Ethics statement

All experimental procedures were approved by the Animal Ethics Committee of Oswaldo Cruz Institute/Fiocruz (licenses L-006/2018, L-002/2023-A2). The in vivo data presented was obtained from two independent experiments (Register Books #49, #53, #66, #71, and #75, LBI/IOC-Fiocruz) conducted according to the Brazilian Federal Law 11.794 (October 8^th, 2008). The drawings were made by hand using licensed PowerPoint and Adobe Photoshop 2024.

Experimental design

The Author´s Checklist (S1 Checklist) describes the experimental checklist. The Institute of Science and Technology in Biomodels (ICTB) of the Oswaldo Cruz Foundation (Fiocruz) provided C57BL/6 strain (H-2^b) 5–7-week-old female mice (107 mice in total). After arrival, mice were numbered and grouped into 3–5 mice per polypropylene cage containing Pinus sawdust. Cages were numbered and randomly sorted for the experimental infection and analysis at the indicated time points. Cages were kept in microisolators, and mice received water and grain-based chew food ad libitum. Mice were adapted for 14 days in a plastic igloo-enriched cage in specific pathogen-free conditions, with light and noise control to minimize stress. After this period, mice were infected and analyzed as described in the Workflow chart (S1 Fig), according to The Arrive guidelines 2.0 [45]. GVP, IRP, AAS and JLV performed two independent experiments. Groups were composed as follows: Kinetic and Pre-therapy: non-infected (NI) controls (10 mice = 5 + 5); T. cruzi-infected (18 mice = 9 + 9), analyzed from 14 to 150 days post-infection (dpi) for NO and TNF serum levels. Therapy groups (therapy 120–150 dpi; analysis at 149–151 dpi – referred to as “150 dpi”): non-infected (NI) controls (10 mice = 5 + 5); and T. cruzi-infected groups: Vehicle-treated (Veh; 19 mice = 9 + 10); pentoxifylline-treated (PTX; 15 mice = 7 + 8), benznidazole-treated (Bz; 18 mice = 8 + 10), Bz + PTX-treated (Bz + PTX; 17 mice = 7 + 10). All infected mice were evaluated for parasitemia in the “Kinetic”, “Pre-therapy” and “Therapy” experiments (S1 Fig).

Infection by Trypanosoma cruzi, parasitemia and clinical follow-up

Mice were infected intraperitoneally with 100 blood trypomastigote (bt) forms of the Colombian T. cruzi strain (DTU-TcI) suspended in 0.2 mL of sterile saline buffer, and parasitemia was performed weekly, as previously described [33]. Weekly, death was recorded, and clinical signs were analyzed (piloerection, apathy, prostration, mobility, posture, aggressive behavior, pain). Body weight loss (which may reveal loss of appetite, diarrhea, and absorption disturbance) was assessed using a rodent weighing scale (Sartorius scale, ED623S-OCE, USA). Exclusion criteria were negative parasitemia (35–45 dpi), any death outside of planned euthanasia and according to pre-established endpoints conditions such as body weight loss (≥ 30% of the initial weight), injuries from fights, pain, posture, ataxia, and immobility. Parasitemia was evaluated microscopically in 5 µL of tail vein blood according to Pizzi-Brener method [46]. At 152 dpi, mice were euthanized under anesthesia, and tissues (encephala and hearts) were collected in RNA later stabilization solution (Invitrogen, USA) for processing prior to molecular assays. Heart tissues were processed for mRNA and miRNA expressions, and the data were previously published [43,44].

Therapy administration

After 120 dpi, when mice show clinical signs of behavioral and cognitive changes [17,18], we started the therapeutic intraperitoneal injection with apyrogenic saline (BioManguinhos/Fiocruz, Brazil) containing PTX (20 mg/Kg/day; Trental, Sanofi-Aventis) and/or Bz (25 mg/Kg/day; LAFEPE) by gavage using apyrogenic water (BioManguinhos/Fiocruz, Brazil), for 30–31 consecutive days [37,43,44]. Non-infected (NI group) and infected mice (Veh group) received an intraperitoneal injection with apyrogenic saline and gavage with apyrogenic water as control groups and groups receiving monotherapies (Bz or PTX), also received complementary Veh treatment, to rationally submit all mice to the stress of daily manipulation as were mice treated with Bz + PTX.

Behavioral tests

Our group established the environmental conditions and behavioral tests according to the literature and fully described them in our previous publications [15,17,18,47]. Briefly, behavioral tests were conducted between 8:00 a.m. and 3:00 p.m. and recorded using a DSC-DVD810 video camera (Sony, USA). The experiments were conducted in a controlled environment with 12 hours of light and 12 hours of dark at a stable room temperature of 22 ± 2 °C and an air conditioner noise level of approximately 40 dB. This setting was designed to reduce stress and enhance the mice’s familiarity with the environment. Mice were submitted to behavioral tests at 149–151 dpi, thereafter, stated as “150 dpi” in “Therapy experiments”. They were not retested, but all were reused in different tests to reduce the number of animals used. Behavioral tests were performed from less stressful tests to the more stressful: spatial habituation memory test (OFT2), novel object recognition memory test (NORT), elevated plus maze test (EPMT), marble burying test (MBT), tail suspension test (TST), and aversive shock evoked test (ASET). The open field test (OFT) is based on mice’s tendency to explore new environments. To evaluate the habituation spatial memory (OFT2), on the first day (day 1, training session), mice are placed in the upper corner of the testing area and allowed to explore for 5 minutes. The long-term memory test (day 2) was performed 24 hours after the training. In this test, the procedure was repeated, the 5-minute session was recorded, and memory retention was evaluated by counting the number of total lines crossed during the test session. Individual baseline differences were corrected using the change ratio score to compare behavior during the training (day 1) and test session (day 2). The data are shown as discrimination index (DI), as follows: DI = number of crossed lines day 2/ (number of crossed lines day 1 + number of crossed lines day 2). The NORT is based on the innate preference of NI normal rodents to explore a novel object rather than a familiar one. The test was initiated two hours after the OFT2. The mouse was allowed to explore two similar objects for five minutes. After 24 hours (day 3), one object was replaced (a novel object), and mice were submitted to a five-minute session [48]. According to a previous publication [17], the DI for NORT was calculated as follows: DI = time exploring the novel object/ (time exploring the novel object + time exploring the familiar object).

Anxiety-like behavior was assessed using EPMT, also based on the natural tendency of mice to explore a new environment. The behavioral parameters related to anxiety were the (i) number of entries (EPMT-entries) and (ii) time spent (EPMT-time) in the open arms. An entry was scored when all four mouse paws were placed on an arm [49]. Mice exhibit repetitive compulsory burying behavior in the presence of aversive stimuli; this is the basis of the MBT. Here, glass beads were used as aversive stimuli, and the number of buried glass beads in sawdust was registered at 5, 10, 15, 20, 25, and 30 minutes, as described [16].

The immobile posture or apathetic behavior in response to short-term moderately stressful or uncomfortable inescapable conditions is interpreted as depressive-like behavior in TST [50]. TST was performed using the tail suspension apparatus (TST, Insight, Brazil), as previously described [47]. Data are expressed as immobility time in seconds (s). As we previously described [17], the aversive shock-evoked test was performed using the inhibitory avoidance apparatus (EP 104MR, Insight, Brazil). In a pre-exposure session, mice were placed on the platform and allowed to explore the box freely for 5 minutes without foot shock. Two hours later, a training session was carried out when mice were placed on the platform, and an automatic device measured their latency to step down on the grid with all four paws. Immediately after stepping down on the grid, the animals received a 3.0-sec scrambled foot shock (0.6 mA). One to five stimulations were required for memory acquisition, and the latency between stimuli was registered (training session, ASET1). In the test session, 24 hours after training, no foot shock was administered, and the step-down latency (maximum 120 seconds) was used to measure memory retention (test session, ASET2). Data are expressed as latency measured in seconds (s), as described [51].

Blood obtention, euthanasia and encephalon tissue obtention

After topical eye drops anesthesia, mice were bled by orbital plexus, and serum was stored at −80 °C. Mice were euthanized at the endpoints (120 dpi and 150 dpi) using CO₂ inhalation in an appropriate chamber, allowing 70% CO₂ saturation for 2–3 minutes, followed by decapitation. Encephala and hearts were collected and weighed. Considering the importance of the cerebral cortex and hippocampus for behavioral and cognitive changes, these areas were dissected. Briefly, after the brain was removed and weighed, it was placed in an ice-cold saline solution to preserve tissue integrity [52]. The cortex and hippocampus were identified using anatomical landmarks [53]. Each region was carefully dissected using fine surgical instruments, ensuring that the boundaries between regions were respected to avoid contamination. The cortex was separated from the hippocampus by making precise cuts along the rhinal and the hippocampal fissures. After dissection, each brain region was preserved in RNA later (AM7021, Thermo Fisher Scientific, USA). The tissues were immediately stored at −80 °C for the following assays: BDNF; lipid peroxidation by evaluation of tissue levels of thiobarbituric acid reactive species (TBARS); GABA/glutamine determination by nuclear magnetic resonance; and parasite load determination by quantitative PCR (qPCR).

Determination of parasite load in the CNS by quantitative PCR (qPCR)

Mice from two independent experiments were euthanized, and the encephalon was removed and dissected [52]. DNA was extracted from dissected cerebral cortex and hippocampus samples, and the parasite load was measured by qPCR [17]. The standard curve was prepared by spiking 10–100 mg of non-infected brain tissue with 10⁵ T. cruzi trypomastigotes (Colombian strain). DNA was extracted and serially diluted (1:10) in TE buffer. This standard curve was subjected to qPCR assays for the T. cruzi satDNA and mouse GAPDH targets. The number of parasite equivalents and the tissue mass (in milligrams) of the samples were calculated by absolute quantification, using the respective standard curves (S2 Fig). Finally, the parasite load was calculated by dividing the number of parasite equivalents by the tissue mass (mg), using the formula:

BDNF gene expression analysis by real-time quantitative RT-PCR

For BDNF studies, the cerebral cortex was removed and processed as previously described [18]. Total RNA was extracted from the same sample used to quantify the parasite load using TRI-Reagent (Sigma-Aldrich, St. Louis, MO, USA). The procedures to obtain RNA, cDNA, and gene expression analysis by RT-qPCR using the TaqMan Assays (Applied Biosystems, USA) were described elsewhere [18]. Assays were performed according to the manufacturer’s instructions for BDNF neurotrophic factor (Mm04230607-s1) as the target gene and GAPDH (Glyceraldehyde-3-phosphate dehydrogenase; Mm99999915g1) and HPRT (Hypoxanthine-guanine phosphoribosyltransferase; Mm01545399m1) as reference genes, using Viia7 (Applied Biosystems, USA), as previously described [18].

Determination of glutamate and GABA neurotransmitters in the CNS by nuclear magnetic resonance

The cortex and hippocampus were analyzed for GABA and glutamate neurotransmitters using nuclear magnetic resonance. All procedures to obtain samples were previously described [18]. Proton Nuclear Magnetic Resonance (¹H NMR) spectra were acquired at 25 °C on a VNMRS-500 (Agilent, USA) spectrometer at 499.78 MHz. Spectra were acquired with a spectral width of 7022 Hz, with 16384 points and 64 accumulations. For quantitative analysis, a relaxation interval (d1) of 35 s was used. Nuclear magnetic resonance spectra were processed with the MestReNova v10.0.1-14719 program.

Lipid peroxidation evaluation in the CNS

Mice were euthanized, encephala removed, the cortex and hippocampus dissected [52], and homogenates prepared. Malondialdehyde and other TBARS, products of degradation of hydroperoxides and lipid peroxides formed during the oxidation of fatty acids were determined through their reaction with thiobarbituric acid (TBA) under heat and acidic pH, forming a pink-colored complex, as previously described [17]. Ten percent (w/v) of tissue homogenate was mixed with 8.1% sodium dodecyl sulfate (SDS), 20% acetic acid pH 3.5, and 0.8% TBA and incubated at 95 °C for 1 hour. After incubation, the reaction product was extracted with n-butanol (1:1), centrifuged for 5 minutes at 800 x g, and supernatants were collected. This MDA-TBA2 adduct exhibits a conjugated double-bond system, which absorbs visible light with a peak at 532 nm, read using a spectrophotometer (Spectra Max M5, Molecular Devices, USA).

NO quantification in sera

Nitrate and nitrite (NO_x) were determined in serum samples after deproteinization and reduction of nitrate by vanadium (III) chloride combined with detection by the acidic Griess reaction, and using a standard curve of 0.8 to 100 μM NaNO₂ and NaNO₃, as described elsewhere [54].

Determination of cytokines in sera

The BD Cytometric Bead Array (CBA) Mouse Inflammation kit (catalog 552364, BD Bioscience) measured serum IL-6, IL-10, MCP-1/CCL2, IFN-γ, TNF, and IL-12p70 protein concentrations using standard calibration curves according to the manufacturer’s recommendations. The samples were analyzed using the 13-Color CytoFLEX-S flow cytometer (BeckmanCoulter, USA). Cytokine concentrations were determined using reference curves and expressed in pg/ mL using the FCAP Array Software.

Analysis of mRNA and miRNA in cardiac tissue, pathway enrichment, and miRNA target filter analysis

This study builds upon previous findings highlighting the absence of cytokine gene expression within the CNS but elevated cytokine levels in serum of chronically T. cruzi-infected C57BL/6 mice [18]. Our study aimed to delineate the systemic inflammatory profile through a detailed examination of TNF expression in cardiac tissues assessed as previously described [37], complemented by mRNA data [43]. This investigation utilized TaqMan probes targeting 92 immune response genes alongside four endogenous reference genes for normalization and quantification. Such an approach allowed for the establishment of a correlation between specific gene expressions in cardiac tissues (including TNF, IFNγ, IL-10, IL-12a, IL-12b, IL-2, IL-6) and serum inflammatory mediators (NO, TNF, IL-6), thus elucidating the ramifications of T. cruzi infection and Bz-based treatments on these markers. Data normalization was conducted using NI controls.

Considering the limitation in accessing miRNA transcriptome data within brain tissues and serum samples, an exploratory approach was adopted to investigate the systemic miRNA profile under T. cruzi infection and therapeutic protocols. The aim was to identify potential targets for differentially expressed miRNAs (DEMs) under various conditions of infection and treatment, drawing upon miRNA datasets provided by Farani et al. [44], that utilized pre-printed TLDA 384 wells microfluidic cards (TaqMan™ Array Rodent MicroRNA A + B Cards Set v3.0, Applied Biosystems, ThermoFisher Scientific, USA/Cat no. 4444909) that contained FAM/NFQ-MGB labelled probes specific to 752 mature miRNAs and four endogenous small nucleolar RNAs candidates for data normalization and relative quantification. The Ingenuity Pathway Analysis (IPA; Qiagen, Redwood City, CA) software was pivotal for this analysis. It facilitated the exploration of direct and indirect relationships between miRNAs and their potential target genes within CNS-related pathways and evaluated the systemic effects of Bz and Bz + PTX therapies. DEMs were mapped to the IPA knowledgebase via miRNA IDs, and a miRNA target filter analysis was employed to pinpoint both experimentally observed targets and those predicted with high confidence. A core analysis was then conducted with the identified targets to discern canonical pathways, predominant diseases, and affected functions. Using the “Grow” and “Connect” tools within IPA allowed for a visualization of the intricate interactions between miRNAs and mRNAs, alongside additional direct and indirect connections amongst molecules at a pre-determined confidence level. This comprehensive analytical framework, enhanced through the “Path Designer” tool for clear visualization, facilitated an in-depth understanding of the molecular interactions at play, offering novel insights into the systemic inflammatory response and therapeutic avenues in this preclinical model of CD.

Statistical analysis

The sample size was determined based on the experience of our group and previous studies using the model of experimental chronic chagasic cardiomyopathy [33,34,37] and behavioral alterations [15,17,18]; therefore, no formal sample size was calculated. The Shapiro-Wilk test was used to assess the normality of the data. For normally distributed data formed by two groups, differences were analyzed with Student’s t-test. For normally distributed data of more than two groups, the difference between groups was analyzed using the parametric one-way ANOVA test, corrected with Tukey post hoc test with multiple comparisons, with a 95% confidence. Non-normally distributed data were analyzed using the non-parametric Mann-Whitney test for groups with two groups or Kruskal-Wallis on ranks for analysis with more than two groups, followed by the post-hoc Dunn’s multiple comparisons tests. Data from two independent experiments were used. All statistical tests were performed with GraphPad Prism 8.0.1 (La Jolla, CA, USA). Data are expressed as mean ± SD. Differences were considered statistically significant when p < 0.05.

To visualize simultaneously the most important variables involved in our analysis, we computed principal-component analysis (PCA), which allowed us to identify graphically the analyzed parameter and recognize the variables associated with the performance of the NI controls and the Veh-, Bz-, PTX-, Bz + PTX-treated T. cruzi-infected groups. To enhance the PCA, we only included variables that showed significant correlation based on the Spearman correlation significant test (p < 0.05). All analyzed parameters of four mice per group were included in PCA analyses. PCA was performed on scaled normalized expression values using the built-in R function PCA from package FactoMineR (v1.34), and the plots were generated using the ggplot2 (v3.3.3), factoextra (v1.0.7) and pca3d (v0.10.2) R packages. All R analysis was carried out using the RStudio environment (v1.4.1103) [55].

Effects of PTX and Bz therapies on parasitemia and parasitism control in chronically T. cruzi-infected mice

Experiments were performed according to the experimental Workflow (S1 Fig). S1 Table shows data used to build graphs and figures. During treatment all mice survived. At 120 dpi, non-infected and T. cruzi groups had similar body weight (S3A Fig). At 150 dpi, mice groups treated with Bz (p = 0.083) and Bz + PTX (p = 0.023) showed reduced body weight compared with age-matched non-infected mice (S3A Fig). Crucially, in Bz-treated groups we noticed a significant reduction of infection-associated splenomegaly (S3B Fig), a highly deleterious aspect of chronic Chagas infection, associated with increased cell activation [33]. No signs of injury, ataxia, or pain were registered. A parasitemia peak was detected at 42 dpi, parasite control was effective, and persistent low or intermittent parasitemia was seen in the chronic phase, at 120 and 150 dpi (Fig 1A). A kinetic study in Colombian-infected C57BL/6 mice showed that most of the behavioral changes are established at 120 dpi [18]. Based on these findings, we settled experiments to test the effects of Bz and PTX initiating therapies at 120 dpi (Fig 1B), allowing us to evaluate their effects in preventing the onset of or reversing behavioral and cognitive changes. At 150 dpi, all C57BL/6 mice infected with 100 bt forms of the Colombian strain survived. As expected, Bz therapy reduced parasitemia and parasite load in the cerebral cortex and hippocampus (Fig 1C, D). PTX treatment did not affect parasitemia (Fig 1C) but reduced parasitism in the cerebral cortex and hippocampus (Fig 1D). Moreover, the Bz + PTX combination controlled parasitemia and parasite burden in the CNS areas (Fig 1C, D).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Effects of PTX and Bz therapies on parasitemia and brain parasitism in chronically infected mice.

(A) The graph shows the kinetic of parasitemia (7-150 dpi). The blue arrow indicates the beginning of therapies (120 dpi). (B) The scheme shows that mice infected with 100 bt forms of the Colombian strain (T. cruzi) were treated with Veh, PTX, Bz, or Bz + PTX, daily from 120-150 dpi. (C) The graph shows parasitemia in Veh-, PTX-, Bz-, and Bz + PTX-treated mice at 150 dpi. (D) The graph shows parasite DNA quantified by qPCR in the cerebral cortex and hippocampus of non-infected control (NI) and T. cruzi-infected mice at 150 dpi. Representative data of two independent experiments. Color code: Grey bars indicate NI, blue bars show Veh-treated, green bars indicate PTX-treated, pink bars show Bz-treated, and yellow bars indicate Bz + PTX-treated infected mice. Each bar shows mean ± SD. ^###, p < 0.001, comparing treated groups with Veh-treated T. cruzi-infected mice; ^&&, p < 0.01, comparing Bz- and Bz + PTX-treated groups with PTX-treated infected mice.

https://doi.org/10.1371/journal.pone.0334708.g001

Effects of therapies with PTX, Bz, and Bz + PTX on behavioral and cognitive changes in chronically T. cruzi-infected mice

At 150 dpi, in contrast with the innate compulsive behavior (assessed by MBT) detected in NI mice, all Veh-treated T. cruzi-infected mice showed reduced ability to bury aversive materials (marbles, in this case) present in their environment (p < 0.001; Fig 2A). Compared with Veh-treated T. cruzi-infected mice, the three treatment schemes ameliorated the innate compulsive behavior (p < 0.001; Fig 2A). However, only Bz + PTX therapy completely abrogated this alteration, as noticed after 20 minutes of the test (p < 0.001; Fig 2A). Compared with NI controls, Veh-treated infected mice showed a reduced number of entries and time spent in the open arms in the EPMT (p < 0.001), suggestive of anxiety-like behavior. After the three therapeutic schemes, the number of entries in the open arms remained reduced (Fig 2B). The time spent in the open arms of the labyrinth remained reduced in Bz-treated mice, while it was partially restored by PTX therapy (p < 0.001). Moreover, Bz + PTX therapy completely restored the time spent in the open arms, compared to Veh- and Bz-treated infected mice (p < 0.001; Fig 2B). At 150 dpi, Veh-treated infected mice showed increased immobility time in TST (p < 0.001). Significantly, the three therapeutic schemes reduced the time of immobility (p < 0.001), now resembling NI controls (Fig 2C), supporting that in chronically infected mice, depressive-like behavior was inhibited by Bz and PTX administered as mono and combined therapies.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Behavioral changes in chronically Trypanosoma cruzi-infected mice are improved by PTX, Bz, and Bz + PTX therapies. C57BL/6 mice were infected with 100 bt of the Colombian T. cruzi strain. Mice received PTX, Bz, or Bz + PTX therapies daily from 120-150 dpi. At 150 dpi, mice were analyzed using standardized behavioral tests. (A) Number of hidden spheres in a 30-minute marble burying test (MBT). (B) Number of entries and time spent in the open arms in an elevated plus maze test (EPMT). (C) Time of immobility in the tail-suspension test (TST). Color code: Grey bars indicate non-infected controls (NI), blue bars show Veh-treated, green bars indicate PTX-treated, pink bars show Bz-treated, and yellow bars indicate Bz + PTX-treated infected mice. The data are shown as the means ± SD. *, p < 0.05 and ***, p < 0.001, T. cruzi-infected compared with NI controls. ^###, p < 0.001, treatments compared with Veh-treated T. cruzi-infected mice. ^ψψψ, p < 0.001, Bz + PTX-treated compared with Bz-treated T. cruzi-infected mice.

https://doi.org/10.1371/journal.pone.0334708.g002

Memory deficits are sequentially established in chronically T. cruzi-infected mice (novel object recognition, spatial habituation, and aversive shock evoked memory) [17,18]. At 150 dpi, NORT revealed a deficit of object recognition in Veh-treated infected mice (p < 0.001), a condition inhibited by the three therapeutic schemes (Fig 3A). At 150 dpi, Veh-treated infected C57BL/6 mice presented disturbed spatial habituation memory at the OFT2 (p < 0.05), compared with age-matched NI controls, a trait completely abrogated by Bz therapy (p < 0.05), while partially ameliorated by PTX and Bz + PTX therapies (Fig 3B). Lastly, compared with age-matched NI controls, Veh-treated infected mice showed decreased latency in the training session for learning and memory consolidation at the ASET1 (p < 0.05) and test session for memory recall at ASET2 (p < 0.001) as shown in Fig 3C, findings compatible with memory dysfunction. PTX and Bz + PTX therapies improved the latency in the training session (p < 0.05). Interestingly, the three therapeutic schemes ultimately hampered (p < 0.01) memory impairment in the test session (Fig 3C). Altogether, these therapeutic schemes improved or resolved the behavioral and mnemonic alterations in chronically T. cruzi-infected mice.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Cognitive alterations in chronically Trypanosoma cruzi-infected mice are resolved by PTX, Bz, and Bz + PTX therapies. C57BL/6 mice were infected with 100 bt of the Colombian T. cruzi strain. Mice received Veh, PTX, Bz, or Bz + PTX therapies daily from 120-150 dpi. At 150 dpi, mice were analyzed using standardized cognitive tests. (A) Discrimination index in the novel object recognition memory task (NORT); Discrimination index = time exploring the novel object/ (time exploring the novel object + time exploring the familiar object). (B) Discrimination index in the two-day exposition to the open field test, registering the crossed lines (OFT2); Discrimination index = number of crossed lines day 2/ (number of crossed lines day 1 + number of crossed lines day 2). (C) Latency (s) in the training (ASET1, striped bars) and test (ASET2, plain bars) sessions of the aversive shock-evoked test (ASET). Results representative of two independent experiments. Color code: Grey bars indicate non-infected controls (NI), blue bars show Veh-treated, green bars indicate PTX-treated, pink bars show Bz-treated, and yellow bars indicate Bz + PTX-treated infected mice. The data are shown as the means ± SD. *, p < 0.05 and ***, p < 0.001, T. cruzi-infected compared with NI controls. ^#, p < 0.05, ^##, p < 0.01 and ^###, p < 0.001, treatments compared with Veh-treated T. cruzi-infected mice.

https://doi.org/10.1371/journal.pone.0334708.g003

PTX and Bz monotherapies and combined Bz + PTX therapy ameliorate neurochemical changes in chronically T. cruzi-infected mice

At 150 dpi, increased TBARS levels, revealing lipid peroxidation and ultimately oxidative stress, were detected in the cerebral cortex (p < 0.001) and hippocampus (p < 0.001) of Veh-treated infected mice, compared with NI controls (Fig 4A, B). Administration of Bz (p < 0.01) and Bz + PTX (p < 0.05) partially reduced the TBARS levels in the cerebral cortex (Fig 4A). In contrast, PTX (p < 0.05) and Bz + PTX (p < 0.05) therapies were more effective in the hippocampus areas (Fig 4B).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. PTX, Bz, and Bz + PTX therapies ameliorate oxidative stress in chronically Trypanosoma cruzi-infected mice. C57BL/6 mice were infected with 100 bt of the Colombian strain. Mice received PTX, Bz, or Bz + PTX therapies daily from 120-150 dpi. Oxidative stress, revealed by the detection of TBARS, was detected in the (A) cerebral cortex and (B) hippocampus of the brain tissue. Representative data of two independent experiments. Color code: Grey bars indicate non-infected controls (NI), blue bars show Veh-treated, green bars indicate PTX-treated, pink bars show Bz-treated, and yellow bars indicate Bz + PTX-treated infected mice. The data are shown as the means ± SD. *, p < 0.05, **, p < 0.01 and ***, p < 0.001, T. cruzi-infected compared with NI controls. ^#, p < 0.05 and ^##, p < 0.01, treatments compared with Veh-treated T. cruzi-infected mice.

https://doi.org/10.1371/journal.pone.0334708.g004

Neurochemical changes linked to canonical neurotransmitter and neurotrophin pathways are mainly found in the cerebral cortex of chronically T. cruzi-infected mice [18]. Here, we show increased concentrations of GABA (p < 0.001) and glutamate (p < 0.05) in extracts of the cerebral cortex of chronically Veh-treated T. cruzi-infected C57BL/6 mice, compared with NI controls (Fig 5A, B). At 150 dpi, PTX and Bz + PTX therapies, but not Bz treatment, reduced GABA (p < 0.001) and glutamate (p < 0.01; p < 0.05) concentrations in the cerebral cortex of infected mice (Fig 5A, B). Compared with NI controls, a decreased BDNF mRNA expression was observed in the cerebral cortex (p < 0.05) of Veh-treated infected mice at 150 dpi (Fig 5C). In mice that received the three therapeutic schemes, BDNF expression was partially recovered, with tissues resembling NI controls (Fig 5C).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. PTX- and Bz-based therapies improve the disbalance of neurotransmitters and BDNF in Trypanosoma cruzi-infected mice. C57BL/6 mice were infected with 100 bt of the Colombian T. cruzi strain. Mice received PTX, Bz, or Bz + PTX therapies, daily from 120-150 dpi. (A) Using the cerebral cortex, nuclear magnetic resonance analyzed GABA and (B) glutamate. (C) BDNF mRNA expression in the cerebral cortex. Color code: Grey bars indicate non-infected controls (NI), blue bars show Veh-treated, green bars indicate PTX-treated, pink bars show Bz-treated, and yellow bars indicate Bz + PTX-treated infected mice. The data are shown as the means ± SD. *, p < 0.05 and ***, p < 0.001, T. cruzi-infected compared with NI controls. ^#, p < 0.05, ^##, p < 0.01 and ^###, p < 0.001, treatments compared with Veh-treated T. cruzi-infected mice. ^ψψ, p < 0.01, Bz + PTX-treated compared with Bz-treated T. cruzi-infected mice.

https://doi.org/10.1371/journal.pone.0334708.g005

The systemic inflammatory profile is more efficiently regulated by Bz-based therapies in chronically T. cruzi-infected mice

Next, we questioned the effects of the therapeutic schemes on the systemic inflammatory profile and its regulatory pathways. The kinetic study revealed a peak of serum NO concentrations coinciding with parasitemia (at 42 dpi). After parasitemia control, NO levels were reduced (at 90 dpi), but as the infection progressed, serum NO levels increased again (at 120 and 150 dpi) (Fig 6A), regardless of parasitemia levels, which became negative or intermittent at these last time points (Fig 1A). At 150 dpi, compared with age-matched NI controls, Veh-treated infected mice showed elevated levels of serum NO (Fig 6B). PTX therapy did not impact NO levels, but Bz treatment and, mainly, Bz + PTX therapy reduced serum NO concentrations in chronically T. cruzi-infected mice (Fig 6B, S2 Table). In comparison with NI controls, TNF concentrations in serum were increased in infected mice, at 120 dpi (pre-therapy), and even higher as infection progressed (Veh, post-therapy) (Fig 6C). At 150 dpi, TNF levels in mice receiving PTX, Bz, and Bz + PTX therapies were less elevated than in Veh-treated infected mice. They remained like the cytokine levels detected in infected mice pre-therapy (Fig 6C, S2 Table). Thus, Bz-based therapies are pivotal in regulating the systemic NO and TNF levels in chronically T. cruzi-infected mice.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Serum NO and TNF levels are regulated by Bz-based therapies in chronically Trypanosoma cruzi-infected mice. C57BL/6 mice were infected with 100 bt of the Colombian T. cruzi strain. Sera were collected from 14-150 dpi. Groups of mice received PTX, Bz, or Bz + PTX therapies daily from 120-150 dpi. (A) Kinetic of serum NO concentrations. (B) Serum NO concentrations were measured after therapeutic schemes at 150 dpi. (C) Serum TNF concentrations were determined pre-therapy (120 dpi) and after therapeutic schemes at 150 dpi. Color code: Grey bars indicate non-infected controls (NI), blue bars show not-treated or Veh-treated, green bars indicate PTX-treated, pink bars show Bz-treated, and yellow bars indicate Bz + PTX-treated infected mice. The data are shown as the means ± SD. ^ϕϕ, p < 0.01, 90 dpi compared with 42 dpi. *, p < 0.05, **, p < 0.01 and ***, p < 0.001, T. cruzi-infected compared with NI controls. ^#, p < 0.05 and ^##, p < 0.01, treatments compared with Veh-treated T. cruzi-infected mice. ^ψ, p < 0.05, Bz + PTX-treated compared with Bz-treated T. cruzi-infected mice. ^&, p < 0.05, Bz + PTX-treated compared with PTX-treated T. cruzi-infected mice. ^δδ, p < 0.01, T. cruzi-infected mice pre-therapy (at 120 dpi) compared with Veh-treated T. cruzi-infected mice (at 150 dpi).

https://doi.org/10.1371/journal.pone.0334708.g006

Previous work showed the absence of detectable cytokine gene expression in the CNS but elevated concentrations of cytokines in the serum of chronically T. cruzi-infected C57BL/6 mice [18]. Here, we showed increased NO and TNF serum levels 120 dpi. To further explore the idea that the Bz-based therapeutic strategies beneficially act on systemic inflammation, we compared the inflammatory profiles in serum obtained in our present study with findings on cytokine gene expression in the heart tissue of Veh-, Bz- and Bz + PTX-treated mice as a complementary readout of systemic inflammation. For that, we reanalyzed our previous data [37,43], focusing on cytokines associated with inflammation (IL-6, IL-10, MCP-1/CCL2, IFN-γ, TNF, and IL-12p70), normalizing the data using serum or heart tissue samples of matched NI controls. Compared with Veh-treated infected mice (Fig 7A), serum of Bz- and, mainly, Bz + PTX-treated mice showed reduced levels of NO, TNF, and IL-6 (Fig 7B, C). Heart tissue of Veh-treated infected mice showed an elevated fold increase of all analyzed cytokines (Fig 7A). The tissue of Bz-treated infected mice showed reduced expression of all inflammatory cytokine genes, while the regulatory IL-10 cytokine was upregulated (Fig 7B). Crucially, Bz + PTX therapy reduced the expression of all analyzed cytokine genes (Fig 7C). Thus, multiparameter analysis supports that the combined Bz + PTX therapy more efficiently dulls the systemic inflammatory profile detected in chronically T. cruzi-infected mice.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. Regulatory role of Bz and Bz + PTX therapies on systemic inflammatory profile in chronically infected mice. C57BL/6 mice were infected with 100 bt of the Colombian T. cruzi strain. Mice received PTX, Bz, or Bz + PTX therapies daily from 120−150 dpi. At 150 dpi, blood was collected, NO concentrations were measured, and cytokines were evaluated in serum by CBA. To obtain data on the expression of cytokine genes in the heart tissue of Bz- and Bz + PTX-treated mice, previously published data [37,43] were reanalyzed, focusing on cytokines associated with inflammation (IL-6, IL-10, MCP-1/CCL2, IFN-γ, TNF, and IL-12p70). All data were normalized using serum or tissue samples of matched NI controls of each set of experiments to allow comparative analysis, and data are expressed as a fold increase. (A) Veh-treated, (B) Bz-treated, and (C) Bz + PTX-treated T. cruzi-infected mice. Dotted lines were used to facilitate data visualization.

https://doi.org/10.1371/journal.pone.0334708.g007

Our investigation into the systemic inflammatory profiles of chronically infected mice has been furthered by evaluating the modulatory effects of Bz-based therapy on cytokine expression (Fig 7). The in-depth transcriptomic standing of the interaction between chronic T. cruzi infection and the systemic regulation of miRNAs, particularly those implicated in inflammation and neurodegenerative processes, expands upon the influence of Bz-based therapeutic regimens. While we acknowledge that accessing the miRNA transcriptome in brain tissues and serum samples of mice poses a challenge, it emphasizes the significance of analyzing the heart tissue for a comprehensive miRNA profile as a marker of systemic processes. Hence, we have reexamined immune response-related mRNA and miRNA transcriptome data originally detailed by Farani et al. [43,44]. Differentially expressed mRNA (DEGs) and miRNAs (DEMs) were identified through comparative analysis between the Veh-treated group and NI controls, offering insights into potential biomarkers for further studies. Regarding the immune response-related genes identified, we explored the activated canonical pathways, where the Veh-treated group includes ‘Th1 and Th2 Activation Pathway’ and ‘Th1 Pathway’, suggesting an immune response skewed towards T helper cell activation and other pathways such as ‘Neuromodulation Signaling Pathway’ is also significant, pointing to ongoing inflammation and potential neurological modulation in mice infected with T. cruzi (S4A Fig). The Bz-treated group exhibited activation of ‘Th1 and Th2 Activation Pathway’, albeit less significantly compared to the Veh-treated group (S4B Fig) and finally, Bz + PTX-treated group revealed remarkable inhibition of pathways related to ‘Th2 Activation’ and ‘Th1 Pathway’, denoting a potential therapeutic effect of the combined treatment in modulating these immune responses, and activation of ‘IL-12 Signaling and Production in Macrophages’ pathway, suggesting a specific immune regulatory impact due to the combined therapy (S4C Fig).

We identified 24 DEMs associated with inflammatory and neurodegenerative pathways that were altered in Veh-treated infected mice (S3 Table), 9 of those were restored after Bz therapy – 3 up- and 6 downregulated in T. cruzi-infected mice (Fig 8A, S4 Table), and other 9 were restored after Bz + PTX therapy, – 5 up- and 4 downregulated in T. cruzi-infected mice (Fig 8D, S5 Table). DEMs ranging from −0.7 to +1.3-fold change underwent functional pathway analysis using Ingenuity Pathway Analysis Software. The elucidation of putative targets for each regulated miRNA through target filter analysis and the core analysis brought to light relevant regulatory pathways. The Top Canonical Pathways impacted by Bz therapy include “organismal injury and abnormalities,” “reproductive system disease” and “neurological disease” (Fig 8B). In contrast, the Disease and Disorders showed alteration of pathways related to “neurological disease” and “psychological disorders” featured prominently in terms of miRNA involvement (Fig 8C). After Bz + PTX treatment, the altered Top Canonical Pathways primarily concerned with “organismal injury and abnormalities,” “gene expression,” and “neurological disease” (Fig 8E). Additionally, Bz + PTX treatment affected regulatory pathways related to Disease and Disorders, wherein “organismal injury and abnormalities,” “reproductive system disease,” and “cancer” were enriched (Fig 8F).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 8. Characterization of miRNA expression profiles and their potential targets upon Bz and Bz + PTX therapy in chronically infected mice. Relative expression was expressed as fold change (2^-ΔΔCt). miRNAs with a fold change greater than 1.3 in the vehicle group and at the 0.75–1.3 range (orange striped area) in the Bz and Bz + PTX therapy groups were considered to have a restored expression. They were selected for further analysis on IPA software. (A) The expression of miRNAs was altered in the infected group and an expression was restored after Bz therapy. (B) Top canonical activated pathways in the infected group with a restored expression upon Bz therapy. (C) Diseases and disorders pathway analysis. (D) The expression of miRNAs was altered in the infected group and restored upon Bz + PTX therapy. (E) Top canonical activated pathways in the infected group with a restored expression upon Bz therapy. (F) Diseases and disorders pathway analysis. Significant intervals of enrichment are shown on the right of the bars.

https://doi.org/10.1371/journal.pone.0334708.g008

In Veh-treated mice, we selected key biological pathways proposed to underpin CD’s pathogenesis and clinical outcome, including the ‘Inflammation of the central nervous system’, ‘Neuroinflammation signaling pathways’, and ‘Psychological disorders’. A group of DEMs (miR-223-3p, miR-21-5p, miR-182-5p, miR-155-5p, miR-148a-3p, miR-146a-5p, miR-133a-5p, miR-132-3p, let-7d-3p) was found to be linked to the aforementioned pathways (S5 Fig), and identified to directly regulate crucial genes such as NOS2/iNOS (Nitric oxide synthase 2/Inducible nitric oxide synthase) and TLR4 (Toll-like receptor 4). In contrast, Bz-treated groups showed miR-146a-5p and miR-223-3p as key miRNAs in restoring the regulatory networks on ‘Inflammation of central nervous system’ and ‘Psychological disorders’ pathways (S6 Fig). miR-146a-5 p’s regulation of TLR4, CCL5 (Chemokine ligand 5), and SERPINA3 (Serpin peptidase inhibitor, Clade A, Member 3), while miR-223-3p targets PLG (Plasminogen), IL6 (Interleukin-6) highlighting their significance in these processes. The combined Bz + PTX therapy not only reinforced the role of miR-146a-5p targeting IL1R1 (Interleukin-1 receptor 1), IL10 (Interleukin 10), CRP (C-reactive protein) and CXCL8 (CXC-chemokine 8 ligand), but also brought to light the involvement of miR-21-5p, miR-132-3p, and miR-148a-3p in modulating genes related to ‘Neurodegenerative diseases’ and ‘Central nervous system inflammation’ (S7 Fig). As an example, miR-21-5p was observed to regulate pro-inflammatory cytokines such as IL12A (Interleukin 12A) and TNF (Tumor necrosis factor). Further, miR-132-3p directly regulates the genes GRM3 (Glutamate metabotropic receptor 3) and MMP9 (Matrix metalloproteinase 9), while miR-148a-3p directly regulates the channel KCNJ6 (Potassium inwardly rectifying channel subfamily J member 6) and PD1A3 (protein disulfide-isomerase A3).

Combined Bz + PTX therapy ameliorates behavioral and cognitive changes regulating the neurochemical changes and systemic inflammatory profile

Next, we ran a multivariate analysis to identify an association between processes and visualize the profile of the groups that received therapies. Initially, we performed a correlation analysis using the Spearman correlation test for data of parasitemia (T. cruzi), behavioral and cognitive changes, neurochemical alterations (TBARS, GABA and glutamate, BDNF), and serum inflammatory profiles (TNF, NO), comparing NI controls, Veh-treated, PTX-, Bz- and Bz + PTX-treated T. cruzi-infected mice. The correlation analysis (Figs 9A and S8) demonstrated a direct association between T. cruzi infection and oxidative stress in the cerebral cortex (TBARS-co) and hippocampus (TBARS-h) and serum NO levels (NO). Further, serum TNF levels were correlated with serum NO levels, the neurotransmitters GABA, and glutamate (GLU) concentrations in the cerebral cortex. Crucially, serum NO levels were directly correlated with oxidative stress (TBARS-co) and GABA concentrations in the cerebral cortex while inversely correlated with BDNF expression in the brain cortex. Regarding behavioral changes, the loss of innate compulsive behavior (MBT – decreased number of marble burying), depressive-like behavior (TST – increased immobility time), and anxiety (EPMT-entries and EPMT-time – reduced number of entries and time in open arms) were associated with the serum NO levels and oxidative stress (TBARS-co and TBARS-h). Likewise, these changes, except for TST, showed a negative correlation with parasitemia (T. cruzi); that is, at higher levels of parasitemia, there is a reduction in the evaluated behaviors (decrease in the number of marble burials and a reduction in the number of entries and time in open arms, indicative of loss of innate compulsive behavior and anxiety as parasitemia increases. Further, learning and memory consolidation at the ASET (ASET1) was inversely related to GABA and glutamate concentrations in the cerebral cortex (GABA-co; GLU-co), while memory recall (ASET2) was inversely related to the oxidative stress in the hippocampus (TBARS-h). Interestingly, except for the readout of the memory of object recognition (NORT), the readouts of other behavioral tests were directly or indirectly related, supporting an association between loss of innate compulsive behavior, anxiety, depressive-like behavior, and habituation and aversive memory deficits (Fig 9A). The correlation values and significance levels are provided in S6 and S7 Tables, respectively.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 9. Relationship between parasitemia, neurochemical and immunological stressors and behavioral and cognitive changes in chronically infected mice. (A) A circle represents a significant correlation, and the intensity of the color indicates the strength of this correlation, which is represented in blue (direct correlation, 0 to +1) or red (inverse correlation, 0 to −1). The “X” indicates the absence of a significant correlation between the variables. (B-C) The three first principal components explain 71.15% of the data variance, and the 2D projection of (B) PC1-PC2 and (C) PC1 and PC3 indicate the association between parasite load (T. cruzi), serum TNF and NO levels, and neurochemical changes in the CNS, and the behavioral and cognitive changes. (D) The 3D projection, with the three first principal components, shows that NI control group and Bz + PTX-treated T. cruzi-infected group show similar behavior and demonstrate the differences between the Veh-treated infected group and NI cluster, and Bz + PTX-treated and Veh-treated cluster.

https://doi.org/10.1371/journal.pone.0334708.g009

Lastly, to visualize the similarities between groups and identify clusters of differentiation, we performed a PCA using data of behavioral and cognitive changes and immunological and neurochemical abnormalities, comparing NI controls, Veh-treated, and groups treated with the three therapeutic schemes (PTX, Bz, and Bz + PTX). The three-first principal components were used to explain 71.15% of the data variance. As shown in Fig 9B, the 2D projections of the samples using the two-first components, which explain 62.18% of the variance, confirmed the association between infection (T. cruzi) levels, increased serum NO levels, and increased TBARS levels in the cerebral cortex and hippocampus. The increase in these variables is correlated with decreased BDNF expression in the cortex and reduced entries into the open arms of the EPMT (anxiety). At the same time, the presence of depressive-like behavior was shown to be related to an increase in the TBARS levels in the cerebral cortex. Also, the increased concentrations of the neurotransmitters GABA and glutamate in the brain cortex were correlated with the increased serum TNF concentrations, which correlated with depressive-like behavior. Moreover, the increase in GABA and glutamate in the brain tissue and TNF in serum are inversely correlated with the decrease in the number of hidden marbles in the MBT, the reduction in the time mice spent in the open arms of the EPMT and the decrease latency in aversive shock evoked test (ASET1) (Fig 9B), sustaining the association of serum TNF concentrations and neurotransmitter changes in the brain tissue with loss of innate compulsive behavior, anxiety and consolidation aversive memory. Additionally, increased parasitemia showed a close relationship with increased TNF and NO serum levels, an increase in TBARS and neurotransmitter levels in the brain cortex, and depressive-like behavior in the group of Veh-treated infected mice (Fig 9C). PC1 and PC3 projection showed similar profiles as in PC1 and PC2 (Fig 9B, C). Moreover, 3D projections, with the three first principal components explaining 71.15% of the variance, confirmed that well-defined clusters were observed separating the Veh-treated infected group, showing completely different profiles than the other groups. At the same time, the NI control group, Bz-treated and Bz + PTX-treated groups were visualized closely, regarding parameters such as decrease in anxiety (EPMT), improvement in innate compulsive behavior (MBT), increase in the consolidation of the aversive memory (ASET1), and increase in the BNDF expression in the cerebral cortex. Indeed, our PCA analysis supports a closer proximity between the Bz + PTX-treated infected group and NI control group (Fig 9D).