Multi-parametric thrombus profiling microfluidics detects intensified biomechanical thrombogenesis associated with hypertension and aging

Abstract

Arterial thrombosis is a leading cause of death and disability worldwide with no effective bioassay for clinical prediction. As a symbolic feature of arterial thrombosis, severe stenosis in the blood vessel creates a high-shear, high-gradient flow environment that facilitates platelet aggregation towards vessel occlusion. Here, we present a thrombus profiling assay that monitors the multi-dimensional attributes of thrombi forming in such biomechanical conditions. Using this assay, we demonstrate that different receptor-ligand interactions contribute distinctively to the composition and activation status of the thrombus. Our investigation into hypertensive and older individuals reveals intensified biomechanical thrombogenesis and multi-dimensional thrombus profile abnormalities, endorsing the diagnostic potential of the assay. Furthermore, we identify the hyperactivity of GPIbα-integrin α_IIbβ₃ mechanosensing axis as a molecular mechanism that contributes to hypertension-associated arterial thrombosis. By studying drug-disease interactions and inter-individual variability, our work reveals a need for personalized anti-thrombotic drug selection that accommodates each patient's pathological profile.

Fig. 1. Combining microfluidic stenosis assay with multi-fluorescence imaging to comprehensively characterize biomechanical platelet aggregation.

a A microfluidic chip with a quarter coin placed adjacently. b Illustration of the experimental setup. c Zoom-in of the dashed box in (b). A hump inside the channel creates 80% stenosis. When blood is perfused over, platelets spontaneously aggregate around the hump. d Shear rate and shear stress at the stenosis area estimated by fluid dynamics simulation. e Left: a layout of the two sets of fluorescently tagged sensors for thrombus profiling. Right: zoom-in of the thrombus shown in (c), illustrating the staining of Sensor Set 1. f Representative fluorescent images of thrombi stained with Sensor Set 1 (left) and 2 (right). The experiment was repeated independently on 28 healthy young subjects’ blood samples. g Representative time courses of signal intensity of biomarkers in Sensor Set 1 (left, detecting platelets (Plt), fibrinogen (Fg), VWF, and P-selectin) and 2 (right, detecting platelets, phosphatidylserine (PS), extended integrin α_IIbβ₃ (E⁺ α_IIbβ₃) and fully activated α_IIbβ₃ (Act. α_IIbβ₃)). AFU: arbitrary fluorescence unit. h, i Scatter plots with mean ± s.e.m. (n = 28) of the signal intensity of all biomarkers (h; expired and refrigerated blood samples were tested as controls, n = 4) and the normalized signal intensity of Fg, VWF, P-selectin, PS, E⁺ α_IIbβ₃ and Act. α_IIbβ₃ (i) 7.5 min after the onset of thrombus formation. The definition of each dimension of the 7-dimension thrombus profile is indicated below the graphs. j Scatter plots with mean±s.e.m. (n = 3 subjects for the first 4 groups and =5 subjects for the last group) of the thrombus residue size in the presence of aspirin (2×) or clopidogrel (2×) or both (2× or 20×), or ALB cocktail. N.S., not significant, compared with no drug treatment, assessed by one-way ANOVA (F-value = 2.87, degrees of freedom = 21) and multiple comparison (p = 0.9848, >0.9999, >0.9999, >0.9999, =0.0447, respectively, from left to right).

Fig. 2. Delineating the respective contribution of GPIbα–VWF, α_IIbβ₃–VWF, and α_IIbβ₃–fibrinogen interactions to biomechanical platelet aggregation.

a (Left) Illustration of key receptor–ligand interactions in a biomechanical thrombus, highlighting the GPIbα-integrin α_IIbβ₃ mechanosensing axis. A panel of monoclonal antibodies and their respective targets are indicated, which were used to inhibit one receptor or ligand at a time. The head of the integrin α_IIbβ₃ colored green or red, respectively denotes the integrin being unrecognizable or recognizable by PAC-1. (Right) Table layout of tested antibodies and their respective antigens and targeting receptor–ligand interactions. b Representative images of DiOC₆(3)-labeled thrombi formed in the presence of different concentrations of NMC4. The experiment was repeated independently on 3 healthy young subjects’ blood samples. c–e Dose-dependency curves of antibodies against GPIbα–VWF (c), α_IIbβ₃–VWF (d), and α_IIbβ₃–fibrinogen (e) interactions in reducing the size of biomechanical thrombi (mean ± s.e.m.; data acquired from three subjects; n ≥ 2 for each data point). f–l Comparing the normalized signal intensities of Fg, VWF, P-selectin, PS, E⁺ α_IIbβ₃, and Act. α_IIbβ₃ in the biomechanical thrombi, in the absence and presence of AK2 (f), NMC4 (g), LJ-P5 (h), 152B6 (i), 7E9 (j), LJ-155B39 (k), and LJ-134B29 (l), respectively (mean ± s.e.m.; n = 5 subjects). P values are results of multiple t-test with points without and with drug treatment paired. m Summarizing the effects of AK2, NMC4, LJ-P5, 152B6, 7E9, LJ-155B39, and LJ-134B29 on the thrombus profile into seven-digit barcodes. A positive, neutral, or negative effect is denoted by a bar being at the top, middle, and bottom of the column, respectively; it is also numerically denoted by “+”, “0”, or “−”, respectively. The antibodies are categorized by their target receptor–ligand interaction, which is indicated by different background colors.

Fig. 3. Testing the effects of inhibitors against integrin α_IIbβ₃ and VWF on biomechanical platelet aggregation.

a Dose dependency of 7E3 and 10E5, in reducing the size of the biomechanical thrombi (mean ± s.e.m.; data acquired from three subjects; n ≥ 2 for each data point). b Dose dependency of two sizes (50 and 510 nm) of polystyrene negatively charged nanoparticles (PS-CNP), in reducing the size of the biomechanical thrombi (mean±s.e.m.; data acquired from three subjects; n ≥ 2 for each data point). c–e Comparing the normalized signal intensities of Fg, VWF, P-selectin, PS, E⁺ α_IIbβ₃, and Act. α_IIbβ₃ in the biomechanical thrombi, in the absence and presence of 7E3 (c), 10E5 (d), and PS-CNP (diameter: 510 nm) (e), respectively (mean ± s.e.m.) (n = 4 or 5 subjects). P values are the results of multiple t-test with points without and with drug treatment paired. f Summarizing the seven-digit effect barcodes of 7E3, 10E5, and PS-CNP. A rule of addition is indicated, demonstrating that the add-up of the barcodes of GPIbα–VWF inhibition and α_IIbβ₃–VWF inhibition equals that of VWF inhibition, and that the add-up of the barcodes of α_IIbβ₃–VWF inhibition and α_IIbβ₃–Fg inhibition equals that of integrin α_IIbβ₃ inhibition.

Fig. 4. Characterizing abnormalities in the thrombus profiles associated with hypertension and aging.

a Scatter plots with mean ± s.e.m. (n indicated above each column) of the size of thrombi generated by healthy subjects grouped by age. P values are the results of one-way ANOVA (F value = 14.11, degrees of freedom = 50) and multiple comparison. b Comparing the time course of thrombus growth (mean ± s.e.m. with fitting lines of a sigmoidal model) between healthy young and hypertensive groups (n = 8). c Scatter plots with mean ± s.e.m. of the thrombus profiles of healthy young, healthy older, hypertensive young, and hypertensive older adult subjects (n = 33, 14, 9, and 13, respectively). P values are the results of two-way ANOVA (F values = 7.85, 106.3, 43.66; degrees of freedom = 18, 6, 3, for interaction, row factor and column factor, respectively) and multiple comparison. d P values of the significance of the impact of aging on the thrombus profile of hypertensive subjects (top), and that of hypertension on the thrombus profile of older subjects. e Scatter plot of normalized E⁺ α_IIbβ₃ signal intensity vs. thrombus size, with blood from healthy young, healthy older, hypertensive young, and hypertensive older subjects (n = 33, 14, 9, 13, respectively). Solid line: linear fitting of all data points, with a two-sided regression slope test performed to show a significant positive correlation. Dash lines: threshold values that best separate healthy young and other groups. f Scatter plots and linear fits (P values: results of two-sided regression slope test) of thrombus size and normalized E⁺ α_IIbβ₃ signal intensity vs. hypertension duration, systolic and diastolic blood pressures and their sum, HbA1C, BMI, total cholesterol, LDL-C, HDL-C, and triglyceride in hypertension patients. Green, yellow, and red background colors indicate normal, borderline abnormal, and pathologically abnormal ranges, respectively. g–l Scatter plots with mean ± s.e.m. (n indicated above each column) of the thrombus size (left) and normalized E⁺ α_IIbβ₃ signal intensity (right) of healthy young (g–i) or hypertensive and/or older (j–l) subjects, grouped by gender (g, j), race (h, k), and ethnicity (i, l). One-way ANOVA and multiple comparison or Student’s t-test was performed for data comparison, with p values, F- or t-values, and degrees of freedom (df) annotated on the figures.

Fig. 5. Hyperactivity of GPIbα and integrin α_IIbβ₃ associated with hypertension.

a Snapshots of healthy young and hypertensive young subjects’ platelets adhering to VWFA1 (upper) or Fg (lower). Experiments were repeated for four subjects. b–d Mean ± s.e.m. (n = 4 subjects) of platelet surface coverage (b, d) and rolling velocity (c) vs. shear rate on a surface pre-coated with 25 μg/mL VWFA1 (b, c) or 100 μg/mL Fg (d). P values are the results of two-way ANOVA (F values = 0.183, 13.16, 42.39 (b), 0.352, 3.345, 69.22 (c), 0.768, 6.518, 27.65 (d); degrees of freedom = 15, 5, 3 (b), 15, 5, 3 (c), 15, 5, 3 (d) for interaction, row factor and column factor, respectively) and multiple comparison. e BFP setup (top) and molecular binding illustration (bottom). f–k Adhesion frequency (Scatter plots with mean ± s.e.m.; n ≥ 6) (f, i), effective avidity and affinity (mean ± standard error; n ≥ 6) (g, j) and bond lifetime vs. force (mean ± s.e.m., n ≥ 50 for each data point) (h, k) of VWFA1- (f–h) or Fg- (i–k) coated beads binding to healthy young (HY; three subjects) or hypertensive (HTN; four subjects) subjects’ platelets. Student’s t-test results are annotated. l fBFP setup. m Representative time course of a hypertensive subject’s platelet’s Ca²⁺ level during repeated VWFA1 pulling at 40-pN clamping force. Peak increase ΔI_max is marked. n Intraplatelet ΔI_max (scatter plot with mean ± s.e.m.; n ≥ 7) of healthy young (left) and hypertensive (right) subjects’ platelets during VWFA1 pulling at different clamping forces. P values are results of two-way ANOVA (F values = 5.601, 6.146, 63.34; degrees of freedom = 3, 3, 1, for interaction, row factor and column factor, respectively) and multiple comparison. f, i, n different colors indicate different subjects. o–r Representative flow cytometry histograms of E⁺ α_IIbβ₃ (o), total α_IIbβ₃ (p), Act. α_IIbβ₃ (q), and P-selectin (r) signals on healthy young (blue) and hypertensive (red) subjects’ platelets. s Scatter plot with mean ± s.e.m. (n = 5 subjects) of flow cytometry MFI of total α_IIbβ₃, E⁺ α_IIbβ₃, Act. α_IIbβ₃, and P-selectin signals on healthy young (green) and hypertensive (red) subject’s platelets. P values, t ratios, and degrees of freedom (df) of Multiple t-test are annotated. t Proposed mechanisms of E⁺ α_IIbβ₃ over-expression in the biomechanical thrombi of hypertension patients.

Fig. 6. Drug–disease interactions and personal thrombus barcodes.

a Individual point plot (n = 5; lines connecting points of the same subjects) of the thrombus profiles of hypertension patients without and with aspirin/clopidogrel (2×), NMC4 (IC50), and 7E3 (IC50) treatment. P values are the results of two-way ANOVA (F values = 12.59, 23.13, 34.27; degrees of freedom = 18, 6, 3, for interaction, row factor and column factor, respectively). b The seven-digit effect barcodes of hypertension without and with NMC4 or 7E3 treatment. A rule of addition is indicated. c Illustration of how values in the personal thrombus profiles being low, normal or high are defined. Thrombus profiles of healthy young subjects were used as the reference, the values of which are fitted to a Gaussian distribution. The mean ± 2 s.d. range is defined as normal, and values lower and higher are defined as abnormally low and high, respectively. d Fractions of abnormally high, normal, and abnormally low values in each dimension of the personal thrombus barcodes from healthy young, healthy older, hypertensive young, hypertensive older, hypertensive+NMC4, and hypertensive+7E3 groups. e Comparing the personal thrombus barcodes of hypertensive subjects without and with NMC4 or 7E3 inhibition. Blood samples from a total of five subjects was tested. For easier visualization, bars indicating “high”, “normal”, and “low” are respectively marked by red, yellow, and green.