Platelets are a rich source of latent TGF-β1, a cytokine with potent profibrotic and immunomodulatory effects, yet little is known about how platelet TGF-β1 becomes activated. Current theory is that TGF-β1 (a 25 kD disulfide-bonded homodimer produced by furin cleavage of a conjoint LAP-TGF-β1 precursor) must dissociate from the LAP (a 80 kD disulfide-bonded homodimer) to be active.

Methods: Platelet-rich plasma (PRP), whole blood, or 5x108 washed platelets in HEPES-buffered modified Tyrode’s buffer with 1 mM each Ca2+ and Mg2+, was placed in an aggregometer cuvette at 37°C with thrombin (0.125 U/ml) or TRAP (10 μM) for 2 h. Samples were then centrifuged and both the pellet and the supernatants were analyzed. TGF-β1 levels were quantified by: (1) Quantikine TGF-β1 assays in which recombinant TGF-β receptor-II is used to capture active TGF-β1. Active TGF-β1 was measured in untreated samples while total TGF-β1 was measured by converting latent TGF-β1 to active by acidification, (2) Mink lung epithelial cell (MLEC) assay in which active TGF-β1 binding results in expression of a luciferase reporter gene fused to a truncated PAI-1 promoter. Samples were incubated with MLECs overnight and cell lysates were assessed for luciferase activity. Non-reduced and reduced samples were also analyzed by SDS-PAGE and immunoblotting with antibodies to TGF-β1, LAP (present in large latent complexes [LLCs] and small latent complexes [SLCs]), and LTBP-1 (present in LLCs).

Results: Latent TGF-β1 was maximally released within 5 minutes after platelet stimulation with thrombin (130 ± 15 ng/ml, n=9), whereas TGF-β1 activity increased slowly and progressively, reaching approximately 1% of total TGF-β1 at 2 hours (1.44 ± 0.69 ng/ml, n=14). A polyclonal antibody reported to be specific for active TGF-β1 reacted with a 25 kD band in nonreduced samples at all time points (5-120 mins) as expected, since SDS dissociates TGF-β1 from LAP in both SLCs and LLCs. There was also, however, an unexpected time-dependent increase in the binding of this antibody to several closely spaced bands of ~110 kD and ~260 kD representing SLCs and LLCs, respectively. Antibodies to LAP also reacted with the ~110 and ~260 kD bands. Reduced samples immunoblotted with the LAP antibody demonstrated a prominent band at 40 kD at all time points representing the LAP monomer. Preincubation of platelets with reagents that inhibit thiol-disulfide exchange reduced the peak generation of active TGF-β1 by 62–81%: bacitracin (165 ± 71 pg/ml vs. control, 848 ± 448 pg/ml, n=8, p=0.005), CMPS (781 ± 362 pg/ml vs. 2060 ± 420 pg/ml, n=5, p=0.01), DTNB (570 ± 140 pg/ml vs. 1810 ± 222 pg/ml; n=3, p=0.005). A maleimide/biotin probe (BMCC) demonstrated time-dependent incorporation into SLCs and LLCs. Chromatography of 2 h platelet supernatants using Superdex-200 demonstrated that high molecular weight fractions containing LLCs had TGF-β1 activity on Quantikine assays (n=4).

Conclusions: 1) Inhibitors of thiol-disulfide exchange inhibit the activation of platelet TGF-β1, (2) Molecular species containing TGF-β1 covalently attached to LAP or other proteins via disulfide bonds appear in the platelet supernatant over time, (3) TGF-β1 activity can be detected in high molecular weight species. Our data suggest that thiol-disulfide exchange may contribute to TGF-β1 activation; this mechanism may have implications for TGF-β activation in other cells.

Author notes

Corresponding author