Male ApoE KO mice (Laboratory Animal Center of Xi’an Jiaotong University), aged eight weeks, were utilized in this study. These mice were intravenously injected via the tail vein with 1 $\times$ 10¹⁰ plaque-forming units of Ad-CTRP9, AAV-shCTRP9, Ad-GFP, or AAV-GFP (as a control). The viral products were purchased from Hanbio Biotechnology Co., Ltd. (Shanghai, China). The mice were euthanized via an intraperitoneal injection of pentobarbital sodium at a concentration of 150 mg/kg body weight. Throughout the experimental period, the mice were maintained on a Western diet, which included 21% fat and 0.15% cholesterol (Vital River Company, Beijing, China). The experimental mice were assigned to two groups (n = 15) and housed in a controlled environment with a 12-hour light/12-hour dark cycle and unrestricted access to water and food. The animal experimental protocol was approved by the Laboratory Animal Administration Committee of Fourth Military Medical University (Approval No. k9-014) in accordance with the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH publication number 85-23, revised 2011).

After an overnight fasting period, blood samples were collected, treated with EDTA, and subsequently centrifuged at 1500 rpm for 10 min at 4 °C for plasma isolation. The levels of high-density lipoprotein cholesterol (HDL-C), triglycerides (TGs), low-density lipoprotein cholesterol (LDL-C), and total cholesterol (TC) in plasma were quantified via commercially available assay kits supplied by Biosino Bio-Technology & Science, Inc. (Beijing, China) [22].

For analysis of oxidative stress markers in the plasma, various parameters, including malondialdehyde (MDA) levels, superoxidase dismutase (SOD) activity, total nitric oxide synthase (TNOS), glutathione peroxidase (GSH-Px), inducible nitric oxide synthase (iNOS), glutathione S-transferase (GST), and total antioxidant capacity (T-AOC), were measured. These measurements were conducted via colorimetric assay kits. Additionally, nitric oxide (NO) levels were determined with a nitrate reductase assay kit. All these assay kits were acquired from Nanjing Jiancheng Bioengineering Institute (Nanjing, China).

For measurement of atherosclerosis, the mice were euthanized via an intraperitoneal injection of pentobarbital sodium at a dosage of 150 mg/kg. The aortic trees were subsequently dissected and subjected to oil red O staining. Adjacent sections on separate slides were immunostained with monoclonal antibodies against macrophages (MOMA-2; 1:100; Cat. No. ab-33451; Abcam, Cambridge, MA, USA). Visualization of the sections was achieved with a DAB substrate kit (Cat. No. GK800511, Gene Tech, Shanghai, China). The en face lesion size was measured via an image analysis system provided by WinRoof Mitani Co. (Tokyo, Japan) [22].

Frozen cross-sections of the aortic root were examined for microscopic analysis of atherosclerotic lesions. Specifically, ten cross-sections from each mouse were evaluated. The sections were subjected to oil red O and hematoxylin‒eosin (H&E) staining to visualize the lesions. The stained sections were subsequently quantified via the image analysis system provided by WinRoof Mitani Co. (Tokyo, Japan) to determine the oil red O-stained area [23].

Total RNA was isolated from aortas via RNAzol (TaKaRa, Tokyo, Japan). The purity and concentration of the extracted RNA were determined via a Nanodrop instrument from Thermo Fisher Scientific (Rockford, IL, USA). Additionally, the integrity of the RNA was verified via an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). The purification of mRNA from the total RNA samples was performed via the NEBNext® Poly(A) mRNA magnetic isolation module. For assembly of the library, which was characterized by an insert size of 400 bp, the NEBNext UltraTM RNA Library Prep Kit was utilized following the guidelines provided by the manufacturer (Illumina San Diego, CA, USA) [24]. The quality of the library was evaluated via the Agilent Bioanalyzer 2100 system. The index-coded samples were subsequently clustered on a cBot Cluster Generation System via a TruSeq PE Cluster Kit v4-cBot-HS (Illumina) [25]. Library sequencing was performed with the Illumina HiSeq 2500 platform, which utilized 100 bp paired-end reads. This equipment was sourced from Illumina (San Diego, CA, USA).

A Perl script was used to remove reads that included contaminated adapters, those with more than 0.25% low-quality bases (indicated by a Phred quality score below 20), or those containing more than 10% Ns. The purified reads were subsequently aligned to the mouse reference genome (GRCm38) via TopHat software. The quantification and normalization of gene expression were performed via Cufflinks software, which uses the reads per kilobase per million (RPKM) method for this process [26]. The analysis of DEGs between the Ad-GFP and Ad-CTRP9 groups was conducted via DESeq software (V1.44.0) [27]. The significance threshold for p values in various tests was established through the false discovery rate (FDR). Genes exhibiting an absolute FDR value less than 0.05 and a fold change $\ge$2 were classified as significantly differentially expressed. Pathway and functional enrichment analyses involved mapping datasets from the Kyoto Encyclopedia of Genes and Genomes (KEGG). KEGG terms that were significantly enriched were identified using a threshold p value of $\le$0.05.

Freshly obtained liver samples were homogenized in lysis buffer and maintained at 4 °C. The resulting supernatants derived from this process were collected and used for Western blot experiments. Lysates (20 µg) were resolved via 10% sodium dodecyl sulfate‒polyacrylamide gel electrophoresis, after which they were transferred onto polyvinylidene fluoride membranes (Sequi-Blot, Bio-Rad, Hercules, CA, USA). The membranes were subsequently incubated with primary antibodies overnight at 4 °C. Specifically, anti-CTRP9 (No. NBP2-46834, Novus, Centennial, CO, USA) was used at a dilution of 1:2000, and $\beta$-actin (No. ab15246, Abmart, Shanghai, China) was used at a dilution of 1:1000. These procedures were conducted in strict accordance with the manufacturer’s prescribed guidelines. Signal detection was achieved via the use of Immobilon reagent (Millipore, Billerica, MA, USA). The signals were captured via an LAS-400 LuminoImage Analyzer manufactured by Fujifilm Co. (Tokyo, Japan).

The proximal aortas were initially fixed in 4% paraformaldehyde (BL539A, Biosharp, Hefei, China) for 20 min and then treated with a permeabilization solution (0.1% Triton $\times$100 and 0.1% sodium citrate) on ice for 2 min. Subsequently, the aortas were blocked in horse serum (Cat. No. 16050130, Gibco, Grand Island, NY, USA) for 2 h [3]. The sections were incubated with specific antibodies, namely, anti-MDA (Cat. No. GB111891, Servicebio, Wuhan, China) and anti-iNOS (Cat. No. GB11119, Servicebio, Wuhan, China), after which the samples were incubated with secondary antibodies, specifically goat anti-rabbit Alexa Fluor 488 (Cat. No. A11008, Invitrogen, Carlsbad, CA, USA). Images were acquired via a Leica DMRE microscope equipped with Spot digital image analysis software and a camera, as well as a laser scanning confocal microscope (Leica, Wetzlar, Hessen, Germany).

The assessment of cholesterol efflux was conducted in accordance with the methods of Yin et al. [28] and Maarit Hölttä-Vuori et al. [29]. Briefly, we injected Ad-CTRP9 or Ad-GFP into ApoE KO mice via the tail vein for 4 weeks. After the mice were humanely euthanized through cervical dislocation, precooled PBS was used to extract macrophages from the abdominal cavities of the mice. The macrophages isolated from the experimental groups were subjected to labeling with medium containing 2.5 µM BODIPY-cholesterol (HY-125746, MCE, Monmouth Junction, NJ, USA). This labeling process was conducted at 37 °C with 5% CO₂ for 60 min. The cells were removed from their environment, subsequently resuspended and subjected to two rounds of purification. Flow cytometry (Calibur; BD Biosciences, Franklin Lakes, NJ, USA) was used to measure the mean fluorescence intensity (MFI) as the baseline. After a 4-h interval, the MFI of BODIPY-cholesterol in the macrophages was once again determined via flow cytometry. For determination of cholesterol efflux, the following formula was used: (baseline MFI – final MFI)/baseline MFI $\times$ 100.

The data are presented as the mean values $\pm$ SEMs. Statistical analyses were performed via Welch’s t test in cases where the p value was unequal or via Student’s t test with an equal F value. A statistically significant difference between two groups was defined at p $\le$ 0.05.

To evaluate the impact of CTRP9 knockdown on the formation of atherosclerotic lesions, we administered shRNA adeno-associated virus (AAV), which targeted mouse CTRP9 (AAV-shCTRP9), or AAV-GFP to ApoE KO mice. The liver CTRP9 protein levels were notably lower in the AAV-shCTRP9 group than in the control group at six days post-treatment (p 0.01) (Fig. 3A,B). The data indicated a substantial increase in body weight in the AAV-shCTRP9 group compared with the GFP group after administration for 12 weeks (p 0.01) (Fig. 4A).

Fig. 3.

CTRP9 knockdown in ApoE KO mice. (A) Protein expression of CTRP9 was determined by Western blotting ($\beta$-actin was used as the loading control). (B) Quantification of the Western blot gels; n = 4 in each group (pooled samples). (C) Illustrative image of ORO staining in aortas. (D) Percentage of the gross lesion area (n = 8 in each cohort). (E) Illustrative micrographs of atherosclerotic lesions of the aortic root. Aortic root sections were stained with ORO and H&E. Scale bar = 200 µm. (F) Quantitative analysis of aortic root lesion areas are shown on the left (n = 10 in each cohort). (G) Quantitative analysis of the ORO-stained area is shown on the right (n = 6~9 in each cohort). (H) Aortic root sections were stained with MOMA-2 to identify macrophages. Scale bar = 200 µm. (I) Quantitative analysis of macrophages in aortic root lesion areas (n = 7~8 in each cohort). Mean $\pm$ SEM. *p 0.05, **p 0.01, and ***p 0.001, AAV-shCTRP9 versus AAV-GFP.

Fig. 4.

Measurement of plasma lipids and tissue weight after knockdown CTRP9 in ApoE KO mice. (A) Body weights of ApoE KO mice. Plasma levels of (B) TC, (C) TGs, (D) LDL-C, (E) glucose and (F) HDL-C. (G) The weight of BAT, iWAT and eWAT. (H) The weight of liver, kidney and spleen. n = 10 in each cohort. Mean $\pm$ SEM. *p 0.05, **p 0.01 and ***p 0.001, AAV-shCTRP9 versus AAV-GFP.

The en face lesion size of the aorta was notably greater (30%) in the AAV-shCTRP9 group than in the control group (p 0.05) (Fig. 3C,D). Moreover, histological observation revealed an increase in aortic root atherosclerotic lesions in the AAV-shCTRP9 group (p 0.001). Compared with that in the control group, the number of microscopic lesions in the aortic root was substantially greater (60%) in the AAV-shCTRP9 group (p 0.001) (Fig. 3E,F). Furthermore, the lipid area within the lesions, as indicated by oil red O staining, was notably increased by 30% in the AAV-shCTRP9 group (p 0.001) (Fig. 3E,G). In addition, the aortic roots of the AAV-shCTRP9 group presented increased MOMA2-positive areas relative to those of the GFP group (p 0.05) (Fig. 3H,I). Further experiments revealed no significant differences in metabolic parameters, such as the levels of TGs, glucose, TC, and HDL-C, between the AAV-GFP- and AAV-shCTRP9 groups (p $>$ 0.05) (Fig. 4B,C,E,F). Notably, LDL-C was significantly greater in the CTRP9 knockdown cohort than in the control cohort (p 0.001) (Fig. 4D). In addition, we analyzed the effects of CTRP9 knockdown on fat content and organ weight. The findings did not reveal significant differences in the weights of the eWAT, iWAT, BAT, or liver between the AAV-GFP and AAV-shCTRP9 groups (p $>$ 0.05) (Fig. 4G,H). The kidney (p 0.001) and spleen weights (p 0.05) were significantly greater in the CTRP9 knockdown group than in the control group (Fig. 4H).

To elucidate the molecular changes associated with the shrinkage of atherosclerotic lesions in the mice injected with Ad-CTRP9, we conducted RNA sequencing analysis on aortic samples from ApoE KO mice. Our transcriptomic analysis revealed a total of 3485 DEGs in the Ad-CTRP9 group compared with the control group. Among these DEGs, 1728 genes were upregulated, whereas 1757 genes were downregulated (Fig. 5A, Supplementary Table 1).

Fig. 5.

Typical characteristics of DEGs in the aortas of ApoE KO mice. (A) Volcano plots showing the expression of differentially expressed genes (DEGs) based on RNA sequencing analysis. (B) Analysis of the DEGs was performed via DESeq software by contrasting the Ad-GFP and Ad-CTRP9 cohorts. The expression levels are denoted by colors, with red (high expression) and blue (low expression), and are proportional to their brightness (see color bar). (C,D) Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment of DEGs in the aorta. The DEGs were charted into the KEGG datasets, and significantly enriched KEGG terms were determined at p 0.05. n = 3 for each cohort.

CTRP9 overexpression induced the differential expression of many genes, including those that mediate oxidative stress, cell adhesion, and cholesterol effusion and efflux. These genes include proteases (MMP3, MMP23, TIMP1, TIMP3,and TIMP4), chemokines (CCL4, CCL5, CCL7, CCL9, CXCL14, CXCL16, and MIF), adhesion molecules (CAV2, ITGA9, ITGB3, ITGB7, ITGB11,and ICAM2), and those that function in cholesterol metabolism (LXR$\alpha$, VLDLr, and LDLRad3).

Atherosclerosis is believed to be a chronic and multifaceted condition characterized by the accumulation of lipids [30]. Lei et al. [31] reported that CTRP9 can promote the expression of pivotal proteins involved in cholesterol efflux. Liver X receptor alpha (LXR$\alpha$) is a critical regulatory factor for cholesterol homeostasis, including the processes of absorption, trafficking, biosynthesis, and efflux [32]. As a nuclear transcription factor, LXR$\alpha$ governs the transcription of the very-low-density lipoprotein receptor (VLDLr) and ATP binding cassette subfamily G member 1 (ABCG1) in macrophages, thereby influencing cholesterol efflux [33]. These findings revealed substantial increases in the levels of ABCG1 and LXR$\alpha$ following the overexpression of CTRP9. Hence, CTRP9 augmented cholesterol efflux in macrophages by increasing LXR$\alpha$/ABCG1 expression, which was consistent with prior research [31].

The overexpression of CTRP9 significantly regulated the signaling pathway involved in leukocyte transendothelial migration (Fig. 5B and Supplementary Fig. 1). Therefore, we analyzed the genes that were upregulated and downregulated in this signaling pathway according to the RNA-seq results and generated a cascade regulation diagram of the signaling pathway on the basis of the signal transduction pathway (Supplementary Fig. 1). In leukocytes, tail retraction genes (My12b, My12 and My17) and a cell motility-related gene (Rac2) were upregulated (Supplementary Fig. 1), which suggests that the movement of leukocytes in the blood was increased. In contrast, the gene expression levels of endothelial cell adhesion molecules (VCAM1 and ICAM1) were significantly decreased after CTRP9 overexpression (Supplementary Fig. 1). Consistent with these findings, the expression level of the leukocyte surface adhesion molecule ligand protein ITGAL was significantly reduced (Supplementary Fig. 1). Therefore, the ability of monocytes to adhere to endothelial cells was weakened. Moreover, the expression levels of the tight junction genes Cldn5 and ESAM were significantly increased in endothelial cells after Ad-CTRP9 administration compared with those after Ad-GFP administration (Supplementary Fig. 1), indicating that the number of tight junctions between endothelial cells was increased and that the resistance of monocytes to crossing the cell‒cell gap was increased.

A KEGG pathway analysis was subsequently conducted to further elucidate the pathways associated with these DEGs in the context of atherogenesis (Fig. 5C,D). The proportion and quantity of DEGs in these signaling pathways are depicted in Fig. 6. The results indicated that the oxidative phosphorylation signaling pathway had the greatest functional representation with 44 DEGs, which corresponded to approximately 1% of the total number of DEGs. The available evidence indicates that these enriched pathways play critical roles in oxidative stress and the inflammatory reactions associated with atherosclerosis [34]. Our results clearly demonstrated that Ad-CTRP9 primarily modulated genes involved in aortic inflammation and cellular migration (Fig. 5B–D).

Fig. 6.

CTRP9 overexpression activates mitochondrial oxidative phosphorylation in the aorta. (A) Diagram of the oxidative phosphorylation signaling pathway. The results of the bioinformatics analysis revealed that, except for Complex II, significant changes were observed in the DEGs related to the other four complexes. Heatmap analysis of DEGs associated with the (B) Nicotinamide adenine dinucleotide (NADH) dehydrogenase complex, (C) cytochrome bc1 complex, (D) cytochrome c oxidase complex and (E) ATP synthase complex. The DEGs were charted into the KEGG datasets, and significantly enriched KEGG terms were determined at p 0.05. n = 3 for each cohort. (F) Representative images of malondialdehyde (MDA) and nitric oxide synthase (iNOS) immunostaining in aortic root samples are shown. Red and DAPI (blue) were merged. Scale bar = 200 µm. (G) Quantitative analysis of the integrated density of immunofluorescence staining (n = 6). Mean $\pm$ SEM. *p 0.05 and ** p 0.01, Ad-CTRP9 versus Ad-GFP. FMN, flavin mononucleotide; UQ, ubiquinone; ISP, iron sulfur protein.

We then investigated the antioxidative effects of CTRP9 by quantifying the plasma levels of proteins related to oxidative stress. As shown in Fig. 7A, the levels of TNOS and iNOS were significantly lower and the level of T-AOC was significantly greater in the CTRP9-overexpressing group than in the control group (p 0.05). Additionally, we performed immunofluorescence staining to assess oxidative stress markers in plaques (iNOS and MDA) after CTRP9 overexpression. Compared with the control group, the Ad-CTRP9 group presented a noteworthy decrease in iNOS expression (p 0.05), whereas no significant difference in MDA expression was found between the two groups (p $>$ 0.05) (Fig. 6F, Supplementary Fig. 2). This observation was consistent with the results of the RNA sequencing (Fig. 6A–E) and plasma protein measurements (Fig. 7). In contrast, after CTRP9 overexpression, the levels of SOD, MDA, GST, GSH-Px, and NO did not significantly differ from those in the control group (p $>$ 0.05) (Fig. 7A). Conversely, GST activity and GSH-Px activity were significantly lower in the CTRP9 knockdown group than in the control group (p 0.05) (Fig. 7B). Nevertheless, the levels of SOD, MDA, TNOS, iNOS, and T-AOC did not significantly differ between the AAV-shCTRP9 group and the control group (p $>$ 0.05) (Fig. 7B). Notably, CTRP9 knockdown led to an increase in NO production compared with that in the GFP group (p 0.05) (Fig. 7B).

Fig. 7.

Oxidative stress-related protein levels were measured in the plasma via biochemical methods. Superoxidase dismutase activity (SOD), MDA, total nitric oxide synthase (TNOS), iNOS, glutathione peroxidase (GSH-Px), glutathione s-transferase (GST), total antioxidant capacity (T-AOC) and nitric oxide (NO) were measured after (A) overexpression of CTRP9 or (B) knockdown CTRP9. n = 3 in each cohort (5 mouse plasma samples were pooled into one sample). Mean $\pm$ SEM. *p 0.05, Ad-CTRP9 versus Ad-GFP, AAV-shCTRP9 versus AAV-GFP.