Humoral innate immunity at the crossroad between microbe and matrix recognition: The role of PTX3 in tissue damage

3.1. Role of fluid-phase PRMs in tissue repair

Components of the humoral arm of innate immunity that play a role in tissue damage principally belong to the complement system, which in addition to defense from infections [31], also regulates migration and activation of immune cells involved in wound healing [32]. Abolition of complement activation in C3-deficient mice and defective leukocyte recruitment in C5a-deficient mice are associated with abrogation of inflammation in the wound site, leading to acceleration of wound closure [33]. In the wound site, C1q localizes without C4 and C3 on endothelium and stroma and affects in vivo blood vessel formation in the granulation tissue. C1q modulates in vitro EC functions including permeability, proliferation, migration and tube formation, contributing to skin healing [34]. Complement cascade components are also involved in activation and propagation of coagulation [32]. Mannose-binding lectin (MBL) is described to affect coagulation and to contribute to arterial thrombosis [35]. Upon binding to altered host cells in ischemia/reperfusion injury, MBL promotes the activation of the lectin complement pathway contributing to inflammation and hence tissue damage [35], [36].

Studies on the involvement of CRP in tissue damage are complicated by the low level of conservation among species in terms of regulation of expression (CRP is not an acute phase protein in the mouse) and functional activities (e.g. complement activation) [27]. A protective role of CRP was observed in tissue damage associated to autoimmunity. However, these anti-inflammatory effects of CRP have not been confirmed by other studies performed using highly pure native human CRP [37]. In models of atherosclerosis, cardiac ischemia/reperfusion and cerebral infarct, CRP was deposited in tissue lesions and activated the complement cascade, thus contributing to tissue damage [38]. Therapeutic inhibition of CRP appeared a promising approach to cardioprotection in acute myocardial infarction [39]. However, studies performed with transgenic human CRP in ApoE-deficient mice showed that CRP is not pro-inflammatory, pro-thrombotic or pro-atherogenic [40].

SAP has been show to control the inflammatory response to injury, by affecting neutrophil recruitment and activation, and exerting an anti-inflammatory effect [41]. By regulating activities of macrophages through FcγRs, SAP inhibits renal fibrosis in vivo [42]. SAP is also a component of normal tissues (e.g. skin and glomerular basement membrane) where it is covalently bound to collagens and other ECM proteins. Administration of SAP reduces pulmonary inflammation and fibrosis in animal models suggesting a role for SAP in anti-fibrotic therapy [43]. In a randomized, blinded, placebo controlled, human clinical trial, use of recombinant human SAP (PRM-151) reduced the number of fibrocytes in pulmonary fibrosis, and improved lung function in patients affected by idiopathic pulmonary fibrosis, by inhibiting alternative activation of macrophages and fibrocyte differentiation via interaction with FcγRs [44]. Recently it has been shown that SAP prevents fibrosis by interacting with DC-SIGN expressed on immune cells. A polycyclic aminothiazole DC-SIGN ligand and anti-DC-SIGN antibodies mimicked SAP effects in vitro and in an in vivo model of pulmonary fibrosis, by increasing the expression of IL-10 in lung epithelial cells and sustaining an anti-inflammatory and anti-fibrotic response [45]. Finally, SAP binds to amyloid fibrils present in tissue amyloid deposits in systemic amyloidosis, Alzheimer's disease and transmissible spongiform encephalopathy, and it contributes to pathogenesis by stabilizing amyloid deposits [29]. Administration in mice of antibodies directed against SAP within amyloid deposits was shown to trigger a potent, complement-dependent, macrophage-derived reaction that removed the massive visceral amyloid deposits [46]. The property of SAP to bind apoptotic cells, chromatin fragments and nuclear debris released under cell death, and to regulate the complement cascade [47] suggests the involvement in handling of injury-induced necrotic cells.

4.1. Gene and protein

PTX3 belongs to a class of molecules identified in early 1990s called long pentraxins on the basis of their structure. The PTX3 gene is localized on the chromosome 3 and organized in three exons with the first and second coding the signal peptide and the N-terminal domain, and the third exon coding the C-terminal pentraxin domain. The proximal promoter includes numerous potential binding sites for transcription factors such as Pu1, AP-1, NF-κB, SP1 and NF-IL-6. PTX3 is a multimeric glycoprotein of eight identical protomers assembled to form an elongated octamer with a molecular weight of 344KDa stabilized by intermolecular disulphide bonds [1]. The C-terminal domain of PTX3 contains the pentraxin signature and is up to 57% homologous to CRP and SAP. In contrast, the N-terminal domain is unrelated with any known protein sequence. A unique N-linked glycosylation site is located in the C-terminal domain at Asn220 [1].

4.2. Regulation and sources

PTX3 is rapidly induced by several stimuli in different cell types, including cells of the myeloid lineage [e.g. monocytes, macrophages, dendritic cells (DCs)] and stromal cells [smooth muscle cells (SMCs), fibroblasts, chondrocytes, ECs, epithelial cells] in response to proinflammatory cytokines, DAMPs and agonists of TLRs, microbial components or intact microorganisms [1]. Neutrophils store PTX3 in secondary granules, and upon appropriate stimulation promptly release it [48].

The expression of PTX3 is differently regulated on the basis of the cell type and/or and stimulus [1]. IL-1 receptor and MyD88 signaling are centrally involved in PTX3 induction both in sterile damage and infectious models [49], [50], [51]. In alveolar epithelial cells, TNF-α induces PTX3 through JNK. HIF-1α and C/EBPβ are involved in the regulation of PTX3 expression during cardiac surgery [52]. IFNγ inhibits PTX3 production induced by LPS in monocytes and IL-6 does not represent a stimulus to induce PTX3, in contrast with CRP and SAP. IL-10 induces PTX3 in myeloid cells and amplifies the production in inflammatory conditions, therefore suggesting that PTX3 is part of a genetic program activated by IL-10, which acts as an essential player in M2 macrophage-dependent resolution of inflammation in tissue damage [1]. PTX3 is epigenetically regulated in selected tumors (e.g., esophageal squamous cell carcinoma, leiomyosarcomas and colorectal cancer) due to hypermethylation of PTX3 gene promoter and of an enhancer [50].

4.3. Ligands and functions

PTX3 functions involve recognition and binding to different ligands, including microbial moieties, complement components, ECM proteins and growth factors [1], [53]. PTX3 is an essential component of the humoral innate response to microbes. PTX3 interacts with selected bacteria, fungi and viruses, including Klebsiella pneumonia, Pseudomonas aeruginosa, uropathogenic Escherichia coli, Aspergillus fumigatus, Paracoccidoides brasilienis, cytomegalovirus (CMV) and influenza virus (IV) type A [1], [53], [54]. PTX3, by acting as an opsonin for bacteria and fungi, activates a collaborative response between different humoral and cellular effector pathways that includes complement activation and FcγRII- and CD11b-dependent facilitated recognition and phagocytosis [55]. In humans, PTX3 single nucleotide polymorphisms (SNPs) associated in particular haplotypes are linked to susceptibility to aspergillosis, pulmonary tuberculosis, Pseudomonas aeruginosa infection in cystic fibrosis patients, fungal infections in solid organ transplanted patients, and urinary tract infections [1], [54], [56], [57].

PTX3 regulates inflammatory responses in infectious and sterile in vivo models, including acute kidney injury [58], severe acute respiratory syndrome or LPS-induced lung injury [59], atherosclerosis [60] and acute myocardial infarction [49], through multifaceted molecular mechanisms. A possible molecular mechanism is provided by binding of PTX3 to the adhesion molecule P-selectin through its N-linked glycosidic moiety, that leads to inhibition of leukocyte recruitment at damaged site [61]. In addition, PTX3 binds key elements of the complement cascade and modulates complement activation [62]. By interacting with C1q through its globular head, PTX3 modulates the classical pathway of complement. PTX3 interacts with the members of the lectin pathway ficolin-1, ficolin-2 and MBL and synergistically amplifies a complement-mediated innate response to Aspergillus fumigatus and Candida albicans [63]. As CRP and SAP, PTX3 also interacts with FH and C4BP [62] on the cell surface or in ECM, thus preventing alternative pathway-mediated exaggerated complement activation. This mechanism is of particular interest in cancer, where the interaction of PTX3 with FH is determinant in regulating complement-dependent tumor-promoting inflammation [50]. Moreover, PTX3 interacts with FGF2 and FGF8b and antagonizes their functions via its N-terminal domain and suppresses FGF2-mediated pro-angiogenic and pro-restenotic activity on ECs and SMCs respectively [64].

The first evidence of the involvement of PTX3 in ECM remodeling emerged from studies on the infertility of PTX3-deficient female mice [1], [65]. PTX3 produced by cumulus cells interacts with the ECM components TNFα-induced protein 6 (TNFAIP6 or TSG-6) and inter-α-trypsin inhibitor (IαI), which are essential to assembly a hyaluronan (HA)-rich matrix around the oocyte. PTX3-deficient female mice display infertility due to a defective assembly of this HA-rich matrix, essential for oocyte fertilization [65].

4.4. Role of PTX3 in tissue repair

The function of PTX3 in tissue repair was recently investigated in different models of non-infectious tissue damage (excisional skin wounding, chemically-induced liver and lung injury, arterial thrombosis) [51]. Results showed that after tissue damage, PTX3 contributed to the orchestration of tissue repair by interacting with plasminogen (Plg) and fibrin matrix, and promoting pericellular fibrinolysis. Local acidic pH, which normally occurs in the early phases of tissue injury [66], governs this interaction and sets PTX3 in a tissue repair mode, while retaining anti-microbial recognition and interaction with complement elements (Fig. 2 ).

PTX3 levels increased following skin injury in blood and locally, where it was detected for almost the entire duration of the repair process. PTX3 localized in the clot and bordered the pericellular ECM of aligned remodeling macrophages and MSCs collectively migrating towards the wound site, as well as the epithelial outgrowth. In later stages, PTX3 was scattered in the scab and newly formed tissue. Cell migration in cluster or ducts represents a hallmark of tissue remodeling events associated with wound repair and cancer invasion [67]. Several cell types were responsible of PTX3 production in the wound site in response to TLR activation and amplification by IL-1, including macrophages, MSCs and smooth muscle cells. Reactive MSCs, which are defined by PDGFRα and fibroblast activation protein (FAP) expression, were the most important producers of PTX3 in damaged skin. Neutrophils, which store preformed PTX3 in their granules [48], reasonably released it during clot formation. PTX3 deficiency was associated with an abnormal local hemostatic response to injury, as indicated by the presence of thicker clots, augmented fibrin deposition and persistence in damaged dermis, augmented content of factors derived from activated platelets and thrombin in wounds (Fig. 2). At later time points, PTX3-deficient wounds showed augmented expression of reactive fibroblasts and collagen deposition, and general signs of delay of mature tissue formation, including defective re-epithelialization and final epithelial hyperplasia. A premature wound contraction accompanied the initial impaired thrombotic response of PTX3-defcient mice. Differently from humans, a normal wound repair in mice includes a phase of wound contraction, which restricts the entry of microbes and facilitates wound closure. The most likely mechanism underlying the wound contraction in mice appears to be the interaction between contractile fibroblasts and the ECM [68], with the contribution of the panniculosus carnosus (a thin muscle layer found in humans only at the neck platysma) [68], [69]. Platelet-derived factors, such as PDGF, TGFβ, serotonin and thrombin have been implicated in wound contraction in animal models [68], [69]. Therefore, in PTX3-deficient wounds, an augmented content of factors reasonably derived from platelets during the clotting phase and the consequent higher expression of markers of differentiated/activated fibroblasts were possible mechanisms underlying the increased contraction and augmented collagen deposition. Administration of pharmacological inhibitors of coagulation and platelet activation acting on different molecular targets (Factor Xa, thrombin), abolished the defects observed in PTX3-deficient mice, including premature wound contraction and collagen accumulation, hence indicating that an altered hemostatic response was responsible for the phenotype observed. In agreement, in a mouse model of arterial thrombosis, PTX3 derived from vascular ECs localized within the thrombus and in the vessel wall, where it dampened thrombogenesis. By targeting fibrinogen (FG) through its N-terminal domain, PTX3 inhibited platelet adhesion and aggregation [70]. These phenotypes were independent from the interaction of PTX3 with P-selectin and leukocyte recruitment at the wound site [70]. Actually, the alterations in skin damage and arterial occlusion were identical in PTX3-deficient and double PTX3/P-selectin-deficient mice.

Increased clotting and fibrin deposition followed by increased fibrosis were also observed in models of liver and lung sterile injury [51]. In these models, genetic deficiencies causing prothrombotic responses are associated with impaired tissue remodeling and fibrosis [71]. PTX3 was observed in necroinflammatory areas associated with neutrophils in acute injury, and in portal tracts associated with remodeling macrophages and MSCs in fibrotic liver. PTX3-deficiency caused increased centrolobular thrombosis and augmented deposits of fibrin in necroinflammatory areas, and augmented liver fibrosis. In line with these studies, PTX3 was shown to play a protective role in a mouse model of brain ischemic injury, being involved in edema resolution and glial scar formation [72].

The increased clotting and fibrin and collagen deposition associated with PTX3-deficiency in all these models suggested the interplay of PTX3 with the hemostatic system. In addition, these phenotypes are reminiscent of those obtained in mice with genetic defects in key molecules of the fibrinolytic system [73], [74], [75], [76]. Indeed, PTX3 was found to interact with FG, fibrin and Plg at acidic pH (optimal range from 6.5 to 5.5), while it did not bind other tested components of the coagulation or fibrinolytic cascade [51]. In agreement with these results, the interaction between PTX3 and members of the ficolin and collectin families via their FG domain and collagen domain was facilitated in an acidic milieu [62], [63], suggesting a conformational chance of the molecule, which potentially exposes new binding sites for specific ligands. In the first phases of skin wound healing, the extracellular pH is acidic and ranges from 5.7 to 6.1, while chronic and highly infected wounds are characterized by non-acidic pH [66]. Acidification of the wound site, which occurs as result of cell metabolic adaptation to trauma-induced tissue hypoperfusion, has functional relevance in the outcome of healing by affecting several processes including cell adhesion, migration and proliferation. An acidic pH is present in inflammatory phases of acute wounds that progress in healing, while chronic and highly infected wounds are characterized by non-acidic pH [66]. The interaction with fibrin and Plg was restricted to PTX3, since short pentraxins did not interact with these factors. Moreover, the second exon-encoded N-terminal domain of PTX3 was responsible for these interactions, whereas the C-terminal pentraxin domain showed a low affinity binding only with Plg. The use of specific antibodies directed against the N-terminal domain completely blocked PTX3 binding to fibrin. The in vivo administration of PTX3 rescued all the alterations associated with PTX3-deficiency, including premature wound contraction, altered clotting response and augmented collagen deposition, and the N-terminal domain of PTX3 alone recapitulated the activity of the whole molecule, suggesting therefore the primary involvement of this domain in tissue repair [51]. The interaction of PTX3 N-terminal domain with FG was also determinant in the attenuation of platelet aggregation [70]. As assessed by the use of selective antibodies and in competition experiments, PTX3 interaction with fibrin and Plg occurred via different sites in its N-terminal domain, and did not affect the interaction between fibrin with Plg. PTX3 was shown to interact with kringle (KR) 5 domain of Plg, which is involved in triggering the Plg conformational change leading to the open form, and not with Plg KR1 domain, which mediates the initial recruitment of the proenzyme to fibrin or cell surface. Ligands of l-lysine binding sites on the KR5 domain trigger the conformational changes of Plg that passes from a closed-inactive to an open-active form essential for its conversion to plasmin, which is central in fibrin matrix turnover and thrombus removal after tissue injury [75]. Whether the interaction of PTX3 with Plg through the KR5 domain has a functional relevance in the Plg activation is presently unknown. In cell-free fibrinolysis assays, the interaction of PTX3 with Plg resulted in potentiation of fibrin gel degradation mediated by plasmin and triggered by uPA and tissue plasminogen activator (tPA). In contrast, no potentiation was observed at near neutral pH, suggesting a critical role of acidic pH in governing this PTX3 function [51]. Plg activators uPA and tPA are neutral proteases and, accordingly, diminished cell-free fibrinolysis was observed at acidic pH compared to neutral pH. Therefore, the pH dependence of PTX3 binding to fibrin and Plg ensures that it would not occur in the circulation but rather at sites of tissue repair, where it sustains fibrin matrix resolution. The acidic environment likely acts as a “switch on” signal for this function of PTX3 in tissue repair. Use of confocal and two-photon microscopy, a pH sensitive probe and fluorophore-conjugated specific antibodies ascertained the occurrence of PTX3 interaction with fibrin and Plg preferentially in more acidic areas in the wound site [51].

At the beginning of new tissue formation, remodeling macrophages and MSCs enter the wound allowing the replacement of the provisional matrix by granulation tissue rich in type I collagen, as well as other ECM proteins [4]. These cells invade the fibrin matrix creating patent passages through pericellular fibrinolysis mainly mediated by Plg activation [77]. In wild type wounds, PTX3 and Plg colocalization bordered the pericellular matrix of these cells, whereas in PTX3-deficient wounds Plg was associated with cells dispersed in the fibrin matrix. In addition, PTX3-deficient macrophages and MSCs did not associate in bundles invading the injured tissue and failed to create passages within the fibrin matrix and showed defective in vitro fibrinolytic activity. Moreover, in PTX3-deficient wounds a higher content of intact fibrin, a lower content of Plg and plasmin associated with lower levels of D-dimer were observed. Therefore, these studies suggest that PTX3 contributes to a normal repair by promoting Plg-mediated pericellular fibrinolysis in the acidic wound site (Fig. 2). This model has been strengthened by a subsequent in vivo study showing that in bone marrow-derived MSC-dependent skin repair, PTX3 released from these cells promoted MSC pericellular fibrinolysis and migration within the fibrin rich wound site [25]. In this study, PTX3-deficient MSCs implanted into skin excisional wounds induced defective wound closure compared to wild type MSCs. Histological evaluation of skin samples treated with PTX3-deficient MSCs showed a reduction of the granulation tissue and excessive accumulation of fibrin. Accordingly, PTX3-deficient MSCs showed a defective ability to degrade the fibrin matrix both in vitro and in vivo. Thus, this study suggests that the fibrinolysis mediated by MSC-derived PTX3 is directly related to the dissolution of clot and remodeling of the fibrin-rich provisional matrix, and is required for the migration of these cells into the wound, which constitutes an important pre-requisite for appropriate wound healing in this model. Moreover, by creating passages into the fibrin-clot, the invading MSCs may promote the recruitment of other remodeling cells, thereby favoring the acceleration of repair processes, such as granulation tissue maturation. The capacity of MSC to reduce inflammation, in addition to their ability of inducing reepithelialization, promoting granulation tissue formation, increasing angiogenesis and regulating ECM remodeling, candidate these cells as an innovative therapy for cutaneous healing.

PTX3 administration was also observed to revert IL-6/STA3 dependent interstitial fibrosis in a mouse model of acute renal injury [78], and a recent in vitro study suggests a possible function of PTX3 in the promotion of human and murine fibrocyte differentiation via FcγRI interaction [79]. These results suggest therefore a potential therapeutic use of PTX3 as modulator of fibrinolysis in abnormal processes of tissue repair, for instance in human pathological hypertrophic scars.

In the model of vascular thrombosis, PTX3 limited the thrombotic response, and the N-domain of PTX3 was responsible of attenuation of FG-induced platelets aggregation, whereas the PTX3 C-terminal domain was shown to inhibit collagen-induced platelet activation [70]. The interaction of PTX3 and members of the ficolin and collectin family occurs via their FG-domain or collagen-domain, which are respectively recognized by the N-domain or C-domain of PTX3 [63], and results in amplification of microbial recognition and effector functions. Therefore, PTX3 has a dual role in defense against pathogens and in regulating thrombotic events and fibrinolysis in response to tissue damage, by interacting with domains conserved in molecules belonging to the innate immunity and hemostatic system, whose functions during evolution overlapped [80], [81]. Thus, PTX3 emerges as a molecule of the humoral innate immunity that plays a regulatory role in the interplay between hemostasis and tissue repair.