Plasma peptidomics: Cutting coagulation down to size

Blood coagulation is a tightly regulated sequential proteolytic cascade that culminates in the generation of fibrin clot. Many of these well-orchestrated limited proteolytic events generate a new plasma peptide, each with its own origin story. In this issue of the Journal of Thrombosis and Haemostasis, Del Castillo Alferez et al. deploy mass spectrometry to define dynamic peptide signatures during blood coagulation[1]. Their data demonstrate that many of these peptides have stories that are still entirely untold or are much more complicated than initially perceived. In a long overdue technical achievement to be celebrated, the study highlights the significant challenges but vast opportunity for such an approach to transform coagulation research and, perhaps, even clinical practice (Fig. 1).

Figure 1. Dissection of blood coagulation by plasma peptidomics.

Tissue factor (TF)-initiated coagulation of plasma generates numerous peptides. Mass spectrometry and computational methods are then used to identify these peptides, some well-known and others previously uncharacterized. Plasma peptidomics can be deployed with variable TF concentrations, kinetic time points, or specific protease inhibitors such as hirudin to gain mechanistic insights. A STRING (Search Tool for the Retrieval of Interacting Genes/Proteins) network highlights the complex web of limited proteolysis extending far beyond the canonical coagulation cascade. Further technical advancement of plasma peptidomic methods is required, but one can envision a future (dotted arrow) with robust clinical applications of these proteolytic signatures in bleeding and thrombotic risk prediction. Created with BioRender.

Whereas a conventional rendering of the blood coagulation cascade would portray a limited number of peptide cleavages to a limited number of zymogens[2], tissue factor (TF)-triggered coagulation of plasma reproducibly induced one or more activation-dependent peptides in 38 unique proteins! Many, such as factor IX, prothrombin and fibrinogen, are expected. But many more, including a variety of proteins implicated in human disease such as apolipoprotein E, fibrillin-1, and transthyretin, are not.

Other proteins identified have more established relationships to blood coagulation but were not known or well appreciated to undergo limited proteolysis during blood coagulation and warrant more detailed investigation. Most interesting are those peptides that were only present after coagulation. Ceruloplasmin is a metalloprotein that binds copper in plasma; it was shown decades ago that ceruloplasmin can be cleaved by plasmin[3] and elevated levels are associated with venous thromboembolism[4]. CXCL7 is derived from cleavage of pro-platelet basic protein, which is released primarily from platelet alpha-granules and supports, among other functions, leukocyte chemotaxis to the platelet thrombus[5]. S100A8 is a calcium-binding protein secreted by neutrophils and monocytes that promotes procoagulant platelet formation and vascular inflammation[6].

Another intriguing group of proteins are those with peptides that are readily detected in normal plasma but increase upon TF-dependent coagulation. Plexin domain containing 2, encoded by PLXDC2, is a transmembrane protein found on neurons but also endothelial cells and macrophages that can be cleaved from the cell surface[7]. PLXDC2 peptide intensity increased in a TF-dependent manner and was completely reversed by thrombin inhibition. SNPs at the PLXDC2 locus are associated with factor V levels, as well as prothrombin time and partial thromboplastin time[8, 9]. Whether the observed cleavage of PLXDC2 and these associations are causally related and by what mechanism remains unclear. Of particular interest are complement-derived peptides, including those derived from C3, C4a, and C4b. Notably, C3f, a peptide liberated during C3b inactivation, demonstrated a time-dependent increase in summed intensity score along a kinetic trajectory that closely parallels prothrombin activation and fibrinopeptide release. The complement and coagulation zymogen activation cascades are closely interrelated and frequently invoked in the pathogenesis of various thromboinflammatory disorders[9], though the connections and physiologic significance of these complement-coagulation interactions have been challenging to dissect – here plasma peptidomics could offer a unique opportunity to map these networks.

Even within archetypal coagulation factors that undergo proteolytic activation during TF-triggered blood coagulation, plasma peptidomics uncovers non-canonical cleavage sites and unexpected activation peptide heterogeneity. For example, the canonical fibrinopeptides A and B were observed, but were accompanied by a distribution of less abundant overlapping fibrinopeptide-like sequences, perhaps reflecting post-cleavage processing in plasma. By contrast, a novel, less abundant activation peptide(R₄₄-R₇₅) immediately following fibrinopeptide B was detected exclusively upon TF-induced blood coagulation and likely reflects an alternative fibrinogen cleavage site on knob B. Multiple unexpected cleavage products in the αC (C-terminal) region of fibrinogen were also observed. Curiously, when thrombin is inhibited with hirudin prior to coagulation initiation, low abundance sequences mapping to fibrinopeptide A and B are still detectable even though the prothrombin activation peptide is not. Given that fibrinogen is at least an order of magnitude more abundant than prothrombin, this likely reflects limitations of the method’s sensitivity, but trace cleavage by non-canonical proteases activated upstream of thrombin may be contributory.

Many important serine protease inhibitors (Serpins) that curtail excessive blood coagulation are inhibitors that are themselves cleaved in the process of irreversibly inhibiting their target proteases. Serpins involved in regulation of coagulation include antithrombin, α1-antitrypsin, and protein C inhibitor among many others[10]. Serpins contain a reactive center loop that acts as a false substrate; importantly, the cleavage event generates a unique peptide that, as demonstrated by the authors, can be efficiently detected for some of these proteases in coagulated plasma. α2 macroglobulin is an abundant, non-selective plasma protease inhibitor that uses a promiscuous bait region to ensure cleavage by circulating proteases, also leading to irreversible sequestration and eventually clearance[11]; bait cleavage peptides were also apparent after plasma coagulation. Because the half lives of many activated coagulation proteases are short, their indirect detection by accumulation of more stable enzyme-inhibitor complexes is widely utilized to study coagulation activation. While the specific protease inhibitor cleavages detected by peptidomics in this study are not surprising, the ability to one day examine many serine protease inhibitors simultaneously during blood coagulation is tantalizing.

Although these data highlight the potential of plasma peptidomics to study blood coagulation, they also exemplify many of the technical limitations these methods must overcome for this potential to be fully realized. Most notable is the observation that many expected peptides are not detected. For example, the well-characterized activation peptides released by blood coagulation factor X and the anticoagulant protein C are not detected. First, this is at least in part due to abundance, with detection of rare peptides amidst a sea of more abundant ones (i.e. fibrinogen) being a longstanding technical challenge in plasma proteomics. Second, coagulation of plasma will selectively impact protein partitioning between the liquid and solid phases, so additional processing steps will be required to ensure the full peptidome has been uniformly sampled across experiments. Third, some peptides may carry post-translational modifications such as glycosylation that preclude detection with standard algorithms. No exogenous proteases (i.e. trypsin) are added in the authors’ method, ensuring that all peptides detected arise from endogenous, potentially physiologically relevant cleavage events; this does, however, disfavor detection of internal cleavage events (i.e. activation of factor VII at R212) and favor detection of liberated N- or C-terminal peptides (i.e. fibrinopeptide A) and products of multiple proteolytic events (i.e. prothrombin cleavage at R314 and R363). Finally, these methods are not yet quantitative, and ultimately addition of isotope-labeled peptide standards will be required to facilitate reliable comparison between diverse human samples.

The resting blood coagulation system exists in an idling state that reflects an equilibrium between coagulation activation, inactivation, and fibrinolysis. This basal activation of blood coagulation is neatly reflected by the presence of peptide biomarkers including factor IX and X activation peptides, prothrombin activation peptide F1+2, and even D-dimer in healthy individuals[12]. The peptidomics studies detailed by Del Castillo Alferez et al. presently have limited sensitivity to detect endogenous functional coagulation peptides and focus on plasma coagulated in vitro. The authors propose many thoughtful approaches that may enhance sensitivity and many others can be imagined to enable robust detection of these plasma peptides marking coagulation activation. One can, however, envision a not-too-distant future where an endogenous plasma peptide signature captured by mass spectrometry comprehensively captures the basal activation of blood coagulation. Such a profile in patients with rare coagulation factor deficiencies or thrombophilia could at last help to decipher the physiologic significance of various coagulation feedback loops and regulatory mechanisms in vivo. Moreover, such a plasma peptide signature could provide profound insight into thrombotic risk, risk of recurrent venous thromboembolism after completion of therapeutic anticoagulation, response to anticoagulant therapy, or the impact of difficult-to-monitor rebalancing hemostatic therapies in patients with hemophilia and other bleeding disorders.

The present work by Del Castillo Alferez et al. marks a bold step towards cutting coagulation down to size to understand it for what it really is: an ordered but disorienting web of limited proteolytic cleavage events. Each cut births a new peptide, and each peptide has a story. Methodological advancements will fill in the holes and improve the resolution. So far, plasma peptidomics offers an emerging portrait of blood coagulation that confirms what is already known yet hints that blood coagulation may be even more complicated than previously thought.