Subjects with discordant or negative serology for T. cruzi infection but epidemiological findings for Chagas disease were recruited at the Chagas disease Unit of Hospital Interzonal de Agudos Eva Perón in Buenos Aires, Argentina. The protocol was approved by the IRB of the Hospital Interzonal de Agudos Eva Peron (Protocol No. 40/14 and 14-0004). Signed informed consent was obtained from all individuals prior to inclusion in the study following the guidelines of the United States Department of Health and Human Services. T. cruzi infection was determined by indirect immunofluorescence assay, hemagglutination, and ELISA techniques (12), by using an in-house epimastigote whole homogenate antigen, at the Instituto Nacional de Parasitología Dr. Mario Fatala Chabén, Argentina. Patients positive in at least two of these tests were considered to be infected, while subjects with positive results in one out of three tests, repeated at least twice, were considered as having discordant serology. Physical evaluation, electrocardiography and echocardiography were performed in all subjects included in the study. Twenty-one subjects with discordant serology, and 40 SP individuals, all born in areas endemic for T. cruzi infection (i.e., five subjects in Paraguay, and the remaining subjects in different endemic provinces of Argentina), but living in Buenos Aires by the time of the study were recruited for analysis. These individuals did not display clinical signs of Leishmania infection, which is endemic in some provinces of Argentina, and the amnanesis questionnaire did not revealed that SD subjects had been infected with Leishmania. Twelve subjects born and raised in Buenos Aires where T. cruzi infection is not endemic were recruited as uninfected control subjects not-exposed to T. cruzi (Table 1). Seven patients with Tegumentary Leishmaniasis infection or leishmaniasis-T. cruzi coinfection born and living in Salta, an argentinean province endemic for leishmaniasis, were also included as controls. Tegumentary Leishmaniasis is the prevalent type of Leishmania disease in the provinces in which SD and SP subjects for T. cruzi infection were born. Individuals with hypertension, ischemic heart disease, cancer, HIV infection, syphilis, diabetes, arthritis, or serious allergies were excluded from this study.

Approximately 50 mL of blood was drawn by venipuncture into heparinized tubes (Vacutainer; BD Biosciences). PBMC were isolated by density gradient centrifugation on Ficoll-Hypaque (Amersham) and stored in liquid nitrogen at 1.5 × 10⁷ million PBMC/mL in new-born bovine serum containing 10% DMSO. On the day of the assay, PBMC were thawed and washed in RPMI media containing 10% new-born bovine serum, 100 units/mL penicillin, 0.1 mg/mL streptomycin, 2 mM l-glutamine, and 10 mM Hepes buffer. Viability of cells was checked by trypan blue staining with a viability range of 80–90%. Another 10 mL of blood was allowed to coagulate at room temperature and centrifuged at 1,000 g for 15 min for serum separation. Due to sample availability the assays were not run for all the samples.

Protein lysate from T. cruzi amastigotes derived from the Brazil strain was obtained by four freeze/thaw cycles followed by sonication, as previously reported (23). Recombinant His-tagged T. cruzi-derived proteins were expressed in Escherichia coli and purified under denaturing conditions using TALON Metal Affinity Resins, as previously described (24), including glycosomal phosphoenolpyruvatecarboxykinase (ATP) (TcCLB.508441.20) (KN104), putative 60S acidic ribosomal protein P0 (TcCLB.508355.250) (KN107), paraflagellar rod protein 2 (TcCLB.511215.119) (KN117), flagellar calcium-binding 24 kDa protein (TcCLB.507491.151) (KN122), putative surface protein TolT (TcCLB.510433.20) (ANOF), and conserved hypothetical protein (TcCLB.506635.130) (ANOH). Leishmania antigen from Montenegro skin test was kindly provided by the Instituto Nacional de Parasitología Dr. Mario Fatala Chabén, Argentina. A Leishmania lysate comprised by a mix of promastigotes derived from the predominant spp. found in the Northwest of Argentina, Leishmania braziliensis and Leishmania amazonensis was obtained as a second source of Leishmania antigen. Promastigotes of L. braziliensis and L. amazonensis were separately processed. Briefly, parasites were washed by centrifugation with phosphate-buffered saline at 4,500 rpm for 10 min. Deionized sterile water was added and five freeze/thaw cycles were performed followed by centrifugation at 4,500 rpm. Supernatant was stored at −80°C until use. Protein content was measured by Bradford method with a final protein concentration of 0.8 mg/mL.

The number of T. cruzi antigen-responsive IFN-γ- and IL-2-secreting T-cells was determined by ex vivo ELISPOT assays, using a commercial kit (ELISPOT Human IFN-γ or IL-2 ELISPOT Set; BD Biosciences, USA), as described elsewhere (23, 25). PBMC were stimulated with 10 µg/mL of a T. cruzi-derived lysate preparation, 10 µg/mL T. cruzi-derived recombinant proteins, 10 µg/mL Montenegro Leishmania antigen, 20 ng/mL Phorbol 12-myristate 13-acetate (Sigma) plus 500 ng/mL ionomycin (Sigma), or media alone for 16–20 hr at 37°C and 5% CO₂. Spot forming cells (SFCs) were automatically enumerated using ImmunoSpot analyzer (CTL). The mean number of spots in triplicate wells was obtained for each condition, and the number of specific IFN-γ and IL-2-secreting T-cells was calculated by subtracting the value of the wells containing media alone from the antigen-stimulated spot count. Responses were considered positive if (1) a minimum of 10 SFC/4 × 10⁵ PBMC were present per well and (2) this number was at least twice the value of wells with media alone (26).

The 4 × 10⁶ PBMC/well were stimulated with 15 µg/mL T. cruzi amastigote lysate or media alone plus 1 µg/mL CD28/CD49d (BD Biosciences) at 37°C in a CO₂ incubator for 16–20 h. Stimulation with Staphylococcal enterotoxin B (1 µg/mL; Sigma Aldrich) served as a positive control. Brefeldin A was added for the last 5 h of incubation and then the cells were harvested and stained with antihuman CD4 allophycocyanin (APC-Cy7, clone RPA-T4) and antihuman CD3 Pacific Blue (clone UCHT1), both from BD Biosciences. After washing, cells were fixed and permeabilized with Cytofix/Cytoperm solution (BD Biosciences) and incubated with anti-human CD154-fluoresceinisothiocyanate (FITC, clone TRAP1), IFN-γ phyco-erythrin (PE, clone 4S.B3), tumor necrosis factor (TNF)-α allophycocyanin (APC, clone 6401.1111), macrophage inflammatory protein (MIP)-1β peridinin chlorophyll (PerCP Cy5, clone D21-1351), and IL-2 (PECy7, MQ1-17H12) (BD Biosciences). Approximately 600,000 events were acquired per sample in a HyperCyAn flow cytometer (Beckman Coulter, Brea, CA, USA). Analysis was performed with FlowJo software (Tree Star Inc., Ashland, OR, USA). Lymphocytes were gated based on forward scattering (FSC) and side scattering (SSC), and CD3⁺ cells were then analyzed for CD4 vs. each marker. Cytokine coexpression profiles were determined using the Boolean gating function of FlowJo software (see Figure S1 in Supplementary Material). A total of 31 different T-cell populations were obtained from the combination of the five different T-cell functions analyzed. Background responses detected in negative control samples with media alone were subtracted from those detected in T. cruzi-stimulated samples for every specific functional combination. Negative cutoff values for each specific functional combination were calculated as the average T. cruzi-specific T-cell response plus three standard deviations in two seronegative subjects. T-cell responses for each functional combination were considered positive when the T-cell response was higher than the cutoff value and at least twice the response with media alone; at least five events were present for single responses or at least three events for polyfunctional responses (27). For analysis only individuals with positive T-cell responses to T. cruzi antigens for at least one coexpression profile were included when calculating the proportion of each cytokine-producing subset contributing to the total T. cruzi-specific response. The mean fluorescence intensity (MFI) of each cytokine for different cytokine coexpression subsets was analyzed to determine the amount of cytokines produced in response to T. cruzi.

The 1 × 10⁶ PBMC were stained with a panel of antihuman-CD4 (PerCP) or anti-CD8 (PerCP), anti-CD45RA (FITC), anti-CD127 Alexa Fluor 647 and anti-CD132 (PE); or a panel of anti-CD4 (PerCP) or anti-CD8 (FITC) and anti-HLADR (PE) for 30 min on ice. Data were acquired on a FACS Calibur flow cytometer (Becton Dickinson, USA) and analyzed with FlowJo software (Tree Star, San Carlos, CA, USA).

Serum specimens were screened for antibodies reactive to a panel of 11–14 recombinant T. cruzi proteins in a Luminex-based format, as previously described (20, 24). The ratio of the specific MFI for each antigen to the MFI of the negative control protein was calculated for each serum. Values above de mean plus four standard deviations of a minimum of 16 true negative sera run in the same assay, and individually determined for each antigen were considered a positive test.

Demographic and clinical characteristics of T. cruzi-infected subjects included in this study were summarized using range and median. The normality of data was evaluated by the Shapiro–Wilk test. Differences among groups were evaluated by ANOVA followed by a Bonferroni test for multiple comparisons or by a test for linear trend. Comparisons of the frequencies of responders in the different categories of ELISPOT responses between groups were evaluated by use of the chi² test or Fisher’s exact test, as appropriate. The Mann–Whitney U test or Student’s t-test was applied to compare the number of spot counts, the number of functions and the MFI of each cytokine between SD and SP individuals. Differences were considered statistically significant at P < 0.05.

Serodiscordant subjects had positive results for T. cruzi-specific antibodies by immunofluorescence (i.e., 15 out 20 SD subjects) or hemagglutination (i.e., 6 out of 21 SD subjects) but negative findings by ELISA (Table 2). Two SD subjects showed some degree of electrocardiographic alterations but had normal left ventricular diameters and normal global systolic function, indicating absence of heart failure (Table 1). The remaining subjects presented no clinical or electrocardiographic abnormalities (Table 1).

We have previously shown that adult subjects with chronic T. cruzi infection have IFN-γ-producing T-cells responsive to T. cruzi antigens whereas IL-2 secreting cells are rare in Chagas disease patients. In contrast, uninfected non-exposed subjects have no T-cell responses specific for T. cruzi (23, 25, 28, 29). Herein, T. cruzi-specific T-cell responses measured by direct ex vivo cytokine production in response to a T. cruzi lysate were evaluated in 19 subjects with discordant findings in conventional serological tests. Most subjects with discordant serology had T-cells that showed simultaneous secretion of both IFN-γ and IL-2 in response to T. cruzi (Tables 2 and 3), contrasting with the low prevalence of IL-2-producing T-cells in SP individuals (Table 3). The number of IFN-γ spot counts among responder subjects by ELISPOT was significantly higher in SD subjects than in SP subjects (Figure 1). Although not statistically significant, a trend to higher IL-2 spot counts was also observed in SD subjects compared with SP individuals.

To confirm the specificity of T-cell responses in SD subjects, the frequency of IFN-γ-producing T-cells in response to a set of T. cruzi-derived recombinant proteins included in a multiplex assay developed by our group was also evaluated. These proteins were selected for being high-abundant in the amastigote and trypomastigote stages or being unique for T. cruzi (24). Eight out of nine (88.9%) SD subjects showed positive responses to 3–5 proteins assessed, while IFN-γ-producing T-cells were present in 11 out of 15 (73.3%) SP subjects in response to 1–5 proteins. Positive responses to T. cruzi-derived recombinant proteins were not observed in the 10 seronegative subjects not exposed to T. cruzi evaluated (data not shown). SD subjects had higher levels of IFN-γ producing T-cells responsive to the KN 104 recombinant protein compared with SP individuals (Figure 2). In addition, higher frequency of IFN-γ-producing T-cells specific for KN122, ANOF, and ANOH was observed in SD subjects compared with SP subjects, although the differences were not statistically significant (Figure 2). Of note, seven SD subjects showed responses to ANOF, a protein unique to T. cruzi. T cells in SD and SP subjects did not recognize Leishmania antigens (Figure 2). In addition, T cells in patients with Tegumentary leishmaniasis did not recognize T. cruzi antigens, while patients with Leishmania-T. cruzi coinfection presented positive T-cell responses to the T. cruzi lysate further corroborating that T-cell responses in SD subjects are specific for T. cruzi (Table 4). Thus, a subset of SD subjects exhibit robust T cells responsive to T. cruzi suggesting prior exposure to the parasite.

To further assess the functional capacity of T. cruzi-specific T-cells in SD subjects in comparison to SP subjects, multiparameter flow cytometry was performed to measure five T-cell functions, IFN-γ, IL-2, TNF-α, MIP-1β, and CD154, independently and simultaneously in these groups. The analysis was focused on CD4⁺ T-cells since this is the main T-cell population induced by the amastigote lysate preparation (23, 25). We first analyzed the contribution of each subset to the total CD4⁺ T-cell response in subjects with positive responses by intracellular staining. Although no statistically significant, SD subjects displayed an increased proportion of monofunctional CD4⁺ T-cells compared to SP subjects (Figures 3A,B). T. cruzi-specific monofunctional CD4⁺ T-cell responses from SD subjects are predominantly comprised of IFN-γ, CD154 and MIP-1β-producing cells, whereas SP subjects had primarily IFN-γ-producing cells (Figures 3A,B).

The composition of polyfunctional T cells in SD subjects was also different from that in SP individuals, with an enrichment in IL-2-producing T-cells (i.e., 52 and 21% of the total CD4⁺ T-cells producing cytokines consisted of IL-2-producing cells, respectively, black arch in Figures 3A,B). This is particularly noticeable in CD4⁺ T-cells exhibiting three functions (i.e., green arch in Figures 3A,B). Moreover, the magnitude of IL-2 production by CD4⁺ T-cells in SD subjects, as determined by the MFI, was higher than those in SP subjects (Figures 4A–E). SD subjects also showed a higher production of IFN-γ and MIP-1β among CD4⁺ T-cells expressing one (Figure 4E) or two (Figure 4D) functions compared with SP individuals.

We have previously demonstrated that T. cruzi-infected subjects have decreased frequencies of T cells expressing the IL-7R components CD127⁺CD132⁺ and a reciprocal gain of CD127^-CD132⁺ in CD8⁺ and CD4⁺ T-cells when compared with uninfected subjects, supporting a process of persistent antigen stimulation (30). T. cruzi-infected subjects also exhibit an increased number of CD4⁺ and CD8⁺ T-cells expressing the activation marker HLA-DR (31). To explore whether SD subjects had a T-cell profile more similar to actively infected or to uninfected subjects, the expression of molecules denoting chronic immune activation and antigen exposure was evaluated in these groups.

The frequencies of CD127⁺ CD132⁺ and CD127^-CD132⁺ cells in the CD4⁺ (Figure 5A) and CD8⁺ (Figure 5B) T-cell compartments in SD subjects were similar to those found in uninfected non-exposed subjects. The expression of HLA-DR in CD4⁺ and CD8⁺ T-cells in SD subjects was also comparable to that in uninfected controls and distinct from SP individuals (Figure 5C). These findings indicate that T cells in SD individuals have a resting phenotype different from that of SP individuals with active infection.

Humoral immune responses were also assessed by multiplex assay (24), which allowed the detection of antibodies specific for a set of T. cruzi-derived recombinant proteins. The majority of SD subjects (i.e., 14/19 subjects assayed) had low antibody responses specific for T. cruzi-derived recombinant proteins compared with SP subjects, in terms of both the number of proteins recognized and the strength of reaction (Table 2, Figures 6A,B) (24). SD subjects had positive reactivity to 1.5 proteins in average, compared with more than four proteins in SP subjects detected herein (Figure 6B) and in former studies (24, 31, 32).

T-cell and antibody responses against T. cruzi antigens in subjects with discordant serology for T. cruzi infection were sustained over time, as observed by the monitoring of five subjects in which these responses were evaluated for 3–8 years (Table 2). However, discordant serological findings by conventional serology might fluctuate between completely negative findings (i.e., three negative tests out of three performed) to positive testing by immunofluorescence or hemagglutination but not by ELISA assay (i.e., two positive tests out of three performed indicated in bold, Table 2). Likewise, T. cruzi-specific antibody levels also oscillated in a low range of responses by the multiplex technique (Table 2; Figure 6A). SD subjects did not show electrocardiographic changes during the follow-up period.