dmatpnpnicttbTlcamripca

g

Lsaia

d4E

Clinical ImmunologyVol. 92, No. 1, July, pp. 56–66, 1999Article ID clim.1999.4716, available online at http://www.idealibrary.com on

1CA

Chagas’ Disease Encephalitis: Intense CD81 Lymphocytic Infiltrate IsRestricted to the Acute Phase, but Is Not Related to the

Presence of Trypanosoma cruzi Antigens

Andrea A. Silva,*,† Ester Roffe,* Ana P. Marino,* Paula V. A. dos Santos,* Thereza Quirico-Santos,†Claudia N. Paiva,*,‡ and Joseli Lannes-Vieira*,1

*Department of Immunology, Oswaldo Cruz Institute–Fiocruz, 21045-900 Rio de Janeiro, Rio de Janeiro, Brazil; †Department ofPathology, Fluminense Federal University, Niteroi, Rio de Janeiro, Brazil; and ‡Department of Immunology,

Institute of Biophysics Carlos Chagas Filho, UFRJ, Brazil

tdtirptaiiComCg

ttiuplTpgvrflgreiplctis

Central nervous system (CNS) damage can occururing Chagas’ disease, especially in children and im-unosuppressed patients. During the acute phase,

mastigotes are rarely found, but inflammatory infil-rates are scattered throughout the CNS. Moreover,eripheral lymphocytes and antibodies recognizingeural components were described, suggesting thearticipation of the immune system in the genesis ofeural lesions. Herein, we performed a histopatholog-cal study of Colombian-infected C3H/He mice,omparing the distribution of CNS-inflammatory infil-rates versus Trypanosoma cruzi antigens. Inflamma-ory infiltrates were observed during the acute phase,ut did not correlate with the presence of detectable. cruzi antigens. Infiltrates consisted mainly of CD81

ymphocytes, although macrophages and a few CD41

ells were observed. In the chronic stage of infection,lthough neuropathies were a common finding, onlyild inflammatory infiltrates could be detected. Our

esults suggest that the presence of CNS inflammatorynfiltrates is not directly related to the presence ofarasite antigens and indicate that, different fromhronic myocarditis, encephalitis resolves during thecute phase of Chagas’ disease. © 1999 Academic Press

Key Words: encephalitis; Trypanosoma cruzi; Cha-as’ disease; central nervous system; CD8 T cell.

INTRODUCTION

Chagas’ disease, one of the leading causes of death inatin America, is caused by transmission of Trypano-oma cruzi. The disease presents an acute stage char-cterized by active infection, immunological abnormal-ties, and histopathological lesions. Most pathologicallterations detected in this stage have been attributed

1 To whom correspondence should be addressed at Departamentoe Imunologia–Fundacao Oswaldo Cruz (FIOCRUZ), Av. Brasil,365, Rio de Janeiro, RJ, Brazil, 21045-900. Fax: 55-21-2801589.

t-mail: lannes@gene.dbbm.fiocruz.br.

521-6616/99 $30.00opyright © 1999 by Academic Pressll rights of reproduction in any form reserved.

56

o the presence of the parasite. In the chronic stage,espite the presence of abundant inflammatory infil-rates in multiple organs, low parasitism is found innfected humans and animals (1), casting doubt on theole played by parasite-induced tissue damage in theathogenesis of chronic Chagas’ disease. Moreover, au-oimmune reactions against the heart, striated muscle,nd neural components occur during this stage, lead-ng some authors to propose the existence of an auto-mmune component in the pathogenesis of chronichagas’ disease (2, 3). The postulated autoimmunerigin of chronic phase cardiomyopathy has receiveduch attention in the past 10 years, but the study ofhagas’ disease encephalitis has been somewhat ne-lected.Since its description by Carlos Chagas in 1909 (4),

he nervous form of Chagas’ disease has been difficulto characterize. Meningoencephalitis rarely occurs, be-ng more frequent during acute infection of childrennder 2 years and in immunosuppressed patients, whoresent reactivated infection (5–8). The neuropatho-ogic picture produced by experimental infection with. cruzi is similar in many aspects to that observed inatients with the acute nervous form (9). In this re-ard, similar to humans, dogs (10) and mice (11) de-elop encephalitis during acute infection, presentingandom distribution and nodular arrangement of in-ammatory mononuclear cells. In this stage, amasti-ote forms and parasite antigens were detected in neu-ons, glia, and microglia in both humans andxperimental animals (12, 13). The involvement of themmune system in the genesis of neural lesions wasroposed on the basis of the existence of peripheralymphocytes and antibodies specific for nervous systemomponents in infected animals (14, 15), but a correla-ion between the presence of parasite antigens andnflammatory infiltrates within the central nervousystem (CNS) has not been attempted so far. Moreover,

he composition of such infiltrates in terms of lympho-

cis

ctda

M

fCwtpap

A

fto1aCpnsAmn

H

aotbLtawi

I

vogdwt(tacbsparptdiaoi

wwCtTdAi

mti

pt

57ENCEPHALITIS IN ACUTE CHAGAS’ DISEASE

yte subsets remains unknown and the persistence ofnflammation during the chronic phase is a controver-ial question (16, 11).In an effort to shed light on these questions, we

haracterized the distribution and nature of inflamma-ory infiltrates present in the CNS during Chagas’isease and compared it to the distribution of T. cruzintigens, mapped immunohistochemically.

MATERIALS AND METHODS

ice and Parasites

Female C3H/He mice (5–7 weeks old) were obtainedrom the Animal Facilities, Bio-Manguinhos, Oswaldoruz Foundation. Mice were infected intraperitoneallyith 100 blood trypomastigote forms of the Colombian

ype III strain of T. cruzi (17), maintained by serialassage in C3H/He mice. Parasitemia was estimatedccording to Brener’s method (18) and employed as aarameter to establish acute and chronic phases.

ntibodies

Specific antiserum recognizing T. cruzi was a giftrom Dr. Rosa Teixeira de Pinho (Oswaldo Cruz Insti-ute, Brazil). Biotinylated monoclonal antibodies rec-gnizing the molecule CD8 (clone 53-6.7) and CD4 (GK.5) were purchased from PharMingen. Anti-macroph-ge (F4/80) monoclonal antibody was purchased fromaltag and anti-T cell receptor ab (clone H57-597) wasurchased from Gibco. Biotinylated antibodies recog-izing rat or rabbit immunoglobulin and peroxidase–treptavidin complex were purchased from Amersham.ppropriate controls were prepared by replacing pri-ary antibodies with purified rat immunoglobulin or

FIG. 1. Parasitemia curve of C3H/He mice infected with 100arasites of the Colombian strain of T. cruzi. Each point representshe mean of five mice. Standard deviation is plotted.

ormal rabbit serum, as indicated in the figures. d

istopathological Studies

Groups of five C3H mice were sacrificed under ethernesthesia at various intervals postinfection. Groupsf three age-matched control mice were sacrificed athe same intervals. The encephalon was removed, em-edded in tissue-freezing medium (Tissue Tek, Milesaboratories) and stored in liquid nitrogen. Serial sec-ions 5- to 7-mm thick were prepared by sagital cutsnd fixed in cold acetone. These sections were stainedith hematoxylin and eosin or submitted to indirect

mmunoperoxidase.

mmunohistochemistry

The indirect immunoperoxidase technique was de-eloped as previously described (19). Briefly, serial cry-stat sections were mounted on poly-L-lysine-coveredlass slides and fixed for 10 min in cold acetone. En-ogenous peroxidase and nonspecific antibody bindingere blocked incubating the specimens with PBS con-

aining 0.1% sodium azide and normal goat serumdiluted 1/50). Next, we performed sequential incuba-ions with primary unlabeled antibodies (anti-T. cruzintigen, anti-macrophage, anti-CD4, anti-CD8, or spe-ies-matched control Ig’s), secondary biotinylated anti-odies (goat anti-rabbit Ig or goat anti-rat Ig), andtreptavidin–peroxidase complex. All incubations wereerformed for 1 h with antibodies diluted in BSA–PBSnd were followed by washes in PBS. The peroxidaseeaction was developed with aminoethylcarbazole inresence of hydrogen peroxide. The material was coun-erstained with Mayer’s hematoxylin and analyzed un-er a light microscope. Sections of hearts from acutelynfected mice were used as positive controls for specificnti-T. cruzi antigen staining, revealing the presencef pseudocysts, while spleen sections were used as pos-tive controls for lymphocyte staining.

RESULTS

Figure 1 shows the course of parasitemia observedhen highly susceptible C3H/He mice were infectedith 100 parasites of the Colombian strain of T. cruzi.irculating parasites were detected 14 days postinfec-

ion (dpi) and the peak of parasitemia occurred 42 dpi.rypomastigotes were rarely found in the blood at 63pi, characterizing the onset of the chronic phase.round 40% of the infected mice died during acute

nfection.Histopathological studies showed edema, enlarge-ent of perivascular spaces, focal meningoencephali-

is, and perivascular and parenchymal mononuclearnfiltrates irregularly distributed throughout the CNS

uring the acute phase. Areas of incomplete blood–

bhti(voBl

ic3cflb

c

(H

58 SILVA ET AL.

rain barrier, such as choroid plexus (Fig. 2b) andippocampus (Fig. 2e), presented intense inflamma-ory infiltrates during acute infection, while only mildnfiltrates were found in chronically infected animalsFigs. 2c and 2f). Normal controls were completely de-oid of cellular infiltrates in these (Figs. 2a and 2d) orther encephalic regions studied (data not shown).rain parenchyma, perivascular spaces, and cerebel-

FIG. 2. Inflammatory infiltrates in choroid plexus and hippocaright) hippocampus. (a, d) Normal mice; (b, e) Acutely infected (42 d&E. Original magnification: (left) 3400; (right) 3250.

um were intensively infiltrated by mononuclear cells t

n acutely infected animals (Figs. 3a–3c), but not inhronic chagasic survivors (90 dpi), as shown in Figs.d–3f. In Table 1, we summarize individual data con-erning occurrence, intensity, and distribution of in-ammatory infiltrates within particular CNS areas inoth acutely and chronically infected mice.A few infected animals presented visible neurologi-

al symptoms, which were not correlated with the in-

us of acutely and chronically infected mice. (Left) Choroid plexus;(c, f) Chronically infected mice (90 dpi). Sections were stained with

mppi);

ensity of CNS inflammation. These symptoms com-

ptspltcc

ni

fwrw

AS d e

59ENCEPHALITIS IN ACUTE CHAGAS’ DISEASE

rised sequential expression of (1) a tendency to fall oro adopt a circular motion, always directed to the sameide, (2) humpback walk, (3) loss of tail tonus, and (4)aralysis, affecting mainly the hind limbs. Neverthe-ess, some individuals showed only some of these symp-oms during disease progression. Different from en-ephalitis, the incidence of neurological symptoms was

FIG. 3. Inflammatory infiltrates in brain parenchyma, perivasculcutely infected mice (42 dpi); (right) chronically infected mice (90 dpections were stained with H&E. Original magnification: (a, b, d, an

rescent from the acute (7 dpi, 0% n 5 97; 28 dpi, 2.6% c

5 77; 42 dpi, 8% n 5 62) to the chronic stage ofnfection (90 dpi, 12% n 5 57).

Previous histological studies showed that amastigoteorms of T. cruzi are either not detected or rarely foundithin inflammatory foci or inside glial cells and neu-

ons (6). In an attempt to visualize parasite antigens,e labeled tissue specimens with the anti-T. cruzi poly-

pace, and cerebellum of acutely and chronically infected mice. (Left)(a, d) Brain parenchyma; (b, e) perivascular space; (c, f) cerebellum.) 3400; (c) 3200; and (f ) 3100.

ar si).

lonal antiserum. The present immunohistochemical

sTTcacrigapatCtpwapnhcesi(peib

flnffc

caCiFCaraattv8cbn

sidiasntaivtnr

6 de

60 SILVA ET AL.

tudies revealed a random and irregular distribution of. cruzi antigen reactivity in the CNS of infected mice.he presence of parasite antigens was restricted toells scattered throughout the CNS–parenchyma of allcutely infected animals. Cells stained positive for T.ruzi antigens were most frequently devoid of sur-ounding inflammatory reaction (Fig. 4). Further, mostnflammatory infiltrates were devoid of T. cruzi anti-ens (Figs. 5a and 5b), although parasite antigens werelso found within inflammatory infiltrates as an amor-hous material (Figs. 5c and 5d). The anti-T. cruzintigen specificity of this labeling was confirmed, sincehe polyclonal serum employed did not cross-react withNS normal sections (Fig. 4c). Different from our con-

rol heart sections of acutely infected mice, where theresence of T. cruzi antigens was frequently associatedith pseudocysts, cells stained positive for T. cruzintigens within the CNS did not contain detectableseudocysts or isolated amastigotes. In fact, we wereot able to detect parasites within the CNS, despite theigh levels of parasitemia and the maximization ofhances provided by anti-T. cruzi antigen staining. Thexhaustive search for amastigotes in the sectionstained with hematoxylin and eosin was negative in allnfected animals except in one 28-day-infected mouseFigs. 5e and 5f). In the chronic stage of infection,arasite antigens were hardly detected and their pres-nce was not associated with the rare inflammatorynfiltrates found. These results are summarized in Ta-le 2.The phenotypical nature of the cells present in in-

ammatory infiltrates was characterized by an immu-ohistochemical assay. During the acute phase of in-ection both antigen-containing (Fig. 6a) and antigen-ree (Fig. 6b) inflammatory infiltrates were mainly

TABSummary of Individual Data Concerning Occurrence

in both Acutely and C

Dayspostinfection

Individualmice

Brainparenchyma Meninges

14 1 2a 12 2 1

28 1 11 112 11 113 111 11

42 1 111 1112 111 1113 11 11

90 1 6 12 2 23 2 1

a The numbers of mononuclear cells infiltrating each encephalic are, very few; 2, no cells were found in the indicated region; ND, not

omposed of macrophages (Figs. 6c and 6d) and CD81 c

ells (Figs. 6e and 6f), although CD41 cells (Figs. 6gnd 6h) were also detected. The predominance of theD81 cell subset was observed since the first stages of

nfection (14 dpi) in all encephalic regions analyzed. Inig. 7, we show that these lymphocytes outnumberD41 cells by a factor of 2–3 in leptomeninges ofcutely infected mice (42 dpi). Cells expressing T celleceptor b chain were detected in the correspondingreas, indicating that most of the CD81 and CD41 cellsre ab T lymphocytes (data not shown). Regardless ofhe CNS areas analyzed, macrophages were also de-ected (Fig. 8c), identifying cells correspondingly de-oid of CD8 or CD4 expression in serial sections (Figs.a and 8b). Similar to those in the acute phase, thehronic stage inflammatory lesions are characterizedy a predominance of CD81 cells and fewer or virtuallyo CD41 (Table 2).

DISCUSSION

In this paper we characterized the inflammatory le-ions present in the CNS during experimental Chagas’nfection, emphasizing the lack of correlation betweenistribution of mononuclear cell infiltrates and local-zation of T. cruzi-antigens. In fact, although both par-site antigens and inflammatory infiltrates do nothow a preferential localization, they frequently didot coincide. This finding suggests that at least part ofhe lymphocytic infiltrates are not directed to T. cruzi-ntigens. One possibility to explain our data is that thenfiltrate is composed of nonspecific bystander acti-ated cells that migrate through the endothelial cells ofhe blood vessels and extravasate into the CNS. Alter-atively, antigenic mimicry between T. cruzi and neu-al components (2) or polyclonal lymphocyte activation

1tensity, and Distribution of Inflammatory Infiltratenically Infected Mice

Choroidplexus Hippocampus

Perivascularspace Cerebellum

1 1 6 22 1 2 2

ND 11 11 1111 111 111 1111 111 11 11

111 11 111 111111 111 111 111111 111 11 11

1 1 2 21 1 2 22 2 2 2

ere categorized into five grades: 111, many; 11, moderate, 1, few;monstrated.

LE, Inhro

a w

aused by infection (20) could result in transient auto-

raasafpfn

dhepsmpsssofto

fCi(apCccctlocTmdCadsi

di

af

TCCM

i

apsaitM

61ENCEPHALITIS IN ACUTE CHAGAS’ DISEASE

eactivity to sequestered neural components, such asutosensitization to myelin in mice (21). In this regard,ntibodies against neural components have been de-cribed (22, 15) and we recently found antibodiesgainst myelin basic protein and its encephalitogenicragment in infected mice (Silva et al., manuscript inreparation). Moreover, splenic CD41 T cells specificor peripheral nerve components that induce demyeli-

FIG. 4. Presence of T. cruzi antigens in brain parenchyma ofcutely infected mice (28 dpi). (a) shows a group and (b) a singleositive cell detected with anti-T. cruzi polyclonal serum. (c) is aection of normal brain parenchyma stained likewise, depicting thebsence of cross-reactive staining. Arrows indicate cells stained pos-tive for T. cruzi antigens, devoid of surrounding mononuclear infil-rates. Immunohistochemistry is described under Materials andethods. Original magnification 3200.

ation and inflammatory lesions in sciatic nerves were 1

escribed in T. cruzi-infected mice (14). On the otherand, it is possible that the antigen-free infiltratesxhibiting high numbers of mononuclear cells mightresent parasite antigens at other levels of the tissueince only 5–10 serial sections of the CNS tissue wereade. Also, we cannot exclude the possibility that our

olyclonal anti-T. cruzi serum fails to detect particularhedded-parasite antigens, which could be able to sen-itize neural cells and trigger an immunological re-ponse, later turned off by antigen clearance. Finally,ne could not exclude the possibility that the antigen-ree perivascular and parenchymal inflammatory infil-rates, at least in part, are the result of ischemia sec-ndary to microvascular compromise (23).The predominance of CD81 lymphocytes in antigen-

ree and antigen-containing infiltrates distinguisheshagas’ disease encephalitis from autoimmune exper-

mental allergic encephalomyelitis, multiple sclerosis24), postinfectious autosensitization to myelin (21),nd African trypanosomiasis (8), in which CD41 lym-hocytes predominate. Different from our findings inhagas’ disease, the encephalitis caused by the intra-ellular protozoan Toxoplasma gondii and by lympho-ytic choriomeningitis virus (LCMV) involves lympho-ytic infiltrates predominantly composed of CD81 cellshat surround pathogen-infected cells (25, 26). In fact,ethal LCMV encephalitis is caused by the recognitionf LCMV antigens on infected target cells expressinglass I MHC, leading to extensive CNS damage (27).hus, the presence of antigen-free infiltrates in ourodel suggests that Chagas’ disease encephalitis has a

istinct pathogenesis, although the role played byD81 cells is still unknown. However, as discussedbove, the possibility that our polyclonal serum fails toetect particular antigens which could be able to sen-itize neural cells and trigger CD81-mediated cytotox-city remains to be tested.

The sharp contrast between the intense scarring car-iac muscle inflammation and the mild resolving CNSnflammation found in Colombian-infected chronic cha-

TABLE 2

Summary of Occurrence and Intensity of T. cruzi Antigensnd Mononuclear Cells in both Acutely and Chronically In-ected Mice

Uninfected

Stage of disease

Acute Chronic

. cruzi antigen a2 111 2D8 2 11 6D4 2 1 2acrophage 6 11 1

a The presence of T. cruzi antigens and mononuclear cells infiltrat-ng nervous tissue was categorized into five grades: 111, many;

1, moderate, 1, few; 6, very few; 2, no positive cells were found.

gtiaaac

tvbsps

mscatdam f) 3

62 SILVA ET AL.

asic C3H/He mice is noteworthy. Analysis of distribu-ion of lymphocytic infiltrates versus parasite antigensn infected hearts from chronic chagasic patients (28)nd mice (29) demonstrated that CD81 lymphocytesre found close to parasite antigens, while CD41 cellsre spread throughout the heart tissue, although asso-

FIG. 5. The presence of T. cruzi antigens is unrelated to the preseice (28 dpi). (a) shows the absence of antigen in an inflammatory in

howed in (a). In (c) the detection of parasite antigens is shown as anounterstained with Mayer’s hematoxylin. The same area is depicted28 days-infected mouse is shown. (f) shows that small rounded str

he corresponding antigen-positive area shown in (e). Sections shownetection of parasite antigens using an anti-T. cruzi polyclonal serumnd (f) were stained with hematoxylin and eosin. Arrows indicatononuclear infiltrates. Original magnification: (a–d) 3250; (e and

iation between the quantity of parasite antigens and c

he number of CD81 T cells was not observed. Con-ersely, CNS inflammation rarely involves proximityetween lymphocytes and T. cruzi antigens, and para-ites could hardly be detected by us in the brain. Theathologic picture of heart versus CNS inflammationuggests different origins and reveals different out-

of inflammatory infiltrates in brain parenchyma of acutely infectedrate, (b) depicts the presence of inflammatory cells in the same areaorphous material stained in red within the inflammatory infiltrate(d). In (e) a positive staining for parasite antigens in cerebellum of

ures resembling amastigote forms of the T. cruzi were identified in(a), (c), and (e) were submitted to an immunohistochemical assay fordescribed under Materials and Methods. Sections shown in (b), (d),

ells stained positive for T. cruzi antigens, devoid of surrounding1000.

ncefiltamin

uctinas

e c

omes for these inflammatory processes. Although we

cpau

63ENCEPHALITIS IN ACUTE CHAGAS’ DISEASE

FIG. 6. Serial sections of the brain tissue of acutely infected mice (42 dpi) showing immunoperoxidase staining for T. cruzi antigens andell surface markers of T lymphocytes infiltrating leptomeninges. The parasite antigens are detected as a red amorphous staining and theositive cells are identified by a partial or complete ring of dark color outlining the cell membrane. Serial sections show (a and b) parasitentigens staining; (c and d) macrophages; (e and f) CD81 lymphocytes; and (g and h) CD41 lymphocytes. Immunohistochemistry is described

nder Materials and Methods. Original magnification 3250.

chcuaflhmsgtiswcdrstmtsinpl

ussCCtp(mttwipaBott

(oimsnn

64 SILVA ET AL.

annot exclude a bias in our observations caused byigh mortality rates in the acute phase among CNS-ompromised mice, leading to survival of those whonderwent mild encephalitis, we must emphasize thatll acutely infected mice analyzed presented CNS in-ammation, implying that chronic phase survivorsave undergone encephalitis to some degree. Further-ore, recent studies developed in our laboratory

howed a predominance of CD81 lymphocytes in anti-en-free and antigen-containing infiltrates found inhe myocardium of chronically infected C3H mice usedn the present study (P. V. Alves dos Santos, manu-cript in preparation). Our results are in agreementith previous data showing a predominance of CD81

ells in inflammatory infiltrates present in the myocar-ium of chagasic patients (30, 28) and T. cruzi-infectedats (31) and mice (32, 33). Other authors demon-trated CD41 cells as the predominant T cell popula-ion in these lesions (14, 16, 29). All these differencesay be due to the different models of infection used in

hese studies. However, one could not exclude the pos-ibility that they reflect technical problems in detect-ng the cell markers. Anyhow, further studies areeeded to fully define the functions of CD81 cellsresent in the CNS and heart muscle inflammatoryesions in experimental infection and Chagas’ disease.

In other experimental models of chagasic infectionsing different T. cruzi strains, mice strains, or otherpecies, pathological alterations in the CNS were es-entially observed during the acute phase (13, 11).onversely, BALB/c mice acutely infected with T. cruziL strain did not present parasitism (34) or inflamma-

ory infiltrates within the CNS, despite their intenseresence in peripheral nerves and autonomous gangliaH. Lenzi, personal communication). However, in C3H

ice infected with CL or Y strains, the mild inflamma-ory lesions found in the brain during late acute infec-ion persisted in the chronic phase (16). The reasonshy C3H mice react differently from BALB/c to CL

nfection remain unknown, but they can be related toeculiarities of C3H immune response to CL-specificntigens. We are presently studying the response ofALB/c mice to infection with the Colombian strain, inrder to define whether this transient encephalitis ishe rule among different mice lineages infected withhe Colombian strain.

FIG. 7. Serial sections of the brain tissue of acutely infected mice42 dpi) showing immunoperoxidase staining for cell surface markersf T lymphocytes infiltrating leptomeninges. The positive cells aredentified by a partial or complete rim of dark color outlining the cell

embrane (arrows). Serial sections show (a) hematoxylin and eosintaining; (b) CD81 lymphocytes; and (c) CD41 lymphocytes. Immu-ohistochemistry is described under Materials and Methods. Origi-

al magnification 3250.

mats

pcstcabs

itiTttbrtprifwswab

Ca

fCsa

65ENCEPHALITIS IN ACUTE CHAGAS’ DISEASE

In our study, neurological symptoms occurredainly in chronically infected animals and were not

ssociated with the presence of active encephalitis inhe CNS. Conversely, encephalitis accompanied by

FIG. 8. Mononuclear cells infiltrating cerebellum of acutely in-ected mice (42 dpi). Serial sections show (a) CD81 lymphocytes; (b)D41 lymphocytes; and (c) macrophages. Arrows indicate positivelytained cells. Immunohistochemistry is described under Materialsnd Methods. Original magnification 3100.

imilar neuropathies was observed during the acute

hase in dogs, but this study was not extended to thehronic phase (10). It is possible that accumulated tis-ue damage during acute infection encephalitis leadso sequelae, which could explain the progressive in-rease in the occurrence of clinical symptoms from thecute to the chronic phase of infection. Another possi-ility is that neuropathies result from demyelination,imilar to experimental allergic encephalitis (24).In this paper we demonstrated the presence of an

ntense and predominantly CD81 inflammatory infil-rate that permeates brain parenchyma during acutenfection and is frequently unrelated to the presence of. cruzi antigens. Inflammation does not persist during

he chronic stage, although mild inflammatory infil-rates, restricted to areas of incomplete blood–brainarrier, could still be detected at this stage. The factorsesponsible for the entrance and transient retention ofhese lymphocytes within the CNS during the acutehase, and later, resolution of inflammatory process,emain unknown. This possibility is currently undernvestigation. Also, it remains to be demonstrated byunctional experiments whether the CD81 cells foundithin the CNS during the infection are mounting a

pecific anti-T. cruzi or an autoimmune response, orhether they still represent nonspecific lymphocytesctivated in the periphery and brought within the CNSy a systemic inflammatory process.

ACKNOWLEDGMENTS

The work was supported by CNPq, CAPES, Instituto Oswaldoruz, and PAPES-Fiocruz. The authors are grateful to Heloisa Diniznd Genilton Vieira for preparing the figures.

REFERENCES

1. Koberle, F., Chagas’ disease and Chagas’ syndromes: The pathol-ogy of American trypanosomiasis. Adv. Parasitol. 6, 63–116,1968.

2. Petry, K., and Van Voorhis, W. C., Antigens of Trypanosomacruzi that mimic mammalian nervous tissues: Investigations oftheir role in the autoimmune pathophysiology of chronic Chagas’disease. Res. Immunol. 142, 151–156, 1991.

3. Kalil, J., and Cunha-Neto, E., Autoimmunity in Chagas diseasecardiomyopathy: Fulfilling the criteria at last? Parasitol. Today12, 396–399, 1996.

4. Chagas, C., Estudos sobre a morfologia e o ciclo evolutivo doSchizotrypanum cruzi, n. gen. n. sp., agente etiologico da novaentidade morbida do homem. Mem. Inst. Oswaldo Cruz, 1, 159–219, 1909.

5. Rocha, A., Meneses, A. C. O., Silva, A. M., Ferreira, M. S.,Nishioca, S. A., Burgarelli, M. K. N., Almeida, E., Turcato-Jr, G.,Metze, K., and Lopes, E. R., Pathology of patients with Chagas’disease and acquired immunodeficiency syndrome. Am. J. Trop.Med. Hyg. 50, 261–268, 1994.

6. Pittella, J. E. H., Central nervous system in Chagas’ disease. Anupdating. Rev. Inst. Med. Trop. (Sao Paulo) 35, 111–116, 1993.

7. Prata, A., Chagas’ disease. Infect. Dis. Clin. North Am. 8, 61–76,

1994.

1

1

1

1

1

1

1

1

1

1

2

2

2

2

2

2

2

2

2

2

3

3

3

3

3

R

66 SILVA ET AL.

8. Pentreath, V. W., Trypanosomiasis and nervous system. Pathol-ogy and immunology. Trans. R. Soc. Trop. Med. Hyg. 89, 9–15,1995.

9. Queiroz, A. C., Estudo das alteracoes encefalicas em casos hu-manos agudos de doenca de Chagas. Rev. Pat. Trop. 7, 13–22,1978.

0. Villela, E., and Torres, M., Estudo histo-pathologico do systemanervoso central em paralysia experimental determinada peloSchizotrypanum cruzi. Mem. Inst. Oswaldo Cruz 26, 77–79,1926.

1. Pittella, J. E. H., Central nervous system involvement in exper-imental trypanosomiasis cruzi. Mem. Inst. Oswaldo Cruz 86,141–145, 1991.

2. Torres, M., and Villaca, J., Encefalite e mielite causadas por umtripanosoma (T. cruzi). Mem. Inst. Oswaldo Cruz 11, 80–89,1919.

3. Pittella, J. E. H., Meneguette, C., Barbosa, A. J. A., and Bam-birra, E. A., Histopathological and immunohistochemical studyof the brain in the acute and chronic phases of experimentaltrypanosomiasis in dogs. Ann. Trop. Med. Parasitol. 84, 615–621, 1990.

4. Hontebeyrie-Joskowicz, M., Said, G., Milon, G., Maechal, G., andEisen, H., L3T4 T cells able to mediate parasite-specific delayed-type hypersensibility play a role in the pathology of experimen-tal Chagas’ disease. Eur. J. Immunol. 17, 1027–1033, 1987.

5. Spinella, S., Liegeard, P., and Hontebeyrie-Joskowicz, M.,Trypanosoma cruzi: Predominance of IgG2a in nonspecific hu-moral response during experimental Chagas’ disease. Exp. Para-sitol. 74, 46–56, 1992.

6. Younes-Chennoufi, A. B., Said, G., Eisen, H., Durand, A., andHontobeyrie-Joskowicz, M., Cellular immunity to Trypanosomacruzi is mediated by helper T cells (CD41). Trans. R. Soc. Trop.Med. Hyg. 82, 84–89, 1988.

7. Andrade, S. G., Caracterizacao de cepas do Trypanosoma cruziisoladas no Reconcavo baiano. Rev. Pat. Trop. 3, 65–121, 1974.

8. Brener, Z., Therapeutic activity and criterion of cure on miceexperimentally infected with Trypanosoma cruzi. Rev. Inst. Med.Trop. Sao Paulo 4, 389–394, 1962.

9. Lannes-Vieira, L., Dardenne, M., and Savino, W., Extracellularmatrix components of the mouse thymic microenvironment on-togenic studies and modulation by glucocorticoid hormones.J. Histochem. Cytochem. 39, 1539–1546, 1991.

0. Minoprio, P. M., Eisen, H., Forni, L., D’Imperio Lima, M. R.,Joskowicz, M., and Coutinho, A., Polyclonal lymphocyte responseto murine Trypanosoma cruzi infection. I. Quantification of bothT- and B-cell response. Scand. J. Immunol. 24, 661–668, 1986.

1. Sell, S., Immune resistance to infection. In “Immunology, Immu-nopathology and Immunity,” pp. 695–700, Appleton & Lange,

Stanford, CT, 1996.

2. Ribeiro-dos-Santos, R., Marquez, J. O., Furtado, C. C. G., Ol-iveira, J. C. R., Martins, A. R., and Koberle, F., Antibodiesagainst neurons in chronic Chagas’ disease. Tropenmed. Parasi-tol. 30, 19–23, 1979.

3. Rossi, M., Microvascular changes as a cause of chronic cardio-myopathy in Chagas’ disease. Am. Heart J. 120, 233–236, 1990.

4. Owens, T., Renno, T., Taupin, V., and Krakowski, M., Inflam-matory cytokines in the brain: Does the CNS shape immuneresponse? Immunol. Today 15, 565–570, 1994.

5. Gazzinelli, R. T., Denkers, E. Y., and Sher, A., Host resistance toToxoplasma gondii: Model for studying the selective induction ofcell-mediated immunity by intracellular parasites. Infect. Ag.Dis. 2, 139–149, 1993.

6. Janeway, C. A., and Travers, P., T cell mediated cytotoxicity. In“Immunobiology: The Immune System in Health and Disease,”pp. 7:32–7:34, Current Biology Garland, New York, 1997.

7. Doherty, P. C., and Zinkernagel, R. M., Capacity of sensitizedthymus-derived lymphocytes to induce fatal lymphocyte chorio-meningitis is restricted by the H2 gene complex. J. Immunol.114, 30–39, 1975.

8. Higushi, M. L., Reis, M. M., Aiello, V. D., Benvenuti, L. A.,Gutierrez, P. S., Bellotti, G., and Pileggi, F., Association of anincrease in CD81 T cells with the presence of Trypanosoma cruziantigens in chronic human chagasic myocarditis. Am. J. Trop.Med. Hyg. 56, 485–489, 1997.

9. Pirmez, C., and Ribeiro-dos-Santos, R., Autoreactivity in chronicexperimental Trypanosoma cruzi infection. Ciencia Cultura 46,418–423, 1994.

0. D’Avila Reis, D., Jones, E. M., Tostes, S. Jr., Lopes, E. R.,Gazzinelli, G., Colley, D. G., and Mc Curley, T. L., Characteriza-tion of inflammatory infiltrates in chronic chagasic myocardiallesions: presence of tumor necrosis factor-a positive cells anddominance of Granzyme A1, CD81 lymphocytes. Am. J. Trop.Med. Hyg. 48, 637–644, 1993.

1. Sato, M. N., Yamashiro-Kanashiro, E. H., Tanji, M. M., Kaneno,R., Higushi, M. L., and Duarte, J. S., CD81 cells and naturalcytotoxic activity among spleen, blood and heart lymphocytesduring the acute phase of Trypanosoma cruzi infection. Infect.Immun. 60, 1024–1030, 1992.

2. Sun, J., and Tarleton, R. L., Predominance of CD81 T lympho-cytes in the inflammatory lesions of mice with acute Trypano-soma cruzi infection. Am. J. Trop. Med. Hyg. 48, 161–169, 1993.

3. Tarleton, R. L., Son, J., Zhang, L., and Postan, M., Depletion ofT-cell subpopulations results in exacerbation of myocarditis andparasitism in experimental Chagas’ disease. Infect. Immun. 62,1820–1829, 1994.

4. Lenzi, H. L., Oliveira, D. N., Lima, M. T., and Gattass, C. R.,Trypanosoma cruzi: Paninfectivity of CL strain during murine

acute infection. Exp. Parasitol. 84, 16–27, 1996.

eceived January 15, 1998; accepted with revision March 3, 1999