The BALB/c mice used in this study were bred and maintained in the animal facility at the Instituto de Investigaciones en Microbiología y Parasitología Médica (IMPaM), Universidad de Buenos Aires–CONICET. All the procedures were approved by the Institutional Committee for the Care and Use of Laboratory Animals (CICUAL, Facultad de Medicina, Universidad de Buenos Aires, CD No. 2271/2014) and are in accordance with the guidelines of the Argentinean National Administration of Medicines, Food and Medical Technology (ANMAT), Argentinean National Service of Sanity and Agrifoods Quality (SENASA) and also based on the US NIH Guide for the Care and Use of Laboratory Animals.

All mice were provided with a 12-h day/night cycle and water and food ad libitum with a standard diet. Seven male mice per group were infected intraperitoneally with 1 × 10⁵ bloodstream trypomastigotes of a lethal RA (pantropic/reticulotropic) subpopulation of T. cruzi (23) and sacrificed by CO₂ inhalation at 10 days postinfection (dpi). Each experiment was carried out at least three times.

1-Methyl-3-hydroxy-4-pyridinecarboxylic acid derivative was resynthesized following the previously reported pathway by Brun et al. (22) with some modifications in the reaction conditions for the final steps of the synthesis and the purification step that led to the desired compound HP₂₄ in the zwitterion form (Figure 1A) instead of the previously described chloride compound. 3-Hydroxy-isonicotinic acid (1 g, 7.18 mmol) was suspended in 5 ml of DMF in a 25-ml round bottomed flask. The resulting suspension was stirred at room temperature, and 10% NaOH (7.5 ml) was added dropwise until complete dissolution of the solid (pH 9–10). Methyl iodide (2.06 g, 14.46 mmol, d = 2.28 g/ml, 0.9 ml) was added under stirring and the solution was then refluxed, monitoring the reaction progress by thin layer chromatography (n-butanol:H₂O:AcOH, 1:1:1). Once the starting material disappeared, the solvent was removed under reduced pressure, obtaining a deep orange colored solid, which was dissolved in boiling water (50 ml). The solution was acidified with 37% HCl (3.5 ml), and 10% H₂O₂ (1 ml) was added. Then, the iodine was exhaustively extracted with CHCl₃ (5× 15 ml) in a separating funnel. The organic phase was concentrated under pressure to dryness, obtaining an orange crude raw powdery solid (1.662 g), which was purified by reversed-phase chromatography in a Biotage Isolera Spektra Flash Chromatography apparatus equipped with prepacked C18 cartridges. The fractions containing the product were pooled and concentrated to dryness by means of a rotary evaporator, yielding a white powdery product (0.956 g, 6.21 mmol).

Mice were treated by oral gavage with HP₂₄ (400 mg/kg/day) suspended in phosphate-buffered saline (PBS), according to Brun et al. (25), for 9 days, since day 1 postinfection (pi), and sacrificed at day 10 pi. Euthanasia was carried out at this time for humanitarian reasons due to the unhealthy conditions of the mice infected with the RA parasite strain.

Parasitemia of T. cruzi-infected mice was analyzed through a small incision at the end of the tail. Blood from T. cruzi-infected mice and T. cruzi-infected HP₂₄-treated mice was obtained at 3, 5, 8, and 10 dpi. Fivefold dilutions were obtained in red blood cell lysis buffer (150 mM NH₄Cl, 0.1 mM EDTA, and 10 mM KHCO₃, pH 7.4). Parasitemia was measured in a Neubauer chamber. Body weight gain or loss was monitored on the days described previously. For survival studies, two independent groups (T. cruzi-infected and T. cruzi-infected HP₂₄-treated mice) of seven animals each were followed up daily until different days postinfection and analyzed by the Kaplan–Meyer method.

To elucidate the binding mode of the HP₂₄ molecule, docking calculations were performed against a murine PPARγ receptor. First, we built an homology model with MODELLER software (26) using human PPARγ structures as template from Protein Data Bank (http://www.pdb.org), PDB id 5DV6 and 2PRG (27). Then, we built and optimized HP₂₄ ligand structure with Molefacture module of VMD software (24) employing default settings for convergence criteria. Finally, AutoDock4 software (28) with Water Site Biased Method (29) was used for docking experiments. The PPARγ binding site (BS) was defined as all atoms within a radius of 10 Å from the co-crystallized ligand rosiglitazone (RSG, PDB id 2PRG). The receptor BS amino acids were treated as rigid structures during the calculations. A total of 100 different docking runs were performed, and the results were clustered according to the ligand-heavy atom RMSD using a cutoff of 2 Å. The genetic algorithm parameters for each conformational search run were kept at their default values. The AutoDock4.2 energy function (OA atom type map) was modified adding an additional energy term for each crystallographic water (CWS) placed on ABS BS in the apo form (PDB id 1PRG) to the original function (29).

Macrophages were obtained by washing the peritoneal cavity of mice with 8 ml of RPMI-1640 culture medium (Invitrogen Life Technologies, Grand Island, NY, USA), supplemented with 10% of heat-inactivated fetal bovine serum (FBS) (Internegocios S.A., Argentina) and antibiotics (50 µg/ml of penicillin, streptomycin, and gentamicin). Cells were left to adhere to the plastic surface of cell culture dishes, 35 mm × 10 mm (Greiner Bio One International AG) for 3 h at 37°C under a 5% CO₂ atmosphere (30).

Peritoneal macrophages were isolated from T. cruzi-infected mice as indicated above. The cells were preincubated for 30 min with specific PPARγ antagonist, T0070907 (50 µM in DMSO) (Sigma-Aldrich Co., St. Louis, MO, USA), and treated afterward with HP₂₄ (100 µM in PBS) for 48 h. Cell viability was examined by Trypan Blue dye exclusion test, for each treatment.

Release of NO by peritoneal macrophages from T. cruzi-infected mice was assessed by the Griess reaction, as previously described (31). The absorbance at 540 nm was compared with a standard curve of NaNO₂.

Macrophage-induced angiogenesis was quantified with an in vivo bioassay. After detaching, the concentration of macrophages was adjusted to 2 × 10⁵ cells/ml in culture medium without FBS. Seven normal syngeneic male mice per group were inoculated intradermally in both flanks with 0.1 ml of cell suspension. Five days after inoculation, mice were sacrificed, and the internal layer of skin was separated from the underlying tissues, and the vascular response was observed with a dissecting microscope (Konus USA Corporation, Miami, FL, USA) at a 7.5× magnification and photographed with an incorporated digital camera (Canon Power Shot A45, Canon USA, Inc., Lake Success, NY, USA). Photos were projected on a reticular screen to count the number of vessels per square millimeters of skin. Angiogenesis was quantified as vessel density, calculated as the total number of vessels divided by the total number of squares (32).

Macrophages (1.5 × 10⁶) from uninfected, T. cruzi-infected and T. cruzi-infected HP₂₄-treated mice were obtained and cultured with heart slices (100 mg/sample) from uninfected and T. cruzi-infected mice in 4 ml RPMI-1640 culture medium (Invitrogen Life Technologies, Grand Island, NY, USA). After 48 h, the culture supernatants were collected, and hearts were homogenized to obtain total proteins as previously described (33).

Hearts of all experimental groups were fixed in 4% paraformaldehyde in PBS, dehydrated and embedded in paraffin. Six non-contiguous sections (5 µm) were stained with hematoxylin–eosin or Masson trichrome stain. Cellular infiltrates, presence of amastigote nests and collagen deposition were examined in using a Nikon Eclipse E600 microscope (Nikon Inc.). Images were captured with a Spot RT digital camera. At least 30 random microscopic fields (400×) were analyzed in each microscopic section, using the open source Image J software (NIH, USA) (34).

Total RNA was extracted from frozen cells by using a QuickZol reagent (Kalium Technologies, Buenos Aires-Argentina). Total RNA was reverse-transcribed using Expand Reverse Transcriptase (Invitrogen Corp., MA, USA). RT-qPCR was performed using a 5× HOT FIREPol^® EvaGreen^® qPCR Mix Plus (ROX) (Solis BioDyneCorp., Estonia) in an Applied Biosystems 7,500 sequence detector. Primer sequences were: 18S: Fw 5′AACACGGGAAACCTCACCC 3′, Rv 5′ CCACCAACTAAGAACGGCCA 3′; connective tissue growth factor (CTGF): Fw 5′ CCTAAAATCGCCAAGCCTGT 3′, Rv 5′ CACCCCGCAGAACTTAGCC 3′, and PPARγ: Fw 5′ ATCTACACGATGCTGGC 3′, Rv 5′ GGATGTCCTCGATGGG 3′; PCR parameters were 52°C for 2 min, 95°C for 15 min, and 40 cycles of 95°C for 30 s and 60°C (for 18S), 63°C (for CTGF) or 54°C (for PPARγ). Quantification was calculated using the comparative threshold cycle (C_t) method and the efficiency of the RT reaction (relative quantity,2−ΔΔCt). The replicates were then averaged, and fold induction was determined, considering the value at time 0 as 1 (35).

TNF-α and IL-6 levels in culture supernatants were quantified by enzyme-linked immunosorbent assays using DuoSet antibody pairs (R&D Systems, Minneapolis, MN, USA).

Total protein extracts were obtained after washing the hearts with PBS and adding 300 ml of RIPA modified lysis buffer (50 mM NaCl, 50 mM Tris–HCl (pH 7.40), 1% Triton X-100, 1 mM EDTA, 1 mM PMSF; 2.5 g/l Protease Inhibitor Cocktail (Sigma-Aldrich Co., St. Louis, MO, USA), 1 mM Na₃VO₄, 1 mM NaF), or washing the cultured cells and scraped off the dishes with 50 µl of the same buffer. Then, the tubes were kept on ice for 30 min with swirling, and the samples were centrifuged at 7,000 g at 4°C for 10 min. The supernatants were stored at −20°C. Protein concentrations were determined by the Bradford method using the Bio-Rad Protein Assay (Bio-Rad, USA) and bovine serum albumin (Sigma-Aldrich Co., St. Louis, MO, USA) as a standard (36). For Western blot analysis, total proteins were boiled in Laemmli sample buffer, and equal amounts of protein (40–50 µg) were separated by 10–12% SDS-PAGE. The gels were blotted onto a Hybond-P membrane (GE Healthcare, Madrid, Spain) and incubated with the following antibodies: anti-NOS2, anti-NOS3, anti-Arginase I (Arg-I), anti-CD31, anti-VEGF-A, and anti-α-actin (Santa Cruz Biotechnology, CA, USA). The blots were revealed by enhanced chemiluminescence in an Image Quant 300 cabinet (GE Healthcare Biosciences, PA, USA) following the manufacturer’s instructions. Band intensity was analyzed using the NIH-ImageJ software (37).

Data are expressed as the mean of three independent experiments ± SEM for each treatment (seven mice/group). The Kaplan–Meier test was used to compare survival curves between groups. One-way ANOVA was used to analyze the statistical significance of the differences observed between the uninfected, uninfected HP₂₄-treated, T. cruzi-infected, and T. cruzi-infected HP₂₄-treated mice. The Tukey post hoc test was performed to compare all pairs of groups. Kruskal–Wallis test and Dunn’s post hoc test was used to analyze the differences in collagen deposition between uninfected, T. cruzi-infected, and T. cruzi-infected HP₂₄-treated mice. Differences were considered statistically significant when P < 0.05. All analyses were performed using the Prism 5.01 software (GraphPad, USA).

Selection of the 3,4-pyridinecarboxylic derivative HP₂₄ was based on its properties as PPARγ ligand and on its ability to reduce the pro-inflammatory response in a Dextran-induced colitis mouse model, as previously reported (22). A simplified synthesis route is shown in Figure 1A.

To elucidate how HP₂₄ interacts with PPARγ, we compared its binding against that of co-crystallized structure of synthetic PPARγ ligand, rosiglitazone (RSG). Figure 1B illustrates the superimposition of the best energy docking result for HP₂₄ (estimated free energy of binding −4.05 kcal/mol), as well as the experimental position of the co-crystallized RSG (estimated free energy of binding −8.95 kcal/mol), present in the 2PRG PDB structure. The most important interaction between RSG and PPARγ is a hydrogen bond network with residues GLN286, SER289, HIS323, and TYR473. Hydrophobic contacts with LEU330, ILE341, MET364, and CYS285 are considered as secondary interactions (27). Besides, a detailed visual inspection of the binding modes for HP₂₄ reveals the same hydrogen bond network, adding an extra-polar contact with HIS449 in the hydrophilic region of PPARγ BS (Figure 1B, panel right).

Since ligand binding can be inhibited by specific antagonists, we designed an in vitro assay to confirm that HP₂₄ binds to PPARγ as modeled by the in silico binding analysis. This involved the use of NO release by T. cruzi-infected peritoneal macrophages, as an indicator system of the inhibitory effect of HP₂₄ on pro-inflammatory mediators release. T. cruzi-infected macrophages release significantly higher amounts of NO than uninfected macrophages. Moreover, HP₂₄ significantly reduced the release of this pro-inflammatory mediator. Notably, preincubation of T. cruzi-infected macrophages with the specific PPARγ antagonist, T0070907 impeded the effect of HP₂₄ on NO release, thus confirming the specific binding and activation of PPARγ by HP₂₄ (Figure 1C).

As previously reported by our group, infection of mice with T. cruzi increases PPARγ expression in the heart and in peritoneal macrophages (30, 38). In this study, we tested the effect of HP₂₄ treatment on PPARγ mRNA levels in macrophages and hearts of uninfected and T. cruzi-infected mice. Figure 2 shows that HP₂₄ treatment increased the PPARγ mRNA expression in uninfected mice, but did not modify the already increased expression in T. cruzi-infected mice.

We evaluated whether the treatment with HP₂₄ modified the parasitemia levels, weight and survival of T. cruzi-infected mice. These parameters were evaluated at 3, 5, 8, 10, and 13 dpi. We found no differences in parasitemia levels, weight or survival between T. cruzi-infected and T. cruzi-infected HP₂₄-treated mice (Table 1). These results are consistent with those previously reported by our group with other PPARγ ligands (34).

We have previously demonstrated that PPARα and PPARγ ligands promote the polarization of macrophages isolated from T. cruzi-infected mice toward an M2 profile (30).

Here, we evaluated the role of the new PPARγ ligand HP₂₄ as an anti-inflammatory ligand. We determined its effect on the expression of NOS2 and pro-inflammatory cytokines. The treatment with HP₂₄ significantly inhibited NOS2 expression as well as TNF-α and IL-6 secretion in macrophages from T. cruzi-infected mice (Figure 3A). The PPARγ agonist did not affect the secretion of pro-inflammatory cytokines in macrophages from uninfected mice (Figure 3A).

We also analyzed whether HP₂₄ was able to promote the participation of macrophages in tissue repair and neovascularization processes. To this end, we analyzed the expression of representative proangiogenic markers like VEGF-A and Arg-I in peritoneal macrophages of T. cruzi-infected mice. Both T. cruzi infection and HP₂₄ treatment significantly increased VEGF-A expression in peritoneal macrophages with respect to uninfected and untreated mice. This effect was potentiated in macrophages from T. cruzi-infected HP₂₄-treated mice (Figure 3B). However, HP₂₄ was able to significantly increase the expression of Arg-I, both in T. cruzi-infected macrophages and in uninfected control cells (Figure 3B).

The formation of new blood vessels requires the sprouting of preexisting ones and their subsequent fusion with others (39). It has been previously reported that peritoneal macrophages from tumor-bearing mice are able to induce a strong neovascular response in the skin of syngeneic normal mice (32). Based on these results, we investigated whether HP₂₄ treatment modulated the ability of peritoneal macrophages to induce neovascularization in the skin of normal syngeneic mice. Peritoneal macrophages from T. cruzi-infected donor mice increased vessel density in comparison with macrophages from uninfected donors. Furthermore, macrophages from T. cruzi-infected mice-treated in vitro with HP₂₄ induced a higher increase in neovascularization than macrophages from T. cruzi-infected donor mice, upon passive transfer in the skin of normal syngeneic mice (Figure 4A). To assess the participation of PPARγ in the effects of macrophage-induced neovascularization, the cells were pretreated with the PPARγ antagonist T0070907 and then treated with HP₂₄. As shown in Figure 4A the PPARγ antagonist reduced the effect of HP₂₄ on new vessel formation in the skin of syngeneic normal mice.

Blood vessel density positively correlated with CD31 and VEGF-A expression. Western blot analysis showed higher expression of CD31 and VEGF-A in skin extracts of normal syngeneic recipient mice, upon transfer of macrophages from T. cruzi-infected mice treated in vitro with HP₂₄, than upon transfer of macrophages from T. cruzi-infected mice. The PPARγ antagonist T0070907 significantly reduced the effects of HP₂₄ on the expression of CD31 and VEGF-A in the skin of normal syngeneic recipient mice upon passive transfer of macrophages from T. cruzi-infected mice (Figure 4B).

As expected, CD31 and VEGF-A expression was higher in the skin of recipient mice transferred with macrophages from T. cruzi-infected mice than in mice transferred with macrophages from uninfected controls (Figure 4B).

To investigate whether macrophages from T. cruzi-infected HP₂₄-treated mice participate in heart angiogenesis, we cocultured macrophages from these mice with heart explants from T. cruzi-infected mice. Then, we evaluated the expression of proangiogenic markers in those explants. Macrophages from T. cruzi-infected mice induced an increase in VEGF-A expression in heart explants, whereas macrophages from control mice were unable to modify VEGF-A expression in the same explants (Figure 5). When heart explants were cocultured with macrophages from T. cruzi-infected mice treated in vitro with HP₂₄, the expression of VEGF-A, CD31 and Arg-I was further increased (Figure 5).

Moreover, the effect of HP₂₄ on the expression of VEGF-A, CD31 and Arg-I in heart explants was reverted when the macrophages from T. cruzi-infected mice were pretreated with the PPARγ antagonist T0070907 (Figure 5).

The same pattern was observed when macrophages were cocultured with heart explants from uninfected mice although the expression levels were lower (Figure S1 in Supplementary Material). Besides, the expression of CD31, VEGF-A and Arg-I did not differ in heart explants from T. cruzi-infected mice cultured alone or in the presence of macrophages from uninfected control mice (Figure S2 in Supplementary Material).

We investigated the effects of HP₂₄ treatment on the inflammatory response in the heart of T. cruzi-infected mice. Mice infected with T. cruzi showed intense inflammatory reaction, consisting of mononuclear cell infiltrates. Treatment with HP₂₄ significantly reduced heart inflammation (number of inflammatory foci/field, T. cruzi vs. T. cruzi-HP₂₄, 0.84 ± 0.41 vs. 0.21 ± 0.05, N = 5, P < 0.05, Figure 6). We did not find significant differences in the number of amastigote nests per field between the T. cruzi-infected and T. cruzi-infected HP₂₄-treated groups (Figure 6).

Fibrosis was observed in heart sections of T. cruzi-infected mice using Masson’s trichrome staining (Figure 7A). The heart area compromised by collagen deposits was significantly reduced in T. cruzi-infected HP₂₄-treated mice (Figure 7B). In addition, we analyzed the mRNA expression of CTGF as a profibrotic marker. RT-qPCR showed that CTGF expression levels were higher in the hearts from T. cruzi-infected mice than in those from uninfected mice. Interestingly, treatment with HP₂₄ significantly inhibited CTGF mRNA in infected mice (Figure 7C).

To demonstrate the proangiogenic role of the HP₂₄ treatment in the hearts of T. cruzi-infected mice, we evaluated the cardiac expression of CD31, VEGF-A, NOS3, and Arg-I. Western blot analysis showed increased expression of CD31 and VEGF-A in hearts upon infection while that of NOS3 and Arg-I remained unchanged. Treatment with HP₂₄ promoted the increase of all proangiogenic markers in the heart of T. cruzi-infected mice. (Figure 8A). Moreover, HP₂₄ treatment in vivo reduced the expression of NOS2 in the hearts of T. cruzi-infected mice in comparison with untreated T. cruzi-infected mice (Figure 8B).