Vascular tissue samples were collected intraoperatively from 23 patients undergoing routine isolated CABG procedures (RA, n = 10; SVG, n = 13). Vein harvesting was done by endoscopically technique, radial artery harvesting by open technique. One graft segment per patient, approximately 2 cm in length, were obtained and divided. One half was treated immediately with the EDI according to manufacturer's instructions, and the other half with heparinized lactated Ringer's solution (RL). The samples were not tested by pressurized injections. Samples were preserved in the respective solutions for 60 min at room temperature and transferred to the laboratory for further processing.

The study was approved by the Ethics Committee of the Medical University of Vienna (EC Number: 2061/2018). Written informed consent was obtained from all patients prior to inclusion in the study. All samples were rendered anonymous by using ID numbers for data analysis and laboratory protocols. The data underlying this article will be shared upon reasonable request to the corresponding author. The use of all human cells and tissues was approved by the institutional ethics review board and complies with the Declaration of Helsinki. Patients with recent (<1 year) or current administration of immunomodulatory agents (e.g., thalidomide, lenalidomide), chemotherapeutical or immunosuppressive agents (e.g., cortisone, cytokines), recent (<6 months) RA puncture for cardiac catheterization, concomitant immune disease, positive virology (HIV, hepatitis-A/-B/-C, and influenza), diabetes mellitus on insulin >6 months prior to CABG, were excluded from participation in the study to avoid potential interactions with the effect of the EDI and to reduce inhomogeneity among the study population.

In this study we used the EDI DuraGraft® (Somahlution, Jupiter, FL, United States). It is an ionically and pH-balanced physiological salt solution containing glutathione, L-ascorbic acid, and L-arginine and other protective ingredients. The components and proposed mechanism of action of DuraGraft® have been previously described (12). In brief, DuraGraft is an intraoperative treatment solution that protects the structural integrity and function of the vascular endothelium, and protects the conduit from the damaging effects of ischaemia during storage and reperfusion. Heparinized (2 IE Heparin/mL) lactated Ringer's solution (RL) (B. Braun, Austria) served as a control. After harvest, either the EDI or RL was used to carefully flush and store the RA or SVG prior to anastomosis.

We analyzed cellular-cytotoxic and cell-growth conditions in EC and vSMC culture treated with EDI or RL. Cell culture medium (EGM^TM-2 Endothelial Cell Growth Medium-2, Lonza, Slough, UK) was used as additional control. Cell isolation from specimens was performed according to previously described protocol (17). Viability was assessed using an EZ4U kit (Biomedica MP, Vienna, Austria). Cell cycle and annexin V-staining were performed as previously described (18).

RNA isolation, cDNA synthesis and RT-qPCR were performed as described previously (18). Samples were normalized to the geometric mean of reference genes. Oligonucleotide-primer sequences are listed in Supplemental Materials.

Cultured cells and tissue were analyzed by Western blot as described previously (18). Detection was performed with ECL Ultra Western blot HRP Substrate (Lumigen, Inc.) and quantified with ImageQuant TL 7.0 (GE Healthcare, Vienna, Austria).

Immunofluorescence (IF) staining has been described previously (18). DNA were counterstained with DAPI for cell nuclei visualization and equally cell analysis, as well internal fluorescence control (ThermoFisher Scientific, Vienna, Austria). Microscopy pictures were taken on a Confocal Laser Scanning Microscope 700 (Zeiss, Vienna, Austria). Images were analyzed with the CellProfiler™ cell image analysis software. For histochemical staining formalin-fixed tissue samples were processed routinely and sections stained by Elastica van Gieson for histopathologic examination. All antibodies used are listed in Supplemental Materials.

RA and SVG segments were mounted onto cryo-SEM stubs with 25% dextran and plunged into liquid nitrogen. Gold sputter coated samples were viewed with a Hitachi S4700 Field Emission SEM using an accelerating voltage of 2 kV and a working distance of 6.5 mm.

The activity of the HO-1 promotor regulatory region was measured as previously described (19). All experiments were performed at least three times for each plasmid, and the average relative luciferase activity is reported. The used sequences were: Wild- TGGAAGGCCTTCTTTCTAGAGAGGGAATT; Mutant- TGGAAGGCCTTCTTAGAAGAGAGGGAATT.

The hypoxic induced ROS increase was analyzed by redox analyser (Redox Sys, Aytu Bioscience, US) following manufacturer's protocol.

Baseline ROS/RNS levels of treated grafts were analyzed by OxiSelect^TM In vitro ROS/RNS Assay Kit according to the manufacturer's protocol (Cell Biolabs, Inc., San Diego, CA). A H₂O₂ standard curve was prepared for each plate and statistical analysis.

Statistical analysis was performed with GraphPad Prism Version 6 (GraphPad Software, CA, USA) and SPSS 15.0 software (SPSS Inc, Chicago, USA). Comparisons between groups were performed by two-tailed Student's t-test or analysis of variance (ANOVA). Comparisons between multiple groups were tested by Multi-Way ANOVA. Variables were evaluated by Spearman's rank correlation. P < 0.05 was considered statistically significant.

RA and SVG samples were obtained from patients undergoing CABG. Patient baseline characteristics are presented in Supplementary Table 1. Baseline levels of ROS/RNS of graft protein lysates assayed using OxySelect^TM method correlated significantly with presence of diabetes and with age in RA, and with age in SVG, and level of eNOS correlated significantly with age (Supplementary Tables 2, 3).

We assessed whether the EDI reduces endothelial cell injury and intimal swelling to examine its effect on endothelial integrity. Histological analysis showed partial endothelial denudation, and an irregular structure of sub-endothelial and intimal layer in control compared with EDI-treated RA grafts (Figure 1A). The adventitial layer showed incipient detachment from the medial layer in RL-treated segments. We measured level of endothelial marker CD31 to confirm endothelial detachment. Partial endothelial denudation appeared responsible for the significant loss of CD31 in control group, which was also associated with a significant reduction of ECs compared to EDI-treated RA grafts (Figures 1B,C). CD34 and von Willebrand factor (vWF) were significantly overexpressed in EDI-treated RA grafts (Figure 1D). Correspondingly, gene expression analysis demonstrated significantly retained high expression of vWF mRNA after EDI treatment (Figure 1E).

The specific morphological structure of subendothelial and intimal layer between EDI-treated and RL-treated RA grafts were visualized by CD31 IF staining and SEM (Figures 1F,G). After RL treatment the endothelial layer showed erratic cellular structure and beginning of detachment from intima and the intimal layer compared to EDI treatment, which showed a balanced compact structure (Figure 1G). Detailed SEM images showed an intact cell-to-cell connection between ECs after EDI treatment (Figure 1H).

We found similar cell protective effects of EDI on endothelial and subendothelial layer integrity in SVG (Supplementary Figures 1, 2).

In summary, the endothelial and sub-endothelial layer was preserved in both types of bypass grafts after EDI treatment, while RL-treated grafts showed damaged endothelial surface and beginning incongruence of intimal structure.

We analyzed the viability of ECs and vSMCs of bypass grafts. Expression of the cell proliferation marker Ki67 was significantly reduced in RL-treated RA (Figures 2A,B) compared to retained proliferation capacity of EDI-treated RA.

We observed reduced cell growth in cultivated vSMCs of RA grafts treated with RL when compared to EDI solution (Figure 2C). This reduced cell proliferation may be a manifestation of hypoxic DNA damage, cell senescence and cell death (20). Therefore, cell senescence and DNA damage were assessed by γH2AX staining. We found a significant increase of γH2AX-expression with RL treatment, together with an increase of apoptosis marker p53-expression (Figure 2D). In contrast, in EDI-treated graft segments suggested a protective effect against DNA damage and cell senescence. A maintained proliferation rate of EDI treated RA-vSMCs was confirmed by cell proliferation assay (Figure 2E) and metabolism assay. Both showed a significant protective effect of EDI treatment on RA-vSMCs and ECs isolated from RA grafts (Figure 2F). Analysis of the cell cycle by flow cytometry was in line with these results (Figure 2G). IF staining revealed a significant increase of γH2AX positive signals in EC at the endothelial layer of RL-treated RA (Figure 2H). The same beneficial results were observed for EDI-treated SVG (Supplementary Figure 3).

In summary, our analyses showed a reduction of proliferation and increased cell senescence in RL-treated RA grafts and SVG, which may be considered a response to intracellular stress by hypoxic conditions after harvesting.

We measured cell death induced by hypoxic damage in RA grafts. The phosphorylated p53 (p-p53) protein was used as a marker of apoptosis. While EDI-treated RA grafts showed normal levels of cell apoptosis signals, RL treatment leads to a significant increase of p-p53 levels (Figures 3A,B). SVG showed similar favorable effects after EDI treatment (Supplementary Figures 3, 4).

OxySelect^TM measurement of ROS/RNS indicated a significant increase of hypoxic induced ROS/RNS production in the control group, whereas EDI treatment had a protective inhibitory effect (Figures 3C,D). Further, while SVG and RA samples showed a reduction of ROS production, RA grafts presented a more pronounced outcome (Supplementary Figure 5).

Next, the non-activated eNOS and activated (phosphorylated) eNOS (p-eNOS) expression, which impairs endothelial function (21, 22), was analyzed in correlation with ROS/RNS levels. IF staining demonstrated a significantly higher expression of p-eNOS in the endothelial layer, but not in the subendothelial layer of the RL-treated RA grafts (Figure 3E), while eNOS protein showed significantly elevated levels in EDI-treated RA grafts in all layers (Figures 3F,G).

For endothelial specific protein expression, we analyzed the ratio of eNOS to the endothelial marker CD31. Our findings indicated activation during hypoxic stress in these sections. In accordance with hypoxic stress induced rapid intimal swelling, we measured VEGF, a cellular signal for activated vascular cell proliferation (23). VEGF expression was significantly higher in RL-treated when compared to EDI-treated RA sections (Figure 3H).

In addition, we analyzed eNOS and VEGF protein expression of EDI vs. RL treated SVG, and then compared it to EDI treatment in RA grafts. While p-eNOS was not detected, we observed high VEGF expression measured by IF in SVG (Supplementary Figure 6).

Taken together, RL-treated samples appeared to trigger a cascade of p-eNOS and VEGF expression after hypoxic stress conditions while hypoxic cell damage was inhibited by EDI in the RA graft sections. Our findings demonstrate improved graft preservation and therefore advantage of RA grafts over SVG.

To detect the p-eNOS involved pathway, we analyzed total-AKT and activated phosphorylated AKT (p-AKT-Ser⁴⁷³) with protein expression experiments. The results demonstrated that the role of ROS in the induction of eNOS is dependent on AKT/p-AKT regulation at the transcriptional level in RA grafts (Figures 4A,B). The results showed significant differences of total-AKT and p-AKT expression in EDI-treated grafts (Figure 4A). Total-AKT was significantly depleted and active p-AKT protein was significantly elevated in RL-treated RA grafts. The elevation in p-AKT in control RA grafts was particularly striking given the reduced levels of total-AKT. Comparison of SVG treated with EDI vs. RL showed significantly higher active p-AKT protein levels in the RL-treated group (Supplementary Figure 6C).

Caveolin-1 (CAV-1) is expressed in EC, macrophages and vSMC (24), and acts as a negative regulator of its activity and thus a regulator of NO production (25, 26). Hypoxia has an immediate effect on the endothelial barrier, therefore we focused on the endothelial layer. CAV-1 protein levels were elevated after EDI treatment measured by Western blot (Figure 4C). Moreover, we found retained CAV-1 expression in intact EC of EDI-treated RA (Figure 4D).

HO-1 is a protein whose expression is systemically induced in pathophysiological states associated with oxidative stress and vascular injury (27). As expected, significantly increased HO-1 protein levels were found after EDI treatment in RA grafts (Figures 4E–G, white arrow).

In summary, our findings indicate the protective effect of EDI treatment by decreased ROS levels and less AKT activation, which is in accordance with the eNOS/p-eNOS results. The activation of eNOS-counterpart CAV-1 and lower HO-1 levels indicate endothelial dysfunction in the RL-treated group, which was absent after EDI treatment, suggesting that the EDI may efficiently inhibit oxidative cell damage.

To test redox potential, we analyzed the static oxidation-reduction potential (sORP) which is proportional to the balance of reductants and oxidants. Low sORP values indicate normal range of oxidative stress, while relative increased values define a higher state of oxidative stress. The capacity of oxidation-reduction potential (cORP) is the measure of antioxidant reserves in the normal range, while lower cORP values describe that the sample is below normal antioxidant reserves (28). sORP and cORP measurements in EDI-treated RA grafts showed significantly higher protective potential against oxidative stress (Figure 5A). The same measurements were performed in SVG samples with a stronger benefit in EDI-treated RA grafts compared to EDI-treated SVG (Supplementary Figure 7).

Platelet-derived growth factor (PDGF) and transforming growth factor-β(TGFβ are important mediators in vascular remodeling, vasculogenesis, and angiogenesis (29). In our study, IF staining showed non-significant changes in PDGFα expression (Figures 5B,C). NO regulates TGFβ signaling in EC (30), and was significantly increased in RL-treated RA grafts (Figure 5D).

An additional important mediator modulating oxidative vascular injury is the induction of HO-1. HO-1 is known to protect against neointimal formation by promoting reendothelialization (31). Therefore, we analyzed mediator induced promotor activation of HO-1 DNA sequence during EDI treatment in our cell culture set by luciferase activity assay. We found a significant increase of promotor activated HO-1 expression in EDI-treated compared to RL-treated ECs and vSMC (Figure 5E). In summary, EDI showed a strong protective ability by activating HO-1 expression, while the RL group demonstrated activation of TGFβ due to oxidative stress and vascular injury.