2.1 Antibodies, Lentiviruses and Chemicals

Collagenase type II was purchased from Sigma. Antibodies against SQSTM1(#5114), ATG7 (#8858), ATG5(#12994), p-ULK1(Ser757) (#14202), 4EBP-1 (#9644T), LC3B (#83506), GAPDH (#5174), α-SMA (#19245), p-AKT (#9018), AKT (#4691), mTOR(#2972), p-mTOR(#2971) and Collagen I (#72026) were obtained from Cell Signaling Technology. ULK1 (#A8529), p70 S6 Kinase (#A2190), Phosoho-p70 S6 Kinase (#AP0502), P-4EBP-1(#AP 0032), PIK3CA (#A0265) were obtained from Abclone. Anti-DDR2 antibody (#ab126773) was obtained from Abcam. Western Secondary antibodies (Peroxidase AffiniPure Goat Anti-Rabbit IgG, #111-35003; Peroxidase AffiniPure Goat Anti-Mouse IgG, #115035003), Alexa Fluor 594-labelled anti-rabbit IgG (#711585152), Alexa Fluor 488-labelled anti-rabbit IgG (#715545150) were supplied by Jackson immune-research. DDR2 shrna lentivirus and DDR2-overexpressing lentiviruses were synthesized by HanbioBiotech Co., Ltd. rapamycin (#HY-10219), Chloroquine diphosphate (#HY-135811A), N-acetylcysteine (NAC, #HY-B0215), WRG-28 (#HY-114169), losartan (#HY-17512), PD 123319 (HY-10259A) and LY294002 (#HY-10108) were obtained from Medchemexpress (USA).

2.2 Animal models and treatments

DDR2^flox/flox (DDR2^fl/fl) mice with loxP sites flanking exons 9 of Ddr2 gene were generated by Shanghai Model Organisms (#NM-CKO-200311). S100a4-CreERT2 (FSP1-Cre) mice were a gift from GemPharmatech (#T006589). Fibroblast-specific DDR2 knockout mice (DDR2-cKO) were created by crossing DDR2^fl/fl mice with S100a4-CreERT2 transgenic mice.¹⁴ To induce Cre-mediated recombination, 8-week-old male DDR2-cKO mice received intraperitoneal injections of tamoxifen (75 mg/kg/day, dissolved in corn oil; Sigma-Aldrich, #T5648) for 5 consecutive days.¹⁵ Animal protocols were approved by Animal Research Committee of Xinhua Hospital, Shanghai Jiaotong University School of Medicine and daily health monitoring confirmed no adverse events or fatalities during the study. Littermate DDR2^fl/fl mice (Cre-negative) were used as controls to ensure genetic and environmental consistency. The Adventitial remodelling model was established as described previously.^{3, 16} Briefly, male wild-type (WT), DDR2^fl/fl, and DDR2-cKO mice (8-week-old) were administrated Ang II (1000 ng/min/kg, Abcam, ab120183) or saline via subcutaneous infusion for 4 weeks through Alzet miniosmotic pumps, while the each group of WT was further divided into two groups (subcutaneous injection with PBS or rapamycin(4 mg/kg/d)).

2.3 Aortic dissociation and single cell preparation

After the adventitial remodelling model was established, mice were euthanized by CO₂ overdose and aorta was harvested. Aortas from five mice in each group (control and Ang II) were pooled, longitudinally opened, and the adventitia were stripped for RNA sequencing. Single-cell suspension of adventitial cells was prepared using an enzymatic digestion protocol previously described.¹⁷ In brief, after rinsing with 1% PBS, the adventitia was cut and digested with an enzyme digestion [2 mg/mL collagenase type I (Invitrogen, #17018-029), 2 mg/mL disapase II (Sigma, #C2-BIOC) in Hank's balance salt solution containing calcium and magnesium] for 2 h at 37°C. The cell suspension was filtered through a 70 µm mesh, washed twice in the 1 × PBS, and resuspended at 1 × 10⁶ cells per ml in 1 × PBS containing 0.04% bovine serum albumin, before being subjected to scRNA-seq.

2.4 Single-Cell RNA Sequencing

The scRNA-seq was conducted using the Standard 10× Chromium Single Cell 3′ v3 protocol by 10X Genomics GemCode Technology. In summary, single cells were individually tagged with a unique 10× Barcode and unique molecular identifier through the encapsulation into Gel Bead-In-Emulsions. Sequences sharing identical 10× Barcodes were then classified as originating from the same cell. The resultant cDNA library was prepared and subjected to sequencing on the Nova PE150 platform.¹⁷

Raw reads were demultiplexed and mapped to the reference genome using the 10X Genomics Cell Ranger pipeline (https://www.10xgenomics.com/support) with default parameters. The Cell Ranger mkfastq command was employed to generate Fastq files, and the data were then mapped to the pre-built mouse reference genome (mm10, version 1.2.0).¹⁸

After aggregating samples from aortic adventitial cells, the Seurat v3.0 software package in R was deployed to carry out a series of data processing steps. These included filtering of cells, data normalization, conducting principal component analysis, detecting variable genes, forming clusters, and generating a t-SNE plot for visualization purposes. Gene expression was visualized using Seurat functions VlnPlot, FeaturePlot and DotPlot for violin plots, feature plots, and dot plots, respectively. Specific cluster markers were identified using the Seurat function FindAllMarkers. Data generated in present study will be made available in the Gene Expression Omnibus repository upon acceptance.

2.5 Histological and immunofluorescence, transmission electron microscopy

The sample preparation and analysis of the adventitia were performed following previously established methods.⁵ Briefly, thoracic aortas were collected and fixed in 4% paraformaldehyde for 24 h. After being embedded in paraffin, 5-µm transverse histological sections were cut and stained with haematoxylin-eosin (H&E), or used for additional special staining and immunostaining. The adventitial thickness and fibrosis area, primarily in the ascending aorta, were quantified using Image J software. Additionally, adventitial samples from the mice were fixed in 2.5% (v/v) glutaraldehyde. Ultrathin sections were prepared and examined using a Philips CM 10 transmission electron microscope.

2.6 Primary AFs cell culture, cell proliferation and migration assays

The methodology for the extraction and characterization of AFs from the healthy thoracic aortas of male C57BL/6J mice has been previously documented in our earlier published study.⁵ The EdU assay was employed to assess cell proliferation, the proliferation rate was determined by calculating the percentage of EdU-positive cells relative to the total cell count.¹⁹ The scratch-wound assay and transwell assay were used to evaluate the capabilities of cell migration.^{5, 19} AFs were transfected with lentivirus for 24 h or pretreated for 1 h with pharmacological agents (LY294002 NAC, rapamycin, CQ). For the scratch-wound assay, a standardized scratch was generated in confluent monolayers using a p200 pipette tip. Cells were maintained in serum-free medium for 24 h at 37°C, and wound closure was quantified as: Migration area (%) = (A₀ - A_n)/ A₀ × 100, where A₀ and A_n represent the initial and residual wound areas. For the transwell assay, 2 × 10⁴ cells in serum-free medium were seeded into 8.0 µm Transwell inserts (Corning, NY, USA).Inhibitors or Ang II were added to the lower chamber and migrated cells were fixed with methanol after 24 h, stained with Crystal Violet (China) and counted in five random fields per membrane under a phase-contrast microscope (Leica).

2.7 Western blot analysis

Tissues from mice and treated or control AFs were collected, washed and lysed as previously reported.⁵ Equal loading of protein was assessed by BCA, and polyvinylidene difluoride membrane was used for transfers, which was blocked with 5% BSA. The membranes were incubated with primary antibodies against a range of proteins including Collagen I, α-SMA, GAPDH, DDR2, p-AKT, AKT, p-mTOR, mTOR, p-p70S6K, t-p70S6K, p-4EBP-1, t-4EBP-1, p-ULK1, t-ULK1, ATG7, ATG5, LC3B, and SQSTM1. All primary antibodies were diluted to 1:1000 using dilution buffer from Beyotime. Subsequently, the membranes were exposed to corresponding secondary antibodies. After thorough washing, protein signals were captured by the Enhanced chemiluminescence Western blot detection system and the bands were quantified by Image J software.

2.8 Co-immunoprecipitation

Co-immunoprecipitation was performed using a Pierce™ Co-Immunoprecipitation Kit (Thermo Fisher, #26149) to assess the interaction between DDR2 and PI3K, following the manufacturer's instructions. Briefly, affinity-purified antibodies against PI3K (25 µg) or DDR2 (25 µg) were immobilized onto the resin for 90 minutes at room temperature. After three washes with 1× coupling buffer, 500 µg of protein lysate was added to the resin, and the mixture was incubated overnight at 4°C. After incubation, the resin was washed four times before the immunoprecipitated proteins eluted and analyzed by Western blotting.

2.9 Reactive oxygen species measurement

For the quantification of intracellular reactive oxygen species (ROS), cells were stained using MitoSOX (5 µM, Thermo Fisher, #M36009) for a duration of 10 min prior to being fixed with 4% paraformaldehyde. Subsequently, coverslips were mounted, and the cells were examined using a confocal fluorescence microscope. Image J software quantified Fluorescence intensity across 4–5 separate fields.

2.10 PI3K activity assay

PI3K activity in cell lysates was measured utilizing the PI3K activity assay kit (Echelon Biosciences, #K-1000s) as per the guidelines provided by the manufacturer. These experiments were conducted in triplicates, and the outcomes from treatments were compared with the baseline values from control samples.

2.11 Transfection

Lentivirus harbouring DDR2 RNA interference (LV-sh-DDR2) and DDR2 over-expression (LV-DDR2) were designed by HanbioBiotech Co., Ltd. AFs underwent viral infection at a multiplicity of infection (MOI) of 50. Prior to initiating the main experiments, the efficacy of the infection was verified using Western blot analysis. The AFs were infected by dual-tagged LC3 (mRFP-GFP-LC3) lentivirus (Shanghai Genechem Co.,Ltd.) for autophagy flux analysis. Briefly, the cells were infected by lentivirus (20 MOI) for 4 h followed by a treatment with Ang II (100 nM) and rapamycin (1 µM) for 24 h. Following fixation using 4% paraformaldehyde, cells were examined under a fluorescence microscope, and the quantities of GFP and mRFP dots were ascertained through manual enumeration of fluorescent puncta across 10 separate fields derived from three distinct fibroblast preparations. The ratio of dots per cell was quantified by dividing the aggregate number of dots by the total count of nuclei within each individual microscopic field.

2.12 Statistical analysis

Statistical evaluations were conducted using Graph-Pad Prism software (version 10.4). All experiments were performed with a minimum of three independent biological replicates. Data are expressed as mean ± standard error of mean (SEM). Comparisons among multiple groups were analyzed using two-way analysis of variance (ANOVA), followed by Bonferroni's post-hoc test for multiple comparisons. Non-parametric Kruskal–Wallis tests with Dunn–Benjamini–Hochberg post-hoc corrections were applied for multi-group comparisons of protein levels. Differences between two groups were compared by Student's t-test (or Mann–Whitney U test for non-parametric data). A p-value less than 0.05 was deemed to denote statistical significance.

3.1 DDR2 is highly expressed in adventitial fibroblasts and upregulated in Ang II induced hypertensive mice

To characterize the adventitial cellular landscape, we isolated and enzymatically dissociated adventitial cells from male WT mice 12 weeks of age for scRNA-seq. After quality control, a total of 6332 cells from the control group and 6058 cells from the Ang II-infused group were selected for further analysis. Unbiased clustering using Seurat canonical correlation analysis revealed 9 clusters, which were visualized with t-distributed stochastic neighbour embedding (Figure 1A). Most cell types were identified, including B cells, T cells, NK cells, Endothelial cells, AFs, smooth muscle cells, Mast cells and Macrophages, with matched transcriptional profiles and putative biological roles (Figure 1A). To decipher the functional differences after Ang II stimulation, we used gene ontology (GO) analysis to examine changes in gene molecular functions, cellular components, and biological processes. GO analysis indicated that after Ang II stimulation, the main effects were on cell proliferation, migration, extracellular matrix adhesion, collagen fibre organization, and positive regulation of angiogenesis (Figure 1B). Single-cell transcriptomic analysis revealed constitutive high expression of Ddr2 in AFs, with a unimodal expression distribution (Figure 1C). Comparative analysis demonstrated that Ang II treatment increased the proportion of Ddr2 (Figure 1D). Building upon single-cell transcriptional evidence of Ddr2 upregulation, our time-course Western blot analysis (Figure S1) revealed that Ang II induced rapid DDR2 protein accumulation in AFs, with detectable increases as early as 5–10 min post-stimulation.

3.2 DDR2 deficiency restores Ang II-induced AFs phenotype switching and autophagy inhibition

Previous studies suggest that abolition of DDR2 reduces lung fibrosis and angiotensin-induced cardiac fibrosis by regulating the expression of collagen and vital cellular processes,^{11, 20} which also indicate that DDR2 might as well be required for adventitial remodelling. Consistent with our prior findings demonstrating Ang II-induced adventitial remodelling,⁵ Western blotting and immunofluorescence analyses revealed that Ang II mechanistically upregulated α-SMA and collagen I expression in AFs. These findings collectively confirm Ang II-driven phenotypic switching of AFs toward a pro-fibrotic myofibroblast state.²¹ Present results also showed that Ang II induced acute activation of DDR2 within minutes (Figure S1). To delineate DDR2's functional involvement in Ang II-induced AFs activation, we systematically assessed phenotype switching, proliferative capacity, and migratory responses through integrated experimental approaches. As shown in the Figure 2A,B, the phenotypic transition of AFs following Ang II stimulation was attenuated by the transfection of LV-sh-DDR2 as determined by the lower expressions of α-SMA and Collagen I in LV-sh-DDR2 + Ang II treated cells in contrast to Ang II treated AFs. And as shown in Figure 2C–F, LV-sh-DDR2 reduced Ang II-mediated proliferation and migration of AFs.

As previous studies have shown that autophagy may mediate AFs activation in atherosclerosis,²² we also investigated the effect of DDR2 on autophagy. Interestingly, autophagy suppression by Ang II was restored in DDR2 deficient AFs (Figure 2G,H). These results suggest that suppression of DDR2 alleviates Ang II-induced AFs phenotype switching, and thus there could be a link between DDR2 and autophagy.

3.3 DDR2 overexpression accelerates Ang II-induced autophagy impairment and AFs phenotype switching

To further explore the role of DDR2 in Ang II treatment, AFs were infected with DDR2-overexpressing lentivirus (LV-DDR2). DDR2 overexpression significantly exacerbated the transformation of AFs into myofibroblasts after Ang II administration (Figure 3A,B). In agreement with the phenotypic alterations, the proliferation and migration of AFs were also strengthened by the up regulation of DDR2, after Ang II treatment (Figure 3C,F). Notably, these LV-DDR2-induced pro-fibrotic effects were effectively blocked by the DDR2-specific inhibitor WRG-28 (Figure S2). Meanwhile, the combination of Ang II treatment and DDR2 overexpression caused more severe damage to autophagy (Figure 3G,H), implying that DDR2 might mediate autophagy inhibition to promote Ang II-induced adventitial remodelling.

3.4 Rapamycin attenuates Ang II-induced autophagy inhibition, AFs phenotype switching, proliferation and migration

To delineate the inhibitory the effects of rapamycin on Ang II-mediated AFs activation, cells were pretreated with or without rapamycin (1 µM) for 1 h prior to 24 h stimulation with vehicle or Ang II (100 nM). As demonstrated in Figure 4A,B, rapamycin pretreatment remarkably alleviated Ang II-induced myofibroblast differentiation, evidenced by reduced α-SMA and collagen I expression. Following 24 h exposure, Ang II stimulation markedly increased EdU-positive cells compared to control group, while rapamycin cotreatment substantially suppressed this proliferative response (Figure 4D,E). Consistent with these findings, rapamycin attenuated Ang II-enhanced AFs migratory capacity, as quantified by transwell (Figure 4F) and scratch assays (Figure 4G). Since Rapamycin stimulates autophagy by inhibiting mTOR signalling, we explored the role of autophagy in Ang II-stimulated AFs. Western blotting revealed that Ang II downregulated the LC3B-II/I ratio and expression of ATG5 and ATG7, while enhancing mTOR phosphorylation (p-mTOR) and inhibiting phosphorylation of ULK1 at Ser757 (p-ULK1). These effects were attenuated by rapamycin co-treatment (Figure 4H–K). This indicates that Ang II does impair autophagy in AFs. Additionally, autophagic flux test further confirmed Ang II also inhibits autophagic flux in AFs, whereas rapamycin partially restored it (Figure 4L,M). Collectively, our results propose that rapamycin may play a key role in the inhibition of Ang II-induced AFs phenotype switching, proliferation and migration via autophagy enhancement.

3.5 Chloroquine (CQ) aggravates Ang II-induced autophagy inhibition and AFs phenotype switching

To further demonstrate the role of autophagy in Ang II-induced AFs, we checked the effect of CQ, an autophagy antagonist, on Ang II treated AFs. As anticipated, CQ addition to Ang II treated cells further increased the protein levels of α-SMA and Collagen I in AFs in contrast to Ang II alone treatment (Figure 5A,B). Consistent with the phenotypic transformation, the AFs treated with CQ and Ang II showed a higher level of proliferation compared to AFs treated with Ang II alone, as seen in the EdU assay (Figure 5C,D). Moreover, an increase in cell migration was seen in AFs treated by a combination of CQ and Ang II (Figure 5E,F). Furthermore, the upregulation of SQSTM1 (p62) and downregulation of LC3B-II/I proved that Ang II restrained autophagic activity in AFs (Figure 5G,H). Taken together, we can conclude that autophagy is certainly involved in regulation of Ang II-induced adventitial remodelling.

3.6 PI3K/AKT/mTOR signalling is required for DDR2-mediated autophagy inhibition

Although we confirmed that DDR2 contributes to Ang II-induced adventitial remodelling in association with autophagy, the underlying mechanisms to these phenomena are still needed to be explored. Previous studies have suggested a role of PI3K/Akt/ mTOR pathway in hypoxia-induced AFs proliferation.²³ Also, DDR2 has been implicated in PI3K pathway in some malignancies.²⁴ To delineate whether Ang II modulates DDR2 through AT1 or AT2 receptor signalling, we performed pharmacological antagonism experiments.²⁵ Pretreatment with the AT1 receptor blocker losartan completely abolished Ang II-induced DDR2 upregulation, whereas the AT2 receptor antagonist PD123319 showed no inhibitory effect (Figure S3). This confirmed AT1 receptor specificity in mediating Ang II-DDR2 signalling. We then performed a co-immunoprecipitation assay to further explore whether PI3K is a target of DDR2. We found that endogenous PI3K was co-immunoprecipitated with DDR2 in AF lysates (Figure 6A). Significantly, PI3K activity was elevated in response to Ang II stimulation, accompanied by enhanced phosphorylation of AKT, mTOR, and downstream effectors including p70S6K, 4E-BP1, and ULK1, thereby inhibiting autophagy. However, the activation of PI3K/AKT/mTOR signalling was neutralized by the inhibition of DDR2 (Figure 6B–G). These results indicated that DDR2 mediates Ang II-induced autophagy inhibition via PI3K/Akt/mTOR pathway.

3.7 PI3K/AKT signal inhibition attenuates Ang II-induced adventitial remodelling

LY294002, an inhibitor of the PI3K/AKT pathway, was used to further elucidate the importance of this pathway in the phenotypic alteration of AFs induced by Ang II. The elevated expression of α-SMA and Collagen I by Ang II, along with intensive proliferation and migration and PI3K/Akt/mTOR pathway, were markedly attenuated after LY294002 intervention (Figure 7). These results further validated our hypothesis.

3.8 DDR2-mediated autophagy inhibition promotes ROS accumulation in Ang II-stimulated AFs

In our recent study, we found that Ang II stimulation leads to massive ROS production partially contributing to the phenotypic switch of AFs. As autophagy is reported to be a tool for the balancing the amount of ROS in a cell, we investigated whether Ang II-induced AFs phenotypic switching through inhibition of autophagy is related to redox homeostasis. We thus examined the ROS level of AFs upon Ang II stimulation and other treatments using MitoSOX staining. Notably, rapamycin and LV-sh-DDR2 both attenuated Ang II-induced ROS elevation to a level comparable to the well-known ROS scavenger NAC. Conversely, CQ and LV-DDR2 further potentiated ROS elevation (Figure 8A). In addition, NAC significantly attenuated Ang II-induced AFs phenotypic switching, cell proliferation and migration (Figure 8B–G). This proved that there is a role of autophagy in maintaining redox homeostasis during Ang II-induced AFs activation.

3.9 Rapamycin partially attenuates Ang II-induced adventitial remodelling and the phenotypic transition of AFs in C57BL/6J mice

To further elucidate whether autophagy is involved in adventitial remodelling, we induced adventitial remodelling in a C57BL/6J mice model infused with Ang II for 4 weeks. The H&E, Masson and Sirius red staining were performed to assess the aorta remodelling. As shown in the Figure 9A–C, rapamycin significantly reduced Ang II infusion-induced adventitia thickening and fibrosis. Similarly, immunohistochemistry staining, and Western blot analysis also exhibited that rapamycin remarkably attenuated the Ang II-induced collagen I and α-SMA expression in mice, both of which are well-recognized marker of AFs phenotype switching (Figure 9D–F). We next examined the changes of autophagy in Ang II infusion-induced adventitia. Coinciding with our in vitro results, electron microscopy and Western blot analysis also showed that Ang II negatively regulates autophagy in adventitia, whereas rapamycin treatment can abolish this effect (Figure 9G–L). Notably, while rapamycin significantly attenuated Ang II-induced adventitial remodelling and fibrosis, our quantitative analysis revealed incomplete restoration to baseline levels (Figure 9B,C, F, J, L; Figure S4). This partial therapeutic efficacy suggested that mTOR-dependent autophagy modulation alone may be insufficient to fully counteract Ang II's pro-fibrotic signalling, potentially due to parallel activation of mTOR-independent pathways that sustain myofibroblast activation.²⁶

3.10 Deletion of Ddr2 alleviates adventitial remodelling and autophagy restriction in mice with Ang II infusion

To clarify the role of DDR2 in adventitial remodelling and autophagy in vivo, we generated Ddr2-cKO mice (Figure S5, Table 1) and exposed them to Ang II for 4 weeks. Deletion of Ddr2 obviously attenuated adventitia thickening and fibrosis in mice with Ang II infusion (Figure 10A–C). The Ang II-induced collagen I and α-SMA expressions in the mice adventitia were also sharply decreased, as confirmed by the immunohistochemistry staining and Western blot analysis (Figure 10D–F). Coinciding with our in vitro results, the expression of DDR2 was significantly increased after Ang II infusion in mice adventitia (Figure 10G,H). Electron microscopy and western blot analysis also demonstrated that the absence of DDR2 alleviated autophagy restriction in Ang II infusion-induced adventitia (Figure 10I–N). Of note, DDR2 deficiency had negligible effect on the body weight and blood pressure (Figure S1A,B).