the timing of sulfadiazine therapy impacts the reactivation of latent toxoplasma infection in...

7
ORIGINAL PAPER The timing of sulfadiazine therapy impacts the reactivation of latent Toxoplasma infection in IRF-8 -/- mice Christian Jost & Ingrid Reiter-Owona & Oliver Liesenfeld Received: 30 June 2007 / Accepted: 23 July 2007 / Published online: 12 September 2007 # Springer-Verlag 2007 Abstract The process of reactivation of latent infection with Toxoplasma gondii in immunosuppressed hosts is yet not fully understood. In the past, a number of murine models of reactivation in immunocompromised mice have been described using sulfadiazine to establish latent infection before withdrawal and subsequent reactivation. We studied the process of reactivation in brains of mice with a targeted mutation in the interferon-regulatory factor (IRF)-8 gene after withdrawal of sulfadiazine therapy. IRF-8 -/- mice were orally infected with five cysts of the ME 49 strain of T. gondii. To allow establishment of latent infection with cyst formation, mice were treated with sulfadiazine starting either 3, 5, 6, or 7 days postinfection. Sulfadiazine was withdrawn after 1421 days to allow reactivation. We observed that timing of sulfadiazine therapy had a marked impact on the course of infection and reactivation. Mice treated late after infection (days 57) showed increased mortality and de- creased time to death compared to mice treated early after infection (group A). In the blood of mice with late (days 57) but not early (day 3) initiation of treatment, T. gondii-specific deoxyribonucleic acid was detected by polymerase chain reaction. Using double staining with stage-specific anti- bodies, tachyzoites were detectable in brains of mice with late initiation of sulfadiazine treatment but rarely within cysts thus indicating continued invasion of parasites across the bloodbrain barrier. Intracerebral cyst rupture or bradyzoitetachyzoite conversion was not detectable in IRF-8 -/- mice when sulfadiazine therapy was initiated late after infection. These results indicate that continued invasion of tachyzoites rather than reactivation of latent cerebral infection impacts the course of infection in this model of reactivated toxoplasmosis. In conclusion, the timing of sulfadiazine therapy is of utmost importance for the course of infection in immunocompromised mice. Introduction T. gondii is a protozoan parasite and causes asymptomatic Toxoplasma infection and severe toxoplasmosis (Montoya and Liesenfeld 2004). Acute infectioncharacterized by proliferation of tachyzoites within nucleated host cellsis self-limiting and benign in immunocompetent hosts. The host immune system mediates conversion of tachyzoites into bradyzoites resulting in cyst formation and establish- ment of latent infection. The parasite persists life-long in various organs including the central nervous system, heart, and sceletal muscle. In immunocompromised hosts, cyst rupture and conversion of bradyzoites into tachyzoites leads to uncontrolled tachyzoite proliferation with severe tissue damage that, if left untreated, results in death of the host by toxoplasmic encephalitis (TE). Reactivation remains a significant cause of death in acquired immunodeficiency syndrome (AIDS) patients, especially among those who do not have access to care, even in the era of highly active antiretroviral therapy (Bonnet et al. 2005). Reactivated toxo- plasmosis is also a serious complication in organ transplant recipients and occurs more commonly in recipients of allogeneic stem cell transplants than previously suggested (Martino et al. 2005). Cyst rupture in brains and subsequent infiltration of tachyzoites into the brain tissue has been Parasitol Res (2007) 101:16031609 DOI 10.1007/s00436-007-0700-y C. Jost : I. Reiter-Owona (*) Institut für Medizinische Mikrobiologie, Immunologie und Parasitologie, Universität Bonn, Sigmund-Freud Str. 25, 53105 Bonn, Germany e-mail: [email protected] O. Liesenfeld Institut für Mikrobiologie und Hygiene, Campus Benjamin Franklin, Charité Universitätsmedizin Berlin, Hindenburgdamm 27, 12203 Berlin, Germany

Upload: christian-jost

Post on 15-Jul-2016

212 views

Category:

Documents


0 download

TRANSCRIPT

ORIGINAL PAPER

The timing of sulfadiazine therapy impacts the reactivationof latent Toxoplasma infection in IRF-8−/− mice

Christian Jost & Ingrid Reiter-Owona & Oliver Liesenfeld

Received: 30 June 2007 /Accepted: 23 July 2007 / Published online: 12 September 2007# Springer-Verlag 2007

Abstract The process of reactivation of latent infection withToxoplasma gondii in immunosuppressed hosts is yet notfully understood. In the past, a number of murine models ofreactivation in immunocompromised mice have beendescribed using sulfadiazine to establish latent infectionbefore withdrawal and subsequent reactivation. We studiedthe process of reactivation in brains of mice with a targetedmutation in the interferon-regulatory factor (IRF)-8 geneafter withdrawal of sulfadiazine therapy. IRF-8−/− micewere orally infected with five cysts of the ME 49 strain ofT. gondii. To allow establishment of latent infection with cystformation, mice were treated with sulfadiazine starting either3, 5, 6, or 7 days postinfection. Sulfadiazine was withdrawnafter 14–21 days to allow reactivation. We observed thattiming of sulfadiazine therapy had a marked impact on thecourse of infection and reactivation. Mice treated late afterinfection (days 5–7) showed increased mortality and de-creased time to death compared to mice treated early afterinfection (group A). In the blood of mice with late (days 5–7)but not early (day 3) initiation of treatment, T. gondii-specificdeoxyribonucleic acid was detected by polymerase chainreaction. Using double staining with stage-specific anti-bodies, tachyzoites were detectable in brains of mice withlate initiation of sulfadiazine treatment but rarely within cyststhus indicating continued invasion of parasites across theblood–brain barrier. Intracerebral cyst rupture or bradyzoite–

tachyzoite conversion was not detectable in IRF-8−/− micewhen sulfadiazine therapy was initiated late after infection.These results indicate that continued invasion of tachyzoitesrather than reactivation of latent cerebral infection impactsthe course of infection in this model of reactivatedtoxoplasmosis. In conclusion, the timing of sulfadiazinetherapy is of utmost importance for the course of infectionin immunocompromised mice.

Introduction

T. gondii is a protozoan parasite and causes asymptomaticToxoplasma infection and severe toxoplasmosis (Montoyaand Liesenfeld 2004). Acute infection—characterized byproliferation of tachyzoites within nucleated host cells—isself-limiting and benign in immunocompetent hosts. Thehost immune system mediates conversion of tachyzoitesinto bradyzoites resulting in cyst formation and establish-ment of latent infection. The parasite persists life-long invarious organs including the central nervous system, heart,and sceletal muscle. In immunocompromised hosts, cystrupture and conversion of bradyzoites into tachyzoites leadsto uncontrolled tachyzoite proliferation with severe tissuedamage that, if left untreated, results in death of the host bytoxoplasmic encephalitis (TE). Reactivation remains asignificant cause of death in acquired immunodeficiencysyndrome (AIDS) patients, especially among those who donot have access to care, even in the era of highly activeantiretroviral therapy (Bonnet et al. 2005). Reactivated toxo-plasmosis is also a serious complication in organ transplantrecipients and occurs more commonly in recipients ofallogeneic stem cell transplants than previously suggested(Martino et al. 2005). Cyst rupture in brains and subsequentinfiltration of tachyzoites into the brain tissue has been

Parasitol Res (2007) 101:1603–1609DOI 10.1007/s00436-007-0700-y

C. Jost : I. Reiter-Owona (*)Institut für Medizinische Mikrobiologie,Immunologie und Parasitologie, Universität Bonn,Sigmund-Freud Str. 25, 53105 Bonn, Germanye-mail: [email protected]

O. LiesenfeldInstitut für Mikrobiologie und Hygiene, Campus BenjaminFranklin, Charité Universitätsmedizin Berlin,Hindenburgdamm 27, 12203 Berlin, Germany

proposed as the mechanism of reactivation of latentinfection with T. gondii (Frenkel and Escajadillo 1987; Luftand Remington 1992). However, the in vivo process of cystwall rupture and bradyzoite-to-tachyzoite stage conversionhas not been documented yet in detail. Extensive postmor-tem studies in brain tissue of AIDS patients with TE havenot revealed sites of cyst wall rupture or intracystic stageconversion (Reiter-Owona et al. 2000).

In the past, a variety of murine models have been devel-oped to study different aspects of TE. It is known thatresistance of mice to the disease is under genetic control(Suzuki 2002). BALB/c mice establish a latent infectionafter infection, whereas mice with a C57BL/6 backgroundare not able to completely control tachyzoite proliferationindependent of the Toxoplasma strain used (Suzuki 2002).It is also known that interferon (IFN)-γ-mediated immuneresponses control the parasite (Suzuki 2002). In the presentstudy, we therefore used mice with a targeted mutation in thegene encoding the IFN regulatory factor (IRF)-8 (IRF-8−/−)previously called INF consensus sequence-binding protein.These mice lack interleukin-12 (IL-12) p40 production andare unable to upregulate IFN-γ production upon infection(Scharton-Kersten et al. 1997). IRF-8−/− mice were reportedto develop reactivated toxoplasmosis similar to immuno-compromised hosts after discontinuation of sulfadiazine(SA) therapy (Scholer et al. 2001; Dunay et al. 2004). Weapplied this mouse model to investigate reactivation andattempted to visualize cyst wall rupture and early stages ofreactivation with bradyzoite-to-tachyzoite stage conversion.

Materials and methods

Mice and experimental infection

T. gondii Cysts of the ME 49 strain of T. gondii wereobtained from brains of NMRI mice (bred at the For-schungsinstitut für Experimentelle Medizin, Charité) 2 to3 months after intraperitoneal infection with ten cysts. Micewere killed by asphyxiation with CO2, and their brains wereremoved and triturated in phosphate-buffered saline (PBS,pH 7.2). An aliquot of the brain suspension was used tocount the number of cysts in the suspension. The suspensionwas diluted in 0.9% NaCl, and five cysts per 200 μl wereadministered orally by gavage.

Mice Inbred IRF-8−/− (C57/BL6 background), kindly pro-vided by I. Horak (Charité; Holtschke et al. 1996), and NMRImice were bred and maintained under specific pathogen-freeconditions. NMRI mice (Institut für Medizinische Mikro-biologie, Immunologie, und Parasitologie, Universität Bonn)chronically infected with the T. gondii strain DX (190 dayspostinfection [p.i.]) served as control animals.

Treatment groups All infected IRF-8−/− mice receiveddrinking water with SA (Sigma-Aldrich, Germany) at aconcentration of 200 mg/l ad libitum. Mice in group Awereadministered SA for 14 days starting at three days p.i.(Table 1). Mice in group B were treated with SA for 14 (B1)or 21 days (B2) starting at 5 days p.i. Mice in group Creceived SA for 14 (C1) or 21 days (C2) starting at 7 daysafter infection, whereas mice in group D received SA for21 days starting at 6 days p.i. Control mice were left untreated.

Histology For histological analyses, mice were chosen atrandom from all groups at indicated time points (Table 2).Brains were obtained and immediately processed forcryosectioning (groups A, B) and/or paraffin embedding(groups A–D). The brains were cross-sectioned, and thecaudal parts were cut longitudinally into two equal pieces.For cryosectioning, organs were mounted on Tissue- Tek®

(OCT Tissue-Tek®, Miles Laboratories, Naperville, IL),frozen in liquid nitrogen, and stored at −40°C. Cryosections(5 μm) were mounted on slides and fixed with cold acetone(10 min). For standard paraffin embedding, brains werefixed in formaldehyde (4% in 0.1 M PBS, pH 7.4) for 24 h.Selected paraffin sections (4–6 μm) were stained byhematoxylin and eosin (H&E).

Immunohistology Immunohistology was performed on se-rial paraffin sections after clearance in xylene and acetoneas described previously (Sahm et al. 1997). Sections werestained with a polyclonal rabbit anti-Toxoplasma serumdiluted 1:300 in PBS, pH 7.6. Tachyzoites and cysts wereexamined with mouse monoclonal anti-SAG1 (P30) anti-body GII (mouse IgG1, BioGenex, San Ramon, CA),diluted 1:50 in PBS, pH 7.6. Rat monoclonal antibody(mAb) CC2 (Bohne et al. 1995), diluted 1:100 in PBS,pH 7.6, was used to identify early cyst stages and a cystwall antigen (Ferguson 2004). After exposure to theprimary anti-Toxoplasma antibody, sections were incubatedwith StrAviGen Super-sensitive Immunodetection System(SS Multilink, BioGenex, Germany) according to themanufacturer’s instructions, followed by incubation with

Table 1 Experimental design and groups of mice used

Group Number ofIRF-8−/− micea

Start oftreatmentb

Treatment period

Control 5 n.a. n.a.A 8 3 days p.i 14 daysB 8 5 days p.i. 14 days (B1) 21 days (B2)C 8 7 days p.i. 14 days (C1) 21 days (C2)D 38 6 days p.i. 21 days

n.a. Not applicableaMice were infected orally with five cysts of the ME49 strain of T. gondiib 200 mg/l sulfadiazine in drinking water

1604 Parasitol Res (2007) 101:1603–1609

peroxidase-labeled secondary antibody (SS Multilink) andby diaminobenzidine (Liquide-DAB, BioGenex) as thechromogen. For double staining, CC2 was used as the firstprimary antibody (incubation 30 min), visualized by theStrAviGen system with Liquide DAB as the chromogen.SAG1 was applied for 45 min as a second primary antibodyfollowed by an alkaline phosphatase-labeled anti-rat second-ary antibody diluted 1:80 in PBS, pH 7.6 (DAKO Diag-nostika, Hamburg, Germany) with Fast Red as a couplingreagent (DAKO). Finally, slides were covered with Super-Mount® (BioGenex) and embedded in DEPEX (Serva,Heidelberg, Germany) after counterstaining with hematox-ylin. The sections were examined with a Leitz Dialux 20microscope, and pictures were taken with a Olympus SC 35camera. Cryosections were first exposed to the primary anti-Toxoplasma antibody (dilution see above, 45 min at 36°C)followed by a secondary, anti-species-specific IgG-fluores-cein isothiocyanate-conjugated antibody (anti-rabbit 1:60).Sections were examined with a Leitz Dialux 22 microscope.Specificity of immunohistological results was controlledwith antibody-negative sera from mouse, rat, and rabbitusing Toxoplasma-infected brains from IRF-8−/− and NMRImice. Between 30 and 100 sections were taken from eachsample (group A>100). After staining, the entire sections wereevaluated microscopically (10×, 25×, 40×, 100×). At the timeof SAwithdrawal, four, four, and three mice were analyzed ingroups A, B, and C, respectively. Twomice each were analyzedevery 12 h after withdrawal of SA in group D (Table 2).

Polymerase chain reaction Deoxyribonucleic acid (DNA)was extracted from 200 μl of mouse blood using theQIAamp blood mini kit (QIAGEN, Germany). Amplifica-tion was based on the 35-fold repetitive B1 gene of T. gondii(Burg et al. 1988) using primer pairs TOXOB22/TOXOB23(Bretagne et al. 1993) and 1/2 and 3/4 (Eggers et al. 1995).

At least two separate polymerase chain reaction (PCR)assays were run for each sample with these sets of primers.Reaction mixtures consisted of 200 μM of deoxyribonu-cleotide triphosphates each, 10 pmol of each of the twoprimers (TOXOB22/TOXOB23, 1/2, or 3/4), 10 mM KCl,10 mM Tris–HCl, pH 8.3, 2 mM MgCl2, 1.25 U ofAmpliTaq DNA-polymerase Stoffel-fragment (Perkin-Elmer), 10 μl of DNA solution, and sterile water to give afinal volume of 50 μl. Negative controls (sterile distilledwater instead of DNA solution as well as saline) and apositive control (QIAamp Blood kit extracted ToxoplasmaDNA, strain BK, corresponding to three tachyzoite DNAequivalents) were included. Samples were cycled as prev-iously described (Gross et al. 1992; Bretagne et al. 1993).Amplification products were separated on a 1.8% agarosegel, ethidium bromide stained, and analyzed under UV light.Intensities of the resulting amplification products wereexpressed as follows: + (low), ++ (medium), and +++ (high).

Results

Influence of sulfadiazine treatment on mortality,time to death, and cerebral parasite loadsin IRF-8−/− mice

To investigate whether the time of initiation and duration ofSA therapy impact the course of infection, we comparedmortality of IRF-8−/− mice treated with SA starting ondays 3, 5, 6, or 7 p.i. (Table 1, Fig. 1). Mortality was 100%in untreated control mice; all untreated mice had succumbedto infection by day 13 p.i. (Fig. 1). One hundred percentmortality was also observed when SA treatment was startedon days 5, 6, and 7. In these mice, late initiation of SA

Table 2 Parasite loads in T. gondii-infected IRF-8−/− mice

Groupsa Parasite brain foci and T. gondii-DNA at the time of sulfadiazinewithdrawal

Parasite brain foci after withdrawal of sulfadiazine

Days p.i. Parasite foci/sectionb (size) T. gondii DNAc

A (3 of 14) 17 0.07±0.02 (small) (n=4) Negative (n=4) Days 1–7 increasing number of small fociB1 (5 of 14) 19 0.13±0 (small) (n=2) ++ (n=2) n.d.B2 (5 of 21) 26 3.15±1.81 (medium–large) (n=2) ++ (n=2)C1 (7 of 14) 21 3.9±0.52 (small) (n=2) +++ (n=2)C2 (7 of 21) 28 15.1±3.77 (medium–large) (n=1d) +++ (n=1d)D (6 of 21) 27 11.0±1.79 (medium) (n=2) +++ (n=2) Days 1–4 number and size of foci increasing

n.d. Not determineda Numbers in parenthesis indicate the day p.i. when treatment started and the duration of treatment in days.b Numbers given represent the mean±standard deviation as determined in entire sections of brains of respective mice/group (for details, see“Materials and methods”).c T. gondii-specific DNAwas detected by PCR in blood samples as described in “Materials and methods.” ++, +++: Intensity of the amplification productd One moribund animal

Parasitol Res (2007) 101:1603–1609 1605

treatment was associated with decreased time to death; micein which treatment was started on days 5, 6, and 7 started todie by days 18, 12, and 9, respectively. Mice in group D beganto show clinical signs of reactivation (ruffled fur, weight loss,enhanced breathing frequency) 3 days after withdrawal of SA(day 30 p.i.). Mice were seriously ill 1 day later, and allanimals had succumbed to the infection by day 4 afterdiscontinuation of SA. When treatment was initiated earlier(day 3 p.i., group A), mice began to die by day 24 p.i., andmortality was 50% at the end of the study period (day 43).

The initiation of SA treatment thus impacts the time todeath and mortality in IRF8−/− mice. Histological analysisrevealed that decreased time to death caused by delayedonset of SA treatment was paralleled by increasing numbersof parasite foci in brains of mice at the time of SAwithdrawal (Table 2). When sections from mice in groupsB2 and C2 were examined, not only did the numbers ofparasitic foci increase but also the size of parasitic fociincreased compared to mice in group A. These results mostlikely indicate ongoing parasite replication in these micedespite SA application.

Presence of Toxoplasma DNA in blood samples

The concentration of Toxoplasma-specific DNA was belowthe detection limit of PCR in blood samples from mice ingroup A, whereas circulating Toxoplasma-specific DNAwas demonstrated in all blood samples from mice in groupsB, C, and D (Table 2). The intensity of amplification pro-

ducts increased when treatment started late after infection.These findings suggest continuous extracerebral parasitereplication.

Histological changes and parasite loadsafter discontinuation of sulfadiazine

When SA treatment was initiated on day 3 p.i. (group A),brains of infected mice harbored only a few small parasitefoci at the time when SA treatment was withdrawn (Table 2).Numbers of (small) cerebral parasite foci graduallyincreased after withdrawal of SA (Table 2). The concentra-tion of Toxoplasma-specific DNA was below the detectionlimit in blood samples of two animals killed 1 day after thefinal SA application. Early onset of SA treatment appears tolimit parasite replication in peripheral blood and thus resultsin low cyst numbers.

To visualize early stages of reactivation, sufficient numb-ers of cysts per brain section are required. Because in brainsof mice with early onset of SA treatment, only low parasitenumbers were found, we investigated brains of mice treatedwith SA for 21 days starting on day 6 p.i. (group D). Mice ingroup D showed high numbers of parasite foci at the time ofSA withdrawal; numbers and sizes of parasite foci increasedafter withdrawal of SA therapy (Table 2). CirculatingToxoplasma-specific DNA was detectable at high concen-trations in blood samples obtained at days 1–4 after SAwithdrawal (data not shown).

Presence of anti-SAG1- and CC2-positive parasitesin brains of IRF-8−/− mice with toxoplasmicencephalitis (group D)

Histological analysis revealed TE in all mice. Labeling withanti-SAG1 and CC2 revealed that CC2+ parasites werepredominant in brain tissues, whereas only few SAG1+ or

Fig. 1 Mortality and time to death in IRF-8−/− mice with latentToxoplasma infection. Mice were orally infected with five cysts of theME49 strain of T. gondii and treated with SA starting at days 3,5, 6, or7 after infection or left untreated. SA therapy was withdrawn atindicated time points. Asterisk, one moribund mouse in group C2 waskilled for histological analysis on day 28

Fig. 2 Immunohistological staining for Toxoplasma antigens. Tachy-zoites were identified with anti-SAG1 (P30) antibody GII. Cystdevelopment (early cystic stages and cyst wall antigen) was detectedusing antibody CC2; am. Amorphous structures

1606 Parasitol Res (2007) 101:1603–1609

SAG1+/CC2+ foci were detectable at the time of SA with-drawal (Fig. 2). Tissue cysts were often present in clumps,and some of them were of unusual size resembling cauli-flowers. Within the majority of these compartments, stain-ing for CC2 was localized between or on the surface ofparasites (most likely bradyzoites; Fig. 3a). Staining of cystwalls—indicating mature cysts (Sahm et al. 1997)—wasobserved in 20% of CC2+ foci. Walled compartments con-taining numerous SAG1−/CC2− organisms were alsodetected (Fig. 3b); these structures could be identified asdegenerated Toxoplasma cysts using H&E staining. Thesurrounding brain tissue appeared vacuolated, but fewmicroglia cells were present. In some mice, CC2+-amor-phous structures were detected that were not surrounded byintact cyst walls and resembled cysts in lysis; surroundingbrain tissues did not harbor individual parasites or inflam-matory infiltrates (Fig. 3c). We observed invasion of micro-glial cells, primarily located in areas where single ormultiple SAG1+ organisms and CC2+ foci were increased.Other areas of brains did not show increased inflammatoryreactions.

In those few areas of tachyzoite invasion, the surroundingbrain tissue had no evidence of necrosis, a common featureof TE in mice when numerous, fast-proliferating tachyzoitesare present.

Increasing numbers of tachyzoites were observed afterdiscontinuation of SA treatment in all sections examined.One and 2 days after withdrawal of SA, individual ormultiple tachyzoites within a parasitophorous vacuole wereoften located close to or even within meningeal blood vessels(Fig. 3d). Tachyzoites were detected at multiple sites of anindividual brain section. When multiple serial sections frombrains were examined, we were unable to detect a commonfocus from which all tachyzoites may have originated. Wetherefore conclude that proliferating tachyzoites most likelyinvaded the brain tissue from other (extracerebral) sites ofinfection. We did not find any evidence for bradyzoite/tachyzoite conversion within an intact cyst wall. Two walledcompartments, which reminded of a “budding” (mother–daughter) cyst, were revealed in serial sections as two sepa-rate cysts with parasites of different developmental stages(Fig. 3e). The smaller cyst most likely resulted from a recentinfection (SAG1+).

In the brain of one moribund mouse (day 4 after with-drawal of SA), brain cell disintegration was observed. Inaddition, cysts were remarkable for their staining pattern andshape: SAG1+ parasites were concentrated near the cyst walland close to the surrounding brain cells (Fig. 3f). These cysts(three to four in 12 of 60 sections) were preferentiallylocalized close to the meninges, and the surrounding cerebral

Fig. 3 Distinct stages of Toxoplasma gondii in the brains of IRF-8−/−

mice after SA treatment. a CC2 labeling in a mature Toxoplasma cyst,b degenerated unlabeled Toxoplasma cyst (c), bradyzoites cannot beidentified, c CC2+ amorphous structure; a cyst wall is not recogniz-

able, d single tachyzoite (arrow) within a meningeal blood vessel (bv),e very small SAG+ compartment (c1) in close vicinity to a mature cyst(c2) close to a meningeal blood vessel (bv), f SAG1+ compartments ina moribund animal at day 4 after withdrawal of SA

Parasitol Res (2007) 101:1603–1609 1607

parenchyma did not show free parasites. Assuming that theanti-SAG1 antibody is of high specificity, the describedpattern would be most compatible with intracystic stageconversion.

Discussion

Murine models employing the T. gondii strain ME 49 andmice of the C57Bl/6 genotype have been widely used tostudy pathogenesis and therapy of TE (Araujo and Slifer2003). The ME 49 strain disseminates and forms signifi-cantly more cysts than other strains before and after immu-nosuppression (Suzuki and Joh 1994). C57BL/6 mice arehighly susceptible to infection with T. gondii and developnecrotizing TE as result of a continuous chronic-progressiveinfection (Suzuki 2002). In the present study, latent infectionwas established by continuous administration of SA (chosenbecause of its reported benefit in maintaining latent in-fection; Scholer et al. 2001; Dunay et al. 2004) to IRF-8−/−

mice who do not produce IL-12p40 and are therefore unable toupregulate IFN-γ production (Scharton-Kersten et al. 1997).

To allow development of latent infection with a highnumber of brain cysts for later reactivation (and immuno-histological detection), we initiated treatment late (days 5–7)after infection with T. gondii in combination with a low SAconcentration in the drinking water. Unexpectedly, numbersand sizes of parasitic foci in brains were increasing underprolonged therapy especially when treatment started late.Immunolabeling with anti-SAG1 and anti-CC2 mAbsrevealed that the majority of the encysted foci detectedduring SA therapy were SAG1−/CC2+, and only few singleSAG1+ parasites were present. These results indicate an anti-parasitic effect of SA on tachyzoites but not on the encystedSAG1− parasites that continued to replicate. Distinct murinemodels of reactivation were reported to require individualtiming of onset of SA treatment to establish latent infection(Wang et al. 2005).

The low number of SAG1+ parasites detected in braintissue is in contrast to the increasing amount of T. gondii-specific DNA in peripheral blood thus suggesting ongoingextracerebral replication of parasites during SA treatment.In untreated immunocompetent mice, parasitemia maypersist for up to 2 months after infection (Miedouge et al.1997). In chronically infected IFN-γ−/− mice, continuousrelease of tachyzoites from cysts may result in theemergence of free tachyzoites and an increased number ofbrain cysts (Denkers and Gazzinelli 1998). In immunocom-promised patients, increased parasite loads in peripheralblood samples may precede TE (Martino et al. 2005). In thepresent study, we found no evidence for tachyzoite releasefrom brain cysts during treatment with SA. Despite thelimitations in sensitivity of histological techniques, we

hypothesize that those few SAG1+ parasites detected inbrains of mice most likely had recently invaded across theblood–brain barrier.

Immunohistology of brains in mice killed after discon-tinuation of SA treatment revealed that the lethal outcome ofinfection was not correlated with a widespread proliferationof tachyzoites in the brain. Even mice in which SA treatmentstarted early (day 3) showed increasing parasite foci despitenegative blood PCR results. The sensitivity of PCR in bloodmay be too low to detect intermittent parasitemia, oralternatively, reactivation occurred in brains of mice treatedearly with SA. It was speculated that the dysfunction ofmultiple organs was the major causes of death in IFN-γ−/−

mice after reactivation (Norose et al. 2001). After discon-tinuation of SA treatment, we expected a synchronousrelease of tachyzoites from cerebral cysts. However, usinganti-SAG1 and anti-CC2 mAbs, reactivation of cerebralinfection was characterized primarily by the presence ofsingle tachyzoites rather than by cyst wall rupture. Parasitesexpressing SAG1 were often detected close to meningealvessels indicating a hematogenous spread across the blood–brain barrier. These findings were supported by positive PCRresults in all blood samples obtained from mice in thosegroups treated late after infection (groups B–D). Extracere-bral bradyzoite-to-tachyzoite conversion has been reported inlungs and hearts of mice after treatment (Belal et al. 2004); incontrast, Suzuki et al. (2000) detected large amounts ofmessenger ribonucleic acid for SAG1 in brains but notlivers or spleens in IFN-γ−/− mice. These authors concludedthat the brain is the major organ in which proliferation oftachyzoites occurs. However, Saeij et al. (2005) recentlydemonstrated differences in the dissemination pattern ofdifferent Toxoplasma strains (genotypes) after reactivation.

Our stage conversion studies may also be limited becausethe antibody CC2 is not regarded as stage specific and is nota bradyzoite marker (Ferguson 2004). The antigen recog-nized by CC2 is distributed throughout the granularmaterial of the cyst matrix (Sahm et al. 1997) that laterforms the cyst wall and colocalizes with the lectin Dolichosbiflorus agglutinin at the tissue cyst wall (Ferguson 2004).CC2 cross-reacts with cellular debris within parasiticlesions (Ferguson 2004) and with parasites in lysis(Reiter-Owona et al. 2000). Bradyzoites were indirectlyidentified by the demonstration of CC2-labeled cysts. Abradyzoite-specific marker would therefore not necessarilychange results from this study using the CC2 antibody.

“Cysts” containing SAG1+ parasites as a marker forearly bradyzoite-to-tachyzoite conversion were onlydetected in the brain of one individual moribund mouse.The brain tissue showed massive tachyzoite infiltration andsigns of brain cell disintegration. In several compartments,SAG1+ parasites overlayed the suspected cyst wall sug-gesting that the wall had been lyzed by the activity of

1608 Parasitol Res (2007) 101:1603–1609

endogenous enzymes released from the necrotic host cell(Frenkel and Escajadillo 1987; Hulinska et al. 1990).

In conclusion, the present study revealed that onset andduration of SA treatment (and probably the concentration ofSA) are major factors that influence the course of Toxoplasmainfection in immunodeficient mice. Late onset of SA therapymost likely results in persistent parasitemia and invasion ofparasite into the brain rather than reactivation of infectionfrom brain cysts.

Acknowledgments We greatly acknowledge the expert technicalassistance of Andrea Maletz and Ildiko R. Dunay. The monoclonalantibody CC2 is a gift from Uwe Gross. This work was in partsupported by a grant from the Deutsche Forschungsgemeinschaft,FG463, Projekt 7A to O.L.

References

Araujo FG, Slifer T (2003) Different strains of Toxoplasma gondiiinduce different cytokine responses in CBA/Ca mice. InfectImmun 71:4171–4174

Belal US, Norose K, Aosai F, Mun HS, Ahmed AK, Chen M,Mohamed RM, Piao LX, Iwakura Y, Yano A (2004) Evaluationof the effects of sulfamethoxazole on Toxoplasma gondii loadsand stage conversion in IFN-gamma knockout mice using QC-PCR. Microbiol Immunol 48:185–193

Bohne W, Gross U, Ferguson DJ, Heesemann J (1995) Cloning andcharacterization of a bradyzoite-specifically expressed gene(hsp30/bag1) of Toxoplasma gondii, related to genes encodingsmall heat-shock proteins of plants. Mol Microbiol 16:1221–1230

Bonnet F, Lewden C, May T, Heripret L, Jougla E, Bevilacqua S,Costagliola D, Salmon D, Chene G, Morlat P (2005) Opportu-nistic infections as causes of death in HIV-infected patients in theHAART era in France. Scand J Infect Dis 37:482–487

Bretagne S, Costa JM, Vidaud M, Tran J, Nhieu V, Fleury-Feith J(1993) Detection of Toxoplasma gondii by competitive DNAamplification of bronchoalveolar lavage samples. J Infect Dis168:1585–1588

Burg JL, Perelman D, Kasper LH, Ware PL, Boothroyd JC (1988)Molecular analysis of the gene encoding the major surfaceantigen of Toxoplasma gondii. J Immunol 141:3584–3591

Denkers EY, Gazzinelli RT (1998) Regulation and function of T-cell-mediated immunity during Toxoplasma gondii infection. ClinMicrobiol Rev 11:569–588

Dunay IR, Heimesaat MM, Bushrab FN, Muller RH, Stocker H,Arasteh K, Kurowski M, Fitzner R, Borner K, Liesenfeld O(2004) Atovaquone maintenance therapy prevents reactivation oftoxoplasmic encephalitis in a murine model of reactivatedtoxoplasmosis. Antimicrob Agents Chemother 48:4848–4854

Eggers C, Gross U, Klinker H, Schalke B, Stellbrink HJ, Kunze K(1995) Limited value of cerebrospinal fluid for direct detection ofToxoplasma gondii in toxoplasmic encephalitis associated withAIDS. J Neurol 242:644–649

Ferguson DJ (2004) Use of molecular and ultrastructural markers toevaluate stage conversion of Toxoplasma gondii in both theintermediate and definitive host. Int J Parasitol 34:347–360

Frenkel JK, Escajadillo A (1987) Cyst rupture as a pathogenic mechanismof toxoplasmic encephalitis. Am J Trop Med Hyg 36:517–522

Gross U, Roggenkamp A, Janitschke K, Heesemann J (1992) Improvedsensitivity of the polymerase chain reaction for detection ofToxoplasma gondii in biological and human clinical specimens.Eur J Clin Microbiol Infect Dis 11:33–39

Holtschke T, Lohler J, Kanno Y, Fehr T, Giese N, Rosenbauer F, LouJ, Knobeloch KP, Gabriele L, Waring JF, Bachmann MF,Zinkernagel RM, Morse HC 3rd, Ozato K, Horak I (1996)Immunodeficiency and chronic myelogenous leukemia-like syn-drome in mice with a targeted mutation of the ICSBP gene. Cell87:307–317

Hulinska D, Sykora J, Zastera M (1990) Effect of cortisone onToxoplasma gondii infection studied by electron microscopy.Folia Parasitol (Praha) 37:207–212

Luft BJ, Remington JS (1992) Toxoplasmic encephalitis in AIDS.Clin Infect Dis 15:211–222

Martino R, Bretagne S, Einsele H, Maertens J, Ullmann AJ, Parody R,Schumacher U, Pautas C, Theunissen K, Schindel C, Munoz C,Margall N, Cordonnier C (2005) Early detection of Toxoplasmainfection by molecular monitoring of Toxoplasma gondii inperipheral blood samples after allogeneic stem cell transplanta-tion. Clin Infect Dis 40:67–78

Miedouge M, Bessieres MH, Cassaing S, Swierczynski B, Seguela JP(1997) Parasitemia and parasitic loads in acute infection and afteranti-gamma-interferon treatment in a toxoplasmic mouse model.Parasitol Res 83:339–344

Montoya JG, Liesenfeld O (2004) Toxoplasmosis. Lancet 363:1965–1976Norose K, Mun HS, Aosai F, Chen M, Hata H, Tagawa Y, Iwakura Y,

Yano A (2001) Organ infectivity of Toxoplasma gondii ininterferon-gamma knockout mice. J Parasitol 87:447–452

Reiter-Owona I, Seitz H, Gross U, Sahm M, Rockstroh JK, Seitz HM(2000) Is stage conversion the initiating event for reactivation ofToxoplasma gondii in brain tissue of AIDS patients? J Parasitol86:531–536

Saeij JP, Boyle JP, Boothroyd JC (2005) Differences among the threemajor strains of Toxoplasma gondii and their specific interactionswith the infected host. Trends Parasitol 21:476–481

Sahm M, Fischer HG, Gross U, Reiter-Owona I, Seitz HM (1997)Cyst formation by Toxoplasma gondii in vivo and in brain-cellculture: a comparative morphology and immunocytochemistrystudy. Parasitol Res 83:659–665

Scharton-Kersten T, Contursi C, Masumi A, Sher A, Ozato K (1997)Interferon consensus sequence binding protein-deficient micedisplay impaired resistance to intracellular infection due to aprimary defect in interleukin 12 p40 induction. J Exp Med186:1523–1534

Scholer N, Krause K, Kayser O, Muller RH, Borner K, Hahn H,Liesenfeld O (2001) Atovaquone nanosuspensions show excel-lent therapeutic effect in a new murine model of reactivatedtoxoplasmosis. Antimicrob Agents Chemother 45:1771–1779

Suzuki Y (2002) Immunopathogenesis of cerebral toxoplasmosis. JInfect Dis 186(Suppl 2):S234–S240

Suzuki Y, Joh K (1994) Effect of the strain of Toxoplasma gondii onthe development of toxoplasmic encephalitis in mice treated withantibody to interferon-gamma. Parasitol Res 80:125–130

Suzuki Y, Kang H, Parmley S, Lim S, Park D (2000) Induction oftumor necrosis factor-alpha and inducible nitric oxide synthasefails to prevent toxoplasmic encephalitis in the absence ofinterferon-gamma in genetically resistant BALB/c mice.Microbes Infect 2:455–462

Wang X, Claflin J, Kang H, Suzuki Y (2005) Importance of CD8(+)Vbeta8(+) T cells in IFN-gamma-mediated prevention of toxo-plasmic encephalitis in genetically resistant BALB/c mice. JInterferon Cytokine Res 25:338–344

Parasitol Res (2007) 101:1603–1609 1609