Introduction
Chagas disease (CD) is a zoonosis caused by the flagellated protozoan Trypanosoma cruzi, which affects about 6–7 million people worldwide [1]. WHO considers it one of the neglected tropical diseases; it is highly endemic in many Latin American countries [Reference Hotez, Dumonteil, Woc-Colburn, Serpa, Bezek, Edwards, Hallmark, Musselwhite, Flink and Bottazzi2, 3]. It has also been observed in non-endemic regions, such as the USA, Europe, Oceania and Asia, due to migratory movements [Reference Rodríguez-Guardado, Rodríguez, Alonso, Seco, Flores-Chavez, Mejuto and Cartón4–Reference Rassi and Marin-Neto6]. The extensive process of rural exodus in Latin America accentuates exposure to different comorbidities, and overlap between CD and HIV increases the number of co-infections [Reference de Almeida, Ramos Júnior, Correia and Shikanai-Yasuda7, Reference Ramos Júnior, Correia, Almeida and Shikanai-Yasuda8]. CD reactivation has been observed in approximately 20% of those co-infected [Reference Ramos Júnior, Correia, Almeida and Shikanai-Yasuda8–Reference Sartori, Lopes, Caramelli, Duarte, Pinto, Neto and Amato Shikanai-Yasuda11] as an opportunistic infection, increasing mortality rate. In endemic areas of Latin America, the rate of CD/HIV co-infection is estimated to be between 1.3% and 27.6% [Reference Almeida, Lima, Lages-Silva, Guariento, Aoki, Torres-Morales and Pedro12–Reference Dolcini, Ambrosioni, Andreani, Pando, Martínez Peralta and Benetucci14]. In non-endemic areas, prevalence varies from 0.0% to 10.5% [Reference Rodríguez-Guardado, Rodríguez, Alonso, Seco, Flores-Chavez, Mejuto and Cartón4, Reference Salvador, Treviño, Sulleiro, Pou, Sánchez-Montalvá, Cabezos, Soriano, Serre, Gómez, Prat, Pahissa and Molina5]. In Brazil, the prevalence of CD/HIV co-infection is between 1.3% and 5% [Reference Almeida, Lima, Lages-Silva, Guariento, Aoki, Torres-Morales and Pedro12, Reference Stauffert, Silveira, Mesenburg, Manta, Dutra, Bicca and Villela15]. CD has different clinical manifestations ranging from the acute form to asymptomatic chronic phase. Cardiomyopathy, mega colon and mega oesophagus are the most frequent clinical manifestations of the chronic disease, making up around 30% [Reference Rassi and Marin-Neto6]. New aspects of CD immunopathology with unusual clinical manifestations, such as meningoencephalitis and marked cardiac damage, have recently occurred as a reactivation in people living with HIV/AIDS (PLWHA), mainly in those presenting severe immunosuppression [Reference Sartori, Ibrahim, Nunes Westphalen, Braz, Oliveira, Gakiya, Lopes and Shikanai-Yasuda9, Reference Sartori, Shikanai-Yasuda, Amato Neto and Lopes10, 16]. In 2008, Brazilian guidelines on CD recommended screening for Trigonoscuta cruzi infection be carried out for all PLWHA, especially for those from endemic areas [16] due to the risk of developing CD reactivation. In Brazil, serological tests for CD screening are based on the total T. cruzi antigen; these may have cross-reactivity with other diseases, including endemic diseases such as leishmaniasis, which is caused by pathogens similar to T. cruzi, thus leading to false-positive results [Reference Umezawa, Nascimento, Kesper, Coura, Borges-Pereira, Junqueira and Camargo17–Reference Caballero, Sousa, Marques, Saez-Alquezar and Umezawa19]. For this reason, tests possessing high sensitivity and specificity are required for serological screening of CD/HIV co-infection. Immunoblotting using trypomastigote excreted-secreted antigen (TESA Blot) is an excellent test for avoiding cross-reactivity and false-positive results. TESA Blot has high sensitivity and specificity in diagnosing acute and chronic CD as it does not present cross-reactivity with Leishmania spp.-infected patients [Reference Umezawa, Nascimento, Kesper, Coura, Borges-Pereira, Junqueira and Camargo17]. Many cases involving CD reactivation due to HIV co-infection have been reported [Reference Sartori, Ibrahim, Nunes Westphalen, Braz, Oliveira, Gakiya, Lopes and Shikanai-Yasuda9–Reference Sartori, Lopes, Caramelli, Duarte, Pinto, Neto and Amato Shikanai-Yasuda11]; however, little is effectively known about the prevalence of co-infection in patients without reactivation. Detecting early co-infection using accurate diagnostic methodologies could allow the monitoring of these patients and prevent more severe and lethal clinical conditions. In this study, we evaluate the prevalence of T. cruzi infection in PLWHA using epimastigote alkaline extract antigen from T. cruzi (ELISA EAE) and TESA Blot to avoid false positives and determine the real prevalence of T. cruzi infection in this population.
Methods
Study design
This is a descriptive cross-sectional study that used ELISA EAE and TESA Blot to evaluate the presence of specific antibodies against T. cruzi in 240 serum samples from PLWHA at the Emilio Ribas Institute of Infectology (IIER), São Paulo, Brazil. HIV infection was confirmed according to a flowchart from the Brazilian Ministry of Health [16] using two different serological methods.
Subject and study design
From a population of 8,500 PLWHA, we included 240 HIV-infected patients from IIER between April 2015 and March 2016, considering a 16% prevalence of Leishmania infection. All patient information was collected from their medical records and the IIER’s electronic system. The following information was collected: (1) antiretroviral therapy (ART), (2) CD4+ T cell count, and (3) HIV-1 viral load values. Twenty serum samples, 10 from healthy individuals and 10 from patients with chronic CD, were used as controls.
Serology from CD
Epimastigote alkaline extract
T. cruzi epimastigote Y strain extract was prepared using fresh parasites cultivated in a liver infusion tryptose (LIT) medium as described previously [Reference Umezawa, Shikanai-Yasuda and Stolf20]. Briefly, 500 mg of epimastigotes was solubilized in 0.3 N NaOH for 18 h at 4°C, then neutralized (pH 7–8) with 0.3 N HCl and centrifuged at 12,000 g for 1 min at 4°C. The supernatant was collected and protein concentration measured using the Macro-BCA protein assay reagent kit (Pierce Co) and stored at −80°C.
Trypomastigote excreted-secreted antigen
TESA from the T. cruzi Y strain was obtained as previously described [Reference Umezawa, Nascimento, Kesper, Coura, Borges-Pereira, Junqueira and Camargo17]. Briefly, the supernatants of LLC-MK2 cell cultures containing 2% fetal calf serum infected with T. cruzi were collected when trypomastigote concentration reached about 10 × 106/mL. After centrifuging at 1800 g for 15 min at 4°C, the supernatant containing TESA was then submitted to a second centrifugation (7000 g for 5 min at 4°C) and used directly without any further treatment, or stored in small aliquots at −80°C.
In-house ELISA EAE
In-house ELISA EAE was performed according to Umezawa et al. [Reference Umezawa, Shikanai-Yasuda and Stolf20]. High-binding polystyrene Costar plates (Corning, USA) were coated with 50 μL EAE (4 μg/mL) in 0.05 M carbonate-bicarbonate buffer, pH 9.6, for 18 h at 4°C. The plates were blocked for 1 h with phosphate‐buffered saline‐Tween 20 (0.05%) containing 5% fat-free milk (Nestlé®). The plates were subsequently incubated with 50 μL diluted sera (1:200) for 1 h at 37°C, then washed and incubated with IgG anti-human peroxidase conjugate (Sigma) for 1 h at 37°C. After a new wash cycle, hydrogen peroxide and O-phenylenediamine dihydrochloride (OPD-tablets; Sigma Co.) were added to each well. The plates were incubated for 30 min at 37 °C in the dark, and the reaction stopped by adding 25 μL 4 N HCl. Absorbance at 492 nm was measured using an ELISA plate reader (Labsystems Multiskan MS).
TESA Blot
In-house TESA Blot was performed according to Umezawa et al. [Reference Umezawa, Nascimento, Kesper, Coura, Borges-Pereira, Junqueira and Camargo17]. Antigenic proteins from TESA were separated by SDS-PAGE, transferred to nitrocellulose sheets, and blocked with PBS containing 5% fat-free milk for 1 h at room temperature. Membrane strips (5 mm) were incubated with serum (1:200), diluted in PBS with 1% milk for 2 h or overnight at room temperature and washed, then bound antibodies were detected with horseradish peroxidase-labelled anti-dog IgG (Sigma Co.). Hydrogen peroxide and 4-chloro-1-naphthol were added for colorimetric detection of the bands. Samples were considered positive when a large 150–160 kDa band and/or five bands between 130 and 200 kDa were observed [Reference Umezawa, Nascimento, Kesper, Coura, Borges-Pereira, Junqueira and Camargo17].
Statistical analysis
A database was generated in Microsoft Excel 2013®. Analysis and interpretations of results were done using Microsoft Excel 2013® and Prism™ version 5.0 (GraphPad Software, Inc.). Graphs of individual (Abs492nm) distribution for each serum were obtained, and a cut-off value was determined from the receiver operating characteristics (ROC) curve using Prism™. For this, we used 20 samples from healthy donors and 10 samples from chronic CD patients (from the biorepository, Laboratory of Protozoology, Institute of Tropical Medicine, Faculty of Medicine, São Paulo University). Samples were considered positive when they presented an Abs492nm >0.38 (cut-off).
Ethics
The use of samples collected for the project ‘Prevalence of infection by Leishmania spp. in HIV/AIDS patients living in an urban area’ was approved by the ethics and research committee of the IIER (CAAE 75757417.5.0000.0061). Informed consent form signed by participants of the above project mentions that their sera can be used in other research projects.
Results
Demographic data
All patients had HIV infection confirmed according to the Brazilian Ministry of Health flowchart. Of the 240 subjects, 172 (71.6%) were male and 68 (28.3%) female; 211 (87.9%) were under antiretroviral therapy (ART); 213 (88.8%) presented CD4+ T cell count >200 cells/mm3 and only 27 (11.2%) <200 cells/mm3; 195 patients (81.3%) presented undetectable viral load; 24 (10%) presented viral loads between 41 and 10,000; 19 (7.9%) presented viral loads between 10,001 and 100,000; and 2 (0.8%) presented viral loads >1,000,000 (Table 1). In our sample, 27.8% presented criteria for AIDS according to the presence of an opportunistic infection and/or a CD4+ T cell count <200 cells/mm3.
Table 1. Data from 240 HIV-infected patients according to ART, CD4+ T cell counts and viral load
Serology
Of the 240 HIV-positive patients evaluated by ELISA EAE, 5 (2.1%) were positive, with absorbances (492 nm) of 0.59, 0.77, 0.86, 2.28 and 2.68, all above the cut-off value (Table 2, Figure 1a). In order to confirm these results, we evaluated these five positive patients using TESA Blot. Two of the five patients had CD confirmed by TESA Blot (0.83%) according to 150–160 kDa protein band reactivity (Table 2, Figure 1b). These patients presented the highest absorbance (492mn) values of 2.28 and 2.68 by ELISA EAE. They came from municipalities close to the city of Montes Claros in the northern region of Minas Gerais where, historically, CD has played an important endemic role. One patient presented cardiac symptoms, and the other was asymptomatic to CD. Comparing infected patients (CD infection confirmed by ELISA EAE and TESA Blot) with non-infected patients, we observed mean CD4+ T cell counts of 588 and 652, respectively. Mean viral load values were 37,698.50 and 33,161.15, respectively. Two hundred and ten T. cruzi-non-infected (87.5%) and only 1T. cruzi-infected patient (50%) were under ART.
Table 2. Evaluation by ELISA (EAE) of positivity for CD in 240 HIV-infected patients from IIER
Figure 1. Detection of antibodies against T. cruzi by ELISA and TESA-Blot. (A) Reactivity by ELISA-EAE IgG (Abs 492nm) from 240 serum samples from PLWHA at IIER; 10 samples from normal individuals (N = negative control) and 10 samples from patients with chronic Chagas’ disease (CH = positive control). The horizontal line represents the cut-off value; Abs 492nm = 0.38). (B) Evaluation of IgG reactivity by TESA BLOT from five patients (lanes 1-5) with IgG reactivity to T. cruzi by ELISA-EAE. Lanes 1 and 2 positive patients; lanes 3-5, non-reactive patients; lane 6, Chagas disease control patient; lane 7, uninfected control. Reactivity to molecular weight bands 150-160 kDa (at the left of figure) indicate serum positivity.
EAE, epimastigote alkaline extract.
Discussion
In recent decades, the spread of HIV infection to rural areas and the movement of patients with CD to urbanized areas has increased the overlapping of these diseases, effectively contributing to an increased occurrence of CD/HIV co-infection [Reference de Almeida, Ramos Júnior, Correia and Shikanai-Yasuda7, Reference Ramos Júnior, Correia, Almeida and Shikanai-Yasuda8]. Despite the recommendation to screen for T. cruzi infection in all PLWHA from endemic areas, there is a lack of information on the real prevalence of CD/HIV co-infection. The aim of our study was to use serological tests to determine the prevalence of infection by T. cruzi in a population of PLWHA treated at IIER, São Paulo, Brazil. Our main sampling strategy for patients probably in the indeterminate/chronic phase was based on choosing different serological tests than those recommended for diagnosis (ELISA, immunoblotting) [Reference Umezawa, Shikanai-Yasuda and Stolf20, Reference Abras, Gállego, Llovet, Tebar, Herrero, Berenguer, Ballart, Martí and Muñoz21]. The (in-house) ELISA EAE test, initially used as a screening tool, had good sensitivity, but less specificity due to false-positive results due to cross-reactivity with leishmaniasis [Reference Vexenat, Santana and Teixeira18, Reference Umezawa, Nascimento and Stolf22]. Our results showed a 2.1% prevalence of co-infection when using ELISA EAE (Table 2); however, only two of these five positive results were confirmed by TESA Blot (Figure 1b). TESA Blot is based on immunoblotting with excreted and secreted molecules from T. cruzi trypomastigotes and has high sensitivity and specificity in acute and chronic CD diagnosis as it does not present cross-reactivity with leishmaniasis, an endemic disease in the same transmission area as CD [Reference Umezawa, Nascimento, Kesper, Coura, Borges-Pereira, Junqueira and Camargo17]. The three samples that were not confirmed by TESA Blot showed low reactivity by ELISA, close to the cut-off point, suggesting false-positive results probably due to cross-reactivity with Leishmania infection. These samples came from patients born in the state of São Paulo, which for decades has presented several municipalities and endemic regions with both CD and leishmaniasis. Cunha et al. [Reference Cunha, Celeste, Kesper, Fugimori, Lago, Ibanes, Ouki, Neto, Fonseca, Silva, Júnior and Lindoso23] using the same samples showed positivity for Leishmania infection in these same three ELISA-positive/TESA Blot-negative samples in our cohort. So, we believe only two samples were truly positive for CD, with a prevalence of 0.83%. The two ELISA-positive samples, which were confirmed by TESA Blot, came from municipalities where CD is endemic [Reference Almeida, Lima, Lages-Silva, Guariento, Aoki, Torres-Morales and Pedro12, Reference Martins-Melo, Ramos, Alencar and Heukelbach24].
The main advantage of ELISA in current research is its high sensitivity, which combined with other qualities such as practicality, analysis of multiple samples in the same reaction, quantification and automation, making it an excellent screening test for T. cruzi infection. In addition, ELISA EAE is an in-house test using total antigens from epimastigotes that seems to have a high sensitivity but low specificity due to cross-reactivity with leishmaniasis [Reference Umezawa, Shikanai-Yasuda and Stolf20, Reference Umezawa, Nascimento and Stolf22]. ELISA EAE associated with TESA Blot seems to substantially increase the reliability of our results, as test sensitivity is important in HIV-infected patients due to immunosuppression as well as excluding any possibility of cross-reactivity with leishmaniasis [Reference Umezawa, Nascimento, Kesper, Coura, Borges-Pereira, Junqueira and Camargo17]. TESA Blot has been already in use for confirmation of immunodiagnosis of CD [Reference Ferreira-Silva, Pereira, Rodrigues-Júnior, Meira, Basques, Langhi-Júnior, Romanelli, Umezawa, Késper-Júnior, Louzada-Neto, Bordin and Moraes-Souza25–Reference Araújo and Berne31]. Here, for the first time we used TESA Blot to measure the prevalence of T. cruzi/HIV co-infection.
In Brazil, previous studies using classic validated serological methods (ELISA, indirect hemagglutination and IFI) had established a co-infection rate between 1.3% and 5% [Reference Almeida, Lima, Lages-Silva, Guariento, Aoki, Torres-Morales and Pedro12, Reference Stauffert, Silveira, Mesenburg, Manta, Dutra, Bicca and Villela15]. In other regions, considering studies on vulnerable populations, co-infection prevalence can reach 7.8%, as was observed in a group of injecting drug users in Buenos Aires, Argentina [Reference Benchetrit, Andreani, Avila, Rossi, de Rissio, Weissenbacher, Martinez Peralta and Dolcini32]. The prevalence of CD/HIV co-infection can vary according to geographical area, due to high or low CD prevalence. Additionally, the accuracy of the test used to confirm T. cruzi infection can be another factor related to high or low prevalence. Clearly, if the study were to be performed in areas with high CD prevalence or with a vulnerable population, the co-infection frequency would have been higher, as was observed in studies from Bolivia and Argentina [Reference Reimer-McAtee, Mejia, Clark, Terle, Pajuelo, Cabeza, Lora, Valencia, Castro, Lozano, Bern, Torrico and Gilman13, Reference Dolcini, Ambrosioni, Andreani, Pando, Martínez Peralta and Benetucci14]. Our data show a lower prevalence of T. cruzi/HIV co-infection. Possibly, the variation in patients’ region of origin (18 different states) does not fully represent the regional reality of co-infection, reflecting the difficulty in establishing areas whose endemicities may constitute risk and pointing to the need for regionalized prevalence studies in more outpatient units. However, the vast majority of our patients (57.5%) come from the state of São Paulo, where an extensive control of CD transmission by triatomine was in place between the 1940s and 1970s [Reference Silveira and Dias33]. Additionally, the use of TESA Blot excludes overestimated results by cross-reactivity with leishmaniasis. The higher prevalence of CD infection in other studies was possibly related to false positives due to cross-reactivity with other disorders such as leishmaniasis.
By evaluating the CD4+ T cell counts in the two truly CD/HIV co-infected patients, we observed a median of 588.5 cells/mm3 compared to a median of 568.4 cells/mm3 in non-infected patients. Almeida et al. [Reference Almeida, Lima, Lages-Silva, Guariento, Aoki, Torres-Morales and Pedro12] reported a lower median CD4+ T cell count, 294.1 cells/mm3, for co-infected compared to their uninfected group. Stauffert et al. [Reference Stauffert, Silveira, Mesenburg, Manta, Dutra, Bicca and Villela15] reported 60% of co-infected patients to have a CD4+ T cell count <350 cells/mm3. Low CD4+ T cell values, especially <200 cells/mm3, is an important predictor of the risk of CD reactivation in HIV patients. We did not have any reactivation, due to our patients presenting CD4+ T cell counts >200/mm3. However, we suggest maintaining vigilance in these co-infected patients, as lower CD4+ T cell counts may favour CD reactivation [Reference Reimer-McAtee, Mejia, Clark, Terle, Pajuelo, Cabeza, Lora, Valencia, Castro, Lozano, Bern, Torrico and Gilman13, Reference Dias, Ramos, Gontijo, Luquetti, Shikanai-Yasuda, Coura, Torres, Melo, Almeida, Oliveira, Silveira, Rezende, Pinto, Ferreira, Rassi, Fragata, Sousa, Correia, Jansen, Andrade, Britto, Pinto, Rassi, Campos, Abad-Franch, Santos, Chiari, Hasslocher-Moreno, Moreira, Marques, Silva, Marin-Neto, Galvão, Xavier, Valente, Carvalho, Cardoso, Silva, Costa, Vivaldini, Oliveira, Valente, Lima and Alves34]. This was reinforced by Shikanai-Yasuda et al. [Reference Shikanai-Yasuda, Mediano, Novaes, Sousa, Sartori, Santana, Correia, Castro, Severo, Hasslocher-Moreno, Fernandez, Salvador, Pinazo, Bolella, Furtado, Corti, Neves Pinto, Fica, Molina, Gascon, Viñas, Cortez-Escalante, Ramos and Almeida35] by showing that a low CD4+ T cell count is intrinsically related to reactivation and lethality in PLWHA and CD.
There is therefore a need for careful monitoring of immunological status and parasitaemia in co-infected individuals, with antiparasitic treatment indicated for asymptomatic co-infected patients with high parasitaemia. Regarding the viral load, a great tendency towards effective virological suppression can be observed in the sample, with 81.3% (195/240) of our cohort having a viral load <40 copies/mL (Table 1). During patient follow-up, increase in viral load is an important factor for CD reactivation, as is transient elevation during reactivation, showing a very harmful interaction with pathogens [Reference Dias, Ramos, Gontijo, Luquetti, Shikanai-Yasuda, Coura, Torres, Melo, Almeida, Oliveira, Silveira, Rezende, Pinto, Ferreira, Rassi, Fragata, Sousa, Correia, Jansen, Andrade, Britto, Pinto, Rassi, Campos, Abad-Franch, Santos, Chiari, Hasslocher-Moreno, Moreira, Marques, Silva, Marin-Neto, Galvão, Xavier, Valente, Carvalho, Cardoso, Silva, Costa, Vivaldini, Oliveira, Valente, Lima and Alves34, Reference de Freitas, da Silva, Sartori, Bezerra, Westphalen, Molina, Teixeira, Ibrahim and Shikanai-Yasuda36]. This virological control process highlights the undoubted importance of ART, seen in 87.9% (211/240) of patients in our study. This was much higher than in other studies with similar designs [Reference de Almeida, Ramos Júnior, Correia and Shikanai-Yasuda7, Reference Almeida, Lima, Lages-Silva, Guariento, Aoki, Torres-Morales and Pedro12]. Although essential for immune restoration, the clear role of ART in the natural history of T. cruzi/HIV co-infection has not been fully characterized in literature. It is believed that, once virological suppression and immune system recovery are achieved, CD should follow the natural course of chronic progression as would be observed in immunocompetent individuals. However, it appears that this relationship still lacks sufficient evidence, especially in prospective studies [Reference de Almeida, Ramos Júnior, Correia and Shikanai-Yasuda7].
Finally, our results indicate the need to apply diagnostic tests with high sensitivity and specificity that allow an assessment of the current status of CD/HIV co-infection in Brazil, considering clinical and laboratory characteristics of patients, in order to stratify the effective risk of reactivation. Our results present some limitations, such as low sample size and that the geographical area covered by the study could have underestimated the prevalence of CD in PLWHA. We strongly suggest studies using TESA Blot to confirm the prevalence of T. cruzi infection in PLWHA because it seems to be a more accurate test for confirming CD diagnosis.
Data availability statement
Data obtained are available in Excel spreadsheets under my responsibility.
Author contribution
Investigation: J.C.I.J., J.A.L.L., M.A.C.; Methodology: J.C.I.J., M.A.C., R.R., N.K.; Writing – original draft: J.C.I.J., N.K.; Supervision: J.A.L.L.; Writing – review & editing: J.A.L.L., N.K.; Data curation: M.A.C.
Open access
