Total cholesterol (TC), triglyceride (TG), alanine transaminase (ALT), and aspartate transaminase (AST) kits were all purchased from Medical System Biotechnology Co., Ltd. (Ningbo, Zhejiang, China). Hematoxylin-Eosin solution, 3,3’-Diaminobenzidine (DAB) solution, and nitric oxide (NO) assay kit were all purchased from Nanjing Jiancheng Technology Co., Ltd. (Jiangsu, China). Oil red O solution was form BBI Life Sciences Corporation (Zhejiang, China). Antibodies against carnitine palmitoyltransferase-1α (CPT1α), fatty acid transporter-2 (FATP2), stearoyl-CoA desaturase-1 (SCD-1), acyl-CoA oxidase 1 (ACOX1), fatty acid synthase (FAS), occludin, HRP conjugated goat anti-rabbit IgG (H+L), FITC conjugated goat anti-rabbit IgG (H+L), and β-actin were purchased from Proteintech Group (Proteintech, USA). HRP conjugated goat anti-mouse/rabbit IgG (PV-6001) was from Zhongshan Goldenbridge Biotechnology Co., Ltd. (Beijing, China). The BCA protein concentration determination kit, RIPA lysate kit, DAPI staining solution, and the chemiluminescent assay kit were from Beyotime Biotechnology (Jiangsu, China).

N^ω-nitro-L-Arginine Methyl Ester (L-NAME) was from Yuanye Biotechnology Co., Ltd. (Shanghai, China). All triheptadecenoyl glycerol (T17:1 TAG) was purchased from Nu Chek, Inc. (Elysian, MN, USA). All other reagents were at least analytical grade and purchased from Fisher Scientific (Pittsburgh, PA), Sigma-Aldrich Chemical Company (St. Louis, MO), or as specified.

Sprague-Dawley Rats (Male) were obtained from Zhejiang Academy of Medical Science (SCXK 2014-0001, Hangzhou, China). The animal operations were in accordance with the Guidelines of Care and Use of Laboratory Animals of Zhejiang University of Technology and the operations were approved by the Ethics Committee of Zhejiang University of Technology.

Eighteen SD rats were divided into the normal group (NG) and the L-NAME-treated group (L-NAME), with nine rats in each group. During the experiment, all rats were fed with the standard diet. The normal group rats were given purified water every day, and the L-NAME group was given L-NAME (40 mg/kg, p.o.) with the volume 1 ml/100 g according to the weight for 8 weeks. At the end of the experiment, the rats were anesthetized and the liver, aorta, and gut tissues were separated as soon as possible. The tissues parts were put into the 10% formalin, 2.5% glutaraldehyde, and 30% sucrose solutions for hepatic pathology. The other parts were stored at −80˚C for the hepatic lipids and Western-Blot analysis.

The blood pressure (BP) was monitored within 1, 3, and 7 weeks. The BP-2010 AUL non-invasive blood pressure monitoring system (Softron Biotechnology, Beijing, China) was used to measure BP of SD rats as described previously (Chen et al., 2015). Briefly, the air tightness of the instrument monitoring channel was examined, the temperature of the incubator was adjusted to 37°C, and kept at room temperature. The rats were then placed in a monitoring environment for 15 min in the fixed bag. Meanwhile, the BP monitoring sensors were round the tail root of rats. When the signal was stable, the data of systolic blood pressure (SBP), diastolic blood pressure (DBP), and mean arterial blood pressure (MBP) were detected and collected.

After treatment for 4 and 6 weeks, the rats were fasted overnight and the blood was taken from the ophthalmic venous plexus, centrifuged at 3,500 rpm/min for 10 min to collect serum. The serum transaminase (ALT and AST) and the serum lipid (TC and TG) levels were measured with the corresponding kits by an automatic biochemical analyzer (HITACHI-7020, Japan).

The serum nitric oxide (NO) level was evaluated by the nitrate reductase method as described earlier (Chen et al., 2015). The blood sample was centrifuged at 3,000 rpm/min for 10 min and the serum was separated to determine the NO concentration at the end of experiment. The reductase nitrate restores all NO^3- to NO^2-, which can be determined by the excellent colorimetric reagent to measure the total nitrite as an indicator of NO production. All the procedures were performed as described in the assay kit.

For gross microcirculation of the liver and intestine, the microcirculatory blood flow was assessed by laser Doppler with the moor FLPI V2.1 software (Moor Instruments Ltd, Millwey, Axminster, Devon, UK). Firstly, the rats were anaesthetized with 10% chloral hydrate, then the abdominal cavity was exposed and the liver was separated on the transparent plastic sheet. Next the distance of the scanner and liver was set to be 13 cm. One 1.4 cm × 1.4 cm area was selected on the liver. Then the distance of the scanner and liver was set to be 13 cm. The reflective signal of laser beam was received by a detector positioned in the scanner head and converted into the electrical signal to represent the blood perfusion condition (Li et al., 2018a). Detection of gross microcirculation in the intestine is the same as in the liver.

After the observation of gross microcirculation, microcirculatory blood flow and microcirculation of the intestine and liver were assessed by microcirculation in the vivo of tracking and observing systems (Gene&I, Beijing, China). Immediately after putting the cotton sheet round the intestine on the transparent plastic sheet, it was placed on the platform of the inverted biomicroscopy (Nikon) with a constant 37°C, and kept wet. Microvessels were selected with a 10-fold eyepiece. Then they were replaced with the eyepiece with 20-fold. The microcirculation was observed and recorded with the CCD color camera connected with the microscope for 1 min, then the eyepiece was 40-fold converted, and the high-speed camera (IDT) connected with the microscope was used to record for 30 s at a speed of 500 frames/s. Then the microcirculatory blood flow and microcirculation of the liver was observed as the intestine.

An appropriate amount and parts of liver, aorta, intestine, and colon tissues were fixed in 10% formalin. The tissue wax was prepared by the gradient ethanol dehydration, xylene transparency, and paraffin embedding. Tissue sections were prepared for 4 μm and stained with the hematoxylin-eosin (H&E) staining according to the previous literature reports (Lei et al., 2019; Li et al., 2019).

The liver tissue was dehydrated with the 30% sucrose solution before embedding into the OCT compound (Sakura, Tokyo, Japan), and cut into 10 μm frozen sections. Oil red O (OR) staining was made according to the procedure described in our papers (Lei et al., 2019; Li et al., 2019). Briefly, the sections were rinsed and then kept in 60% isopropanol, finally stained with the 0.5% Oil red O solution for 15 min, and in 60% isopropanol for differentiation. Then the nuclei of cells were processed for hematoxylin staining. All H&E and OR staining was photographed with the biological microscope (BX43, Olympus, Japan) at a total magnification of ×400 and analyzed with the Image-Pro Plus software.

The liver and intestine ultrastructures were observed using the transmission electron microscopy (TEM) as described previously (Lei et al., 2019). Briefly, the livers and intestines were cut into 1 mm³ spaced pieces, fixed in 2.5% glutaraldehyde phosphate buffer (pH 7.4), and postfixed in 1% osmium tetroxide. Then the tissues were dehydrated in a graded ethanol series, followed by embedding in the epoxy resin for blocks preparation. The ultrathin sections were cut into 50~70 nm with the ultrathin microtome. The liver ultrastructures were observed using the transmission electron microscopy (HT7700, HITACHI) at a total magnification of ×1200.

The immunohistochemistry (IHC) staining was similar to that described previously (Li et al., 2018a). The expression and localization of endothelial nitric oxide synthase (eNOS) in the aorta and occludin in the intestine and colon were determined. The deparaffinized tissue sections were incubated with the corresponding primary antibody eNOS or occludin (1:200, dilution). A secondary antibody (HRP conjugated goat anti-mouse/rabbit IgG) was added. The signals were visualized by DAB staining and the nuclei were counterstained with hematoxylin. Positive staining revealed the yellow color under the microscope.

For the immunofluorescent labeling, the deparaffinized tissue sections were incubated with antibodies against CPT1α, FATP2, SCD-1, ACOX1, and FAS overnight at 4°C after antigen retrieval and blocked. The secondary antibody FITC conjugate goat anti-rabbit IgG (H+L) was added and incubated in a light-tight shield. Then the nuclei were incubated with DAPI for 5 min at room temperature in a light-tight shield. The experiment was performed following the manufacturer’s protocol and were photographed with the laser scanning microscope (Olympus, Japan) at a total magnification of ×400. The data of protein expressions were semi-quantitatively analyzed as the integrated option density (IOD) in the positive area of the microphotograph with the Image-Pro Plus software.

A powder sample (~25 mg) from each liver tissue was weighed and further homogenized in 0.5 ml of ice-cold, diluted PBS using sonification (Branson Ultrasonic Bath). A protein assay was performed on individual homogenates applying a bicinchoninic acid protein assay kit (Pierce, Rockford, IL) using the bovine serum albumin as the standard. Lipids were extracted from each sample by employing a modified procedure of Bligh and Dyer extraction in the presence of internal standards that were in a premixed solution for the global lipid analysis and added into each liver tissue sample based on its protein content. Therefore, the lipid levels of liver samples can be normalized to the protein concentration and directly quantified. The lipid extracts were dried with nitrogen, redissolved in 1:1 CHCl₃/MeOH with a volume of 200 μl/mg protein (present in the original sample) capped, and stored at −20°C for the mass spectrometric (MS) analysis as previously described (Yang et al., 2009).

The MS analyses of lipids and data processing were based on the literature (Hu et al., 2015; Hu et al., 2016). In brief, a triple-quadrupole mass spectrometer (Thermo TSQ Quantiva) equipped with an automated nanospray ion source (TriVersa NanoMate, Advion Bioscience Ltd.) and operated using the X calibur system software, was used in the study. Identification and quantification of different lipid classes and individual species were conducted by MDMS-SL as described based on the lipidomics principles, mainly including triglycerides (TAG), fatty acids (FAs), and phosphatidylcholine (PC).

In brief, appropriate liver tissues were lysed at 4°C in the RIPA buffer. The tissue homogenates were centrifuged at 12,000 rpm/min for 15 min at 4°C, and the resulting supernatants were used for the Western-Blot analysis. The total proteins in the supernatants were determined by the BCA assay. Then SDS-PAGE was used to transfer the protein onto the nitric acid fiber membrane. The protein was sealed with the 5% skimmed milk powder, and the first antibody (5% milk, β-actin 1:10,000, CPT1α 1:2,000, FATP2 1:1,000, FAS 1:1,000, and ACOX1 1:1,000) incubated the membrane at 4°C overnight. After being washed three times with TBST, the membrane was incubated with the second antibody, HRP conjugated goat anti-rabbit IgG (H+L) (1:5,000 dilution), for 120 minutes. Finally, the blotted protein bands were detected by the chemiluminescent assay kit and the protein expression levels were normalized to β-actin by densitometry with the Image J software.

The fecal bacterial analysis was evaluated by 16S rRNA amplification of V3-V4 region and Illumina sequencing as previously described (Lei et al., 2019). Briefly, the V3-V4 hypervariable regions of the bacteria 16S rRNA gene were amplified with primers 338F (5′-ACTCCTACGGGAGGCAGCAG-3′) and 806R (5′-GGACTACHVG GGTWTCTAAT-3′) by the thermocycler PCR system (GeneAmp 9700, ABI, USA). The resulted PCR products were extracted from a 2% agarose gel and further purified using the AxyPrep DNA Gel Extraction Kit (Axygen Biosciences, Union City, CA, USA) and quantified using QuantiFluor™-ST (Promega, USA) according to the manufacturer’s protocol. The purified DNA from the individual samples was sequenced by the Illumina MiSeq platform (Illumina, San Diego, USA) according to the standard protocols by Major bio Bio-Pharm Technology Co. Ltd. (Shanghai, China). The operational taxonomic units (OTUs) were clustered with 97% similarity cutoff using UPARSE (version7.1 http://drive5.com/uparse/) and chimeric sequences were identified and removed using UCHIME. The taxonomy of each 16S rRNA gene sequence was analyzed by the RDP Classifier algorithm (http://rdp.cme.msu.edu/) against the Silva (SSU123) 16S rRNA database using the confidence threshold of 70%. The data were scoped on the I-Sanger cloud platform (http://www.i-sanger.com) developed by Majorbio Bio-Pharm Technology Co. Ltd. The community of generic level as well as the calculated alpha diversity (Chao1) and the beta diversity on the generic level were analyzed. Statistical comparisons of the results were performed using the Student’s t-test.

All data were expressed as the means ± SEM and subjected to the t-test. When compared with the same period of the model group, the data were subjected to the independent-sample t-tests. The two-sided Student’s t-test were used to examine differences in microflora between two groups. The Pearson correlation coefficient was used in the examination of correlations between the SBP level and the TAG species. The Spearman correlation coefficient was used to examine the correlations between the relative abundance of the gut microbial community at the genus level and the TAG and FA species, as well as the serum NO and liver microcirculation. All analyses were performed using the updated version of SPSS software. Diagrams were prepared applying the Graph Prims and Excel software.

Compared with the NG, there was no significant effect on the body weight of the L-NAME-treated rats continuously molded for 7 weeks (data not shown), which indicated that L-NAME did not affect the normal growth of rats. SBP, DBP, and MBP in the L-NAME-treated rats were increased by 14.8, 17.1, and 16.2%, respectively compared with the normal rats after the 1-week modeling. The increased rates of SBP, DBP, and MBP were 19.8%, 23.6%, 22.1% after the 3^rd week and went up by 22.7%, 27.4%, 25.6% after the 7^th week ( Figures 1A–C ) (P < 0.01). This indicated that increasing the exposure time results in greater increase rate was observed in the blood pressure of rats.

Further, we also found that the expression of eNOS in aorta endothelium was significantly reduced with thickening of vascular wall ( Figure 1D ), and L-NAME caused a significant decrease of serum NO level in rats ( Figure 1E ) (P < 0.05). This suggested that the increased BP may be related to the endothelial nitric oxide synthase-nitric oxide (eNOS-NO) pathway imbalance in the L-NAME-treated rats.

Firstly, in order to confirm that L-NAME-induced NO deficiency in hypertensive rats can cause liver injury, as well as changes in the serum lipid profiles. In contrast to the NG, the levels of serum TC, ALT, and AST in the L-NAME-treated rats were significantly increased after the 4^th week, and the levels of serum TG and ALT after the 6^th week ( Figures 2A–D ) (P < 0.01, 0.05). Moreover, the ALT level increase rate was 58% after the 4^th week and increased by 72% after the 6^th week compared with the NG. The results suggested that there might be abnormal lipids metabolism and liver injury in the L-NAME-treated rats, which was consistent with other studies (Abd-Elsalam et al., 2017; Bilanda et al., 2017).

Further study of the effect of L-NAME on the liver, we observed the hepatic gross microcirculation and microvasculature abundance to assess the change of hepatic microcirculatory perfusion. Compared with the NG, L-NAME could induce significant decline of perfusion in the liver with a decrease in the number of microvessels ( Figures 2F, G ) (P < 0.01). H&E staining and OR staining revealed that there was a wide range of abnormal lipid accumulation in the liver of the L-NAME-treated rats ( Figure 2E ). Besides, the ultrastructural analysis of hepatocytes with TEM showed that in the normal hepatocytes, the nucleus was round and central, the nuclear membrane was clear, and the rough endoplasmic reticulum was abundant and closely arranged with numerous mitochondria. While the L-NAME-treated rats exhibited significant lipid droplets accumulation, mitochondrial damage, and disordered endoplasmic reticulum structure ( Figure 2E ). These data indicated that there were abnormal hepatic microcirculation and histopathology in the rats chronically treated with L-NAME.

In the present study, it was found that NO deficiency could cause the liver lipid droplets accumulation and liver injury, but there was no evidence whether L-NAME-induced NO deficiency could induce disorder of lipid molecules, which is closely related to liver injury. So, the method of MDMS-SL was used to investigate these changes. The total of 207 lipid molecules were detected by MDMS-SL, including 30 kinds of triglycerides (TAG), 20 kinds of fatty acids (FAs), 10 kinds of ceramide (Cer), 21 kinds of phosphatidylinositol (PI), 50 kinds of phosphatidylcholine (PC), 13 kinds of lysophosphatidylcholine (LPC), 17 kinds of lysophosphatidic (LPE), 19 kinds of sphingolipid (SM), 9 kinds of cardiolipin (CL), as well as smaller amount of other lipid molecules. Among them, the most abundant were PC and TAG. Compared with the NG, 65 kinds of lipids significant changes were found after 8 weeks of L-NAME modeling, which indicated that L-NAME could cause the liver lipid metabolism disorder ( Figure 3A ). The principal component analysis (PCA) describes the largest variation in the data using a few orthogonal latent variables. PCA was carried out and the score plot was obtained for the TAG and FA data, including trends and groupings. For TAG and FA, the PC1 separated the samples with respect to the hepatic lipid levels and accounted for 78% and 82.6% of the total variance, respectively, whereas the PC2 accounted for 17.1% and 14.6%, respectively ( Figure 3B ).

There were 30 types of TAG and 20 FA specifically detected in the liver. The total content of TAG in the L-NAME-treated rats was significantly increased (P < 0.01), from 31.60 ± 2.82 in the NG rats to 144.74 ± 15.02 nmol/mg protein, increasing of 358%. Except for C47:0/C48:7, C48:6, C48:5, and C48:0/C49:7, the other lipids of TAG were significantly increased ( Figures 3C, D ) (P < 0.05, 0.01). Further analyses of the correlation between the blood pressure and the top four contents of TAG indicated that SBP was positively correlated with the main lipid content, showing a significant difference ( Figure 3G ). As FA are an important component of TAG, we further analyzed long-chain fatty acid content to accurately analyze changes in TAG species. The results showed that the total FA content in the model group increased significantly (P < 0.01), which increased from 94.81 ± 8.46 in the NG to 434.23 ± 45.07 nmol/mg protein, with an increase of 358%. Except for C22:0, other FA were all significantly increased ( Figures 3E, F ) (P < 0.05, 0.01). These data suggested that there was an abnormal hepatic fatty acid metabolism in the rats treated chronically with L-NAME.

In the study, for further evaluation of lipids changes in the L-NAME induced rats, the MDMS-SL was also employed to determine the amounts of other phospholipid classes including PC, LPC, and LPE. Compared with the control group, the MDMS-SL analysis showed significant changes of the amount of many PC species, D16:1–18:2, D16:0–18:2, D16:0–18:0, A16:0–20:4, P18:0–18:1, D18:1–18:2, D18:0–18:2, D18:0–18:1, and D18:0–20:3 in the L-NAME-treated rats ( Figure 4A ) (P < 0.05, 0.01). At the same time, there were some LPE species, 20:4, 20:2, and P22:2 significant changed ( Figure 4B ) (P < 0.05). The further analysis demonstrated that the LPC species did not change significantly in the L-NAME-treated rats ( Figure 4C ). These data suggested that L-NAME treatment might slightly alter hepatic levels of phospholipid classes, which also indicated that the L-NAME treatment might induce mainly changes of hepatic fatty acid metabolism.

Further investigations were carried out to determine whether the metabolic enzymes were involved in the underlying mechanisms of hepatic fatty acid metabolism in the L-NAME-treated rats. Some of the proteins involved in the fatty acid synthesis, desaturation, uptake, and oxidation in the liver were measured with the immunofluorescence and Western-Blot method ( Figure 5 ). The results showed that the expression of FAS and SCD-1, the key rate limiting enzyme in the synthesis and desaturation of FA, was significantly enhanced in the liver of the L-NAME-treated rats ( Figures 5A, B, F ), the result of FAS corresponded to the Western-Blot assay ( Figures 5G, H ) (P < 0.05). The FA lipidomics showed that the components of saturated FA and monounsaturated FA increased significantly, and the expression of fatty acid β-oxidation related enzymes ACOX1 and CPTA1α was significantly decreased after the L-NAME treatment ( Figures 5D–F ), the results corresponded to Western-Blot assay ( Figures 5G, H ) (P < 0.05, 0.01). Moreover, the expression of FATP2 did not show significant changes. These results indicated that the L-NAME treatment might increase the fatty acid synthesis and desaturation, as well as reduce β-oxidation in the liver, which may be essential factors in regulating the hepatic fatty acid metabolism.

There is increasing evidence of the role of host-gut interactions on hypertension and liver injury. L-NAME-induced gut microflora disorder in hypertensive rats may lead to liver lipid metabolism disorder, further aggravating liver damage. First of all, it was investigated whether the gut pathology, a significant role of host-gut interactions in the pathogenesis of hypertension, altered in the L-NAME-treated rats. It was observed that the villi in the small intestine and colon of the L-NAME-treated rats were much shorter and appeared stunted ( Figure 6B ), with the perfusion decline compared with the NG ( Figure 6A ) (P < 0.05). The TEM showed that villi in the L-NAME-treated rats became a shorter and looser arrangement, with a large number of disrupted tight junction proteins (TJPs) ( Figure 6B ). The TJPs, including occludin, were very important for the intestinal barrier. There were decreased on occludin expression with the IHC in the small intestine and colon of the L-NAME-treated rats ( Figure 6D ). Accordingly, using the microcirculation in vivo tracking and observing system, we found an increase in the white blood cells roll and stick and mast cells degranulation associations with the intestinal inflammatory status, as well as a decrease in intestinal venous plexus associations with the intestinal perfusion ( Figure 6C ). These data provided evidence to support the proposal that the L-NAME-treated rats with established hypertension have pathophysiological alterations in the gut, with a leaky intestinal barrier, an inflammation reaction, and decline of microcirculation perfusion.

The above data indicate that hypertension is associated with increases in gut permeability and inflammatory status. Then to investigate whether there are also differences in the gut bacterial taxa identified through sequencing between the NG and the L-NAME-treated rats. The relative abundance of genus of all samples showed that the L-NAME treatment could decrease the gut microbiota community ( Figure 7A ), with the same decrease in the Ace index, Chao index, Sobs index, and Shannon index ( Figure 7B ) (P < 0.01). At the phylum level, the L-NAME-treated rats associated gut microflora showed a dramatic decrease in the relative abundance of Proteobacteria and Tenericutes compared with the NG ( Figure 7C ) (P < 0.05). Furthermore, the relative abundance of most gut microflora, such as Ruminococcaceae_UCG-014, [Eubacterium]_xylanophilum_group, norank_o_Mollicutes_RF9, Roseburia, Desulfovibrio, Parabacteroides, Ruminococcaceae_UCG-013, Intestinimonas, and Ruminococcaceae_UCG-010, showed a sharp decrease at the genus level of the L-NAME-treated rats. Also, Romboutsia, Blautia, Coprococcus_1, Ruminococcaceae_NK4A214_group, Ruminococcaceae_UCG-008, and Erysipelatoclostridium increased significantly ( Figure 7D ) (P < 0.05, 0.01).

The taxonomic groups that are differentially abundant between the NG and L-NAME-treated rats were identified using the linear discriminant analysis effect size (LEfSe) with α  =  0.05, LDA > 2 ( Figure 7E ). It was found out that 62 features had significantly different abundance between the NG and L-NAME-treated rats, these biomarkers accounted for 26.7% of taxonomic groups. At a genus level, gut microbiota of NG was differentially enriched with Firmicutes, Proteobacteria, Tenericutes, Actinobacteria, and Bacteroidetes, whereas the L-NAME-treated rats were differentially enriched with Firmicutes. In addition, at the generic level (LDA > 3.2), search was made for the significant differences between the NG and L-NAME-treated rats group. It was found out that Ruminococcaceae_UCG_014, Ruminococcus_1, Lachnospiraceae_NK4A136_group, Eubacterium_xylanophilum_group, unclassified_f_Ruminococceae, norank_o_Mollicutes_RF9, and Roseburia were highly enriched in the NG, while Romboutsia, Blautia, Allobaculum, and Coproccus were highly enriched in the L-NAME-treated hypertensive rats. Interestingly, it was also observed that Mollicutes was highly enriched in the NG, while Erysipelotrichia in the L-NAME-treated rats at the class level ( Figure 7E ).

The finding of our present study was that L-NAME could cause serum NO deficiency, liver microcirculation disorders, changes in liver lipid metabonomics, and microbiota dysbiosis, but it is not clear whether there are certain correlations between these changes. Therefore, we performed correlation analyses between the serum NO and intestinal bacteria, and the correlation between intestinal bacteria and liver injury tag, liver microcirculation and liver FA and TAG species, by Spearman’s correlation coefficient method.

The correlations between the relative abundance of the gut microbial community and important metabolic parameters associated with hypertension and liver injury were analyzed Spearman’s correlation coefficient heat map. At the genus level, Ruminococcaceae_UCG-014 and Lachnoclostridium exhibited significant positive correlations with serum NO and the liver gross microcirculation perfusion (FLUX) (P < 0.05, 0.01). The norank_o_Mollicutes_RF9 showed significant positive correlations with the liver FLUX (P < 0.05). Blautia displayed the opposite trend with the liver FLUX (P < 0.05) ( Figure 8A ). We analyzed the relationship between the gut microbial community and liver TAG species at the genus level. Marvinbryantia, Turicibacter, Ruminococcaceae_UCG-008, Ruminococcaceae_NK4A214_group, and Romboutsia showed significant positive correlations with the liver total TAG, C52:3/C52:10, C52:4/C53:1, C52:2/C53:9, and C54:5/C55:12 (P < 0.05, 0.001), while [Ruminococcus]_gauvreauii_group, [Eubacterium]_xylanophilum_group, Lachnospiraceae_NK4A136_group, norank_f_Lachnospiraceae, unclassified_f_Lachnospiraceae, and Phascolarctobacterium had inverse correlations with them correspondingly ( Figure 8B ) (P < 0.05, 0.001).

Moreover, the relationship between the gut microbial community and liver FA species at the genus level displayed that Prevotellaceae_NK3B31_group, Blautia, Romboutsia, Turicibacter, Ruminococcaceae_UCG-008, Lactobacillus, and Marvinbryantia were significantly positive correlations with the total FA, C14:0, C16:0, C16:1, C18:0, C18:1, C18:2, C18:3, C20:1, and C21:1 ( Figure 8C ) (P < 0.05, 0.001), but Ruminococcaceae_UCG-014, unclassified_f_Ruminococcaceae, Lachnospiraceae_NK4A136_group, norank_f_Lachnospiraceae, [Eubacterium]_xylanophilum_group, unclassified_f_Lachnospiraceae, and [Ruminococcus]_gauvreauii_group inverse correlations with them correspondingly ( Figure 8C ) (P < 0.05, 0.001). It is suggested that the changes of gut microbiota may lead to the difference of liver lipid metabolism in the L-NAME-treated rats. Although there is no detailed study on how the gut microbiota affects the liver lipid metabolism, research has shown that the change of gut microbiota can affect liver fatty change, and probiotics could relieve liver lipid (Albillos et al., 2019).