ABC | Volume 110, Nº4, April 2018

Original Article Nascimento et al Vimentin and Anti Vimentin Antibodies in Chagas’ Disease Arq Bras Cardiol. 2018; 110(4):348-353 vimentin from bovine lens was obtained commercially (Sigma Aldrich, Saint Louis, Missouri, USA). A serum from patient with cardiac chronic American Trypanosomiasis was used as anti T.cruzi antibody. Human sera from American Trypanosomiasis patients and controls were used from the biorepository of T.cruzi patients samples from E.S.Umezawa, Lab.Protozoology, IMTSP, serologically characterized in TESA specific serology tests and published previously in several articles, were recovered and comprising 26 sera from acute disease, 33 from isolated cardiac disease, 17 from isolated digestive disease, 20 without clinical disease (asymptomatic disease) and 40 sera from patients outside endemic area. All clinical data were maintained by the attendant physician and not available for this study. Antigen expression and morphology All morphological assays were performed in a Zeiss Axioplan epifluorescent microscope with fluorescein filters. For antigen detection, we fixed LLC-MK2 control cells, T.cruzi infected LLC-MK2 cells and T.cruzi epimastigotes and permeated cell surface with Triton X-100 7 with either anti-Vimentin mAb or anti- T.cruzi antibodies as elsewhere described. After this step, bound antibodies were revealed with adequate fluorescein conjugate, carefully washed and mounted in glycerin for observation. Representative Fields were digitalized at high power field using a Canon camera. TESA and vimentin ELISA T.cruzi trypomastigotes excreted secreted antigen was obtained as elsewhere described 8 . TESA (1/80) and Vimentin (0.06ug/ml) in carbonate 0.05 M pH9.6 were adsorbed overnight to wells of 96 wells high binding ELISA plates (Corning Inc. New York, USA). After washing and blocking with PBS Tween 20, 0,05% plus 5%milk or BSA 0.5%, adequate dilution of sera (1/50 vimentin and 1/200 TESA) were incubated for one hour. After new washings, adequate dilution of peroxidase conjugate were added for another hour, washed and bound conjugate revealed by 1 h with orto-phenylenediamine and hydrogen peroxide. After 30 min in 37°C, reaction was stopped with 4N HCl and 492 nm absorbance determined in a microplate reader (Multiskan-Titertek II). Statistical analysis All quantitative data, such as O.D. ELISA, were analyzed using ANOVA after the Levene test for variance check, with intragroup comparisons by the Bonferroni's test, if there are uniformity of variances. In the absence of this homogeneity, data were analyzed by Kruskal-Wallis tests with Dunns post-tests. We opt for graphical representation of individual data in dot plot with association of mean and SEM for comparison. Qualitative analysis, as frequency of positive sera in the group, was analyzed by Fisher exact tests in two group analysis. We also included 95% confidence interval of estimated proportion. Significant difference was considered when the probability of equality (H1 = H0) was less than 0.05(p≤0.05), using two-tailed analysis and power greater than 90%. We used the statistical package GraphPad Prism 7.0 for all statistical analysis and plotting. Results We analyzed the distribution of Vimentin in culture cells using indirect fluorescent assays as described in Methods, using as external controls anti- T.cruzi sera, derived from chronic infected patients for identification of the parasites in the same model, as could be seen in figure 1. LLC-MK2 cells, the host cell used for intracellular growth of T.cruzi , showed a discrete and uniform cytoplasmic staining, uniform in most cells (Figure 1A). Those cells are no reactive to anti- T-cruzi antibodies, without any staining (Figure 1B). After infection and growth of T.cruzi amastigotes, those cells express larger amounts of vimentin, with heavy staining of cytoplasm outside the parasitophorous vacuole and some particle shadowing patterns, suggesting that vimentin are associated with cell cytoplasm (Figure 1C). Vimentin could involve unstained cytoplasmic parasites, but no specific staining of parasites was seen. Those parasites were easily identified by anti- T.cruzi antibodies showing a typical morular pattern in the cytoplasm of infected cells (Figure 1D). No staining of those parasites was observed with anti-vimentin mAbs, which demonstrate the absence of antigen mimicry, both for amastigotes (Figure 1C) or extracellular parasites (Figure 1E). Those extracellular parasites are heavily stained by anti T.cruzi antibodies as well as intracellular amastigotes (Figure 1F). Anti-vimentin auto antibodies We search for anti-vimentin antibodies in human sera from controls or American trypanosomiasis patients. Our sample was composed by patients with well-defined clinical forms as described in Methods. Vimentin ELISA was prepared with commercial protein and antibody binding was revealed by commercial conjugates. Standardization was easily as controls were adequately negative, allowing an adequate cut-off definition. We also tested all samples in a reported high specificity ELISA assay, TESA, which uses an excreted secreted antigen in solid phase, with high reactivity in all clinical forms of American trypanosomiasis. Those assays could be seen in Figure 2. We clearly demonstrate that all patients from our sample of American trypanosomiasis react very well in TESA assay, without any false positive or dubious sample in control groups. All clinical forms presented a similar reactivity for parasite antigens, including those with asymptomatic infection. Anti-vimentin antibodies were present in most American trypanosomiasis samples, but notably, they are much more present in acute or clinical defined syndromes (76,9%), especially cardiac disease (87,9%) and digestive form 70,5%. Paradoxically, they were relatively infrequent in asymptomatic infected patients (25%), which had a clearly positive serological reaction to parasite antigens, but had low frequency of anti-vimentin antibodies, similar to controls (2,5%), p>0.05, Bonferroni’s ANOVA post-test) or frequency of positive samples (Table 1) with similar conclusions. The main reactivity of these autoantibodies appears to be more intense in active cardiac or acute disease, which were associated to large parasite burden and inflammatory response than digestive or asymptomatic disease. The proportion of reagent sera was also shown in Table 1 assuming cutoff value estimated as defined in methods. The anti-vimentin ELISA reactivity of 349