Early Progression of Abdominal Aortic Aneurysm is Decelerated by Improved Endothelial Barrier Function via ALDH2‐LIN28B‐ELK3 Signaling

Abstract The involvement of endothelial barrier function in abdominal aortic aneurysm (AAA) and its upstream regulators remains unknown. Single‐cell RNA sequencing shows that disrupted endothelial focal junction is an early (3 days) and persistent (28 days) event during Angiotensin II (Ang II)‐induced AAA progression. Consistently, mRNA sequencing on human aortic dissection tissues confirmed downregulated expression of endothelial barrier‐related genes. Aldehyde dehydrogenase 2 (ALDH2), a negative regulator of AAA, is found to be upregulated in the intimal media of AAA samples, leading to testing its role in early‐stage AAA. ALDH2 knockdown/knockout specifically in endothelial cells (ECs) significantly increases expression of EC barrier markers related to focal adhesion and tight junction, restores endothelial barrier integrity, and suppresses early aortic dilation of AAA (7 and 14 days post‐Ang II). Mechanically, ELK3 acts as an ALDH2 downstream regulator for endothelial barrier function preservation. At the molecular level, ALDH2 directly binds to LIN28B, a regulator of ELK3 mRNA stability, hindering LIN28B binding to ELK3 mRNA, thereby depressing ELK3 expression and impairing endothelial barrier function. Therefore, preserving vascular endothelial barrier integrity via ALDH2‐specific knockdown in ECs holds therapeutic potential in the early management of AAAs.

Total RNA was extracted with Trizol and assessed with the kaiaoK5500®Spectrophotometer (Kaiao, Beijing, China) and the RNA Nano 6000 Assay Kit of the Bioanalyzer 2100 system (Agilent Technologies, CA, USA).
Total RNA samples that met the following requirements were used in subsequent experiments: RNA integrity number (RIN) > 7.0 and a 28S:18S ratio > 1.8.The sequencing quality was assessed with FastQC (Version 0.11.5) and then low-quality data were filtered using NGSQC (v0.4), a quality value of over 50% bases of the read was less than 5, and trimmed the adapter sequence, and removed the reads containing poly-N using inhouse Perl scripts.Sequence libraries were generated and sequenced by CapitalBio Technology (Beijing, China).RNA from 6 human AAD intimal tissues and 3 control intimal sample was extracted for RNA sequence study.The differential genes list is available in Excel S1, Supporting Information.

Echocardiography
After Ang II infusion, the maximal diameter of the abdominal aorta was measured with a high-resolution ultrasound imaging system (VisualSonics, Vevo3100) in a double-blind way.Aneurysm incidence was quantified based on a definition of an external width of the suprarenal aorta that was increased by 50% or greater compared with aortas from saline-infused mice.The ruptured aorta was excluded from the analysis of aortic diameter but included in the analysis of aneurysm incidence.Mice were anesthetized with 1.2-1.5 Vol % isoflurane.

Single-cell RNA sequencing
Regular digestion of mouse aorta samples.The methods of the aortic aneurysm mouse model are as described before.Aortas were harvested at day 0 without Ang II exposure, as well as 3 and 28 days after Ang II exposure.
Preparation of the single-cell suspension of the aortic cells was performed following a previously described enzymatic digestion protocol [1] .Briefly, the isolated whole aorta pooled from five mice in each group were cut and digested with an enzyme solution (450 U/mL collagenase type I, 125 U/mL collagenase type XI, 60 U/mL hyaluronidase type I-s, and 60 U/mL DNase I) for 1.5 hours at 37°C.The cell suspension was strained through a 30 μm filter and washed twice with PBS.The cells were resuspended in 0.4% BSA-PBS and the cell viability was greater than 80%, as confirmed by trypan blue staining.The resuspended cells were subjected to scRNA-seq.
Cell filtering.Dimensional reduction, clustering, and analysis of single-cell RNA sequencing data were performed using the R package Seurat.Cells were filtered out if more than 4,000 genes or less than 200 genes were detected, the percentage of mitochondrial gene counts were higher than 25% and hemoglobin genes counts were higher than 10%.
Single-cell clustering.Clustering was conducted using Seurat (v.4.1.1).To avoid batch effects among samples and experiments, we used Harmony to integrate the samples.We used the FindClusters function of Seurat on the basis of the shared nearest neighbor (SNN-cliq) clustering method with parameters (k = 20, dimuse = 1:20, res = 0.1).The Seurat functions DotPlot were used to visualize the gene expression with dot plot.

Definition of marker genes of cell groups.
Marker genes for each cluster were determined using the Wilcoxon rank-sum test via the FindAllMarkers function in Seurat.The FindAllMarkers function was used with the parameters min.pct=0.25,logfc.threshold=0.25,only.positive=TRUEand return.thresh=0.01.Based on the known marker genes, cell types were assigned to each cluster, including smooth muscle cells, fibroblasts, macrophages, ECs, T cells, B cells, granulocyte cells, proliferating cells and adipocytes.
Endothelial cluster gene set signatures.The FindAllMarkers Seurat function was utilized to find positive markers of EC population using the Wilcoxon rank-sum test.Pathway enrichment analysis was performed using enrichKEGG function to identify gene sets in clusterProfiler (V4.4.4).The differential genes list versus between day 28 and day 0 is available in Excel S2, Supporting Information.The differential genes list versus between day 3 and day 0 is available in Excel S3, Supporting Information.

Cell cultures
Different cell lines were chosen according to different purposes.
Primary Human Aortic Endothelial Cells (HAECs) and Primary Human Umbilical Vein Endothelial Cells (HUVECs) were purchased from ScienCell Research Laboratories, which were cultured in Endothelial Cell Medium (ScienCell, 1001).The HUVEC cell line were purchased from Pricella and maintained with RPMI 1640 medium (Sigma, 8758).Primary HAECs, primary HUVECs and the HUVEC cell line were used to determine endothelial barrier function.
HEK293T cells were purchased from the American Type Culture Collection (Manassas, VA, USA) and were cultured in high-glucose DMEM supplemented with 10% FBS, 100 units/mL penicillin and 100 μg/mL streptomycin.HEK293T cells were used in cell biology research in this research, because of their reliable growth and propensity for transfection.

siRNA-mediated gene knockdown
ELK3-siRNA, FOS-siRNA, TEAD1-siRNA, LIN28B-siRNA, and scramble-siRNA were purchased from GenePharma (Shanghai, China).Lipofectamine RNAiMAX reagents (Thermo Fisher Scientific,13778150) were used for transient transfection of siRNA duplexes into HUVECs cell line.Sequences of siRNAs are available in Table S2, Supporting Information.After 24 hours of transfection, cells were ready for further experiments.

Plasmid construction and transfection
The human ALDH2, ELK3 and LIN28B cDNA was cloned into pcDNA3.1 by Biosune biotechnology.
Desired DNA fragments of the human JAM-A、claudin5、VE-cadherin and ELK3 promoter were PCRamplified and cloned into the pGL3-Basic luciferase reporter vector (Promega, E1751).All constructs were confirmed by DNA sequencing.Lipofectamine 2000 reagents (Thermo Fisher Scientific, 11668030) were used for transfection of plasmids into HUVEC cell line or HEK293T.

Construction of adeno-associated virus
AAV serotype 1 (AAV1) vector serotypes (with an ICAM2 promoter) is the preferred viral vector to target aortic endothelial cell after injection.AAV1 shRNA ALDH2 sequence was cloned into the pAV-ICAM2-GFP-mir30-shRNA to get ALDH2-RNAi plasmid [2] .AAV1 shRNA ELK3 sequence was cloned into the pAV-ICAM2-GFP-mir30-shRNA to get an ELK3-RNAi plasmid.Sequences of shRNA are available in Table S2, Supporting Information.AAV1 ELK3 sequence was cloned into the pAV-ICAM2-P2A-GFP to get an ELK3 overexpression plasmid.Specific sequences and nonspecific controls were separately constructed in one vector, and AAV1 harboring these sequences were generated by Vigene Biosciences (Jinan, China).Mice were injected with one of the above viruses (5×10 11 pfu/kg) for 4 weeks before they were randomly grouped and infused with Ang II.

Real-time quantitative PCR (RT-qPCR)
Tissues or cells were mechanically homogenized and total RNA was extracted using Trizol Reagent (Sigma, T9424).Reverse transcription into cDNA was performed using the Reverse Transcription System kit (Vazyme, R223-01).The synthesized cDNA was amplified with a standard quantitative polymerase chain reaction (qPCR) protocol including the use of ChamQ Universal SYBR qPCR Master Mix (Vazyme, Q711-02).Amplification products were normalized against GAPDH mRNA, which was amplified in the same reaction as an internal control.Data were analyzed according to the 2 −∆∆Ct method.The fold change in mRNA expression was quantified relative to control.RT-qPCR primers are listed in Table S3, Supporting Information.

Western blot
Tissue or cell lysates with the same protein content (determined by the BCA method; Boster, AR0146) were prepared.Proteins were separated by 10% SDS-PAGE or 12% SDS-PAGE and transferred to polyvinylidene difluoride (PVDF) membranes.After being blocked for 1.5 hours in 5% non-fat milk, the bands were incubated overnight at 4°C with primary antibodies.After washing, secondary antibodies were incubated for 1.5 hours.The bands were scanned and detected by a chemiluminescence instrument (General Electric Company, AI600RGB).The relative intensity of immunoreactive bands was assessed by Image J software.

Immunoprecipitation (IP) and Co-immunoprecipitation (Co-IP) assays
For IP in HUVEC cell line, cells were washed thrice and scraped into ice-cold PBS.The pellets were resuspended with lysis buffer.The protein supernatant fluid was respectively incubated with 2 μg of antibody against LIN28B (Abcam, ab19188) and rabbit IgG separately at 4 °C overnight incubation and rotation.The immune complexes were incubated with 40 μL of protein A/G sepharose beads for 3 hours rotation and washed with the lysis buffer for 3 times.20 μL of 4×SDS loading buffer was added into each sample, which was then subjected to western blot analysis.High-resolution mass spectrometry-based proteomics was performed to determine proteins interact with LIN28B.One hundred and seventy-eight proteins were identified in the protein mixture enriched by LIN28B antibody and KEGG enrichment analysis was generated by Metascape (https://metascape.org).Annotation and enrichment results are listed in Excel S4, Supporting Information.
For Co-IP, cells were lysed with IP lysis buffer.The primary antibody was coupled with protein samples and incubated at 4°C overnight.Add Protein A/G beads to each sample and incubate the lysate beads mixture at 4°C under rotary agitation for 3 hours.After IP, the samples were washed with IP lysis buffer thrice.Proteins were eluted with 1x SDS sample buffer.
Thirdly, these 25 TFs were experimentally verified in aortic tissue and HUVEC cell line.

Chromatin immunoprecipitation (ChIP) assay
HUVEC cell line transfected with vector or ELK3 plasmid followed by serum starvation for 24 hours, were first incubated with 1% formaldehyde for 15 minutes to cross-link DNA protein complexes.Cells were then rinsed, harvested and lysed.Cell lysates were sonicated to produce genomic DNA fragments 300-1000 bp in length.The lysates were pre-cleaned with a mixture of salmon sperm DNA-protein A/G, followed by immunoprecipitation with an anti-ELK3 (Novus Biologicals, NBP1-83960) or normal IgG (Novus Biologicals, NBP2-24891) antibody overnight at 4°C.Then, single-strand salmon sperm DNA saturated with protein A/G plus-agarose beads (Santa Cruz Biotechnology, sc-2003) was added and incubated for 1 hours.
The immunocomplex beads were then washed sequentially with wash buffer and TE buffer and eluted with elution buffer (100 mM NaHCO3, 1% SDS).The eluted immunocomplexes were incubated with proteinase K buffer at 65°C to release the cross-linked DNA, which was further purified by phenol/chloroform/isoamyl alcohol and amplified by PCR to measure the ELK3-bound sequences.

RNA immunoprecipitation (RIP) assay
The RNA isolated from HUVEC cell line was incubated in immunoprecipitation buffer (150 nmol/L NaCl, 0.1% NP-40, 10 mmol/L Tris-HCl, pH 7.4) containing anti-LIN28B antibody (Abcam, ab191881) and IgG (Cell Signaling Technology, 2729) at 4 °C for 12 hours.Then, magnetic protein A/G beads was added and continuously incubated at 4 °C for 2 hours.After being centrifuged five times and washed with IP buffer, the LIN28B-bound RNA was enriched and subjected to RT-qPCR analysis.

Measurement of mRNA stability
To measure the stability of ELK3 mRNA, HUVEC cell line were pretreated with actinomycin D (5 μg/mL, MedChemExpress, HY-17559) for 30 minutes.Total RNA of cells was then collected at 0, 4, 8, 12, and 24 hours after actinomycin D treatment, followed by RT-qPCR analysis using primers for the ELK3 coding sequence.The primers for RT-qPCR are listed in Table S3, Supporting Information.

In vivo Miles permeability assay
On day 7, 14 or 28 post-AAA surgery, mice were injected with 100 μl of 1% Evans blue solution in 1× PBS through tail vein injection.One hour later, mice were perfused with saline in the left ventricle to remove intravascular Evans blue dye, and the heart and aorta tissues were excised.Then the surrounding adipose tissues were removed, and images of the whole aorta were captured using a digital camera.The aortic crosssections were mounted by Prolong Gold Antifade Reagent with DAPI and used for imaging.Due to the autofluorescence characteristic of Evans blue (excitation at 620 nm, emission at 680 nm), the amount of Evans blue that permeated into the aortic wall was visualized (red) using the fluorescent microscope as mentioned above.

Immunofluorescence and confocal microscopy
After being washed with PBS thrice, the cells coverslip or tissue section were fixed with 4% paraformaldehyde in PBS for 15 minutes at room temperature.Aspirate fixative, rinse three times in PBS for 5 minutes each.
Permeabilize samples with 0.1% Triton X-100 in PBS for 10 minutes.Then the cells coverslip or tissue section were blocked with 5% bovine serum albumin for 30 minutes and then incubated with the following primary antibodies at 4°C overnight: anti-VE-cadherin (Abcam, ab33168), anti-p120-catenin (Santa Cruz Biotechnology, sc-23873) and anti-claudin-5 (Santa Cruz Biotechnology, sc-374221).Subsequently, the cells coverslip or tissue section were incubated with the appropriate secondary antibodies at room temperature for one hour.Finally, the cells coverslip or tissue section were mounted with DAPI.Images of cells were taken under the confocal laser scanning fluorescence microscopy (LSM710; Carl Zeiss AG, Jena, Germany).

Histological analysis
The entire aorta was fixed with 10% formalin for 24 hours at room temperature before being embedded in paraffin for sectioning.Tissues were sectioned at 5 μm and underwent H&E staining, Elastic tissue fibers-Verhoeff's Van Gieson (EVG) staining, Masson staining or immunohistochemical staining.After heatmediated antigen retrieval, paraffin sections were blocked with 5% BSA and subjected to immunohistochemical staining.The sections were incubated overnight at 4°C with primary antibodies against ALDH2 (Abcam, ab133306) and ELK3 (Santa Cruz Biotechnology, sc-134401).After several washes with PBS, the sections were incubated with biotin-labeled goat anti-rabbit and goat anti-mouse secondary antibodies at room temperature for 1 hours.Labeling was visualized with a 0.05% diaminobenzidine solution, and images of cells were taken under the confocal laser scanning fluorescence microscopy (Carl Zeiss AG, LSM710).

Figure S1: Quality Control (QC) plots. (A)
The number above the figure is the Pearson correlation coefficient between the number of UMI and genes in three samples.The closer the correlation coefficient was to 1, the stronger the correlation was.(B) The distribution of the number of genes, number of counts, percentage of mitochondrial fraction, percentage of hemoglobin genes fraction and percentage of ribosome fraction in three samples were showed by violin plots.Cells were filtered out if more than 4,000 genes, less than 200 genes were detected, the percentage of mitochondrial gene counts higher than 25% and hemoglobin genes counts higher than 10%.(C) A list of top 20 downregulated pathways from analysis of gene ontology terms related to biological processes in endothelial cells on day 28 versus day 0. (D) A list of top 20 downregulated pathways from analysis of gene ontology terms related to biological processes in endothelial cells on day 3 versus day 0. (E-F) Representative immunoblots and quantification for vinculin, JAM-A, VE-cadherin, p120-catenin and claudin-5 in aortas from ApoE -/-mice at serial time points of Ang II infusion (n=3 per group).Data are presented as mean ± SEM.One-way ANOVA followed by Bonferroni post hoc test was applied in (F).A.U. indicates arbitrary unit.(J-K) Representative immunoblots and quantification for p-ELK3 and ELK3 in HUVECs transfected with negative control shRNA (scramble) or ALDH2 shRNA after 24 hours of Ang II (10 -6 M) treatment (n=3 per group).Data are presented as mean ± SEM.One-way ANOVA followed by Bonferroni post hoc test was applied in (E).Mann-Whitney U test was applied in (F).Unpaired two-tailed Student's t tests with Welch's correction were used in JAM-A of (G) and claudin-5, p120-catenin of (I).Unpaired two-tailed Student's t tests were applied in (H), the rest markers of (G) and (I).Two-way ANOVA followed by Bonferroni post hoc analyses were applied in (K).A.U. indicates arbitrary unit.

Figure S2 :
Figure S2: Construction of ALDH2 ECKO mouse and genotype identification.(A) Overview of EC-specific ALDH2 knockout (ALDH2 ECKO ) mouse construction.(B) Mouse genotype identification.PCR analysis of ALDH2 flox mice.The red arrow refers to loxp (Mutant) band.The green arrow refers to wild type (WT) band.(C) Mouse genotype identification.PCR analysis of Tek-CreERT2 mice.The red arrow refers to Cre band.

Figure S4 :
Figure S4: (A-E) Effect of ECs-specific ALDH2 knockout on Ang II-induced changes in plasma lipid profiles.(A) Triglyceride levels, (B) Total cholesterol levels, (C) High-density lipoprotein cholesterol (HDL-C) levels, (D) Low-density lipoprotein cholesterol (LDL-C) levels, and (E) Blood glucose levels.A through E,

Figure S6 :
Figure S6: ALDH2 knockout alleviates Ang II-induced endothelial barrier dysfunction in vivo.(A) Schematic protocol exhibiting infusion of Ang II or saline in C57BL/6J mice for 2 weeks.(B) Permeability of aortic intimal barrier (to Evans blue dye) and quantification the Evans blue dye in WT and ALDH2 -/-mice, and the autofluorescence of the dye in aortic cross-section images.WT+ Ang II, n=6; and ALDH2 -/-+ Ang II, n=6.Scale bar, 5mm.(C-D) Representative immunoblots and quantification for vinculin, JAM-A, VE-cadherin, p120-cadherin, claudin-5 and ALDH2 in the indicated groups (n=3 per group).(E) Representative images of claudin-5 (red) in WT and ALDH2 -/-mice 2 weeks post-Ang II infusion.Scale bar of low-magnification images, 100μm.Scale bar of zoom in images, 20μm.Data are presented as mean ± SEM.Unpaired two-tailed Student's t tests were used in (B).Two-way ANOVA followed by Bonferroni post hoc analysis was applied in (D).A.U. indicates arbitrary unit.

Figure S7 :
Figure S7: Optimal doses and time lengths of Ang II treatment in endothelial cells as in vitro model.(A) Representative immunoblots and quantifications for Paxillin, p120-catenin, JAM-A, vinculin, VE-cadherin, γ-catenin, β-catenin and α-catenin in HUVECs subjected to different concentrations gradient of Ang II treatment for 24 hours (n=3 per group).(B) Representative immunoblots and quantifications for JAM-A, VEcadherin and vinculin in HUVECs subjected to different concentration gradient of Ang II treatment for 12 hours (n=3 per group).(C) Representative immunoblots and quantifications for JAM-A, VE-cadherin and vinculin in HUVECs subjected to different concentration gradient of Ang II treatment for 48 hours (n=3 per group).Data are presented as mean ± SEM.Kruskal-Wallis test followed by Bonferroni post hoc test was applied in JAM-A of (A).One-way ANOVA followed by Bonferroni post hoc tests were applied in (B) (C) and the rest markers of (A).A.U. indicates arbitrary unit.

Figure S8 :
Figure S8: Expression of other ECs adhesion molecules in vitro.(A-B) Representative immunoblots and quantification for p-β-catenin, β-catenin, p-paxillin and paxillin in the indicated groups for 24 hours (n=3 per group).Data are presented as mean ± SEM.Two-way ANOVA followed by Bonferroni post hoc analysis was applied in (B).A.U. indicates arbitrary unit.

Figure S11 :
Figure S11: ELK3 preserved endothelial barrier function in the MDA-MB-231 cell line.(A) Gene set enrichment analysis (GSEA) was performed on the ELK3 knockdown regulated expression profile using the GSE83325 data set.(B) Top 20 enriched downregulated gene ontology (GO) terms related to cellular component in ELK3 knockdown target genes.(C) Heat map showing the genes related to endothelial barrier regulated by ELK3 knockdown.

Figure S12 :
Figure S12: (A) Vector construction of AAV1-shELK3.(B) Representative images of immunofluorescence staining of ELK3 in the aortic intima following tail vein injection of ICAM2 promoter-driven AAV1 carrying ELK3-shRNA or scramble-shRNA.Scale bar, 50μm and 12.5μm.(C) ELK3 expression in the aortic intima by western blot in 3 animal groups following tail vein injection of ICAM2 promoter-driven AAV1 carrying ELK3-shRNA or scramble-shRNA.(D) Representative ascending aortic (AsAo) and abdominal aorta (AbAo) ultrasound images in 3 animal groups at 4 weeks following Ang II infusion.Data are presented as mean ± SEM.One-way ANOVA followed by Bonferroni post hoc test was applied in (C).A.U. indicates arbitrary unit.

Figure S13 :
Figure S13: (A) Vector construction of AAV1-ELK3 overexpression plasmid.(B) Representative images of immunofluorescence staining of ELK3 in the aortic intima following tail vein injection of ICAM2 promoterdriven AAV1 carrying ELK3 or scramble.Scale bar, 50μm and 12.5μm.(C) ELK3 expression in the aortic intima by western blot in 4 animal groups following tail vein injection of ICAM2 promoter-driven AAV1 carrying ELK3-shRNA or scramble-shRNA.(D) Representative ascending aortic (AsAo) and abdominal aorta (AbAo) ultrasound images in 4 animal groups after 7 days Ang II infusion.(E) Representative H&E, EVG staining and Masson staining of mouse abdominal aorta after 7 days post-Ang II treatment.The black arrows refer to degraded elastic fibers.Low-magnification images show the entire vascular wall at the site of analysis.Scale bar of high-magnification images in E, 100μm and 25μm.(F) Grade of elastin degradation in the aortic wall.Data are presented as mean ± SEM.Two-way ANOVA followed by Bonferroni post hoc analyses were applied in (C) (F).A.U. indicates arbitrary unit.

Figure S14 :
Figure S14: (A) Immunofluorescence staining for ALDH2 in HUVECs receiving saline and Ang II treatment.Scale bars, 10μm.(B) Immunofluorescence staining for ALDH2 in HUVECs receiving vehicle and Daidzin treatment.Scale bars, 10μm.(C) A list of KEGG enrichment analysis pathways in HUVECs identified in the protein mixture enriched by LIN28B antibody.(D) Co-immunoprecipitation assay of LIN28B and EIF3A、 EIF4E、RHA、EIF3D in HUVECs cell line (n=3 per group).(E) RT-qPCR of LIN28B in HUVECs cell line treated with LIN28B siRNA (n=4 per group).(F) Representative immunoblots and quantification for LIN28B in HUVECs cell line treated with Ang II (n=3 per group).(G) Co-immunoprecipitation assay of ALDH2 and LIN28B in HUVECs cell line treated with vehicle and Daidzin (n=3 per group).Data are presented as mean ± SEM.One-way ANOVA followed by Bonferroni post hoc test was applied in (E).Unpaired two-tailed Student's t test was used in (F).A.U. indicates arbitrary unit.