Characterization of Stages of Hepatozoon americanum and of Parasitized Canine Host Cells

Abstract

American canine hepatozoonosis is caused by Hepatozoon americanum, a protozoan parasite, the definitive host of which is the tick, Amblyomma maculatum. Infection of the dog follows ingestion of ticks that harbor sporulated H. americanum oocysts. Following penetration of the intestinal mucosa, sporozoites are disseminated systemically and give rise to extensive asexual multiplication in cells located predominantly in striated muscle. The parasitized canine cells in “onion skin” cysts and in granulomas situated within skeletal muscle, as well as those in peripheral blood leukocytes (PBL), were identified as macrophages by use of fine structure morphology and/or immunohistochemical reactivity with macrophage markers. Additionally, two basic morphologic forms of the parasite were observed in macrophages of granulomas and PBLs. The forms were presumptively identified as merozoites and gamonts. The presence of a “tail” in some gamonts in PBLs indicated differentiation toward microgametes. Recognition of merozoites in PBLs supports the contention that hematogenously redistributed merozoites initiate repeated asexual cycles and could explain persistence of infection for long periods in the vertebrate host. Failure to clearly demonstrate a host cell membrane defining a parasitophorous vacuole may indicate that the parasite actively penetrates the host cell membrane rather than being engulfed by the host cell, as is characteristic of some protozoans.

American canine hepatozoonosis (ACH) was first recognized in Texas in 1978. ⁵ Geographic distribution of the disease has expanded to include numerous cases in Louisiana, Alabama, Georgia, Oklahoma, Florida, and Tennessee. ^{13, 18, 19, 23} The causative agent of ACH was initially believed to be a particularly virulent strain of Hepatozoon canis, ^{5, 18} but is now recognized as a new species, H. americanum. ³⁰

Dogs become infected by ingesting sporulated oocysts of H. americanum that have developed in the definitive host tick, Amblyomma maculatum. ²¹ Following ingestion, oocysts rupture, releasing sporocysts which, in turn, excyst under influence of bile, releasing numerous sporozoites. Sporozoites penetrate the gut mucosa and are disseminated to various organs and tissues. Merogony, however, occurs in host cells located principally in striated muscle. ^{5, 24, 26} To the authors' knowledge, the route by which sporozoites are transported from the alimentary tract to peripheral tissues (muscle) is yet to be defined. ^{10, 27}

The earliest recognized lesion of ACH in experimentally infected dogs is a large “modified” host cell, posited to be a macrophage that harbors a trophozoite. ²⁵ The interval between oral exposure and appearance of the initial lesion is as short as 3½ weeks. ^{21, 25, 27} The most frequently recognized histologic lesion of natural and experimentally induced ACH is the so-called “onion skin cyst,” which is composed of a multilamellar acid mucopolysaccharide deposit that surrounds a centrally located host cell, the latter containing the trophozoite (Fig. 1). ^{7, 25, 30} The parasite undergoes merogony within the cell, and on completion of asexual multiplication, the host cell degenerates, the surrounding mucopolysaccharide material disperses, and merozoites are freed into the surrounding tissue where they incite an acute, localized, neutrophilic inflammatory reaction that progresses to a granuloma (Fig. 2). ^{24, 25} Many large mononuclear cells within the granulomas harbor parasites. ^{7, 26, 30} Parasitized macrophages appear to escape the granuloma by traversing the wall of sinusoidal blood vessels within the granuloma. ²⁶ Some merozoites develop into gamonts that are infective for host ticks, whereas others are believed to be hematogenously transported as merozoites to other sites where they undergo repeat merogonic cycles. ^{24, 26} In the study reported here, we used transmission electron microscopy (TEM) and immunohistochemical (IHC) analysis to determine the identity of host cells within three tissue environments: 1) the encysted stage in skeletal muscle, 2) granulomas in skeletal muscle, and 3) peripheral blood leukocytes (PBL) in buffy coat preparations. Furthermore, we studied the parasite-host cell relationship and gathered data that support the occurrence of repetitive asexual generations and prolonged infection. ¹¹

Fig. 1.

“Onion skin cyst” (OSC), skeletal muscle; experimentally infected dog. At center of cyst is a canine host cell (C), the nucleus (arrow) of which contains a prominent nucleolus. The cytoplasm is highly vacuolated and contains a trophozoite of H. americanum (arrowhead). HE. Bar  =  10 µm.

Fig. 2.

Pyogranuloma, skeletal muscle; naturally infected dog. Notice numerous parasite-infected macrophages (arrows). HE. Bar  =  40 µm.

Materials and Methods

Specimens of blood and muscle used in this study were obtained by biopsy and necropsy from three naturally infected and eight experimentally infected dogs. The experimentally infected dogs were cared for in the tick-free animal facilities of Laboratory Animal Resources (LAR) at Oklahoma State University in accordance with conventional laboratory animal practices. Six of the eight experimentally infected dogs were 17-week-old, mixed-breed littermates that were born and raised at LAR. They were subjects of a multifaceted study that included co-relationships between stage of parasite development, clinical signs of disease, hematologic and biochemical data, and pathomorphologic, radiographic, and scintigraphic observations. ⁸ Skeletal muscle for histologic and immunohistologic examinations, and electron microscopy was collected at intervals between 3 and 10 weeks. Tissue obtained from two experimental animals and from the three naturally infected dogs was collected in the course of previously reported transmission experiments and studies of the tissue stages of H. americanum. ^{21, 24} Muscle biopsy was performed using standard procedures. ² Experimental infections were induced when dogs ingested sporulated oocysts derived from experimentally infected ticks. ²¹

Specimens of skeletal muscle for electron microscopy were placed in cold 2.5% glutaraldehyde fixative in 0.2 M sodium cacodylate buffer and were refrigerated until processing. Buffy coat specimens were prepared in capillary tubes filled with EDTA-anticoagulated blood obtained just prior to euthanasia. Following centrifugation, buffy coats were fixed in 2.5% glutaraldehyde in 0.2 M sodium cacodylate buffer and were refrigerated until processing. Both fixed muscle and buffy coat samples were further fixed in 2% osmium, dehydrated through a graded series of ethanol or acetone, and embedded in polybed resin. ¹⁶ Thick sections (approx 1 µm) of each sample were stained with Mallory's Richardson's blue and were examined by use of light microscopy. Ultrathin (silver-gold reflective) sections of selected samples were prepared using an ultramicrotome and a diamond knife, collected on a 300-mesh copper grid, stained with uranyl acetate and lead citrate, and examined using a JEOL 100CX II STEM transmission electron microscope operated at 80 kV. Grids of 12 encysted stages, 35 granulomas, and three fixed buffy coat preparations were examined in detail.

Specimens of skeletal muscle collected for IHC analysis and confocal microscopy were fixed in neutral-buffered 10% formalin at room temperature for 24 to 48 hours, then were processed in routine manner for paraffin sectioning. Five-micron-thick sections of each sample were stained with hematoxylin and eosin (HE); serial sections were subjected to immunohistochemical staining in a DAKO Autoimmunostainer Model LV (Dako Corporation, Carpenteria, CA). Primary antibodies included MAC 387 (DAKO M0760), CD3 (DAKO A0452), CD79a (DAKO M7051), α-1-antitrypsin (DAKO A0012), desmin (DAKO M0760), smooth muscle actin (DAKO M0851), vimentin (DAKO M0725), von Willebrand factor VIII (DAKO A0082), and S-100 (DAKO Z0628). In addition, rabbit-origin polyclonal antibody to H. americanum sporozoites was used to delineate the parasite within host cells. ²⁶ The procedure followed for each antibody, including ACH immunostaining, was performed using the recommended specification for each antibody and the accepted labeled streptavidin-biotin (LAB-SA) technique. ⁹ Dilutions for each antibody and their appropriate incubation times were previously determined in the histotechnology laboratory at the Oklahoma Animal Diagnostic Laboratory. Appropriate controls, both negative and positive antisera, were used to ensure antibody specificity. Following immunostaining, each slide was counterstained with Mayer's hematoxylin and examined.

For confocal microscopy, serial sections of affected skeletal muscle were deparaffinized and rehydrated in phosphate-buffered saline containing 0.2% fishskin gelatin (PBS-FSG). Sections were heated in antigen-unmasking solution, treated with 10% normal goat serum (diluted with PBS-FSG) for 20 minutes, and subjected to DAKO biotin-blocking system solutions. Slides were washed in PBS-FSG and treated with polyclonal antiHepatozoon antibody at a 1:100 dilution for 1 hour at room temperature. Biotinylated anti-rabbit antibody (1:200 dilution) raised in goats was applied to the slide for 30 minutes, followed by streptavidin conjugated with Alexa 488 (green) or 568 (red) diluted 1:1,000 (Molecular Probes Inc., Eugene, Oregon) for 30 minutes. Slides were washed and treated with the second primary antibody (MAC 387, Vimentin, Lysozyme), and a serial section was incubated with the appropriate negative control antibody for 1 hour. Secondary antibody (horse antimouse) conjugated with Alexa 568 or 488 (Molecular Probes) was applied, and slides were treated for 30 minutes. Confocal microscopy was performed using a Leica TCS SP laser-scanning microscope equipped with three lasers (Leica Microsystems, Exton, PA). The fluorescence of individual fluorochromes was captured simultaneously after optimization to reduce bleed through between channels (Photomultiplier tubes), using the Leica software. Individual confocal slices represent 0.5 µm, and optical slides were collected at 512 × 512-pixel resolution. The image files were transferred to a Macintosh G3 (Apple Computer, Cupertino, CA), and NIH-image v1.62 (ftp://codon.nih.gov/pub/nih-image/) and Photoshop v4.0 (Adobe, CA) were used to assign correct colors to the channels collected. Co-localization of antigens is indicated by the addition of colors as indicated (Figs. 4–6).

Fig. 3.

Peripheral blood film; experimentally infected dog. Macrophage contains ovoid Hepatozoon americanum gamont (arrow). Wright/Giemsa. Bar  =  10 µm.

Fig. 4.

Two cystic lesions (OSC), skeletal muscle; experimentally infected dog. The centrally located host cells (macrophages) react positively (green) for lysozyme. Indirect immunofluorescence (Alexa 488) and confocal microscopy. Bar  =  50 µm.

Fig. 5.

Host macrophage, skeletal muscle; experimentally infected dog. Early-stage trophozoite (red) within host macrophage, the cytoplasm of which reacts positively for vimentin (green). Indirect immunofluorescence (Alexa 568 and Alexa 488) and confocal microscopy. Bar  =  5 µm.

Fig. 6.

Parasite-induced granuloma, skeletal muscle; naturally infected dog. Macrophages react positively with MAC 387 (green). Macrophages contain parasites (red). Indirect immunofluorescence (Alexa 488 and Alexa 568) and confocal microscopy. Bar  =  25 µm.

Results

Encysted stages (“onion skin cysts”) were consistently observed between muscle fibers. In optimal planes of a section, cysts had a centrally located canine host cell that contained a developing zoite (Figs. 1, 5). The plasma membrane of the host cell was irregular, with multiple undulations. The host cell cytoplasm was replete with electron-dense aggregates of monoparticulate glycogen that frequently overshadowed other cytoplasmic organelles. Electron-dense vesicles were observed beneath and often fused with the plasma membrane. Vesicles were commonly observed secreting product into the adjacent ground substance (Figs. 7, 8). The multilamellar material surrounding the host cell was composed of amorphous electron-lucent ground substance and randomly scattered, fine, electron-dense particles. These ultrastructural qualities are consistent with mucopolysaccharide ground substance. Scattered capillaries, fibroblasts, and collagen fibers were often embedded within the lamellar material; fibrous tissue and skeletal muscle fibers adjacent to cysts were compressed. Inflammatory cells were rarely seen associated with “onion skin cysts.”

Parasites in encysted host cells were characterized by a prominent pellicle, central nucleus, variably sized vesicles, small osmiophilic micronemes, mitochondria, lipid vacuoles, and dilated rough endoplasmic reticulum (RER) (Fig. 8), all features of a developing apicomplexan trophozoite. A membrane defining a parasitophorous vacuole was not observed.

Fig. 8.

Higher magnification of parasite in Fig. 7. The parasite contains a central nucleus (N), prominent pellicle (P), micronemes (arrows), and scattered vesicles (V). Uranyl acetate and lead citrate. Bar = 2 µm.

Fig. 7.

Parasite-infected canine host cell in an OSC. The cell has features of a macrophage that include an irregular plasma membrane with undulations and multiple, electron-dense cytoplasmic vesicles, some of which appear to be in process of extrusion into cyst material (arrow). A parasite (trophozoite) is indicated (arrowheads). Uranyl acetate and lead citrate. Bar = 5 µm.

Within granulomas, parasite-infected macrophages had a plasma membrane that was irregular, with numerous short (filopodia) and elongate (lamellipodia) undulations. The nucleus was oval to indented, with dense heterochromatin dispersed throughout the nucleus associated with the nuclear envelope (Fig. 9). The cytoplasm of these cells contained numerous elongate mitochondria that were often clustered around the developing parasite. The cells also had numerous dilated Golgi complexes, electron-dense lysosomes, and scattered, dilated RER. Parasitized neutrophils were not recognized in the examination of granulomas.

Fig. 9.

Parasite-infected macrophages within a skeletal muscle granuloma. Cells have a deeply indented nucleus (HCN) and cell membranes that are irregular, with multiple projecting processes of variable size. Parasites, merozoites, or gamonts are marked (arrows). Uranyl acetate and lead citrate. Bar = 2 µm.

Two stages of H. americanum were observed within macrophages of granulomas. Merozoites were characterized by elements of an apical complex, micronemes, microtubules, pellicle, large nucleus with prominent nucleolus, variable-sized vesicles, and various cytoplasmic organelles (Fig. 10). Pellicles consisted of three layers: the outer and inner layers were electron dense, with an electron-lucent middle layer. A few parasites were surrounded by an irregular pale zone that could represent the remnant of a parasitophorous vacuole, though a distinct membrane was not recognized. A second parasitic stage, presumably a developing gamont, also was present in granulomas. This form had a prominent pellicle and a large round nucleus with irregularly dispersed heterochromatin, as well as numerous vesicles and distended RER.

Fig. 10.

Parasite (merozoite) in macrophage within a skeletal muscle granuloma. Notice prominent apical complex (A), numerous micronemes (MN), microtubules (MT), and host cell nucleus (HCN). Uranyl acetate and lead citrate. Bar = 0.5 µm.

Results of immunostaining of host cells in cysts and in granulomas are delineated in Table 1. Peripheral blood leukocytes were not subjected to immunostaining. Although host cells in granulomas reacted positively with MAC 387, encysted host cells did not react.

Table 1.

Results of immunohistochemical studies of parasitized host cells in two locations.

Location of host cell	Vimentin	CD3	CD 79a	α-anti Tryp	MAC 387	S-100	Desmin	SM Actin	von Willebrand	Lysozyme
Cyst	+	−	−	+	−	−	−	−	−	+
Granuloma	+	−	−	+	+	−	−	−	−	+

Tryp = trypsin; SM = smooth muscle actin.

Infected cells in buffy coat preparations appeared to be macrophages, morphologically similar to cells in cysts and granulomas, and characterized by oval, bilobed, or trilobed nuclei containing nucleoli and euchromatin, as well as heterochromatin that was condensed at the nuclear envelope. Cytoplasmic organelles consisted of numerous mitochondria, lysosomes, digestive vacuoles, well developed and often swollen Golgi complexes, dilated RER, and loosely aggregated ribosomes. Cytoplasmic granules characteristic of canine neutrophils were not observed.

Infected macrophages in buffy coat sections also contained two morphologically distinct forms of H. americanum. Each had a single nucleus. One form (Fig. 11) was folded, with the posterior and anterior ends coming to a distinct point. There was a thick pellicle composed of an outer limiting membrane and an inner membrane; the nucleus was oval, with multiple areas of heterochromatin at the nuclear envelope. The cytoplasm of the parasite was heavily granulated, and contained RER and clusters of mitochondria. In addition, infected macrophages contained round, electron-dense bodies, lipid inclusions, small, clear vacuoles, and numerous dilated Golgi complexes.

Fig. 11.

Parasite-infected canine peripheral blood (buffy coat) macrophage; experimentally infected dog. The parasite (gamont) is acutely folded, has a nucleus (PN), numerous mitochondria (M), Golgi complex (G), and sharply pointed ends. Uranyl acetate and lead citrate. Bar = 1 µm.

The second morphologically distinct form was smaller than the first and had an elongated “tail-like” appendage (Fig. 12). Features of the appendage included central axial microtubules and a bilayer consisting of an electron-dense outer membrane and an intermediate pale zone (Fig. 13). Although space was observed between the parasite and the host cell cytoplasm, a membrane that would define a parasitophorous vacuole was not observed.

Fig. 12.

Parasite-infected peripheral blood (buffy coat) macrophage; experimentally infected dog. The parasite (P), apparently a developing or mature male gamont, has a well-formed tail-like appendage. Uranyl acetate and lead citrate. Bar = 1 µm.

Fig. 13.

High magnification of a posited male gamont, peripheral blood macrophage; experimentally infected dog. The appendage has a prominent electron-dense bilayer membrane. Notice vesicle (V) and mitochondria (M). Uranyl acetate and lead citrate. Bar = 0.25 µm.

Discussion

On the basis of light and electron microscopic findings and with limited immunohistochemical evidence, we concluded that cells hosting the parasite are of monocytic origin whether located in the “onion skin cysts” where merogony occurs, granulomas where gametogeny commences and where merozoites may be redistributed hematogenously, or in PBLs where gametogeny continues. ^{4, 12} Evidence for that contention is relatively complete regarding host cells in granulomas where morphologic and immunohistochemical data indicate their identity as macrophages. Though encysted host cells have similar morphologic features and have been identified by others to be macrophages, ⁷ they did not react with the macrophage marker, MAC 387. Failure of encysted host cells to react with that marker might be anticipated because MAC 387 is a marker of macrophages recently derived from blood; reactivity is lost at later stages of differentiation. ³¹ If, as we suspect, sporozoites reach the locus of merogony after intracellular transport from the intestine, the host cell (a monocyte) left the circulation a minimum of 3½ or many weeks prior to the IHC procedure. On the other hand, the interval between extravasation of macrophages in granulomas and IHC analysis is shorter ²⁵ ; thus, MAC 387 positivity would be expected in the latter instance but not the former.

It is not clear whether the parasite invades yet another host cell after entering the peripheral blood; nevertheless, there is clear evidence that parasite-containing macrophages in the extravascular compartment of granulomas traverse vessel walls to reach the vascular lumen and peripheral circulation. ^{24, 25} The few gamont-containing cells found in the many PBL preparations we examined had morphologic features of macrophages. We did not conduct IHC analysis on PBLs. Though it is commonly reported that neutrophils harbor hepatozoon gamonts, as demonstrated in routinely stained blood smears, careful documentation of that belief has not been established. ^{5, 22} In fact, in the case of H. canis, there is morphologic and histochemical evidence that gamonts actually circulate in monocytic cells rather than neutrophils. ²⁰ Further, considering the brief life span of an extravasated neutrophil, it is unlikely that such a cell could survive long enough to be a successful host. Time is required for the progression of stages in the life cycle of the parasite, including survival of sufficient duration as a gametocyte to be part of a tick vector's blood meal. ¹⁰

Theories as to the mechanism of persistent infection include hematogenous redistribution and recycling of merozoites, quiescent sporozoites, or sporozoite polymorphism. ^{1, 3} The recognition of two morphologic forms of the parasite within host cells of the same granuloma and in PBLs is significant. The existence of merozoites in circulating macrophages adds validity to the hypothesis that hematogenous redistribution of merozoites is a mechanism by which repeated cycles of merogony and prolonged hepatozoon infection occur. The latter phenomenon is exemplified by infection that persisted longer than 5 years in a naturally infected dog housed in a tick-free environment. ¹¹ The form with a tail has been reported previously. ^{7, 30} We observed microtubules in the tailed form, but were unable to confirm the microtubular arrangement characteristic of cilia. Nevertheless, we suspect the structure represents a developing apicomplexan gamont destined to become a microgamete.

Whereas others have reported a parasitophorous vacuole surrounding hepatozoon parasites, we were unable to demonstrate a membrane that would validate the existence of a parasitophorous vacuole. ^{6, 14} Absence of the vacuole would suggest that entry of the parasite into the host cell is by penetration rather than by phagocytosis. Zoites of species belonging to several different families of the Apicomplexa are known to occupy parasitophorous vacuoles; alternatively, zoites of species within other families of this large aggregate of parasites appear to be free within the host cell cytoplasm. ^{1, 28} It has been suggested, for example, that Babesia microti of the class Piroplasmea enters a host cell surrounded by a segment of host cell membrane, but that the host membrane disintegrates rapidly resulting in absence of confirming evidence of a parasitophorous vacuole. ²⁹ The actual means by which H. americanum enters the cell either as a sporozoite or as a merozoite remains to be determined but our findings indicate that the host cell is a macrophage and that the parasite actively penetrates the cell.

We provide further evidence that H. americanum induces the host macrophage to produce and secrete the mucopolysaccharide cyst material that shields the parasite from host defense mechanisms. ^{15, 17} The ability to more finitely identify developmental states of the parasite in granulomas and in peripheral blood would improve chances to develop a clear understanding of repeat cycles and to confirm unequivocally that neutrophils are not successful host cells.