Complement C3 is the main effector protein of the complement system and plays a major role in innate immunity. A growing body of evidence indicates complex interactions between the complement and coagulation cascades (Oikonomopoulou et al, 2012), which are likely to be beneficial in the context of protection following injury. We previously identified C3 as a novel clot component and demonstrated that C3 binds to fibrin with high affinity and prolongs fibrinolysis in a purified system and plasma milieu (Howes et al, 2012), consistent with results of several clinical studies (Schroeder et al, 2010; Hess et al, 2012; Howes et al, 2012).
In the present study we further explored the mechanisms by which C3 becomes incorporated into clots, by evaluating the interactions between C3 and factor XIII (FXIII). Using 5-(biotinamido)pentylamine (BPNH2) in microplate-based cross-linking assays (full details of all methods are provided Appendix S1), BPNH2 was incorporated into immobilized C3 and fibrinogen (positive control) in the presence of thrombin-activated FXIII (FXIIIa) but not zymogen FXIII (FXIIIA2B2) in a concentration-dependent manner (Fig 1A). Time-dependent incorporation of BPNH2 to C3 in the fluid phase was also observed in the presence of FXIIIa but not FXIIIA2B2 (Fig 1B). C3 in a plasma milieu was cross-linked to immobilized fibrinogen in a time-dependent and FXIIIa-dependent manner (Fig 1C). BPNH2 was cross-linked to both the C3α and C3β chains, with incorporation increasing with incubation time (Fig 1D), indicating the presence of accessible glutamine residues in both C3 chains. These data indicate that C3 is a substrate for FXIIIa that is cross-linked to fibrinogen even in the presence of physiological concentrations of other plasma substrates for FXIIIa.
Cross-linking of C3 to itself and to fibrin within a clot environment was also evaluated. No high molecular weight (HMW) species were observed upon incubation of C3 with FXIIIa in the absence of fibrinogen, indicating C3 is not cross-linked to itself (Fig 2A, B). Characteristic HMW fibrin multimer formation was observed in all samples containing fibrinogen and FXIIIa (Fig 2A, C). Fibrinogen and C3 incubated together in the absence of FXIIIa did not produce HMW products (Fig 2A, B), whereas in the presence of FXIIIa HMW multimers cross-reacting with anti-C3c antibody were identified (Fig 2D–F). Three HMW bands appeared after 5 min, two of which also cross-reacted with anti-fibrin(ogen) antibody and remained over 24 h (Fig 2E, F, arrows). The third band was absent after 5 min, suggesting incorporation into larger HMW cross-linked products. A fourth C3-containing HMW band, observed at 24 h (Fig 2E), did not cross-react with anti-fibrin(ogen) antibody (Fig 2F), suggesting C3 may cross-link to itself but only after prolonged incubation periods. Only a relatively small proportion of C3 became cross-linked to fibrin, whereas the majority of fibrin γ- and α-chains were cross-linked under identical conditions.
To evaluate cross-linking of C3 to fibrin in the presence of FXIIIa substrates, perfused plasma clots were solubilized and the incorporation of C3 and α2-antiplasmin (positive control) evaluated (Fig 2G, H). C3 was observed in clots washed with Tris-buffered saline (TBS), but was virtually absent in clots washed with NaCl, indicating non-covalent binding of C3 within clots, consistent with our previous study demonstrating binding of C3 to fibrin by surface plasmon resonance (Howes et al, 2012). C3-containing products were identified in the loading wells of both the TBS- and NaCl-washed plasma clots (Fig 2G), supporting C3 cross-linking into very large HMW multimers in plasma clots. HMW bands cross-reacting with anti-α2-antiplasmin were observed in the TBS- and NaCl-washed clots (Fig 2H), consistent with the cross-linking of α2-antiplasmin to fibrin α-chain multimers. The time-course for C3 cross-linking into plasma clots was assessed in non-perfused solubilized plasma clots (Fig 2I, J). Four C3-containing HMW products were observed after 5 min and three bands remained present after 4 h (Fig 2I). Three of the four bands cross-reacted with the anti-fibrin(ogen) antibody (Fig 2J, arrows). C3 was observed in the loading wells and in HMW components after 10 min, indicating C3 is readily cross-linked to plasma clot components, although the majority of C3 was not cross-linked after 4 h.
The results of the present study indicate that complement C3 is a substrate for FXIIIa and becomes cross-linked to fibrin during clot formation. These data are consistent with those recently reported by Nikolajsen et al (2012) who confirmed our original finding that C3 is associated with plasma clots and characterized the FXIIIa-reactive glutamine residues in C3. Importantly, our results extend their findings to identify that C3 becomes cross-linked to fibrin during clot formation in purified and plasma-based systems. These data support specific interactions between C3, FXIII and fibrin, and, with our previous results indicating the presence of two high affinity binding sites for C3 on fibrin (Howes et al, 2012), suggest that C3 may become cross-linked to fibrin α-chains, consistent with the cross-linking sites of other FXIIIa substrates (Ritchie et al, 2000). FXIIIa initially binds to glutamine containing-substrates prior to cross-linking and, given the structural similarity between the FXIII-B subunit and complement regulatory proteins (Rodriguez de Cordoba et al, 1988), the interactions between C3 and FXIII-B may be important for initial substrate recognition. Whether C3 interacts with the free FXIII-B present in excess in plasma in addition to FXIIIA2B2 requires further evaluation, as free FXIII-B has the potential to modulate C3-fibrin cross-linking but was constant in our experimental approach.
Our data add to the body of evidence supporting functionally relevant interactions between complement and coagulation in the context of responses to vascular injury, particularly the recent reports of a specific role for FXIIIa in relation to innate immune responses, which identified FXIIIa-dependent entrapment of Escherichia coli, Staphylococcus aureus and Streptococcus pyogenes in fibrin clots as important in prevention of systemic infection and bacterial release (Wang et al, 2010; Loof et al, 2011). These data and the results of the present study support FXIIIa-dependent interactions between C3 and fibrin as physiologically relevant because incorporation of C3 into a clot would localize C3 to sites of injury and consequently sites of potential pathogen invasion.