Description of an Intracellular Stage in the Experimental Infections of Albino Mice with a Venezuelan Isolate of Trypanosoma evansi Using Scanning and Transmission Electron Microscopy Techniques

Abbreviations: AAT: African Animal Trypanosomosis; VAT’s: Variant Antigen Types; CNS: Central Nervous System; Isolated EcHF91: Isolated from Equus caballus at the Hato el Frío (Venezuela) on 1991; TEM: Transmission Electron Microscopy; SEM: Scanning Electron Microscopy; PCR: Polymerase Chain Reaction; PBS: Phosphate buffered Saline; NMRI-IVIC: Naval Medical Research Institute-Instituto Venezolano de Investigaciones Científicas; IP: Intraperitoneal; OS: Overall Survival; Pa: Pascal; CO2: Carbon dioxide; Tryp: Trypanosomes; SER: Smooth Endoplasmic Reticulum; RER: Rough Endoplasmic Reticulum; HhHF92: Isolated from Hydrochoerus hydrochaeris at the Hato el Frío (Venezuela) on 1992; EcM96: Isolated from Equus caballus at Mantecal (Venezuela) on 1996; kDa: Kilodalton; CHO: Chinese Hamster Ovary cells; PIP3: Phosphoinositol Triphosphate; Ca2+: Calcium cation; ITS1: Internal Transcribed Spacer 1; TEVA1: Trypanosoma evansi strain 1; LDLr: Low Density Lipoprotein receptor; FONACIT: Fondo Nacional de Ciencia, Tecnología e Innovación; IBE-UCV: Instituto de Biología Experimental-Universidad Central de Venezuela; DSc: Doctor of Science.


Introduction
Trypanosomosis is a complex of diseases that affect man and his livestock in Africa, Asia, Central and South America. The African Animal Trypanosomosis (AAT) caused by Trypanosoma brucei, Trypanosoma evansi, Trypanosoma vivax, Trypanosoma congolense and Trypanosoma simiae is probably the more serious problem for etiological agent of equine trypanosomosis known as "Derrengadera" [4] with high seroprevalence rates especially in geographic locations where it is enzootic such as the Apure State (81.7%) [5]. The transmission of the infection is carried out by biting flies of Tabanidae family [6][7][8], affecting horses, donkeys, bovines [9,10] and wild animals as capybara (Hydrochoerus hidrochaeris) in which it causes a subpatent or asymptomatic infection [11]. Recently, human infections caused by T. evansi have been reported in India [12,13] and southern Vietnam [14]. Although these infections are associated with a lack of apolipoprotein L-I, the number of cases could be underestimated. For these reasons, the evaluation and improvement of new diagnostic tests and field researches are required for detection and confirmation of these atypical cases [15].
Infections caused by T. evansi in domestic animals can be acute with fever, emaciation, anemia and sudden death in some cases or chronic, with a progressive weakness, emaciation and a notable decrease in the productive capacity [16]. The pathology of T. evansi infections has been extensively studied by different authors in experimental and field conditions [17][18][19][20][21][22][23]. As a consequence of its location inside host body, T. evansi has been included together with T. brucei, T. brucei rhodesiense, T. brucei gambiense, T. equiperdum and T. cruzi, within a trypanosomegroup whose infection occurs in the blood plasma, but may also occur extravascularly and in certain body fluids (cerebrospinal fluid, peritoneal and synovial liquids and aqueous humor), triggering not only anemia, but also extensive inflammatory, necrotic and degenerative processes [24].
Our electron microscopic studies about the ultrastructural pathology of the murine infections with T. evansi were able to detect trypomastigotes in the blood vessels of all tissues studied in adrenal cortex [25], skeletal muscle [26], liver [27], spleen, kidneys and testes, as well as in the connective tissues of liver and adrenal gland. The extravascular presence of T. evansi should not constitute a new surprising fact, because the very first extravascular observation of trypanosomes from brucei-group (T. brucei, T. brucei rhodesiense, T. brucei gambiense, T. evansi and T. equiperdum), was made in 1912 by Wolbach and Binger, who described that the infections with these trypanosomes cause severe damages in the perivascular connective tissue, characterized by disorganization and breaking of collagen bundles, as well as fibroblasts destruction [28].
Luckins et al. [19] recognize the presence of T. evansi in the connective tissues of the skin from rabbits experimentally infected and propose that the extravascular location of the trypanosomes at the site of inoculation should be responsible for the aggressive inflammatory response, with collagen destruction, edema and necrosis. In addition, they proposed that the extravascular growth of the trypanosome populations and the consequent increase in the number of variant antigen types (VAT´s), would allow the colonization of the bloodstream by T. evansi, as a consequence of the evasion of host immune system through the mechanism of antigenic variation.
Similar observations were described by Uche and Jones [20] in the spleen, lymphatic nodes, vulva and ears of experimentally infected rabbits, and Biswas et al. [23] in the liver of Bandicoot rats. These results confirm that the presence of extravascular trypomastigotes in kidneys, heart, lungs, liver, and central nervous system (CNS), instead to constitute indicators of a protective immune response, constitute a destructive response capable to cause hosts death. The abundance of these extravascular forms of T. evansi during the experimental and natural infections besides being responsible for the higher mortality for deer (Cervus porcinus) [18,22], could be associated to the genetic predisposition of the host to the infection and the development of an inadequate immune response among other factors As it has also been described for T. brucei, the extravascular presence of T. evansi could be a consequence of the ultrastructural damage observed in the capillary, arterioles and veins of the infected animals. Damage to the vasculature described for the infection of the adrenal glands [25], skeletal muscle [21,26,29] and liver [27] with the same Venezuelan isolate of T. evansi , were severe enough to allow the infiltration of blood cells, trypomastigotes but also biologically active molecules secreted by live T. evansi trypomastigotes [30] or from dead trypanosomes into tissues and extravascular spaces.
The main contribution of this work is the confirmation that the occurrence of intracellular stages in experimental infections of albino mice with the Venezuelan isolated EcHF91 of T. evansi is more frequent than what was initially reported in the preliminary works of our group [25][26][27]29]. In this regard extensive observations of a greater number of ultrathin sections allowed us to verify that cell invasion is not only limited to cells of the adrenal cortex, hepatocytes and vascular endothelium, but also to cells of the immune system such as plasma cells and lymphocytes. In addition, the ultrastructural and morphological findings obtained by Transmission Electron Microscopy (TEM) and Scanning Electron Microscopy (SEM) provide suggestive evidence on the occurrence of an active invasion process as well as the confirmation that bloodstream and intracellular trypomastigotes share the same ultrastructural features. The results are discussed in the context of the impact of these stages on the diagnosis-treatment of T. evansi trypanosomosis, its evasion of immune system, the persistence of this hemoparasite in natural hosts as well as its seroprevalence and detection by the polymerase chain reaction technique (PCR) in clinical and parasitological healthy animals.

Trypanosoma evansi
The population of T. evansi used in this work was derived from a heterogeneous line originally isolated in 1991 from an infected horse (Equus caballus) at the Hato El Frío (7° 56' North and 68° 57' West) in the Apure State of Venezuela (Ec/1991/Faculty of Veterinary Sciences-UCV/Stock 3/IBE). This heterogeneous population (EcHF91) was classified as T. evansi according to the clinical manifestations, its morphology, parasitological behaviour and ultrastructural feature (low number of coated vesicles in the flagellar pocket in relation to T. equiperdum a co-endemic specie of trypanosome) [31]. It was maintained in the laboratory through passages in albino rats and cryopreservation at -196°C, using 1% glucose PBS pH 7.4 with 10% (v/v) dimethyl-sulphoxide as cryoprotector.

Experimental design and infections
Two groups of 10 heterozygous male mice (NMRI-IVIC) with a body weight of 20-25 g were used for the electron microscopic and parasitological studies. They were kept in metal cages, using rice-shell as a bed in the animal facility of the Laboratory of Immunology and Chemotherapy (Instituto de Biología Experimental-UCV). They were fed on Ratarina® and ozonized water ad libitum, and experimentally infected by intraperitoneal (IP) inoculation with 0.1 ml of diluted rat infected blood containing 10 4 trypanosomes (400 trypanosomes/g of body weight). Parasitaemia of both groups was daily checked as described by Brener method [32] and 6 animals for electron microscopic studies were sacrificed on day 5 post-infection under Ketamine anesthesia (80-120 mg/kg IP), when they were dying and parasitaemia reached levels of approximately 10 9 trypanosomes/ml. A third group of healthy mice (n=4) were also sacrificed and used as controls.
The overall survival (OS) of groups was defined as the number of live mice present in the cages at each day. All the dying mice (in a convulsing and prostrated state with nose-bleeding) present in cages were euthanized under anesthesia as previously described in order to avoid their suffering.

Statement of animal rights
During experimental infection, all animals were maintained under veterinary supervision to safeguard health and minimize animal suffering. Protocols used were approved by the Ethical Committee for Laboratory Animal Use under number 013-11 according to the Ethics Chart of animal experimentation.

Transmission electron microscopy
Pieces of adrenal glands, liver and spleen (2 mm 3 each) were fixed with Karnovsky´s solution (2.5% glutaraldehyde, 4% p-formaldehyde in 0.1 M Millonig phosphate buffer pH 7.4) and 1% osmium tetroxide, dehydrated in increasing ethanol concentrations and embedded in LX-112 epoxy resin (Ladd Research Inc., Burlington) according to Rossi et al. [25]. Ultrathin sections were obtained with a diamond knife in a Porter-Blum MT2-B ultramicrotome, stained with uranyl acetate and lead citrate [33], and examined with Hitachi models H500 and H7100 transmission electron microscopes with an acceleration voltage of 75-100 kV.

Scanning electron microscopy
Pieces of the aforementioned organs were fixed and freeze fractured. Briefly, tissue pieces were fixed with Karnovsky´s solution, cryoprotected by incubation with 25% (v/v) glycerol in 0.1 M Millonig phosphate buffer pH 7.4 for 45 minutes at ambient temperature and quickly frozen under super cooled isopropane (-196°C). Frozen samples were freeze-fractured with a steel knife in a JEOL 9010C freeze-fracture device (1 × 10 -6 Pa, -120°C), dried by evaporation on a critical point dryer (Hitachi HCP-2) using carbon dioxide (CO 2 ) as liquid transition fluid, mounted on a sample holder and shadowed with platinum ions (Electron Microscopy Science EMS-350) [34]. Microscopic observations were done with Hitachi model S800 scanning electron microscope with an acceleration voltage of 5 kV.

Results
The course of parasitaemia of mice experimentally infected with T. evansi is shown in Figure 1. The prepatent period was 2 days and maximal parasitaemia (1.26 × 10 9 Tryp/ml) was reached when the overall survival of both experimental groups was 10% (1 out of 10 mice).  Transmission electron micrographs obtained from hepatocytes of healthy mice (Figure 3), showed the normal ultrastructural feature of cells with high metabolic function as judged by the abundance of mitochondria, rough endoplasmic reticulum (RER) and SER, lysosomes, glycogen particles and lipid stores (lipid droplets) ( Figures 3A and 3B). Normal bile canaliculi between hepatocytes were also seen ( Figure 3B).
White pulp from control mice ( Figure 4) showed numerous lymphocytes (characterized by their high ratio nucleus/cytoplasm), in close contact between them and macrophages, while the spaces in red pulp (results not showed) contained lymphocytes, reticular cells, erythrocytes and platelets.  Figure 11B). No intracellular trypomastigotes were observed in adrenal medulla (results not showed). In the liver, intracellular trypanosomes were detected inside the cytoplasm of normal ( Figure 12) and necrotic hepatocytes (Figures 13 and 14).
The presence of intracellular trypomastigotes was not limited to     the adrenal cortex and liver parenchyma; they have been also detected less frequently inside plasma cells (Figure 15), endothelial cells of blood vessels from adrenal cortex ( Figure 10A) and interacting with large activated spleen lymphocytes.
These intracellular stages of T. evansi showed to have the same ultrastructural feature of bloodstream trypomastigotes, with the exception of some forms observed in liver ( Figure 14). These forms showed an ultrastructure that resembles the epimastigote stage of African trypanosomes, as judged for the position of the kinetoplast in relation with the anterior body end. Additionally, intracellular Although in most of the electron micrographs herein presented ( Figures 5, 6A, 6B, 7A, 7B, 8, 13 and 14), the ultrastructure of the T. evansi trypomastigotes was normal, as judged by comparisons with findings reported by Hernández-Páez [35], the ultrastructure of trypomastigotes in some electron micrographs ( Figure 10A, 11A, 11B  and 15), resembles the ultrastructural feature of T. brucei cell-death, characterized by apoptotic ultrastructural features such as swollen mitochondrion ( Figures 10A, 15A and 15B), an apparent increase in glycosome electron density ( Figures 11A and 11B), an apparent decrease on cytoplasm electron density (loss of cytoplasmic material) ( Figure 15A), and alteration of nuclear chromatin ( Figure 15B).
As it will be discussed in the next section, a working hypothesis        about a feasible mechanism for the internalization of T. evansi in these tissues is proposed based on the results herein presented, as well as results previously published by our group [25][26][27]29,30] and from other authors working with T. evansi and other trypanosome species ( Figure  16).

Discussion
The occurrence of intracellular stages of T. evansi is something relatively new. Only a letter to the editor, previous to our works describes the presence of amastigote and spheromastigote stages in the brain of a cat infected with T. evansi [36]. This communication contrasts with findings reported to the date about morphology of trypanosomes of subgenus Trypanozoon (Salivary Section). It is important to highlight that with the exception of the epimastigote-like stage that was herein presented; all of the intracellular trypanosomes detected in our work were trypomastigotes. Moreover, in all of studied samples we have not observed the presence of trypomastigotes in different stages of rolling as it has been described by Ormerod and Venkatesan [37] for T. brucei. In addition trypomastigotes were observed freely in the cytoplasm and not enclosed in a parasitophorus vacuole, as it has been described for Leishmania spp and the initial stages of infection of cardiomyocytes with T. cruzi [38].
This observation previously reported for T. brucei in the CNS of infected rats and mice [39][40][41][42], was described for the very first time for T. evansi by our group [25] as a casual finding during studies about the ultrastructural pathology in murine infected with this trypanosome.
From the year 1999 to the date, we have accumulated more substantial ultrastructural evidences about the occurrence of this phenomenon in the experimental infections of albino mice with EcHF91 isolate.
Studies performed by our group (manuscript in preparation) with other Venezuelan isolates of T. evansi, such as the isolated HhHF92 from a Venezuelan capybara (Hydrochoerus hydrochaeris) captured in 1992 at the same locality of the EcHF91 isolated, and EcM96 isolated from an infected horse at Mantecal (Apure state) in 1996, not only have demonstrated differences in the parasitological behavior, but also the inability to develop this intracellular stages in albino mice experimentally infected, results that were in agreement with studies performed with different isolates of T. evansi [43][44][45], in which it was impossible to detect the presence of intracellular trypanosomes.
Although we have found trypomastigotes of T. evansi inside necrotic spaces lacking of plasma membrane limits, we have also detected intracellular trypomastigotes inside spaces with membrane limits that show a normal looking ultrastructure in adrenal cortex cells, hepatocytes, endothelial cells and plasma cells. In these spaces trypanosomes observed, show the typical ultrastructural features of trypanosomes belonging to the subgenus Trypanozoon as it has been described for a Venezuelan isolate of T. evansi [35] and for African trypanosomes belonging to brucei-group [46]. In all of the cases trypanosomes were surrounded by mitochondria, lipid droplets, glycogen granules and RER with normal ultrastructure.
The appearance of intracellular trypomastigotes in the cytoplasm of these cells should be a consequence of mechanisms responsible for the increased vascular permeability, as well as necrotic process and large gaps seen in the endothelial walls [25,27], together with the presence of receptors to T. evansi surface ligands in the affected host cells. The mechanism through which T. evansi infects or penetrates these cells is unknown, however this event would be explainable if an active invading process mediated by parasite's biologically active molecules as well as ligands and receptors in host cells, are present.
As it can be seen in Figure 16, a working hypothesis about a feasible mechanism for the internalization of T. evansi in these tissues is proposed according to the results herein presented, and other published by our group [25][26][27][29][30], as well as from other authors working with T. evansi and other trypanosome species. It is important to clarify that the working hypothesis never intends to assert that intracellular forms described in this paper occur as described in the figure. This hypothesis should be tested through experimentation.
According to Figure 16A, in the first step, the secretion or releasing of different trypanosome enzymes (proteases, peptidases, phospholipases, etc.) to the bloodstream during the infection, could act in a synergistic way causing damages of the sinusoid walls, vascular endothelium, blood cells and tissues as it has been described for T. brucei [47] and T. vivax [48]. In addition a 29 kDa T. evansi cysteine-proteinase with gelatinolytic activity, secreted extracellulary [30] and detected as a somatic antigen in trypomastigotes of T. evansi and T. equiperdum [49], could act by removing sialic acid residues in the form of glycopeptides [50], which along with trypanosomal phospholipases can induce vasculitis and an increase of the vascular permeability [51][52][53].
On the other hands, the high levels reached by these antigens from trypanosomes in bloodstream and its reactions with specific antibodies and complement, could be responsible for the formation of immune complexes [54] to produce tissue damages and inflammation due the release of kallikrein that together with phospholipases, are capable to activate kininogen to produce bradykinin and kallidin, which in turn increases vascular permeability by combining with acidic groups from mucopolysaccharides and endothelial glycoproteins [53].
The interaction of T. evansi with the endothelial cells in addition to the necrotic process, the large gaps seen in the capillary walls [25,26] and liver sinusoids [27], and the mechanic effects associated to the motility of trypanosomes [55,56], could explain the presence of trypanosomes and blood cells in the Disse`s and sub-endothelial spaces of liver and adrenal cortex, but also the interaction of trypanosomes with the parenchymatous cells of these tissues ( Figure 16B). In the hypothetic process proposed, the interaction-invasion of host cells by T. evansi can be divided into three stages: adhesion and recognition, signaling, and invasion. As it has been described for T. cruzi entry process, many molecules that are present in the membrane of the host cells such as lectins and bradykinin receptors, could be potential partners for recognition depending on the cell type involved [38].
Lectins present in mammalian cells are represented by sugarbinding proteins, which are highly specific for their sugar moieties and are involved in attachment between pathogens and host cells [57][58][59][60]. In a second step, presence of lectins on the plasma membrane of these cells, as it has been described for hepatocytes and Küpffer cells [61] could help the adhesion of T. evansi to endothelial cells, cells of the adrenal cortex, plasma cells and lymphocytes. These evidences are supported by findings about the binding of murine hepatocytes to acrylamide sheets containing D-galactose, one of the most abundant monosaccharides on the Variable Surface Glycoprotein of the surface coat of African trypanosomes [62]. On the other hand, secretion of sialidases whose levels are marked increased in plasma of bucks experimentally infected with African isolates of T. evansi [63], could produce cleavage of surface sialic acids from host cells, exposing β-galactosyl residues which in turn could be breakdown by β-galactosidase or been recognized by β-D-galactose specific lectins on the surface of adrenal cortex cells, hepatocytes, white blood cells, macrophages and Küpffer cells from hepatic sinusoids, leading to the adhesion of T. evansi to host cells and the abnormal extravascular erythrophagocytosis observed during experimental infections of mice [64].
This first interaction could be stabilized by the specific recognition of a 120 kDa T. evansi protein, through cell surface receptors as it has described for peritoneal macrophages, lymphocytes and epithelial cells from mice kidney (Zhao et al., unpublished data), and presented at the First International Seminar on Non Tsetse-Transmitted Animal Trypanosomosis in Annecy (France) in October 14-16, 1992 [65].
During adhesion-recognition process, these interactions parasitehost cells could also represent external signals which are transduced intracellularly determining a cell signaling event with the consequent increase on intracellular calcium ions level (Ca 2+ ) ( Figure 16C) as it has been described for T. cruzi.
The intracellular calcium levels may also remain high by the effect of the 29 kDa cysteine-protease secreted by T. evansi [30]. This protease could stimulate an influx of Ca 2+ that determine the invasion of the cells, in a similar way described for the brucipain, a 30 kDa cysteineprotease released during the infection of the CNS cells and endothelial cells with T. brucei gambiense [66][67][68].
The bradykinin receptors are another class of receptors that T. evansi could use to penetrate host cells, as it has been described for the invasion of Chinese hamster ovary (CHO) cells by T. cruzi trypomastigotes. These receptors are coupled to the heterotrimeric protein G and are formed by two subtypes: the bradykinin-2 receptor, which is constitutively, expressed by cardiovascular cells and the bradykinin-1 receptor whose expression is up-regulated in injured tissues [69].
The invasion of host cells by T. evansi could be achieved by the actin dependent pathway ( Figure 16D). According to this pathway, trypomastigotes penetrate into a host cell through plasma membrane expansions and invaginations that accumulates phosphoinositol triphosphate (PIP3), a phosphoinositide which is produced by class I PI3 Kinase, as a consequence of the increase in the intracellular levels of Ca 2+ [38]. This event culminates with the assembly of a parasitophorous vacuole formed from the plasma membrane containing internalized T. evansi trypomastigotes ( Figure 16E). Subsequently, trypomastigotes could produce the lysis of the parasitophorous vacuole membrane ( Figure 16E) through the releasing of phospholipases [70], and/or the action of a membrane-bound neuraminidase that was isolated from bloodstream trypomastigotes of T. evansi [71].
The enzymatic breakdown of the parasitophorus vacuole, could lead to the presence of free T. evansi trypomastigotes in the cytoplasm of cells exhibiting different degrees of ultrastructural alterations, as it has been described in previous works ( Figure 16F) [25][26][27]29].
The intracellular stages of T. evansi could be one of the many biological characters shared by particular genotypes of trypanosomes belonging brucei-group [31,72] and capable to develop cryptic stages [37,39,41], as it has been described for their ultrastructure, their mammalian hosts, way of transmission, pathogenicity, biochemical [49] and molecular characteristics.
From the biological and pathological point of view, these intracellular trypomastigotes could be implicated as a source of relapse after chemotherapy [40,72] and must contribute seriously to the pathogenesis of the infection by producing direct and indirect ultrastructural alterations of liver, adrenal cortex, vascular endothelium, red blood cells [25][26][27]64], as well as immunosuppression as a consequence of its tropism for lymphocytes and plasma cells. These are important facts to be considered by research groups that are working on the development of new drugs against T. evansi, especially if drugs cannot cross the plasma membrane of the infected host cells.
In addition, intracellular trypomastigotes could be responsible for the recolonization of the vascular bed once the first VAT's was destroyed by specific antibodies, as it has been proposed by Seed et al. [73] for antigenic variation during experimental infections of Microtus montanus with T. brucei gambiense (Wellcome TS strain).
As it has been described for T. brucei, the occurrence of intracellular stages of T. evansi EcHF91 in the murine models could be explained as a capability or peculiarity of this genotype of T. evansi, because these intracellular stages it has not been described yet in other Venezuelan isolates of T. evansi studied such as EcM96, HhHF92 and TEVA1.
On the other hand, because the invasion process could be determined by the size of the developing trypanosome population and the induction of the overexpression of receptors for parasite surface macromolecules, as it has been described for the Low Density Lipoprotein receptor (LDLr) overexpression in cardiac cells of mice experimentally infected with T. cruzi [74], more assays should be performed in order to determine if the invasion process constitute an parasitaemia-associated event during the experimental infection with the isolated EcHF91.
In the intracellular trypomastigote stages of T. evansi described in this work, the ultrastructural apoptotic features detected were similar to those described by Barth et al [78] in T. brucei staurosporine-induced cell death such as significant increase in glycosomes, acidocalcisomes and lysosomes number and electron density, increase in flagellar pocket size and volume, appearing of vesicles and constrictions inside the flagellar pocket among other.
Because pathogenic trypanosomatids are able to induce apoptotic responses in the populations of parasitized host cells, but also in their own populations during the course of infection [79,80], apoptosis must surely be playing an important role in the host-parasite relationship. In this sense, parasitic apoptosis seems to be involved in the evasion of host's inflammatory response (inhibiting or promoting their survival), as well as in the auto regulation of the uncontrolled growth of parasitic population, to avoid the early death of host and thus contribute to maximize their own dispersion and survival [79,81].
The occurrence of these intracellular stages in susceptible hosts under natural or field conditions could have important consequences at epidemiological level. In this regard, they could be responsible for cryptic parasitaemia described during infections of bovines [9], but also of the high levels of prevalence (30-50%) that could only be revealed by PCR in areas where the T. evansi infection is enzootic [10].
More research is needed not only to demonstrate the occurrence of these stages in horses, cattle, bubaline and capybara under field conditions, but also to know the mechanisms through which T. evansi invades host cells, as well as to assess the correlation of its presence with the high levels of prevalence in enzootic areas