Mice

All mouse experiments followed procedures outlined by the Council Directive of the European Union (86/609/EEC), and prior to experimental work, approval was granted by Semmelweis University’s Institutional Animal Care and Use Committee (PE/EA/1363-8/2019). Ten week old male, wild type (WT), PCSK9−/− and LDLR−/− mice were used with groups (n = 6, all were C57BL/6J background). Mice were supplied by the Jackson Laboratory (JAX catalogue numbers: 000664, 005993 and 002207, Bar Harbor, USA). Mice were maintained at the animal facility of the Institute of Genetics, Cell- and Immunobiology, with a normal light cycle (12 h light– 12 h dark) and were given free access to food and water. Blood was taken from the retrobulbar venous plexus at experimental start point (week 10) while animals continued normal feeding. At week 11, mice were starved for 5 h after which retrobulbar venous plexus blood was collected and mice were fed high fat diet (HFD) consisting of 45 kcal% fat (catalogue # D12451, Research Diets, New Brunswick, USA) from this point forward for 12 weeks.

At the midpoint of HFD (week + 6), mice were starved for 5 h after which retrobulbar venous plexus blood was collected. At 23 weeks, mice were anesthetised using isoflurane (1 mL/100 g body mass) and sacrificed through terminal bleeding via the inferior vena cava. Mice were weighed on a weekly basis from age 10 weeks through week 23.

For long-term aging experiments, male, WT, PCSK9−/− and LDLR−/− mice were used with group sizes n = 6 (all were C57BL/6J background). All mice were kept on normal chow diet for 22-months. At 22-months, surviving mice were anesthetised using isoflurane (1 mL/100 g body mass) and sacrificed through terminal bleeding via the inferior vena cava. Mice were weighed at termination.

Genotyping of mice

Genotyping confirmation was performed every 12–18 months or when a new breeding cage was set up, to ensure homozygous breeding cages and offspring. Typically, 2 females and 1 male from each knock-out (KO) strain was genotyped simultaneously, plus 1 WT animal as a negative control. Fifty bp DNA ladder (catalogue # SM0373, Thermo Scientific, Waltham, USA) was used. PCR primers and PCR protocols of the Jackson Laboratory were followed. For PCSK9−/− genotyping protocol 27,858: Standard PCR Assay - Pcsk9 < tm1Jdh > Version 4.2 was followed and for LDLR−/− genotyping protocol 27,075: Standard PCR Assay - Ldlr < tm1Her>-alternate 1 Version 1.2 was followed.

Human samples

All human platelet free plasma (PFP) samples were prepared from venous blood samples taken from patients admitted to the Semmelweis Városmajor Clinic. The patients were emergency patients with chest pain diagnosed not to have an acute cardiac ischemic event. Approval was granted by Semmelweis University’s Regional and Institutional Committee of Science and Research Ethics (67/2022). Blood samples from male patients were used for measurement, with patient data displayed in Supplementary Table 1. Group A were patients with normal cholesterol levels and Group B were patients with hypercholesterolaemia (n = 6). TC, LDL-C, high-density lipoprotein cholesterol (HDL-C), triglycerides, Glucose, and HbA1C were all measured using the DxC 700 AU (Beckmann Coulter, Brea, USA).

Mouse blood collection, handling, and storage

Blood was collected from the retrobulbar venous plexus using Corning® Pasteur pipettes L 5 3/4 in. (146 mm), standard tip (catalogue # CLS7095B5X-1000EA, Merck Sigma, Darmstadt, Germany) on all occasions except when mice were sacrificed. When sacrificing the mice, blood was taken from the inferior vena cava using a syringe and needle for careful extraction. All blood samples were gently mixed with 15% ACD-A anti-coagulant and processed within two hours from collection. Blood was centrifuged 2,500 g for 15 min at 4 °C, platelet poor plasma (PPP) was removed, 45 µL was taken and left at 4 °C for cholesterol measurements. For EV flow cytometry measurements, the rest of the PPP was re-spun 2,500 g for 15 min at 4 °C– to produce PFP, which was aliquoted into 1.5 mL Eppendorf tubes, diluted 1:1 in 0.1 μm filtered 0.9% NaCl (saline) solution, flash frozen in liquid nitrogen and stored at − 80 °C.

Assessment of cardiovascular function

All cardiovascular parameters were measured before termination of animals on the final day of experimental timeline, as described earlier [1]. A Vevo 3100 high-resolution in vivo imaging system (Fujifilm VisualSonics, Toronto, Canada) was used with an ultrahigh frequency transducer (MX400, 30 MHz, 55 fps) and evaluation was performed using VevoLAB software (version 5.6.1, Fujifilm VisualSonics, Toronto, Canada). Animals were anesthetized with isoflurane (for induction: 5 V/V% in O2; for maintenance: 2 V/V% in O2), with spontaneous breathing. Throughout the echocardiographic measurements, body temperature was kept at 37 ± 0.5°C, and electrocardiographic activity was obtained continuously. Chest hair was removed, and two-dimensional cines were obtained from the long-, short-, and apical four-chamber views of the heart (LAX, SAX, and APIC4, respectively). Left ventricular end-systolic and end-diastolic volumes (LVESV and LVEDV, respectively, derived from the rotational volumes of the left ventricular trace at diastole and systole, around the long axis line of the spline), left ventricular stroke volume (LVSV), left ventricular ejection fraction (LVEF) and cardiac output (CO) measurements were obtained from the LAX view. Left ventricular diameters in end-systole and end-diastole (LVESD, LVEDD, respectively), left ventricular fractional shortening (FS), left ventricular mass (LV mass), and left ventricular anterior or posterior wall thicknesses in systole and diastole (LVAWTs, LVAWTd, LVPWTs, and LVPWTd, respectively) were acquired from the SAX view. Relative wall thickness (RWT) was calculated as 2*LVPWTd/LVEDD; left ventricular remodelling index (LVRi) was calculated as LV mass/LVEDD. Iso-volumetric relaxation time (IVRT), and E/e’ ratio (where E: peak transmitral blood flow velocity in early diastole measured by pulsed wave Doppler imaging; and where e’: mitral annular velocity in early diastole assessed by tissue Doppler imaging) were derived from the APIC4 view, indicative for diastolic function.

Quantification of aortic atherosclerotic plaques

The aortic arches were isolated after termination of the mice using a stereomicroscope. Plaques were stained and imaged following an earlier protocol [2]. Here, aortic arches were stained using Oil-Red-O (Catalogue # 31170, Serva, Heidelberg, Germany, ) with a 70% ethanol wash prior to staining and an 80% ethanol wash post stain. Stained specimens were stored in PBS until imaged. Images were taken using the S9D microscope with objective lens 10450528 using the attached FLEXACAM C1 camera (Leica, Wetzlar, Germany).

ImageJ software was used to analyse the Oil-Red-O+ area/whole tissue ratio (%) of the plaques. Wand tool was used to select and measure the total isolated whole tissue area. The following steps were used to determine the plaques present on each isolated aortic arch:

1.

The “Image” tab was selected and from there “colour” and then “split channels” was used.

2.

The “process” tab was used to get the “image calculator” where red was subtracted from green.

3.

In the “image” tab, “adjust” and then “threshold” was selected (for Oil-Red-O+ particles) (each determined by eyes of two independent researchers).

4.

Here, watershed of the image (of Oil-Red-O+ particles) was performed by selecting the process” tab and then “binary” and finally watershed.

5.

After which, the total particle area was obtained using the “analyse” tab and selecting “measure particles”, ensuring any size or shaped particle was measured (0-infinity, 0.0–1.0).

All images and groups were initially randomised and blinded by a first researcher before independent analysis by two separate researchers.

Quantification of mouse blood cholesterol levels

Measurements were performed using the HDL and LDL/VLDL Quantitation Kit from (catalogue # MAK045-1KT, Merck Sigma, Darmstadt, Germany) following the manufacturers protocol. Per manufacturers protocol, PPP samples were used to measure cholesterol parameters. All measurements were completed within 7 days of sample collection from fresh PPP maintained at 4 °C.

Flow cytometry analysis of circulating EVs

PFP samples thawn from– 80 °C were analysed using the CytoFLEX S N2-V3-B5-R3 Flow Cytometer from (product # B78557, Beckmann Coulter, Brea, USA).

Measurements were carried out using the “slow” flow rate (10 µL/min) for 2 min and events were obtained from a previously optimized gate for EVs detection that was set up following a protocol outlined in [3, 4] using Flow Cytometry Sub-micron Particle Size Reference Kit (catalogue # F13839, Invitrogen, Carlsbad, USA) and shown in Supplementary Fig. 1. Antibodies were diluted 10-fold in 0.1 μm filtered Annexin Binding Buffer (ABB: 10 mmol HEPES, 140 mmol NaCl; 0.25 mmol CaCl2; pH: 7.4–7.5) and centrifuged at 12,600 g for 10 min at 4 °C, the supernatants were carefully transferred (ensuring that aggregated antibody pellets are not touched) to new labelled tubes.

For mice, 1 µL PFP and for humans, 4 µL PFP was used. These volumes were selected after preliminary optimization tests were performed with concentration curves ensuring that particle numbers did not saturate the detectors. With the selected concentration, sufficient particle numbers were present without causing a swarm effect. Each sample was placed into an Eppendorf tube along with antibody mixes shown in Supplementary Table 2. Tubes were protected from light and placed at 4 °C for 30 min. After the initial PFP antibody incubation, 30 µL count check beads from Sysmex (Order no: 05-4010 Lot: KW190620, 37920 particles/mL (± 10%)) were added and further 0.1 μm filtered ABB was added to reach a final volume of 300 µL.

Flow cytometry data were analysed using CytExpert2.1 (Beckman Coulter, Brea, USA) and exported to office365 Excel (Microsoft, Redmond, USA) for processing before analysis, using the following formula:

$$ \begin & } \\ & \quad = \left( } - } }}}}}} \right) \\ &\qquad \times}\times\left( }~}}}}}} \right) \end$$

Bead-based analysis of human PFP-derived lEVs

The MACSPlex EV Kit IO, human (catalogue # 130-122209, Miltenyi Biotec, Bergisch Gladbach, Germany) was used to measure PFP from three patients from group A and from three patients from group B (for human samples). The MACSPlex EV Kit IO, mouse (catalogue # 130-122-211, Miltenyi Biotec, Bergisch Gladbach, Germany) was used for mice samples, PFP samples were pooled from 6 animals from each group (WT, PCSK9, and LDLR) after HFD. PFP was thawn from − 80 °C, and 1 mL from each human sample and 0.5 mL from pooled mouse samples were centrifuged for 40 min at 20,000 g at 4 °C. Supernatant was discarded and pellets were each re-suspended in 100 µL PBS, and re-centrifuged (40 min at 20,000 g at 4 °C). Supernatants were discarded, and pellets were each re-suspended in 100 µL PBS.

The total protein concentration of the pellet was determined by Micro BCA assay (catalogue # A55864 ThermoFisher, Waltham, Massachusetts, USA) and 10 µg total protein was used in each MACSPlex reaction. The “overnight” protocol was followed as specified in the manufacturer’s manual. As a negative control, a sample containing only MACSPlex Reaction Buffer, Detection antibody cocktail, and Capture Beads was used as a negative control. Measurements were performed using the CytoFLEX S N2-V3-B5-R3 Flow Cytometer (product # B78557, Beckmann Coulter, Brea, California, USA). Measurement was taken on medium speed (30 µL/minute). Gating strategy was made based on negative control sample and can be seen in Supplementary Fig. 2.

Each sample was analysed by subtracting the negative control signal from each marker with any negative values zeroed. For each sample, the isotype control with the highest signal was then subtracted from each marker. Averages were then calculated from each marker. All markers were then normalised to the average of standard EV markers (CD9, CD63, and CD81) and graphed in GraphPad Prism 8.2.0.

Statistical analysis

Statistical analysis was performed using GraphPad Prism version 8.2.0 (GraphPad Software, Inc.). Values are presented as mean ± standard deviation with individual point plots. Shapiro-Wilk normality test was used to assess normality of all groups. For groups with normal distribution, unpaired student t test was performed and for groups that were non-normal, the non-parametric Mann-Whitney U test was applied. For comparisons of data sets within one group, a paired t test or a non-parametric Wilcoxon signed-rank test was selected depending on the normality of the samples.