Cellular stress induces erythrocyte assembly on intravascular von Willebrand factor strings and promotes microangiopathy

Microangiopathy with subsequent organ damage represents a major complication in several diseases. The mechanisms leading to microvascular occlusion include von Willebrand factor (VWF), notably the formation of ultra-large von Willebrand factor fibers (ULVWFs) and platelet aggregation. To date, the contribution of erythrocytes to vascular occlusion is incompletely clarified. We investigated the platelet-independent interaction between stressed erythrocytes and ULVWFs and its consequences for microcirculation and organ function under dynamic conditions. In response to shear stress, erythrocytes interacted strongly with VWF to initiate the formation of ULVWF/erythrocyte aggregates via the binding of Annexin V to the VWF A1 domain. VWF-erythrocyte adhesion was attenuated by heparin and the VWF-specific protease ADAMTS13. In an in vivo model of renal ischemia/reperfusion injury, erythrocytes adhered to capillaries of wild-type but not VWF-deficient mice and later resulted in less renal damage. In vivo imaging in mice confirmed the adhesion of stressed erythrocytes to the vessel wall. Moreover, enhanced eryptosis rates and increased VWF binding were detected in blood samples from patients with chronic renal failure. Our study demonstrates that stressed erythrocytes have a pronounced binding affinity to ULVWFs. The discovered mechanisms suggest that erythrocytes are essential for the pathogenesis of microangiopathies and renal damage by actively binding to ULVWFs.


Microfluidic adhesion experiments.
Microfluidic experiments were performed using the BioFlux 200 air pressure-driven microfluidic system (Fluxion Biosciences, South San Francisco, CA). Channels of a 48well BioFlux plate were biofunctionalized for 3 hours at 37°C with either 2 % BSA serving as negative control or plasmatic VWF (100 µg/ml) and rinsed with 0.5% BSA. Channels were perfused for 10 min with different erythrocyte or artificial vesicle preparations at 2 dyne/cm² and 10 dyne/cm², respectively. Reflection interference contrast microscopy (RICM) was used to visualize the interactions of the erythrocytes with the coated channel surfaces.

Reflection Interference Contrast Microscopy (RICM). For RICM visualization, an inverted Microscope
Observer Z1 (Zeiss, Jena, Germany) with a Colibri LED illumination system was used at two excitation wavelengths, 470 nm and 555 nm. Images were captured with a monochromatic camera (Axiocam MRm, Zeiss, Jena, Germany). The acquisition time was 200 ms at 350 ms intervals.
Endothelial cell preparation. Human Umbilical Vein Endothelial Cells (HUVEC) were isolated and cultivated as described previously 4,5 . HUVEC were maintained at 37°C and 5% CO2 and seeded into gelatine-coated microchannels after the second passage in accordance with the protocols recommended by Fluxion Biosciences. HUVEC were cultured inside the flow channels using Endothelial Cell Growth Medium 2 (EGM-2, Lonza, Cologne, Germany) until confluence was achieved. Secretion of VWF was stimulated using 100 nM phorbol myristate acetate (PMA; Sigma-Aldrich, Taufkirchen, Germany) in EGM-2 for 10 min before starting the flow experiments. Human Dermal Microvascular Endothelial Cells (HDMEC) were obtained from PromoCell (Heidelberg, Germany) and cultivated in IBIDI flow channels (IBIDI, Martinsried, Germany) using EGM-2 as described previously 5 . After reaching confluency 50 nM histamine was used to stimulate WPB exocytosis. Perfusion time was 10 min at 2 dyne/cm² before fixation with 4% PFA at 4°C.

Immunostaining flow experiments.
To quantify the adherent erythrocytes in the VWF-coated channels, FITC-labeled Annexin-V-Fluos (Roche Diagnostics, Mannheim, Germany) was used during sample preparation. Immunostaining of the adherent erythrocytes on HUVECs and stimulated HDMECs was carried out as follows: fixation was performed with 4% paraformaldehyde (PFA) for 30 min at 4°C, followed by washing with 0.5% BSA and blocking with 2% BSA. The primary and secondary antibodies (1:200 in 0.5% BSA) were incubated for 45 min at room temperature. All steps were carried out at 2 dyne/cm² for 5 min in the flow direction used in the initial experiment. Images were captured immediately after a final rinsing step. The following antibodies were used: FITC-polyclonal sheep anti human VWF (GeneTex Inc., Irvine, CA, USA), Monoclonal mouse anti human Band 3 (Sigma-Aldrich, Taufkirchen, Germany), Texas red goat anti mouse (Invitrogen, Eugene, Oregon, USA).

CRF Patients.
Five patients with CRF, mild anemia and erythropoietin substitution in a saturated state were included in the study. Informed consent was obtained from all subjects before inclusion. The study was conducted according to the ethical guidelines of our institution and the Helsinki Declaration. A 7.5 ml sample of heparin blood was taken from the CRF patients (n=5) and healthy donors (n=5) and prepared as described above.
Mouse experiments. All experimental protocols were conducted in accordance with the German law on animal protection and were approved by the Regierungspräsidium in Tübingen. Homozygous VWF -/mice 7 (B6.129S2-Vwf tm1Wgr /J) (Jackson Laboratory/Charles River, Sulzfeld, Germany) were compared with C57Bl6 control mice. Renal ischemia/reperfusion procedures were performed using 4-to 5-month old male VWF -/and control mice as described 8 . Briefly, animals were anesthetized using a mixture of midazolam (5 mg/kg b.w.), medetomidine (0.5 mg/kg b.w.) and fentanyl (0.05 mg/kg b.w.) and placed on a controlled temperature heating pad (37°C). The right kidney was excised through an incision in the right flank. The left kidney was dissected from the surrounding tissue and placed in a lucite cup before being covered with a moist swab. The left renal artery was occluded by an 8-0 (Ethicon) suture in a hanging weight system. Following 45 minutes of ischemia, the kidney was reperfused by removal of the hanging weight and the filament. Reperfusion was confirmed visually. The incision was closed, and the animals were treated with buprenorphine for analgesia. After 24 h, the animals were sacrificed by being re-anaesthetized and perfused at a constant flow rate of 5 ml/min of PBS, followed by fixation in 4% PFA.
Morphological evaluation of renal histology. For histological analyses, the kidneys were perfused and fixed with 4 % PFA and then dehydrated in ethanol and xylol before being embedded in paraffin. Paraffin embedded tissues were cut into sections of 2 µm thickness and stained with periodic acid Schiff (PAS) and Sirius red stains. Renal morphology was investigated by light microscopy as described below, with the investigator blinded to the genotype of the mice. Tubulointerstitial damage, i.e., tubular necrosis, tubular atrophy and tubular dilation, was assessed in PAS-stained paraffin sections at a magnification of 200x using a semiquantitative scoring system (acute tubular necrosis [ATN] score, displayed in Supplementary Table 2) 9 . To determine the ATN score, 15 fields were randomly sampled per kidney and graded as follows: grade 0, no change; grade 1, necrosis involving less than 25% of the area; grade 2, necrosis affecting 25-50% of the area; grade 3, necrosis involving more than 50% of the area, and grade 4 necrosis involving almost the entire area. Erythrocyte attachment was monitored in the inner stripe (IS) of diseased kidneys in 12 fields at 200x magnification and graded using a 0-4 grading system (grade 0, no erythrocytes attached; grade 1, sporadic erythrocytes attached; grade 2, attached erythrocytes in more than 25% of the capillaries; grade 3, attached erythrocytes in more than 50% of the capillaries; grade 4, attached erythrocytes in more than 75% of the capillaries, displayed in Supplementary Table 1).

Localization of VWF after I/R. Localization of VWF within the injured kidney was investigated by
immunofluorescence double staining using confocal laser scanning microscopy (Zeiss LSM 710 scanning unit equipped with an Argon, a HeNe 633 laser and a DPSS 561-10 laser on an Axio Observer Z1 inverted microscope Zeiss Plan-Acromat 63x/1,4 Oil DIC M27) at room temperature. Formalin fixed paraffin sections of 2 µm were boiled for antigen retrieval in a pressure cooker for 2,5 min using target retrieval solution (DAKO, Glostrup, Denmark). After blocking with 1% BSA sections were incubated with polyclonal rabbit anti-human VWF (Dako, Copenhagen, Denmark) and monoclonal rat anti-mouse CD42b (GPIb, emfret Analytics, Eibelstadt, Germany) for 1 hour at 37°C, followed by washing steps and incubation with secondary donkey anti-rabbit IgG Alexa Fluor488 and donkey anti-rat Alexa Fluor594 (Life Technologies GmbH, Darmstadt, Germany). For signal enhancement, both primary as well as secondary antibodies were incubated in signal-enhancing solution (Pierce Immunostain Enhancer, Thermo Scientific Pierce Protein Research, Rockford, IL, USA). Before embedding the slides in moviol (Merck, Darmstadt, Germany), nuclei were stained with DAPI (Merck). Negative controls for immunostaining included either deletion or substitution of the primary antibody with equivalent concentrations of an irrelevant preimmune rabbit IgG. We used the acquisition software Zen 2009 (Carl Zeiss Microscopy GmbH) for image acquisition. No additional software was used to process the image data.
In vivo imaging of erythrocyte accumulation. Mice were anesthetized by an intraperitoneal injection of ketamine (80 mg/kg; Ceva) and xylazine (14 mg/kg; Ceva), and surgical preparation of dorsal skinfold chambers (DSCs) was performed as described previously 10 . Mice were allowed to recover from surgery for 24 h. For in vivo microscopic observations, the DSC was attached to the microscope stage, and intravital microscopy was performed using an epifluorescence microscope (AxioImager.Z2; Zeiss) with a 10x (numerical aperture 0.3) Plan-Neofluar magnification objective. Before IVM was performed in the DSC, 10 9 erythrocytes, extracted from wild-type donor mice, were treated with either normal or hyperosmolar Ringer solution, labeled with CFSE (20 µM) and intravenously transferred into wild-type recipient mice. Erythrocyte localization was recorded by attached cameras (AxioCam MRm/MRc; Zeiss), and data acquisition was performed with AxioVision software (version 4.8; Zeiss). Fields of equally exposed microscopic images were analyzed for erythrocyte accumulation.
Statistical analyses. STATA12 for Windows was used for statistical analysis. Data are presented as the mean ± s.e.m. Two-sided tests were used throughout, and the differences were considered statistically significant at P < 0.05. Pairwise (univariate) comparisons were performed using Student's t test or the Mann-Whitney U test as appropriate.