The experiments were performed on male C57BL/6 mice, 7–13 weeks, resistance prototype to T. cruzi infection, weighing between 20 and 30 g from TECPAR (Paraná Technology Institute), PR, Brazil. The choice of exclusively male mice for the study is based on previous findings indicating gender dimorphism in nociception and no effect of glial inhibitors in females in this specie (32). Pathogen-free mice were housed in standard clear plastic cages appended to a ventilated rack (Alesco Indústria e Comércio LTDA, Monte Mor, Sao Paulo, Brazil) with free access to water and food, a light/dark cycle of 12/12 h, and controlled temperature. Shaving and feed provision for the animals was always controlled by a sterilization process. The mice were separated into a maximum of five per cage and were maintained in the vivarium of the Department of Pathology at the State University of Londrina for at least 2 days before the experiments. Mice were used only once and were acclimatized to the testing room at least 1 hour before the experiments, which was conducted during the light cycle. At the end of the experiments, mice were anesthetized with isoflurane 5% only once by inhalation overdose, and terminally euthanized by cervical dislocation followed by decapitation. Animals were monitored regarding welfare-related assessment before, during, and after the experiments. Clinical signs of severe illness, including changes in body weight, erection of the back hairs (reflecting irritation or agitation), diarrhea, lethargy, rapid/shallow breathing, and paralysis, were also properly recorded. Animals presenting clinical signs of severe disease before the end of the model were immediately euthanized using isoflurane 5% followed by cervical dislocation as per the guidelines of the Ethics Committee on Animal Use (CEUA) of the State University of Londrina. The euthanasia process used in the study for sample collection also followed the protocol of anesthesia of animals with isoflurane 5% followed by cervical dislocation according to the CEUA guidelines. Efforts were maintained to narrowly minimize the total number of animals used and their discomfort or suffering.

Isoflurane was obtained from Abbot Park (Lake Bluff, IL, USA). L-2-aminoadipic acid (α-aminoadipate; A7275), minocycline hydrochloride (M9511), ammonium pyrrolidine dythiocarbamate (PDTC; P8765), dimethyl sulfoxide (DMSO; D8418), and tween 80 (P5188) were purchased from Sigma-Aldrich (St. Louis, MO, USA). Mouse neutralizing antibody anti-CX₃CL1 (AF472) was obtained from R&D System (Minneapolis, MN, USA). Etanercept was obtained from Wyeth (Enbrel^®, São Paulo, SP, Brazil). IL-1 receptor antagonist was obtained from the National Institute for Biological Standards and Control (IL-1ra; NIBSC, South Mims, Hertfordshire, UK). Vehicle was obtained from LBS Laborasa (sodium chloride 0.9%; São Paulo, SP, Brazil). α-aminoadipate, minocycline, PDTC, etanercept, IL-1ra, antibody anti-CX₃CL1, and vehicle were injected only once by the intrathecal route (L₄-L₆ segment) in a final volume of 5 μL to obtain a local effect. We avoid repeated intrathecal injections because they cause increased nociceptive response via prostanoid production (33) and could, thus, interfere with the analysis of the present model. Doses of treatments used in the present model were chosen because they have analgesic effects as demonstrated previously (32). Intrathecal injections were performed in animals under a brief period of anesthesia induced by inhalation of isoflurane 5%.

Protozoan parasite T. cruzi (Kinetoplastea; Trypanosomatida; Trypanosomatidae; Trypanosoma) infection was induced using Y strain (TcII) and was maintained by weekly intraperitoneal inoculation in Swiss mice with 2x10⁵ blood trypomastigotes as described previously (34). Briefly, the blood of animals infected anteriorly was obtained by cardiac puncture under anesthesia without the use of anticoagulants. Samples were then centrifuged followed by standing in 37°C, and serum supernatant containing most parasites was centrifuged again for the resuspension of the pellet using appropriate culture and compounds (GIBCO, Grand Island, NY, USA). Trypomastigotes were derived from the supernatant of T. cruzi–infected LLC-Mk2 culture cells (ATCC CCL-7; American Type Culture Collection, Rockville, MD, USA) and grown in appropriate culture (GIBCO, Grand Island, NY, USA). Subconfluent cultures of LLC-Mk2 were infected with 5x10⁶ trypomastigotes. Free parasites were removed after 24 h, and cultures were maintained in medium. Five days after infection, free trypomastigote forms could be found in the cell supernatant, and the infection of animals was performed between the 7th and 8th days afterward by the intraperitoneal route. Blood parasitemia was assessed under standardized conditions by direct microscopic observation at 400x magnification of 50 fields. Counting the number of circulating parasites was performed in a volume of 5 μL of blood collected from the tail vein of infected animals. This result was expressed as the number of parasites per mL^-1 in the blood samples. Blood parasitemia was evaluated every 7 days, starting on the 7th day of infection, and survival rates were monitored daily. Infection outcomes in mice were monitored daily to assess their health conditions during the experimental period.

Mice were infected with T. cruzi (5x10³ infective trypomastigotes) by one intraperitoneal inoculation, and the following parameters at the respective time points evaluated in comparison to uninfected mice are described in sequence. Mechanical hyperalgesia and thermal hyperalgesia were assessed every 2 days for a period of 28 days. Survival of infected animals was recorded daily, and survival rate plotted every 2 days as a percentage. Blood parasite load was monitored every 7 days (7, 14, 21, and 28 days). Blood of the animals was also used to determine TNF-α and IL-1β protein levels on the same days described for the assessment of parasitemia. Temporal profile (7–28 days p.i.) of glial fibrillary acidic protein (Gfap) and ionized calcium-binding adapter molecule 1 (Iba1) mRNA expression or protein levels/staining (Western blot and immunofluorescence) were evaluated in spinal cord samples of uninfected and infected mice. The 7th day was selected because, at this time point, we could evaluate the participation of astrocytes (GFAP) and microglia (Iba-1) at the same time using all three approaches if necessary as per the results section. Next, infected animals were treated with vehicle, glial (α-aminoadipate and minocycline), and NFκB (PDTC) inhibitors by the intrathecal route for the evaluation of Gfap and Iba1 mRNA expression, blood parasitemia, and mechanical and thermal hyperalgesia (1–7 h after the treatments) on day 7 p.i.; survival percentage was also assessed during this period to ensure that no animals died during the spinal treatment protocol, so certifying intrathecal treatments would not be harmful to the animals. We opted to evaluate drug treatment effects during 1–7 h because we have previously standardized this approach to study the acute effect of treatments in ongoing hyperalgesia (32). Posteriorly, the effects of intrathecal treatments described above (most effective dose of each inhibitor) upon NFκB activation (7 h after the treatments) as well as double immunofluorescence analysis using antiglial and -p-NFκB antibodies 7 days p.i. were performed. The evaluation of the temporal profile of Cx3cr1, Tnfα, and Il1β mRNA expression in the spinal cord were conducted 7–28 days after the infection. In another round of experiments, mice were treated by the intrathecal route with vehicle, antibody anti-CX₃CL1, etanercept, and IL-1ra and mechanical and thermal hyperalgesia (1–7 h after the treatments), and parasitemia was investigated at day 7 after the infection; again, survival percentage was assessed during this period to ensure that no animals died during the spinal treatment protocol and that the intrathecal treatments would not be harmful to the animals. After this measurement, infected animals were treated with vehicle, glial (α-aminoadipate and minocycline), and NFκB (PDTC) inhibitors (most effective doses) by the intrathecal route to be evaluated; this time, the Cx3cr1, Tnfα, and Il1β mRNA expression in the spinal cord samples at the 7th day p.i. 7 h after the treatments. The activation of DRG neurons in response to infection alone and co-stimulus with capsaicin were examined using the calcium influx test through confocal microscopy as well as mRNA expression of Trpv1 (transient receptor potential cation channel subfamily V member 1) at the 7th day p.i. The effects of intrathecal treatments with vehicle, glial (α-aminoadipate and minocycline), and NFκB (PDTC) inhibitors (most effective doses) were also evaluated over Cx3cl1, Cx3cr1, Gfap, Tnfα, Il1β, and Cox2 mRNA expression in DRG cells at the day 7 after infection 7 h after the treatments.

Mechanical hyperalgesia was tested in mice as previously described (35). Briefly, in a quiet room, mice were placed in acrylic cages (12 X 10 X 17 cm) with wire grid floors, for at least 30 min before the test for adequate environmental adaptation. Stimulations were performed only when the animals were quiet without exploratory movements or defecation and not resting on their paws. The test consisted of evoking a hind paw flexion reflex with a hand-held force transducer (electronic von Frey esthesiometer; Insight, Ribeirão Preto, SP, Brazil), adapted with a 4.15-mm² (referred to as large probe) contact area polypropylene tip (35). The experimenter applies the probe perpendicularly to the central area of the hind paw with a gradual increase of pressure. The applied pressure to the hind paw surface induces movement of the ankle and knee joints, evoking stretching and/or shortening of leg muscles, promoting a muscle contraction response, which is enough to trigger muscle nociceptive activity (movement-induced hyperalgesia) when the latter is sensitized. The end point was characterized by the removal of the hindlimb followed by clear flinch movements. The measurements were standardized to be always measured on the right paw of the mice. The electronic pressure meter apparatus automatically recorded the intensity of the pressure applied when the paw was withdrawn. The value for the response was an average of three measurements. Only one experimenter started and ended the same test in order to avoid differences in the application of the pressure transducer force. The experimenter was always blinded to the experimental groups. The results are expressed by mechanical threshold in grams (g).

The mensuration of thermal hyperalgesia in mice was conducted as described previously (32). Mice were positioned on the heated surface of a hot plate apparatus isolated by containment of a transparent acrylic material (EFF 361, Insight, Ribeirão Preto, SP, Brazil). The temperature on the surface was always maintained at 55 ± 1°C. The reaction time was registered using a conventional chronometer when the mouse presented the behaviors of licking or flinching one of the hind paws. A time limit of 15 s of maximum latency was defined as a cutoff with the intention of avoiding potential tissue injuries. The assessment of thermal hyperalgesia was performed before and after the infection process. The experimenter was always blinded to the experimental groups. The results were expressed by thermal threshold in seconds (s).

The blood of infected mice was collected in sterile microtubes containing the anticoagulant ethylenediamine tetraacetic acid (EDTA) by cardiac puncture following anesthetization with isoflurane 5%. Samples were then centrifuged for separation of plasma, and the resultant supernatant was separated for the tests. TNF-α and IL-1β levels were determined through an enzyme-linked immunosorbent assay (ELISA) test using paired antibodies following the manufacturer’s instructions (eBioscience Inc.; Thermo Fisher Scientific, Waltham, MA, USA) (36). In summary, 96-well plates were first coated overnight in a refrigerator mixed with an immunoaffinity-purified polyclonal sheep antibody specifically for each cytokine. Plates were then blocked for 2 h. Subsequently, recombinant murine TNF-α and IL-1β standards were added to the plates using a serial dilution protocol for the standard curve. For sample testing, 100 μL of samples in duplicate were added and incubated for 1 h at room temperature followed by the addition of rabbit biotinylated immunoaffinity-purified antibodies anti-TNF-α and anti-IL-1β and incubation for 1 h at room temperature. In the next step, 50 μL of avidin-HRP reagent were added to each well in a dilution of 1:5000 for 30 min of incubation at room temperature. Plates were then washed, and the o-phenylenediamine dihydrochloride substrate was added to the plates at a volume of 200 μL per well to produce measurable signals. Finally, after 15 min, the reaction was interrupted with 1 M H₂SO₄ and measured spectrophotometrically at 450 nm. The detection limit of TNF-α (cat. no. 147325) and IL-1β (cat. no. 147012) kits is 8 picograms (pg)/milliliter (mL). The results were expressed as pg of target cytokine per mL of plasma (32).

RT-qPCR was carried out following the protocol as described previously (32). Spinal cord (L₄-L₆ entire segments) and DRG (bilateral L₄-L₆) samples were collected using TRIzol™ reagent (#15596026, Invitrogen; Thermo Fisher Scientific, Waltham, MA, USA) 7–28 days after the infection, depending on the experiment. The purity of total RNA was measured with a spectrophotometer (MultiSkan GO Microplate Spectrophotometer, ThermoScientific, Vantaa, Finland), and the wavelength absorption ratio (260/280 nm) was between 1.8 and 2.0 for all preparations. Reverse transcription of total RNA to cDNA and qPCR were performed using the Go Taq^® 2-Step RT-qPCR system (Promega Corporation, Madison, WI, USA) according to the manufacturer’s guidelines. The relative gene expression was measured using the comparative 2^-(ΔΔCq) method. The expression of β-actin RNA was used as a reference gene to normalize data. The mouse primer pairs used were as follows:

Western blot analyses of the spinal cords were processed according to the previous study by our group (32). L₄–L₆ entire segments of spinal cord were accurately dissected at the 7th day after the infection, and the whole samples were homogenized in radioimmunoprecipitation assay (RIPA) buffer containing a protease and phosphatase inhibitor cocktail (100X, #5872S, Cell Signaling Technology, Danvers, MA, USA). The lysates were then homogenized and centrifuged. The protein extracts were separated by SDS-PAGE and transferred onto a nitrocellulose membrane (GE Healthcare-Amersham, Pittsburgh, PA, USA). Membranes were then incubated in blocking buffer 95% nonfat milk in Tris-buffered saline with Tween 20 or 1% bovine serum albumin (BSA) for different times for each antibody at 4°C in the presence of the primary antibody. β-actin and GFAP were purchased from Cell Signaling Technology (Danvers, MA, USA). Iba-1 and the secondary antibody were purchased from Thermo Fischer Scientific (Waltham, MA, USA). Catalog numbers of products are indicated below. The antibodies and dilutions used were β-actin (#4970, Cell Signaling Technology, 1:1000) and blocked with 5% BSA; GFAP (#12389, Cell Signaling Technology, 1:10,000) on 8% gel and blocked with 5% BSA; and Iba-1 (#16-20001, Wako Chemicals, 1:800) on 15% gel and blocked with 5% BSA. The molecular masses of protein were confirmed by Precision Plus Protein Standards (Bio-Rad, Hercules, CA, USA). After washing in tris buffered saline (TBS) with Tween 20, the membrane was incubated with secondary antibody (peroxidase-conjugated AffiniPure goat anti-rabbit IgG (H+L), #111-035-003, Jackson Immuno Research, 1:5000) on 5% BSA in TBS-T for 2 h at room temperature. Protein was visualized by chemiluminescence with an enhanced chemiluminescence ECL detection reagent (Luminata™ Forte, Millipore, USA). The membranes were stripped and re-probed with an antibody against β-actin for use as a loading control in addition to loading the same amount of protein in the case of Iba-1 protein. For GFAP protein analysis, a separate membrane using an equivalent mirror gel was prepared for the quantification of β-actin because the chemiluminescent signal for GFAP was still detected after the stripping phase as the stripping buffer failed to completely remove the GFAP antibody and ECL reagent from the membrane in the tested dilutions (dilution between 1:1000 and 1:10,000). Results are presented as the mean of six animals per group and compare noninfected and T. cruzi–infected groups at each time point (7–14 days) because experiments were performed separately. The images and analyses were performed in Image Lab 6.1 software (Bio-Rad Laboratories).

The spinal cord immunofluorescence assay was performed at the 7th day of infection. For this, mice were perfused through the ascending aorta with phosphate buffered saline (PBS) followed by 4% of paraformaldehyde (PFA), and L₄–L₆ segments of the spinal cord were accurately dissected out and postfixed in PFA 4% for 24 h. After this period, samples were replaced with 30% saccharose solution for 3 additional days. The spinal cord segments were then washed with PBS and embedded in optimum cutting temperature (O.C.T.) using Tissue-Tek^® reagent (Sakura^® Finetek USA, Torrance, CA), and 10-micrometers (μm) sections were cut in a cryostat (CM1520, Leica Biosystem, Richmond, IL, USA) and processed for immunofluorescence (four slides per mouse/four animals per group). All the sections were initially blocked with a buffer solution (500 μL per slide containing PBS plus 0.1% tween 20 plus 5% BSA) for 2 h at room temperature and subsequently incubated overnight at -4°C with a solution containing primary antibodies against the target protein. Next, a new incubation with secondary antibodies was performed for 1 h at room temperature. For double immunofluorescence, a mixture of primary antibodies for target proteins as well as secondary conjugated antibodies following the same sequence was performed. GFAP (#180063; 1:500 dilution; Invitrogen, Life Technologies, Carlsbad, CA, USA), Iba-1 (#PA5-27436; 1:500 dilution, Invitrogen, Life Technologies, Carlsbad, CA, USA), and phosphorylated NFκB p65 subunit (sc-136,548, 1:200 dilution, Santa Cruz Biotechnology, Dallas, TX, USA) primary antibodies, and Alexa Fluor 488 (#A-110088, 1:1000 dilution, Thermo Fischer Scientific, Waltham, MA, USA) and Alexa Fluor 647 (#A28181, 1:1000 dilution, Thermo Fischer Scientific, Waltham, MA, USA) secondary antibodies were used in the study. The slide assembly was conducted using ProLong™ Gold Antifade Mountant with 4’,6-diamidino-2-phenylindole, dihydrochloride (DAPI) melting media (#P36931, Thermo Fischer Scientific, Waltham, MA, USA). Analysis using slides with secondary antibodies alone was conducted in parallel with the controls to ensure that unspecified staining did not occur. Immunofluorescence analyses were performed in the dorsal horn of the spinal cord at a magnification of 20x. The images and analysis were performed using a confocal microscope (TSC SP8, Leica Microsystems, Mannheim, Germany) (32, 37). Results of the fluorescence quantification in the samples (control noninfected versus 7-day infected animals) are presented as the mean of four animals per group.

The evaluation of NFκB activation in spinal cord samples was performed following the protocol as described previously (32). For this objective, spinal cord samples (L₄–L₆ entire segments) were collected at the 7th day after the infection and the whole sample homogenized in ice-cold lysis buffer (Cell Signaling Technology, Beverly, MA, USA). The homogenates were centrifuged for 10 min (16,000 g and 4°C), and the resultant supernatants used to assess the levels of total and phosphorylated (p) NFκB p65 subunits by ELISA using two PathScan kits (Cell Signaling Technology, Beverly, MA, USA) specific for total and phosphorylated forms of NFκB p65 according to the manufacturer’s instructions. The test indicates the proportion between total NF-kB and p-NF-kB in the analyzed samples. A low ratio indicates a great amount of p-NF-kB relative to total NF-kB, which indicates NF-kB activation. On the other hand, a high ratio indicates lessened activation because there is less p-NF-kB relative to total NF-kB. The results were expressed as a total-p65/phospho-p65 ratio measured spectrophotometrically (MultiSkan GO Microplate Spectrophotometer, ThermoScientific, Vantaa, Finland).

The dissection and culture of DRG neurons for calcium imaging was performed as previously described (38). Bilateral DRGs (six mice per group) were dissected at the 7th day p.i. into neurobasal‐A medium (Life Technologies, Thermo Fisher Scientific) and dissociated in collagenase A (1 mg·ml−1)/dispase II (2.4 U·ml−1; RocheApplied Sciences, Indianapolis, IN, USA) in HEPES‐buffered saline (Millipore Sigma) for 70 min at 37°C. After trituration with glass Pasteur pipettes of decreasing size, DRG cells were centrifuged over a 10% BSA gradient and plated on laminin‐coated cell culture dishes. DRGs were then loaded with 1.2 μM of Fluo‐4AM in neurobasal‐A medium, incubated for 30 min at 37°C, washed with HBSS, and imaged in a Confocal Microscope (TCS SP8, Leica Microsystems, Mannheim, Germany). To assess TRPV1 activation, DRG plates were recorded for 6 min, which was divided into 2 min of initial reading (0‐s mark, baseline values), followed by stimulation with TRPV1 agonist capsaicin for 2 min at the 120‐s mark (1 μM) and KCl for 2 min at the 240‐s mark (40 mM, activates all neurons). Calcium flux was analyzed from the mean fluorescence intensity measured with the LAS X Software (Leica Microsystems, Mannheim, Germany).

Results are presented as means ± SEM of measurements made on four to six mice in each group, depending on the analysis, per experiment, and are representative of two independent experiments. ANOVA (two-way analysis of variance) was always used for comparison between groups and different doses when responses were measured at varied times after the parasite or treatments and/or stimulus injection. Analyzed factors were treatments, time, and time vs. treatment interaction, and when interaction was significant, one-way ANOVA preceded by Tukey’s post hoc was performed for each time point. Differences between responses were evaluated by one-way ANOVA followed by Tukey’s post hoc for data of a single time point. Differences between two groups were analyzed by t-test. Statistical differences were considered significant when P < 0.05.

Mice were infected with T. cruzi and mechanical hyperalgesia, thermal hyperalgesia, blood parasitemia, survival, and TNF-α and IL-1β levels were evaluated from 2 to 28 days p.i. ( Figure 1 ). Increased mechanical and thermal hyperalgesia in infected mice were detected from the 2nd to the 28th days p.i. ( Figures 1A, B ). The peak of blood parasitemia occurred at the 7th day p.i. with a gradual decrease in subsequent days, and at the 28th day, the parasites were no longer detected in the blood ( Figure 1C ). Interestingly, in infected mice, the hyperalgesia persists even with the decline of parasites in the blood ( Figures 1A, B ), suggesting that not only the parasites per se are responsible for the pain behavior, but possibly plastic changes are involved in the prolonged hyperalgesia. No animals died during the experiment due to T. cruzi infection, which is an expected outcome due to the use of a nonlethal parasite load for the C57BL/6 mouse strain ( Figure 1D ). TNF-α plasma levels were not significant at the 7th day p.i. ( Figure 1E ). TNF-α plasma levels were significant at the 14th day with a gradual reduction from the 14th day onward; however, they remained significant until the 28th day p.i. compared to uninfected animals ( Figure 1E ). IL-1β plasma levels were significant at the 14th day without significant levels in the other time points ( Figure 1F ). Altogether, these results indicate that T. cruzi infection at the tested dose induces chronic pain. However, factors other than solely blood parasite load and circulating levels of TNF-α and IL-1β might explain T. cruzi infection-induced pain overtime.

We evaluated whether T. cruzi infection activates glial cells in the spinal cord of mice. First, we determined the temporal profile of Gfap and Iba1 mRNA expression (7–28 days p.i.; Figures 2A–F ), which are used as markers of astrocytes and microglial activation in the spinal cord, respectively. The activation profile of these two cells did not occur in a similar manner. Gfap mRNA presented a biphasic increase, and for Iba1 mRNA time-dependent upregulation was observed until the 21st day p.i., when it reached its peak ( Figures 2A, F ). At the 7th day, infected mice presented increased Gfap and Iba1 expression when compared to uninfected mice. However, thenceforward Gfap expression returned to the control level at days 14 and 21 p.i. and increased again at day 28 ( Figure 2A ). Instead, Iba1 mRNA expression presented a sustained upregulation until the 21st day p.i., decreasing and ceasing to be significant compared to uninfected animals at day 28 p.i. ( Figure 2F ). A time course of GFAP and Iba-1 protein levels was also performed by Western blot analysis ( Supplementary Figure S1 ), which showed that GFAP was significantly upregulated in the infected group at the 7th day ( Figures 2 B–G ). GFAP was also upregulated at days 21 and 28, thus, partially aligning with the mRNA data ( Supplementary Figure S1A ). A contrast between mRNA and protein data for GFAP was observed at the 21st day in which mRNA was not induced, but protein was increased ( Supplementary Figure S1A ). Iba-1 protein was increased in the infection group at all time points ( Supplementary Figure S1B ) contrasting with the mRNA data in which, at the 28th day, no upregulation was observed in the infected group ( Figure 2F ). An immunofluorescence assay in spinal cord samples for GFAP and Iba-1 was also conducted and showed a classical morphology of activated astrocytes and microglia in infected animals 7 days p.i., which was not observed in uninfected animals ( Figures 2C–E, H–J , respectively). Time-response analysis (7–28) of GFAP and Iba-1 immunofluorescence detected significant activation of both glial cells in infected animals when compared to those uninfected at all evaluated times ( Supplementary Figures S2A, B ). Considering that concomitant activation of astrocytes and microglia was observed at the 7th day p.i. for all tests used (RT-qPCR, Western blot, and immunofluorescence), this time point was chosen for the next experiments to focus on the contribution of spinal cord glial cells to T. cruzi infection-induced pain in mice. These methods (RT-qPCR, Western blot, and immunofluorescence) quantitate different parameters (mRNA vs. protein), and the methods’ sensibilities are different, which might explain the variations observed and that mRNA and protein data did not exactly align in the same way.

To investigate whether inhibiting spinal cord glial cells reduces T. cruzi–induced hyperalgesia, mice were treated by the intrathecal route with vehicle (saline); a selective astrocyte inhibitor, α-aminoadipate (30–100 nmol); or a selective microglial inhibitor, minocycline (50–150 μg), at day 7 after T. cruzi infection, and mechanical and thermal hyperalgesia were measured 1–7 h after the treatment ( Figure 3 ). To validate α-aminoadipate and minocycline effects, mRNA expression of Gfap and Iba1 were also evaluated at the 7th day p.i. The two doses of α-aminoadipate inhibited Gfap mRNA expression; on the other hand, only the highest dose of minocycline inhibited Iba1 mRNA expression ( Figures 3A–D ). The doses of 30 and 100 nmol of α-aminoadipate inhibited mechanical hyperalgesia for up to 7 h after the treatment; however, the effect of the dose of 100 nmol was significantly higher than 30 nm between 3 and 7 h ( Figure 3B ). On thermal hyperalgesia, both doses of α-aminoadipate presented similar activity and inhibited thermal sensitivity at 5 and 7 h after treatment ( Figure 3C ). A similar profile was observed by the treatment with the microglia selective inhibitor, minocycline, regarding mechanical ( Figure 3E ) and thermal ( Figure 3F ) hyperalgesia. The fact that the lowest dose of minocycline did not inhibit the mRNA expression of Iba-1 in the spinal cord but inhibited hyperalgesia can be explained by the fact that Iba-1 or even GFAP are used as markers of glial cell activation and are not necessarily activity measures. Further, α-aminoadipate and minocycline are inhibitors of glial cell metabolism (39, 40), and it is reasonable that a higher dose will have faster activity than a lower dose as well as that makers of activation might change after actual downregulation of activity. Intrathecal treatments with the tested doses of glial inhibitors did not affect blood parasitemia ( Figure 3G ), and no animal died due to infection or treatment protocols ( Figure 3H ). These data corroborate the results presented in Figure 2 and reinforce the concept of a role for spinal cord astrocytes and microglia in T. cruzi–induced hyperalgesia.

The participation of the transcription factor NFκB in T. cruzi–induced spinal cord glial activation and hyperalgesia was evaluated. Mice were treated by the intrathecal route with vehicle (saline) or NFκB inhibitor PDTC (300 μg) at day 7 after T. cruzi infection, and mRNA expression of glial markers and mechanical and thermal hyperalgesia were measured. Additionally, the effects of glial inhibitors upon spinal cord NFκB activation and double staining using the phosphorylated (p) NFκB p65 subunit and glial markers in spinal cord samples were also performed ( Figure 4 ). PDTC inhibited the increased expression of Gfap and Iba1 in infected animals ( Figure 4A ) as well as the mechanical and thermal hyperalgesia at all time points evaluated (1–7 h) upon T. cruzi infection ( Figures 4B, C ). PDTC treatment did not affect blood parasitemia ( Figure 4D ). Furthermore, no animal died due to infection or treatment protocols ( Figure 4E ). Extending these experiments, we notice that the most effective doses of glial inhibitors (100 nmol of α-aminoadipate and 150 μg of minocycline) inhibited T. cruzi–induced NFκB activation in the spinal cord (demonstrated here by the ratio between total p65 subunits per phosphorylated p65 subunit), reaching similar levels of inhibition as positive control PDTC ( Figure 4F ). Finally, using a confocal immunofluorescence assay, we showed the concomitant expression of GFAP ( Figure 4G , two top rows) or Iba-1 ( Figure 4G , two lower rows) and phosphorylated pNFκB p65 ( Figure 4G , last two columns on the right and respective inserts) in infected animals, which was not observed in uninfected mice ( Figure 4G ). As GFAP and Iba-1 are cell membrane molecules, they do not co-localize with NFκB, which has cytoplasmic or, in case of phosphorylation, nuclear localization. The 3-D images (inserts in Figure 4G ) and videos ( Videos S1 and S2 ) allow visualization that part of pNFκB is surrounded by GFAP and Iba-1 staining. Further supporting that NFκB activation aligns with glial cell activation, increased fluorescence intensity of glial markers and pNFκB were detected in infected animals when compared to those uninfected ( Figures 4H, I ). Therefore, these results evidence that T. cruzi infection induces NFκB activation in spinal cord astrocytes and microglia, contributing to hyperalgesia as a result of the disease.

NFκB is a central transcription factor in regulating cytokine production as they are important signaling molecules in neuro-immune regulation and pain (17). The chemokine receptor and cytokine mRNA expression profile were investigated between 7 and 28 days after T. cruzi infection ( Figure 5 ). T. cruzi infection induced a progressive increase in mRNA expression of the chemokine receptor Cx3cr1 ( Figure 5A ) and pro-inflammatory/hyperalgesic cytokines Tnfα ( Figure 5B ) and Il1β ( Figure 5C ) until the 21st day p.i. with a statistical difference compared to those uninfected. After this period, their expression declined and returned close to the baseline values (day 28) ( Figures 5A–C ). These data, together with the glial and NFκB activation (Figs. 2–4), depict spinal cord neuroinflammation in response to T. cruzi infection. Microglia is the main cell expressing CX₃CR1 in the spinal cord, and both astrocytes and microglia are well known sources of cytokines in the spinal cord (17).

In addition to the mRNA expression data, we wanted to ascertain the physiopathological contribution of spinal cord CX₃CL1/CX₃CR1 signaling, TNF-α, and IL-1β in T. cruzi–induced hyperalgesia. For this purpose, mice were treated by the intrathecal route with vehicle (saline), anti-CX₃CL1 antibody (0.25–2.5 mg), soluble decoy receptor (sTNFR2) etanercept (3-10 ng) and Interleukin-1 receptor antagonist (IL-1ra, 30-100 pg) at day 7 after T. cruzi infection. Mechanical and thermal hyperalgesia were measured 1–7 h after the treatment ( Figure 6 ). The dose of 0.25 mg of anti-CX₃CL1 antibody did not affect mechanical or thermal hyperalgesia; however, treatment with the dose of 2.5 mg inhibited mechanical hyperalgesia between 3 and 7 h and thermal hyperalgesia between 5 and 7 h after the treatment ( Figures 6A, B ). The dose of 3 ng of etanercept inhibited only thermal hyperalgesia at 5 h without presenting effects on mechanical hyperalgesia. On the other hand, treatment with the dose of 10 ng of etanercept inhibited both mechanical and thermal hyperalgesia from 3 to 7 h after the treatment ( Figures 6C, D ). Treatment with the dose of 30 pg of IL-1ra inhibited only mechanical hyperalgesia at 1 h without showing effects upon thermal hyperalgesia, and the anti-hyperalgesic effect of the dose of 100 pg of IL-1ra was detected between 1 and 7 for mechanical hyperalgesia and between 5 and 7 for thermal hyperalgesia ( Figures 6E, F ). All doses of the compounds tested in these experiments did not affect blood parasitemia ( Figure 6G ), and no animal died as a result of infection or treatment protocols ( Figure 6H ). These data ( Figure 6 ) line up well with the mRNA expression results presented in Figure 5 and demonstrate a role for CX₃CL1/CX₃CR1 signaling, TNF-α, and IL-1β in hyperalgesia induced by T. cruzi infection.

The results of Figs. 1–6 show that T. cruzi infection induces pain that is amenable by inhibition of spinal cord glial cells, NFκB, chemokines, and cytokines and that the activation of glial cells is NFκB dependent. In this section, we addressed whether chemokine and cytokine expression would be dependent on NFκB and glial cells ( Figure 7 ). Mice were treated by the intrathecal route with vehicle (saline) or the most effective dose of each inhibitor tested earlier, which include 100 nmol of α-aminoadipate, 150 μg minocycline, and 300 μg of PDTC. Treatments with all inhibitors significantly reduced T. cruzi–induced mRNA expression of Cx3cr1 ( Figure 7A ), Tnfα ( Figure 7B ), and Il1β ( Figure 7C ) in the spinal cord. These data corroborate the results presented in Figs. 3–6 and indicate that spinal cord astrocytes, microglia, and NFκB activation are responsible for the production of major molecules related to the sensitization and activation process of nociceptor sensory neurons in T. cruzi infection.

The response of DRG neurons to experimental T. cruzi infection was evaluated through calcium imaging and Trpv1 mRNA expression (an ion channel involved in nociceptor neuron activation). DRG neurons from infected mice presented a higher baseline level of calcium influx as well as a higher response to capsaicin than DRG neurons of uninfected mice ( Figures 8A–H ). In addition to a higher calcium baseline, DRG neurons of infected mice presented a higher intensity of response and higher percentage of responsive neurons upon capsaicin stimulation than DRG neurons of uninfected mice ( Figures 8H, I ). An explanation for a higher baseline and response to capsaicin is that DRG neurons of infected mice presented increased Trpv1 mRNA expression compared to DRG neurons of uninfected mice ( Figure 8J ). Therefore, the behavioral responses triggered by T. cruzi infection parallel with activation of TRPV1-sensitive DRG neurons further confirming the nociceptive nature of behavioral responses.

Intrathecal treatments reach the DRG, indicating that these treatments could affect DRG cells (41) and spinal cord neuroinflammation causes retrograde nociceptor neuron sensitization (42), indicating that spinal cord treatments can affect DRG cells through neuronal signaling. Therefore, we evaluated whether intrathecal treatments with α-aminoadipate, minocycline, and PDTC would affect DRG neuroinflammation. Mice were treated by the intrathecal route with vehicle (saline) or the same doses of glial and NFκB inhibitors used in Figure 7 experiments (100 nmol of α-aminoadipate, 150 μg minocycline, and 300 μg of PDTC). The mRNA expression for Cx3cl1, Cx3cr1, Gfap (marker for satellite glial cells activity), Tnfα, Il1β, and Cox2 mRNA expression were quantitated ( Figure 9 ). Intrathecal treatments targeting spinal cord astrocytes, microglia, and NFκB reduced T. cruzi–induced Cx3cl1 ( Figure 9A ), Cx3cr1 ( Figure 9B ), Gfap ( Figure 9C ), and Tnfα ( Figure 9D ) mRNA expression in DRG samples. Regarding Il1β, only treatments with minocycline and PDTC inhibited its mRNA expression ( Figure 9E ). There was a trend toward inhibition (P = 0.1236) by α-aminoadipate. This result indicates that α-aminoadipate that selectively targets astrocytes could not significantly affect Il1β mRNA expression. We did not observe significant changes of Cox2 mRNA expression in the DRG ( Figure 9F ). These data indicate that intrathecal treatments targeting astrocytes, microglia, and NFκB result in reduced DRG neuroinflammation after T. cruzi infection.