|Home | About | Journals | Submit | Contact Us | Français|
Infections by the protozoan parasite Toxoplasma gondii are widely prevalent worldwide in animals and humans. This paper reviews the life cycle; the structure of tachyzoites, bradyzoites, oocysts, sporocysts, sporozoites and enteroepithelial stages of T. gondii; and the mode of penetration of T. gondii. The review provides a detailed account of the biology of tissue cysts and bradyzoites including in vivo and in vitro development, methods of separation from host tissue, tissue cyst rupture, and relapse. The mechanism of in vivo and in vitro stage conversion from sporozoites to tachyzoites to bradyzoites and from bradyzoites to tachyzoites to bradyzoites is also discussed.
Infections by the protozoan parasite Toxoplasma gondii are widely prevalent in humans and animals worldwide. T. gondii has emerged as one of the most common opportunistic infections in patients with AIDS. Toxoplasmosis in AIDS patients is considered to be a result of reactivation of latent infection, but the mechanisms of reactivation are unknown. This review focuses on the structure and biology of T. gondii stages (tachyzoites, bradyzoites, and tissue cysts) in intermediate hosts (humans) and the resistant (oocyst) stage outside the host.
There are three infectious stages of T. gondii: the tachyzoites (in groups or clones), the bradyzoites (in tissue cysts), and the sporozoites (in oocysts). These stages are linked in a complex life cycle (Fig. (Fig.1).1).
The term “tachyzoite” (tachos = speed in Greek) was coined by Frenkel (73) to describe the stage that rapidly multiplied in any cell of the intermediate host and in nonintestinal epithelial cells of the definitive host. The term “tachyzoite” replaces the previously used term “trophozoite” (trophicos = feeding in Greek). Tachyzoites have also been termed endodyozoites or endozoites. Aggregates of numerous tachyzoites are called clones, terminal colonies, or groups.
The tachyzoite is often crescent shaped, approximately 2 by 6 μm (Fig. (Fig.2),2), with a pointed anterior (conoidal) end and a rounded posterior end. Ultrastructurally, the tachyzoite consists of various organelles and inclusion bodies including a pellicle (outer covering), apical rings, polar rings, conoid, rhoptries, micronemes, micropore, mitochondrion, subpellicular microtubules, endoplasmic reticulum, Golgi complex, ribosomes, rough and smooth endoplasmic reticula, micropore, nucleus (Fig. (Fig.22 to to10),10), dense granules, amylopectin granules (which may be absent), and a multiple-membrane-bound plastid-like organelle which has also been called a Golgi adjunct or apicoplast (24–26, 103, 145, 176). The nucleus is usually situated toward the central area of the cell and contains clumps of chromatin and a centrally-located nucleolus.
The pellicle consists of three membranes, a plasmalemma and two closely applied membranes that form an inner membrane complex (Fig. (Fig.33 and and4)4) which is formed from a patchwork of flattened vesicles (121, 136, 145, 176). The inner membrane is discontinuous at the anterior tip above the polar rings, at micropores which are situated laterally, and at the posterior pore at the extreme posterior tip of the zoite. Polar ring 1 is an electron-dense thickening of the inner membrane complex at the anterior end of the tachyzoite, which encircles a cylindrical, truncated cone called the conoid, which consists of six to eight microtubular elements wound like a compressed spring (Fig. (Fig.44 and and5).5). Twenty-two subpellicular microtubules originate from polar ring 2 and run longitudinally almost the entire length of the cell just beneath the inner membrane complex. In addition, two inner microtubules terminate in the conoid (Fig. (Fig.5).5). The microtubules are like a rib cage and are arranged in a gentle spiral. Individual microtubules have prominent transverse striations (121). Between the anterior tip and the nucleus, there are 8 to 10 club-shaped organelles (164) called rhoptries (Fig. (Fig.33 and and66 to to10).10). Rhoptries are excretory structures, each consisting of an anterior narrow neck up to 2.5 μm long that extends into the interior of the conoid, and a saclike, often labyrinthine posterior end (up to 1 μm long). Micronemes are rod-like structures which occur mostly at the anterior end of the parasite (Fig. (Fig.33 and and6).6).
Although tachyzoites can move by gliding, flexing, undulating, and rotating, they have no visible means of locomotion such as cilia, flagella, or pseudopodia. The functions of the conoid, rhoptries, micropores, and micronemes are not fully known but are probably associated with host cell penetration and creation of an intracellular environment suitable for parasite growth and development. The conoid can rotate, tilt, extend, and retract as the parasite probes the host cell plasmalemma immediately before penetration (19). Rhoptries have a secretory function associated with host cell penetration, secreting their contents through the plasmalemma just above the conoid to the exterior (123). They contain a proteolytic enzyme (140a). The micropore is a cytosome-like structure formed by the invagination of the outer membrane of the pellicle (19, 123, 124).
Tachyzoites enter host cells (Fig. (Fig.10)10) by actively penetrating through the host cell plasmalemma or by phagocytosis (15, 54, 95, 120, 123, 125, 147, 152, 160, 182). After entering the host cell, the tachyzoite becomes ovoid and is surrounded by a parasitophorous vacuole (PV), which appears to be derived from both the parasite and the host cell. Soon after penetration, a tubulovesicular membranous network (TMN) develops within the PV (Fig. (Fig.8).8). Some of the TMN membranes are connected to the parasitophorous vacuolar membrane (145, 149–151). The TMN appears to be derived from the posterior end of the tachyzoite (150). However, convoluted tubules, structurally similar to the TMN, were observed at the end of tachyzoites by Nichols et al. (123), and we have also observed similar structures in in vivo-cultured tachyzoites (Fig. (Fig.10).10).
Tachyzoites multiply asexually within the host cell by repeated endodyogeny (Fig. (Fig.77 and and8),8), a specialized form of reproduction in which two progeny form within the parent parasite (Fig. (Fig.3),3), consuming it (145). In endodyogeny, the Golgi complex divides first, becoming two complexes at the anterior end of the nucleus. Next, the anterior portions of the inner membrane complexes and the subpellicular microtubules of the progeny cells appear as two dome-shaped structures anteriorly. The parasite nucleus becomes horseshoe shaped, and the ends of the nucleus move into the dome-shaped anterior ends of the developing progeny. The inner membrane complex and subpellicular microtubules continue to extend posteriorly and surround one half of the nucleus, which eventually pinches into two. The progeny continue to grow until they reach the surface of the parent. The inner membrane complex of the parent disappears, and its outer membrane becomes the plasmalemma of the progeny cells. Tachyzoites continue to divide by endodyogeny (145). In vivo, most groups of tachyzoites are arranged randomly due to asynchronous cycles of endodyogeny. However, rosettes are occasionally formed due to synchronous division. In rapidly dividing tissue culture-adapted strains, T. gondii within a vacuole may divide synchronously (14, 140), but this is not the norm. Rarely, tachyzoites of certain strains divide by binary fission (63, 139). The host cell ruptures when it can no longer support the growth of tachyzoites (Fig. (Fig.88).
The rates of invasion and growth vary depending on the strain of T. gondii and the type of host cells (1, 101). After entry of tachyzoites into a host cell, there is a variable lag period before the parasite divides, and this lag phase is partly parasite dependent (1). Mouse virulent strains of T. gondii grow faster in cell culture than do “avirulent” strains, and some strains of T. gondii form more rosettes than others (1). Although T. gondii isolates have been classified genetically into types I, II, and III (90, 94, 148), there are no appreciable structural differences among different isolates of T. gondii.
The term “bradyzoite” (brady = slow in Greek) was also coined by Frenkel (73) to describe the organism multiplying slowly within a tissue cyst. Bradyzoites are also called cystozoites.
Tissue cysts (Fig. (Fig.11)11) grow and remain intracellular (Fig. (Fig.11C)11C) as the bradyzoites divide by endodyogeny (64, 65). Tissue cysts vary in size; young tissue cysts may be as small as 5 μm in diameter and contain only two bradyzoites (Fig. (Fig.11B),11B), while older ones may contain hundreds of organisms (Fig. (Fig.11E).11E). Tissue cysts in the brain are often spheroidal and rarely reach a diameter of 70 μm, whereas intramuscular cysts are elongated and may be 100 μm long (27, 36). Although tissue cysts may develop in visceral organs, including the lungs, liver, and kidneys, they are more prevalent in the neural and muscular tissues, including the brain, eyes, and skeletal and cardiac muscles (35). Intact tissue cysts probably do not cause any harm and can persist for the life of the host without causing a host inflammatory response.
The tissue cyst wall is elastic and thin (<0.5 μm thick) (Fig. (Fig.11),11), and it encloses hundreds of crescent-shaped bradyzoites (Fig. (Fig.1212 and and13),13), each approximately 7 by 1.5 μm in size (118). The tissue cyst develops within the host cell cytoplasm. The cyst wall is argyrophilic (77, 153), but results vary depending on the silver impregnation method used. According to Sims et al. (153), the cyst wall stains intensely with Bodian protoargol and Palmgren silver but not with methenamine silver. The lack of staining with methenamine silver indicates that the cyst wall contains no glycogen or other polysaccharides. The cyst wall is composed of host cell and parasite materials (64, 65, 153). It is ultimately lined with granular material, which also fills the space between the bradyzoites. Some bradyzoites degenerate (Fig. (Fig.13),13), especially in older tissue cysts (129). The cyst wall is only faintly periodic acid-Schiff (PAS) positive (Fig. (Fig.11D).11D).
Bradyzoites (Fig. (Fig.33 and and1212 to to16)16) differ structurally only slightly from tachyzoites. They have a nucleus situated toward the posterior end, whereas the nucleus in tachyzoites is more centrally located. The contents of rhoptries in bradyzoites are usually electron dense, whereas those in tachyzoites are labyrinthine. However, the contents of rhoptries in bradyzoites vary with the age of the tissue cyst. Bradyzoites in younger tissue cysts may have labyrinthine rhoptries, whereas those in older tissue cysts are electron dense (Fig. (Fig.12).12). Also, most bradyzoites have one to three rhoptries, which are looped back on themselves (Fig. (Fig.3).3). Bradyzoites contain several amylopectin granules which stain red with PAS reagent; such material is either in discrete particles or absent in tachyzoites (Fig. (Fig.3).3). Bradyzoites are more slender than are tachyzoites. Bradyzoites are less susceptible to destruction by proteolytic enzymes than are tachyzoites (93), and the prepatent period in cats following feeding of bradyzoites is shorter than that following feeding of tachyzoites (48).
Cats shed oocysts after ingesting any of the three infectious stages of T. gondii, i.e., tachyzoites, bradyzoites, and sporozoites. Prepatent periods (time to the shedding of oocysts after initial infection) and frequency of oocyst shedding vary according to the stage of T. gondii ingested (37, 48, 80). Prepatent periods are 3 to 10 days after ingesting tissue cysts, ≥18 days after ingesting oocysts (37), and ≥13 days after ingesting tachyzoites (44). Fewer than 30% of cats shed oocysts after ingesting tachyzoites or oocysts, whereas nearly all cats shed oocysts after ingesting tissue cysts (37, 48).
After the ingestion of tissue cysts by cats, the cyst wall is dissolved by proteolytic enzymes in the stomach and small intestine. The released bradyzoites penetrate the epithelial cells of the small intestine and initiate the development of numerous generations of T. gondii (Fig. (Fig.1717 and and18).18). Five morphologically distinct types of T. gondii develop in intestinal epithelial cells before gametogony begins (47). These stages are designated types A to E instead of generations, because there are several generations within each T. gondii type (Fig. (Fig.17).17). Little has been added to the structure or biology of types A to C since the original description by Dubey and Frenkel (47).
So far, only late stages (presumably type D) have been studied by transmission electron microscopy (TEM). These forms multiply by a specialized form of schizogony (Fig. (Fig.19).19). As in normal schizogony, the nucleus divides two or more times without cytoplasmic division (61, 133, 144). Whether daughter organism (merozoite) formation begins after four or more nuclei have been formed is uncertain. Before or simultaneous with the last nuclear division, merozoite formation is initiated near the center of the schizont by the development of dome-shaped merozoite anlagen; whether one or two anlagen are formed near each nucleus is uncertain (61, 144). The merozoites eventually move towards the periphery of the schizont, and the schizont plasmalemma invaginates around each merozoite, forming the plasmalemma of the merozoite. The merozoites separate from the schizont at their posterior ends, sometimes leaving a residual body (47).
The schizogony observed in T. gondii type D organisms differs from schizogony in conventional coccidia (Eimeria species) in that in T. gondii merozoites are formed internally and immature merozoites do not protrude from the schizont surface. It should be pointed out that many details of schizogony of T. gondii are not clear even in type D organisms; nothing is known of the ultrastructures of T. gondii types to A to C described by Dubey and Frenkel (47).
Piekarski et al. (133) proposed the term “endopolygeny” to describe the division of T. gondii schizonts in feline enterocytes. They believed that merozoite formation began after two nuclear divisions. Vivier (175) had earlier used the term “endopolygeny” to describe observed divisional formation of more than two daughter T. gondii tachyzoites in the peritoneum of mice. We would like to propose that the term “endopolygeny” should not be used to describe schizogony in T. gondii but should be restricted to the merozoite formation in Sarcocystis and Frenkelia spp. (51). In Sarcocystis endopolygeny, the nucleus becomes multilobed but does not divide into separate nuclei. Each nuclear lobe is eventually incorporated into developing merozoites. The number of merozoites in Sarcocystis schizonts varies from 4 to 100 or more (51).
After the asexual development (types A to E), the sexual cycle starts 2 days after tissue cysts were ingested by the cat. The origin of gamonts has not been determined, but the merozoites released from schizont types D and E probably initiate gamete formation. Gamonts are found throughout the small intestine, but most commonly in the ileum, 3 to 15 days after inoculation (Fig. (Fig.18).18). They are found above the nucleus of the host epithelial cell near the tips of the villi of the small intestine (Fig. (Fig.20).20). Female gamonts are subspherical, and each contains a single centrally located nucleus and several PAS-positive granules. Ultrastructurally, the mature female gamete contains several micropores, rough and smooth endoplasmic reticulum, numerous mitochondria, double-membraned vesicles, and wall-forming bodies (WFB) (Fig. (Fig.20B).20B). The double-membraned bodies are located near the nucleus and are probably derived from it (67). WFB are of two types: type I and type II. Type I WFB are about 0.35 μm in diameter and electron dense, and they appear before type II WFB (67). Type II WFB are moderately electron dense, less abundant than WFB, and larger than WFB (1.2 μm in diameter) (67).
Mature male gamonts (microgamonts) are ovoid to ellipsoidal in shape (Fig. (Fig.20A).20A). During microgametogenesis, the nucleus of the microgamont divides to produce 10 to 21 nuclei (47). The nuclei move toward the periphery of the parasite and enter protuberances formed in the pellicle of the microgamont. One or two residual bodies remain in the microgamont after division into microgametes (Fig. (Fig.20A).20A). Microgametes are elongated and consist mainly of nuclear material. The anterior end is a pointed structure called the perforatorium, within which lie two basal bodies. Two long, free flagella originate from the basal bodies and project posteriorly. A large mitochondrion is situated near the basal bodies and just anterior to the nucleus. Five microtubules originate near the nucleus and extend posteriorly alongside it for a short distance.
Microgamonts have up to 21 microgametes. Microgametes use their flagella to swim to and penetrate and fertilize mature macrogametes to form zygotes. After fertilization, an oocyst wall is formed around the parasite (Fig. (Fig.20B).20B). Infected epithelial cells rupture and discharge oocysts into the intestinal lumen. Five layers of the oocyst wall are formed at the surface of the zygote (67). No extensive cytoplasmic changes occur in the zygote while layers 1, 2, and 3 are formed, but type I WFBs disappear with the formation of layer 4 and type II WFBs disappear with the formation of layer 5.
Unsporulated oocysts are subspherical to spherical and are 10 by 12 μm in diameter (Fig. (Fig.21A).21A). Under light microscopy, the oocyst wall consists of two colorless layers. Polar granules are absent, and the sporont almost fills the oocyst. Sporulation occurs outside the cat within 1 to 5 days of excretion depending upon aeration and temperature.
Sporulated oocysts are subspherical to ellipsoidal and are 11 by 13 μm in diameter (Fig. (Fig.21B21B and C). Each oocyst contains two ellipsoidal sporocysts without Stieda bodies. Sporocysts measure 6 by 8 μm. A sporocyst residuum is present; there is no oocyst residuum. Each sporocyst contains four sporozoites (Fig. (Fig.2121C).
The ultrastructure of sporulation was described by Ferguson et al. (56–60). The cytoplasm of the unsporulated oocyst (zygote) has a large nucleus with amorphous nucleoplasm and a distinct nucleolus. The zygote is limited by a unit membrane with few micropores. The nucleus divides twice, giving rise to four nuclei, which are situated at the periphery of the zygote; at this stage a second limiting membrane is formed. After the cytoplasm divides, two spherical sporoblasts are formed, each with two nuclei (57).
As the sporulation continues, the sporoblasts elongate and sporocysts are formed. The two outer membranes of the sporoblasts become the outer layer of the sporocyst wall, and the plasmalemma of the cytoplasmic mass becomes the inner layer (58). Ultimately, as the sporocyst develops, four curved plates form the innermost layer of the sporocyst. The plates are joined by four liplike sutures (described below) with a depression on the surface of the sporocyst wall at the junction point. Sporozoite formation begins when two dense plaques (anlagen) appear at both ends of the sporocyst. Each nucleus divides into two and is incorporated into elongating sporozoite anlagen. Thus, four sporozoites are formed in each sporocyst (59). A prominent residual body is left after sporozoite are formed; the residual body is enclosed in a single-unit membrane (59).
Ultrastructurally, the oocyst wall of sporulated oocysts (Fig. (Fig.22)22) consists of three layers: an electron-dense outer layer, an electron-lucent middle layer, and a moderately electron-dense inner layer (159). The middle layer consists of remnants of two membranes that evidently were laid down between the inner and outer layers during oocyst wall formation. Treatment of oocysts with 1.3% sodium hypochlorite (Clorox) removes the outer layer. The oocyst wall contains a single micropyle, which is relatively small and is located randomly in the oocyst wall (Fig. (Fig.23).23). The micropyle is a 350-nm-diameter indentation that consists of three layers that are continuous with the three layers in the oocyst wall. However, the outer layer of the micropyle is thin and moderately electron dense and the inner layer is electron dense, disk shaped, and slightly thicker than the inner layer of the oocyst wall. Although the function of the micropyle is not known, it might represent a permeable site in the oocyst wall that is susceptible to the actions of CO2 and various enzymes that allow the entry of bile salts and trypsin, which stimulate the excystation of sporozoites from sporocysts.
The sporocyst wall consists of two distinct layers with a thin, electron-dense outer layer and thicker, moderately electron-dense inner layer. The inner layer consists of four curved plates (20, 58) (Fig. (Fig.2121 and and22).22). At sites of apposition between two plates, there are two apposing liplike thickenings and an interposed strip in the inner layer (20, 159). During excystation in the presence of bile salts and trypsin, the sporocyst ruptures suddenly, evidently at the sites of apposition between plates, releasing the sporozoites. As the plates separate, they curl inward to form conical coils (Fig. (Fig.2323 and and24).24). The sporocyst residuum consists of amylopectin granules and lipid bodies.
Ultrastructurally, the sporozoite is similar to the tachyzoite, except that there is an abundance of micronemes, rhoptries, and amylopectin granules in the former. Sporozoites are 2 by 6 to 8 μm in size with a subterminal nucleus (Fig. (Fig.2424 and and25).25). There are no crystalloid bodies or any refractile bodies in T. gondii sporozoites (Fig. (Fig.2424 and and25).25).
The ultrastructural differences of some of the organelles in tachyzoites, bradyzoites, and sporozoites of the VEG strain of T. gondii are compared in Tables Tables11 and and2;2; the VEG strain was isolated from the blood of an AIDS patient and is mildly virulent to mice depending on the stage of the parasite inoculated (42, 49).
Sporozoites, tachyzoites, and bradyzoites of T. gondii are similar ultrastructurally but differ in certain organelles and inclusion bodies (Tables (Tables11 and and2).2). All three zoites have similar numbers of rhoptries but the rhoptries differ in appearance between the zoites. For example, the rhoptries of tachyzoites are uniformly labyrinthine; sporozoites usually contain both labrynthine and uniformly electron-dense rhoptries; and bradyzoites usually contain uniformly electron-dense rhoptries, some of which are folded back on themselves (Fig. (Fig.33 and and15).15). Tachyzoites have few micronemes, sporozoites have an intermediate number, and bradyzoites have many. Dense granules are more numerous in sporozoites and tachyzoites than in bradyzoites (Fig. (Fig.33 and and25).25). Amylopectin granules are numerous and relatively large in sporozoites and bradyzoites but are few and small or absent in tachyzoites. Lipid bodies are numerous in sporozoites, rare in tachyzoites, and absent in bradyzoites (Fig. (Fig.33 and and2525).
In general, the attachment and penetration of host cells by T. gondii zoites (tachyzoites, bradyzoites, sporozoites, and merozoites) appear similar to those described for other coccidian parasites. The mechanical events involved in zoite attachment and penetration include (i) gliding of the zoite, (ii) probing of the host cell with the conoidal tip of the zoite, (iii) indenting the host cell plasmalemma, (iv) forming a moving junction that moves posteriorly along the zoite as it penetrates into the host cell, and (v) partially exocytosing micronemes, rhoptries, and dense granules. Zoites of T. gondii can penetrate a variety of cell types from a wide range of hosts, indicating that the biochemical receptors involved in attachment and penetration are probably common to most animal cells. Host cell receptors consisting of laminin, lectin, and SAG1 are involved in T. gondii tachyzoite attachment and penetration (100). T. gondii zoites can, however, enter cells by means other than receptor-mediated penetration. Speer et al. (160) found that after passing completely through cells, some sporozoites of T. gondii carried an envelope of host cell membranes and cytoplasm but were still capable of penetrating other cells. Zoites of T. gondii can also enter cells by being endocytosed (160).
Recently, a great deal of research on the formation of the PV has been conducted (96). As tachyzoites penetrate host cells, they are surrounded by a membrane that is evidently derived from the host cell plasmalemma minus host proteins. This membrane is destined to become the PV membrane (PVM), and a number of parasite proteins associate with it, including rhoptry proteins ROP2, ROP3, ROP4, and ROP7. ROP2 is located on the host cell cytoplasmic side of the PVM, suggesting that it plays a role in host-parasite biochemical communication (2). Within a few minutes after penetration, tachyzoites modify the newly formed PV and the PVM with parasite proteins and a TMN forms within the PV. The PVM acquires pore structures that freely allow charged molecules up to 1,200 kDa to diffuse bidirectionally between the PV and the host cell cytoplasm (142). Dense-granule proteins (GRA) are secreted into the PV after tachyzoite penetration, with GRA3 and GRA5 localizing on the PVM and GRA1, GRA2, GRA4, and GRA6 associating with the TMN. Collectively, these modifications establish a parasite-friendly environment within the host cell cytoplasm that is conducive to parasite replication.
To date, there is only a single report involving the expression of parasite antigens in T. gondii sporozoites. Speer et al. (161) found that GRA3 and SAG1 are developmentally regulated; they are not expressed in the sporozoite or the sporozoite-infected cell until 12 to 15 h after sporozoite inoculation of cell cultures. The first parasite multiplication occurs after the expression of GRA3 and SAG1, indicating that although the sporozoite is competent for cell penetration, it is not immediately able to establish an intracellular environment which can support replication. Therefore, many of the proteins necessary for parasite growth appear to be down-regulated in the sporozoite, perhaps due to dormancy within the oocyst.
Most of the available information concerning the interaction of T. gondii zoites with host cells has been derived from in vitro cultivation with tachyzoites. There is limited information on in vivo zoite-host cell interactions. Recent studies with sporozoites have shown that T. gondii interacts with host cells substantially differently in vivo from the interaction in vitro. For example, Speer et al. (161) found that in cultured cells, T. gondii sporozoites induced the formation of two types of PVs. Type 1 PVs formed first, were relatively large (20 to 30 μm in diameter), had an indistinct PVM, and contained no exocytosed dense granule material or a TMN. After 12 to 18 h, sporozoites in type 1 PVs actively penetrated into the host cell cytoplasm and established type 2 PVs, which contained exocytosed dense granules, a TMN, and a distinct PVM. Parasites multiplied by endodyogeny in type 2 PVs but did not multiply in type 1 PVs. In follow-up studies in mice, it was found that sporozoites passed through enterocytes and goblet cells of the mouse intestinal epithelium and entered the lamina propria, where they infected all cells of the host except erythrocytes and underwent endodyogeny to form tachyzoites (52, 158). In contrast to the in vitro studies, sporozoites did not form type 1 PVs at any point while they were in the intestinal epithelium or the lamina propria. Even though the sporozoites were just passing through ileal enterocytes, they still formed PVs that contained exocytosed dense granule material and well-developed TMNs, which appear necessary for parasite multiplication (102, 106, 149, 150). The presence of exocytosed material and TMNs associated with the PVs of sporozoites in transit across the intestinal epithelium indicates that parasite replication does not always follow the formation of a parasite-modified PV. After entering cells in the lamina propria, PVs equivalent to the type 2 PVs were again formed, but here the sporozoites multiplied by endodyogeny. These findings therefore indicate that the type 1 PV seen in cultured cells might represent an anomaly of in vitro cultivation.
Bjerkås (5) and Sibley et al. (150) suggested that the TMN was formed in association with a transient sac-like structure at the posterior end of the tachyzoite. However, as stated above, TMN can develop early near the anterior of tachyzoites even before the parasite completely enters the host cell (Fig. (Fig.1010).
Lainson (104) reviewed earlier literature on the development of tissue cysts. Levaditi et al. (107) apparently were the first to report that T. gondii may persist in tissues for many months as “cysts.” However, considerable confusion between the terms “pseudocysts” (group of tachyzoites) and “cysts” existed for many years. Frenkel and Friedlander (77) and Frenkel (70) cytologically characterized cysts containing organisms with a subterminal nucleus and PAS-positive granules surrounded by a argyrophilic cyst wall. Wanko et al. (177) first described the ultrastructure of the T. gondii cysts and its contents.
Lainson (104) provided evidence that cysts were formed in mice as early as 8 days after the inoculation of tachyzoites. His illustrations of a cyst with four organisms on day 8 postinfection (p.i.) and with 20 organisms on day 10 p.i. provided the first convincing evidence of young cysts of T. gondii. Older cysts were up to 60 μm in diameter and contained approximately 3,000 organisms.
Jacobs et al. (93) first provided a biologic definition of cysts when they found that cystic organisms were resistant to digestion by gastric juice (pepsin-HCl) whereas tachyzoites were destroyed immediately. Thus, cysts became important in the life cycle of T. gondii because carnivorous hosts could become infected by ingesting cysts.
Dubey and Frenkel (48) made the first in-depth study of the development of the tissue cysts and bradyzoites and defined cysts biologically and morphologically. They found that the cysts were formed as early as 3 days after inoculation of mice with tachyzoites. Cats shed oocysts with a short prepatent period (3 to 10 days) after ingesting bradyzoites, whereas after they ingested tachyzoites or oocysts, the prepatent period was longer (≥14 days).
Ferguson and Hutchison (65) reported the first in-depth ultrastructural studies of the development of T. gondii cysts.
Dubey and Beattie (45) proposed that cysts should be called tissue cysts to avoid confusion with oocysts.
T. gondii tissue cysts occur in many organs and cell types. Most observations have been made with tissue cysts in the brains of mice, but the acid-pepsin digestion procedure and bioassay in cats have shown that tissue cysts occur in many extraneural organs. Tissue cyst distribution is in part controlled by the host and strain of T. gondii. The distribution of tissue cysts in different organs of animals fed the GT-1 strain of T. gondii is shown in Table Table3.3. In pigs, dogs, cats, and rodents fed oocysts of the VEG strain of T. gondii, more tissue cysts were found in muscular tissues than in the brain in cats, dogs, and pigs whereas more tissue cysts were found in the brain than in other organs in rats and mice (38–40, 49, 110). Infections with T. gondii in experimental animals appear to be essentially the same as in some naturally infected animals as reviewed by Dubey and Beattie (45). For example, Jacobs et al. (92) isolated T. gondii from 18 diaphragms and 9 brains of 31 seropositive sheep killed at a slaughterhouse in New Zealand.
Tissue cyst size is dependent on cyst age, the type of host cell parasitized, and the cytological method used for measurement. Young tissue cysts may contain as few as two bradyzoites surrounded by a distinct cyst wall and measure about 5 μm in diameter (Fig. (Fig.11B).11B). Tissue cysts in myocytes are two to three times longer than those in neural cells (36). Because tissue cysts are most numerous in the brains of mice, most observations were made with neural tissue cysts. Beverley (4) measured the volume of tissue cysts liberated from the mouse by grinding brains with a mortar and pestle in saline (0.85% NaCl). Tissue cysts grew uniformly up to 10 weeks, after which there was considerable variability in tissue cyst size, perhaps due to a second generation of tissue cysts. The tissue cysts were up to 58 μm in diameter. He also stated that a tissue cyst may contain 60,000 organisms. Van der Waaij (174) made an in-depth study of the growth of tissue cysts in mouse brains in which 100 to 500 free (liberated from mouse brains by homogenization in saline), unstained tissue cysts were measured in the brains at 4, 8, 12, 16, and 24 weeks from mice inoculated subcutaneously (s.c.) with tissue cysts. The tissue cysts grew uniformly in size up to 12 weeks p.i., and then the growth levelled off. The tissue cysts were up to 70 μm in diameter, but the mean diameter of 100 tissue cysts at 16 weeks p.i. was 42 μm.
Ferguson and Hutchison (65) measured tissue cysts in thin (≤1-μm) sections from mouse brain fixed in glutaraldehyde. They ultrastructurally observed 140 tissue cysts with a minimum of 10 tissue cysts at 11, 21, and 28 days and 3, 6, 12, 18, and 22 months. The tissue cysts were up to 20 μm in diameter at 28 days, up to 30 μm at 3 months, and up to 50 μm at ≥6 months.
Dubey (38) measured tissue cysts in the brains of rats 72 to 75 days after feeding the rats oocysts. The brains were fixed in 10% buffered neutral formalin and sectioned at 5-μm thickness. Although the tissue cysts (n = 224) were up to 50 μm in diameter, most of them were approximately 30 μm in diameter.
In addition to these reports, Dubey (44a) has never seen T. gondii tissue cysts larger than 70 μm in diameter in formalin-fixed paraffin-embedded sections of brains of hundreds of naturally or experimentally infected animals. The measurement of tissue cysts in formalin-fixed, paraffin-embedded sections provides a standard means of reporting results. The sizes of the tissue cysts vary a great deal when unstained tissue cysts are examined between a glass slide and coverslip, depending on the homogeneity of the brain suspension, the amount of fluid, and the pressure applied. A highly flattened tissue cyst that initially passed through a 63-μm filter is shown in Fig. Fig.26.26.
The size of the T. gondii tissue cyst may vary with the strain of T. gondii. Although there are no firm data, one of us (44a) has observed up to 300% variability in tissue cyst size at 2 months after oral inoculation of mice with oocysts of different isolates of T. gondii. The tissue cysts of some isolates were only 20 μm in diameter, whereas others were up to 60 μm in diameter.
There are no firm data on the number of bradyzoites in a tissue cyst. Most of the information is from reviews (4, 137). In a large, flattened tissue cyst illustrated by Huskinson-Mark et al. (91), 990 bradyzoites are clearly visible. The 60,000 bradyzoites in a tissue cyst mentioned by Beverley (4) appears unrealistic.
Cornelissen et al. (21) described a method to separate tissue cysts from the brains of mice after suspensions of brain homogenates were run on discontinuous Percoll gradients. According to these authors, T. gondii tissue cysts have a specific gravity of 1.056. This method has been widely used. Tissue cysts can also be separated on discontinuous gradients with 25 to 30% Percoll (7, 132) or a 20% dextran solution (79). The degree of success in purifying tissue cysts depends on the host tissue and the amount of blood contamination. To minimize tissue and erythrocyte contamination, the mice should be bled out before the tissue is harvested and the brain homogenate should be passed through a 90-μm wire sieve. Tissue cysts of T. gondii are smaller than 90 μm and pass through the sieve. They can be stored in saline at 4°C for 2 months (41, 93). However, the mortality rate of bradyzoites stored at 4°C for prolonged periods has not been determined.
Mice are often used to obtain tissue cysts of T. gondii for experimental purposes, and the most frequently used T. gondii strains are Beverley and ME-49. Tissue cyst numbers in mouse brains vary depending on the strain of mouse, the strain of T. gondii, the route of inoculation, and the number of organisms inoculated. More tissue cysts are produced if mice become mildly ill but without obvious clinical signs. With the original Beverley strain, outbred mice inoculated s.c. with tissue cysts developed mild illness (weight loss during weeks 2 and 3) but survived. Hundreds of large tissue cysts were seen when the mice were killed at 12 weeks p.i. However, this strain has been passaged frequently in mice and has become more pathogenic for mice. Similarly, the pathogenicity of some lines of the ME-49 strain has increased since its original isolation. Therefore, to prevent mortality, one may have to use prophylaxis (sulfadiazine sodium or sulfamerazine in drinking water at 15 to 100 mg/100 ml of water) when the mice become ill. The dosage and duration of chemotherapy should be adjusted for each T. gondii strain, stage of the parasite inoculated, and strain of mice; there is no standard formula.
The number of tissue cysts in mouse brain is genetically regulated (6, 16–18, 117). Brown et al. (17) reported that the tissue cyst burden was regulated by mouse chromosome 17 containing the class I gene Ld. More tissue cysts were produced in congenic mice.
Tissue cyst persistence may vary with the duration of infection, depending on the strain of T. gondii and the host. In one experiment with the ME-49 strain, the number of tissue cysts recovered from the brains of CBA/Ca mice at 4, 8, 12, and 16 weeks p.i. were 3,720, 2,158, 3,133, and 1,538, respectively (62). Therefore, the optimal time of harvest should be determined for each strain of T. gondii in a given host.
Suzuki et al. (169) compared tissue cyst formation and mortality in two inbred strains of mice inoculated intraperitoneally (i.p.) with tissue cysts of the ME-49 strain. At 3 weeks after inoculation of 20 tissue cysts or more, nearly 10 times as many tissue cysts were found in the brains of CBA/Ca mice, all of which survived, as compared with BALB/c mice, two of which had died. After i.p. inoculation with 80 tissue cysts, the mortality rate during acute infection increased to 75% in BALB/c mice, whereas all CBA/Ca mice survived. However, during chronic infection, 50% of CBA/Ca mice died between 2 and 6 months p.i. whereas all BALB/c mice survived. Therefore, the optimal time to harvest tissue cysts may vary with the strain of T. gondii and the strain of the mouse.
It is well known that T. gondii tissue cysts persist in organs of infected hosts for several months and perhaps for life, depending on the host and parasite strains. The localization of tissue cysts also varies with the host and the strain of T. gondii. Although more tissue cysts are found per gram of tissue in mice than in other hosts, the fate of tissue cysts is difficult to study in mice because mice are never completely immune to T. gondii and new tissue cysts (Fig. (Fig.12)12) are formed even in chronic infections (4, 46, 62, 174). Even tachyzoites are present in the brains of chronically infected mice (52, 62). Because T. gondii tissue cysts are small and tissue cyst rupture is unpredictable, Frenkel (70) studied tissue cyst rupture in hamsters by using T. gondii and a related parasite, Besnoitia jellisoni, where intact tissue cysts were not associated with inflammatory reaction. Leaking and ruptured tissue cysts were accompanied by marked inflammation. Microglial nodules were common in chronically infected hamsters, and in some T. gondii antigen could be shown to be present. He proposed that some tissue cysts rupture in chronically infected animals and that the released bradyzoites are destroyed by the immunocompetent host. However, in immunosuppressed animals, the released bradyzoites are believed to reactivate T. gondii infection.
The factors affecting tissue cyst rupture are largely unknown. T. gondii tissue cyst rupture has been documented only rarely. Ferguson et al. (66) quantitatively studied tissue cyst rupture in mice chronically infected with the STR strain of T. gondii. Only 2 (0.27%) ruptured tissue cysts were seen among 750 tissue cysts examined, although T. gondii-positive debris was found in eight (1.4%) glial nodules. Tissue cyst rupture was documented in a Panamanian night monkey (Aotus lemurinus) that had been vaccinated three times with an attenuated (ts-4) strain of T. gondii and then challenged orally with tissue cysts of a complete T. gondii strain (55, 76). Glial nodules were found around degenerating tissue cysts. Degenerating tissue cysts were found in rats inoculated orally with oocysts of the VEG strain in the absence of tachyzoites or formation of new tissue cysts (38).
The mechanism of formation of new generations of tissue cysts in chronically infected mice is unknown (Fig. (Fig.2727 to to29).29). Clusters of tissue cysts are found in mice infected with certain strains of T. gondii (163), sometimes in clinically normal mice (Fig. (Fig.28).28). Whether bradyzoites leak from intact tissue cysts is uncertain (174). The rupture of tissue cysts and subsequent multiplication of tachyzoites can lead to fulminating toxoplasmosis even in chronically infected mice (Fig. (Fig.29).29).
Little information about the mechanisms of relapse is available. Treatments with corticosteroids, antilymphocyte serum, and anti-interferon (IFN) antibody are known to induce immunosuppression and relapse due to toxoplasmosis (3, 78, 81, 82, 119, 126, 162, 165–167). The animal model, route of inoculation and stage of T. gondii used for primary infection, and criteria used for evaluation of relapses are all important considerations. For example, hamsters are more corticosteroid sensitive than are mice or rats, and orally administered corticosteroids are less effective than parenterally administered corticosteroids (74). Chronically infected hamsters initially immunized with the RH strain died of overwhelming toxoplasmosis following corticosteroid treatment (78). Similar relapsing fatal toxoplasmosis has not been induced in mouse models. Odaert et al. (127) examined the brains of chronically infected mice 6, 9, and 12 days after oral administration of dexamethasone. They reported seeing more foci of necrosis and more tissue cysts in cortisone-treated mice than in controls. However, there was no convincing evidence that relapse had occurred in this short period following corticoid administration (74). In experiments involving chronically infected mice, Sumyuen et al. (165) found that there was an increased mortality rate after immunosuppressive therapy (cortisol acetate or azathioprine, or azathioprine plus cortisol acetate) but that the number of T. gondii organisms in the brains and lungs of drug-treated mice remained similar to that found in untreated mice. Similar results were reported recently by Nicoll et al. (126), who found increased numbers of necrotic and gliotic foci, but no increase in parasite numbers, in dexamethasone-treated and chronically infected T. gondii mice. However, many tachyzoites die in the necrotic foci and do not infect new cells.
Miédougé et al. (119) compared the clinical course and parasite numbers in mice injected i.p. with tissue cysts of the Beverley strain of T. gondii. Starting 50 days p.i., group A and B mice were injected weekly with anti-IFN rabbit antiserum and group C were controls. Group B mice also received antitoxoplasmic drugs (pyrimethamine and sulfadiazine in drinking water). On day 68 (i.e., 18 days after treatment), five mice from each group were killed and homogenates of lungs and brains were bioassayed in tissue culture. All the mice remained clinically normal. The T. gondii numbers were dramatically higher in group A mice. All five mice in group A had 856 to 1,337 T. gondii organisms/g of lung, whereas only one of five mice in group B and no mice in group C had organisms. Because histologic examination was not performed, the results may be misleading because rupture of a single tissue cyst while making tissue homogenates for bioassay can result in a ≥1,000-fold difference in T. gondii infectivity titers. The authors also reported that there was parasitemia 53, 61, and 68 days after infection even in mice not given anti-IFN antibody. Thus, the mouse model does not appear to be an accurate representation for relapsing toxoplasmosis in AIDS patients (3). One should also take into account that clusters of tissue cysts of different sizes, numbers of tachyzoites, and reactivated lesions occur in chronically infected, nonimmunosuppressed mice (52).
Although it has been known for 30 years that immunity to T. gondii is cell mediated (71), the precise mechanism of relapse is unknown. Clinical toxoplasmosis in AIDS patients is thought to be due to reactivation of a chronic infection, probably mediated by CD4+ lymphocyte deficiency (82). Production of the cytokine IFN-γ in mice is considered to be the main mediator of immunity to toxoplasmosis in both acute and chronic infection (167, 168). Mice chronically infected with the ME-49 strain of T. gondii died 10 to 18 days after administration of anti-IFN antibody, and depletion of both CD4+ and CD8+ lymphocytes was needed to induce mortality in chronically infected mice (82). Depletion of CD8+ lymphocytes alone induced mortality in some mice, but mortality was not induced in those depleted of only CD4+ lymphocytes.
Lesions, similar to those seen in AIDS patients, were seen in mice that died after the administration of anti-IFN antibody; they consisted of necrosis and infiltration of neutrophils associated with the presence of tachyzoites (82). Because a few tachyzoites might be present even in chronically infected asymptomatic mice, it remains to be determined whether the reactivation is initiated by these “dormant” tachyzoites or bradyzoites released from tissue cysts.
Even though Hogan et al. (89) were the first to report the presence of T. gondii tissue cysts in cell culture, little interest in in vitro cultivation of tissue cysts was shown over the next 25 to 30 years, and few reports appeared in the literature (88, 97, 98, 114, 146). Hoff et al. (88) demonstrated that in vitro-produced tissue cysts led to oocyst excretion in cats, indicating that the tissue cysts were biologically the same as those produced in vivo. A renewed interest in studying the biology of tissue cysts and bradyzoites was evidenced once the importance of toxoplasmic encephalitis in AIDS patients was recognized in the mid-1980s. Improved methods for the induction and detection of tissue cysts were developed in the early 1990s. Many researchers are now using in vitro systems to investigate many exciting aspects of bradyzoite-tachyzoite and tachyzoite-bradyzoite interconversion and the development of tissue cysts.
Studies concerning the development of tissue cysts and conversion of tachyzoites to bradyzoites and of bradyzoites to tachyzoites are not always clear-cut because of the terminology used to describe stages and the gradualness of the conversion process (85). For example, is a PV that contains tachyzoites and bradyzoites, as indicated by immunostaining, a tissue cyst, and if TEM demonstrates a tissue cyst wall surrounding what are structurally tachyzoites, is it a tissue cyst? The results of observations in vitro should always be compared to what is already known about tissue cyst production in vivo (74).
The methods used to determine whether tissue cysts and bradyzoites are present in vitro are widely varied.
Light microscopy is probably the least definitive of all the methods used to demonstrate tissue cysts in vitro because it is impossible to distinguish bona fide tissue cysts from large groups of tachyzoites. It is somewhat easier to identify free-floating tissue cysts that originated from host cell rupture and are floating in the media in cell cultures, but confusion with groups of tachyzoites still cannot be ruled out. Phase-contrast microscopy may provide a more definitive identification of these stages because the tissue cyst wall is phase lucent (178, 179).
Identification of tissue cysts in cultures stained with Giemsa, PAS, or silver stains can also be highly subjective; PAS and silver stains are superior to Giemsa. The age at which tissue cysts become PAS or silver positive has not been examined critically. Immunostaining with specific monoclonal or polyclonal antibodies is highly specific and is the technique currently used by most researchers (see below). Quantitative studies that use only unstained or histochemically stained cultures must be interpreted with caution.
The tachyzoites of T. gondii are more susceptible to digestion by acid-pepsin solution than are the bradyzoites (93). Resistance to acid-pepsin digestion was used to determine the numbers of bradyzoites in a cell culture in a modified plaque assay (134). However, this method cannot consistently distinguish tachyzoites from bradyzoites (see below).
TEM can be used to demonstrate tissue cysts and the localization of antigens in bradyzoites and tissue cysts. The matrix of tissue cysts produced in vitro is often absent or greatly reduced (113, 116) compared to tissue cysts in mouse brain (65). Free-floating tissue cysts that originated from host cell rupture during processing for TEM are most likely to exhibit a lack of matrix material. Tissue cysts that form in vitro are generally smaller than those found in vivo, and the identity of the host cell is important (87). Young tissue cysts many contain odd numbers of organisms (Fig. (Fig.30)30) (89, 154, 179). Generally, it is difficult to obtain quantitative data by TEM because of the small amount of material used for examination. Immunoelectron microscopy has been used to determine the expression and localization of bradyzoite-specific antigens in vitro (85, 86, 172, 179).
Only the bradyzoite stage of T. gondii will reliably produce oocyst excretion in cats. The prepatent period is short in orally acquired bradyzoite-induced infections and can be used to determine if tissue cysts are present in cell cultures (88, 109). By using TEM and by bioassay in cats fed T. gondii, it was determined that tissue cysts were present on day 3 p.i. but were not infectious for cats until day 6 p.i. (109). Repeated passage of T. gondii in cell cultures results in the loss of the ability of bradyzoites to induce oocyst excretion in cats after about 40 passages in vitro (109), a phenomenon similar to that which occurs after repeated passage of tachyzoites in mice (74a). Functional bradyzoites of the oocystless T263 strain of T. gondii have been produced in cell culture and used to orally vaccinate cats against oocyst excretion (134).
With the identification of stage-specific antigens and the subsequent generation of bradyzoite and tissue cyst stage-specific monoclonal antibodies, methods for closely examining the developmental biology of T. gondii tissue cysts became available (9, 10, 13, 86, 99, 171, 178, 181, 183). By using both tachyzoite- and bradyzoite-specific monoclonal antibodies, it was determined that tachyzoites and bradyzoites could both be present in the same PV (10, 155, 156, 179), indicating that stage conversion from tachyzoite to bradyzoite is asynchronous. Immunostaining permits accurate quantification of specific stages and is important in studies designed to determine the ability of agents to induce tissue cyst formation in vitro.
The types of host cells and T. gondii strains used to study tissue cyst and bradyzoite development in vitro have been widely varied (Table (Table4).4). The majority of studies have been done with fibroblast cell types (mainly of human origin) because these cells are easy to grow and usually survive for extended periods as an intact monolayer. Most cell lines will support tissue cyst development, and the type of cell line probably does not contribute greatly to the presence or absence of tissue cyst formation (109, 116). Most strains of T. gondii will produce tissue cysts in vitro (113, 116). Strains with low replication rates, which are less pathogenic for mice (i.e., VEG, ME-49, Beverley, Prugniaud, and NTE), usually produce more tissue cysts than do the more rapidly dividing, pathogenic strains (i.e., RH and BK) (83, 114). Before the use of methods to induce tachyzoite-to-bradyzoite conversion, strains of T. gondii that were less pathogenic for mice produced tissue cysts with larger numbers of zoites (both bradyzoites and tachyzoites) than did highly pathogenic strains (11, 83).
Most strains of T. gondii will spontaneously develop tissue cysts in cell culture with no manipulation (22, 109, 111, 113, 156). However, the number of tissue cysts produced spontaneously is small and manipulation of the cell culture system is needed to increase this number (11, 83, 155, 156, 178).
Early on, the use of T. gondii antiserum in the culture medium was used to promote tissue cyst formation in vitro (88, 98, 146). The results of these studies are difficult to interpret because questionable methods of identification of tissue cysts were used (98, 146) or tissue cyst numbers were not quantifiable (88). Popiel et al. (134) achieved an enhanced rate of tissue cyst formation in cultures treated with 5% rabbit anti-T. gondii serum but attributed the enhancement to a reduction in the replication of tachyzoites. It remains unclear if anti-T. gondii serum actually induces bradyzoite and tissue cyst formation.
Tachyzoite-to-bradyzoite conversion can be induced by applying external stress to various types of infected cell lines. Many investigators use pH manipulation to induce in vitro stage conversion. Both acidic (pH 6.6) and basic (pH 8.0 to 8.2) manipulations will lead to the conversion of tachyzoites to bradyzoites (155, 179). Most researchers use the pH 8 treatment. Exposure of extracellular tachyzoites to medium of pH 8.1 for 1 h increases the formation of bradyzoites and tissue cysts (180). Temperature stress (40°C) and chemical stress (sodium arsenite) will also induce tachyzoite-to-bradyzoite transformation (157). However, sodium arsenite treatment did not result in the production of tissue cysts when examined by TEM (155).
IFN-γ may act to inhibit tachyzoite replication in static cell cultures and permit the spontaneous development of tissue cysts (97). Although IFN-γ does not cause increased expression of bradyzoite-specific antigens or tissue cyst production in human fibroblasts (155), it is highly effective in inducing tissue cyst formation in cultured macrophages. IFN-γ treatment of murine macrophages induces the release of nitric oxide (NO), which reduces tachyzoite replication and induces the expression of bradyzoite-specific antigens (9, 11). If IFN-γ induction of NO release is inhibited by treatment with polymyxin or anti-tumor necrosis factor, no inhibition of tachyzoite replication and no increased expression of bradyzoite antigens occur. The highest level of bradyzoite antigen expression is found in macrophages where the NO release is manipulated to be 40 to 65% of maximum (11). Exogenous NO supplied by sodium nitroprusside will also inhibit tachyzoite replication and induce bradyzoite antigen expression in murine macrophages and in host cells with nonfunctional mitochondria (12). Released NO reacts with iron-sulfur centers of proteins and therefore reacts with several proteins involved in electron transport. The activity of NO in host cells lacking functional mitochondria indicates that its effect is exerted on parasite mitochondria and not on host cell mitochondria.
Mitochondrial inhibitors (oligomycin, antimycin A, atovaquone, rotenone, myxothiazol, and carbonyl cyanide m-chlorophenylhydrazone), which affect different aspects of mitochondrial function, will also induce tachyzoite-to-bradyzoite stage conversion (11, 69, 170). Studies involving host cells with nonfunctional mitochondria indicate that the inhibitors are affecting the T. gondii mitochondrian and not the host cell mitochondria (11, 170).
A common mode of action of the pH manipulations, sodium arsenite treatment, 43°C treatment, NO, and mitochondrial inhibitors is that they can be associated with the induction of heat shock proteins (HSP) (9, 180). The relationship of HSP to these stress inducers is being actively investigated in several laboratories. Bohne et al. (8) have cloned and characterized a bradyzoite-specifically expressed gene (hsp30/BAG-1), which is related to genes encoding small HSP of plants. BAG-1 is located in the cytoplasm of bradyzoites, and expression of the 30-kDa antigen appears to be regulated at the mRNA level (8). Weiss et al. (180) have shown that a monoclonal antibody to the inducible form of HSP70 reacts specifically with bradyzoites and recognizes a 72-kDa antigen. If quercetin is used to inhibit the synthesis of HSP, induction of bradyzoite antigen expression is also inhibited. The roles that HSPs play in bradyzoite development and induction will be an active area of research over the next several years.
Sporozoites, tachyzoites, and bradyzoites can develop into tissue cysts in vitro. If sporozoites are used as the inoculum, the appearance of tissue cysts seems to be delayed by several days (109). Ultrastructurally, tissue cysts have been observed as early as 2 days after inoculation of a mixture of tachyzoites and bradyzoites (111), and several TEM studies have documented tissue cysts by 3 days p.i. (109, 179). Tissue cysts present by 3 days p.i. are not biologically mature, because they do not induce oocyst excretion in cats.
Immunostaining studies of cultured cells have identified a small population of tachyzoites that express bradyzoite- and tachyzoite-specific antigens (10, 155), which most probably represent transitional stages. In the absence of any induction treatment, tachyzoites obtained from the peritoneal cavity of mice do not react with bradyzoite-specific antibodies, but by 24 h in culture, a small population will express both tachyzoite and bradyzoite antigens, and cells containing only bradyzoite-reactive organisms are present by 3 days (156). Maximal expression of bradyzoite-specific antigens occurs 4 days after inoculation of NTE tachyzoites in pH 8-induced cell cultures (85).
When mouse brain-derived bradyzoites are used as the inoculum in the absence of any induction treatment, the bradyzoites begin to express tachyzoite-specific antigens as early as 15 h (156). Mixed populations were observed at 24 h, and cells containing only tachyzoites or only bradyzoites were observed at 48 h. The observation of groups of parasites that were reactive with only bradyzoite-specific antibodies indicates that the inoculated bradyzoites can produce additional bradyzoite-reactive stages.
The source of tachyzoites and the strain of T. gondii may influence the development of bradyzoites. According to Appleford and Smith (1), a small population of organisms from peritoneal exudates of mice contain bradyzoites as revealed by staining with a bradyzoite-specific monoclonal antibody (Pb36).
Few data are available on the isolation of T. gondii tissue cysts from cell cultures. Soête et al. (155) isolated pH 8- or 43°C-induced tissue cysts of the RH strain from Vero or human fibroblast cells by scraping the cells from the monolayer and rupturing host cells with a Dounce homogenizer. Low-speed centrifugation resulted in a sediment that contained tissue cysts and free zoites. Bohne et al. (9) used magnetic cell sorting to isolate pH 8-induced bradyzoites of the NTE strain of T. gondii grown in L929 fibroblasts. A surface-reactive bradyzoite-specific antibody was used to purify the bradyzoites from a mixture of bradyzoites and tachyzoites. The final product consisted of 95 to 98% pure bradyzoites free of host cell contaminants.
The period needed for stage conversion of T. gondii varies with the inoculum, route of inoculation, and method of examination (Table (Table5).5). In these studies, tissue cyst formation was sought in the brains of mice. To circumvent the process by which the parasite reaches the brain tissue from the site of inoculation and is then converted to bradyzoites, Dubey and Frenkel (48) inoculated tachyzoites directly into the brains of mice and found that some tachyzoites converted to bradyzoites between 2 and 3 days p.i. At daily intervals, T. gondii-injected mice were fed to cats and the feces of the cats were examined for oocyst shedding. Cats fed mice infected 1 or 2 days previously never shed oocysts with a short prepatent period (<10 days). The results were not related to the number of T. gondii organisms inoculated, because the cats were fed homogenates of many mice inoculated simultaneously by 4 routes (i.p., s.c., intramuscular, and intracerebral). Dubey and Frenkel (48) also demonstrated tissue cysts by conventional histology with special stains (PAS and silver impregnation method).
Recently, tissue cyst formation in mice inoculated orally with tissue cysts or bradyzoites or with oocysts or sporozoites was examined (42, 52). Tissues of mice fed tissue cysts or oocysts were bioassayed in cats, in mice after pepsin digestion, and by immunohistochemical staining with a bradyzoite-specific antibody (BAG-5).
After the mice were fed tissue cysts or bradyzoites, bradyzoites penetrated through enterocytes, entered various cell types in the lamina propria, and divided into tachyzoites by 18 h p.i. The infection disseminated to extraintestinal organs, with parasitemia occurring at 24 to 48 h p.i. Bioassay and histologic examination showed that tissue cysts formed as early as 6 days p.i. in the brain and other tissues. In general, the detection of BAG-5 antigen paralleled that in the cat bioassay. The acid-pepsin digestion procedure was unreliable because it gave inconsistent results with tissue cysts produced during acute infection. Evidence for direct conversion of bradyzoites to bradyzoites was not found.
After the mice were fed oocysts, sporozoites penetrated enterocytes but developed only in lamina propria cells during the first 12 h p.i. The conversion of sporozoites to tachyzoites to bradyzoites required 7 days. Thus, there was a delay of 1 day in bradyzoite formation after oocyst feeding with respect to that after tissue cyst feeding.
In both bradyzoite-induced and sporozoite-induced oral infections, T. gondii organisms were not seen in histologic sections of the brain until 6 days p.i., and at that time individual BAG-5-positive organisms were seen mixed with BAG-5-negative organisms (Fig. (Fig.31).31). Whether individual BAG-5-positive organisms seen in the brain had migrated from other tissues or were released from tissue cysts could not be determined. The stage of the tachyzoite-to-bradyzoite conversion at which the tissue cyst wall is formed in acute infection is uncertain. In chronically infected mice, a tissue cyst wall is clearly visible around two bradyzoites, both of which also have a terminal nucleus (Fig. (Fig.11).11). Recent studies by Sahm et al. (141) suggest that bradyzoites invading the host cell secrete the ground substance of the cyst wall and that during the invasion process the parasite determines whether it will form tissue cysts or tachyzoites.
The recent development of stage-specific antibodies makes it possible to study in vitro conversion of bradyzoites to tachyzoites and in vitro conversion of tachyzoites to bradyzoites. Soête et al. (156, 157) inoculated bradyzoites onto MRC-5 cell cultures and monitored the appearance of bradyzoite-specific (B+) or tachyzoite-specific (P30) antigens. At 15 h after inoculation, the organisms had both B+ and P30 antigens. Organisms began to divide at approximately 24 h p.i. and vacuoles containing two doubly labelled organisms were seen. By 48 h p.i. some organisms had lost B+ antigens and were considered to be tachyzoites. A quantitative analysis was not possible after 48 h p.i. because some parasites had ruptured and reinvaded other host cells. These authors concluded that not all bradyzoites transform into tachyzoites, since they found multiplying bradyzoites (156); these results were different from in vivo transformation of all bradyzoites into tachyzoites by 48 h p.i. as determined by the cat bioassay. The same authors also observed that cell cultures inoculated with tachyzoites from the peritoneal exudate of mice had completely lost tachyzoite-specific markers by 72 h p.i. Bohne et al. (9, 10) obtained similar results in cell cultures inoculated with bradyzoites. The bradyzoite-specific antigens declined from 100 to 15% from day 1 to 3 after inoculation of cultures. Conversely, the bradyzoite-specific antigens increased steadily from day 1 to 6 in cultures inoculated with tachyzoites.
By using stage-specific monoclonal antibodies, bradyzoite-specific antigens were detected in the brains of mice on day 9 (127). Mice were examined 6, 7, 9, 12, and 14 days after being fed 20 tissue cysts of the 76K strain. The brains of the mice were stained with tachyzoite-specific antibodies (P30) and bradyzoite-specific antibodies (P36). On days 6 and 7, only P30-specific organisms were present. On day 9, groups of parasites were labelled doubly with P30 and P36 antibodies. On days 12 and 14, the number of P30-positive organisms decreased whereas the number of P36-positive organisms increased.
BAG-5 develops early in vitro and is probably a small HSP-like molecule associated with development. Weiss et al. (179), using the BAG-5 bradyzoite-specific antibody, found that tissue cysts had formed by day 3 after inoculation of human foreskin fibroblasts with bradyzoites of the ME-49 strain; it appears that some bradyzoites formed tissue cysts directly without conversion to tachyzoites. Using a monoclonal antibody specific for the tissue cyst wall, Halonen et al. (87) found tissue cysts in human fetal neuronal culture beginning on day 2 after inoculation with the ME-49 strain. The development of bradyzoites in vitro occurs in a series of steps, whereas the development of mature tissue cysts is less frequent and requires a minimum of 6 days in culture. This is supported by studies of antigen expression as ascertained by the various bradyzoite-specific sera available. Most bradyzoite-specific monoclonal antibodies and recombinant polyclonal sera available are expressed after 1 day in culture under conditions which induce bradyzoite switching (9–11, 155, 156). The late-appearing antigen is not seen until 5 to 6 days of culture. In addition, results of feeding experiments suggest that even at 6 days, not all of the observed tissue cyst-like structures are mature, since only a subset behave as tissue cysts in animal bioassays (88, 109). Thus, to some extent, the data on in vitro cystogenesis mimic the data seen in this study. Early on, one would expect tachyzoites that are in process of becoming mature bradyzoites to express BAG-5 antigen (day 1 of transition), but full biologic maturity would take several more days. It is clear from the above discussion that for in vitro work, markers of mature functional cysts are needed.
While these in vitro studies are helpful in elucidating stage conversion, the results do not completely agree with those of in vivo studies. The 15 h needed to convert bradyzoites to tachyzoites in cell culture agrees with the data obtained with mice, where functional tachyzoites were not found 18 h after bradyzoites were fed to mice (42). The 72 h needed for tachyzoites to completely lose their tachyzoite-specific markers also agrees with the 3 days needed to form biologically functional tissue cysts after inoculation of tachyzoites into mice (48). However, when bradyzoites were inoculated into mice by any route, the minimum period needed to form biologically functional tissue cysts was 6 days (48). Unlike in vitro studies, all bradyzoites had converted to biologically defined tachyzoites by 18 h p.i. (42). Therefore, biologic measurements should be examined in the context of parasitologic and host factors (74), and caution should be used in interpreting in vitro phenomena as having biological significance.
Until recently, acid-pepsin digestion was a generally accepted method to distinguish tachyzoites from bradyzoites and to recover T. gondii from tissues. Biologically, bradyzoites are resistant to gastric digestion and thus remain orally infective whereas tachyzoites are often destroyed by gastric juice. This resistance of bradyzoites to digestion by gastric juice has been known for over 36 years. Jacobs et al. (93) found that bradyzoites can survive in acid-pepsin solution for 2 h or more whereas tachyzoites are killed within 1 h. These authors digested homogenates of liver, spleen, and lungs of mice, inoculated i.p. with RH strain tachyzoites, for 60 min in acid-pepsin solution and found that tachyzoites were rendered noninfective to mice. Direct microscopic examination of tachyzoites in acid-pepsin revealed that they were immediately damaged; they became more granular, less refractile, and ghostlike within 15 to 30 min (93). However, Jacobs et al. (93) tested the infectivity of tachyzoites for mice after digestion in acid-pepsin for only a 60-min period. Subsequently, most other researchers accepted that tachyzoites are immediately destroyed by acid-pepsin. Unlike acid-pepsin, tachyzoites survived in 1% trypsin for 3 h (93).
Sharma and Dubey (143) quantitatively studied survival of tachyzoites and bradyzoites in acid-pepsin and trypsin solutions. They reported that bradyzoites survived in acid-pepsin for 2 h but not for 3 h.
According to Pettersen (130), the destruction of tachyzoites in acid-pepsin was due to acid, not to pepsin, because no differences were found in survival rates when tachyzoites were incubated in acid-pepsin at room temperature or at 37°C (pepsin is active only at 37°C and at low pH). Pettersen (130) also reported that tachyzoites of two virulent strains (RH and 119) survived in acid-pepsin for 20 min but not 25 min. He thought that bradyzoites were present in the peritoneal exudate of mice inoculated with bradyzoites of two avirulent strains of T. gondii. The peritoneal exudate was obtained 4 or 6 days after i.p. inoculation with bradyzoites of strain DUE; the organisms in the peritoneal exudate survived 90 min in acid-pepsin at 37°C.
In a follow-up paper, Pettersen (131) proposed that bradyzoites can be excreted in the milk of mice 5 days after i.p. inoculation of lactating mice with 1,000 bradyzoites. The evidence for this was that milk treated with HCl for 60 min at room temperature produced T. gondii infection in bioassayed mice. This result disagrees with the conclusion reached by Dubey and Frenkel (48), who found that bradyzoites were not formed in any tissue of mice until 7 days after bradyzoite inoculation.
Popiel et al. (134) used the acid-pepsin digestion procedure to quantify the development of bradyzoites in cell cultures by using cell culture as a bioassay. Tachyzoites of the T-263 strain, obtained by a 2-day cultivation in cell culture, were killed after a 10 min digestion in acid-pepsin. These authors used the same concentration of acid as used by Jacobs et al. (93) but only 10% of the pepsin used by Jacobs et al. (93). Bradyzoites produced in cell culture survived acid-pepsin digestion for 30 to 60 min. Popiel et al. (134) concluded that organisms that resisted 30 min of acid-pepsin digestion were bradyzoites.
Lindsay et al. (113) compared the appearance of acid- pepsin-resistant organisms with the development of tissue cysts by TEM in cell cultures inoculated with the RH strain and a temperature-sensitive (ts4) mutant derived from it. Tissue cysts were not seen in cell cultures inoculated with these two strains, but the organisms survived acid-pepsin digestion.
To resolve whether the inoculum used to test the effect of pepsin digestion, the strain of T. gondii used, the method used to obtain tachyzoites, and the source of tachyzoites affected results, Dubey (43) conducted experiments with extracellular tachyzoites from the peritoneal exudate obtained 3 to 9 days after i.p. inoculation of mice with tachyzoites. The following conclusions were drawn from this study: (i) tachyzoites occasionally survived acid-pepsin digestion for 2 h, which was not due to protection within host cells; (ii) the strain of T. gondii did not affect the susceptibility of tachyzoites to acid-pepsin; and (iii) even extracellular tachyzoites were infective to mice orally, but the infectivity was dose dependent (the infective dose of tachyzoites by the oral route in mice was 1,000). Therefore, it was concluded that one cannot rely on oral infectivity in mice or digestion in acid-pepsin as the sole criterion to distinguish between tachyzoites and bradyzoites.
†Contribution J5148 from the Montana State University Agricultural Experiment Station.