A Nrf2-OSGIN1&2-HSP70 axis mediates cigarette smoke-induced endothelial detachment: implications for plaque erosion

Abstract Aims Endothelial erosion of plaques is responsible for ∼30% of acute coronary syndromes (ACS). Smoking is a risk factor for plaque erosion, which most frequently occurs on the upstream surface of plaques where the endothelium experiences elevated shear stress. We sought to recreate these conditions in vitro to identify potential pathological mechanisms that might be of relevance to plaque erosion. Methods and results Culturing human coronary artery endothelial cells (HCAECs) under elevated flow (shear stress of 7.5 Pa) and chronically exposing them to cigarette smoke extract (CSE) and tumour necrosis factor-alpha (TNFα) recapitulated a defect in HCAEC adhesion, which corresponded with augmented Nrf2-regulated gene expression. Pharmacological activation or adenoviral overexpression of Nrf2 triggered endothelial detachment, identifying Nrf2 as a mediator of endothelial detachment. Growth/Differentiation Factor-15 (GDF15) expression was elevated in this model, with protein expression elevated in the plasma of patients experiencing plaque erosion compared with plaque rupture. The expression of two Nrf2-regulated genes, OSGIN1 and OSGIN2, was increased by CSE and TNFα under elevated flow and was also elevated in the aortas of mice exposed to cigarette smoke in vivo. Knockdown of OSGIN1&2 inhibited Nrf2-induced cell detachment. Overexpression of OSGIN1&2 induced endothelial detachment and resulted in cell cycle arrest, induction of senescence, loss of focal adhesions and actin stress fibres, and disturbed proteostasis mediated in part by HSP70, restoration of which reduced HCAEC detachment. In ACS patients who smoked, blood concentrations of HSP70 were elevated in plaque erosion compared with plaque rupture. Conclusion We identified a novel Nrf2-OSGIN1&2-HSP70 axis that regulates endothelial adhesion, elevated GDF15 and HSP70 as biomarkers for plaque erosion in patients who smoke, and two therapeutic targets that offer the potential for reducing the risk of plaque erosion.

cassette was shuttled into pDC511 (Microbix Biosystems, Canada) for adenoviral vector production as previously described 2 .

RNA extraction and reverse transcriptase-polymerase chain reaction (RT-PCR)
Total RNA was extracted from HCAECs using the RNeasy Mini Kit (Norgen) according to the manufacturer's protocol. 250ng RNA were reverse transcribed into cDNA with random primer by reverse transcriptase (RT) (Qiagen). cDNA (relative 1ng RNA) was amplified by standard PCR with Taq DNA polymerase (Sensifast, SybrGreen, LOW-ROX Kit, Bioline) and primers. Primers were designed using NCBI software. For each gene, SYBR Green was used in place of a labelled probe (primers sequences Table S4). GAPDH RNA was used as the internal control for each gene and the target gene was amplified in duplex in PCR mixtures (10 μl final volume) containing 4 μl Sybr® Green PCR Master Mix, cDNA template 1 μl, optimised primers 2 μl and 3 μl of H2O. PCR thermal cycle parameters were: 5 minutes 95°C 30 seconds between at 65°C to 70°C depending on primers optimisation, 30 seconds at 72°C 40 cycles of 95°C for 15 seconds and 60°C for 1 minute. Reactions were performed, and fluorescence was monitored in an Applied Biosystem detector (Applied Biosystem). Relative mRNA expression level was defined as the ratio of target gene expression level to GAPDH mRNA expression. Primers sequences can be found in Table S4. Table S4. Primer sequences used for real-time PCR

RNASeq data analysis on HCAEC with OSGIN1&2 overexpression
Strand-specific RNA-seq libraries were prepared using the Illumina workflow with the TruSeq® Stranded mRNA Sample Preparation Kit. Paired-end reads of 65bp were generated from each sample on the Illumina platform of HiSeq4000. The fastq files generated were analysed with FastQC 5 , any low quality reads and contaminated barcodes were trimmed with Trimmomactic 6 . All libraries were aligned to the hg38 assembly of human genome using STAR-2.5.3a 7 and only the unique alignments were reported for each read. The mapped reads were also counted with STAR at gene level against gencode.v25.annotation.gtf. R was used for all the statistical analysis of data 8 . The counts data were normalized and donor effect removed using the R package RUVSeq 9 . Differentially expressed genes were detected with the R package of DESeq2 10 between groups of experimental data sets. Cluster analysis was carried out on the DE genes identified with DESeq2 using a padj cut off of 0.05 with gplots 11 .The predicted upstream regulators and altered canonical pathways were generated through the use of IPA (QIAGEN Inc., https://www.qiagenbioinformatics.com/products/ingenuity-pathway-analysis) 12 . The cluster analysis was carried out on the differentially expressed genes identified with DESeq2 using a p-adjusted cut off of 0.05, absolute log2 fold change cut off of 0.5, and base Mean cut off of 50. The Pearson distance was clustered with the hclust function and plotted with an R package of gplots.

Patient study population and blood samples
Patients presenting with ST-segment elevation myocardial infarction (STEMI) underwent primary percutaneous coronary intervention (PCI) at the Second Affiliated Hospital of Harbin Medical University were prospectively enrolled in this study. The diagnosis of STEMI was made according to the following criteria: detection of cardiac troponin values with at least one value above the upper reference, typical symptoms of acute myocardial ischemia and new ischemic electrocardiogram changes. Thrombus aspiration was performed in all patients before optical coherence tomography (OCT) examination. OCT imaging of culprit lesion were performed after antegrade flow restoration without any intervention. Patients were divided into plaque erosion and plaque rupture group according to OCT image of culprit lesion. The plaque rupture and plaque erosion were defined by OCT based on our previously established definition 13 . OCT-erosion was identified by the presence of attached thrombus overlying an intact and visualized plaque, while OCT-rupture was identified by disruption of fibrous cap and cavity formation in the plaque. The blood samples around the culprit lesion were collected from intracoronary aspirates during the primary PCI. The serum was separated and stored at -80℃ immediately. Written informed consent was obtained from all patients. This study was approved by the institutional research ethics committee of the Second Affiliated Hospital of Harbin Medical University and conforms to the principles outlined in the Declaration of Helsinki.
Baseline characteristics of ACS patients analysed in this study that experienced either plaque rupture (PR) or plaque erosion (PE) Immunofluorescence on Mouse Aortas 8µm frozen sections of aortas from mice exposed to air (control mice) or cigarette smoke for 3 months (as described 14 ) were fixed in ice cold acetone before immunofluorescence was performed using the antibodies described above. The mouse tissue used in this manuscript was not generated for this study, but was surplus to the described study 14 , and therefore utilised in this study in line with 3Rs principles. All experimental procedures were carried out in accordance with relevant guidelines and regulations and approved by the Ethical Committee of Animal Experiments of the KU Leuven. Figure S1. Top 50 genes upregulated and downregulated compared to elevated shear stress (ESS) control.  Figure S2: Cell Adhesion of HCAECs exposed to a combination of elevated shear stress, TNFα and CSE together (ESSTC), with additional Z-VAD-FMK treatment (20µM), vs untreated and DMSO controls at ESS for 72 hours. Cell number was quantified using picogreen assay, expressed as mean fold change against control ± S.E. n=3. Figure S3: Cell number of HCAECs exposed to a combination of TNFα and CSE together (T+C), with additional GM6001 treatment (10µM), vs untreated and DMSO controls at ESS for 72 hours. Cell number was quantified using picogreen assay, expressed as mean fold change against control ± S.E. n=3. Figure S4: Cell Adhesion of HCAECs exposed to a combination of elevated shear stress, TNFα and CSE together (ESSTC), with additional 3µM Rosuvastatin treatment, vs untreated and DMSO controls at ESS for 72 hours. Cell number was quantified using picogreen assay, expressed as mean fold change against ESSTC ± S.E. n=3. Figure S5: Cell Adhesion of HCAECs exposed to a combination of elevated shear stress, TNFα and CSE together (ESSTC), with additional 200µM Apocynin treatment, vs untreated and DMSO controls at ESS for 72 hours. Cell number was quantified using picogreen assay, expressed as mean fold change against control ± S.E. n=3. Figure S6: Cell Adhesion of HCAECs exposed to a combination of elevated shear stress, TNFα and CSE together (ESSTC), with additional Necrostatin-1 treatment (10µM), vs untreated and DMSO controls at ESS for 72 hours. Cell number was quantified using picogreen assay, expressed as mean fold change against control ± S.E. n=3. and H3K4me3 at the promoter of OSGIN1, together with the higher levels of H3K4me1 on OSGIN1 compared to OSGIN2. While POLII is bound to promoters of both genes, difference in H3K4 methylation suggests that OSGIN1 is actively transcribed and OSGIN2 is poised for transcription. We found an NRF2 binding site antioxidant response elements (ARE) under the NRF2 binding peak located at 130bp upstream of the transcription start site (TSS) of OSGIN1 (C) and predicted four AREs 4kb upstream of the OSGIN2 TSS. This supports our previously published findings 15 .      RSAD2  GBP1  SAMD9  GCH1  SAMD9L  HERC6  SERPINH1  IFI35  SLC15A3  IFI44L  SLC37A1  IFIH1  SP110  IFIT1  TAP1  IFIT2  TRIM21  IFIT3  TRIM26  IFIT5  TRIM69  IFITM1  UBC  IRF7  UBE2L6  ISG15  XAF1  LY6E  Supplementary Tables S7 Cluster 3, Genes and top 10 canonical pathways and regulators.  ATF4  CSF2RB  PCDH17  CTHRC1 PLCB2  CX3CL1  PLCD1  DDR2  PLCL1  DGKA  POSTN  DKK2  PSMB9  DOCK5  PTPRU  DOCK8  SAT1  DPP4  SELENBP1  ERAP1  SEMA7A  FAM129A SLITRK4  FRAS1  ST6GALNAC1  GATSL3  STXBP2  HSPA12B TACSTD2  IGFBP5  TFAP2A  IL18R1  TRPV2  ITGB3  TUSC3  KLF4 VCAM1 KRT19
A B Figure S20. HSP70 quantification in serum of ACS serum. A) No significant differences were observed between patients with OCT-defined plaque rupture (PR) and plaque erosion (PE). B) subgroup analysis of PR and PE in non-smokers (NS) and smokers (S), indicating a significant difference between PR and PE only in smokers. C) comparison between smokers and non-smokers between PR and PE groups didn't observe a significant interaction. D) There was no overall effect of smoking on HSP70 levels in ACS patients.  Figure S21: a) Overexpression of OSGINs was evaluated through western blotting. b) PARP cleavage antibody didn't show any cleave of the PARP protein confirming it was no apoptotic pathway related. c) Total protein (e) quantification was used to determine p62 and HSP70 accumulation a) b) c) d) Figure S22: Rescue experiment was carried on orbital shaker using Metformin and VER155008. Following orbital shaker experiment ECs were lysate and total protein (d) quantification was used to determine (a)p62, (b)HSP70 and (c)LAMP1(western blotting analysis).