Sickle red blood cell-derived extracellular vesicles activate endothelial cells and enhance sickle red cell adhesion mediated by von Willebrand factor

Abstract

Endothelial activation and sickle red blood cell (RBC) adhesion are central to the pathogenesis of sickle cell disease (SCD). Quantitatively, RBC-derived extracellular vesicles (REVs) are more abundant from SS RBCs compared with healthy RBCs (AA RBCs). Sickle RBC-derived REVs (SS REVs) are known to promote endothelial cell (EC) activation through cell signalling and transcriptional regulation at longer terms. However, the SS REV-mediated short-term non-transcriptional response of EC is unclear. Here, we examined the impact of SS REVs on acute microvascular EC activation and RBC adhesion at 2 h. Compared with AA REVs, SS REVs promoted human pulmonary microvascular ECs (HPMEC) activation indicated by increased von Willebrand factor (VWF) expression. Under microfluidic conditions, we found abnormal SS RBC adhesion to HPMECs exposed to SS REVs. This enhanced SS RBC adhesion was reduced by haeme binding protein haemopexin or VWF cleaving protease ADAMTS13 to a level similar to HPMECs treated with AA REVs. Consistent with these observations, haemin- or SS REV-induced microvascular stasis in SS mice with implanted dorsal skin-fold chambers that was inhibited by ADAMTS13. The adhesion induced by SS REVs was variable and was higher with SS RBCs from patients with increased markers of haemolysis (lactate dehydrogenase and reticulocyte count) or a concomitant clinical diagnosis of deep vein thrombosis. Our results emphasise the critical contribution made by REVs to the pathophysiology of SCD by triggering acute microvascular EC activation and abnormal RBC adhesion. These findings may help to better understand acute pathophysiological mechanism of SCD and thereby the development of new treatment strategies using VWF as a potential target.

Keywords: adamts 13; endothelial inflammation; extracellular vesicles; sickle cell disease; von willebrand factor.

Figure 1.. SS RBCs generate increased level of REVs than AA RBCs and SS REVs carry higher concentration of heme than AA REVs. (A&B)

Size distribution of AA REVs and SS REVs generated from RBCs in pooled-samples from healthy donors or pooled-samples from patients with SCD. Black and red lines indicate the mean number of AA and SS REVs from 5 individual tests and gray and light-red shaded area indicate the standard deviation. AA REV and SS REV share similar size distribution (Mean size ± SD: AA REV = 205.4 ± 223.7 nm vs. SS REV = 181.7 ± 200.2 nm). (C) Concentration of AA REV and SS REV generated in vitro per microliter of purified RBCs in samples from either healthy donors (gray) and patients with SCD (red). SS RBCs generated statistically significantly higher quantity of REVs than AA RBCs (Mean concentration ± SD: SS REV = 3.17 × 10⁷ ± 3.91× 10⁶/mL vs. AA REV = 6.35 × 10⁶ ± 5.68× 10⁵/mL, p = 0.011, 2-Group t test). (D) Heme concentration carried by equal number of AA REVs (gray) and SS REVs (red). SS REVs carried statistically significantly higher concentration of heme than AA REVs (Mean concentration ± SD: SS REV = 17.3 ± 1.1 μM vs. AA REV = 6.2 ± 0.3 μM, p = 0.012, 2-Group t test).

Figure 2.. SS REVs activate HPMECs, increase vWF expression and mediates enhanced adhesion of SS RBCs, which can be reduced by hemopexin and ADAMTS13.

(A, B): HPMECs were treated with SS REVs (A) and AA REVs (B) for 2 hours in 37 °C and incubated with fluorescently labeled antibodies against vWF following a fixing step with 4% PFA. The SS REV treated HPMECs demonstrated significantly increased vWF expression comparing to HPMECs incubated with AA REVs (C, n=5, p=0.012, 2 Sample t test). (D) SS RBC adhesion on HPMECs activated with SS REVs. (E) SS RBCs adhesion on HPMECs treated with AA REVs. (F) AA RBCs adhesion on HPMECs treated with SS REVs. Insets are closer view of RBCs adhered to HPMECs. (G) Interaction between SS RBC and SS REV activated HPMECs is significantly stronger than the other test groups, and this interaction is reduced by hemopexin and ADAMTS13 (mean ± SD: SS REV-SS RBC = 157 ± 42, AA REV-SS RBC = 16 ± 12, AA REV-SS RBC = 19 ± 12, AA REV-AA RBC (5 ± 4), SS REV-SS RBC with hemopexin = 14 ± 8, SS REV-SS RBC with ADAMTS13 = 19 ± 10, and heme-SS RBC = 140 ± 49, p < 0.05 for all groups except for heme-SS RBC, n = 5 in each group, 2-Group t test). *: p = 0.002 **: p = 0.001 NS: p = 0.058

Figure 3.. Heme- and REV-induced vaso-occlusion in SS Towns mice are inhibited by ADAMTS13.

(A) Heme-induced vaso-occlusion (measured in percentage of stasis) in vivo in SS Towns mice is inhibited by ADAMTS13 (mean ± SD for 1, 2, 3, and 4 hours: Vehicle: 31.8% ± 7.1%, 26.8% ± 2.9%, 24.5% ± 4.5%, 20.8% ± 2.9% vs. ADAMTS13: 6.1% ± 4.7%, 4.9% ± 4.1%, 2.4% ± 2.8%, 2.4% ± 2.8%, p < 0.05 for all time points, n = 4 for each group at each time point). (B) REV-induced vaso-occlusion in vivo in SS Towns mice is inhibited by ADAMTS13 (mean ± SD for 1, 2, 3, and 4 hours: Vehicle: 31.5% ± 5.0%, 25.8% ± 4.1%, 23.4% ± 4.7%, 18.7% ± 4.7% vs. ADAMTS13: 7.2% ± 3.0%, 6.0% ± 2.7%, 3.6% ± 2.4%, 2.3% ± 2.7%, p < 0.05 for all time points, n = 4 for each group at each time point).

Figure 4.. RBC adhesion to REV-activated HPMECs correlates with subject clinical phenotype including hemolytic and inflammatory biomarkers.

(A): A subpopulation (Group 1, N = 6) with distinct hemolysis markers of lactate dehydrogenase (LDH) levels and absolute reticulocyte count (ARCs) comparing to the rest (Group 2, N = 9) via k-means clustering analysis. RBCs from subjects in Group 2, with significantly higher LDH levels and ARCs, have greater adhesion to REV-activated HPMECs compared to the RBCs from subjects in Group 1 (B, Mean ± SD: 267 ± 125 vs. 117 ± 72, p = 0.012, 2-Group t test). The gray and green shaded areas indicate normal ranges for ARC and LDH, respectively. (C): Subjects in Group 2 with higher LDH and ARC and enhanced RBC adhesion have significantly higher WBC counts, than subjects in Group 1 (Mean ± SD: 11.4 ± 3.6 vs. 8.5 ± 1.2, p = 0.47, 2-Group t test). Shadowed area: WBC count range from 4.5 to 10 × 10⁹. (D): Subjects in Group 2 with higher LDH and ARC and enhanced RBC adhesion have higher ferritin levels, although not statistically significant, than subjects in Group 1 (2841 ± 2804 vs. 876 ± 1527, p = 0.106, 2-Group t test). Shadowed area: normal range of ferritin level (2 to 1000 μg/L) [71, 72]. (E) Subjects with confirmed deep vein thrombosis (DVT) had significantly higher RBC adhesion to REV activated HPMECs than the ones without confirmed DVT (Mean ± SD: 361 ± 119 vs. 151 ± 78, p = 0.033, 2-Group t test). Six out of six subjects in Group 1 patient with lower RBC adhesion, and three out of six subjects in Group 2 did not have DVT. Four out of eight subjects in group 2 with higher RBC adhesion had DVT. DVT status of two out of eight subjects in Group 2 were not available (1 patient diagnosed as ‘unclear’, 1 patient record not accessible). *In Fig. 4C 2 patients had the exact same WBC count at 7.1×10⁹/L

Figure 5.. Impact of REV on microvascular endothelial cell response.

SS REV are capable of promoting HPMEC VWF expression within 2 hours, likely through the acute stress sentinel pathway of Weibel-Palade body degranulation. SS RBC adhere specifically to SS REV-activated HPMECs, likely mediated by endothelial VWF. The adhesion is decreased with VWF cleaving protease ADAMTS13, and with heme-binding protein hemopexin. Enhanced SS RBC adhesivity was observed in patients with elevated biomarkers of hemolysis and inflammation and thrombophilia.