Patient 1: chronic chagasic/ HIV positive patient, from Minas Gerais State, Brazil, presenting indeterminate form acquired from natural transmission before the HIV infection. The CD4 lymphocyte count was 102 mm³, when the patient died of intracranial hemorrhage 15.

Patient 2: hemophiliac/ HIV positive patient from Bahia State, Brazil, manifesting cardiac form of Chagas disease acquired from natural transmission. The CD4 lymphocyte count was higher than 500 mm³ 16.

Four T. cruzi reference strains: representing the major zymodeme groups 17: Z1 – Dm28c and Colombian (T. cruzi I), Z2 – CL strain, Z2b – Y strain (T. cruzi II) were used as control in the molecular typing. The two T. cruzi strains from Chagas disease/HIV co-infection human cases were prior isolated by xenodiagnosis and propagated in biphasic cultures (Novy-Nicolle-McNeal – NNN/ Brain Heart Infusion-BHI/ 10% Fetal Bovine Serum, complemented with 20 mg/ml ampicillin, 0.1 mg/ml streptomycin and 100 U/ml penicillin at 28°C. The parasite growth was accompanied during five passages. The cell mass was prepared and the parasites characterized by molecular markers according to protocols previously described 181920. Concomitantly, groups of ten male Swiss mice weighting 8–10 g were inoculated, intraperitoneally, with culture forms and examined through fresh blood slides after 20 days. Mice, presenting patent parasitemia, were maintained through passages with 10⁴ bloodstream trypomastigotes, on the 21^st day of infection for five months. After passage in mice, both strains were re-isolated in biphasic cultures (without antibiotics) at 28°C with serial passages on the 21^st day. The primary isolates were kept in culture for twelve months and cryopreserved in liquid nitrogen.

In the chronic phase of the disease (six months of infection), mice were immunosuppressed with 50 mg of cyclophosphamide per kg of body weight for eighteen consecutive days 21. After immunosuppression parasites were recovered from the blood and characterized using molecular approaches.

The experiments in mice were conducted using a protocol approved by the Center for Biological Evaluation and Care of Research Animals (CEUA-FIOCRUZ, Program n° P0067-00) and in accordance with the code of ethics of the Brazilian College of Animal Experimentation (COBEA).

T. cruzi populations obtained from in vivo and in vitro conditions were compared with the primary uncloned isolates. The cell mass was prepared washing the parasites twice by centrifugation at 700 × g in SE buffer (0.15 M NaCl / 0.1 M EDTA, pH 8,0) for 10 min at 4°C. The cell mass was stored at -70°C before being used for multilocus enzyme electrophoresis and kDNA analyses.

Isoenzymatic analyses were performed on 1% agarose gel according to protocols previously described 1720. The following nine enzymes were used: glucose 6 phosphate dehydrogenase (G6PDH; E.C. 1.1.49), glucose phosphate isomerase (GPI; E.C.5.3.1.9), isocitrate dehydrogenase (IDH; E.C. 1.1.1.42), malate dehydrogenase (MDH; E.C.1.1.1.37), malic enzyme (ME; E.C.1.1.1.40), nucleotidase (NH, E.C.3.2.2.1), peptidase D (PEP-D, E.C.3.4.13.9), 6-phosphogluconate dehydrogenase (6PG; E.C. 1.1.44), phosphoglucomutase (PGM; E.C.2.7.5.1).

The kDNA extraction and restriction enzyme analyses were carried out according to the methodology previously described 1819.

Approximately 1 μg of purified kDNA was digested by the restriction enzymes Hinf-I at 37°C for 3 hrs in the appropriate buffer. kDNA fragments (maxi and minicircles) were separated through electrophoresis in 1.5% agarose gel in TBE buffer (0.89 M Tris-base; 0.89 M boric acid; 0.024 M EDTA). The electrophoretic pattern was visualized under UV light after ethidium bromide staining.

Fragments of kDNA were transferred to a nylon membrane according to Southern 22, the gel having previously been denatured in 1.5 M NaCl/0.5 M NaOH and neutralized in 1 M Tris-HCl pH 8.0. Each step was carried out during 30 min with mild stirring.

The total kDNA from primary isolate of Patient 1, named as kDNA profile I and populations recovered from in vivo (kDNA profile II) and in vitro conditions (kDNA profile III) were radiolabeled with α³² P dATP by the random primer reaction 23 and used as probes. After hybridization, the filters were washed three times during 30 min in 0.1x SSC/ 0.5% SDS at 60°C. Autoradiography was performed using X-ray film (Kodak X-OMAT) at -70°C overnight.

The isoenzymatic profiles were compared using Simple Matching (SM) coefficient of similarity to determine the proportion of mismatched bands. The matrix was transformed into a dendrogram by the UPGMA (unweighted pair-group method with arithmetic averages) algorithm 24 using NTSYS program (version 2.0, Exeter Software, 100 North Country Rd, Setauket, NY).

The primary isolates from both patients were prior characterized by MLEE and restriction analyses of kDNA revealing distinct zymodemes and kDNA restriction profiles. Numerical analysis based on patterns generated by nine enzymatic systems has allowed the division of the strains into two clusters separated by a coefficient of similarity of 0.73. The first cluster was formed by the strain isolated from Patient 1 and T. cruzi strains representative of domestic zymodemes (Z2) sharing different coefficients of similarity (from 0.82 to 0.94). No T. cruzi strains analyzed was found to be 100% similar (SM = 1). The strain isolated from Patient 1 was identified as isoenzymatic variant of the reference strains (CL and Y) associated with the domestic cycle (Z2). The sylvatic Z1 zymodeme (represented by the T. cruzi strains Dm28c and Colombian) was included in the second cluster sharing approximately 85% of common characteristics. The strain from Patient 2, considered the most differentiated, was linked to the two main clusters by a coefficient of 0.60 (Fig. 1). The isoenzymatic pattern observed for such strain was compatible with strains belonging to sylvatic zymodemes in four (PGM, GPI, PEP-D and 6PG) out of the nine enzymes tested.

Hinf-I restriction analysis of kDNA has demonstrated the presence of three different genotypic profiles when the primary isolate from Patient 1 (kDNA profile I) and parasite populations from in vivo (kDNA profile II) and in vitro (kDNA profile III) conditions were compared. In fact three different parasite subpopulations were selected from Patient 1 (Fig. 2A, lanes 1, 2 and 3). The genotypic profiles of the populations that emerged in vivo and in vitro were distinct. However, only one genetic pattern was detected after immunosuppression in mice, independent of the initial inoculum (Fig. 2A, lanes 4 and 5). In figure 2B the region of minicircle fragments ranging from 600 to 1500 bp was amplified to better visualize the kDNA polymorphisms. Polymorphisms were not detected in kDNA fragments of the parasite isolated from Patient 2, before and after immunosuppression (not shown).

The genotypic differences observed in the kDNA restriction profiles (kDNA profiles I, II and III) were also evaluated in cross hybridization experiments using the same kDNA profiles as homologous probes (Fig. 3). When kDNA profile I (Patient 1) was used as probe, a strong sequence homology with the homologous kDNA and with sequence from profile II populations was detected (Fig. 3B, lanes 1 and 2). However an absence of cross-homology with kDNA from in vitro populations (profile III) was perceived (Fig. 3B, lane 3). Curiously, the probe prepared from kDNA profile III only hybridizes with its homologous sequences and with those from kDNA profile I (Fig. 3C, lanes 1 and 3). Interestingly the kDNA probe prepared from parasites recovered after passage in immunosuppressed mice (kDNA profile II) did not recognize homologous sequences in kDNA profile III (Fig. 3D, lane 3).

Analysis of nine enzymatic systems (G6P, GPI, IDH, MDH, ME, NH, PEP-D, 6PG, PGM) revealed no evident isoenzymatic variation among the primary isolate from Patient 1 and populations recovered after passage in immunosuppressed mice. Distinct isoenzymatic variation was observed in T. cruzi populations from in vitro conditions. The primary isolate from Patient 2 and parasites isolated after passage in immunosuppressed mice displayed identical isoenzymatic patterns (not shown).