Trypanosoma musculi Infection in Mice Critically Relies on Mannose Receptor–Mediated Arginase Induction by a TbKHC1 Kinesin H Chain Homolog Free

Parasites have developed various strategies to escape the immune response and take advantage of host growth factor production. Therefore, parasites need to reduce the production of toxic molecules, including NO and its derivatives, that are synthesized by the immune system, in particular macrophages (1–3).

Macrophages are activated and programmed into different functional subtypes in response to signals delivered by the microenvironment. Distinct macrophage-activation states have been widely acknowledged, with classically activated and alternatively activated states as extremes (4). Classically activated macrophages produce effector molecules (reactive oxygen species, NO) and inflammatory molecules (TNF-α, IL-1β, IL-6). The various forms of alternatively activated macrophages generally exhibit high levels of mannose receptor (MR) and arginase 1 upregulation (4). In trypanosomiasis, macrophages are characterized by the induction of arginase (5–7). Arginase hydrolyzes l-arginine to l-ornithine and urea, decreasing the intracellular pool of l-arginine and, thus, also reducing trypanotoxic NO production by NO synthase 2. Moreover, l-ornithine is a precursor for the synthesis of polyamines and trypanothione, which are essential for trypanosomatid survival (8). Thus, increased arginase activity is likely to counteract the host immune response and favor parasite growth.

In human African trypanosomiasis caused by Trypanosoma gambiense, arginase serum levels increase and return to control values after treatment (9). Accordingly, in experimental murine infection by Trypanosoma brucei, host macrophage arginase activity represents a marker of susceptibility to the disease (10). In T. brucei, arginase activity is induced by excreted/secreted factors (ESFs); in particular, TbKHC1, a kinesin H chain, has been identified as the arginase-inducing factor (11). Arginase induction in macrophages was shown to promote infection by providing polyamine nutrients for trypanosomes, which cannot generate their own source of l-ornithine through the activity of a functional arginase enzyme (11, 12).

We investigated the role of arginase activity induction during a natural murine trypanosomiasis caused by Trypanosoma musculi. In this infection, blood parasitemia lasts 20–25 d, comprising a growth phase, a plateau, and an elimination phase. Although the control of the dynamics of T. musculi infections is, in part, nonimmunological, elimination of this parasite requires Abs and macrophages (13–16). Yet, the immune mechanisms controlling early T. musculi proliferation are poorly characterized. Kidney forms persist during all host life. Blood parasites are not eliminated in T cell–deprived mice, blood parasite evolution is modified by nutrition and, curiously, previously infected pregnant mothers quickly became parasitemic (13, 17, 18). T. musculi modulates the host immune response in coinfection with various infectious agents (19). T. musculi can be considered an appropriate model of a host–parasite relationship. Moreover, T. musculi coexpress several receptors binding the Fc regions of IgM, IgG, and IgE, with the receptor common to IgM and IgE being of particular interest (20). The unsuspected immunomodulatory role of an orphan kinesin, TbKHC1, in mice infected with T. brucei, a bovine parasite, led us to investigate the presence and the role of a homolog of this kinesin in this natural murine trypanosome and to examine the presence of homologs in other trypanosomes.

Female Swiss mice, aged 6–8 wk, were purchased from Charles River Laboratories (Saint Germain Nuelles, France). SIGN-R1–knockout (KO) and congenic control (BALB/c background), C57BL/6 (B6), and MR C-type 1 (Mrc1)-KO B6 mice were bred in the Laboratory of Molecular Parasitology, Université Libre de Bruxelles.

Experiments, maintenance, and care of mice complied with the guidelines of the European Convention for the Protection of Vertebrate Animals used for Experimental and other Scientific Purposes (CETS no. 123). Experiments were approved by the Department for the Protection of Animals and Plants of the Préfecture de la Gironde (Identification no. A33-063-324) and the Ethical Committee for Animal Experiments at the Université Libre de Bruxelles (laboratory accreditation no. LA2500482).

d-Mannose, d-galactose, S-(2-boronoethyl)-l-cysteine (BEC), l-arginine, l-ornithine, protease inhibitor mixture, α-isonitrosopropiophenone, DEAE cellulose, Triton X-100, urea, Tris, propidium iodide, and MnCl₂ were obtained from Sigma-Aldrich Chemie (Saint-Quentin Fallavier, France). TRIzol reagent was obtained from Invitrogen (Illkirch, France), FITC anti-mouse CD206 (Mrc1) was from BioLegend (San Diego, CA), allophycocyanin-rat anti-mouse CD14 was from BD Biosciences (Le Pont de Claix, France), and PE–anti-mouse CD209 (SIGN-R1) mAb and control Ab were from Miltenyi Biotec (Bergisch Gladbach, Germany). For cell culture and RNA analysis, all reagents were obtained from Life Technologies (Villebon-sur-Yvette, France).

T. musculi (Partinico II strain) was originally obtained from The London School of Hygiene and Tropical Medicine (21). Parasites were stored in liquid nitrogen or maintained in vivo by a peritoneal injection (5 × 10⁴ parasites per mouse) into naive Swiss mice. When required, animals were treated i.p. with 0.2 ml of physiological saline, mannose (50 mM), or galactose (50 mM) three times a day, beginning day 1 postinfection. Mice were also given anti-TbKHC1 Ab (10), an IgG2b isotype control (250 μg in 0.5 ml of saline), or physiological saline (0.5 ml). Parasitemia was monitored by tail blood puncture from days 5 to 24 of infection.

Parasites from 10-d–infected mice were purified from blood by chromatography on DEAE cellulose (22). Soluble parasite ESFs were prepared in a protein-free culture medium, according to a previously described method (23). Purified parasites (2 × 10⁸ per milliliter) were incubated in this medium for 2 h at 37°C. Parasite viability was initially assessed by microscopic examination and then by flow cytometry after propidium iodide staining. Supernatant containing ESFs was separated from trypanosomes by centrifugation (1000 × g, 10 min, 4°C) and filtered using a 0.22-μm low-binding protein filter (Millipore, Molsheim, France). Protein concentration was estimated by the Bradford method, using a kit from Bio-Rad (Marnes-la-Coquette, France).

SDS–PAGE was carried out under nonreducing conditions in 10% polyacrylamide gels (24) and subsequently stained with silver. Protein samples were transferred to a nitrocellulose membrane (Amersham, Buckinghamshire, U.K.) and subsequently blocked, as previously described (25, 26). The membrane was incubated with mouse mAbs at a 1/100 dilution at 37°C for 2 h. After three washings, the membranes were immersed in peroxidase-conjugated anti-murine Ig (Dako, Les Ulis, France), diluted at 1:10,000, at room temperature for 2 h. After washing, membranes were exposed to Clarity ECL reagent (cat. no. 170–5061) at room temperature for 2 min and visualized using a ChemiDoc MP (cat. no. 170–8280; both from Bio-Rad). Band intensities were detected and quantified using Image Lab 5.0 software (Bio-Rad).

The T. musculi KHC sequence was amplified from gDNA with primers designed from the alignment of KHC sequences of T. brucei (Tb927.6.4390) and T. theileri (provided by Dr. Ivens and Prof. Matthews). The resulting fragment, corresponding to nt 275–2780 of T. brucei, was fully sequenced after subcloning in Zero BluntTOPO vector (Invitrogen). Primers specific for the T. musculi sequence were then designed to amplify the 5′ and 3′ extremities of the TmKHC transcript from total RNA by 5′ and 3′ RACE (First Choice RLM-RACE Kit; Ambion, Merelbeke, Belgium). The resulting PCR fragments were cloned into the Zero BluntTOPO vector and sequenced. The following primers were used: TthKHC forward, 5′-AATTTTGATGGTGCGTTTG-3′; TthKHC reverse, 5′-ATTGACATATAATCATCCATTGCAAT-3′; TmKHC 5′ reverse out, 5′-AAACGCGTGGTCAACGCACG-3′; TmKHC 5′ reverse in, 5′-GCAACACTCTCGAAGATCTGC-3′; TmKHC 3′ forward out, 5′-GTGAACTTCGGGAAATAATTGCGTCAGC-3′; and TmKHC 3′ forward in, 5′-GTCAGCGCATTTATCTGTTCCGGAATC-3′.

Peritoneal macrophages were cultured in 96-well plates (10⁵ per well) in RPMI 1640 medium supplemented with 2% FBS, 20 mM HEPES, 2 mM sodium pyruvate, 10⁻⁴ M 2-ME, 10 μg/ml gentamicin, and 3% glucose in a humidified 5% CO₂ atmosphere at 37°C, as previously described (4). Viability, as determined by trypan blue dye exclusion, was >98%, and May–Grünwald–Giemsa and nonspecific esterase staining revealed that >96% of the cells were macrophages. ESFs (10 μg/200 μl), BEC (100 μM), mannose and galactose (up to 50 mM), l-ornithine (2.5 mM), and Abs (3 μg/200 μl) were added to the culture medium when required.

The macrophage culture medium was removed 48 h later to obtain a macrophage-conditioned medium (MCM) (27). A medium, set under the same conditions but not exposed to macrophages, was used as the control medium. After collection, these media were centrifuged at 1000 × g for 5 min and stored at −80°C until use.

Macrophages from naive mice (10⁵ per well) were cocultured with purified T. musculi (400 per well) in culture medium (28). Parasites were counted after 3 d in coculture.

Arginase activity was measured in macrophages cultured with T. musculi or ESFs for 48 h or in ex vivo macrophages from uninfected mice or infected animals at days 7, 10, 15, 20, and 24 postinfection. Arginase activity was measured by a colorimetric method, as previously described (29). Briefly, 10 mM MnCl₂ and 0.5 M l-arginine were added in turn to macrophage lysates and incubated at 37°C for 1 h. The reaction was stopped by adding an acid solution, and the urea formed by arginase was revealed by adding α-isonitrosopropiophenone at 100°C for 45 min. The colored product was quantified by absorption at 540 nm.

Mouse peritoneal macrophages were plated in 60-mm dishes at a density of 2 × 10⁶ cells. Total RNA was extracted using TRIzol reagent. Total RNA (2 μg) was retrotranscribed using the Transcriptor First Strand cDNA Synthesis Kit, according to the manufacturer’s instructions. Amplification of specific PCR products was detected using FastStart Universal SYBR Green Master (ROX) (Roche, Basel, Switzerland) and the following primers: Arg1, 5′-AAGAAAAGGCCGATTCACCT-3′ and 5′-CACCTCCTCTGCTGTCTTCC-3′, 201 bp; Arg2, 5′-ACAGGGTTGCTGTCAGCTCT-3′ and 5′-TGATCCAGACAGCCATTTCA-3′, 298 bp. PCR was performed using an ABI PRISM 7900HT (Applied Biosystems, New York, NY). Fold differences in gene expression were calculated as 2−ΔΔ cycle threshold using β-actin as the housekeeping gene (10).

Peritoneal cells from T. musculi–infected mice were removed at 13 d postinfection. Peritoneal cells from noninfected mice were used as control. Cells (10⁶ per milliliter) were incubated with allophycocyanin–anti-mouse CD14 and FITC–anti-mouse CD206 (MR), PE-anti-mouse CD209 (SIGN-R1), or the corresponding control Abs. After 20 min, cells were washed and resuspended in PBS containing 0.5% FCS. Cells were analyzed by flow cytometry (BD FACSCanto and FACSDiva software version 6.1.2).

The protein sequences of T. musculi, T. brucei, and Trypanosoma theileri KHC were compared with homologs in GenBank using the algorithm “protein BLAST” (Basic Local Alignment Search Tool) from the National Center for Biotechnology Information (https://blast.ncbi.nlm.nih.gov/Blast.cgi). To reconstruct a phylogenetic tree, we selected the sequences from the nonredundant protein sequences database according to three criteria: identity >50%, query cover >50%, and the sequence species were previously characterized as T. brucei, Trypanosoma vivax, Trypanosoma equiperdum, Trypanosoma congolense, Trypanosoma evansi, Trypanosoma cruzi, T. theileri, and Trypanosoma grayi and their subspecies. MEGA6 (30) was used for sequence alignment and tree reconstruction. The best amino acid substitution model was estimated by the Bayesian Information Criterion and used to build a maximum likelihood tree itself tested by the bootstrap method with 100 replications.

Quantitative data are expressed as mean ± SEM. The Student t test was used for comparison between two groups. Prism 5.0 (Graph Pad, La Jolla, CA) was used for analyses.

Arginase activity was measured in macrophages from infected mice at different days postinfection. This activity peaked between days 13 and 17 and paralleled the number of circulating parasites (Fig. 1A) (regression analysis between arginase activity and parasitemia in Supplemental Fig. 1).

Levels of Arg1 and Arg2 mRNA expression in peritoneal macrophages were determined by RT-PCR. The level of Arg1 was ∼12-fold higher in macrophages from mice 13 d postinfection with T. musculi compared with macrophages from uninfected mice, whereas Arg2 expression was not induced (Fig. 1B).

In macrophages cocultured with T. musculi, arginase activity increased compared with controls (Fig. 2). Arginase activity was still induced, even after separation of parasites from macrophages by a 0.22-μm filter support. Finally, arginase activity induction was also observed in macrophages cultured with ESFs (Fig. 2).

T. musculi was grown in the presence of macrophages or in a 2-d parasite-free MCM (27, 28). Adding BEC, an arginase inhibitor, to the parasite–macrophage culture dramatically reduced parasite growth, which was restored by adding l-ornithine (Fig. 3). MCM promoted parasite growth, whereas control medium had no effect. Addition of BEC to MCM just prior to incubation with parasites had no effect on parasite growth, indicating that BEC had no direct effect on parasites (data not shown).

Using Western blot, a mAb directed to TbKHC1, a previously described arginase-inducing factor from T. brucei (11), recognized a band in T. musculi ESFs corresponding to the molecular mass of TbKHC1 (122 kDa) (Fig. 4).

To identify a putative T. musculi KHC gene, we first aligned the TbKHC1 sequence with the T. theileri homolog, which is more closely related to T. musculi. Two primers were designed in conserved regions to amplify a 2505-nt fragment from T. musculi gDNA. Sequencing of this fragment showed 70 and 63% identity with TbKHC1 at the DNA and protein levels, respectively. This sequence was used to design two sets of primers for amplification of the 5′ and 3′ ends of the T. musculi gene by retrotranscription of total RNA, followed by nested PCR. Assembly of these 5′ and 3′ sequences with the first PCR product revealed that the TmKHC gene encoded a protein of 1145 aa that had 59% identity with TbKHC1 at the protein level (Fig. 5). The protein N-terminal domain contained the highly conserved kinesin motor domain with a typical ATP binding site.

When anti-TbKHC1 mAb was added to T. musculi–macrophage cocultures, arginase induction and parasite growth were markedly reduced compared with the isotype-control Ab (Fig. 6A).

In addition, when anti-TbKHC1 Ab was injected i.p. into mice on day 2 post–T. musculi infection, the parasite load strongly decreased, whereas isotype-control Ab treatment had no effect (Fig. 6B). Arginase activity also decreased in macrophages from 13-d infected mice in the anti-TbKHC1 Ab–treated group (38 ± 11 mU per 10⁵ macrophages) compared with those from the isotype-control group (117 ± 20 mU per 10⁵ macrophages) or the saline group (125 ± 18 mU per 10⁵ macrophages).

In cells obtained from the peritoneal cavity of an unmanipulated mouse, ∼50–60% are B cells, ∼30% are macrophages, and 5–10% are T cells (31). In murine trypanosomiasis, C-type lectin receptor expression increases in macrophages (6). The number of peritoneal macrophages was increased in T. musculi–infected mice, and granulocytes did not exceed 1% (14). In the peritoneal cavity from 14-d infected mice, CD14^high cells have been considered macrophages, and expression levels of CD206 (Mrc1) on these cells (Fig. 7) increased 4-fold compared with control macrophages (mean fluorescence intensity, 1987 ± 89 versus 476 ± 51), whereas CD209b (SIGN-R1) expression was unchanged (mean fluorescence intensity, 685 ± 102 versus 648 ± 91).

Mannose or galactose was added to parasite–macrophage cocultures. Mannose inhibited parasite growth in a dose-dependent manner, whereas galactose had only a marginal effect (Fig. 8). Supplementing MCM with mannose or galactose just before adding it to purified parasite cultures had no effect on parasite growth, indicating the absence of a direct toxic effect of sugars on the parasites (data not shown).

In addition, mannose had a more marked inhibitory effect than galactose on parasite- and ESF-induced arginase activity (Fig. 8B).

Intraperitoneal injection of mannose from day 1 post–T. musculi infection dramatically decreased the parasite load in mice, whereas galactose had a less pronounced effect (Fig. 9). At 14 d postinfection, macrophage arginase activity had decreased in mice receiving mannose (34 ± 15 mU per 10⁵ cells) compared with infected mice receiving saline (122 ± 17 mU per 10⁵ cells) or galactose (88 ± 21 mU per 10⁵ cells).

T. musculi parasitemia was reduced 10-fold in Mrc1-KO mice compared with wild-type mice, whereas parasitemia decreased marginally in SIGN-R1 KO mice compared with controls (Fig. 10).

Only eight GenBank sequences fulfilled the selection criteria (i.e., >50% homology and >50% cover from sequences belonging to the selected species): one T. brucei gambiense, one T. equiperdum, one T. congolense, one T. vivax, one T. grayi, two T. cruzi, and one Trypanosoma cruzi marinkellei. There was no protein sequence of KHC for T. brucei rhodesiense, T. evansi, or T. theileri. The protein sequences of T. brucei, T. musculi, and T. theileri KHCs, of 1111, 1148, and 1123 aa, respectively, were aligned with these eight GenBank sequences; after elimination of the noncommon parts at both ends, we obtained a final alignment of 1071 aa sites, including gaps. The maximum likelihood tree is shown in Fig. 11. TbKHC1 (strain TREU927 of T. brucei brucei, GenBank reference XP_845556) presented 100% identity with T. equiperdum (SCU65329), 99% identity with T. brucei gambiense (XP 011774228), and 78% identity with T. congolense (CCC91128). KHC of T. musculi presented 75–76% identity with the strains belonging to the T. cruzi clade and <66% identity for the other species, whereas KHC of T. theileri presented its maximum identity (69%) with T. grayi (XP 009307339) and <65% with the other species.

This study revealed that T. musculi ESFs induced macrophage arginase 1 activity via a pathway involving mannose-binding receptor(s), promoting parasite growth. A mAb specific to the kinesin H chain TbKHC1, an orphan kinesin present in ESFs from T. brucei, inhibited arginase activity and reduced T. musculi load in vitro and in vivo, suggesting that the TbKHC1 homolog identified in T. musculi, with 59% aa identity with TbKHC1, was the active component of T. musculi ESFs. However, although TbKHC1 facilitates T. brucei parasitemia via the SIGN-R1 receptor (11), MR was apparently the main target of T. musculi ESFs. This suggested that kinesin H chain–related proteins played similar roles in promoting infection in two distant trypanosomes, via macrophage arginase induction, following distinct C-type lectin receptor targeting.

Macrophages have been widely used as feeder layer cells to favor parasite cell proliferation in vitro (27, 28). Adding an arginase inhibitor reduced their parasite growth-promoting activity, which was restored by l-ornithine supplementation (Fig. 3), although synergy with the various components of FCS cannot be ruled out. The essential requirement for l-ornithine is linked to the absence of the functional enzyme arginase in trypanosomes (12), resulting in their dependence on host arginase activity for polyamine and trypanothione synthesis, which are essential for parasite survival, growth, and differentiation (7, 32). Difluoromethylornithine, a structural analog of l-ornithine, is used alone or, more recently, in combination with nifurtimox as an effective drug in human African trypanosomiasis (33). However, its administration is difficult and requires large quantities of i.v. injected fluids, which limits its use in remote areas. Therefore, it would be useful to find easier ways to target polyamine synthesis, including delivery of cytotoxic polyamine analogs, to improve resistance to trypanosome infection. Alternatively, inhibitors of the pathway triggering arginase activity are expected to reduce parasite load in infected animals, suggesting that strategies targeting trypanosome ESFs would be useful for increasing resistance to trypanosomiasis.

Parasites have to adapt and develop in successive hosts by modulating various host environments and defense mechanisms. Communication between host cells and parasites may take the form of cell–cell contact, soluble factors, and/or extracellular vesicles (exosomes or ectosomes) from parasites, involving membrane receptors of host cells. C-type lectin receptors, expressed in large quantities by dendritic cells and macrophages, play important roles in numerous basic phenomena, as well as various aspects of the immune response. These receptors induce highly diverse immune responses to pathogens (34). They are involved in protection against various fungi, such as Cryptococcus neoformans (35). Several pathogens exploit these receptors to suppress an efficient immune response (36). Distinct C-type lectin receptors can contribute positively and negatively to the macrophage response to Leishmania infantum, a parasite related to trypanosomes (37). MR binds to many microorganisms, primarily intracellular, that are coated with carbohydrates, including viruses, bacteria, fungi, and parasites (38). We showed in this article that T. musculi KHC1, a secreted kinesin, induced arginase via Mrc1, which is upregulated in macrophages from T. musculi–infected mice. This arginase induction was inhibited by a mAb directed to TbKHC1 (Fig. 6), as well as by mannose, an MR inhibitor (Fig. 8) (11, 39). In mice infected with T. cruzi, a parasite related to T. musculi, MR expression was upregulated in macrophages. The T. cruzi Ag cruzipain enhanced MR recycling, which favored arginase induction and T. cruzi survival in macrophages, whereas receptor blockade decreased arginase activity and parasite growth (40). All of these data highlighted the importance of arginase induction for extra- and intracellular trypanosomes and identified the parasite molecules and host receptors involved.

ESFs from different parasite strains induce macrophage arginase as a function of strain virulence (23); however, their contents are poorly known. A mAb directed to TbKHC1 reduced parasite load in T. musculi–infected mice (Fig. 6). This indicated that a homolog to TbKHC1 was present in T. musculi and possessed similar properties, such as macrophage arginase induction (Figs. 4 and 5). Moreover, sequence analysis revealed the presence of homologs of this gene in other trypanosomes (Fig. 11), including pathogens affecting humans (causing African trypanosomiasis [sleeping sickness] or Chagas disease) and animals (reducing production and increasing morbidity and mortality in livestock). It is worth noting that the hierarchical structure of the KHC phylogenetic tree was consistent with previous reports on other markers: monophyly of T. cruzi and T. brucei/T. equiperdum and intermediary positions for the others species making KHC as a potential phylogenetic marker for trypanosomes (41, 42).

The dysregulation of host l-arginine–inducible metabolism by ESFs represents an effective parasite mechanism to hamper host immune response and modify host molecule production to favor parasite invasion and growth. Preventing this host metabolism dysregulation by immunizing against active components in ESFs or locking partner host molecules is a promising approach. A selective targeting of TbKHC1 kinesin H chain homologs in trypanosomiasis deserves further investigation. Antigenic variations in surface glycoproteins hamper vaccination against African trypanosomiasis. Conserved epitopes in nonvariant Ags may be selected and combined and then associated with an adjuvant in an appropriate vaccine formulation to induce adequate immune responses (43, 44).

We thank Patricia Nabos for animal care, Vincent Pitard and Santiago Gonzalez (cytometry platform, University of Bordeaux), and Dr. Alasdair Ivens and Prof. Keith Matthews (University of Edinburgh) for providing T. theileri genome sequence information.

The authors have no financial conflicts of interest.