Two photosynthetic microorganisms (PMs), the cyanobacterium A. platensis (Utex, 1926) and the microalgae D. tertiolecta (Utex, 1644), were obtained from the University of Texas culture collection. Arthrospira platensis was cultivated in a 250-ml Erlenmeyer flask containing 100 ml of Schlösser medium (19) in an orbital shaker at 75 ± 5 rpm, while D. tertiolecta was cultivated in a 1-L Erlenmeyer flask containing 400 ml of f/2 medium (20), under constant aeration. Both cultivations used an initial biomass concentration of 50 mg/L, temperature of 27°C ± 1°C, and continuous light intensity of 52 ± 4 μmol photons m⁻² s⁻¹ (21).

After reaching the stationary growth phase, the cultures were centrifuged at 430×g for 10 min (Hermle Labortechnik 326 HK, Wehingen, Germany), lyophilized (SP Scientific BenchTop Pro, Warminster, USA), resuspended (100 mg ml⁻¹) in 0.1 M of Tris–HCl buffer pH 7.0, and sonicated at 20 pulses in an ice bath with intervals of 1 min between each pulse (22). The homogenate was centrifuged at 10,000×g for 5 min at 4°C, and the supernatant filtered at 0.22 µm was denominated bioactive extract and used for further analysis. Protein concentration was determined by the method of Smith et al. (23) using the Micro BCA Protein Assay Kit (Pierce).

The promastigotes of L. infantum (strain MHOM/BR/1972/BH46) were culture-expanded in Schneider medium (Sigma-Aldrich) pH 7.2, containing 10% of fetal bovine serum (FBS—Vitrocell^®, Brazil) and 1% of antibiotic (penicillin 200 U/ml). The peaks were performed between the third and fourth day of cultivation and incubated in a biochemical oxygen demand greenhouse (BOD) at a temperature of ±26°C, to be used in the leishmanicidal effect tests.

To obtain the antigenic fractions, the parasitic mass was subjected to rapid freezing (−197°C) in liquid nitrogen, followed by thawing (40°C). Soon after, the cell suspension was subjected to the ultrasonication process (60 s/40 W), three times, and centrifuged at 10,000×g for 20 min at 4°C. The resulting supernatant was a soluble antigen fraction (L. infantum soluble antigen: LSA). Proteins were quantified by the Bradford method (24), and the fraction was stored at −80°C. LSA was used in the cell culture assays.

Whole blood was collected from healthy human individuals by venipuncture using heparin tubes (18 ml) to perform the separation of peripheral blood mononuclear cells (PBMCs). Phosphate-buffered saline (PBS) (pH 7.2) was added to heparinized blood in a proportion of 1:2. This mixture was added to Ficoll-Hypaque at a ratio of 1:3 (Ficoll-Hypaque:blood) to obtain the ring of PBMCs after centrifugation at 900×g for 30 min without brake. Cells were washed twice in PBS and then resuspended in 1 ml of RPMI 1640 medium supplemented with 10% of FBS and 1% of penicillin/streptomycin (200 U/ml). Cells were counted in a Neubauer chamber using Trypan Blue dye. The cell value obtained was adjusted to a concentration of 1 × 10⁶ cells/ml by adding RPMI 1640 medium supplemented with 10% of FBS, to be used in the assays.

The selectivity index (SI) was calculated as the ratio between the 50% cytotoxic concentration (CC₅₀) in human cells and the 50% inhibitory concentration (IC₅₀) in parasite cultures. All cells were treated with bioactive extracts of photosynthetic microorganisms and reference drugs (N-methylglucamine antimoniate and miltefosine).

PBMCs from healthy humans were cultivated in 96-well plates and treated with different stimuli to evaluate the CC₅₀. Cell treatment (containing 1 × 10⁶ cells/ml) was performed in triplicate for each stimulus, incubated for 24 h, using concentrations of 62.5 to 1,000 µg/ml of bioactive extracts of PMs (25, 26). After the culture time, the supernatant was discarded and MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide] was added at a concentration of 1 mg/ml, solubilized in RPMI 1640 medium without phenol red, and incubated in an oven at 37°C for 3 h. Subsequently, the formazan crystals formed were solubilized with dimethylsulfoxide (DMSO) and then read in a spectrophotometer at 540 nm (27).

Cells of L. infantum in the promastigote form were cultivated for the evaluation of the IC₅₀ of the PM bioactive extracts. Cell treatment was performed in triplicate for each stimulus, incubated for 48 h in a 96-well plate containing 1 × 10⁶ cells/ml (28). After the culture time, the cells were counted and evaluated for their morphology under a light microscope. For natural extracts, protein concentrations of 15.63 to 500 µg/ml were used (29). For N-methylglucamine antimoniate and miltefosine, the concentration curves were 15.63 to 250 µg/ml and 0.1 to 4 µg/ml, respectively (30, 31). Three independent experiments were carried out in triplicates.

Whole blood was collected from healthy humans to obtain the PBMCs. PBMC cultures were performed in 48-well polystyrene plates, containing 5 × 10⁵ cells/well, and stimulated with the mitogen phytohemagglutinin (PHA) (10 µg/ml) (positive control of the culture), the soluble antigen of L. infantum (LSA) (25 µg/ml) (32), N-methylglucamine antimoniate, miltefosine, and natural extracts. The concentration values—CC₅₀, ½ CC₅₀, ¼ CC₅₀, and IC₅₀—of the reference drugs and natural extracts were determined. Unstimulated wells served as a negative control.

Probit regression analyses were performed using the IBM SPSS Statistic 25 software program in order to determine the values of CC₅₀ and IC₅₀. Also, comparative analyses were performed using descriptive statistics techniques with absolute and percentage distributions. The Kolmogorov–Smirnov normality test was used, and according to the analysis, Student’s t-test (parametric) or the Mann–Whitney test (non-parametric) was performed for comparisons between two categories, or one-way analysis of variance (ANOVA) with Tukey post-test (parametric) or Kruskal–Wallis with Dunn’s post-test (non-parametric) was performed for comparisons between more than two categories. Correlation analyses were also performed using Pearson’s test (parametric) and Spearman’s test (non-parametric). All tests were performed using the GraphPad Prism 5 program (GraphPad Software Inc. 2007, San Diego, CA, USA), and a p-value <0.05 was considered significant.

Before the collection, human individuals included in the study were invited to participate and sign an informed consent form, authorizing the use of the collected material for scientific purposes. This study is part of the project “Immunology applied to the development of new control strategies for leishmaniasis” which has the approval of the Research Ethics Committee: CAAE 89972718.8.0000.5190 No. 4,077,060 (Fiocruz/PE).

Assays for the determination of CC₅₀ and IC₅₀ were performed, and it was observed that the extract of A. platensis had the lowest toxicity for human PBMCs, presenting a CC₅₀ of approximately 1,000 µg/ml. In turn, D. tertiolecta presented higher toxicity to L. infantum, with a lower IC₅₀.

By analyzing the reference drugs Sb^V and MTF, we observed that MTF was more toxic to L. infantum cells as well as to the host cells. Despite the high toxicity, the MTF was over 100 times more selective than other treatments (SI = 129.6). On the other hand, Sb^V had the lowest SI (2.11), demonstrating that the extracts of A. platensis and D. tertiolecta were more selective for the parasite (SI = 3.8 and 4.7, respectively) than this reference drug, which is currently used for the treatment of VL in Brazil. Table 1 shows the CC₅₀ and IC₅₀ values as well as the SI.

For the extracts, we observed that both A. platensis and D. tertiolecta induced the production of cytokines of the Th1 profile by PBMCs, at all concentrations evaluated, demonstrating a positive stimulus against VL ( Figure 1 ). We emphasized IL-10, which gradually decreased its production as the concentration of the extracts increased, with statistically significant differences between the concentrations of the A. platensis (p-value = 0.031) and D. tertiolecta (p-value = 0.007) extracts.

Comparing the results obtained from the extract of A. platensis with LSA, we observed the presence of a Th1-protective response for all concentrations, with IFN-γ stimulation higher than that observed with the natural infection mimicked by LSA, with a statistically significant difference for the IC₅₀ concentration (p-value = 0.032) ( Figure 1D ). For all concentrations ( Figures 1A–D ), the A. platensis extract did not stimulate TNF production that was superior to LSA, with statistically significant differences, but it stimulated the production of this cytokine more than the cells without stimulation. Regarding the extract of D. tertiolecta, compared to LSA, we detected that, for all concentrations, a Th1-protective response was found, with the mean of TNF stimulation higher than that observed with LSA ( Figures 1E–H ). Furthermore, at the concentration of CC₅₀, a lower stimulus for IL-10 production (p-value = 0.002) and for IFN-γ (p-value = 0.008) was found than LSA ( Figure 1E ). Table 2 shows the mean values of cytokine production (in pg/ml) after cell stimulation with the four concentrations of each extract.

When analyzing the correlation of cytokines and the extracts, we observed that between TNF and IL-4, in the CC₅₀ and ½ CC₅₀ concentrations of the A. platensis extract, there were a very strong and significant negative correlation (r = −0.92; p-value = 0.02) and a significant positive correlation (r = −0.97; p-value = 0.01), respectively. For the extract of D. tertiolecta, we observed very strong but not significant correlations. Supplementary Table 1 shows the correlation data of cytokines induced by the extracts.

Regarding the reference drugs, all concentrations of Sb^V, except IC₅₀, showed an immunological suppression, with only IL-10 being stimulated. As for MTF, all concentrations, except IC₅₀, showed immunological suppression, with stimulation of only TNF at the ¼ CC₅₀ concentration. Table 3 shows the mean values of cytokine production stimulated by the concentrations of each drug. Regarding the correlation between cytokines, for Sb^V, there were very strong but not significant correlations. For MTF, we observed very strong and significant negative correlations between IL-4 and IL-2 in the CC₅₀ concentration (r = 0.90; p-value = 0.03) and between TNF and IL-10 in the IC₅₀ concentration (r = 0.91; p-value = 0.01). Supplementary Table 2 shows the correlation data of cytokines induced by the reference drugs.

For the extracts, we observed a slightly higher NO production when compared to cells without stimulus and stimulated with LSA (which demonstrated a reduction in NO production, when compared to cells without stimulus). We also identified that the nitrite dosage was directly proportional to the stimulus concentration ( Figure 2 ). Statistically significant differences were observed between the highest concentration (CC₅₀) of each extract compared with the LSA (A. platensis: p-value = 0.017; D. tertiolecta: p-value = 0.005).

In turn, the reference drugs also obtained a measurement of nitrite that was a little higher than the cells without stimulation and stimulated with LSA, but no proportionality between dosage and concentration was detected. For Sb^V, a statistically significant difference was found between the concentrations of ½ CC₅₀ (p-value = 0.028) and ¼ CC₅₀ (p-value = 0.023) with the LSA, pointing out a higher NO induction of these concentrations compared to the antigen. Moreover, for MTF, a statistically significant difference was also found between the CC₅₀ and LSA (p-value = 0.041).

From the stimulated cells, messenger RNA was extracted and reverse-transcribed to cDNA for quantification, targeting the genes Tbx21, GATA3, RORc, and FOXP3, which are related to the induction of T-cell differentiation for the Th1, Th2, Th17, and Treg profiles, respectively.

When analyzing the results of gene expression stimulated with natural extracts ( Figure 3 ), we observed that in all concentrations of A. platensis extract, a balance between the Th1/Th2 profile was present, with the average expression of the Tbx21 gene being higher than that of GATA3 at the highest stimulus concentration, with no statistical difference observed, similar to the results of gene expression stimulated with LSA. Notably, the FOXP3 gene at the highest concentration showed a negative expression ( Figure 3A ), and the RORc gene varied from concentration to concentration, showing a greater expression at the ¼ CC₅₀ concentration ( Figure 3C ; Table 4 ).

Regarding the extract of D. tertiolecta, we observed that at the lowest concentrations ( Figures 3G, H ), there was a balance between all the inducing genes, with the average expression of the GATA3 gene being lower than the others. As shown by the results, the ½ CC₅₀ concentration ( Figure 3F ) stimulated a balanced and positive expression between the Th1×Th17×Treg profiles, with a negative expression of GATA3, inducing the Th2 profile, and a statistically significant difference was observed when compared to LSA (p-value = 0.009). At the highest concentration, there was only a positive expression of the RORc gene ( Figure 3E ). Analyzing the correlation between genes, for A. platensis, a very strong positive correlation was observed between Tbx21 and RORc at the ½ CC₅₀ concentration (r = 0.95; p-value = 0.047). Regarding the correlation analysis of the extract of D. tertiolecta, at the concentrations tested, no significant correlations were detected. Supplementary Table 3 shows the correlation data of gene expression induced by the extracts.