previous studies
   observed both oxidative damage and inflammatory response in the liver
   of mice exposed to PM[2.5] (20 mg/mL) using an intratracheal
   instillation method.^[195]32 Therefore, the liver inflammation
   exacerbated by PM[2.5] exposure might aggravate oxidative stress in
   both the liver and the serum of mice fed HFCD, which was responsible
   for the elevated serum oxLDL levels observed above.

Conclusion

   This study explored the impacts and underlying mechanisms of PM[2.5]
   and HFCD on the cardiovascular risk in mice. PM[2.5] exposure elevated
   the serum levels of non-HDL-C and oxLDL, raising the cardiovascular
   risk in mice fed with HFCD. As summarized in [196]Figure [197]6
   , PM[2.5] exposure increased the cholesterol level and decreased the
   HDL-C level in mice fed with HFCD by regulating the expression of
   related genes and proteins, such as Hmgcr and Cyp7a1. The perturbation
   of lipid metabolism observed by lipidomics analysis was also associated
   with the enhanced non-HDL-C level. In addition, the cotreatment of
   PM[2.5] and HFCD created a pro-oxidation environment in the liver, as
   evidenced by the disturbance of purine metabolism and glutathione
   metabolism. Oxidative stress combined with mutually reinforcing
   inflammation resulted in oxidative modification of LDL and the
   accumulation of oxLDL. Our results provided new insights into the
   interaction mechanisms between PM[2.5] exposure and cardiovascular risk
   in susceptible individuals consuming excess cholesterol and fat. It is
   crucial to enhance vigilance toward the cardiovascular risk among
   individuals with HFCD living in PM[2.5]-polluted areas.

Figure 6.

   Figure 6
   [198]Open in a new tab

   Proposed working model of interaction between more severe nonalcoholic
   steatohepatitis and increased cardiovascular risk caused by PM[2.5]
   exposure in mice fed with HFCD.

Acknowledgments