Emerging functional microfluidic assays for the study of thromboinflammation in sickle cell disease

Abstract

Purpose of review

This review briefly summarizes the significant impact of thromboinflammation in sickle cell disease in relation to recent advances in biomarkers that are used in functional microfluidic assays.

Recent findings

Sickle cell disease (SCD) is an inherited hemoglobinopathy that affects 100 000 Americans and millions worldwide. Patients with SCD exhibit chronic haemolysis, chronic inflammation and thrombosis, and vaso-occlusion, triggering various clinical complications, including organ damage and increased mortality and morbidity. Recent advances in functional microfluidic assays provide direct biomarkers of disease, including abnormal white blood cell and red blood cell adhesion, cell aggregation, endothelial degradation and contraction, and thrombus formation.

Summary

Novel and emerging functional microfluidic assays are a promising and feasible strategy to comprehensively characterize thromboinflammatory reactions in SCD, which can be used for personalized risk assessment and tailored therapeutic decisions.

INTRODUCTION

Inherited haemoglobin (Hb) disorders are carried by nearly 7% of the world’s population, with most structural Hb variants having the recessive b-globin gene mutations, β^S or S and β^C or C [1–3]. Sickle cell disease (SCD) arises when these mutations are inherited homozygously (Hb SS or SCD-SS) or paired with another b-globin gene mutation, such as Hb C (Hb SC or SCD-SC). Chronic inflammatory processes, haemostatic alterations and thrombotic events are common in SCD [4,5]. Under hypoxic conditions, HbS polymerizes in red blood cells (RBCs) and provokes a complex pathophysiology of acute and chronic organ damage [5,6]. Sickle RBCs exhibit reduced deformability [7–10], increased adhesion to the endothelium [9,11–15], and are prone to haemolysis, which contributes to endothelial inflammation and coagulation activation [4]. This complex crosstalk between thrombosis and vascular inflammation, termed thromboinflammation, ultimately plays a significant role in major complications of SCD, including venous thromboembolism (VTE), vaso-occlusion, ischemia-reperfusion and chronic organ damage, resulting in increased morbidity, healthcare utilization and reduced life expectancy [16,17,18^■■]. Understanding the heterogeneity of pathogenic thromboinflammatory responses in SCD may assist in providing timely, tailored therapeutic interventions to patients with SCD [18^■■,19,20]. However, the crosstalk between thrombosis and vascular inflammation in SCD has not been thoroughly characterized and predictive biomarkers of disease severity and response to therapies remain under-developed [21].

To date, studies of thromboinflammatory responses in SCD have either utilized animal models or in-vitro assays. The main limitations of available in-vitro assays include among others the lack of components of the microcirculation such as endothelial cells and the inability to precisely regulate the oxygen content or shear stress, which can directly impact sickle RBC behaviour. Therefore, a comprehensive assessment of how cellular, plasma-derived and vessel wall components contribute to the thromboinflammatory milieu in SCD cannot be thoroughly captured.

ENDOTHELIAL ACTIVATION AND INFLAMMATION IN SICKLE CELL DISEASE

The pathophysiologic mechanisms in SCD have been extensively reviewed in prior publications [4,22,23], but we would like to draw attention to the concept that proinflammatory and prothrombotic signals reciprocally activate each other to promote vascular damage, vaso-occlusion and end-organ damage. During intravascular haemolysis, SCD RBCs release damage-associated molecular patterns, including free heme and heme-laden red blood cell-derived extracellular vesicles (REVs), which activate endothelial cells through Toll-like receptor 4 (TLR4) and initiate the translocation of nuclear factor kappa B (NF-κB) from the cytoplasm to the nucleus, wherein it induces the transcription of genes for adhesion molecules [24,25] (Fig. 1). As a result, endothelial expression of vascular cell adhesion protein 1 (VCAM-1) [26,27] and intercellular adhesion molecule 1 (ICAM-1) [28,29] is significantly enhanced, an event that propagates platelet capture, SCD RBC and leukocyte firm adhesion. These adhered leukocytes release free radical species that inactivate nitric oxide, a homeostatic mediator that reduces platelet activation and thrombosis and restricts endothelin-1 and VCAM-1 overexpression [30]. Recruited neutrophils also engage in heterotypic cell interactions with platelets, which were previously shown to contribute to vascular thrombosis in SCD [31]. Release of SCD REVs can activate the intrinsic pathway of coagulation in a mechanism that is partially dependent on FXI, and is associated with elevated plasma levels of prothrombin fragment 1·2 and D-dimer [32–35]. In murine models of SCD, leukocytes expressed significantly higher levels of tissue factor (TF) and had increased TF activity compared with control (AA) mice [36]. Systemic inhibition of TF reduced inflammatory cytokine and myeloperoxidase levels in the lungs of treated SCD mice, and attenuated endothelial cell injury as demonstrated by reduced plasma levels of IL-6, serum amyloid P and soluble VCAM-1. In contrast, endothelial cell specific deletion of TF had no effect on coagulation but selectively attenuated plasma levels of IL-6 injury in SCD. In murine studies of SCD vaso-occlusive crisis, thrombin activation of protease-activated receptor-1 (PAR-1) on endothelial cells contributed to lipopolysaccharide (LPS)-induced pulmonary vaso-occlusion mediated by arteriolar neutrophil-platelet microthrombi [35]. In our preliminary studies, REV-treated endothelial cells expressed higher levels of von Willebrand factor (vWF) and enhanced formation of vWF multimers, which have been shown to mediate sickle RBC and platelet recruitment (Fig. 1) [37]. The above processes support that several proinflammatory and prothrombotic mediators have distinct contributions to activation of coagulation, vascular inflammation and endothelial cell injury in SCD.

FIGURE 1.

Overview of thromboinflammation in sickle cell disease.

EXISTING MICROFLUIDIC ASSAYS FOR ASSESSING ABNORMAL CELLULAR ADHESION IN SICKLE CELL DISEASE

Red blood cell [38] and white blood cell (WBC) [39] adhesion to the endothelium can be used as biomarkers of microvascular inflammation [40,41], and likely contribute to SCD pathogenesis [15,42–46]. Sickling-dependent damage of the RBC membrane promotes the expression of molecules that are either not expressed in healthy RBCs [e.g. very late activation antigen-4 (VLA-4)], or are physiologically shielded from the cell surface, including negatively charged phosphatidylserine, CD36, intercellular adhesion molecule-4 (ICAM-4) and basal cell adhesion molecule and Lutheran blood group (BCAM/LU) [47–55]. Early studies of ex-vivo blood cell adhesion relied on open systems under static conditions and parallel plate flow chambers [56]. Flow cytometry of aberrant surface molecule expression or activation has served as a surrogate to directly measure abnormal cell adhesion in individuals with SCD [57–59], but quantitative changes in surface molecules (e.g. BCAM/LU) do not always faithfully recapitulate changes in cell adhesive properties [51,60]. Key clinical and experimental studies employing flow chambers or ex-vivo rat mesocecum [11,12,61,62] have shown that sickle red cell and WBC adhesion [39], plausibly contribute to the pathogenesis of vascular occlusion [15,42,43,61] and may correlate with disease severity [38,46,63]. However, few measurements of blood cell adhesion in SCD at baseline and with therapy have been reported, and none of these studies were conducted under pathologically relevant conditions (i.e. hypoxia, activation of endothelium) [12,25,29].

To address these unmet needs, our laboratory has focused on the development and utilization of microfluidic technologies such as the SCD-BioChip, which is integrated with a micro-gas exchanger, biomolecular probes, laminated endothelium, high-throughput microscopic imaging and data analysis methods to study blood-endothelium interactions and adhesion dynamics under both normoxic and hypoxic conditions (Fig. 2a,b). We have focused on standardizing these microfluidic assays to measure abnormal blood cell adhesion in whole blood samples and have uncovered significant associations with clinical phenotypes and response to treatment in SCD [56,64–73].

FIGURE 2.

Functional microfluidic assay for assessing cellular adhesion and endothelial activation. (a) SCD-BioChip microfluidic chips are manufactured using glass slides, double-side adhesive and polymethyl methacrylate (PMMA), containing three parallel microchannels of 50 μm height. The outlet tubing has the option to be connected to a micro-gas exchanger demonstrated in (b). The micro-gas exchanger is composed of an outer gas impermeable tubing and an inner gas permeable tubing enabling controllable gas exchange. The SCD-BioChip assay is versatile allowing quantitative assessment of endothelial cell contraction, blood cell interactions with endothelial cells, cell adhesion to adhesion molecules (i.e. P-selectin, E-selectin, ICAM-1, VCAM-1) and extracellular matrix proteins (i.e. laminin and fibronectin). (c) Thrombin-mediated endothelial contraction (scale bar represents 100 μm). (d) RBC adhesion on REV-activated endothelium (scale bar represents 100 μm). (e) Adhesion of purified peripheral neutrophils on E-selectin coated microchannels (scale bar represents 50 μm).

Utilizing our SCD-BioChip assays (Fig. 2a), we demonstrated that adhesion profiles of SCD RBCs on laminin-coated microfluidic surfaces capture patient clinical Hb composition, wherein patients with SCD (HbSS) demonstrated higher adhesion profiles compared with patients with HbSC, beta-thalassemia and healthy donors (HbAA) [64]. Moreover, we incorporated a micro-gas exchanger (Fig. 2b) and demonstrated that SCD RBC adhesion profiles to laminin and fibronectin were significantly increased under hypoxia compared with normoxia [66]. Further analysis identified a unique patient sub-population whose RBCs showed maximal adhesion to laminin and fibronectin in response to hypoxia, which in turn correlated with a more severe clinical SCD phenotype [66]. Using ICAM-1 laminated in microchannels, we found that RBC-ICAM-1 interactions were mediated by fibrinogen bound to the RBC membrane [68]. Specifically, SCD RBCs exhibited rolling adhesion and firm attachment to ICAM-1 at highershearstressover3000/s, and lower shear stress below 2000/s, respectively. In separate studies, we demonstrated that SCD RBC adhesion on heme-activated endothelialized microfluidic channels varied amongst individual patients and was associated with biomarkers of haemolysis and inflammation, age and a recent history of transfusion [67]. In another study, we showed the in-vitro efficacy of an FDA-approved Syk tyrosine kinase inhibitor for chronic myeloid leukaemia (Imatinib) in blocking Band 3 tyrosine phosphorylation on SCD RBCs and reducing RBC adhesion to heme-activated human pulmonary microvascular endothelial cells under combined hypoxia (83% SpO₂) and controlled shear stress (1dyne/cm²) [74]. Thrombin activation of endothelial cells interferes with their barrier function, promotes cell contraction and leads to increased expression of ICAM-1, VCAM-1 and vWF, which results in circulating SCD RBC capture and firm adhesion (Fig. 2c). An endogenous serine protease inhibitor, antithrombin-III (AT-III), capable of inhibiting several enzymes in the coagulation cascade, including thrombin, was demonstrated to mitigate the endothelial cell contraction and reduced surface expression of vWF and VCAM-1, downregulating SCD RBC adhesion [69,73]. Alternatively, SCD REV-treated microvascular endothelial cells demonstrated elevated levels of ICAM-1 [29] and vWF[65] compared with REVs derived from normal RBCs and these effects correlated with biomarkers of haemolysis and incidence of deep vein thrombosis [65]. Targeting vWF with its endogenous protease ADAMTS13 or neutralizing heme with hemopexin, each resulted in significant reduction of REV-mediated adhesion events [65].

Monocytes [75–78] and neutrophils [79–81] are increased in number, abnormally activated and associated with adverse outcomes in SCD [24,78,82–85]. Abnormal interactions between leukocytes and endothelial selectins (i.e. P-selectin and E-selectin) have been targeted in vitro and in vivo with antibodies and small molecules [70,71,86–89], and have been the focus of Phase II and III clinical studies in SCD. In our previous work, heterogeneous leukocyte adhesion was observed in P-selectin coated microchannels, and these adhesion events were inhibited dose-dependently following pretreatment with Crizanlizumab (ADAKVEO; Novartis International AG) [70,90]. In addition, we demonstrated that pretreating whole blood with Crizanlizumab attenuated leukocyte rolling but did not affect firmly adhesion of cells. When we used E-selectin coated microfluidic channels, we found that neutrophils from SCD individuals exhibited significantly higher adhesion under normoxic and hypoxic conditions compared with neutrophils from healthy AA donors (Fig. 2e). Similarly, inhibiting E-selectin with a blocking monoclonal CD62E antibody significantly reduced endothelial adhesion of neutrophils from patients with SCD.

Together, these results demonstrate that microfluidic assessment of abnormal blood cell adhesion on sub-endothelial matrices, endothelial adhesive molecules or on activated endothelium, is a powerful approach to diagnose and quantitate abnormalities in cell activation state and for assessing the therapeutic efficacy of disease modifying therapies targeting blood cells. In addition, we anticipate that blood cell adhesion on endothelial cells preactivated with patient blood components, such as plasma or circulating extracellular vesicles, can be used as a functional biomarker of vascular inflammation for individual patients with SCD.

EMERGING THROMBUS FORMATION ASSAYS FOR ASSESSING THROMBOSIS IN SICKLE CELL DISEASE

Thrombin generation assays and thromboelastometry have demonstrated that patients with SCD exhibit increased levels of thrombin-antithrombin complexes, D-dimer, prothrombin fragment 1.2, factor VII and fibrinogen [32,91–94]. However, these assays lack key components, cellular fractions or endothelial cells that are known to contribute to fibrin deposition and the structure, properties and dynamics of sickle thrombus formation [95^■,96]. Further, current thrombus formation assays lack the capability of conducting tests under hypoxia, which was previously shown to augment endothelial cell TF expression and promote thrombosis in SCD [96,97] among other mechanisms.

Using the SCD-BioChip technology laminated with E-selectin, we found that hypoxia leads to the formation of large RBC aggregates, specifically in SCD individuals with lower baseline haemoglobin content and higher LDH levels. Importantly, this subgroup of SCD patients exhibited reduced neutrophil adhesion to E-selectin, suggesting that the degree of neutrophil activation in SCD is inversely proportional to leukocyte L-selectin expression (manuscript under review). Ongoing studies will focus on the diagnostic utility of this assay in steady-state SCD [71,98–100].

Recently, we have developed a functional microfluidic assay for the quantitative and dynamic measurement of thrombus formation using combined particle image velocimetry (PIV) and wavelet-based optical flow velocimetry (wOFV) [72]. This thrombus formation assay was tested using citrate-anti-coagulated whole blood supplemented with 100mmol/l CaCl₂. Whole blood flow was introduced using a constant pressure pump. Higher degree of coagulation represented by increased thrombus formation within microchannels altered the blood flow resistance, which can be visualized and quantitated by measuring local flow velocity field alterations under the provided constant pressure. Both PIV and wOFV were utilized to analyse microscopic images of whole blood flow in a microchannel during coagulation. PIV was used to compute the average velocity over the entire image domain, while wOFV was used because it provides orders of magnitude higher spatial resolution and is well suited at determining RBC velocity from captured images of whole blood (Fig. 3a–f) [72].

FIGURE 3.

Thrombus formation microfluidic assay [72]. (a) Photo of microfluidic channel with formed thrombus. (b) Temporal evolution of the horizontal component of velocity along the flow channel is used for quantification of the thrombus formation process including kinetic rate of thrombus formation and percentage of occlusive thrombus within the entire microchannel. (c, e) Representative microscopic images at 230 and 380 s during a thrombus formation assay. (d, f) Wavelet-based optical flow velocimetry (wOFV) determines the velocity field from blood flow streamlines. Higher blood flow velocity is indicated by the bright yellow velocity field and streamlines are shown in green. Blood coagulation is induced by recalcification of whole blood with 100 mmol/l CaCl₂. As a thrombus is formed, the blood flow velocity field changes significantly from the 230 to 380 s timepoint.

CONCLUSION

Experimental evidence is mounting that a complex crosstalk between inflammatory and thrombotic pathways become operative in SCD and is critical for major clinical manifestations of SCD, including stroke, venous thromboembolism, acute chest syndromeand ischemia-reperfusion injury [20]. Despite our improved understanding of pathophysiologic mechanisms in SCD, there are persistent challenges in the diagnostic assessment and management of SCD patients that relate to the inherent heterogeneity of the disease, the lack of reliable biomarkers of disease severity and thrombotic risk, and the lack of in-vitro diagnostics that comprehensively capture thromboinflammatory events ex vivo [18^■■,20]. As a result, future studies in the SCD field should also include the development of clinical microfluidic technologies for the assessment of heterotypic cell–cell interactions, thromboinflammation, organ-on-chip applications and in-vitro preclinical drug screening endeavours.

KEY POINTS.

• The complex crosstalk between inflammation and thrombosis (thromboinflammation) is incompletely understood in SCD, in part due to lack of available biomarkers of disease and because of limited diagnostic tools.

• Recent advances in functional microfluidics can capture changes in white blood cell and red blood cell adhesion, cell aggregation, endothelial degradation and contraction, and thrombus formation for comprehensive assessment of microvascular thrombosis mechanisms in SCD.

• Future efforts should focus on the refinement and optimization of microfluidic tools to better define the functional components of thromboinflammatory processes across various diseases that will allow for individualized risk assessment and tailored therapeutic decisions.