2.1. JURY device for pressure- and retraction-controlled balloon denudation in rabbits
The JURY device for pressure- and retraction-controlled vessel wall injury was developed for the use of balloon tip catheters in rabbits. Various Fogarty embolectomy catheters (3F40, 3F80, 4F40 and 4F80; mpö pfm GesmbH, Austria) were tested and selected according to the average size of the adult rabbit aorta abdominalis. In vitro tests showed that the Fogarty 4F40 catheter exhibits the best compliance for mechanical pressure control in the experimental setting and that a threshold pressure of 1.2 bar is required to expand the balloon of the 4F40 catheter.
2.2. Animal experiments
All animal experiments were approved by the Austrian Federal Ministry of Education, Science and Research (BMWF-66.010/01 1 I-II/3b/2012 and 66.010_0070-V_3b_2018). For the study, 4- to 6-month old male rabbits (New Zealand White rabbits (NZW rabbits), Charles River (France), Schitkowitz (Austria) and Hyla/Wildtype crossbreds (cbHyla, Schitkowitz Austria)) were used. Prior to the studies, the rabbits were acclimatized for 2-4 weeks. Rabbits were set on special diets after or before surgery. The diets used included an unpurified complete chow diet (standard diet, SD) for full nutritional conditions and purified special diets either enriched with 1% cholesterol (HCD), deficient in vitamins and choline required for Hcy metabolization (no vitamin B12, 20% folate (2 mg/kg), 20% vitamin B6 (6 mg/kg) and 10% choline (0.128 mg/kg) (VCDD)) and combined VCD/HCD. Special purified diets were acquired from Sniff, Germany. Average daily diet intake was determined by weighing the remaining food after two days once per week for each animal. Weight was measured once a week. Six or eight weeks later, rabbits were sacrificed and the aortas dissected. Same sections were used for myography, MRI and other analyses. All specimens from myography and MRI were used for histology.
2.3. Vessel wall injury
Injury of the adult rabbit aorta abdominalis was performed by three techniques: the automated JURY-controlled injury, classical “blind” injury without pressure monitoring and manual pressure-adjusted injury (a modification of classical “blind” injury with monitoring and manual adjustment of the pressure). A detailed description of the JURY device and the surgery is provided in supplement 1 and 2.
- Classical “blind” injury: manual injury, during which the pressure is regulated manually by thumb press on a syringe; no display of pressure data for the operator; manual retraction with undefined speed. To compare pressure curve in classical “blind” injury with pressure curves in manual pressure-adjusted injury and JURY-controlled injury the catheter during classical “blind” injury was connected to a 1 ml Luer-Lock syringe and the JURY device with a pressure sensor in between. Only during the inflation of the balloon the operator was informed by the assistant when the target pressure was reached, otherwise the operator could only maintain the pressure according to his feeling (“blind”).
- Manual pressure-adjusted injury: a modification of the classical “blind” injury with monitoring and manual adjustment of the pressure by the operator based on the displayed pressure curve; manual retraction with undefined speed. The catheter was connected to a 1 ml Luer-Lock syringe and the JURY device with a pressure sensor in between. The balloon pressure was displayed on a computer monitor in real-time.
- JURY-controlled injury: fully automated JURY-controlled injury with pressure being displayed, recorded and adjusted automatically, JURY-controlled retraction with preset speed. For this technique the catheter is locked to a pressure sensor (Pendotech, USA) on a 1 ml Luer-Lock syringe that is firmly clamped to the sled of the JURY device (Fig. S2). The desired target pressure and the retraction speed are entered in the graphical user interface of the JURY controller software (Arduino, USA). After reaching the target pressure, which takes a few seconds, the retraction of the catheter is initiated. After reaching the desired end position, the balloon is deflated and the catheter is gently removed. Initiated by the operator, balloon inflation, retraction at constant pressure and constant speed as well as deflation are all performed automatically by the JURY device.
To compare the injury protocols described above 3 NZW rabbits were used. In the first animal, the operator performed the balloon injury “blind” without monitoring the pressure curve. In the second animal, the operator monitored the pressure curve during the denudation and manually adjusted the pressure to 1.8 bar. The catheter was retracted manually (undefined speed) in the first and second animal. In the third animal, the JURY device was pre-adjusted to maintain a pressure of 1.8 bar and a constant retraction speed of 5 mm/s. In all cases, pressure and, if applicable, motor step data were recorded by the JURY device in real-time. The comparison of the methods is shown in Fig. 1. Due to high variation in the target pressure, the classical “blind” injury protocol was not further used.
Subsequently, abdominal aortas were injured in 6 NZW rabbits at a target pressure of 1.2 bar (minimal pressure for inflating the balloon) and a single retraction: 3 rabbits were treated according to the manual pressure-adjusted injury protocol and 3 rabbits according to the automated JURY-controlled injury protocol (constant retraction speed of 5 mm/s). Aortas from two untreated rabbits were used as controls. The following day, rabbits were sacrificed and the aorta abdominalis dissected in the area of the denudation to assess the injury.
In the next step, atherosclerotic plaque formation was examined in rabbit groups of 4-6 (NZW rabbits and cbHyla). Aortas were injured by manual pressure-adjusted or JURY-controlled injury techniques. The injury conditions used were either 1.2 bar (mild injury: minimal pressure; 1x pull-back) or 1.8 bar (severe injury: higher pressure; 3x pull-backs). After the injury, all rabbits were fed HCD for 6 weeks.
NZW rabbits were used to compare the effects of the modified diets on atherosclerotic transformation of the aorta. The rabbits were divided into 4 groups of 8 animals each. All vessel wall injuries were performed automatically with one retraction by the JURY device at 1.8 bar and a 60% reduced retraction speed of 2 mm/s to intensify single-applied injury. In the HCD, VCDD and VCD/HCD groups, the animals were preconditioned before the surgery (VCDD was initiated two weeks and HCD one week before vessel wall injury). One day after the injury, the special diets were resumed and fed for additional 8 weeks.
2.4. Plasma Hcy, total cholesterol and triglyceride levels
Methanol (MeOH), water (MS-grade) and acetonitrile were purchased from Merck, USA. 1,4-dithiothreitol (DTT), trifluoroacetic anhydride and ammonium acetate were purchased from Sigma, USA. Formic acid was purchased from Roth, Germany. L-Hcy from Sigma was used as the standard. DL-Hcy-d4 from CDN isotopes, Canada was used as an internal standard. All standard stock solutions were prepared in H2O + 0.1% formic acid and stored at -80°C until use. Blood was taken from the ear vein or by heart puncture after sacrifice, collected into EDTA or Lithium Heparin Vacuette® tubes (Greiner, Germany) and centrifuged for 15 min at 4°C and 3,200 rpm. Blood plasma was immediately aliquoted and frozen at -80°C. In order to analyse the plasma Hcy levels, 100 μl of each plasma sample were mixed with 25 μl of 20 mmol/l DL-Hcy-d4 as internal standard and 10 μl of 0.5 mol/l DTT and incubated for 15 min at RT. Then 300 μl of precipitation reagent (acetonitrile + 0.1% formic acid + 0.05% trifluoroacetic acid) were added, the samples were centrifuged for 5 min at 13,000 rpm at room temperature (RT) and 125 μl of the supernatant were transferred into an autosampler vial. The solvent was then evaporated for about 30-45 min under a stream of nitrogen and the samples were reconstituted in 125 μl of H2O/MeOH 9:1 (v/v) + 0.1% formic acid. For the absolute quantification of Hcy in the plasma, a dilution series for a calibration curve in the range from 1.4138 to 113.1034 µmol/l Hcy was prepared, followed by the preparation of quality controls (QC) at 94.25 µmol/l, 11.31 µmol/l and 1.885 µmol/l Hcy. Plasma Hcy levels were analyzed using a 1290 Infinity UHPLC coupled to a 6470 Triple-Quadrupole mass spectrometer (Agilent, USA) using a BEH C18 column (3.0 mm x 150 mm; 1.7 μm) (Waters, USA) with 50°C column temperature, 5 μl injection volume and a constant flow rate of 200 μl/min under the control of Agilent MassHunter Workstation Data Acquisition software. H2O + 0.1% formic acid (A) and methanol + 0.1% formic acid (B) were used as solvents. 95% solvent A was held for 2 min, followed by a change to 100% solvent B over the next 2 min, which was held for an additional 3.5 min. Re-equilibration was carried out by changing to 95% solvent A within 5 seconds, followed by 3 min at 95% solvent A. Total running time was 11 min. Hcy and Hcy-d4 as internal standards were analyzed in the MRM mode. The dwell time for all transitions was 60 ms, the fragmentor voltage 80 and the cell acceleration voltage 4. The transitions m/z 136 to 90 for Hcy and m/z 140 to 94 for Hcy-d4, both with a collision energy of 15, were used as quantifiers. The transitions m/z 136 to 73 and 56 were the qualifiers for Hcy, m/z 140 to 77 and 59 for Hcy-d4, respectively, all with a collision energy of 19. Plasma total cholesterol and triglyceride levels were analysed using a cobas 8000 modular routine analyser (Roche, Austria).
2.5. Lipoprotein analysis
Plasma from rabbits of each dietary group was pooled and diluted with PBS if necessary. 200 μl of each plasma pool was subjected to fast protein liquid chromatography (FPLC) on a Pharmacia FPLC system (Pfizer Pharma, Germany) equipped with a Superose 6 (30x1cm) column (Amersham Biosciences, USA). The lipoproteins were eluted with 10 mM Tris-HCl, 1 mM EDTA, 0.9% NaCl and 0.02% NaN3 (pH 7.4) in fractions of 0.5 ml/min for 60 min. Total cholesterol and triglyceride concentrations were determined enzymatically. To enhance sensitivity, the reaction buffers were supplemented with sodium 3,5-dichloro-2-hydroxy-benzenesulfonate.
2.6. Biaxial extension tests
The biomechanical properties of the aortas were determined by biaxial strain tests of square-shaped aortic specimens with a side length of approx. 8 mm, the sides being aligned in the circumferential and axial directions of the aorta. The specimens were mounted to the carriages of the biaxial extension stage and immersed in a phosphate-buffered physiological solution (PBS) at 37°C. During testing, the specimens were subjected to various loading protocols in order to determine the physiological deformations of the tissue. Different stretches (1.1–1.6 in 0.1 increments) were applied consecutively, starting with the lowest stretch until the tissue failed. For each stretch level, four preconditioning cycles and one measuring cycle were performed with different loading ratios (1(circ.):1(axial), 1 : 0:75; 1 : 0:5; 0:75 : 1, 0:5 : 1) with an actuator speed of 3 mm/min. Each measuring cycle consists of a loading and an unloading path. It is crucial to use different load ratios between the circumferential and axial directions in order to cover the physiological deformations and to capture the direction-dependent material response of the aortas by means of biaxial extension tests. In addition, this so-called ‘true’ biaxial test approach results in a unique set of constitutive parameters for the specimen to be tested. The forces of load cells, positions of carriages and distances between special tissue markers (which act as gage markers) recorded by a videoextensometer were recorded continuously. Cauchy stress versus stretch ratio plots were generated from these data sets. The authors found that the results were very sensitive to the initial preloads. In order to obtain reproducible results, a prestress of about 0.1 kPa in terms of engineering stress (corresponds to 5 mN) was applied to each specimen.
2.7. Myography
Aortic rings were prepared from the abdominal aorta. The aortic segments were rapidly cleaned from the surrounding tissue, cut into 2 mm rings and transferred to the physiological buffer described below. The rings were mounted on pins connected to a micrometer and a force transducer. The experiments were carried out in myograph chambers (620M Multi Wire Myograph System; Danish MyoTechnology, Denmark) filled with a modified Krebs-Ringer bicarbonate buffer solution (KRS) consisting of 119 mM NaCl, 4.7 mM KCl, 1.2 mM KH2PO4, 1.2 mM MgCl2, 2.5 mM CaCl2, 25 mM NaHCO3, 0.03 mM EDTA-Na2 and 5.5 mM D-glucose. The rings were kept in an open bath at 37°C, while the KRS was continuously oxygenated with 95% O2 and 5% CO2 to keep the pH at 7.4. The maximal active tension was achieved by contracting the rings with the modified KRS with high K+ concentration (65 mM NaCl, 59 mM KCl, 1.2 mM MgSO4, 1.2 mM KH2PO4, 2.5 mM CaCl2, 25 mM NaHCO3, 0.03 mM EDTA-Na2 and 5.5 mM D-glucose). After reaching maximal isometric tension, the high K+ content was washed out and the rings were allowed to relax in a physiological buffer. The rings were then gradually contracted with increasing concentrations of norepinephrine (1 nM – 0.3 µM) to reach 80% of the maximum norepinephrine-induced contraction and then relaxed with increasing concentrations of acetylcholine chloride (1 nM – 0.3 µM) to assess endothelium-dependent relaxation. In order to examine endothelium-independent relaxation, the rings were washed and constricted with norepinephrine to 80% of the maximal constriction, and subsequently exposed to increasing concentrations of the nitric oxide donor sodium nitroprusside (0.1 nM – 30 nM). Relaxation values were expressed as a percentage of the norepinephrine-induced contraction.
2.8. MRI analysis
Evaluation of plaque area: For the assessment of plaque area a 5 cm piece of dissected aorta was washed with PBS (pH 7.4) and immersed in perfluoropolyether (Fomblin®, Solvay, Italy) in a 1 ml Luer-Lock syringe without air bubbles. Fomblin®, a contrast agent without MR signal, was added to minimize magnetic susceptibility effects at air-tissue boundaries (39). The specimens were scanned with a 15.2 Tesla small animal MRI (Biospec 152/11, Bruker, Germany). For segmentation of the aortic sections, a proton density-weighted, fat-saturated 2D fast spin echo sequence with the following parameters was used: TR/TE = 1500/2.9 ms, turbo factor = 1, slice thickness = 500 µm, in-plane resolution = 100 µm x 100 µm, imaging matrix = 200 x 200, the number of segments was between 59 and 120 slices (Fig. 2). A slightly different setting was used to assess plaque area in rabbits fed different diets (Fig. 3). For the assessment of plaque area in rabbits fed different diets aortic sections of 1 cm in length were taken from approximately the same position from each animal and immersed in Fomblin®. A 7 Tesla small-animal scanner (Biospec 70/20, Bruker, Germany) equipped with a cryogenic cooled transceiver was used. A proton density-weighted, fat-saturated 2D fast spin echo sequence with the following parameters: TR/TE = 1900/6.64 ms, turbo factor = 2, slice thickness = 500 µm, in-plane resolution = 60 µm x 60 µm, imaging matrix = 200 x 200, 40 slices per aorta. For both studies, the mean cross-sectional area of the vessel wall across the entire specimens was analyzed by semi-automated segmentation with a custom build tool created in Matlab (MathWorks Inc, USA).
Diffusion tensor MRI: Within the same imaging session, a fat-saturated segmented diffusion tensor imaging protocol was used with the key parameters: EPI readout with 4 segments, TR/TE = 2000/21.13 ms, slice thickness = 800 µm, in-plane resolution = 80 µm x 80 µm, imaging matrix = 128 x 128, 9 slices. Water diffusion was analyzed in 120 isotropically distributed directions at a b-value of 1000 s/mm². The average of 10 non-diffusion weighted scans was used as a reference for calibrating the diffusion tensor. Diffusion tensor images including fractional anisotropy maps of the aortic cross-sections were calculated with the software supplied with the MRI (Paravision 6.1, Bruker, Germany). After the scans, the syringes were emptied and refilled with PBS containing 4% paraformaldehyde (PFA) for preservation and later histological analysis.
2.9. Histology
For cryosectioning, the aortic rings were washed in PBS (pH 7.4), after MRI or myography, embedded in Tissue Tek O.C.T T COMPOUND (VWR International, Austria) and frozen at −20°C in a cryotom (Microm, Germany). Slices of 8 μm thickness were cut, transferred to SuperFrost® Plus glass slides (Thermo Scientific, Germany) and dried overnight.
For CD31 endothelial and RAM11 macrophage immunostaining, procedures were performed at RT using a Polink-2 Plus Mouse Detection System kit (GBI Labs, USA). The slides were rinsed twice with PBS (5 min each time) and blocked with sterile filtered PBS + 5% BSA for 30 min. The slides were then incubated with a mouse anti-human CD31 antibody (clone JC/70A, Abcam, UK) 1:20 dilution or an anti-rabbit RAM11 antibody (Dako/Agilent, USA) 1:50 dilution in PBS + 0.5% BSA, according to the Polink staining protocol. As substrate BCIP/NBT solution (GBI Labs, USA) was used as substrate and the slides were counterstained with filtered neutral red solution (1% acetic acetate in dH2O, 10 mg/ml neutral red dye (Fluka Analytical, Switzerland) for 5 min. Finally, the slides were dried, and covered with Aquatex (Merck, USA) and a cover slip.
Oil red O neutral lipid staining was performed at RT. The slides were each rinsed with PBS, distilled water and 2-propanol solution (60% + 40% dH2O) for 5 min. The slides were then stained in freshly prepared and filtered oil red O solution (60% 2-propanol, 40% dH2O, 10 mg/ml oil red O dye, Sigma-Aldrich, USA) for 15 min, briefly de-stained with the same 2-propanol solution and washed in dH2O. Finally, the slides were counterstained for nuclei with Haemalaun nach Mayer (Gatt-Koller, Austria) for 2 min, washed with dH2O, dried, and covered with Aquadex (Merck, USA) and a cover slip. After drying, the slides were visualised in transmitted light mode of an Olympus BX51 Basic Fluorescence Microscope (Germany) with a DP71 camera. All images were taken with an UPlan Apo Infinity Corrected 4× objective. Liver tissue was fixed in PBS pH 7.4 + 4% PFA, embedded in paraffin blocks, and 3 µm sections were cut using a microtome (Histocom, Micron HM440E, Switzerland). The sections were stained with hematoxylin and eosin according to standard protocols.
2.10. Sample preparation and multi-photon microscopy
Ring-like pieces of rabbit aortas were cut longitudinally under the stereomicroscope using racer blades and subsequently mounted between two coverslips so that the aortic intima was flat to the objective. Imaging was performed at the IMB-Graz Optical Imaging Resource using a picosecond laser (picoEmerald; APE, Germany) integrated in a Leica SP5 confocal microscope (Leica Microsystems, Germany). The laser delivers temporally and spatially overlapping pulse trains: a stable 1064 nm line, a tunable signal beam and an idler beam. For coherent anti-Stokes Raman scattering (CARS) imaging of densely packed neutral lipids in lipid droplets, the signal beam was tuned to 816.4 nm (2840 cm-1). The sample was illuminated simultaneously with the 1064 nm line, which led to the CARS signal of symmetrical CH2 stretching vibrations of esterified fatty acids in lipid droplets. For second harmonic generation (SHG) imaging of collagen, the signal beam was tuned to 880 nm. A two-channel, non-descanned detector (NDD) in epi-mode was used to detect CARS and SHG signals simultaneously. An SP 680 barrier filter (excitation light filter), an RSP 495 beamsplitter for two-channel separation of emitted light towards a 465/120 (‘SHG channel’) and a 650/210 (‘CARS channel’) bandpass filter was used. The wide range of the bandpass filters allows the additional detection of two-photon excited autofluorescence (TPAF) mainly of elastin in the ‘CARS channel’ and of elastin and cellular structures in the ‘SHG channel.’ The SHG signal from collagen and the CARS signal from lipid droplets were typically significantly stronger than the autofluorescence signal from elastin and cellular structures under the same imaging conditions. Optical sections were acquired using HCX IRAPO L 25× NA 0.95 water immersion and a sampling interval of 125 nm x 125 nm. Eight- to sixteen-times line averaging was applied to reduce image noise.
2.11. Electron microscopy
Aortic tissue was fixed in 2.5% (w/v) glutaraldehyde and 2% PFA (w/v) in 0.1 M cacodylate buffer, pH 7.4, for 2 h, and then post-fixed in 2% (w/v) osmium tetroxide for 2 h at room temperature (RT). After dehydration (in graded series of ethanol), tissues were infiltrated (ethanol and TAAB Embedding Resin, pure TAAB Embedding Resin) and placed in TAAB Embedding Resin (8 h), transferred into embedding moulds, and polymerised (48 h, 60°C). Ultrathin sections (70 nm) were cut with a UC 7 Ultramicrotome (Leica Microsystems, Austria) and stained with lead citrate for 5 min and platinum blue for 15 min. Electron micrographs were taken using a Tecnai G2 transmission electron microscope (FEI, Netherlands) with a Gatan Ultrascan 1000 charge coupled device (CCD) camera (-20°C, acquisition software Digital Micrograph, Gatan, Germany and Serial EM). Acceleration voltage was 120 kV.
2.12. Statistical analysis
Data are presented as line graphs with error bars showing means and 95% Confidence Interval or as box plots partially with single data points superimposed (e.g. vessel wall area), respectively. Indicated sample sizes in figure legends refer to biological replicates (independent animals). Comparisons between two groups were done by Mann Whitney U test, if data violating normality distribution. In case of multiple group comparisons, analysis of variance (ANOVA) or Kruskal-Wallis test followed by Bonferroni post hoc tests were applied. In case of serial measurements (concentration), Greenhouse-Geisser–corrected two-way repeated-measures ANOVA was used. Generally, significant factorial designs were followed by pairwise comparisons that were corrected in case of multiple comparisons by Bonferroni post hoc analyses. Data residual distribution was confirmed by Shapiro-Wilk’s test and visual data inspection by use of Q-Q plots, whereas homogeneity of variance was verified by Levene’s test. All reported P values are two-sided, and an a level of 0.05 was used throughout. Analyses were performed with SPSS 27.0 (SPSS Inc., Chicago, USA) or GraphPad Prism 8 (GraphPad Software LLC, Massachusetts, USA). The statistical analysis is shown in S7.