Synergistic inflammatory signaling by cGAS may be involved in the development of atherosclerosis

Mitochondrial DNA damage in atherosclerosis

It is well known that mtDNA damage correlates with atherosclerosis progression [15, 18, 19]. DNA with oxidative damage is resistant to degradation by cytosolic nuclease three prime repair exonuclease 1(TREX1), and enhances STING-dependent immune sensing [20]. Released oxidized mtDNA acts as a powerful inflammatory stimulus [16]. In accordance with previous findings, immunofluorescence revealed that 8-Oxo-2′-deoxyguanosine (8-OH-dG), a marker for oxidative damage to DNA, can be detected in the plaque of ApoE ^-/- mice and strong 8-OH-dG staining was discovered mainly in mitochondria, irrespective of diet. In addition, part of the 8-OH-dG staining overlapped with neither nucleus nor mitochondria (Figure 1A), suggesting that oxidized DNA can be released into the cytoplasm or outside the cell. Additionally, human atherosclerotic plasma contained higher concentrations of dsDNA than those in healthy controls (Figure 1B), consistent with results demonstrated by previous studies [11, 13]. This released DNA could be the actuator for the development of atherosclerosis.

Differences in cGAS expression identified by database analysis based on the Gene Expression Omnibus

Although cGAS acts mainly as a DNA sensor and the role in broadly eliciting immune responses and inflammation has also been reported, whether cGAS contributes to atherosclerosis is yet to be uncovered. In the present study, the dataset GSE40156 from the Gene Expression Omnibus (GEO) was re-analyzed by bioinformatics methods and it was found that the expression of cGAS was increased in the total aorta of ApoE ^-/- mice. Furthermore, cGAS expression in ApoE ^-/- mice was significantly increased with age (Table 1). It is well known that macrophages are of great importance in the initiation, progression and regression of atherosclerosis and different subtypes of macrophages affect the outcome of atherosclerosis [21]. To explore the difference of cGAS in macrophage polarization, GSE57614 from the GEO was assayed and it was found that cGAS expression remarkably increased in M1 macrophages when compared with M2 macrophages (Table 2). Collectively, the results of these bioinformatics analyses results mentioned above imply that cGAS may play a pivotal role in atherosclerosis.

cGAS as a novel factor for modulating atherosclerosis

To explore the effects of cGAS on atherosclerosis, ApoE ^-/- mice were kept on a western-type or chow diet for 16 weeks. Results demonstrated that cGAS was expressed in the plaque of ApoE ^-/- mice regardless of diet, and was positively correlated with plaque area, and mainly distributed in macrophages (Figure 2A, 2B), whereas, the protein expression of cGAS in the aorta of ApoE ^-/- mice fed a western-type diet was higher than that of ApoE ^-/- mice fed a chow diet (Figure 2C, 2D). Importantly, cGAS inhibition reduced lipid deposition and foam cell formation in RAW264.7 macrophages (Figure 3A, 3B). These data demonstrated that cGAS contributed to atherosclerosis, and suggested that exploring the underlying mechanism could be of great significance.

Transcriptome profiling of RU.521-treated RAW264.7 macrophages

In order to have a better view of the mechanisms of atherosclerosis mediated by cGAS, we performed RNA-seq transcriptomic analysis of the underlying DEGs and the mechanisms in RAW264.7 macrophages treated for 12 h with or without RU.521, a cGAS inhibitor. The genes expressed were sequenced using the Illumia platform. DEGs between the cGAS inhibition group and the control group were detected. First, the intra-group samples showed a strong correlation (Figure 4A). Further, the repeatability of intra-group data based on principal component analysis (PCA) was acceptable, where the distances of intra-group samples were close. (Figure 4B).

A total of 1,462 DEGs were identified with the criteria of log2 (Fold Change, log2FC) ≥ 2 and Padj < 0.05, including 845 upregulated genes and 617 down-regulated genes (Figure 4C, 4D and Supplementary Table 2). Good distinction of DEGs between different groups was presented in the hierarchical clustering heatmap (Figure 4E).

DEGs and functional enrichment analysis associated with atherosclerosis

In order to understand whether cGAS inhibition changes expression of the genes involved in modulating atherosclerosis, we identified DEGs in cGAS inhibition probably linked to atherosclerosis according to research published in the PubMed database. A total of 275 DEGs associated with atherosclerosis were found when cGAS was inhibited (Supplementary Table 3). These DEGs included colony-stimulating factors (Csfs such as Csf2, Csf2rb, Csf1r, and Lif.), interleukins (Ils, including Il1b ,Il7, Il18, Il10.), tumor necrosis factor family members (such as TNF superfamily members–Tnfsf–Tnfsf10, Tnfsf8, and Tnfsf11a.), toll-like receptors (Tlrs including Tlr3, Tlr5, and Tlr7), interferon regulatory factors (Irfs, Irf5, Irf7, and Irf9), genes related to cholesterol metabolism (Msr1, Acat2, Apoa1, Abca1 and Abcg5), signal transducers and activators of transcription (Stats, such as Stat1, and Stat2), IFN and IFN-stimulated genes (Isgs, Ifna, Ifnb, Ifit1, Ifit2, and Isg15.) and so on. In addition, expression of Il6, Tlr9, Irf3, Cd36, Acat1 and Sting were also decreased despite absolute log2FC < 2.

Gene Ontology (GO) analysis showed that the main functional enrichments occurred in biological processes (BP), including inflammatory response, immune response, cholesterol biosynthetic process and so on, which are of great importance in the formation and progression of atherosclerotic plaque. In addition, GO terms linked to molecular function (MF) and cellular compartments (CC) were also discovered, such as extracellular space, cell surface, endoplasmic reticulum, and cytokine activity (Figure 5A, 5B and Table 3). Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis showed comparable results, with the most significant KEGG terms associated with atherosclerosis being cytokine–cytokine receptor interaction, Jak–STAT signaling pathway, steroid biosynthesis, chemokine signaling pathway, TNF signaling pathway, fat digestion and absorption, PPAR signaling pathway and TLR signaling pathway (Figure 5C, 5D and Table 4).

Moreover, Metascape analysis showed remarkable enrichment in regulation of the cytokine-mediated signaling pathway (Figure 6A–6C). There were 22 overlapped DEGs in both GO enriched terms “Inflammatory response” and “Immune system process”/ “Immune response”/ “Innate immune response” and seven overlapped DEGs in the KEGG analysis terms “Cytokine–cytokine receptor interaction”, “TNF signaling pathway”, “Jak–STAT signaling pathway”, “ Chemokine signaling pathway” and “Toll-like receptor signaling pathway”(Figure 6D, 6E and Tables 5, 6).

Protein–protein interaction network of DEGs associated with atherosclerosis

Next, the protein–protein interaction (PPI) network of DEGs was built as shown below (Figure 7A). The hub genes identified by the MCC method were identified as follows: Stat1, Irf7, Ifih1, Isg15, Ifit1, Rsad2, Stat2, Irf9, Ifi44, and Dhx58, exhibiting lower expression in cGAS inhibition (Figure 7B). A functional subnet module containing hub genes was selected from the PPI network and functional enrichment analysis was performed using ClueGO. The nodes in this module related to atherosclerosis were mainly enriched in the “response to interferon” (Figure 7C–7E).

Verification of the genes expressed in synergistic inflammatory signaling

We sought to verify and extend the main findings suggested by RNA-seq. RAW264.7 cells were treated with or without the cGAS inhibitor RU.521(2 μg/mL) for 12 h and then real-time quantitative polymerase chain reaction (RT-qPCR) was performed. As shown in Figure 8A, a robust anti-inflammatory effect can be induced by cGAS inhibition via increasing the expression of anti-inflammatory factors (Il10, Arg1, Lif, Il1rn) and reducing inflammatory factors (Il1b, Il7, Tnfsf10, Il18). Additionally, Il6 expression was also significantly decreased despite absolute log2FC < 2 in the RNA-seq results. In general, macrophages are classified as classical M1 or alternative M2. M1 macrophages are pro-inflammatory with high levels of cytokines such as IL 1B, IL6, and TNFα. Meanwhile M2 macrophages show anti-inflammatory properties, secreting anti-inflammatory factors, such as IL10 and ARG1. When stimulated by TLR ligands and IFNs, macrophages can transform into the M1 phenotype [22]. Furthermore, it is widely known that abnormal TLR signaling contributes to chronic inflammation [23]. Our results demonstrated that cGAS inhibition reduced the mRNA expression of Tlrs and, Ifns as well as the IFN-responsive genes (Figure 8B, 8C). These indicated that cGAS inhibition can restrain M1 polarization, promote M2 polarization, and that TLR or IFN signaling mediated by cGAS may participate in this process. Evidence suggests that M1 polarization can also be regulated by STAT1, IRF3, and IRF5 [24, 25], and inflammatory genes can be regulated by STAT1–IFN–IRF through over-expression in their promoters [26]. Previous studies have reported that STAT and IRF in the atherosclerotic tissues are displayed at significantly higher levels when compared with the matched normal tissues [27–31]. Here, Stat (Stat1, Stat2) and Irf (Irf3, Irf5, Irf7, Irf9) mRNA were significantly down regulated under cGAS inhibition (Figure 8B), consistent with the down regulation of inflammatory factors. RAW264.7 macrophages were pretreated with RU.521 for 12 h and then incubated for another 6 h with lipopolysaccharide (LPS, 10 ng/mL) plus interferon gamma (IFNγ, 20 ng/mL) that polarizes macrophages to the M1 phenotype. The results showed that cGAS significantly inhibited the M1 phenotype by decreasing Il1b, Il6, Tnfa, and Cd86 while increasing Il10, Arg1, and Nos2 mRNA expression which are characteristic of the M2 phenotype. Consistently, cGAS inhibition also inhibited the up-regulation of Stat and Irf in response to LPS plus IFNγ stimulation (Figure 8F). These results indicated that STAT and IRF signaling mediated by cGAS may be the key point of macrophage polarization through inflammatory regulation.

Given that foam cell formation acts as a hallmark of atherosclerosis, the expression of genes contributing to cholesterol uptake, cholesterol esterification and hydrolysis as well as cholesterol efflux was evaluated via RT-qPCR. cGAS inhibition blocked expression of Cd36, Msr1, Acat1, and Acat2, which mediate cholesterol uptake, cholesterol esterification and hydrolysis. In contrast, cGAS inhibition significantly increased the mRNA expression of Apoa1 and Abcg5, which contribute to cholesterol efflux, consistent with the RNA-Seq results. However, cGAS failed to increase the expression of Abca1 that promotes cholesterol efflux (Figure 8D). These results indicated that cGAS inhibition could suppress foam cell formation by hindering cholesterol uptake. Obviously, this assumption was confirmed in the RAW264.7 in vitro model (Figure 3C, 3D). In addition, the mRNA expression of colony-stimulating factors as well as chemokines was also verified (Figure 8E).

Animals

Eight-week-old ApoE ^-/- male mice on a C57BL/6 background from Hua Fukang Experimental Animal Center (Beijing, China) were kept under standard conditions with constant humidity (approximately 40%) and temperature (approximately 22° C) as well as a 12-h light/dark cycle. A standard chow diet was offered for 1 week prior to all procedures, after which a western-type diet containing 0.21% cholesterol and 41 kCal% fat (Special Diet Services, Huafukang, China) or continuation on the standard chow diet was offered for 16weeks. Hearts and aortas were harvested for assessment of atherosclerotic burden.

Cell culture and treatment

RAW 264.7 macrophages purchased from the Cell Resource Center of Shanghai Institute for Biological Sciences were grown in Dulbecco’s Modified Eagle’s Medium (DMEM, HyClone, Logan, UT, USA) containing 10% heat-inactivated fetal bovine serum (FBS, ScienCell Research Laboratories, Carlsbad, CA, USA), 1%penicillin and 1% streptomycin (HyClone). Cells were incubated under standard conditions with constant temperature (37° C) and an atmosphere of 5% CO₂. Cells were grown to 60–80% confluence before use. In order to perform RNA-seq, cells were treated for 12 h with RU.521 (cGAS inhibitor 2 μg/mL, InvivoGen, San Diego, CA, USA) or with DMSO as vehicle control. In addition, RAW264.7 macrophages were pretreated with RU.521 (2 μg/mL) for 12 h and then incubated for another 6 h with LPS (10 ng/mL, HY-D1056, MedChemExpress, Monmouth Junction, NJ, USA) plus IFNγ (20 ng/mL, HY-P7071, MedChemExpress) that polarize macrophages to M1.

Histologic and morphometric analysis

As previously described, mouse hearts were embedded in OCT and frozen at −80° C after perfusion, fixation, and dehydration. Then serial sections were cut at 10μm thickness using a cryostat (CryoStar NX50, Thermo Scientific, Waltham, USA). Hematoxylin and eosin (HE) staining of sections was performed for quantification of lesion areas.

Immunofluorescent staining

Immunofluorescent staining was performed as previously described [65]. The primary antibodies used were as follows: cGAS (Mouse, 1:100, sc-515777, Santa Cruz Biotechnology, Santa Cruz, CA, USA), MOMA-2 (Rat, 1:100, ab33451, Abcam, Cambridge, UK), CD31 (1:100, AF3628, R&D Systems, Minneapolis, MN, USA) and α-SMA (Rabbit, 1:100, GB13044, Servicebio, Wuhan, China). 4′,6-diamidino-2-phenylindole (DAPI, Invitrogen, Carlsbad, CA, USA) was applied to stain nuclei. Fluorescent images were captured with a Nikon A1Si confocal microscope (Nikon, Tokyo, Japan), and analyzed with NIS Elements AR Imaging Software 4.10 (Nikon) and ImageJ 1.41 software (NIH, Bethesda, MD, USA).

Dil-oxLDL uptake assay

Cells were treated with or without RU.521(2 μg/mL) for 24 h, followed by incubation with 20 μg/mL Dil-LDL (YB-0010, Yiyuan Corporation, Guangzhou, China) for 6 h or 8 h at 37° C. Then the cells were fixed and analyzed by immunofluorescent staining.

Cholesterol efflux

Cholesterol efflux assays were performed as previously described [66]. Briefly, RAW264.7 macrophages were labeled with NBD (N1148, Life Technologies, Carlsbad, CA, USA) in the presence of RU.521 (2 μg/mL). After the cholesterol pools were equilibrated, cells were incubated with HDL (50 μg/mL, YB-003, Yiyuan Corporation) for 6 h. The fluorescence intensity of NBD was measured at 485nm excitation and 535nm emission. Efflux was measured as a percentage of fluorescence intensity in medium/(fluorescence intensity in medium + fluorescence intensity in cells) ×100%.

Foam cell formation assays

RAW264.7 macrophages were incubated in DMEM containing oxLDL (100 μg/mL, YB-002, Yiyuan Corporation) with or without RU.521 (2 μg/mL) for 24 h. After fixation for 20 min, cells were blocked with 1% BSA, followed by staining with BODIPY 493/503 (D3922, Invitrogen) for 30 min. Next, nuclei visualized by staining DAPI for 10 min. Finally, antifade mounting medium (P0126, Beyotime Institute of Biotechnology, Jiangsu, China) was applied to prevent fluorescence quenching. Images were acquired using a Nikon A1Si confocal microscope. Analysis of different images was performed using ImageJ and Adobe Photoshop CC. Oil Red O staining was carried out as previously described [66, 67]. RAW264.7 macrophages were incubated in DMEM containing oxLDL (100 μg/mL) with or without RU.521 (2 μg/mL) for 24 h. After fixation for 30 min, cells were stained with oil red O (G1015, Servicebio) for 30 min, and then visualized by light microscopy (Olympus, Tokyo, Japan).

Extraction and measurement of plasma dsDNA

Individuals (healthy volunteers and patients with atherosclerosis) were enrolled from Union Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, China. Atherosclerosis was confirmed through ultrasonography or MRA/CTA/DSA imaging of the carotid or cerebral artery. Individuals with autoimmune diseases were excluded. Blood was collected and the plasma was isolated. Extracellular dsDNA in plasma was extracted using a DNA Extractor SP Kit (ZWK-296-60501, Wako Pure Chemical Industries Ltd., Osaka, Japan) and quantified using the Quant-iT™ PicoGreen ® dsDNA Reagent Kit (P7589, Invitrogen) following the manufacturer’s instructions.

RNA extraction and real-time quantitative PCR analysis

Total cellular RNA was isolated and extracted using the TRIzol reagent (Vazyme, Nanjing, China). Quantitative RT-qPCR was carried out using the SYBR Green detection chemistry (Vazyme, Nanjing, China) in a 10μL reaction volume with an ABI 7500 Fast real-time PCR system. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) expression was used as the internal normalized reference gene. All samples were measured in triplicate, and the ΔΔCt method was applied to relative quantitative measurements. Primers are shown in Supplementary Table 1.

Western blotting analysis

Western blotting analysis was conducted as previously described [65]. Primary antibodies against cGAS (Mouse, 1:1000, sc-515777, Santa Cruz Biotechnology) and, GAPDH (GAPDH, A01020, Abbkine, Redlands, CA, USA) were used. ImageJ software (NIH) was used to analyze protein expression levels which were normalized to GAPDH.

Microarray data and data processing of DEGs

The microarray data of GSE40156 and GSE57614 derived from Gene Expression Omnibus (GEO) databases(https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE40156 and https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE57614) which were based on the GPL1261 [Mouse430_2] Affymetrix Mouse Genome 430 2.0 Array and GPL6480 Agilent-014850 Whole Human Genome Microarray 4x44K G4112F. DEGs were picked out using GEO2R online tools. A P value <0.05 and | log2FC | >1 were chosen as cut-off standards.

RNA sequencing

RNA sequencing was performed following a previously published method [67]. Briefly, for each sample, a total amount of 1μg RNA was used as input material. The NEBNext^® Ultra™ RNA Library Prep Kit for Illumina^® (New England Biolabs (NEB), Ipswich, MA, USA) was applied to prepare sequencing libraries, followed by generation of clusters with the TruSeq PE Cluster Kit v3-cBot-HS (Illumina Inc., San Diego, CA, USA). The Illumina Novaseq platform was utilized to sequence libraries and then about 150 bp paired-end reads were obtained. Analysis of differential expression was performed using the edgeR R package. The Benjamini and Hochberg method was used to adjust P value. A Padj < 0.05 and | log2FC | ≥ 2 were set as the thresholds for significant differential expression.

Functional enrichment analysis of DEGs

GO enrichment analysis and KEGG pathway analysis of DEGs was performed using the online tool DAVID (version: 6.8, https://david.ncifcrf.gov/) with thresholds of count ≥ 2 and P value < 0.05. Furthermore, Metascape (https://metascape.org/gp/index.html#/main/step1) was also used to obtain functional enrichment results.

Protein–protein interaction network analysis of DEGs

The PPI network analysis of DEGs was based on the STRING database (https://string-db.org/), which functions as a tool for predicting protein–protein interactions. A threshold of PPI score (medium confidence) ≥ 0.4 was used to predict the DEG-encoded proteins. Then, Cytoscape software (version: 3.6.0) was used to visualize the PPI network. Module and hub gene analyses were carried out with MCODE, cytoHubba plugin as well as ClueGO of Cytoscape software.

Statistical analyses

Results are expressed as mean ± standard deviation (SD). The analysis was completed using GraphPad Prism 6. In addition, the Student's t-test was used for comparison between the two groups. P values < 0.05 or Padj < 0.05 were considered statistically significant. R (3.6.2) software was used to display the distribution of DEGs as well as functional enrichment analysis.

Ethics statement

This study was approved by the Medical Ethics Committee of Tongji Medical College and the Institutional Animal Care and Use Committee, Huazhong University of Science and Technology (HUST), Wuhan, China (S947, S2342). Written informed consents were obtained from Individuals (healthy volunteers and patients with atherosclerosis).