Thrombin-Induced Endothelial Cell Damage Is Mitigated By Human Anti-Thrombin III in a Microfluidic Device

Introduction

Sickle cell disease (SCD) is a pleiotropic disease that results in a hypercoagulable state due to the depletion of natural anticoagulants and abnormal activation of the coagulation cascade. Vascular inflammation is a characteristic of SCD and can lead to endothelial cell activation and abnormally enhanced adhesion of hematopoietic cells. These adhesive events can lead to vaso-occlusive crises and are thus a target of pharmaceutical exploration. The complex pathophysiology of SCD suggests that there are many potential therapeutic targets, particularly during activation of the coagulation cascade. Presently, we describe the use of a microfluidic device to investigate the effectiveness of anti-thrombin III (AT-III) in mitigating thrombin-mediated effects on and red blood cell adhesion (RBC) to endothelial cells. Our microfluidic device, endothelium-on-a-chip (Fig. 1A), incorporates microchannels functionalized with endothelial cells (ECs), which allows for investigation of SCD pathophysiology under physiological shear stress. Here, we examine the extent to which thrombin causes cellular contraction in human umbilical vein endothelial cells (HUVECs) and the capacity of AT-III to mitigate this contraction. Similarly, we quantify the adhesion of sickle RBCs to thrombin-activated HUVECs and the adhesion to HUVECs pre-treated with AT-III.

Methods

Microchannels seeded with human umbilical vein endothelial cells (HUVECs) were cultured to confluency under flow for 48 – 72 hrs at 37 °C in a 5% CO₂ atmosphere. Culture media containing 10% fetal bovine serum (FBS) was perfused through the microchannels at a shear stress of 1 dyne/cm² which corresponds to a typical value seen in human post-capillary venules. Approximately 18 hours prior to the start of the experiment, perfused media was exchanged for ITS+1 media in order to condition cells to a low-serum environment and to avoid neutralization of thrombin by residual serum factors. At the start of each experiment, devices were disconnected from the dynamic flow culturing system and incubated with experimental reagents under static conditions. Endothelialized microchannels were treated with one of three conditions: Control channels were incubated with ITS+1 media for 5 hrs; Thrombin channels were incubated with ITS+1 media for 1 hr, followed by 5 IU/mL human alpha thrombin for 4 hrs; Anti-thrombin III channels were incubated with 7 IU/mL human anti-thrombin III (AT-III) for 1 hr, followed by 5 IU/mL human alpha thrombin for 4 hrs. AT-III was donated by Grifols USA (Los Angeles, California). Following each experiment, the endothelialized microchannels were fixed with 3.7% formalin. Microchannels were then imaged and confluency was measured using the FIJI plug-in PHANTAST with default settings. For adhesion experiments, whole blood samples were collected from 6 patients with homozygous SCD (HbSS). RBCs were isolated and suspended at 20% hematocrit in basal media, then perfused through treated, endothelialized microchannels. Microchannels were then imaged and adherent RBCs were quantified.

Results

The results of contraction and adhesion experiments are shown in Figures 1B&C, respectively. Control channels had an average confluency of 97.9% ± 1.7 (Fig. 1B&D), while thrombin-treated channels resulted in HUVEC contraction causing an average confluency of 63.7% ± 8.9 (Fig. 1B&E). Microchannels treated with AT-III prior to thrombin incubation had an average confluency of 92.5% ± 5.3 (Fig. 1B&F). Adhesion of sickled RBCs followed a similar trend: control channels had an average of 16.0 ± 22.7 RBCs per analyzed region (Fig. 1C&G), thrombin channels had an average of 222.8 ± 86.7 RBCs per analyzed region (Fig. 1C&H), and AT-III channels had an average of 19.6 ± 12.3 RBCs per analyzed region (Fig. 1C&I).

Discussion

We observed that thrombin treatment resulted in significant endothelial cell contraction. This thrombin induced cell contraction could be mitigated by pre-treating the microchannels with AT-III. Similarly, RBC adhesion significantly increased in thrombin-treated microchannels, and could be similarly mitigated by pre-treatment of the endothelium with AT-III. Taken together, our in vitro model suggests that AT-III may be worthy of investigation as a treatment for SCD associated thromboinflammation and vaso-occlusive crises.