This study was approved by the Research and Ethics Committee of the Federal University of Sergipe (UFS)-University Hospital, Aracaju, Sergipe State (SE), Brazil (CAAE 0162.0.107.000-09). The experiments were performed according to the ethical standards of the Declaration of Helsinki and followed the guidelines and regulations of the Brazilian National Council of Health resolution 196/96. All participants gave their written informed consent.

The study was performed in the Laboratory of Immunology and Molecular Biology of the UFS University Hospital, SE, Brazil, between January 2018 and December 2019. Blood samples were collected from 18 VL patients whose age ranged from 20 to 59 years (mean = 39 years). Patients were diagnosed based on the classical VL symptoms (fever, weight loss, spleen and liver enlargement, anemia, and pancytopenia). Clinical diagnosis was confirmed by serum reactivity to the rK39 antigen (KalazarDetect^®) Rapid Test, INBIOS International Inc., Seattle, WA) and by the growth of Leishmania parasites in cultures of blood or bone marrow samples, using NNN media (Sigma-Aldrich) (32). After diagnosis, patients were treated daily intravenously with meglumine antimoniate (MA; 20 mg/kg/day) (Glucantime^® Sanofi), for 21 days, or with liposomal amphotericin B (LAMB; 3 mg/kg/day) for 5–7 days (AmBisome, Gilead). Only patients displaying more severe symptoms of VL were eligible for treatment with LAMB (32, 33). Liver and spleen sizes were evaluated through palpation by two observers and reported as the distance to the rib border in cm at diagnosis (D0) and 7 (D7), 14 (D14), 21 (D21), 30 (D30), 60 (D60), 90 (D90), and 180 (D180) days after the beginning of treatment.

Blood samples were used for hemogram and leukogram assessments and for PBMC fractionation. Peripheral blood was collected before specific Leishmania treatment (day 0) and on days 7, 14, 21, 30, 60, 90, and 180 after beginning the chemotherapy treatment. We also collected blood from 13 healthy controls who live in the same VL endemic area and whose ages ranged from 20 to 60 years. Control individuals had no previous history of infectious diseases for at least a month before this study and were not under treatment with immunosuppressive drugs at the time of blood collection.

PBMC isolation was performed by Ficoll Paque^® (Sigma, USA) density gradient centrifugation on. Samples of 10⁶ PBMCs/well were cultured in Roswell Park Memorial Institute (RPMI) 1640 (Gibco^®, USA) medium supplemented with 10% fetal calf serum and 1% of penicillin/streptomycin. Cells were stimulated with soluble Leishmania antigen (SLA) of L. (L.) infantum chagasi (10 µg/ml) or with no addition and cultured at 37°C for 6 h in a humidified atmosphere with 5% CO₂, followed by the addition of Brefeldin at 1 µg/ml final concentration following the manufacturer’s instructions (GolgiPlug™ BD Biosciences Pharmingen, USA) and an additional 12-h incubation. After this incubation, cells were washed with phosphate-buffered saline (PBS), centrifuged at 405 g for 5 min at 4°C, and blocked with 2% fetal goat serum and 2% fetal bovine serum in PBS for 20 min (11).

The cells were treated with the following antibodies for surface staining: Anti-human CD3 PeCy7 (Clone HIT3a, BioLegend^®, USA), Anti-human CD4 V500 (Clone RPA-T4, BD Horizon™, USA), Anti-human CD8 Pe-Cy™5 (Clone HIT8a, BD Pharmingen™, USA), Anti-human CD45 APCCy7 (Clone HI100, BioLegend^®), and Anti-human CCR7 PE (Clone GO43H7, BioLegend^®) for 30 min at 2°C–8°C in the dark, after which they were fixed and permeabilized with Cytofix/Cytoperm buffer (BD Biosciences Pharmingen) for 20 min. Intracellular staining was carried out by incubating the cells with Rat Anti-human IL-2 BV421 (clone MQ1-17H12, BD Horizon™), Anti-human TNF-α Alexa Fluor 647 (Clone Mab11, BioLegend^®), and anti-IFN-γ Alexa Fluor 488 (Clone 4S.B3, BioLegend^®), for 30 min at 4°C in the dark. Then they were washed with Perm Wash buffer (BD Biosciences Pharmingen) and PBS. Finally, the cells were centrifuged and suspended in PBS. Samples were acquired in a FACSCanto II™ flow cytometer with at least 30,000 events and analyzed using FlowJo software version 7.0 (Becton Dickinson, San Jose, CA, USA).

Following the recommendations of Seder et al. (26) for multiparameter cytometry analysis, we used several hierarchical steps to identify the functional subpopulations of CD4 and CD8 T cells based on the expression of the following markers: CD3, CD4, CD8, CD45RA, CCR7, IL-2, TNF-α, and IFN-γ. First, we identified the single cells and then selected the lymphocyte cells among them ( Supplementary Figure 1 ). After that, we gated the CD3-positive cells and among them identified their CD4 and CD8 subsets. Additionally, from each CD4⁺ and CD8⁺ subset, we excluded the CCR7⁺ and CD45RA⁺ cells, which are the naïve T lymphocytes. Then only the effector and memory T cells remained, and so we evaluated the production of the pro-inflammatory cytokines IL-2, TNF-α, and IFN-γ by T cells. Boolean gating was used to generate combinations of cytokine expression and types of lymphocytes in order to identify lymphocytes expressing one cytokine and any combination of two cytokines or three cytokines ( Supplementary Figure 1 ).

In this study, the frequencies and the integrated mean fluorescence intensity (iMFI) of each type of cytokine-producing cells were analyzed. iMFI is a widely used parameter that is composed of the product of the frequency of a particular cytokine expressing T cell and the mean fluorescence intensity (MFI) of the same cell. Thus, this multiplication is the ideal parameter to measure the quality of cytokine production in relation to the number of producing cells.

The differences between the frequencies and iMFI cytokine secretion of CD4⁺ and CD8⁺ T cells of the controls and each time of treatment of all patients were compared by Mann–Whitney. Additionally, the Wilcoxon matched-pairs signed rank test was used to compare different times of treatment. We also used the IC95% for the specific comparison of total frequencies and iMFI of patients treated either with MA or LAMB. The two-way ANOVA Sidak test was used to compare the frequencies of the CD8^High and CD8^Low populations along the treatment, and Student’s t-test was used to compare each time of treatment with the frequencies of controls. We used GraphPad Prism 9 software, and differences were considered statistically significant for p < 0.05. Furthermore, we used Spearman’s two-tailed test for correlation analysis and plotted the results using the corrplot package (v0.84) running under R (v3.6.1) in RStudio (1.1.456).

Patients with VL were evaluated before treatment (D0) and on days 7 (D7), 14 (D14), 21 (D21), 30 (D30), 60 (D60), 90 (D90), and 180 (D180) after the beginning of the treatment. Treatments with LAMB and MA lasted until D7 and D21, respectively. All patients showed clinical improvement after treatment, as shown in Table 1 . We do not show the complete clinical results of the 18 patients because it was not possible to collect blood for hemogram at all times, from six to seven patients who had profound anemia and cachexia. Also, nine patients did not return for control at day 180. Typical VL symptoms, such as increases in spleen and liver sizes and decreased counts of hemoglobin, hematocrit, platelets, leukocytes, neutrophils, lymphocytes, eosinophils, and monocytes, were observed in patients at D0. The chemotherapy treatment was successful and resulted in a decrease in the sizes of the spleens and livers and increases in the values of all other variables. The sizes of the spleens and livers started to decrease and were significantly different from those of untreated patients on D7. Similarly, the platelet, leukocyte, and lymphocyte counts had increased by D7. On the other hand, the hemoglobin and hematocrit levels only started to increase on D14, the monocytes increased on D21, and the neutrophils and eosinophils only increased as of D30 after beginning of the treatment. All parameters, which were monitored along the time, recovered their normal values by D180, when patients were considered cured from VL ( Table 1 ). Our results indicate that the chemotherapy treatment, which lasted from 7 to 21 days, was very successful since improvements in clinical parameters were noted as early as at the end of the first week.

The total frequencies and iMFI of IL-2-, TNF-α-, or IFN-γ-producing CD4⁺ T cells were assessed in patients with VL along their treatment (D0–D180) and in healthy controls ( Figure 1 ). Stimulation of lymphocytes was performed with SLA. No significant variations among any treatment times or between treatment times and healthy controls were observed for IL-2 total frequencies or iMFI, suggesting that the basic production of IL-2 was preserved during treatment as in healthy controls ( Figure 1 ).

In contrast, production of TNF-α increased along the treatment towards the cure on D180, when the frequencies of total producing CD4 T cells returned to normal levels, whereas iMFI still remained high. In fact, a rapid and time-dependent increase of TNF-α total frequency was observed starting from D7, reaching its maximum on D90, and decreasing to normal levels on D180. The iMFI of CD4⁺TNF-α producing T cells was, in contrast, still high on D180 ( Figure 1 ).

However, the production of IFN-γ by CD4 T cells also increased alongside the healing process but started earlier than that of TNF-α ( Figure 1 ), on D7, and reached a peak on D21, when both MA and LAMB therapy had finished. The intensity of IFN-γ production (iMFI), however, was already noted on D0 ( Figure 1 ). In fact, the frequency of CD4⁺IFN-γ⁺ T cells on D21 was 4.7 times higher than on D0 (p = 0.016) and 16 times higher, if compared with healthy controls (p = 0.001) ( Figure 1 ).

In addition, we used the multiparameter flow cytometry strategy for the analysis of the T-cell responses. Through this analysis, we are able to distinguish the frequency of CD4 T cells secreting only one (CD4⁺IL-2⁺, CD4⁺TNF-α⁺, and CD4⁺IFN-γ⁺), two (CD4⁺IL-2⁺TNF-α⁺, CD4⁺TNF-α⁺IFN-γ⁺, and CD4⁺IL-2⁺IFN-γ⁺), or three cytokines (CD4⁺IL-2⁺TNF-α⁺IFN-γ⁺) simultaneously ( Figure 2 ).

As detected for total CD4⁺ cells producing IL-2 ( Figure 1 ), all combinations of cells producing IL-2 were stable along the time and did not differ from those of healthy controls ( Figure 2 ). In contrast, and as detected also for iMFI values and total frequencies of CD4 TNF-α-producing T cells ( Figure 1 ), significant increases of the frequencies of CD4⁺ T cells producing only TNF-α⁺ were observed on D60 after treatment in comparison with D0. Furthermore, higher percentages of CD4⁺TNF-α⁺IFN-γ⁺ T cells were detected at all treatment times, if compared with healthy controls ( Figure 2 ), and also there were higher percentages on D21 and D30, if compared with D0. The increase of the frequencies of CD4⁺TNF-α⁺IFN-γ⁺ T cells was impressive, and patients cured of VL showed a statistically significant enhancement of 78.9 times (p = 0.001) of this subpopulation if compared with healthy controls on D180 ( Figure 2 ). In fact, the frequencies of CD4⁺TNF-α⁺IFN-γ⁺ T cells increased progressively in a time-dependent manner towards the cure, on D180. On the other hand, the frequencies of CD4⁺ T cells secreting only IFN-γ increased on D7 and D21 after the beginning of treatment, similar to what was detected for total frequencies and iMFI ( Figure 1 ). However, they had decreased slightly by D60 ( Figure 2 ). No significant variation was detected in the frequencies of CD4⁺IL-2⁺IFN-γ⁺ ( Figure 2 ).

Finally, the frequencies of multifunctional CD4⁺IL-2⁺TNF-α⁺IFN-γ⁺ T cells increased significantly on D60, D90, and D180 as compared with those of controls, and on D90 as compared with those on D0, D30, and D21. Remarkably, on D180, the frequency of functional T cells was 5.7 times higher than that of healthy controls and 4.2 higher than in untreated patients ( Figure 2 ), suggesting that the increase in multifunctional CD4⁺ T cells is correlated with cure. Our data suggest that the total response of CD4⁺ T cells promotes an increase in the quality of the Th1 response during the treatment, along with an increase in functional T-cell phenotypes ( Figure 2 ).

The individual relative contribution of each cytokine-expressing phenotype population to the whole CD4⁺ T-cell response, for each treatment time, is summarized in the pie charts of Figure 2 . The relative proportions of multifunctional CD4⁺IL-2⁺TNF-α⁺IFN-γ⁺ and the CD4⁺TNF-α⁺IFN-γ⁺-producing T cells increased along the treatment towards the cure, while in the cells of the healthy controls, the CD4⁺ T single-producer cells of IL-2⁺, TNF-α⁺, or IFN-γ⁺ were predominant. These results show that the treatment of VL determines the differentiation of CD4 T cells expected in the development or recovery of the Th1 response. In fact, in treated patients, besides the single-cytokine producers, double-cytokine producers are also present in minor proportions. In fact, the relative proportions of CD4⁺IL-2⁺TNF-α⁺ and CD4⁺IL-2⁺IFN-γ⁺ T cells increased until D14 and then decreased as the treatment advanced.

The evolution of the relative proportions of each CD4⁺ cytokine-expressing subtype along the treatment times is represented in Supplementary Figure 2 . The relative proportions of cells producing only one cytokine decreased towards the end of treatment at D180, suggesting that the single producers converted into the double and triple producers, which reveals the recovery of the CD4 differentiation immune response. The frequencies of the double-cytokine producers CD4⁺IL-2⁺TNF-α⁺ remained high until D90, and those of the CD4⁺IL-2⁺IFN-γ⁺ cells were high on D7 and D14, but both decreased on D180. On the other hand, only the CD4⁺TNF-α⁺IFN-γ⁺ and the multifunctional T cells show a progressive time-dependent increase in frequencies along the treatment and reached their maximum values with the cure ( Supplementary Figure 2 ). Our results suggest that the acquisition of a long-term memory T-cell CD4 response in combination with a strong effector-memory T-cell response is associated with the cure of VL.

As previously described for CD4 T cells ( Figure 1 ), the frequencies and iMFI of total IL-2-producing CD8⁺ T cells did not differ from those of controls at any of the treatment times, thus suggesting background basal and sustained stimuli of IL-2 in clonal expansion. In contrast, the proportions of total TNF-α⁺-secreting cells were 14 times higher on D60 than on D21, and the frequencies of total IFN-γ-secreting CD8⁺ lymphocytes and iMFI had increased six to seven times by D30 in comparison with controls and D0. A decrease in brightness was observed on D60 ( Figure 3 ).

As observed for the CD4⁺ T cells ( Figures 1 and 2 ) and for the CD8⁺ total frequencies and iMFI ( Figure 3 ), the multiparameter analysis confirmed that the frequencies of the IL-2 single producers CD8⁺ T cells did not change along the time ( Figure 4 ), whereas the frequencies of single producers of TNF-α were 3.6 times higher on D60 than on D0. Furthermore, the frequency of single producers of IFN-γ was 14.6 times greater on D30 as compared with healthy controls.

Regarding the CD8⁺ T cells producing two cytokines, the frequencies of CD8⁺IL-2⁺TNF-α⁺ T cells had increased by 41.2 times by D60 compared with D0 ( Figure 4 ), and the frequencies of CD8⁺TNF-α⁺IFN-γ⁺ T cells had increased by 9.6 times on D60 and by 14.7 times on D90 as compared with controls. In contrast, and as detected for single producers of IFN-γ, an earlier rise of the proportions of the CD8⁺IL-2⁺IFN-γ⁺ T cells was observed with increases of 28.2 and 81.4 times, on D21 and D30, respectively, when compared with D0 ( Figure 4 ).

However, similar to what was described for CD4 T cells ( Figure 2 ), the proportions of the multifunctional CD8⁺IL-2⁺TNF⁺IFN-γ⁺ T cells increased significantly, if compared with controls, on D60 and D90, suggesting the involvement of memory T cells in the cure process ( Figure 4 ).

The pie chart shows that soon after the beginning of the treatment, on D14, the frequencies of double producers of TNF-α⁺IFN-γ⁺ and IL-2⁺TNF-α⁺ lymphocytes start to increase, indicating the expected evolution of the CD8 Leishmania-specific T-cell response. After cure, on D180, the proportion of multifunctional memory T cells is at a maximum. As predicted by Seder’s model (26), the multifunctional central-memory CD8⁺ T cells seem to interconvert with the TNF-α⁺IFN-γ⁺ effector-memory T cells during the treatment. Furthermore, as expected from Seder’s model, when cure is achieved by D180, both types of cells are present in high proportions, and there is also an increase in the frequencies of the single producers of IFN-γ, which are considered terminal effector CD8 T cells ( Figure 4 ).

An increase in the quality of the CD8⁺ T cells response is observed when there is a concomitant increase of frequencies of the central-memory (multifunctional CD8⁺IL-2⁺TNF-α⁺IFN-γ⁺), effector-memory T cells (CD8⁺TNF-α⁺IFN-γ⁺), and terminal effector CD8 T cells (CD8⁺IFN-γ⁺) ( Figure 4 ).

The evolution of the relative proportions of each CD8⁺ cytokine-expressing phenotype along the time of treatment is represented in Supplementary Figure 3 . Confirming the data of Figure 4 , healthy controls individuals always showed lower proportions of the single producers of the IFN-γ⁺ terminal effector, and double producers of the TNF-α and IFN-γ effector memory CD8 T cells compared with VL patients during the treatment. In agreement, except for those at D7, the multifunctional CD8⁺IL-2⁺TNF-α⁺IFN-γ⁺-TCM cells showed frequencies above those of healthy controls during the whole treatment and reached maximum values at the cure (D180) ( Supplementary Figure 3 ).

Two different subpopulations of CD8⁺ T cells were identified by the flow cytometry analysis, according to their brightness intensity. Cells showing a higher expression of CD8⁺ molecules were classified as CD8^High, while those showing a lower expression were considered as CD8^Low ( Figure 5A ). The differences between the two populations were highly significant starting from D30 (p < 0.0001) ( Figure 5B ). CD8^High frequencies were high in healthy controls and impressively low in untreated patients (D0 and D7). However, they progressively recovered to normal values starting from D21 and reached the highest values at the cure, on D180 after the beginning of the treatment ( Figure 5B ). In contrast, the frequencies of CD8^Low cells were higher in untreated patients than in healthy controls at D0, and these decreased along the treatment and did not recover to their normal values even at D180. In fact, the evolution of the frequencies of CD8^High and CD8^Low phenotypes was negatively correlated (R = −0.676; p < 0.0001). Our results indicate that the progress of chemotherapy treatment goes along with an increase of CD8^High and a decrease of CD8^Low frequencies that represent markers of the cure for VL.

In VL, an increase in the size of the spleen and decrease in hematocrit, hemoglobin, monocyte, and leukocyte counts are typical markers of this disease ( Table 1 ). We observed that during the treatment, the frequencies of CD8 single producers of TNF-α and of CD8^High correlated negatively with spleen sizes, while the frequencies of CD8^Low T cells correlated negatively with the leukocyte and monocyte counts, indicating that CD8 single producers of TNF-α and of CD8^High are strong correlates of cure, while CD8^Low is a marker of disease ( Table 2 ).

In addition, the total frequencies of CD4⁺TNF-α⁺ T cells, the frequencies of CD8 triple secretors of IL-2, TNF-α, and IFN-γ and of CD8^High T cells correlated positively with the increases of hematocrit and hemoglobin counts, while the frequency of CD8 single producer of TNF-α is associated only with the increase in hemoglobin counts ( Table 2 ). Furthermore, during the treatment, the increase of monocyte counts was positively correlated with the increases of the CD4⁺IL-2⁺, the CD4⁺IFN-γ⁺ total frequencies, and CD4⁺IFN-γ⁺ iMFI. Monocyte frequencies also correlated with the increases of the frequencies of the CD4⁺ T cells as single producers of IL-2, TNF-α, and IFN-γ; double producers of IL-2 and IFN-γ; and triple secretors of IL-2, TNF-α, and IFN-γ, and with the frequencies of CD8^High T cells ( Table 2 ). The frequencies of the CD8 single producers of TNF-α correlated positively with neutrophils, while CD8^High correlated with the counts of leukocytes, neutrophils, lymphocytes, and eosinophils ( Table 2 ). Thus, these variables indicate that the recovery of the T CD4 response can also be considered as a marker of the cure. In contrast, only the increases in the frequencies of CD8 single secretors of TNF-α or triple secretors are associated with a decrease of symptoms. Remarkably, the increases of CD8^High and the decreases of CD8^Low frequencies are definitively strong correlates of cure.

Spearman’s correlation matrix heatmap demonstrated the association of the profiles between the clinical outcomes, hematological counts, and the cytokine-producing CD4 and CD8 T cells, for each treatment time separately. We also showed the corrplots for D0, D7, D30, and D180 ( Figure 6 ). The correlation patterns between all these factors changed gradually during the treatment. For instance, on D180, when the patient was considered completely cured, both liver and spleen sizes were negatively associated (red dots) with the frequencies of CD4⁺IL-2⁺, CD4⁺TNF-α⁺, CD4⁺IFN-γ⁺, CD4⁺IL-2⁺TNF-α⁺, CD8⁺TNF-α⁺, CD8⁺IL-2⁺TNF-α⁺, and CD8⁺IL-2⁺TNF-α⁺IFN-γ⁺ T cells. In addition, the spleen sizes were also negatively associated with the frequencies of CD4⁺IL-2⁺TNF-α⁺IFN-γ⁺, and the liver sizes were negatively correlated to the proportions of CD4⁺TNF-α⁺IFN-γ⁺, CD8⁺IL-2⁺, and CD8⁺TNF-α⁺IFN-γ⁺. Notably, on D180, the CD4⁺TNF-α⁺IFN-γ⁺ T cells correlated strongly positively with leukocytes, lymphocytes, and monocytes and weakly positively with hemoglobin, platelets, and neutrophils. This analysis confirms that the complete development and differentiation of the CD4 and CD8 Leishmania-specific T cells is associated with the healing process and that these T-cell phenotypes are strong correlated with VL cure.

In VL, the immune response is directly affected by the drugs that act mainly to improve the clinical outcome. Therefore, we analyzed the chemotherapeutic effect of LAMB and of MA separately. Patients who exhibited more severe symptoms associated with age, comorbidities, unexpected hemophagocytosis, relapse, etc. were treated with LAMB, whereas less severe cases received MA.

The total frequencies and iMFI of cytokine producers of CD4⁺ T cells are represented in Figure 7 . Before treatment, on D0, the levels of iMFI-IL-2 and total frequencies and iMFI-TNF-α⁺ CD4⁺ T cells were higher in patients treated with LAMB ( Figures 7B–D ). In contrast, the total frequencies of CD4⁺IL-2 response were 12.4 and 4.5 times higher on D14 and D180, respectively; and the iMFI IL-2 response was 3.9 higher on D30 in patients treated with MA ( Figures 7A, B ). The TNF-α iMFI response remained higher on D7–D14 in patients treated with LAMB but increased towards the cure only in patients treated with MA ( Figure 7D ). Additionally, the total frequencies of IFN-γ-secreting CD4⁺ T cells were 2.3 and 2.0 times higher in patients treated with MA, on D7 and D21, respectively ( Figure 7E ). Our results suggest that the MA treatment stimulates higher secretions of IL-2 and TNF-α along the treatment and up to the cure, and of IFN-γ until D21. Treatment with LAMB, on the other hand, had no impact on the IL-2 iMFI response and decreased the frequencies of total TNF-α-secreting T cells along time.

As for the CD8⁺ T-cell response, and also in agreement with their higher severity, the levels of iMFI and total frequencies of IL-2⁺, TNF-α⁺, and IFN-γ⁺-secreting T cells on D0 were higher in the patients who were going to receive treatment with LAMB ( Figures 8A–F ). Cytokine secretion was predominant in LAMB-treated patients mainly until D7 ( Figures 8D, F ) or D14 ( Figures 8C, D ). On the other hand, MA treatment determined 10.6–16.1 times higher responses of IL-2 on D14 and D180 ( Figures 8A, B ); 15–10.6 higher responses of TNF-α on D60 ( Figures 8C, D ); and 3–10.1 times higher responses of IFN-γ on D21, D90, and D180 ( Figures 8E, F ). We conclude that the treatment with MA generated long-lasting predominant responses of cytokine secretion. More than the CD4⁺ T cells, the IL-2 and IFN-γ-secreting CD8 T-cell predominance appears to be associated with the cure.

Additionally, we analyzed the correlations between the clinical outcomes and the frequencies and iMFI of cytokine-secreting responses, considering the patients treated with MA or LAMB separately ( Table 2 ). More important clinical parameters of cure were associated with immunological variables in patients treated with MA. In fact, the CD4⁺ IFN-γ⁺ total frequencies and iMFI of MA treated patients increased as their liver sizes and monocyte counts decreased. The size of the livers is an important hallmark of VL. Furthermore, the intensity of TNF-α secretion by CD4⁺ T cells was highly correlated with the decrease of liver and spleen sizes and with the increase in platelets and leukocyte counts. The correlations with the decrease of spleen and liver sizes exhibited the highest R values. In addition, and importantly, the total frequencies of CD4⁺TNF-α⁺ were positively correlated to the hemoglobin and hematocrit values and the platelet, leukocyte, and neutrophil counts. The intensity of the secretion of TNF-α by CD4⁺ T cells also increases in correlation to platelets and leukocytes. In contrast, the total frequencies of CD8⁺IFN-γ⁺ of MA-treated patients were negatively correlated to the platelet counts ( Table 2 ). This result might be an indication of the association of terminal effector CD8⁺ T cells in VL.

In contrast, only three variables were associated with the clinical signs in patients treated with LAMB. The total frequencies of CD4⁺IL-2⁺ and iMFI of CD8⁺IL-2⁺ cells correlated negatively with lymphocyte counts, while the CD4⁺IL-2⁺iMFI values were correlated with eosinophils. However, studies with larger cohorts of patients are needed to confirm this hypothesis.

In order to clarify the impact of the MA and LAMB treatments on the immune responses, we further analyzed the frequencies of cytokine-expressing CD4⁺ T cells, but considering separately the patients treated with LAMB or MA ( Supplementary Figure 4 ). Similarly to what was observed when all patients were considered together ( Figures 1 and 2 ), no variation of frequencies of single secretors of IL-2 was detected. In contrast, and as detected for the analysis of the total frequencies ( Figure 7 ), the MA treatment promoted a main pronounced response in CD4 T cells, which secrete IFN-γ. In fact, the MA effect was observed in the increase of the frequencies of terminal-effector single secretors of IFN-γ throughout the whole treatment and until D180, of double secretors of IFN-γ and IL-2 on D7 and D14, and of the multifunctional CD4⁺IL-2⁺TNF-α⁺IFN-γ⁺ on D7. MA also induced increased proportions of double secretors of IL-2 and TNF on D14. In contrast, LAMB treatment impacted more prominently the secretion of TNF-α by single secretors on D7 and D90; by double secretors of IL-2 and TNF-α on D21, D30, and D180; by double secretors of TNF-α and IFN-γ on D14; by double secretors of IL-2 and IFN-γ on D21 and D60; and by multifunctional CD4⁺IL-2⁺TNF-α⁺IFN-γ⁺ on D30 and D60.

Additionally, the treatment with MA showed a stronger impact on the development of the CD8⁺ cytokine-secreting T-cell response ( Supplementary Figure 5 ), by increasing the frequencies of single producers of IL-2 on D14, D30, and D180; of single producers of TNF-α on D21 and D60; of single producers of IFN-γ on D14 and D21; of double producers of IL-2 and TNF-α by D14 and D60; of double producers of TNF-α and IFN-γ by D7 and D21; and of double producers of IL-2 an IFN-γ by D14, D90, and D180. In contrast, LAMB treatment enhanced the frequencies of single producers of IL-2 on D7; of single producers of TNF-α on D14; of double producers of IL-2 and TNF-α by D7 and D21; of double producers of TNF-α and IFN-γ by D14, D90, and D180; and of multifunctional CD8⁺IL-2⁺TNF-α⁺IFN-γ⁺ on D14, D21, D30, and D180.

Therefore, and different from what was suggested by the analysis of the total frequencies and iMFI ( Figures 7 and 8 ), the detailed analysis of the cytokine responses disclosed that both the MA and LAMB therapies promote immune responses. MA treatment had more impact on the secretion of IFN-γ, LAMB had stronger influence on the secretion of TNF-α by CD4⁺ T cells, but both generated memory CD4⁺ T-cell responses. In addition, the therapy with MA promoted increased frequencies of all types of single- and double-cytokine secretors, but only LAMB promoted higher responses of multifunctional CD8⁺IL-2⁺TNF-α⁺IFN-γ⁺ T cells.