The total Ldlr gene knockout (Ldlr^–/–) rat was previously established by the CRISPR/Cas9 system and maintained in our laboratory (Zhao et al., 2018). In the current study, a heritable total G2a knockout (G2a^–/–) rat was generated in the same way. Briefly, two sgRNAs targeting the exon 3 of G2a, CTGCCACACGTTGTCCTACGAGG, and CGGCGAGCACGTTTCTCTGTAGG, were constructed by overlapping PCR. The transcribed sgRNA and Cas9 mRNA were purified and a mixture of them was microinjected into zygotes of Sprague Dawley rats (SLAC Laboratory Animal Co., Ltd., Shanghai, China). Three pups (founder 1–3) were born after 50 injected zygotes were transferred to 1 pseudopregnant female rat. All three pups carried G2a deletion. Sequence analysis revealed that these three founders had frameshift mutations (Supplementary Figures 1A,B). In the end, founder one was chosen and a stable colony of G2a^–/– rat was established, which carried 199-bp deletion from No. 137654370bp to 137654379bp and No. 137654088bp to 137654276bp in the genome (NC_005105.4), and resulted in a termination codon TAA and deletion of 346 amino acids. Rat tail clips were used for genotyping as usual (Supplementary Figure 1C). PCR primers for identifying G2a mutation were 5′- CCTCATCTTGTCAGGGTC -3′ (sense) and 5′- CCGCAGGTAGTAGTAGCC -3′ (antisense). Homozygous G2a^–/– rats showed a markedly reduced G2a expression in related tissues (Supplementary Figure 1D).

Genetically matched Ldlr^–/– and Ldlr^–/–G2a^–/– rats were derived from cross-breeding G2a^–/– and Ldlr^–/– rats. Both male and female animals were maintained and monitored as described previously, in a SPF facility on 12 h light/12 h dark cycles. All rats were supplied water and a normal diet freely accessible. Animals were transferred to Western Diet at 8 week-old to induce atherosclerosis (Zhao et al., 2018). All animal procedures and techniques were approved by the Animal Ethics Committee of East China Normal University with a permit number (R20170202). A mix of male and female rats was used in this study. Most animal experiments were originally conducted by dividing male and female rats. As both genders displayed similar results, they were combined. Representative images of histological analysis and immunoblotting were from male rats. A scheme depicting the time protocol of rat treatments and time points of measuring various parameters was provided in Supplementary Figure 2.

Bone Marrow-Derived Macrophages (BMDMs) were induced from bone marrow derived cells of 8 or 36 week-old animals as previously described (Weischenfeldt and Porse, 2008; Muschter et al., 2015), and about 90% of the resulted cells were CD68 positive. For foam cell induction, macrophages were incubated with 10 μg/ml oxLDL for 24–48 h. In some experiments, macrophages were stimulated with 100 ng/ml LPS (Sigma) or 20 ng/ml IL-4 (PeproTech) and their expression of inflammatory or anti-inflammatory cytokines were then evaluated by quantitative real-time PCR.

Migration assay was performed using the transwell chamber (Corning Falcon) in a 24-well cell culture plate. Cell suspensions containing 1 × 10⁵ cells in 500 μl serum-free DMEM medium with 10 μg/ml oxLDL were placed in the top chamber. The lower chamber was filled with 500 μl complete DMEM medium containing 10 μg/ml oxLDL and 10% FBS. After 36 h incubation in a 37°C, 5% CO₂ incubator, migrated cells were fixed with 4% paraformaldehyde, washed then stained with 0.1% crystal violet for 30 min. Images were captured through an inverted microscope (Olympus), and cells counted with the ImageJ software.

The primary human umbilical vein endothelial cell (HUVEC) was from Dr. Xinli Wang (Baylor College of Medicine, Houston, TX), and cultured as described previously (Wu et al., 2016). Expression of G2A in the cells was knocked down by siRNA (about 60% knockdown assessed by real-time PCR). ROS was detected by means of an oxidation-sensitive fluorescent probe (DCFH-DA) (ROS Assay Kit, S0033S, Beyotime Biotechnology, China). After treated with 20 μg/ml oxLDL for 36 h, cells were washed then incubated with 10 μmol/L DCFH-DA according to the manufacturer’s instructions. The dichlorofluorescein (DCF) fluorescence was detected by FlexStation 3 (Molecular Devices) at an excitation wavelength of 488 nm and emission wavelength of 525 nm. The oxidative stress marker, malondialdehyde (MDA) in serum was conducted by a plate reader assay, based on reaction between MDA and thiobarbituric acid, as described in Lipid Peroxidation MDA assay kit (S0131S, Beyotime Biotechnology, China). OD535nm was detected by Cytation 5 imaging reader (BioTek, United States).

At 8, 24, and 36 week-old, rats were fastened overnight. Blood samples were collected under anesthesia through intraperitoneal injection of 1.25% (m/v) avertin (Sigma-Aldrich). Serum was obtained by centrifugation at 3,000 rpm for 15 min at 4°C, and then kept at –80°C until analysis. Lipids including total cholesterol (TC), triglyceride (TG), low-density lipoprotein cholesterol (LDL-c) and high-density lipoprotein cholesterol (HDL-c) were analyzed using AU680 Automatic Biochemistry Analyzer (Beckman Coulter, United States). For atherosclerosis risk prediction, the atherosclerotic index was calculated as (TC- HDL-c)/HDL-c (Zhao et al., 2018; Zhang et al., 2020). Levels of oxLDL, free fatty acid, insulin and leptin were determined by respective ELISA assays from Shanghai Hengyuan Biological Technology Co., Ltd. (Shanghai, China).

Total RNA was extracted from cells or tissues using Trizol (Invitrogen). Reverse transcription was performed with the cDNA Prime Script Reverse Transcription kit (Takara) according to the manufacturer’s instructions. Real-time PCR was performed in triplicate by SYBR Green PCR Master Mix (Takara) on a MX3005p system (Stratagene, United States). 500 ng RNA was used for the reverse transcription, and 500 ng cDNA for real-time PCR. The expression of target genes was normalized to the house keeping gene β-actin.

The histological analysis was performed as previously published (Zhao et al., 2018). Briefly, rats were sacrificed at 36 or 54 weeks old by CO₂ asphyxiation under avertin anesthesia. Whole aortas were excised and adipose tissue and adhesion tissue around the arteries were removed. Aortas were fixed in 4% paraformaldehyde at 4°C overnight. The fixed tissues were paraffin-embedded, and sections were prepared using a paraffin slicer (Leica) then subjected to hematoxylin-eosin (H&E) staining. The en face aortas or frozen sections were subjected to Oil red O staining, Masson staining or immunohistochemical staining. Stained sections were photographed by an inverted microscope (Leica) or Caseviewer2.0 (3D HISTECH), and the resulted images were analyzed by Image-Pro^® Plus version 6.0 software.

Ice-cold SDS lysis buffer (0.08 mM Tris–HCl, 10 SDS, 0.25% bromophenol blue, 50% glycerol and 25% β-mercaptoethanol) was used to collect proteins from cells. Heat denatured proteins were equally loaded and fractionated by 12% SDS-PAGE, then transferred to a nitrocellulose membrane (Schleicher and Schuell MicroScience). According to the target protein, membranes were incubated with the following respective primary antibodies: rabbit anti-PI3K (1:1,000 dilution, Cat. No. 4257), rabbit anti-phospho-PI3K (1:1,000, Cat. No. 4228), rabbit anti-AKT (1:1,000, Cat. No. 9272), rabbit anti-phospho-AKT (1:2,000, Cat. No. 4060), and rabbit anti-Bcl-xL (1:1,000, Cat. No. 2764) from Cell Signaling Technology (CST), rabbit anti-Bcl-2 (1:1,000, sc-56015) from Santa Cruz and mouse anti-β-actin (1:1,000, A5441) from Sigma. Secondary antibodies were donkey anti-mouse or donkey anti-rabbit (1:2,000) from Thermo-Invitrogen. Detection of positive bands were conducted using fluorescence labeled secondary antibodies IRDye^® 800CW goat anti-mouse IgG (Cat. No. 926-32210) and IRDye^® 800CW goat anti-rabbit IgG (Cat. No. 926-32211) from Li-COR. Images of the blots were analyzed on the Odyssey Infra-Red Imaging System (Li-COR).

All data are expressed as mean ± SD. Statistical analysis was performed by Graphpad prism 6 software. Differences between two groups were determined by an unpaired Student’s t-test. For multiple comparisons, two-way ANOVA with Tukey’s multiple comparison test was used. Values of p < 0.05 were considered statistically significant.

To evaluate its possible involvement in atherosclerosis, G2A expression in atherosclerosis related tissues and cells were first analyzed in WT rats. Similar to mice G2A, rat G2A was found highly expressed in immune organs including thymus and spleen, and widely expressed in immune cells, such as lymphocytes and different subpopulations of macrophage, with the bone marrow cells showed the highest expression. There was also some G2A expression in the visceral adipose tissue (VAT) and aorta (Supplementary Figure 3).

Next, phenotypes associated with atherosclerosis development including lipid metabolism, oxidative stress and macrophages were compared between G2a^–/– and WT rats. Serum lipid profile analysis did not show a significant difference in the two genotypes in either normal diet or Western diet-fed groups (Supplementary Figure 4). Neither was an obvious change discovered in cytokines secreted by macrophages, except a minor increase in the inflammatory cytokine TNFα by LPS stimulated G2a^–/– macrophages (Figures 1A,B). However, consistent enhancement was detected in macrophage migration. As shown in Figure 1C, bone marrow derived macrophages from 8 week-old G2a^–/– rats displayed enhanced migration when induced by ox-LDL. Moreover, the reactive oxygen species (ROS) producing enzyme nitric oxide synthase 2 (iNOS) was markedly increased in arteries of G2A deficient rats (Figure 1D), indicating oxidative stress and cell injury. The expression of apoptotic genes was significantly enhanced while survival genes were reduced in G2a^–/– macrophages (Figure 1E).

We further analyzed the effects of G2A deficiency on ATP binding cassette transporters and lipid scavenger receptors. ABCA1 did not show significant change, but a decreased ABCG1 mRNA level was noticed in G2a^–/– macrophages (Figure 1F). No significant difference in LDLR-related proteins (LRP1, LRP5, and LRP6) was detected. There were some changes in the expression of lipid scavenger receptors (SRs), though, with increases in SRB1 and lectin-like oxidized LDL receptor-1 (LOX-1) but a reduction in SRA1 (Figure 1G). Taken together, G2A knockout in WT rats did not affect lipid metabolism but led to arterial oxidative stress as well as enhancement in macrophage migration and apoptosis.

Although G2A deficiency did not alter lipid profiles in WT rats, an apparent difference was observed between sera from 24 week-old Ldlr ^–/– and Ldlr^–/–G2a^–/– rats with the former clearer and the latter more turbid (Figure 2A). Not surprisingly, greatly augmented TC and TG were detected in Ldlr^–/–G2a^–/– rats. A significant increase for TC even began at 8 week-old, before Western diet induction (Figure 2B). The HDL-c level did not show significant change, but a marked elevation of LDL-c was detected at 36 week-old, along with a trend of increase in LDL/HDL ratio (Figure 2C). Interestingly, a greatly up-regulated atherosclerotic index was manifested in 24 and 36 week-old Ldlr^–/–G2a^–/– rats as compared to respective Ldlr^–/– controls (Figure 2D), predicting higher atherosclerosis risk. No significant alteration was noticed in free fatty acids, but a very consistent increase was found in the serum level of oxLDL, even at 8 weeks old before the Western diet intervention (Figures 2E,F). To analyze whether the aggravated lipid disorder had anything to do with food intake and body weight, we checked the two indexes, as well as water intake, leptin, and insulin. Only water intake in male Ldlr^–/–G2a^–/– displayed a moderate enhancement as compared to the male Ldlr^–/– counterpart (Supplementary Figure 5). No significant difference in these indexes was detected between female rats of the two genotypes (data not shown). Therefore, G2A deficiency severely aggravated lipid disorder in Ldlr^–/– rats which had no apparent relation to food intake or body weight.

After finding the exacerbated dyslipidemia in Ldlr^–/–G2a^–/– rats, we next examined the direct effect of G2A deficiency on atherosclerosis development. Not surprisingly, male and female Western diet-fed Ldlr^–/–G2a^–/– rats at 36 and 54 weeks of age showed significantly more aortic plaques by en face analysis, as compared to Ldlr^–/– rats. Representative en face aortic images were presented in Figure 3A and lesion areas accounted for 7.2–15.4% of total aortic intima in 36 week-old and 15.3–17.3% in 54 week-old Ldlr^–/–G2a^–/– rats, average 1.9 and 1.5-fold increase of lesion area in respective Ldlr^–/– counterparts. Besides, histological analysis showed more lipid deposition in different segments of the Ldlr^–/–G2a^–/– aorta, especially in the aortic sinus and coronary arterial orifice (Figures 3B,C). To analyze whether G2A deficiency affected other tissues, we examined the liver, spleen, heart, kidney, visceral adipose tissue (VAT), and lung. No significant change was discovered in their relative tissue weight (data not shown). However, there was more lipid deposit in the Ldlr^–/–G2a^–/– liver without significant change in aspartate aminotransferase (AST) or alanine aminotransferase (ALT) levels. A similar result was noticed in the kidney of Ldlr^–/–G2a^–/– rats, with more lipid deposits but no significant difference in uric acid or creatinine level. Moreover, no apparent histological change was found in the liver or kidney (Supplementary Figure 6). The collective data means G2A deficiency in Ldlr^–/– rats led to increased atherosclerosis development, and enhanced lipid deposition in the liver and kidney.

As macrophage infiltration plays an important role in atherosclerosis development (Martens et al., 1998), we proceeded to perform immunohistochemical analysis to see whether there is any change in aortic macrophages. Indeed, increased macrophage content was found in the aortic sinus of Ldlr^–/–G2a^–/– as compared to Ldlr^–/– rats (Figure 4A), together with markedly elevated CD68 mRNA level (Figure 4C). In addition, increased staining of the adhesion molecule VCAM1 was also found in the aortic sinus of Ldlr^–/–G2a^–/– rats, together with increased mRNA levels of VCAM1, ICAM1, and E-selectin, although the VCAM1 change did not reach significance (Figures 4B,D). Moreover, Masson staining revealed more collagen deposition in the vessel wall of Ldlr^–/–G2a^–/– rats (Figure 4E). Hence the aggravated atherosclerosis in G2A deficient Ldlr^–/– rats was accompanied by increased macrophage recruitment and vascular fibrosis.

The above results and G2A single deletion’s effect on macrophages prompted us to analyze functional changes in monocytes and macrophages. Both bone marrow-derived monocytes and macrophages from Ldlr^–/–G2a^–/– rats exhibited a higher migration ability than those from Ldlr^–/– rats (Figures 5A,B). Concomitantly, there were consistently and significantly increased aortic chemokines related to monocyte and macrophage recruitment (Figure 5C), indicating increased interactions between these cells and endothelium. The increased macrophage chemotaxis may also attribute to matrix metalloproteinase (MMP) enhancement, as increased expression of MMP1 and MMP9 was found in Ldlr^–/–G2a^–/– macrophages (Figure 5D). Since the significantly increased CCL2 is related to inflammation, the inflammatory phenotype of macrophages was also examined. After 24 h stimulation by oxLDL, Ldlr^–/–G2a^–/– macrophages displayed a minor increase in inflammatory IL-1β and reduction in anti-inflammatory IL-10 as compared to Ldlr^–/– macrophages. In polarization conditions, G2A deficient macrophages showed a significant increase of IL-1β when stimulated by LPS, and decreased TGF and IL-10 when stimulated with IL-4 (Supplementary Figure 7).

The expression of genes encoding ATP-binding cassette transporters, responsible for cholesterol efflux, was unaffected by G2A deficiency as examined by real-time PCR (Figure 5E). However, the expression of scavenger receptors (SRs) including LOX1, CD36, SRA1, and SRB1 were all significantly increased (Figure 5F). Among them, LOX-1 displayed the greatest elevation. SRs mediated oxLDL internalization was also enhanced by G2A deletion (Figure 5G). Intriguingly, the cell number was declined when macrophages were incubated with Dil-oxLDL for 24 h, as compared to 18 h incubation (data not shown), implying possible cell apoptosis.

Further examination revealed G2A regulating macrophage apoptosis which is related with oxidative stress. As in WT, the absence of G2A in Ldlr^–/– rats greatly up-regulated iNOS mRNA expression, both in the aorta and oxLDL stimulated macrophages, suggesting stress associated with increased blood oxLDL and oxLDL uptake. This was confirmed by the elevated serum MDA level in Ldlr^–/–G2a^–/– rats and ROS generation in G2A knockdown HUVEC cells (Figures 6A,B). As shown in Figure 6C, the expression of apoptotic genes p53 and caspase 12 was enhanced while survival genes (Bcl2, Dad, and Bcl-xl) were reduced in Ldlr^–/–G2a^–/– macrophages. In addition, flow cytometry analysis (FACSCalibur, Becton Dickinson, United States) manifested significantly elevated Annexin V-positive apoptotic cells (Figure 6D). The result was corroborated by in situ TUNEL assay which showed increased corresponding fluorescence in section preparations of Ldlr^–/–G2a^–/– aortas (Supplementary Figure 8). Apoptotic pathway-related genes were then probed by immunoblotting. There was significantly reduced phosphorylation of PI3 kinase (PI3K) and AKT (Figure 6E and Supplementary Figure 9), as well as a decrease of downstream gene expression including Bcl2 and Bcl-xl. The data suggested possible participation of PI3K/AKT signaling pathway in G2A regulation to macrophage apoptosis. Finally, to examine whether there was any alteration in efferocytosis, the clearance of apoptotic cells, LDLR-related proteins in macrophages were analyzed significantly suppressed expression of LRP1 and LRP2 was detected. Moreover, Gas6 and Mfge8 were inhibited although only the latter reached significance (Figure 6F). Altogether, the above results indicate that G2A deletion induced atherosclerosis aggravation was associated with macrophage changes including increased chemotaxis and apoptosis.