Chagas’ disease or American Trypanosomiasis represents a serious burden for millions of people worldwide. This parasitic infection, considered a neglected reemerging tropical disease, causes more than 10,000 deaths each year, being a main cause of heart failure in Latin America, where it is endemic (Ferreira et al., 2016). However, Chagas disease is currently expanding to non-endemic countries (United States and Canada) and continents (Europe, Asia, and Oceania) due to migration of infected people (Sbaraglini et al., 2016), affecting nearly 6 million people worldwide (WHO, 2015).

Trypanosoma cruzi, the etiological agent of Chagas disease, exhibits a variety of mechanisms to evade the host immune response. These mechanisms enable the parasite to establish an infection and persist in the host, determining a chronic stage (Freire-de-Lima et al., 2016). The T. cruzi life-cycle is extremely complex, involving an invertebrate hematophagous triatomine vector and an extensive range of mammalian hosts, including humans (Chacón et al., 2016). T. cruzi metacyclic trypomastigotes, one of the infective forms of the parasite, are released in the feces or urine of triatomines after a blood meal. These infective forms encounter mucosa or discontinuous regions of the epithelium, infecting mammalian host cells (Cardoso et al., 2015). To invade these cells, trypomastigotes use a variety of virulence factors that interact with host components to help the parasite to invade mammalian cells. In trypomastigotes, gp90 (Dorta et al., 1995; Yoshida, 2006), mucins (Villalta and Kierszenbaum, 1984; Yoshida et al., 1989), Tc85, trans-sialidase family (gp85, gp82 and TSA-1) (Cross and Takle, 1993), gp35/50 (Teixeira and Yoshida, 1986), TS (Previato et al., 1985), gp83 (Lima and Villalta, 1988; Villalta et al., 2008); penetrin (Ortega-Barria and Pereira, 1991); and proteases such as cruzipain, oligopeptidase B and Tc80 (Murta et al., 1990), have been described. Additionally, we have shown that T. cruzi calreticulin (TcCRT) is an important virulence factor on the T. cruzi surface (Ramirez et al., 2011b).

In the host cell, trypomastigotes are confined in a parasitophorous vacuole, from which they escape to differentiate into amastigotes, the replicative form of the parasite in mammalian cells. After several rounds of replication, the amastigotes differentiate into another infective form of the parasite, the bloodstream trypomastigotes, which are released upon rupture of the host cell membrane and infect neighboring cells or enter the bloodstream (Cardoso et al., 2015). Long-term persistence likely involves episodic reinvasion as well as continuous infection that extends to different tissues (Lewis and Kelly, 2016). Once the trypomastigotes reach the bloodstream, the parasite bypasses the complement system mediated lysis and opsonization, with the support of surface proteins (Cardoso et al., 2015) or by capturing host plasma molecules (Pangburn, 2000). Thus, the parasite disseminates through the bloodstream to many tissues during the acute phase that may veer toward a chronic phase. During this stage, about 30% of those infected will develop symptoms of this disease (Cardoso et al., 2015).

Trypanosoma cruzi infection activates both innate and adaptive immune responses during the acute phase. These responses involve complement system activation, macrophages, natural killer cells, B cells, CD4⁺ and CD8⁺ T cell activation, as well as the production of pro-inflammatory cytokines such as interferon-gamma, tumor necrosis factor, interleukin-12 and interleukin-17 (Cutrullis et al., 2017). Despite its activation, the immune response fails to completely eliminate the parasite, because T. cruzi has acquired a variety of strategies to evade the host immune response. Thus, trypomastigotes possess membrane-bound proteins that provide them with a certain degree of resistance to the complement-mediated lysis, enabling them to survive for longer periods in the bloodstream (Tomlinson and Raper, 1998). Early studies have demonstrated that trypomastigotes are highly complement resistant, while non-infective epimastigotes are highly susceptible (Muniz and Borriello, 1945; Rubio, 1956). This strategy to resist the complement system activation could be extremely important to define the evolutionary spectrum of possible hosts. Thus, birds are refractory to T. cruzi infection (Minter-Goedbloed and Croon, 1981; Teixeira et al., 2006). After inoculation, parasites cannot be recovered from bird blood (Teixeira et al., 1987). However, parasites inoculated to fertile chicken eggs infect embryo cells until the 10th day post-fertilization (Nitz et al., 2004). Thus, the elimination of the parasite could depend, in an important part, on the innate immune system and its maturation. Other studies indicate that metacyclic trypomastigotes are lysed by avian serum (Lima and Kierszenbaum, 1984), in an antibody-independent manner (Kierszenbaum et al., 1981). The administration of normal chicken serum to T. cruzi infected mice generates a marked decrease in their parasitemias (Kierszenbaum et al., 1976). These results indicate that, in addition to the maturation of the immune response, the avian serum and its complement system destroy the parasite. However, the molecules and mechanisms involved are unknown.

The complement system consists of about 40 proteins circulating in the blood/plasma or anchored to cell surfaces. This system is an important component of innate immune response which recognizes microbial pathogens, promoting opsonization, inflammation, clearance of immune complexes and cellular lysis (Kemper and Hourcade, 2008). The complement system recognizes different patterns on the aggressors or antibodies bound to pathogenic antigens. Three different complement system activation pathways have been identified: Classical (CP), activated by antigen:antibody complexes and other molecular patterns which are recognized by C1q sometimes in conjunction with adaptor molecules, such as C-reactive protein (Jiang et al., 1991); Lectins (LP), which responds to mannose containing polysaccharides through the collectin Mannan Binding Lectin (MBL), and to other targets such as acetyl sugars via the Ficolins and other Collectins such as CL-11 (Garred et al., 2016); AP, activated by a variety of microbial surfaces (Kemper et al., 2010). The activation of these three pathways concludes the formation of C3 and C5 convertase enzymes (Walport, 2001a,b), and generation of opsonins C3b and C4b (Kemper et al., 2010). In the CP, C1s, in the context of C1, a complex of three proteins C1q, C1r, C1s, is activated by binding of C1q to a target. C1s, a protease, will cleave both C4 and C2, forming the cell surface- bound C3 convertase (C4b,2b) that will activate C3, generating C3a and C3b fragments. C3b, similar to C4b, has a highly reactive thiolester bond that mediates binding to a nearby C4b,2b, thus initiating the assembly of C5-convertases (C4b,2b,3b), which cleaves C5 to generate C5a and C5b (Walport, 2001a,b; Kemper et al., 2010). In the C5-convertas both C4 and C3 are covalently bound to the target membrane. Then, C5b associates with C6 and C7 in the fluid-phase (plasma), resulting in the C5b-7 complex, which inserts into the membranes of target cells where it binds to C8 and C9, forming a MAC. This step from C5-C9 is also named the TP (Kemper et al., 2014). This MAC complex promotes membrane disruption and cell lysis (Walport, 2001a,b; Kemper et al., 2010). In the LP, the danger recognition lectin modules, Ficolins and MBL, will recognize a variety of pathogen associated molecular patterns (PAMPS or danger signals), with subsequent activation of their serine proteases (MASPs). The following activation steps are similar to those of the CP. The AP is activated by default, detecting the lack of certain molecules (e.g., sialic acid) that are prominent on the host cell surface. C3 (C3bBb) and C5 (C3bBbC3b)-convertases are generated. Given the structural similarities that exist between C3 and C4, and between C2 and factor B, the AP and CP convertases share extensive properties such as the capacity to cleave C3 and C5. The steps and molecules involved in the TP are common to all activation routes.

The complement system is tightly controlled by regulatory proteins present in plasma and on the cell membranes (Morgan, 2010), including those of selected pathogens. Thus, the CP is regulated by C1-INH, a serine protease inhibitor in plasma that binds activated C1, removing C1r and C1s from the complex and inactivating them (Davis, 1988, 1989). The AP downregulation is provided by FI, a serine protease that cleaves C3b in the convertases (Morgan, 2010). In fluid-phase, there are two proteins acting as cofactors for FI: FH and C4BP (Vik et al., 1990). Both break up convertases, a property termed “decay acceleration” (Gigli et al., 1979; Reid and Turner, 1994). On membranes, decay-accelerating factor protein (DAF) binds to and disassemble the CP and AP convertases. The membrane cofactor protein (MCP) acts as a cofactor for the cleavage of C4b and C3b by FI. Another cell-surface protein, complement receptor 1 (CR1) inactivates CP and AP convertases, causing dissociation and acting as a FI cofactor. On the other hand, the AP possesses a protein that stabilizes the C3 and C5 convertases, named Properdin (P), the only positive complement regulatory protein described so far (Smith et al., 1984). This protein is also a danger recognition module for pathogens such as Chlamydia pneumonia (Cortes et al., 2011) and Neisseria gonorrhoeae (Spitzer et al., 2007).

Other complement regulatory proteins, present in the fluid-phase and on membranes, tightly regulate MAC formation. The fluid-phase C5b-7 complex is a target for S-protein and clusterin, abundant serum proteins. On the membrane, CD59 binds to C8 in the C5b-8 complex and blocks the incorporation of C9, preventing MAC formation (Lachmann, 1991; Davies and Lachmann, 1993; Morgan, 2010).

All three complement pathways can be activated in the absence of antibodies, and so the complement system acts as the first barrier preventing infection. However, certain microorganisms have developed mechanisms to escape attack and survive in the host bloodstream (Iida et al., 1989). Thus, particular pathogens possess complement regulatory proteins on their surfaces or recruit them from host plasma, as a strategy to resist the complement activation. In T. cruzi, several complement regulatory proteins have been described. Some of them have been known for many years (Tomlinson and Raper, 1998) while others have recently been revealed.

During its life-cycle, T. cruzi experiences multiple regulated morphological and physiological modifications to survive within the triatomine vector and mammalian host cells. Thus, bloodstream trypomastigotes (infective form in mammalian host), amastigotes (intracellular replicative stage in host cells) and metacyclic trypomastigotes (infective forms present in the vector’s feces) resist the complement-mediated lysis and the macrophage defense mechanisms. Differently from trypomastigotes, epimastigotes, the replicative and non-infective form of the parasite, present in the mid-gut of triatomine vectors, are highly susceptible to the complement-mediated lysis and cellular destruction (Muniz and Borriello, 1945; Rubio, 1956; Cestari Idos et al., 2008).

Early experiments indicated that mice infected with virulent bloodstream trypomastigotes present a marked exacerbation of T. cruzi infection with higher parasitemia and mortality when the complement system is depleted with cobra venom factor (Budzko et al., 1975). This finding indicates that complement is important to control the acute infection. The complement system may cause lysis by activation of the cascade or facilitate the parasite destruction by phagocytic cells. Infective forms of T. cruzi (trypomastigotes) in the presence of specific antibodies plus complement, show a certain proportion of surviving parasites (Kierszenbaum and Ramirez, 1990). In an attempt to select a trypomastigote subpopulation with increased resistance to lysis by specific antibodies plus complement, the parasites sensitive to immune lysis were eliminated by three successive passages in mice. The resulting parasites maintained their level of sensitivity and produced similar levels of parasitemia (Kierszenbaum and Ramirez, 1990). This finding indicates that the complement resistance is more dependent on phenotypic than genetic variations. In other words, it is likely that the number of complement regulatory proteins on the parasite surface follows a phenotypic normal distribution, and this will affect their complement system susceptibility. Consequently, the ability to resist the complement differs not only among different T. cruzi strains (Cestari Idos et al., 2008), but also among individuals in the same strain. Thus, the Colombiana (TcI) strain is more susceptible than the Y strain (TcII). Lysis is apparently due to the CP and LP in early stages of activation (Cestari Idos et al., 2008).

The role of complement in T. cruzi immune lysis has been studied for a long time. This susceptibility varies considerably, among other factors, with the parasite developmental stage (Anziano et al., 1972). During its cycle in the insect, T. cruzi assumes various forms and its infectivity is associated with the acquisition of resistance factors (Iida et al., 1989). As previously mentioned, while metacyclic and bloodstream trypomastigotes are resistant, the epimastigotes are extremely sensitive to the complement system (Muniz and Borriello, 1945; Rubio, 1956). Amastigotes, the replicative form of the parasite, are not exclusively intracellular, being found in the bloodstream during the infection acute stages (Ley et al., 1988). In circulation, amastigotes are extremely efficient activators of the complement system cascade, binding the terminal components (C9 polymerization). Although this parasite stage activates the AP, lysis does not occur because the C5b-9 complexes bind but do not insert into the lipid bilayer of the plasma membrane (Iida et al., 1989). This kind of activation in the absence of final lysis also occurs in other parasites such as promastigotes of Leishmania, which activate the complement system, but with a minimal surface deposition of C9 (Puentes et al., 1988). Bacteria such as Borrelia burgdorferi produce an 80 kDa CD59-like protein, which has affinity for C8β subunit and C9, and prevent MAC formation (Pausa et al., 2003; Lambris et al., 2008). A possible explanation of amastigote resistance to the complement system is that hydrophobic domains present on parasite surface molecules serve as non-specific “traps” for nascent C5b-7 complexes that will inefficiently bind C8 and C9. Another possible explanation is the presence of an unidentified specific inhibitor, homologous to a host factor that will prevent the incorporation of C8 and C9 on amastigotes (Iida et al., 1989). This resistance, both in trypomastigotes and amastigotes, involves the expression of surface and soluble components playing a role in complement activation.

Several molecules present on the parasite mediate resistance at different levels of the complement system cascade. In addition, trypomastigotes capture inhibitory host components, which are used to inhibit complement activation on the parasite surface, such as: PMV (Figure 1A) (Cestari et al., 2012), TcCRT (Figures 1B,C) (Ferreira et al., 2004), T. cruzi trypomastigote-decay accelerating factor (T-DAF) (Figure 1D) (Kipnis et al., 1988; Tambourgi et al., 1993), T. cruzi complement regulatory protein (TcCRP) (Figure 1E) (Norris et al., 1991; Norris and Schrimpf, 1994; Beucher et al., 2003); FH (Figure 1F) (Schenkman et al., 1986), gp58/68 (Figure 1G) (Fischer et al., 1988) and C2 receptor inhibitor trispanning (CRIT) (Figure 1H) (Cestari Idos et al., 2008). The molecular inhibitory mechanisms of these proteins are only partially known for some of these molecules, but the central role is in the inhibition (decay) of C3 and/or C5 convertases. The inhibition of these key enzymes may have important biological consequences, such as: (i) inhibition of complement-mediated lysis, (ii) decrease in C3a and C5a (anaphylatoxins) generation (these small complement fragments are essential in the recruitment of blood cells to the infection site), and (iii) decreased opsonization, which mediates phagocytosis of pathogens during the infection (Cestari et al., 2012).

Trypomastigotes cover their surface with sialic acid residues taken from the host by TS (Frasch, 2000). Sialylated molecules downregulate AP activation (Pangburn et al., 2008; Zipfel and Skerka, 2009). FH, the AP negative regulatory protein, could also play a role on parasite complement resistance. FH is composed of 20 SCR domains (Pangburn, 2000) with different functions. FH possesses three sites (Sharma and Pangburn, 1996) that interact with unique domains on C3b (Alsenz et al., 1985; Lambris et al., 1988; Sharma and Pangburn, 1996; Jokiranta et al., 2000), participating in: (i) decay-accelerating activity for the AP C3/C5 convertases, (ii) cofactor for FI activity, and (iii) interaction with sialic acids (Alsenz et al., 1984; Kuhn et al., 1995; Pangburn, 2000). T. cruzi possesses TS to transfer α(2-3)-linked sialic acid from serum glycoconjugates to acceptor sites on the parasite surface (Schenkman et al., 1991; Tomlinson et al., 1994). These glycoconjugates restrict complement activation on the microbes surface, and are considered a virulence factor (Edwards et al., 1982; Marques et al., 1992; Wetzler et al., 1992; Edwards et al., 1993). This polyanion on the trypomastigote surface may also be critical for parasite survival in the circulation (Kipnis et al., 1981; Sher et al., 1986). In parallel, FH binds to T. cruzi, with higher affinity to metacyclic trypomastigotes than to epimastigotes (Schenkman et al., 1986). Since, FH binds to trypomastigotes covered by sialic acid via TS action, it could participate as an important complement regulatory protein to control the AP activation (Figure 1F).

Plasma membrane-derived vesicles play important roles in several infectious- and non-infectious diseases and stress (Combes et al., 1999, 2005; Al-Nedawi et al., 2008; Bebawy et al., 2009; Faille et al., 2009; Ansa-Addo et al., 2010). Bloodstream and ECs release PMV from their cellular membranes (Gasser et al., 2003; Pilzer et al., 2005; Coltel et al., 2006; Ansa-Addo et al., 2010). T. cruzi trypomastigotes are exposed to PMVs from several cell types. T. cruzi induces an increase in PMV released from blood cells in a Ca²⁺-dependent manner. T. cruzi surface molecules such as gp82 and oligopeptidase B induce a transient increase of intracellular Ca²⁺ in host cells, which could induce PMV release. These vesicles bind to C3 convertase, inhibiting its catalytic activity of both CP and LP (Figure 1A). The molecules on the PMVs interacting with C3 convertases are still unknown (Cestari et al., 2012). On the other hand, TGF-β-bearing PMVs promote host cell invasion via the lysosome-independent route. This phenomenon is dose- and parasite- infective stage dependent and non-specific for parasite strains or host cell types (Cestari et al., 2012). In agreement with this TGF-β function, chronically infected patients have higher levels of circulating TGF-β (Araújo-Jorge et al., 2002). Thus, TGF-β-bearing PMVs could activate the TGF-β signaling pathway to promote parasite infectivity (Cestari et al., 2012). Similar PMVs-mediated mechanisms have been described to other infectious agents such as HIV (Mack et al., 2000) and Plasmodium sp. (Faille et al., 2009). At the same time, T. cruzi produces exosomes that stimulate different host cells to produce PVMs to modulate the host immune response (Cestari et al., 2012; Mantel and Marti, 2014). The T. cruzi exosomes contain several surface components, like glycoproteins, such as gp85/transialidases, alphaGal-containing molecules, proteases, cytoskeleton proteins, mucins, and associated to GPI-anchored molecules.

Our laboratory has been working for over 20 years with TcCRT, a multifunctional ER-resident protein that the parasite translocates to the external milieu, mainly via the area of flagellum emergence. Although TcCRT is located mainly in the ER, histochemical studies have found it also in the Golgi, reservosomes, flagellar pocket, cell surface, cytosol, nucleus and kinetoplast (Ferreira et al., 2004; Souto-Padron et al., 2004; Gonzalez et al., 2015). However, the mechanisms involved in these heterogeneous TcCRT locations are still unknown.

Similar to its human counterpart HuCRT, TcCRT is a key protein for T. cruzi. Thus, epimastigotes bearing a monoallelic deletion of the TcCRT gene present a reduced rate of differentiation to trypomastigotes in an in vitro assay (Sanchez-Valdez et al., 2013).

In spite of the long evolutionary distance, TcCRT shares 50% homology with HuCRT (Ramirez et al., 2011a), a molecule with more than 40 functions described (Eggleton et al., 2016). HuCRT has three domains: globular N-terminus (N), proline-rich (P), and acidic C-terminus (Michalak et al., 2009). An S-domain (aa 160–283), within N and P, interacts with C1q (Stuart et al., 1996, 1997), the first component of the CP. TcCRT shares these domains and is involved in a variety of host-pathogen interactions, such as immune evasion, T. cruzi infectivity and tumor growth inhibition (Figure 2).

Trypanosoma cruzi has developed different adaptive strategies to avoid destruction by the host immune system, in particular by complement. In many conditions, the parasite has acquired these abilities only at the stage of development that are infective for the host (trypomastigotes and amastigotes). Thus, the non-infective forms of the parasite are susceptible to the complement in the presence of human serum while trypomastigotes are highly resistant. Recently, it has been demonstrated that this resistance varies also between strains. The complement system is a physiological barrier during infection which the infective forms must evade in early stages. These infective forms express on their surfaces different complement regulatory proteins or capture them from the host plasma to inhibit the complement system. Most of these proteins have a decay-accelerating activity, inhibiting the C3 convertase formation (such as T-DAF, CRP, gp58/68 and FH) and/or participate as a cofactor for FI (such as FH). Other components inactivate the catalytic activity of C3 convertase (PMV), or bind and inactivate specific complement components (TcCRT and CRIT). Some of these proteins are multifunctional. Thus, TcCRT disrupts the initial step of CP and LP, reducing the C3 convertase formation. However, TcCRT is also an important virulence factor in association with C1. Thus, the TcCRT/C1 interaction on the parasite surface promotes infectivity in vitro and in vivo. On the other hand, native endogenous parasite TcCRT inhibits tumor development in infected animals, since this effect is fully reproduced by recombinant TcCRT and reversed by specific antibodies. This role is in agreement with previous reports indicating a possible protective role of T. cruzi infection in cancer development.

GR-T and AF designed and performed experiments, interpreted data, generated key reagents, wrote, revised, edited and approved the manuscript. GR-T designed the figures.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.