Label-free microfluidic sorting of platelets and leukocytes for studying inflammation and thrombosis in blood vessel-on-chip models

Understanding blood-endothelial interactions is essential for advancing diagnostic and therapeutic strategies in inflammatory and cardiovascular disease. Conventional centrifugation-based approaches for leukocyte and platelet isolation may introduce shear induced activation, potentially confounding downstream analyses of their pathophysiological roles in the vascular microenvironment. To address this limitation, this thesis presents a Dean Flow Fractionation (DFF) microfluidic platform integrated with a leukocyte concentrator for simultaneous, label-free isolation of leukocytes and platelets from small whole blood volumes (~100 µL). The system was first optimized using PDMS devices before translating into CNC milled PMMA chips, achieving ~80% leukocyte and 45% platelet recovery, yielding ~0.3 x 106/mL leukocytes and 6 x 106/mL platelets. The gentle, centrifuge-free workflow minimized platelet activation and enabled accurate flow cytometry quantification of leukocyte-platelet aggregates. Clinical validation with healthy (n=18) and fever patient samples (n=20) showed a 4-5 fold increase in leukocyte counts compared to the clinical standard protocol (1:100) diluted blood with strong correlations for platelet (R2 = 0.6459) and leukocyte (R2 = 0.8012) measurements. The platform was further applied to evaluate antisense oligonucleotide (ASO) induced thrombosis, revealing significant neutrophil-platelet aggregation with cytosine phosphate-guanine (CpG) ASO. A small-volume (100-200 µL) DFF workflow was developed for longitudinal mouse studies, which showed elevated platelet activation in streptozotocin induced diabetic wild-type mice and increased neutrophil counts and neutrophil-platelet aggregates in ApoE-/- + STZ mice. To complement cellular analyses, a 3D stenosis-on-chip model (~450 µm lumen) and a capillary network (~100 µm) were developed to study ASO induced leukocyte and platelet adhesion under physiologically relevant flow. Overall, the developed microfluidic platforms provide rapid (~10 min), activation-minimizing, centrifugation-free blood cell fractionation and vascular models that can support drug screening, preclinical evaluation of RNA therapeutics, and point-of-care diagnostic applications.