Original article

The Syrian hamster as a model for the dilated cardiomyopathy of Chagas’disease: a quantitative echocardiographical and histopathological analysis

Angelina M.B. Bilatea,f, Vera M.C. Salemib, Felix J.A. Ramiresb, Thales de Britoc,Ana M. Silvac, Eufrosina S. Umezawad, Charles Madyb, Jorge Kalila,e,f,

Edecio Cunha-Netoa,e,f,*a Laboratory of Immunology, Heart Institute (Incor), University of São Paulo Medical School,

Av. Dr. Enéas de Carvalho Aguiar, 44-Bloco II-90 andar, 05403-000 Sao Paulo, SP, Brazilb Division of General Cardiopathies, Heart Institute (Incor), University of São Paulo Medical School, São Paulo, Brazil

c Laboratory of Immunopathology, Institute of Tropical Medicine, University of São Paulo Medical School, São Paulo, Brazild Laboratory of Protozoology, Institute of Tropical Medicine, University of São Paulo Medical School, São Paulo, Brazil

e Division of Clinical Immunology and Allergy, Department of Medicine, University of São Paulo Medical School, Sao Paulo, Brazilf Institute for Investigation in Immunology (iii)/Millenium Institute, Brazil

Received 1 April 2003; accepted 6 July 2003

Abstract

Chronic Chagas’ disease cardiomyopathy (CCC) is caused by the protozoanTrypanosoma cruzi, and it affects 30% of the 16–18 millionpeople infected in Latin America. A good rodent model that develops a dilated cardiomyopathy closely resembling human CCC afterT. cruziinfection is still needed. We compared the cardiomyopathy developed byT. cruzi-infected Syrian hamsters with human Chagas’ diseasecardiomyopathy using quantitative methods. Female hamsters were infected with 3.5 × 104 (G1,n = 10) or 105 (G2,n = 10)T. cruziY strainblood trypomastigotes. Control animals (C,n = 10) were injected with saline solution. Cardiac function was assessed by echocardiography at4, 8 and 12 months post-infection. Heart sections were submitted to histopathological/morphometric analysis 12 months post-infection.At thistime, ventricular dysfunction and diffuse or multi-focal myocarditis were observed in 91% and 100% of G1 and G2 infected groups,respectively. Median interstitial collagen volumes in groups C, G1 and G2 were 1.2%, 1.9% and 3.9%, respectively, and were significantlyhigher in group G2 than in group C. Among infected animals, myocarditis showed a positive correlation with interstitial fibrosis. Deaths in thechronic phase (8–12 months post-infection) were more frequent among G2 than G1, and were associated with macroscopic ventriculardilation, severe myocarditis and increased fibrosis values, along with an earlier onset of ventricular dysfunction. TheT. cruzichronicallyinfected Syrian hamster develops a cardiomyopathy which resembles human Chagas’ disease cardiomyopathy, and might be an adequate toolto investigate pathogenic mechanisms of this disease and to search for novel therapeutic strategies.

© 2003 Éditions scientifiques et médicales Elsevier SAS. All rights reserved.

Keywords: Trypanosoma cruzi; Hamster; Chagas’ disease cardiomyopathy; Myocarditis; Fibrosis; Echocardiography

1. Introduction

Chronic Chagas’ disease cardiomyopathy (CCC), causedby the protozoanTrypanosoma cruzi, is a significant cause of

morbidity and mortality in Central and South America,where 8–9 million people are infected and 25 million peopleremain at risk[1]. Due to intense immigration from endemicareas, transfusion-related infection has recently been per-ceived as a potential threat in the United States[2]. Themajority of infected individuals remain asymptomatic andfree from heart disease, in the so-called “indeterminate” formof the disease (65–70%), while up to one-third of infectedindividuals eventually develop a life-threatening inflamma-tory dilated cardiomyopathy 15–30 years after initial infec-tion [3]. CCC is thought to be the single most common causeof congestive heart failure and sudden death among young-

Abbreviations:CCC, chronic Chagas’ disease cardiomyopathy; DD,diastolic diameter; FS%, fractional shortening; H&E, hematoxylin and eo-sin; ICVF, interstitial collagen volume fraction; LV, left ventricular; LVDD,left ventricular diastolic diameter; LVSD, left ventricular systolic diameter;SD, systolic diameter.

* Corresponding author. Tel.: +55-11-3069-5914;fax: +55-11-3082-9350.

E-mail address:edecunha@usp.br (E. Cunha-Neto).

Microbes and Infection 5 (2003) 1116–1124

www.elsevier.com/locate/micinf

© 2003 Éditions scientifiques et médicales Elsevier SAS. All rights reserved.doi:10.1016/j.micinf.2003.07.001

to-middle-aged adults in endemic areas in Brazil[4]. CCCpatients with severe congestive heart failure have a poorerprognosis and 50% shorter survival period than a similar,non-chagasic, hospital cohort in Ribeirão Preto, Brazil[5].

The hallmark of CCC heart lesions is a multi-focal ordiffuse myocarditis in the virtual absence of parasites andextensive fibrosis[6,7], consistent with a myocardial remod-eling process[8]. Previous reports showed that myocardialfibrosis in CCC patients is higher than that observed inpatients with idiopathic dilated cardiomyopathy[9]. Thecontribution of myocardial fibrosis to CCC severity wasunderscored by the significant correlation between the de-gree of interstitial collagen deposition and left ventricular(LV) systolic dysfunction[8]. The major mechanisms re-sponsible for the pathogenesis of CCC and factors whichgovern differential susceptibility to the development of di-lated cardiomyopathy uponT. cruzi infection are not yetcompletely understood. Several pathogenic mechanismshave been postulated, such as immune responses to parasiteantigens[10], autoimmunity to heart self-antigens[7,11–13],autonomic denervation[3] and microvascular spasm[14],but their relative roles remain elusive, although immune-mediated inflammation certainly plays a major role[12,15,16].

Chronically infected mice and rats may develop an in-flammatory infiltrate and fibrosis in the heart[15,17–19,20],but the development of a cardiomyopathy closely resemblingthe human CCC, with extensive fibrosis, segmental myocar-dial abnormalities and macroscopic ventricle dilation, after alatent period with no significant abnormalities (the ‘indeter-minate’phase), has not been reported yet. Some authors haveshown that rabbits and monkeys could develop dilated CCCuponT. cruzi infection[21,22], although the cardiac lesionswere not subject to quantitative methods and were not alwaysreproducible. It has been reported that cyclophosphamide-treated mongrel dogs consistently develop chronic heart dis-ease with diffuse myocarditis, fibrosis and cardiomegaly[23]. Furthermore, larger animals require long and expensiveexperiments.

The Syrian golden hamster (Mesocricetus auratus) hasproven to be an adequate experimental model for visceral andtegumentary leishmaniasis[24,25], as well as for congestiveheart failure induced by coxsackievirus-induced chronicmyocarditis or in spontaneous cardiomyopathic mutantstrains [26–28]. These mutants carry deletions in thed-sarcoglycan gene[29,30], which cause impairment of sar-colemmal integrity and subsequent cardiomyopathy[31].Taken together, these reports indicate that the Syrian hamsteris a suitable model for studying inflammatory or non-inflammatory cardiomyopathy and heart failure. A previousanatomopathological study showed that after 3–10 months ofT. cruzi infection, hamsters developed a chronic cardiomy-opathy similar to that observed in CCC patients, includingmyocarditis and fibrosis, leading to dilation in most of theanimals studied[32]. However, due to the lack of quantitativemethodology and accurate timing for the assessment of his-

topathological lesions, this work failed to provide data ondisease evolution. The use of non-invasive transthoracicechocardiography in small rodents allows serial examina-tions of LV function in a single animal[33,34]. The directassessment of structural/functional changes during diseaseprogression may contribute to a better understanding ofpathophysiology and evaluation of new therapeutic agents[35–37].

In order to evaluate whether theT. cruzi-infected Syrianhamster is a suitable model for further studies of pathogen-esis and specific therapy for CCC, we analyzed the develop-ment of cardiomyopathy upon chronicT. cruzi infection,using serial echocardiography, morphometry to measuremyocardial collagen content (fibrosis) and scoring of myo-carditis. By infecting animals with various numbers of para-sites, we were able to assess the influence of parasite load ondisease progression.

2. Methods

2.1. Hamsters

Ten- to 12-week-old female outbred Syrian golden ham-sters (M. auratus) were obtained from University of SãoPaulo Medical School Central Facility and maintained withfood and water ad libitum and handled according to localregulations and research protocols approved by our internalreview base.

2.2. Parasites and infection

Twenty hamsters were infected via the intraperitonealroute with 3.5 × 104 (G1,n = 10) or 105 (G2,n = 10)T. cruziY strain blood trypomastigotes.After 12 months of infection,all surviving animals were euthanized by intracardiac 3 MKCl injection after intraperitoneal anesthesia with 50 mg/kgof pentobarbital. Ten non-infected hamsters were injectedwith saline solution and utilized as controls (C) and eutha-nized in the same time limit and way as infected animals.After 12 months of infection, the serum of each animal wascollected for further analysis by ELISA for the presence ofIgG antibodies againstT. cruzi whole antigenic extract, toconfirm the infection[38]. Serum samples from non-infectedanimals (group C) were used as controls for ELISA.

2.3. Survival studies

Survival was monitored by daily observation. Animalsthat spontaneously died before 12 months of infection andwere recovered prior to rigor mortis were necropsied forhistopathological analysis. Deaths from the first day of infec-tion until 4 months, from 4 until 8 months, and from 8 until12 months post-infection were considered to have takenplace in the acute, intermediate and chronic stage, respec-tively.

1117A.M.B. Bilate et al. / Microbes and Infection 5 (2003) 1116–1124

2.4. Echocardiography

To evaluate whether chronicT. cruzi infection wouldaffect LV function, we performed echocardiography after 4,8 and 12 months of infection. For this study, anesthesia wasinduced by intraperitoneal injection of pentobarbital(25 mg/kg). The chest of each animal was carefully shaved,they were positioned in the supine position with front pawswide open, and an ultrasound transmission gel was applied tothe precordium. Transthoracic echocardiography was per-formed using an Acuson Sequoia model 512 echocardio-grapher equipped with a 13-MHz linear transducer. The heartwasfirst imaged in the two-dimensional (2D) mode in theparasternal long-axis view, followed by short-axis view. Leftventricular systolic diameter (LVSD) and left ventriculardiastolic diameter (LVDD) were measured at the level of thepapillary muscles, using 2D guided M-mode imaging. Eachmeasurement was obtained with the average of three con-secutive heart beats. All measurements were done accordingto the American Society of Echocardiography guidelines[39] by the same experienced observer. LV systolic functionwas assessed by fractional shortening (FS%) of left ventricle,which was calculated from the M-mode using the followingequation:

FS (%) = [(LVDD – LVSD)/LVDD] × 100.

2.5. Histopathological examination

Five-micrometer sections of ventricles from all survivinganimals were transversely cut at midportion of the heart forleft and right ventricle simultaneous analysis, fixed in buff-ered formalin solution pH 7.2 and embedded in paraffin, andsubsequently, stained with hematoxylin and eosin (H&E)and picrosirius red.

2.5.1. Myocarditis scoreHeart sections were stained with H&E and examined

blindly by an experienced pathologist. Inflammatory infil-trate was analyzed and semiquantitatively graded as focal ormild myocarditis (+), moderate (++) and severe (multi-focalor diffuse) myocarditis (+++)[10].

2.5.2. Fibrosis quantification by morphometric analysisAreas of collagen were stained by fibrillar collagen-

specific picrosirius red, and interstitial and perivascular col-lagen volume fractions were determined by videomorphom-etry, using an image processing and analysis system(Quantimet 570, LeicaQwin, Cambridge Instruments, Cam-bridge, UK). Briefly, for interstitial fibrosis, a numericalvalue of the total and the picrosirius red-stained volume wereobtained by adding the values of each area from 15 to 20 dif-ferent 50× microscopic fields in a section; the fraction of totalvolume of collagen deposition per area of the entire sectionanalyzed was calculated for each animal. For perivascularfibrosis, a numerical value corresponding to collagen depo-sition and to the lumen was obtained for each vessel with

100× magnification, and the volume fraction of perivascularcollagen of each section was calculated. For this analysis, weconsidered only vessels with regular lumen.

2.5.3. In situ detection ofT. cruzi antigens byimmunohistochemistry

Sections from paraffin blocks were previously deparaf-finized and incubated overnight at 4 °C with polyclonalanti-T. cruzi serum raised in rabbits after three immuniza-tions withT. cruzifrozen/thawed whole homogenate. Bioti-nylated goat anti-mouse/rabbit IgG and Duet Strepto AB-Complex (Dako, Denmark) were used at the amplificationstep of the reaction. Diamino-benzidine (DAB, SigmaChemical Co., USA) was used as chromogen. The sectionswere lightly counterstained with H&E. Heart sections ofT. cruziY strain acutely infected mice with myocarditis wereused as positive controls. Negative controls included omis-sion of the primary anti-T. cruziantibody and its replacementby normal rabbit serum.

2.6. Statistics

Descriptive statistics are represented by median andlower/upper 95% confidence intervals or maximum andminimum values. Statistical comparisons were performed bynon-parametric analysis of variance (ANOVA-Kruskal Wal-lis) followed by multiple comparison test (Dunn’s test). Cor-relations between continuous variables were analyzed by thenon-parametric Spearman test.P < 0.05 was consideredstatistically significant. For analysis of individual datapoints, the median ± 95% confidence interval from the con-trol group was considered the normal range, and data pointsfalling outside this range were considered abnormal.

3. Results

3.1. Survival duringT. cruzi infection

T. cruzi infection resulted in deaths at the acute (1st–4thmonth) and chronic (8th–12th month) phases of infection inboth infected groups (Fig. 1). During the first 4 months of

Fig. 1. Survival rates amongT. cruzi chronically infected hamsters.C, control group; G1, infected with 3.5 × 104 parasites and G2, infected with105 parasites.n = 9 (C);n = 10 (G1);n = 10 (G2).

1118 A.M.B. Bilate et al. / Microbes and Infection 5 (2003) 1116–1124

infection, three out of 10 animals died in each group, butthose infected with 105 parasites (G2) died before the secondmonth, while those infected with 3.5 × 104 parasites (G1)died between the third and fourth month after infection.Deaths at the chronic phase were significantly more frequentamong animals from group G2 (four out of seven acute-phasesurvivors) than G1 (one out of seven acute-phase survivors)(P < 0.05) (Fig. 1). All infected animals displayed high levelsof anti-T. cruziIgG antibodies, as measured by ELISA (datanot shown).

3.2. Echocardiographic changes: chronicT. cruzi infectionresults in significant LV dysfunction

Transthoracic echocardiography was performed on allsurviving animals at 4, 8 and 12 months after infection. Thetotal number of hamsters in each group available for studyafter 4 and 8 months of infection was nine in the controlgroup and seven each in group G1 and G2; after 12 months ofinfection, there were nine in the control group, seven in G1and four in G2. Global hypocontractility could be observedin a progressively higher number of animals after 4,8 and 12 months of infection (data not shown); a representa-tive M-mode echocardiogram can be seen inFig. 2b. Onthe other hand, we did not observe any segmentalhypocontractility—including apical aneurysms—in any in-fected animal at any time. The values of LVSD after

12 months of infection were significantly higher in bothinfected groups than in group C(Table 1).Even though someanimals displayed an increased LVDD value after 12 monthsof infection, the LVDDs from the infected groups were notsignificantly different from those of controls at any time(Table 1).Fig. 3shows a progressive increase in the numberof animals showing decreased FS% values after 4 and8 months of infection (1/7 and 3/7 in G1; 3/7 and 4/7 in G2).Furthermore, the FS% of G1 or G2 were significantly lowerthan those of group C at the 12th month post-infection, where

Fig. 2. Representative transthoracic LV M-mode echocardiograms and macroscopic transversal sections of heart. (A) Non-infected control hamster (C.4),showing normal diameters and (B) hamster after 4 months ofT. cruziinfection (G2.9), showing LV dilation. LVSD, left ventricle systolic diameter; LVDD, leftventricle diastolic diameter. (C) Control hamster C.6, no ventricular dilation; (D) hamster after 12 months of T.cruzi infection G2.10, right and left ventricledilation and decreasing thickness of the LV free wall. H&E staining. LV, left ventricle; RV, right ventricle.

Table 1Echocardiographical analysis of diastolic and systolic diameters from leftventricle ofT. cruzichronically infected hamsters

Months post infection Groups LVDDa LVSDa

(mm) (mm)4 C 4.8 (4.4 – 5.7) 2.8 (2.2 – 3.6)

G1 5.1 (4.2 – 5.7) 3.1 (2.4 – 3.5)G2 5.0 (3.9 – 6.2) 2.9 (2.3 – 5.2)

8 C 5.6 (4.6 –5.9) 3.4 (2.6 – 3.6)G1 5.5 (4.5 – 5.8) 2.9 (2.7 – 3.6)G2 5.9 (5.3 – 6.9) 4.0 (3.0 – 5.4)

12 C 5.6 (4.9 – 8.1) 3.6 (2.7 – 5.8)G1 6.4 (5.4 – 7.1) 4.9 (3.9 – 6.0)*G2 6.4 (5.6 – 6.7) 4.9 (4.2 – 5.0)*

a median (maximum – minimum value); * p < 0.05 vs group C; LVDD;left ventricle end-diastolic diameter; LVSD, left ventricle end-systolic dia-meter. C, control group; G1, infected with 3.5 × 104 T. cruziparasites; G2,infected with 105 T. cruziparasites

1119A.M.B. Bilate et al. / Microbes and Infection 5 (2003) 1116–1124

6/7 animals in G1 and 4/4 in G2 showed decreased FS%(Fig. 3). Despite the fact that a subgroup of G2 animalsdeveloped an earlier and more severe impairment of LVfunction (animals 4, 9 and 10,Fig. 6), we observed that someof the G2 hamsters (2, 8, 6 and 5) showed FS% values similarto those of the animals from group G1 at the three time points(Fig. 6).

3.3. Histopathological findings: chronicT. cruzi infectionresults in myocardium damage

Histopathological analysis was performed on the hearts ofall euthanized animals and also on those who died spontane-

ously in the chronic phase and whose hearts could be recov-ered prior to rigor mortis: seven control non-infected animals(C group), seven animals from G1 and five from G2. Allinfected animals displayed myocarditis, which was moderateor severe in nine out of 12 animals (Fig. 6). Mild myocarditiswas only found in two out of seven control animals (data notshown). Immunohistochemistry analysis with anti-T. cruziantisera revealed thatT. cruziantigens were only found in asingle high-powered field of one section from one infectedanimal out of the 36 heart sections analyzed from the 12 in-fected animals analyzed (data not shown).

The median interstitial collagen volume fraction (ICVF)of group G2 (3.9%) was threefold higher than for controls(1.2%), a statistically significant difference, and twofoldhigher than that of G1 (1.9%). Group G1 showed a 1.5-foldhigher ICVF than the control group. Perivascular collagenvolume remained similar in all three groups (Fig. 5). Micro-scopic scars were found scattered throughout both left andright ventricles from infected animals, and were more promi-nent among those from group G2 who had died spontane-ously in chronic phase (G2.4 and G2.10). Macroscopic ven-tricle dilation and decreased wall thickness could beobserved at necropsy in both left and right ventricles fromtwo out of five animals from group G2 and one out of sevenanimals from group G1 (Fig. 2). We found a positive corre-lation between the ICVF and the intensity of myocarditisafter 12 months of infection (r = 0.73,P = 0.05). In spite ofthe absence of statistical correlation between LV dysfunctionand myocarditis or fibrosis, animals with diminished LVfunction after the 12th month post-infection exhibited ahigher myocarditis score and elevated interstitial fibrosis; asubgroup of those (animals G1.10, G2.4 and G2.10) showedmacroscopic ventricular dilation (Fig. 2D) and died before12 months of infection. On the other hand, animals withpreserved LV function displayed less intense myocarditis andICVF similar to that of the non-infected controls, remainingalive until the end of the 12 months ofT. cruzi infection(Fig. 6). Significantly, most animals showing increased fibro-sis at the 12th month post-infection or premature death haddisplayed LV dysfunction as early as 4 or 8 months afterinfection (Fig. 6).

4. Discussion

The present study employed quantitative measurements todemonstrate that the Syrian hamster develops aT. cruzi-induced cardiomyopathy resembling human CCC: multi-focal and/or diffuse myocarditis in the virtual absence ofT. cruzior its antigens, increased interstitial fibrosis accom-panied by microscopic scars, mild to severe LV dysfunction.Furthermore, macroscopic ventricular dilation and decreasedventricle wall thickness were associated with death in thechronic phase.

The fact that chronicallyT. cruzi-infected hamsters devel-oped LV dysfunction with diffuse hypokinesis, as judged bybi-dimensional echocardiography (Figs. 2 and 3), is in line

Fig. 3. LV systolic function during chronicT. cruzi infection, measured byFS% derived M-mode echocardiography. Each symbol stands for one ani-mal. The dotted lines are the upper and lower 95% confidence intervals fornon-infected controls, horizontal bars are the median for each group.*P < 0.05 vs. C. One animal from group G2 (G2.4) died 1 day after the lastechocardiography and before the date of euthanasia.

1120 A.M.B. Bilate et al. / Microbes and Infection 5 (2003) 1116–1124

with findings in patients with chronic Chagas’ cardiomyopa-thy [40]. It was previously shown that symptomatic CCCpatients can display diffuse hypokinesis and ventricular dila-tion indistinguishable from those observed in cardiomyopa-thies of other etiologies[41]. On the other hand, apicalaneurysm and segmental wall motion abnormality, which arefrequently found among CCC patients[42–44], were notobserved by echocardiography in any infected animal at anytime. The absence of apical aneurysm may be explained bythe time course of the study and differences in hamster andhuman heart morphogenesis, including the disposition ofmuscle bundles and vessels, but the involvement of distinctpathogenic mechanisms—specifically in the response ofheart tissue to inflammation—cannot be excluded. The serialassessment of LV function allowed us to observe the differ-ential progression of LV dysfunction in chronically infectedanimals (Fig. 3). The findings that most animals displayingearly LV dysfunction after 4 months of infection died in thechronic phase suggest an association between early dysfunc-tion and a worse outcome. Furthermore, Mady and Nacruth[4] also reported that earlier clinical manifestations of humanCCC predict a worse outcome in human CCC, and severe LVdysfunction was significantly associated with shorter sur-vival among CCC patients with heart failure[45]. LV dys-function occurred even in the absence of significant LVdilation, since LVDD remained similar between infected andnon-infected animals (Table 1). The impairment of LV func-tion was dependent on the significant increase in LVSD after12 months of infection. The fact that eight infected animals

developed LV dysfunction without dilation, while the otherthree developed an aggressive disease with severe LV dys-function accompanied by macroscopic ventricular dilation,decreased ventricle wall thickness and death (Fig. 6) sug-gested that, in the hamster model, the extent of heart damagecan occur in different degrees. CCC patients can show im-paired LV function that ranges from mild to severe, and asmall group with severe LV dysfunction can develop ven-tricular dilation[40,43].

In our study, chronically infected animals displayed dif-fuse and/or multi-focal myocarditis (Figs. 4 and 6), resem-bling human Chagas’ disease[7]. These animals also dis-played significant myocardial interstitial fibrosis (Fig. 5) andmicroscopic scars, in line with recent findings of increasedcollagen deposition in the myocardium of CCC patients,which is higher than that observed in hearts of patients withidiopathic dilated cardiomyopathy[9]. Macroscopic ven-tricle dilation and decreased wall thickness were also ob-served in a significant proportion of chronically infectedanimals (Fig. 2), similar to the anatomopathological picturein human CCC[3]. Furthermore, the statistically significantcorrelation between intensity of myocarditis and interstitialfibrosis observed in our study suggests that the heart inflam-matory infiltrate could be responsible for the progressivedamage of the cardiac tissue, triggering the remodeling pro-cess that culminates in extensive fibrosis. The associationbetween myocarditis and severity of human CCC was shownby a higher incidence and severity of myocarditis in patientswith congestive heart failure than in those without heart

Fig. 4. Histological sections of ventricles transversely cut at mid-portion of heart. H&E staining. (A) Control hamster C.9; (B) infected hamster G1.2 with focalmyocarditis (200× original magnification); (C) infected hamster G2.8 with diffuse myocarditis (100× original magnification); (D) heart section stained withpicrosirius red from infected hamster G2.10. The infiltrated areas are indicated by arrows, and collagen is stained in red.

1121A.M.B. Bilate et al. / Microbes and Infection 5 (2003) 1116–1124

failure[16]. Our finding of increased fibrosis among the threeanimals with the most severe LV dysfunction suggests thatfibrosis contributed to LV dysfunction inT. cruzi-infectedhamsters. Mady et al.[8] have shown a statistical correlationbetween the magnitude of LV dysfunction and the degree ofinterstitial collagen deposition in the myocardium of CCCpatients. Taken together, these results reinforce our findingsthat the LV dysfunction in chronically infected hamsters maybe a consequence of the remodeling process triggered by theinflammatory process.

The finding that the number of deaths in the acute phasewas similar in both infected groups but occurred earlier ingroup G2 suggests that a higher inoculum is more pathogenicin the acute phase. Furthermore, the fact that animals infectedwith the highest inoculum of parasites developed an earlierand more severe LV dysfunction (Fig. 3), a higher frequencyof LV dilation, diffuse myocarditis (Fig. 4), increased inter-stitial fibrosis (Fig. 5) and had significantly more deaths inthe chronic phase (Fig. 1) than the animals infected with thelowest inoculum suggests a relationship between parasiteload and severity of chronic heart disease. Previous studieshave shown that the parasite load during the acute phase ofexperimentalT. cruzi infection in mice affects the para-sitemia, intensity of heart and striated muscle inflammatoryinfiltrate and activation of the immune system after 1 year ofinfection[19]. However, the fact that some hamsters infectedwith 105 parasites developed heart disease as mild as thattypically observed in animals infected with 3.5 × 104 para-

Fig. 5. Quantification of fibrosis in chronicallyT. cruzi-infected hamsters.Sections of the picrosirius staining were analyzed by PC-based Quantimet570 image analyser. Dotted lines in the graph stand for the lower and upper95% confidence intervals for non-infected controls, and horizontal bars arethe median. *P < 0.01 C ×G2.

Fig. 6. Association between progression of LV dysfunction, histopathological lesions and death during chronicT. cruziinfection. (A) Animals from group G1;(B) animals from group G2. The table at right shows the identification of each animal by number and symbol, interstitial fibrosis, myocarditis score and deathduring infection.

1122 A.M.B. Bilate et al. / Microbes and Infection 5 (2003) 1116–1124

sites (Fig. 6) suggests that, despite the high parasite load,individual genetic differences influence the development ofCCC, as observed in studies with Chagas’ disease cardiomy-opathy patients in our laboratory (Drigo et al., unpublishedobservations).

In this paper, the use of quantitative methods allowed theprecise evaluation of the extent of heart damage in eachindividual animal. Although not all features of CCC wereobserved (e.g. apical/segmental wall hypocontractility), andelectric conduction abnormalities were not sought for, thechronically T. cruzi-infected hamster stood as a suitablemodel for the development of severe disease with ventriculardysfunction and heart failure. The quantitative measurementsof LV function and myocardial histopathology may contrib-ute to the understanding of pathogenesis and the evaluationof novel therapeutic agents, an urgent need, given the poorprognosis and uncertain efficacy of conventional therapiesfor CCC patients. Furthermore, this animal model may en-able testing of the effects of intervention on CCC progressionin 12 months’ time, instead of the decades of follow-uprequired in trials with Chagas’ disease patients, due to itsslow progression.

Acknowledgements

This work was supported by FAPESP, CNPq (300857/94-2-NV) grants from Brazil and HHMI from the USA. Wethank Dr Veronica Coelho, Laboratory of Immunology, andDr João Sbano, Division of Image Diagnostics, Heart Insti-tute (Incor), University of São Paulo Medical School, forcritical review of this manuscript.

References

[1] World Health Org. Tech. Rep., Control of Chagas’ disease (2002)905:i.

[2] I.A. Shulman, M.D. Appleman, S. Saxena, A. Hiti, L.V. Kirchhoff,Specific antibodies toTrypanosoma cruziamong blood donors in LosAngeles, California, Transfusion 37 (1997) 727–731.

[3] F. Köberle, Chagas disease and Chagas’ syndromes: the pathology ofAmerican trypanosomiasis, Adv. Parasitol. 6 (1968) 63–116.

[4] C. Mady, R. Nacruth, Natural history of chronic Chagas’ heartdisease: prognosis factor, Rev. Paul. Med. 113 (1995) 791–796.

[5] R.B. Bestetti, G. Muccillo, Clinical course of Chagas’ heart disease: acomparison with dilated cardiomyopathy, Int. J. Cardiol. 60 (1997)187–193.

[6] A.C. Pereira-Barretto, C. Mady, E. Arteaga-Fernandez, N. Stolf,E.A. Lopes, M.L. Higuchi, G. Bellotti, F. Pileggi, Right ventricularendomyocardial biopsy in chronic Chagas’ disease, Am. Heart J. 111(1986) 307–312.

[7] M.A. Rossi, Pathogenesis of chronic Chagas myocarditis, Rev. Paul.Med. 113 (1995) 750–756.

[8] C. Mady, B.M. Ianni, E. Arteaga, G.S. Montes, E.G. Caldini,G. Andrade, M.C.P. Giorgi, P.H.N. Saldiva, Relation between intersti-tial myocardial collagen and the degree of clinical impairment inChagas’ disease, Am. J. Cardiol. 84 (1999) 354–356.

[9] M.L. Higuchi, S. Fukasawa, T. De Brito, L.C. Parzianello, G. Bellotti,J.A. Ramires, Different microcirculatory and interstitial matrix pat-terns in idiopathic dilated cardiomyopathy and Chagas’ disease: athree dimensional confocal microscopy study, Heart 82 (1999) 279–285.

[10] M.L. Higuchi, T. De Brito, M.M. Reis, A. Barbosa, G. Bellotti,A.C. Pereira-Barreto, F. Pileggi, Correlation betweenTrypanosomacruzi parasitism and myocardial inflammatory infiltrate in humanchronic chagasic myocarditis: light microscopy and immunohis-tochemical findings, Cardiovasc. Pathol. 2 (1993) 101–106.

[11] J. Kalil, E. Cunha-Neto, Autoimmunity in Chagas diseasecardiomyopathy: fulfilling the criteria at last? Parasitol. Today 12(1996) 396–399.

[12] E. Cunha-Neto, J. Kalil, in: M.W. Cunningham, R.S. Fujinami (Eds.),Molecular Mimicry, Microbes and Autoimmunity, ASM Press, USA,2000, pp. 257–274.

[13] J.S. Leon, D.M. Engman, Autoimmunity in Chagas heart disease,Inter. J. Parasitol. 31 (2001) 555–561.

[14] J.O. Mengel, M.A. Rossi, Chronic chagasic myocarditispathogenesis: dependence on autoimmune and microvascular factors,Am. Heart J. 124 (1992) 1052–1057.

[15] L. Pontes-De-Carvalho, C.C. Santana, M.B. Soares, G.G. Oliveira,E. Cunha-Neto, R. Ribeiro-Dos-Santos, Experimental chronic Cha-gas’ disease myocarditis is an autoimmune disease preventable byinduction of immunological tolerance to myocardial antigens, J.Autoimmun. 18 (2002) 131–138.

[16] M.L. Higuchi, F. De Morais, A.C. Pereira-Barreto, E.A. Lopes,N. Stolf, G. Bellotti, F. Pileggi, The role of active myocarditis in thedevelopment of heart failure in chronic Chagas disease: a study basedon endomyocardial biopsies, Clin. Cardiol. 10 (1987) 665–670.

[17] E. Chapadeiro, P.S. Beraldo, P.C. Jesus, W.P. Oliveira, L.F. Junqueira,Cardiac lesions in Wistar rats inoculated with various strains ofTry-panosma cruzi, Rev. Soc. Bras. Med. Trop. 21 (1988) 95–103.

[18] S.G. Andrade, A. Carneiro Filho, A.J. de Souza, E.S. de Lima,Z.A. Andrade, Influence of treatment with immunosuppressive drugsin mice chronically infected withTrypanosoma cruzi, Int. J. Exp.Pathol. 78 (1997) 391–399.

[19] C.R. Marinho, M.R. D’Imperio Lima, M.G. Grisotto, J.M. Alvarez,Influence of acute-phase parasite load on pathology, parasitism andactivation of the immune system at the late chronic phase of Chagas’disease, Infect. Immun. 67 (1999) 308–318.

[20] M. Chandra, J. Shirani, V. Shtutin, L.M. Weiss, S.M. Factor, S.B. Pet-kova, M. Rojkind, J.A. Dominguez-Rosales, L.A. Jelicks, S.A. Mor-ris, M. Wittner, H.B. Tanowitz, Cardioprotective effects of verapmilon myocardial structure and function in a murine model of chronicTrypanosoma cruziinfection (Brazil strain): an echocardiogrpahicstudy, Intern. J. Parasitol. 32 (2002) 207–215.

[21] A.R. Teixeira, F. Figueiredo, J. Rezende Filho, V. Macedo, Chagas’disease: a clinical, parasitological, immunological, and pathologicalstudy in rabbits, Am. J. Trop. Med. Hyg. 32 (1983) 258–272.

[22] C.A. Falasca, M. Gili, D. Grana, E. Gómez, J. Zoppi, E. Maresco,Chronic myocardial damage in experimentalT. cruzi infection of anew world primate, Cebus sp. Monkey, Rev. Inst. Med. Trop. SãoPaulo 32 (1990) 151–161.

[23] Z.A. Andrade, S.G. Andrade, M. Sadigursky, Enhancement of chronicTrypanosoma cruzimyocarditis in dogs treated with low doses ofcyclophosphamide, Am. J. Pathol. 127 (1987) 467–473.

[24] P.C. Melby, V.V. Tryon, B. Chandrasekar, G.L. Freeman, Cloning ofSyrian hamster (Mesocricetus auratus) cytokine cDNAs and analysisof cytokine mRNA expression in experimental visceral leishmaniasis,Infect. Immun. 66 (1998) 2135–2142.

[25] P.C. Melby, B. Chandrasekar, W. Zhao, J.E. Coe, The hamster as amodel of human visceral leishmaniasis: progressive disease andimpaired generation of nitric oxide in the face of a prominent Th-1 likecytokine response, J. Immunol. 166 (2001) 1912–1920.

1123A.M.B. Bilate et al. / Microbes and Infection 5 (2003) 1116–1124

[26] H. Morita, Experimental coxsackie B-3 virus myocarditis in goldenhamsters. Light and electron microscopy findings in a long-termfollow-up study, Jpn. Circ. J. 45 (1981) 713–729.

[27] H. Morita,Y. Kitaura, H. Deguchi, M. Kotaka, K. Kawamura, Experi-mental coxsackie B3 virus myocarditis in golden hamsters. II. Evalu-ation of left ventricular function in intact in situ heart 14 months afterinoculation, Jpn. Circ. J. 48 (1984) 1097–1106.

[28] S. Goineau, S. Nisse-Durgeat, D. Pape, P. Guillo, M.P. Ramee, E. Bel-lissant, Systemic and regional hemodynamic and cardiac remodelingeffects of candesartan in dilated cardiomyopathic hamsters withadvanced heart failure, J. Cardiovasc. Pharmacol. 40 (2002) 189–200.

[29] G. Hasenfuss, Animal models of human cardiovascular disease, heartfailure and hypertrophy, Cardiovasc. Res. 39 (1998) 60–76.

[30] Y. Ikeda, J. Ross Jr, Models of dilated cardiomyopathy in the mouseand the hamster, Curr. Opin. Cardiol. 15 (2000) 197–201.

[31] V. Straub, F. Duclos, D.P. Venzke, J.C. Lee, S. Cutshall, C.J. Leveille,K.P. Campbell, Molecular pathogenesis of muscle degeneration in thedelta-sarcoglycan-deficient hamster, Am. J. Pathol. 153 (1998) 1623–1630.

[32] L.E. Ramirez, E. Lages-Silva, J.M. Soares-Júnior, E. Chapadeiro, Thehamster (mesocricetus auratus) as experimental model in Chagas’disease: parasitological and histopathological studies in acute andchronic phases ofTrypanosoma cruziinfection, Rev. Soc. Bras. Med.Trop. 27 (1994) 163–169.

[33] N. Tanaka, N. Dalton, L. Mao, H.A. Rockman, K.L. Peterson,K.R. Gottshall, J.J. Hunter, K.R. Chien, J. Ross Jr, Transthoracicechocardiography in models of cardiac disease in the mouse, Circula-tion 94 (1996) 1109–1117.

[34] P.J. Schwartz, Do animal models have clinical value? Am. J. Cardiol.81 (1998) 14D–20D.

[35] I. Lemire,A. Ducharme, J.C. Tardif, F. Poulin, L.R. Jones, B.G.Allen,T.E. Hebert, H. Rindt, Cardiac-directed overexpression of wild-typealpha1B-adrenergic receptor induces dilated cardiomyopathy, Am. J.Physiol. Heart Circ. Physiol. 281 (2001) H931–H938.

[36] M.A. Cavasin, X.P. Yang, Y.H. Liu, D. Mehta, R. Karumanchi,M. Bulagannawar, O.A. Carretero, Effects of ACE inhibitor, AT-1antagonist, and combined treatment in mice with heart failure, J.Cardiovasc. Pharmacol. 36 (2000) 472–480.

[37] Y.Y. Li, Y.Q. Feng, T. Kadokami, C.F. McTiernan, R. Draviam,S.C. Watkins, A.M. Feldman, Myocardial extracellular matrix remod-eling in transgenic mice overexpressing tumor necrosis factor alphacan be modulated by anti-tumor necrosis factor alpha therapy, Proc.Natl. Acad. Sci. USA 97 (2000) 12746–12751.

[38] A.M. Lissaldo, S. Hoshino-Shimizu, E.S. Umezawa, A.M. Stolf,Alkaline solubleTrypanosoma cruziepimastigote antigen (ASEA)applied to Dot-ELISA, Rev. Inst. Med. Trop. Sao Paulo 36 (1994)163–166.

[39] D.J. Sahn, A. DeMaria, J. Kisslo, A. Weyman, Recommendationsregarding quantitation in M-mode echocardiography: results of asurvey of echocardiographic measurements, Circulation 58 (1978)1072–1083.

[40] M.V. Simões, O.C. Almeida-Filho, A. Pazin-Filho, R.B.P. Castro,A. Schimdt, B.C. Maciel, J.A. Marin-Neto, Insuficiência cardíaca nadoença de Chagas, Rev. Soc. Cardiol. Est. São Paulo 10 (2000) 50–64.

[41] H. Acquatella, N.B. Schiller, J.J. Puigbó, H. Giordano, J.A. Suarez,H. Casal, N. Arreaza, R. Valecillos, E. Hirschhaut, M-mode andtwo-dimensional echocardiography in chronic Chagas’ heart disease,Circulation 62 (1980) 787–799.

[42] J.S. Oliveira, J.A. Mello-De-Oliveira, U. Frederigue Jr, E.C. Lima-Filho, Apical aneurysm of Chagas’ heart disease, Br. Heart J. 46(1981) 432–437.

[43] E.A. Nogueira, O.M. Uetti, in: A.G.M.R. Sousa, A.J. Mansur (Eds.),SOCESP: Cardiologia, Atheneu, Brazil, 1996, pp. 620–625.

[44] E.J.N. Camara, Segmental changes in contractility of the left heartventricle in Chagas’cardiomyopathy with and without ventriculardilatation, Arq. Bras. Cardiol. 60 (1993) 51–55.

[45] C. Mady, R.H. Cardoso, A.C. Pereira-Barretto, P.L. da Luz, G. Bel-lotti, F. Pileggi, Survival and predictors of survival in patients withcongestive heart failure due to Chagas’ cardiomyopathy, Circulation90 (1994) 3098–3102.

1124 A.M.B. Bilate et al. / Microbes and Infection 5 (2003) 1116–1124