J. Euk. Microbiol.. 41(1), 1994, pp. 47-54 0 1994 by the Society of Protozoologists

Distribution of a-Galactosyl-Containing Epitopes on Trypanosoma cruzi Trypomastigote and Amastigote Forms from Infected Vero Cells

Detected by Chagasic Antibodies THAIS SOUTO-PADRON,* IGOR C. ALMEIDA,** WANDERLEY de SOUZA* and LUIZ R. TRAVASSOS**.' *Institute de BioJisica Carlos Chagas Filho. Universidade Federal do Rio de Janeiro, Rio de Janeiro, RJ 21 941, and

**Disciplina de Biologia Celular, Escola Paulisla de Medicina, SZo Paulo, S P 04023-062, Brazil

ABSTRACT. Reactivity of different Trypanosoma cruzi developmental forms with purified Chagasic anti-a-galactosyl antibodies (anti- Gal) was studied using epimastigotes from axenic cultures, trypomastigotes and amastigotes from infected Vero cell cultures, and an immunogold labeling method as observed by electron microscopy. Epimastigotes were poorly labeled, whereas extracellular trypomas- tigotes and amastigotes bound heterogeneously to the antibody with many cells being intensely labeled at the cell surface, including the membrane lining the cell body, the flagellum and the flagellar pocket. Parasites with poor labeling at the cell surface generally had several gold particles within the cell, mostly in cytoplasmic vacuoles. The Golgi complex of trypomastigotes was strongly labeled. Intracellular parasites were labeled at the parasite cell surface or within vacuolar structures. The expression in T. cruzi-infected Vero cells of a-galactosyl antigenic structures acquired from the parasite was shown by moderate labeling with Chagasic anti-Gal of the membrane lining parasite-free outward cell projections. The reactivity with purified anti-Gal from healthy individuals at the same concentrations of Chagasic anti-Gal was poor, with gold particles appearing in the nucleus and cytoplasm but not at the cell surface. I t paralleled the labeling with Bandeireae simplicifolia IB-4 lectin. The results provide a basis for autoimmune reactions involving anti- Gal from chronic Chagasic patients.

Supplementary key words. Anti-a-galactosyl antibodies, Chagas disease, immuno-gold labeling.

ATURAL antibodies, designated anti-Gal or anti-a-galac- N tosyl IgG, constitute as much as one percent of all cir- culating immunoglobulin in healthy humans [ 11, 141. It is be- lieved that bacteria of the normal intestinal flora provide the constant antigenic stimulation for the synthesis of natural anti- Gal. These antibodies react specifically with the carbohydrate structure Gala( 1 - 3)-Gal/3( 1 + 4)-GlcNAc [ 131 which is found in glycoproteins (e.g. murine laminin) and glycolipids from mammalian cells, except those of man and Old World monkeys and apes [ 15, 27-29]. Patients with Chagas disease and Amer- ican leishmaniasis exhibit high titers of anti-Gal which react with the same carbohydrate structures recognized by natural anti-Gal [4, 161, apparently with the same affinity. Chagasic (Ch) anti-Gal from patients with acute [ 17, 181 or chronic [ 1, 261 disease, both binding to mouse laminin, lyse Trypanosoma cruzi trypomastigotes in complement-independent or comple- ment-mediated reactions, respectively. Almeida et al. [ 11 ob- served, however, that anti-Gal from chronic Ch patients, when compared on a weight basis with anti-Gal from healthy indi- viduals, show an increased affinity for trypomastigote antigens as determined by agglutination, immunofluorescence and im- munolysis reactions. This seems to indicate that the a-galactosyl epitopes of the parasite differ in chemical structure or are dif- ferently expressed in trypomastigotes as compared to other mammalian cells. Alternatively, the antigen-driven process in- duced by T. cruzi might elicit antibodies bearing idiotypes re- lated to, but not identical with those of the natural anti-Gal, or an altogether new idiotype [30].

In the present work we investigated the distribution of a-ga- lactosyl epitopes in intra- and extra-cellular parasites growing in Vero cells, by reaction with anti-Gal isolated from sera of chronic Ch patients. In previous studies [6], a-galactosyl resi- dues in T. cruzi and Leishmania spp. have been detected using immunocytochemical methods with lectin and antibodies. The specific reagent was a mouse monoclonal antibody (Gal- 13) raised against rabbit red cells rich in a-galactosyl glycosphin- golipids [ 121.

The relevance of the present line of investigation is based on the following observations: a) antibodies that lyse trypomasti- gotes protect against T. cruzi [2 1-23] and are induced in patients

I To whom correspondence should be addressed.

with active infection [20, 241; and b) a high proportion (>70%) of the lytic power of the Ch serum is lost after removal of anti- Gal antibodies [ 11. Moreover, the presence of a-galactosylated products [l 11 or whole T. cruzi a-galactosyl antigens at the surface of infected cells, as revealed by high affinity anti-Gal from Ch patients, would provide a molecular basis for auto- immune response in Chagas' disease.

MATERIALS AND METHODS Parasites. The Y strain of Trypanosoma cruzi grown axe-

nically or in mammalian cell culture was used for immunocy- tochemical localization of a-galactosylated antigens. Epimasti- gotes were cultivated for five days at 28" C in liver-infusion trypticase (LIT) medium containing 10% fetal calf serum [7]. Trypomastigotes and amastigotes were obtained from the su- pernatant fluids of Vero cell cultures infected with bloodstream trypomastigote forms of T. cruzi as previously described [S].

Vero cells. Kidney fibroblasts ofAfrican green monkey (Vero cells) were maintained in Falcon plastic flasks using the 199 medium with 5% fetal calf serum and transferred to new flasks after trypsinization, when the cell density nearly formed a con- fluent monolayer. One day before being used, Vero cells were plated on Falcon multiwell tissue culture plates (15 mm di- ameter/well) with coverslips, and maintained at 37" C, over- night, in 5% CO,.

Isolation of anti-Gal antibodies. Anti-a-galactosyl antibodies were isolated from sera of healthy individuals and patients with Chagas' disease as described previously [ 11. Briefly, anti-Gal were obtained by chromatography on Synsorb matrix (silica particles; ChemBiomed, Edmonton, Canada) linked to Gala( 1 + 3)-Gal@( 1 -+ 4)-GlcNAc (0.88 pmol oligosaccharide/g ofSyn- sorb). Nonspecific antibody binding was minimized by adding 0.5 M NaCl to the washing buffer. Bound antibodies were eluted with 50 mM citric acid, pH 2.8 and immediately neutralized with 1 M Tris.HC1, pH 8.5. The antibody solution was then purified in Sepharose-protein A for isolation of the IgC fraction, dialyzed against phosphate-buffered (0.15 M) saline (PBS, pH 7.4), and concentrated to I mg protein/ml. Cell labeling was carried out with 100 pdml of antibody. Control of anti-Gal reactivity was evaluated by agglutination of trypomastigotes from Vero cell culture at a minimum antibody concentration of 5-1 0 pg/ml. These antibodies bind to a-galactosyl N-acetyl- lactosamine or melibiose [ I ] and specifically recognize a-galac-

47

48 J . EUK. MICROBIOL., VOL. 41, NO. 1 , JANUARY-FEBRUARY 1994

Fig. 1 4 . Labeling of epimastigote and amastigote forms with anti-Gal from Chagasic patients. 1. Labeling of epimastigote forms with anti- Gal antibody. Note that there IS a slight labeling in the cytoplasm (arrows) and cytoplasmic compartments (*). The membrane lining the cell body and flagellum presents very few gold particles (small arrows). N, nucleus. x 28,000. 2 4 . Amastigotes obtained from the supernatant of infected cells. All amastigote forms were labeled with anti-Gal but labeling was heterogeneous. An intense surface labeling was observed in some amastigotes,

SOUTO-PADRON ET AL. --a-GALACTOSYL EPITOPES IN TRYPANOSOMA CRUZI 49

Table 1. Inhibition of Chagasic (Ch) anti-Gal reactivity with cell- derived trypomastigotes as measured by agglutination (AG) and indirect immunofluorescence (IIF) reactions.

Minimal Ch anti-Gal AG concentration IIF with 10 wghl

Additions h d m l P Ch anti-Ga

0 a-galactosidase (1 U) a-methylGalp (mM)

2 32 64

2 32 64

(3-methylGalp (mM)

4.7 Positive -b Negative'

9.4 75

150 Negative

4.7 9.4

18.8 Positive

taining the anti-Gal antibody. After incubation, the grids were washed three times in PBS-BSA-TW, pH 8, and incubated for 60 rnin at room temperature in PBS-BSA-TW with gold-labeled protein A, diluted 1:40. Control grids were made incubating a) infected cells and parasites with anti-Gal from healthy individ- uals or only with gold-labeled protein A; trypomastigotes with anti-cysteine proteinase antibody, provided by J. J. Cazzulo, Fundaci6n Campomar, Argentina; b) normal cells with anti-Gal isolated either from Chagasic patients or healthy individuals; c) T. cruzi-infected and normal cells with lectins (Bandeireae sim- plicifolia I-B4 and Arachis hypogaea or peanut agglutinin, both from Sigma), complexed to colloidal gold particles (10 nm). After this step, the grids were washed with distilled water, stained with uranyl acetate and lead citrate and observed in a Zeiss EM 902 or EM 900 transmission electron microscope.

RESULTS AND DISCUSSION

with a-galactosidase at acid pH.

tained even with anti-Gal at 160 &ml.

and melibiose but not by 0-methylgalactopyranoside or lactose [ 13. With cell-derived forms ( 5 x 107/ml), Chagasic anti-Gal at 5-1 0 wglml caused strong agglutination and this effect was spe-

c Negative IIF after a-galactosidase treatment of fixed cells was ob-

tosyl structures on the parasite cell surface as shown by elimi- nation of their reactivity by a- but not P-glycosides.

Immunofluorescence staining of extracellular forms. Free try- pomastigotes and a few amastigotes derived from infected cells in culture were stained by the indirect immunofluorescence method using antiGal from Ch patients (at 10 &mi), bioti- nylated anti-human IgG (Gibco-BRO, Gaithersburg, MD) 1 :500, and Extravidin-FITC (Sigma Chemical Co., St. Louis, Mo) 1 : 100. Incubations of slides with p-formaldehyde-fixed parasites after addition of each reagent were for 30 rnin at room temperature, followed by three-four PBS washings. Observations were made with a Nikon fluorescent microscope.

Parasite-host cell interaction. Interaction of trypomastigotes and nonconfluent Vero cells was allowed for 18 h in the 199 medium supplemented with 5% heat-inactivated fetal calf se- rum. Cells were then washed twice with PBS (pH 7.2) to remove extracellular parasites and the coverslips were processed for transmission electron microscopy (TEM) or light microscopy at days three, four, and five post-infection.

Electron microscopy and immunocytochemistry. Parasites were collected by centrifugation (2,000 g/10 min/4' C), washed twice with PBS and fixed for 60 min in a solution containing 0.1 O/O glutaraldehyde, 4% formaldehyde (freshly prepared from paraformaldehyde), 0.02% picric acid, in 0.1 M phosphate buff- er, pH 7.2. To analyze intracellular parasites, infected Vero cells were fixed after different post-infection periods. Cells were fixed for 60 rnin at room temperature in 1 O/o glutaraldehyde, 4% form- aldehyde, 5 mM CaCI,, 0.02% picric acid, in 0.1 M cacodylate buffer, pH 7.2. After washing in the same buffer, cells weregently scraped off with a rubber policeman. After fixation, parasites and normal and infected cells were rinsed twice in PBS, dehy- drated in methanol, and embedded in Lowicryl K4M at -20" C [ 5 ] . Thin sections were collected on 300-mesh nickel grids, incubated subsequently for 30 rnin at room temperature in PBS, pH 8, with 1.5% bovine serum albumin (BSA) and 0.019'0 Tween 20 (PBS-BSA-TW), and for 60 min in the same solution con-

cifically inhibited by a- but not P-methylgalactopyranoside (Ta- ble 1). Abolition of antibody reactivity by treatment with a-ga- lactosidase could only be tested with fixed cells by indirect immunofluorescence, since under the conditions used (1 U of enzyme from green coffee beans, pH 5-6, and incubation for 18 h), live cells were badly affected and could not be used in ag- glutination reactions. Negative immunofluorescence of a-galac- tosidase-treated fixed trypomastigotes even at anti-Gal concen- trations as high as 160 pg/ml, as well as inhibition of agglutination of live parasites by a- but not 0-galactoside, confirms the spec- ificity of the antibody for a-galactosyl epitopes expressed at the cell surface of trypomastigotes.

Labeling of epimastigotes. Epimastigotes grown in LIT me- dium were poorly labeled with Ch anti-Gal and gold-labeled protein A, on the membrane lining the cell body and the fla- gellum (Fig. 1). Few particles could be observed inside the fla- gellar pocket (not shown), in the cytoplasm and within intra- cellular vesicles. Accordingly, by using monoclonal antibody Gal- 13, specific for Gala( 1 - 3)-Gal epitopes on glycolipids, anti-nidogen antibodies and Bandeiraea simplicifoiia IB4 lectin (BSI-B4), Bretafia and coworkers [6] could not detect receptors on intact epimastigotes that would bind these reagents. When swollen or degenerated epimastigotes were used to react with Gal- 13 or the lectin, a clear staining was observed with intra- cellular structures. These results suggest that a-galactosyl epi- topes in epimastigotes are only accessible to anti-Gal antibodies or lectin after alteration of the membrane structure. Previously, a reaction with anti-Gal from normal human sera was obtained with trypsinized PMSF-EDTA-treated epimastigotes dried on polystyrene plates [3]. Different treatments can, therefore, affect the expression of anti-Gal receptors in epimastigotes.

Labeling of amastigotes. All amastigote forms obtained from the supernatant fluid of Vero cells were labeled with Ch anti- Gal. The intensity of labeling varied among cells of the same population (Fig. 2-4). Strongly labeled amastigotes (Fig. 2, 3) concentrated most of the gold particles on the membrane en- closing the cell body and the short flagellum. Few particles were seen in the cytoplasm, some in intracellular vesicles.

t concentrated on the membrane lining the cell body, the flagellum and the flagellar pocket. 2. Surface labeling on the membrane lining the cell body; K, kinetoplast. x 2 1,000. 3. Surface labeling of the flagellum and flagellar pocket; F, flagellum. x 34,000. 4. In contrast, some amastigotes were poorly labeled with few particles distributed on the cell surface. x 3 1,000.

50 J. EUK. MICROBIOL., VOL. 41, NO. 1 , JANUARY-FEBRUARY 1994

Fig. 5-9. Labeling with anti-Gal of cell-derived trypomastigotes. 5. Trypomastigotes derived from Vero cell culture, labeled with Chagasic anti-Gal. A heterogeneous labeling was observed. Note the intense labeling of one of the trypomastigote forms as compared to the other. The strongly labeled cell has a large number of particles on the cell surface and on the flagellum (arrows), and also on the membrane lining the flagellar pocket (arrowheads). x 55,000. 6-7. Some trypomastigotes had very few or no particles on the cell surface. 6. Intense labeling of the flagellar pocket (arrows). 7. Intense labeling of cytoplasmic vacuoles (*). F, flagellum; K, kinetoplast. x 55,000. 8. Intense labeling with Chagasic anti-Gal of different stacks of the Golgi complex in a trypomastigote form. x 102,000.9. Labeling of a trypomastigote with anti-GaI obtained from normal human serum. Labeling of the parasite cell surface was virtually absent. A few particles were seen inside the cell (arrows). x 50,000.

SOUTO-PADRON ET AL.-a-GALACTOSYL EPITOPES IN TRYPANOSOMA CRUZI 5 1

Fig. 10-13. Labeling of intracellular parasites and infected cells. 10-12. All intracellular forms showed gold particles in the cytoplasm, cytoplasmic structures and vacuoles (arrowheads). 10. Amastigote with gold particles in the cytoplasm. x 12,000. 11. Some evolutive forms had only a few gold particles on the cell surface whereas others were intensely labeled (arrows). Gold particles are also seen in the cytoplasm and vacuoles. x 30,000. 12. Intracellular evolutive form with gold labeling in cytoplasmic structures. x 35,000. 13. Labeling of T. cmzi-infected cells. Infected cells showed particles localized on the membrane lining parasite-free cell surface projections (arrows). x 82,000.

52 J. EUK. MICROBIOL., VOL. 41, NO. I , JANUARY-FEBRUARY I994

SOUTO-PADRON ET AL.-a-GALACTOSYL EPITOPES IN TR YPANOSOMA CRUZZ 53

Labeling of trypomastigotes. Labeling with Ch anti-Gal of extracellular trypomastigotes from Vero cell culture was also heterogeneous. Figure 5 shows two parasites: one is intensely labeled at the cell surface, including the membrane lining the cell body, the flagellum and the flagellar pocket; the other is poorly labeled at the cell surface. Some trypomastigote forms from the same sample showed few or no particles at the cell surface but were intensely labeled in the flagellar pocket and in cytoplasmic vacuoles which were found near the flagellar pocket (Fig. 6) or were distributed along the cytoplasm (Fig. 7). In general, trypomastigotes that were poorly labeled at the cell surface were intensely labeled inside the cell, probably reflecting stages of antigen processing. The Golgi complex, as observed in thin sections of trypomastigotes, was strongly labeled (Fig. 8) with Ch anti-Gal. Several particles could be seen along the different Golgi stacks. Apparently, the glycoconjugates express- ing a-galactosyl epitopes in trypomastigotes are processed in the Golgi complex and then included in several vesicles or vacuoles before insertion in the plasma membrane. The parasite a-ga- lactosyl transferase may be involved in the synthesis of both gl ycolipid and glycoprotein complexes which are recognized with great affinity by Ch anti-Gal. In comparison, the reactivity with anti-Gal from healthy individuals was very poor (Fig. 9). Few particles could be seen inside the parasite, with virtually no labeling at the cell surface. This result confirms our previous work [I], which showed a remarkable difference in the affinity for T. cruzi a-galactosyl glycoconjugates of Ch as compared to normal anti-Gal. The heterogeneity of labeling of Vero cell- derived trypomastigotes was also observed using the monoclo- nal antibody Gal- 13. Although trypomastigote labeling with anti- laminin antibodies usually coincided with that of Gal-I 3, the latter seemed to attach more loosely to the parasite 161. This may reflect the recognition of different a-galactosyl molecules by these reagents. In fact, possible structures for epitopes being recognized by anti-Gal include the more universal Gala(l - 3)-Galfl(l - 4)-GlcNAc but also Galpa(1 - 3)-Galfp and Galpa( 1 --+ 6)-Galpa( 1 + 3)-Galfp as present in glycosylinosi- tolphospholipids (GIPL) (251 of Leishmania major, and pos- sibly also of T. cruzi, or even the presumed Gala( 1 -+ 3)-Man structure [ 2 ] . Recently, we have indeed shown, using a chemo- luminescence method, that Ch but not normal antiGal binds to GIPL 2 and 3 of L. major. Also, the proximity of sialic acid and a-galactosyl residues as proposed for the oligosaccharide chain of T. cruzi 85-kDa glycoprotein [9] could modify antibody affinity or constitute a new epitope, as described for certain mammalian sialyloligosaccharides recognized by monoclonal antibodies [ 191. Normal and Ch anti-Gal may bind nondiscri- minating substrates such as murine laminin or Synsorb-im- mobilized Gala( 1 + 3)-Galp( 1 -+ 4)-GlcNAc with comparable affinity. Specific structures on T. cruzi expressing terminal a-Galp are, however, recognized with greater affinity by Ch anti-Gal.

Labeling of intracellular parasites. Incubation with Ch anti- Gal of thin sections of Vero cells infected with trypomastigotes was followed by intracellular labeling of amastigote and try-

pomastigote forms (Fig. 10-1 2). Labeling again was irregular for different parasites infecting the same cell, ranging from few cytoplasmic gold particles to intense surface labeling. Both ev- olutive forms observed inside Vero cells showed labeling of cytoplasmic vacuoles. Possibly, the different labeling of intra- cellular forms with anti-Gal reflects the various stages of ex- pression and activity of the parasite a-galactosyl transferase in a host cell that does not synthesize this enzyme, as well as the subsequent processing and relocation of the corresponding gly- cogonjugates in parasites of different growth and developmental phases. Generally, a strong surface reaction accompanied a weak cytoplasmic labeling. Conversely, a more intense intracellular or vacuolar labeling was observed in parasites with a more discrete surface labeling, suggesting different stages of processing of the carbohydrate epitope in different parasites.

Labeling of infected host cells. Insertion in the host cell mem- brane of parasite antigens has been recognized in several in- stances. In the case of T. cruzi-infected cells, a membrane in- sertion of a-galactosyl glycoconjugates from intracellular forms of the parasite, or the introduction of terminal a-galactosyl units on acceptor molecules of the host cell membrane by action of the parasite a-galactosyl transferase could, in both cases, cause the abnormal expression of a-galactosyl epitopes. Such abnor- mal expression might trigger an autoimmune process in man analogous to that hypothesized by Galili [ 101 involving a patho- logical deregulation of the human gene encoding a-galactosyl transferase activity, which is normally suppressed in human tissues. Autoimmunity would occur mainly due to the elevated concentrations in man of anti-Gal antibodies. As observed in the present investigation, infected Vero cells incubated with purified Ch anti-Gal clearly showed the presence of gold par- ticles at the cell surface mainly located on the membrane lining cell projections as shown in Fig. 13. This is an evidence for the surface expression of a-galactosyl units in infected cells that normally do not contain Gala( 1 -+ 3)-Gal structures. The ab- sence of parasites in these cell projections suggests that the la- beled components must have been processed for their external plasma membrane expression via internal camers, probably vesicles, free of the intracellular parasite itself. Parasite products, however, could have interacted with newly synthesized or re- cycled membrane of the host cell especially at sites of high turn- over as expected in outward cell projections.

Since binding of Ch anti-Gal to cell surface antigens of in- fected cells apparently involved fewer accessible epitopes than those present at the parasite cell surface or cytoplasm, we com- pared the anti-Gal reactivity with that ofpeanut agglutinin (PNA), which recognizes various P-galactosyl-structures in mammalian cells (Fig. 14). As shown, the intensity ofthe reaction was similar to that obtained with Ch anti-Gal, with few gold particles bind- ing to the surface of infected cells. That the reactivity of Ch anti-Gal depends on the density ofa-galactosyl epitopes exposed at the cell surface, is clearly shown by antibody binding to cell- derived trypomastigotes (Fig. 5,7), which was even more intense than that observed with antibodies directed against another spe-

t

Fig. 14-19. Positive and negative control systems for comparison with Chagasic anti-Gal labeling of parasites, and of normal and infected cells. 14. Labeling of T. cruzi-infected cells with PNA-colloidal gold complex. A few particles were seen on the cell surface (arrowhead) and in the cytoplasm of infected cells. x 3 1,000. 15. Vero cell-derived trypomastigotes incubated with anti-cysteine-proteinase antibody. Two parasites showed gold particles on the cell surface (arrowhead) and cytoplasm whereas the other cell was very weakly labeled. x 24,000. 16. Normal cells incubated in the presence of anti-Gal from healthy individuals. x 39,000. 17. Normal cells incubated wiht Chagasic anti-Gal. No particles were seen on cell surface projections of these cells. x 39,000. 18. Labeling of T. cruzi-infected cells with BSI-B4 lectin-colloidal gold complex. No particles were seen on the cell surface. A few particles were seen in the cytoplasm (arrowhead) and nucleus (arrow). x 20,000. 19. Labeling of T. cruzi-infected cells with anti-Gal from healthy individuals. Note the absence of particles on the cell surface but the presence of gold particles in the nucleus (arrows) and cytoplasm (arrowhead). x 13,000.

54 J. EUK. MICROBIOL., VOL. 41, NO. 1, JANUARY-FEBRUARY 1994

cific T. cruzi marker, the cysteine-proteinase (Fig. 15). The sur- face reactivity of T. cruzi-infected cells with anti-Gal clearly differed from that of normal noninfected cells that did not bind this antibody (Fig. 16, 17).

Natural anti-Gal reactivity. As stressed above, the reactivity of natural anti-Gal from healthy individuals with cell-derived T. cruzi is much weaker than that observed with Ch anti-Gal a t the same concentrations. The same occurred with T. cruzi- infected Vero cells. Gold particles were not seen a t the cell surface using either BSI-B4 colloidal gold complex (Fig. 18) or natural anti-Gal (Fig. 19). A few particles were detected in the nucleus and cytoplasm.

Except for the more generally recognized expression of 0-linked single N-acetylglucosamine units in nuclear and cy- tosolic macromolecules, other carbohydrate structures have not, so far, been recognized external t o the Golgi-endoplasmic-re- ticulum system. Nuclear labeling with monospecific a-galactosyl reagents, as shown here, should therefore be further explored to identify the molecular species involved.

The present results confirm and extend previous data [ 1, 61 on the presence of a-galactosyl epitopes in T. cruzi and propose a mechanism for the start of an autoimmune process which depends on the host cell abnormal expression of parasite-de- rived a-galactosyl epitopes of still undefined structure. Anti- bodies involved in the initiation of this process could be the high affinity anti-Gal from chronic Ch patients, rather than the natural anti-Gal from healthy individuals.

ACKNOWLEDGMENTS The present work was supported by CNPq (Brazil) and the

United Nation Development Program/World BanWWHO Spe- cial Program for Research and Training in Tropical Diseases and the Rockefeller Foundation.

LITERATURE CITED 1. Almeida, I. C., Milani, S. R., Gonn, P. A. J. & Travassos, L. R.

199 1. Complement-mediated lysis of Trypanosoma cruzi trypomas- tigotes by human anti-a-galactosyl antibodies. J. Immunol., 148:2394- 2400.

2. Avila, J . L. & Rojas, M. 1990. A galactosyl (al-3)mannose ep- itope on phospholipids of Leishmania mexicana and L. brasiliensis is recognized by trypanosomatid-infected human sera. J. Clin. Microbiol., 28:1530-1537.

3. Avila, J . L., Rojas, M. & Galili, U. 1989. Immunogenic Gala1 - 3Gal carbohydrate epitopes are present on pathogenic American Trypanosoma and Leishmania. J . Immunol.. 142:2828-2834.

4. Avila, J. L., Rojas, M. & Towbin, H. 1988. Serological activity against galactosyl a- I -+ 3 galactose in sera from patients with several kinetoplastida infections. J. Clin. Microbio/., 26: 126-1 32.

5. Bendayan, M., Nanci, A. & Kan, F. W. K. 1987. Effect of tissue processing on colloidal gold cytochemistry. J. Histochem. Cyfochem.,

6. Bretaiia, A., Avila, J. L., Contreras-Bretaiia, M. & Tapia, F. J. 1992. American Leishmania spp. and Trypanosoma cruzi: Galactosyl a( 1 + 3) galactose epitope localization by colloidal gold immunocy- tochemistry and lectin cytochemistry. Exp. Parasitol., 74:27-37.

7. Camargo, E. P. 1964. Growth and differentiation in Trypano- soma cruzi. I. Origin of metacyclic trypanosomes in liquid media. Rev. Inst. Med. Trop. SGo Paulo, 6:93-100.

8. Carvalho, T. U. & De Souza, W. 1983. Separation ofamastigotes and trypomastigotes of Trypanosoma cruzi from culture cells. Z . Par- asitenkd., 69:57 1-5 75,

9. Couto, A. S., GonGalves, M. F., Colli, W. & Lederkremer, R. M. 1990. The N-linked carbohydrate chain of the 85-kilodalton glycopro- tein from Trypanosoma cruzi trypomastigotes contains sialyl, fucosyl and galactosyl (a 1 -3)galactose units. Mol. Biochem. Parasitol., 39: 10 1- 108.

35:983-996.

10. Galili, U. 1989. Abnormal expression of a-galactosyl epitopes in man. Lancet, II:358-361.

1 1. Galili, U. 1991. The human natural anti-gal antibody and in- fectious agents. Infect. Immun. (Life Sci. Adv., India), 10:4146.

12. Galili, U., Basbaum, C. B., Shohet, S. B., Buehler, J. & Macher, B. A. 1987. Identification of erythrocyte Gala1 -.+ 3Gal glycosphin- golipids with a mouse monoclonal antibody, Gal-13. J. Biol. Chem., 262:46834688.

13. Galili, U., Macher, B. A,, Buehler, J. & Shohet, S. B. 1985. Human natural anti-a-galactosyl IgG. 11. The specific recognition of a( 1 -3)-linked galactose residues. J. Exp. Med., 162:573-582.

14. Galili, U., Rachmilewitz, E. A., Peleg, A. & Rechner, L. 1984. A unique natural human IgG antibody with anti-a-galactosyl specificity. J. Exp. Med., 160:1519-1531.

15. Galili, U., Shohet, S. B., Kobnn, E., Stults, C. L. M. & Macher, B. A. 1988. Man, apes and Old World monkeys differ from other mammals in the expression of a-galactosyl epitopes on nucleated cells. J. Bid. Chem., 263: 17755-1 7762.

16. Gazzinelli, R. T., Galvlo, L. M. C., Cardoso, J. E., CanCado, J. R., Krettli, A. U., Brener, Z. & Gazzinelli, G. 1988. Anti-Trypanosoma cruzi and anti-laminin antibodies in Chagasic patients after specific treatment. J. Clin. Microbiol., 26: 1795-1800.

17. Gazzinelli, R. T., Galvlo, L. M. C., Dias, J. C. P., Gazzinelli, G. & Brener, Z. 1988. Anti-laminin and specific antibodies in acute Cha- gas disease. Trans. R. SOC. Trop. Med. Hyg., 82:574-576.

18. Gazzinelli, R. T., Pereira, M. E. S., Romanha, A., Gazzinelli, G. & Brener, Z. I99 1. Direct lysis of Trypanosoma cruzi: a novel effector mechanism of protection mediated by human anti-gal antibodies. Par- asite Immunol., 13:345-356.

19. Guo, N., Her, G. R., Reinhold, V. N., Brennan, M. J., Siraganian, R. P. & Ginsburg, V. 1989. Monoclonal antibody AA4, which inhibits binding of IgE to high affinity receptors on rat basophilic leukemia cells, bind to novel a-galactosyl derivatives ofganglioside G,,,. J. Biol. Chem.,

20. Kierszenbaum, F. 1976. Cross-reactivity of lytic antibodies against blood forms of Trypanosoma cruzi. J . Parasitol., 62: 134-135.

2 1. IOerszenbaum, F. & Howard, J. G. 1976. Mechanisms of re- sistance against experimental Trypanosoma cruzi infection: the impor- tance of antibodies and antibody-forming capacity in the Biozzi high and low responder mice. J. Immunol., 116: 1208-1 2 1 1.

22. Krettli, A. U. & Brener, Z. 1976. Protective effects of specific antibodies in Trypanosoma cruzi infections. J. Immunol., 116:755-760.

23. Krettli, A. U. & Brener, Z. 1982. Resistance against Trypano- soma cruzi associated to anti-living trypomastigote antibodies. J. Im- munol., 128:2009-2012.

24. Krettli, A. U., CanCado, J. R. & Brener, Z. 1982. Effect of specific chemotherapy on the levels of lytic antibodies in Chagas's dis- ease. Trans. R. SOC. Trop. Med. Hyg., 76:334-340.

25. McConville, M. J., Homans, S. W., Thomas-Oates, J. E., Dell, A. & Bacic, A. 1990. Structures of the glycoinositolphospholipids from Leishmania major. A family of novel galactofuranose-containing gly- colipids. J. Biol. Chem., 265:7385-7394.

26. Milani, S. R. & Travassos, L. R. 1988. Anti-a-galactosyl an- tibodies in Chagasic patients. Possible biological significance. Braz. J . Med. Biol. Res., 21:1275-1286.

27. Suzuki, E. & Naiki, M. 1984. Heterophilic antibodies to rabbit erythrocytes in human sera and identification of the antigen as a gly- colipid. J. Biochem., 83: 103-1 08.

28. Thall, A. & Galili, U. 1990. Distribution of Gala1 - 3GalPl - 4GlcNAc residues on secreted mammalian glycoproteins (thyroglob- ulin, fibrinogen, and immunoglobulin G) as measured by a sensitive solid-phase radioimmunoassay. Biochemistry, 29:3959-3965.

29. Towbin, H., Rosenfelder, G., Wieslander, J., Avila, J. L., Rojas, M., Szarfman, A., Esser, K., Nowack, H. & Timpl, R. 1987. Circulating antibodies to mouse laminin in Chagas disease, American cutaneous leishmaniasis and normal individuals recognize terminal galactosyl (a 1 - 3) galactose epitopes. J. Exp. Med., 166:419432.

30. Unterkircher, C. & Blotta, M. H. S. L. 1991. Polyclonal acti- vation and Chagas's disease. Ci. Cult.. J. Braz. Assoc. Adv. Sci., 43:358- 362.

264:13267-13272.

Received I I-18-92, 4-26-93, 7-23-93; accepted 8-16-93