Differential contribution of FXa and thrombin to vascular inflammation in a mouse model of sickle cell disease

Abstract

Activation of coagulation and vascular inflammation are prominent features of sickle cell disease (SCD). Previously, we have shown that inhibition of tissue factor (TF) attenuates activation of coagulation and vascular inflammation in mouse models of SCD. In this study, we examined the mechanism by which coagulation proteases enhance vascular inflammation in sickle BERK mice. To specifically investigate the contribution of FXa and thrombin, mice were fed chow containing either rivaroxaban or dabigatran, respectively. In addition, we used bone marrow transplantation to generate sickle mice deficient in either protease activated receptor-1 (PAR-1) or protease activated receptor-2 (PAR-2) on nonhematopoietic cells. FXa inhibition and PAR-2 deficiency in nonhematopoietic cells attenuated systemic inflammation, measured by plasma levels of interleukin-6 (IL-6). In contrast, neither thrombin inhibition nor PAR-1 deficiency in nonhematopoietic cells affected plasma levels of IL-6 in sickle mice. However, thrombin did contribute to neutrophil infiltration in the lung, independently of PAR-1 expressed by nonhematopoietic cells. Furthermore, the TF-dependent increase in plasma levels of soluble vascular cell adhesion molecule-1 in sickle mice was not mediated by FXa or thrombin. Our data indicate that TF, FXa, and thrombin differentially contribute to vascular inflammation in a mouse model of SCD.

Figure 1

Characterization of anticoagulation by dabigatran and rivaroxaban. (A-B) aPTT (A) and plasma concentration (B) of dabigatran (active compound) in mice fed for 4 days with chow containing placebo or different amounts of inactive dabigatran (n = 3-5 per group). The overall P value from a 1-way ANOVA is P < .0001). (C-D) PT (seconds) (C) and plasma concentration (D) of rivaroxaban in mice fed for 10 days with chow containing placebo or different amounts of rivaroxaban (n = 5 per group). The overall P value from a 1-way ANOVA is P = .0134. (E-F) Hemostasis time (E) and the number of disruptions (F) in WT mice fed with chow containing placebo (n = 6), rivaroxaban (0.4 mg/g chow; n = 6), or dabigatran (10 mg/g chow; n = 5) for 5 days. Bleeding was observed for 30 minutes, and any formed clots were disrupted. Average bleeding time was calculated from the individual bleeding times determined during the 30-minute period; the number of clots formed is also reported (disruption number). Asterisks directly above the bars indicate the statistical significance of dabigatran- or rivaroxaban-treated mice compared with placebo-treated mice (*P < .05, **P < .01, and ***P < .001). Dab, dabigatran; Con, control placebo; Riv, rivaroxaban.

Figure 2

Role of thrombin in vascular inflammation in SCD. (A-E) aPTT (A) and plasma levels of TAT (B), sVCAM-1 (C), IL-6 (D), and MPO (E) in the lungs of BERK^AA (n = 9-10) and BERK^SS (n = 10-12) mice that received dabigatran (Dab) or control placebo (Con) for 10 days. Formalin-fixed lungs were stained for neutrophils, and neutrophils were counted in 10 high-power fields (HPF; original magnification ×400) for each mouse. (F) Average number of neutrophils in 10 HPFs (×400) of lungs. (G) Representative lung sections demonstrating neutrophil infiltration in four different groups of mice. Neutrophils stain as brown. Asterisks directly above the bars indicate statistically significant difference between BERK^SS compared with BERK^AA mice within the same treatment group (Dab or Con) (*P < .05, **P < .01, and ***P < .001).

Figure 3

Role of PAR-1 in vascular inflammation in SCD. (A-D) Plasma levels of TAT (A), sVCAM-1 (B), IL-6 (C), and MPO (D) in the lung of PAR-1^+/+ or PAR-1^−/− mice transplanted with bone marrow from WT (n = 9-11) or BERK (n = 15-16) mice. Formalin-fixed lungs were stained for neutrophils, and neutrophils were counted in 10 HPF (×400) for each mouse. (E) Average number of neutrophils in lungs. Neutrophils stain brown. Asterisks directly above the bars indicate statistically significant difference between BERK^BM mice compared with WT^BM mice within the same PAR-1 genotype (*P < .05, **P < .01, and ***P < .001). There were no differences between PAR-1^+/+ and PAR-1^−/− mice.

Figure 4

Role of FXa in vascular inflammation in SCD. (A-F) PT (A) and plasma levels of TAT (B), sVCAM-1 (C), IL-6 (D), and MPO (E) in the lungs and average number of neutrophils (F) in lungs of BERK^AA (n = 16) and BERK^SS (n = 13-15) mice fed with chow containing rivaroxaban (Riv) or placebo control (Con) for 10 days. Asterisks directly above bars indicate statistical significance between BERK^SS mice compared with BERK^AA mice within the same treatment group (Con or Riv) (*P < .05, **P < .01, ***P < .001).

Figure 5

Role of PAR-2 in vascular inflammation in SCD. (A-E) Plasma levels of TAT (A), sVCAM-1 (B), IL-6 (C), and MPO (D) and average number of neutrophils (E) in the lung of PAR-2^+/+ or PAR-2^−/− mice transplanted with bone marrow from WT (n = 8-9) or BERK (n = 8-12) mice. (F) Representative lung sections demonstrating neutrophil infiltration in four different groups of mice. Neutrophils stain as brown. Asterisks directly above the bars indicate statistically significant difference between BERK^BM mice compared with WT^BM mice within the same PAR-2 genotype (*P < .05, **P < .01, and ***P < .001).

Figure 6

Proposed role of FXa and thrombin in vascular inflammation in SCD. FSP, fibrin split product.