Thrombospondin 1 (TSP-1) Partially Contributes to Shear or Stirring- Dependent TGF-beta1 Activation In Vitro and During Platelet-Rich Thrombi Formation In Vivo.

Thrombospondin 1 (TSP-1) has been shown to activate latent TGF-beta1 in vitro, but its in vivo role is unclear because TSP-1-null mice have a much less severe phenotype than TGF-beta1-null mice. TSP-1 is stored in platelet alpha-granules and released, along with latent TGF-beta1, with activation. Thus, it is strategically located to contribute to TGF-beta1 activation at sites of platelet deposition. We recently demonstrated that stirring and/ or shear could activate latent TGF-beta1 released from platelets, probably in part through thiol-disulfide exchange and we have identified TSP-1 in human platelet releasates by immunoblotting and observed that it labeled with a thiol-reactive biotin-maleimide probe (MPB) by LC/MS/MS analysis, indicating the presence of free thiols. Remarkably, we found that the amount of TSP-1 detected by immunoblotting in human platelet releasates decreased with stirring or shear and that pretreatment of platelet releasates with the thiol-reactive compounds MPB, NEM or BMCC largely prevented the loss of TSP-1 with shear or stirring. To further assess the role of TSP-1 in TGF-beta1 activation, we have now studied the effect of stirring and shear on TGF-beta1 activation in samples from WT and TSP-1-null mice. Unstirred sera or platelet releasates from WT and TSP-1-null mice had similar levels of total TGF-beta1 [serum, 91 ± 15 ng/mL in WT (n=23) and 108 ± 15 ng/ mL in TSP-1-null mice (n=23; p=0.13); platelet releasates, 58 ± 14 ng/mL in WT (n=14) and 53 ± 16 ng/mL in TSP-1-null mice (n=14; p=0.45)]. With either stirring or shear, the TSP-1-null samples demonstrated less activatibility, but the defect was only partial [stirring increased active TGF-beta1 in serum from 0.6 ± 0.2 ng/mL to 2.2 ± 0.8 ng/mL in WT mice and from 0.5 ± 0.2 ng/mL to 1.6 ± 0.5 ng/mL in TSP-1-null mice (n=23) (p=0.05 using ANOVA analysis for the linear model comparing two-way interactions of stirred samples between WT and TSP-1-null mice); stirring increased active TGF-beta1 in platelet releasates from 0.3 ± 0.2 ng/mL to 2.4 ± 1.5 ng/mL in WT mice (n=14) and from 0.2 ± 0.2 ng/mL to 1.1 ± 0.6 ng/mL in TSP-1-null mice (n=14) (p=0.005)]. To assess in vivo TGF-beta1 activation we induced carotid artery thrombi with FeCl₃ in WT and TSP-1-null mice. When thrombi were removed 5 min after total occlusion, the thrombus extracts from WT and TSP-1-null mice contained similar amounts of total and active TGF-beta1 [total TGF-beta1, 5.6 ± 2.4 ng/mL in WT and 5.5 ± 1.6 ng/mL in TSP-1-null mice (n=9) (p=0.9) and active TGF-beta1 (expressed as a percentage of total), 1.7 ± 0.7% in WT and 2.5 ± 1.4% in TSP-1-null mice (n=9) (p=0.15)]. When thrombi were removed 120 min after total occlusion, however, the percentage of active TGF-beta1 in the thrombus extracts was higher in WT than TSP-1-null mice [4.5 ± 2.0% in WT mice (n=9) and 3.3 ± 1.5% in TSP- 1-null mice (n=9) (p=0.02 using ANOVA analysis comparing interactions between time and type of mice)]. We conclude that:

1. TSP-1 released from human platelets undergoes a shear-induced, thiol-dependent conformational change,

2. TSP-1 contributes partially to shear or stirring-dependent TGF-beta1 activation in vitro in mice, and

3. Time-dependent activation of TGF-beta1 released from thrombi formed in vivo in mice is also partially dependent on TSP-1.

These data suggest a significant, but minor role of TSP-1 in TGF-beta1 activation in vivo and thus provide an explanation for the less severe phenotype of TSP-1-null mice than TGF-beta1-null mice.