1. Introduction

The protozoan parasite Toxoplasma gondii is the etiologic agent of toxoplasmosis [1, 2]. Currently, this infection afflicts a third of the world population, with a seroprevalence ranging from 0.8 to 92%, depending on the region and habits of the populations [3, 4]. For instance, in Brazil and USA the seroprevalence is of 92% and 22.5%, respectively [4, 5]. The parasite may invade any cell type and infect all body tissues showing tropism for the central nervous system (CNS) [3, 6]. The acute phase of T. gondii infection is characterized by presence of tachyzoite forms and the chronic phase is defined by the presence of encysted bradyzoite forms, called tissue cysts [7]. The drugs currently available for etiological treatment of T. gondii infection are more effective in the acute phase, however they do not offer a parasitological cure in the chronic phase [8]. Thus, the cysts cannot be eliminated, remaining in the CNS throughout the life of the host, apparently in silent state [7] and with sustained neuroinflammation during the acute and chronic phase in mouse models of infection [9, 10]. In humans, neuroinflammation has been linked to the development of neurodegenerative illnesses also associated with behavioral alterations as Alzheimer disease [11]. Several studies have investigated the role of T. gondii infection as a risk factor for neurological and behavioral disorders. In humans, infection by T. gondii has been related to mental illnesses such as schizophrenia, bipolar disorder, depression and obsessive-compulsive disorder, as well as mood, personality and other behavioral changes [12], as increased rate of involvement in traffic accidents [13], suicide attempts [14] and alteration of cognitive functioning [15]. Alike humans, mice are intermediate hosts of T. gondii [16] and show behavioral abnormalities. In experimentally infected mice, among the reported behavioral changes are loss of predator fear [17, 18], decreased anxiety [10], increased exploratory behavior [19, 20] and impairment of long-term memory during chronic infection [21]. All these findings contributed to the hypothesis of host manipulation by the parasite, contributing to exposition to definitive host predator and T. gondii cycle maintenance.

Behavioral alterations have been proposed to be independent of persistent neuroinflammation [18] and the apparent presence of T. gondii cysts [18, 22] in the CNS. In non-infectious and infectious experimental models, behavioral alterations have been linked to neuroinflammation or independent of it [23–25]. Further, systemic inflammatory profile associated with increased circulating cytokine levels raised as a contributor to underpin behavioral alterations in infectious diseases as hepatitis and Chagas disease [24, 26]. Therefore, in the present work we carried out a kinetics study of infection of C57BL/6 mice with the ME-49 type II strain of T. gondii to settle initially a model of long-term infection to study behavioral changes. For that, we used standardized tests to evaluate the presence of anxiety, depressive-like behavior, and hyperactivity. Further, trying to shed light on the biological factors associated with infection-induced behavioral abnormalities, we assessed cyst numbers and presence of neuroinflammation as well as their topographical localization in the CNS areas, cytokine expression in the brain tissue and serum, and BBB integrity.

2. Materials and methods

2.2 Experimental design

Experimental check list is described in Author´s Check List (S1 Checklist). A total of 116 female mice of the C57BL/6 (H-2^b) lineage, 3-4-week-old, was provided by the Institute of Science and Technology in Biomodels (ICTB) of Oswaldo Cruz Foundation and housed in the Experimental Animal Facility (CEA-CF/IOC unit) under specific pathogen-free conditions, in polypropylene cages lined with pine sawdust and kept in microisolators, under noise and light controlled conditions (12 hours light/12 hours dark). Animals were randomly grouped into groups of 3–5 mice per cage and received water and grain-based ad libitum. To minimize the effects of stress and allow the adaptation process to the new environment, mice were kept without manipulation for 15 days in the cages, provided with environmental enrichment (igloo). After the adaptation period, the animals were infected and analyzed according to the experimental protocols (Fig 1).

Download:

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

Fig 1. Flow chart showing the experimental protocol with the number of animals used.

Death registered, and mice included in 2 (endpoint registered at “30 and 60 dpi”) or 1 (endpoint registered at “90 dpi”) independent experiments.

https://doi.org/10.1371/journal.pone.0258199.g001

At arrival, experimental groups were formed. The cages were numbered and randomly classified for experimental infection and analysis at the indicated timepoints. Experimental groups were formed with a total of 96 mice divided into two replications: Experiment 1 (i) 30 days postinfection (dpi): 5 non-infected (NI) controls and 12 infected; (ii) 60 dpi: 6 NI controls and 15 infected; and (iii) 90 dpi: 6 NI controls and 24 infected. Experiment 2 (i) 30 dpi: 4 NI controls and 10 infected; and (ii) 60 dpi: 4 NI controls and 10 infected. Additionally, 20 mice divided in two independent experiments of 10 mice each were used to study cytokine expression in the CNS at 20 dpi (2 NI, 3 infected) and at 45 dpi (2 NI, 3 infected), respectively, before and after neuroinflammation onset S1 Fig.

3. Results

3.1 Long-term chronic Toxoplasma gondii infection in C57BL/6 leads to loss of muscle strength with preservation of locomotor capacity

Female C57BL/6 mice were infected with 5 cysts of the ME-49 T. gondii strain. Kinetic of infection was evaluated at 30, 60 and 90 dpi (Fig 2A). The survival rate of infected mice was 100% up to 60 dpi, and 83% (5/24) survival was registered at 90 dpi (Fig 2B), compared to 100% in the NI control group. The analysis of all studied parameters showed no difference among the NI control groups run concurrently to T. gondii-infected mice, at 30, 60 and 90 dpi. Thus, for simplification, the collected data of all NI mice were gathered and referred as NI in graphs and figures. Body weight and clinical evolution were monitored weekly. The infected mice showed piloerection accompanied by body weight loss during the acute phase of the infection (up to 15 dpi). After this period, the weight loss ceased, nevertheless the body weight gain in infected mice remained lower than NI control mice (Fig 2C). In addition, the groups of infected mice showed loss of muscle strength at all timepoints of analysis (Fig 2D).

Download:

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

Fig 2. C57BL/6 mice chronically infected with the ME-49 Toxoplasma gondii strain survive and show weight and muscle strength loss.

(A) Mice were infected with 5 cysts of the ME-49 T. gondii strain, the clinical follow-up and the mortality were assessed and recorded weekly, and the kinetics of infection was evaluated at 30, 60 and 90 dpi. (B) The survival curve shows a survival of 100% at 30 and 60 dpi, and 83% (19/24) at 90 dpi, compared to 100% in the age- and sex-matched NI control group. (C) Infected mice showed body weight loss during the acute phase of the infection (up to 15 dpi), after this period, the weight loss ceased. (D) Muscle strength was compromised in infected mice; values of muscle strength are shown as gram force (gf) / body weight (g). Each experimental group consisted of 4–6 NI mice and 10–19 T. gondii-infected mice. Each circle represents an individual mouse. Data are expressed as means ± SEM, and were analyzed using Welch’s test (C), and ordinary one-way ANOVA (D). **, p<0.01. ***, p<0.001, comparing T. gondii-infected and NI mice.

https://doi.org/10.1371/journal.pone.0258199.g002

To explore whether the locomotor capacity was preserved, the gait pattern was also assessed using the footprint test (Fig 3A). At the three studied timepoints, no differences were detected in the width of the right and left toe spread (Fig 3B) and in the right and left step overlap (Fig 3C), when compared to NI mice. However, the forelimb and hindlimb stride were altered as the infection progresses (Fig 3D). Lastly, the width of the front base increased at 30, 60 and 90 dpi, contrasting with the preservation of the width of the hind base (Fig 3E). Therefore, the locomotor capacity of T. gondii-infected mice was preserved.

Download:

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

Fig 3. Mice with long-term chronic Toxoplasma gondii infection show preserved locomotor capacity.

(A) The gait analysis was performed using the footprint test and were evaluated as such: the spread of the fingers and overlapping distance between the anterior and posterior limbs, right and left, and the stride length of limbs, width of the front and rear base. (B) The toe spread was normal in the right and left limb. (C) |No alteration was observed in the right and left step overlap of infected mice. (D) The forelimb and hindlimb stride were increased as the infection progresses. (E) The width of the front base was enlarged in infected mice, but the hind base was preserved. Each experimental group consisted of 4–6 NI mice and 8–12 T. gondii-infected mice. Each circle represents an individual mouse. Data are expressed as means ± SEM, and were analyzed using ordinary one-way ANOVA. *, p<0,05, **, p<0.01. ***, p<0.001, comparing T. gondii-infected and NI mice.

https://doi.org/10.1371/journal.pone.0258199.g003

3.2 Long-term chronically Toxoplasma gondii-infected mice exhibit anxiety, depressive-like behavior, and hyperactivity

To assess anxiety, we used the OFT and analyzed the time expended in the central zone of the apparatus [29]. Our data showed that when compared to NI controls chronically infected mice exposed to the OFT remained reduced time in the central zone, at 30, 60 and 90 dpi (Fig 4A). More, increased time was expended in the peripheral area of the apparatus, revealed the intensity of the registered lines near the apparatus’ walls (Fig 4B). The presence of depressive-like behavior was evaluated using TST and revealed as enhanced time of immobility [31]. Consistently, when compared to NI controls chronically T. gondii-infected mice showed increased immobility time at early (30 dpi), and long-term (60 and 90 dpi) chronic infection (Fig 4C). Therefore, considering the standardized parameters in the early (30 dpi) and long-term (60 and 90 dpi) chronic T. gondii-infected mice showed depressive like-behavior and anxiety, compared with sex- and age-matched NI controls.

Download:

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

Fig 4. Anxiety and depressive-like behavior are detected in long-term chronically Toxoplasma gondii-infected C57BL/6 mice.

(A) Chronically infected mice showed reduced time in the central zone in OFT, and (B) increased time exploring the peripheral area of the open field. (C) Depressive-like behavior was revealed as enhanced time of immobility in TST. Each experimental group consisted of 4–6 NI mice and 10–15 T. gondii-infected mice. Each circle represents an individual mouse. Data are expressed as means ± SEM, and were analyzed using t-Student test. *, p<0,05, **, p<0.01. ***, p<0.001, comparing T. gondii-infected and NI mice.

https://doi.org/10.1371/journal.pone.0258199.g004

Curiously, we observed that T. gondii-infected mice were apparently more active than NI mice. Thus, we refined our analysis and evaluated the time that infected mice spent immobile, and the distance and speed of travel inside the OFT. Initially, we did not notice differences in terms of immobility time in the OFT when chronically infected mice were compared to NI controls (Fig 5A). The analysis of the walked distance revealed that compared with NI controls 18%, 32% and 21% of infected mice traveled longer distances at 30, 60 and 90 dpi, respectively (Fig 5B). Indeed, the group of long-term chronically infected mice (at 60 dpi) walked longer distances than NI controls (Fig 5B). Moreover, compared with NI controls the walk speed was increased in early (30 dpi) and long-term (60 and 90 dpi) chronically infected mice (Fig 5C), thus suggesting aggravation of this change as the infection progresses. Altogether, these findings support an increase in the locomotor activity in chronically T. gondii-infected mice. Next, we tested the performance of infected mice in FST, a test to evaluate mice activity in adverse environments [33]. Compared to NI controls, we detected a significant decrease in the immobility time in all groups of infected mice submitted to FST. Further, this behavioral alteration is visibly aggravated with the course of infection, comparing 60 dpi with 90 dpi (p < 0.01) and 30 dpi with 90 dpi (p < 0.001) (Fig 5D). Altogether, increase in locomotor activity in OFT and reduced time of immobility in FST, also registered as increased time of activity, are suggestive of impulsive hyperactivity in early and, mainly, in long-term chronic T. gondii infection in C57BL/6 mice.

Download:

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

Fig 5. Hyperactive behavior is observed in chronically Toxoplasma gondii-infected C57BL/6 mice.

To determine hyperactivity, we evaluated the time spent immobile, the distance and the speed of travel in OFT, and the activity in FST. (A) Infected mice did not show differences in immobility time when in the OFT. (B) At 60 dpi, mice showed increased distance traveled in OFT. At 30 (18%) and 90 (21%) dpi only a reduced percentage of mice showed increase in the distance traveled. Red dot line shows mean distance of NI group + 2 standard deviation. (C) The walk velocity in OFT was increased as infection progressed. (D) In FST, immobility time decreased as infection progressed. Each experimental group consisted of 4–6 NI mice and 8–15 T. gondii-infected mice. Each circle represents an individual mouse. Data are expressed as means ± SEM, and were analyzed using t-Student test (A-C) and the Mann-Whitney test (D). *, p<0.05, **, p<0.01. ***, p<0.001, comparing T. gondii-infected and NI mice.

https://doi.org/10.1371/journal.pone.0258199.g005

3.3 The number of cysts in the CNS decreases significantly throughout Toxoplasma gondii infection

Herein, the parasite load was determined by analyzing the number and size of T. gondii cysts present in the CNS of ME-49-infected mice. The number of cysts decreased as the early (30 dpi) chronic infection coursed to long-term chronic infection at 60 dpi (p < 0.05 compared with 30 dpi) and at 90 dpi (p < 0.001 vs 30 dpi; p < 0.01 vs 60 dpi), suggesting gradual control of infection (Fig 6A). However, the size of the cysts increased as infection progressed from early (30 dpi) to long-term (60 and 90 dpi) chronic phase (Fig 6B). Indeed, a more detailed analysis of the cysts size using the digital morphometric apparatus NIS Elements BR version 4.3 software (Nikon Co., Japan), revealed 18 size classes to be considered. Further, these data showed increase in the relative frequency of cysts > 42–46 μm in 60 and 90 dpi, compared with 30 dpi (Fig 6C).

Download:

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

Fig 6. The numbers of parasite cysts in the brain decrease and the sizes increase as Toxoplasma gondii infection progresses in mice C57BL/6.

(A) The number of cysts decreased as the early (30 dpi) infection coursed to long-term chronic infection (60 and 90 dpi). (B) The size of the cysts increased as infection progressed from early to long-term chronic phase. (C) The graphs show the relative frequencies of cysts in each class at the three timepoints analyzed. Each experimental group consisted of 8–12 T. gondii-infected mice. Each circle represents the number of cysts in the brain of each animal (A), and the diameter of each cyst (B). Data are expressed as means ± SEM, and analyzed using the Mann-Whitney test. #, p <0.05, ##, p <0.01. ###, p <0.001, comparing different timepoints.

https://doi.org/10.1371/journal.pone.0258199.g006

Next, we tried to establish associations between the number of cysts in the CNS with the behavioral features analyzed, and no correlation was found in most of the analysis performed. Correlation (p < 0.05) was detected between the number of cysts in the CNS and immobility time in FTS S2A Fig. Further, correlation was observed between the number of cysts and the left and right forelimb strides (p < 0.05) and left and right hindlimb strides (p < 0.01) in the footprint test S2B and S2C Fig.

3.4 Cysts of Toxoplasma gondii prevailed in some regions of the encephalon

Next, to evaluate a putative differential accumulation of T. gondii cysts in specific areas of the CNS, we analyzed two hematoxylin-eosin-stained histological sections per mouse. We divided the brain areas into: Olfactory areas (OLF), Isocortex (ICTX), Cerebral Nuclei (CNU), Hippocampal formation (HPF), Thalamus (TH), Hypothalamus (HY), Midbrain (MB), Pons (P), Medulla (MY) and Cerebellum (CB), according to Allen Institute for Brain Science [41], as shown in Fig 7A. At all timepoints evaluated (30, 60 and 90 dpi), the cysts were found in all areas of the CNS except for HPF at 90 dpi, when no cyst was observed in this area of any of the studied mouse. A larger number of cysts was detected in the ICTX, TH and MB areas, and this pattern was preserved in the three analyzed timepoints (Fig 7B and S3A Fig), supporting the existence of differential accumulation of T. gondii cysts in these CNS tissue areas. The size of the cysts does not vary due to the region in which they are located S3B Fig. Further, most cysts were observed isolated, but groups of 2 to 7 cysts were also found, regardless the studied areas of the brain (Fig 7C).

Download:

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

Fig 7. Some regions of the brain of C57BL/6 Toxoplasma gondii-infection mice are intensely infected.

(A) The stereotaxic coordinates of the bregma of the mouse´s brain were used to construct representative maps of the topographic location of the cysts in each brain area: Olfactory areas (OLF), Isocortex (ICTX), Cerebral Nuclei (CNU), Hippocampal formation (HPF), Thalamus (TH), Hypothalamus (HY), Midbrain (MB), Pons (P), Medulla (MY) and Cerebellum (CB). (B) The cysts were localized in all brain regions, but the isocortex, thalamus and midbrain were colonized more intensely in the early (30 dpi) and long-term (60 and 90 dpi) chronic infection (each color represents an individual mice). The histograms show the percentage of cysts per area studied. (C) Representative pictures show individual cysts or multiple cysts per microscopic field in different brain areas, associated or not with inflammatory foci. Bar = 60 μm. Each experimental group consisted of 4–10 T. gondii-infected mice. Data were analyzed using ordinary one-way ANOVA.

https://doi.org/10.1371/journal.pone.0258199.g007

3.5 Behavioral changes are concomitant with generalized neuroinflammation in long-term chronic Toxoplasma gondii infection

In the three analyzed timepoints, infected mice presented inflammatory foci with meningoencephalitis and perivascular inflammatory cuffs composed of mononuclear inflammatory cells in all evaluated areas of the CNS (Fig 8A and S3C Fig). Rare apparently silent cysts devoid of inflammatory process were observed at 30 dpi (4.71%; 9/191) and at 60 dpi (4.09%; 5/122), while at 90 dpi no silent cyst was observed. Most of the cysts were surrounded or closer to inflammatory foci. A similar pattern was observed in the three evaluated timepoints in a way that the average distance between cysts and inflammatory foci ranged from 61.4 to 195.2 μm at 30 dpi, 104.9 to 182 μm at 60 dpi group, and 46.5 to 254.3 μm at 90 dpi S3C Fig. In addition, the association of cysts and inflammatory foci patterns was similar in different CNS areas.

Download:

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

Fig 8. Toxoplasma gondii-infected C57BL/6 mice showed generalized neuroinflammation in early and long-term chronic phase of infection.

(A) Infected mice presented inflammatory foci with meningoencephalitis and perivascular inflammatory cuffs in all evaluated areas of the CNS in early (30 dpi) and long-term (60 and 90 dpi) chronic infection. Bar = 330 μm. (B) Cysts devoid of inflammation were firstly detected in the CNS at the acute phase (20 dpi, left panel), and the inflammation is already settled in the chronic phase (45 dpi, right panel). Bar = 100 μm. (C) Increased expression of the pro-inflammatory cytokines IFNγ and TNF was detected in the acute phase, and TNF expression was sustained at the chronic phase of infection. (D) In the acute phase, the expression of MIP1α/CCL3, MIP1β/CCL4, RANTES/CCL5 and MCP-1/CCL2, was increased. The upregulation of the expression of TNF and CC-chemokines were sustained at 45 dpi. Each experimental group consisted of 2–4 NI mice and 2–10 T. gondii-infected mice, in two independent experiments. Data are expressed as means ± SEM, and were analyzed using ordinary one-way ANOVA. *, p<0,05 comparing T. gondii-infected and NI mice, and #, p<0.05 comparing acute and chronic groups of T. gondii infected mice.

https://doi.org/10.1371/journal.pone.0258199.g008

Considering that mononuclear cells prevailed in neuroinflammatory processes at 30, 60 and 90 dpi, we settled an experiment to shed light on the pattern of cytokines and CC-chemokines driving putatively the migration of these cells. For that, a group of mice was analyzed at 20 dpi, when cysts devoid of inflammation were firstly detected in the CNS, and at 45 dpi, when inflammation was already settled (Fig 8B). Compared to NI controls, increased expression of proinflammatory cytokines IFNγ and TNF was detected at 20 dpi, therefore, preceding neuroinflammation. Further, TNF expression was sustained at 45 dpi (Fig 8C). At 20 dpi, the expression of the four CC-chemokines analyzed namely MIP1α/CCL3, MIP1β/CCL4, RANTES/CCL5 and MCP-1/CCL2, was enhanced, when compared to NI controls. Although at 45 dpi the expression of MIP1β/CCL4 and JE-MCP-1/CCL2 was reduced in comparison with the expression at 20 dpi, these CC-chemokines remained upregulated in this timepoint of chronic infection compared to NI controls (Fig 8D).

3.6 Systemic cytokine expression is upregulated at early and long-term chronic Toxoplasma gondii infection

To assess systemic cytokine serum levels, blood of NI controls and T. gondii-infected mice was collected. The obtained sera were stored and submitted to simultaneous detection of cytokines, using the Mouse Inflammation CBA kit (IL-12, IL-6, TNF, IFN, MCP-1/CCL2, IL-10). Representative data plots of the FACS analysis are shown (Fig 9A), supporting that at 60 dpi all analyzed cytokines were upregulated, when compared to NI controls. In general, chronically T. gondii-infected mice increased serum levels of pro-inflammatory cytokines. The levels of TNF, IFNγ and MCP-1/CCL2 were elevated significantly at 30 and 60 dpi. At 90 dpi, levels of all cytokines showed a tendency to decrease or, even, cytokines levels were alike those found in sera of NI controls, except for IFNγ and MCP-1/CCL2 levels that remained upregulated at this timepoint (Fig 9B).

Download:

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

Fig 9. Systemic cytokine expression is upregulated in the early and long-term chronic Toxoplasma gondii infection.

(A) The images show representative data plots of the FACS analysis of CBA. (B) Chronically T. gondii-infected mice showed increased levels of pro-inflammatory cytokines at 30 and 60 dpi. At 90 dpi, all cytokine levels showed a tendency to decrease or, even, exhibited cytokines levels like those found in sera of NI controls, except for IFNγ and MCP-1/CCL2 levels. Each experimental group consisted of 2–3 NI mice and 2–5 T. gondii-infected mice, in two independent experiments. Data are expressed as means ± SEM, and were analyzed using ordinary one-way ANOVA. *, p<0,05, **, p<0.01. ***, p<0.001, comparing T. gondii-infected and NI mice.

https://doi.org/10.1371/journal.pone.0258199.g009

3.7 Toxoplasma gondii infection induces disruption of the blood-brain barrier and brain edema

Based on the property of EB dye that binds proteins, mainly plasma albumin and on the physiological ability of the preserved BBB to be impervious to this protein, NI controls and infected mice were injected with EB and the encephala analyzed for EB extravasation, assumed as a biomarker of BBB disruption [42]. Representative images of brains depict localized and spotted EB extravasation at 30 dpi, while it is more evident in the whole brain at 60 dpi, and less noticeable at 90 dpi (Fig 10A). Representative sagittal sections of the encephala at 30 and 60 dpi corroborated this description (Fig 10B). When compared to NI controls, the quantitative data disclosed increased concentrations of EB in the encephala of infected mice at the three analyzed timepoints, with maximum levels achieved at 60 dpi, therefore, revealing BBB significant vascular permeability in T. gondii-infected (Fig 10C). Further, infected mice also presented a significant increase in relative brain weight in all evaluated timepoints (Fig 10D). Thus, associated with increased EB extravasation, these data are suggestive of cerebral edema.

Download:

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

Fig 10. The increase in BBB permeability is concomitant with edema in Toxoplasma gondii-infected C57BL/6 mice.

(A) Representative images of brains depicting EB extravasation at the timepoints assessed. (B) Representative sagittal sections of the encephala, at 30 and 60 dpi. (C) The concentrations of EB in the brain tissue of infected mice increased at the three studied timepoints, indicating the increase in BBB permeability. (D) Cerebral edema, manifested by a significant increase in the relative brain weight (brain weight in milligram / whole body weight in gram), was present in the timepoints evaluated. Each experimental group consisted of 4–6 NI mice and 8–19 T. gondii-infected mice, in two independent experiments. Each circle represents an individual mouse. The data are expressed as means ± SEM, and were analyzed using Mann-Whitey test (C) and t-Student test (D). **, p<0.01. ***, p<0.001, comparing T. gondii-infected and NI mice.

https://doi.org/10.1371/journal.pone.0258199.g010

4. Discussion

The present study evaluated the kinetics of the behavioral alterations of T. gondii-infected C57BL/6 mice, using a battery of standardized behavioral tests. Our data revealed that during the early (30 dpi) and long-term (60 and 90 dpi) chronic infection mice showed anxiety, depressive-like behavior and hyperactivity, concomitant with neuroinflammation, systemic inflammatory environment and BBB disruption. Further behavioral changes occurred regardless of the progressive decrease in the number of cysts and were independent of the localization of the cysts and inflammatory foci in the CNS areas. Thus, multiple factors may contribute to the observed behavioral abnormalities.

The experimental model we used, consisting of C57BL/6 mice infected with the ME-49 strain, allowed survival and chronic phase onset. Although our focus is on chronic infection, we monitored the clinical evolution of mice after the first day postinfection up to the end point. The mice showed a decline in body weight and the presence of piloerection in the first 15 days of infection, features proposed to be associated with the period of widespread parasite multiplication, typical of the acute T. gondii infection [43]. In the chronic phase, weight loss ceased, and body weight was steadied, alike previously described in other studies [9, 20, 44, 45]. Neuromuscular strength decreased in ME-49 strain-infected C57BL/6 mice, thus corroborating previous data that showed reduced neuromuscular strength in this mode [45] and in C57BL/6 mice infected with 10³ bradyzoites of ME49 strain [46]. Also, in chronically ME-49-infected Swiss Webster mice, decreased muscle strength was associated with myositis and reduction in muscle mass, being proposed that the inability to recover the weight during the chronic phase (> 30 dpi) is due, in part, to the continuous damage of the skeletal muscle [9, 45]. Likewise, muscle loss in the anterior tibialis, gastrocnemius and quadriceps muscles have been shown in chronically (37 dpi) ME-49-infected CBA/J mice [43]. Therefore, these data support that chronic infection characterized by cyst formation in the CNS, was associated with body weight loss, as reproduced in our experimental model.

Gait changes can represent a pathological state in many neurological diseases and disorders [47]. To analyze the locomotor capacity, the footprint test was performed, this is a gait analysis method, sensitive to neurodegenerative changes, and can be used in locomotor assessment of different disorders [48]. Here, we evaluated the spatial components of gait: stride length, width of the front and rear base, overlapping distance of the paw and spreading of the toes of right and left limbs. Stride length has been described as a reliable index of motor disorders caused by dysfunction of the basal ganglia in mice [49]. The increase in stance width generally indicates a compensatory gait pattern to avoid gait instability resulting from pain [50]. The overlapping distance of the limbs assesses the existence of greater flexibility of hip rotations or strength of the limbs to reach longer strides [47] as well as the uniformity of the alternation of steps [34] and, the toe spread can be used to form functionality indices for the sciatic, tibial, and peroneal nerves [50]. Here, the T. gondii-infected mice showed increase in the forelimb and hindlimb stride length at early and long-term chronic infection. The increase in stride length is influenced by the speed at which the mouse walks or runs, without affecting its gait [48, 49, 51]. The reduction in this parameter has been described in patients with neurodegenerative diseases such as Parkinson disease [52], and in murine models of Huntington disease [34, 53] and arthritis as a manifestation of pain [48, 54], but unbalanced gait was not observed in the model we described here. T. gondii-infected mice showed increase in the stance forelimb width. The increase in the width of the rear base, accompanied by a decrease in the length of the steps, was reported in a murine model of antigen-induced arthritis, maintained even after intra-articular treatment with AMPA/kainate glutamate receptors antagonist, which prevents pain and reduces pathology [55]. Thus, wider stance widths generally indicate a compensatory gait pattern used to protect a member of a painful movement and avoid gait instability [50]. In our infected mice, there were no signs of ataxia or foot drag detectable by the footprint test, and rear base width, toe spread and overlap of step were preserved. More, when compared to NI controls, the infected mice showed similar exploratory behavior, similar immobility time, and covered similar or longer distances in the OFT, besides the higher speed. So, we consider that these features support that our model maintains preserved locomotor capacity. In a previous study, chronically T. gondii-infected C57BL/6 mice with motor coordination deficits, impaired gait (mis-steps, stride length short, foot dragging in the rear paws), and reduced muscle strength, showed a reduction in exploration in the OFT, but no changes in social behavior or impaired cognitive function were observed, as object recognition test and spatial memory tests require displacement in the open field [46]. Interestingly, Trypanosoma cruzi-infected C57BL/6 mice show preserved muscle strength but alterations in cognitive function, also evaluated in tests that require displacement in the open field [25], supporting that these are dissociated features. Thus, our data suggest that decreased neuromuscular strength did not affect the motor function and the general locomotor activity of chronically ME-49-infected C57BL/6 mice.

Mice tend to explore a new environment, however anxiety behavior reduces time spent expended in central area and increases time consumed closer to the wall or peripheral area of the OFT apparatus [28]. The presence of anxiety, revealed as decreased time expended in the central area of the OFT, was detected in the early and long-term chronic infection of C57BL/6 mice with the ME-49 strain. Recently, similar results were shown in the model of early (30 dpi) chronic T. gondii infection of Swiss Webster mice [19]. In experimental models, TST and FST are widely used to assess depressive-like behavior in infectious and non-infectious diseases [24, 56–58]. It is expected that, when compared to their controls, afflicted mice subjected to both tests will increase immobility time as an indication of a depressive-like behavior [31, 33]. Here, in early (30 dpi) chronic T. gondii infection, mice showed an increase in the immobility time in TST, supporting a depressive-like behavior that is sustained as the infection progresses to the long-term chronic phase (60 and 90 dpi). The FSL rats, a lineage susceptible to behavioral alterations, showed depressive-like behavior when infected with the ME-49 T. gondii strain, and evaluated by sucrose preference test and FST [44]. Contrasting with the finding in FSL rats assessed in FST and added to our results of increased immobility in TST, we bring evidence that ME-49-infected C57BL/6 mice showed a decrease in the immobility time in FST, when compared to the NI controls. Thus, these conflicting results obtained in the two behavioral tests initially proposed to study depressive-like behavior applied to the same experimental groups, leading us to query the applicability of the FST to describe other behavior alterations. Truly, FST has been used previously to examine hyperactivity [59], stress coping behavior [60], while other authors suggest that the FST can be used to assess learning capability [61] and anxiety [62]. Discordant results have also been reported in non-infectious models, when no changes were exhibited in TST and reduced immobility time was shown in FST [57, 58]. Again, in a model of schizophrenia in Swiss Webster mice treated with a NMDA antagonist in the acute (single dose) or the chronic (10 consecutive days) process and submitted to FST and TST, the results were dissenting. In the acute protocol, increased immobility in TST was observed, and no changes were detected in FST. Conversely, in the chronic protocol immobility was increased in FST, whereas decreased immobility was detected in TST [56], supporting that these behavioral tests reveal different physiological mechanisms underlying the immobility process [56–58]. Chronic infection with T. gondii has been noticed as an anxiety reducer in mice exposed to elevated plus-maze test (EPMT) where they visited the open arms more frequently, even showing a decrease in the exploration of the central area of the OFT [10]. However, similar results in EPMT were also interpreted as changes in risk behavior, in model of chronic T. gondii infection [20]. We consider that like the FST, the results obtained in the EPMT may be misinterpreted, whenever EPMT allows to explore the impulsive behavior, one of the symptoms of attention-deficit/hyperactivity disorder (ADHD), which is expressed by the increase in exploratory activity in the open arms [63–67]. Likewise, we propose that the FST test can be a valid tool in the study of impulsivity in a murine model, though further studies are necessary. Anyhow, the reduced immobility in the FST, added to the increase in exploratory behavior and increased velocity in the OFT, led us to conclude that ME-49-infected C57BL/6 mice show a complex pattern of behavioral alterations with depressive-like behavior and hyperactivity in the early and long-term chronic infection, with aggravation of the process as the infection progresses. Interestingly, a recent study showed a relation between the severity of ADHD and IgG positivity for T. gondii infection in 6 to 18-year-old children, suggesting a role for the parasite infection in the exacerbation of ADHD [68]. Therefore, our results supporting hyperactivity in the OFT, depicted as increased velocity and total distance covered, corroborate the description of hyperactivity in the type II Prugniaud strain-infected BALB/cJ mice [22, 69], C57BL/6J infected with genetically modified ME-49 tachyzoites [20], and Swiss Webster infected with the ME-49 strain [19]. Notably, in our model regardless of the reduction in the number of brain cysts, the described behavioral changes were sustained at 60 and 90 dpi. In contrast, a relationship between the load of cysts and the severity of behavioral changes has been reported [10]. Indeed, our results revealed that the reduction in the number of brain cysts as infection progresses was not associated with reduction of hyperactivity, corroborating data in BALB/cJ mice chronically infected with the Prugniaud strain and treated with the antiparasitic guanabenz [22].

To shed light on possible biological contributors to the observed behavioral changes and knowing that the positioning of cysts in specific areas of the brain could give rise to certain behavioral changes [20], we investigated the topographical distribution of cysts and the proximity with inflammatory foci in the brain tissue. Since the cortex occupies 56% of the brain area [39], we divided it into different regions: olfactory areas, isocortex and hippocampal formation, to avoid bias in our results. Our model showed an accumulation of cysts in the isocortex, thalamus and midbrain in the three timepoints analyzed. The existence of tropism of T. gondii cysts to specific regions of the CNS is not entirely clear [70] and previous studies reported the heterogeneous distribution of cysts in the brain of mice. In C57BL/6 mice chronically infected with 30 cysts of the ME-49 strain, a larger infection of the hippocampus and telencephalon was observed, an area comprising the cerebral cortex and the cerebral nucleus [71]. In chronically ME-49-infected B6CBAF1/J mice, a heterogeneous distribution of brain cysts was observed, with enrichment of cysts in the frontal cortex and brain stem structures [10], as well as a higher parasite load in diencephalon, cortex and hippocampus in ME-49-infected Swiss Webster mice [9]. In Long-Evans rats infected with tachyzoites of the Prugniaud strain, a random distribution was observed throughout the forebrain, with enrichment in the preoptic area and the paraventricular hypothalamic nucleus [72]. CD1 mice infected with the HIF strain also showed an accumulation of cysts in the telencephalon [73]. These data led us to the support the existence of tropism of ME-49 strain to specific regions of the CNS, especially the cerebral cortex, and that tropism is dependent on the murine model and the T. gondii strain, indicating that further studies are required to explore this point. Cortical areas, limbic regions and basal ganglia are implicated in patients with depressive disorders [74], anxiety [75] and hyperactivity [76]. In addition, changes in the cerebellar brain region have been reported in patients with ADHD [76]. Despite the higher frequency of cysts in the isocortex, thalamus and midbrain showed in T. gondii- infected mice, the cysts were found to colonize all areas evaluated, although less frequently. Thus, it is not possible to link the location of the brain cysts detected in our model with the behavioral changes reported here. Once the influence of cysts´ localization in the brain on behavioral changes has been ruled out, we evaluated the influence of the size of ME-49 strain cysts, as cyst diameter, evaluating the kinetics of progression of the infection, and the influence of the size of the cyst in the affected brain area. We observed the presence of brain cysts of various sizes in all groups, indicative of an alternation of the growth and rupture phases underlying the population renewal of cysts [10], which did not allow correlation with behavioral changes. In addition, the size of the cysts by area of the brain was homogeneous, so it was not possible to observe any correlation. The variable observed in the three timepoints was neuroinflammation, detected in all brain regions, independently of the presence of cysts. Indeed, neuroinflammation was consistently described in previously analyzed experimental models of chronic toxoplasmosis [9, 10, 22]. Hyperactivity in a murine model of chronic T. gondii-infection was reduced with the use of guanabenz, a drug with potency to reduce neuroinflammation and perivascular cuff, but not the number of cysts in the CNS [22]. In addition, in ME-49-infected BALB/c mice, anxiety and short- and long-term memory impairment of new object recognition were reversed in mice treated with rosuvastatin. This therapy reduced the burden of tissue cysts in the brain and attenuated, but not resolve, the signs of neuroinflammation, including meningitis, perivascular cuffs, microglial proliferation, inflammatory cell infiltration and tissue damage [77]. The combination of sulfadiazine and pyrimethamine, the conventional therapy against toxoplasmosis [8], prevents the presence of parasites in the brain and the development of toxoplasmic encephalitis in murine models, but not the development of mild inflammatory lesions [78], which could still contribute to behavioral abnormalities, a question to be further explored. In humans, there are studies that demonstrate the efficacy of combined therapy in cognitive function in infants [79] and children [80] that developed normally the CNS and preserved intellectual function after treatment. However, the efficacy of this combined therapy regardng the cognitive functions of murine models of T. gondii infection is a matter to be further explored. In our study, mononuclear cells prevailed in the neuroinflammatory processes at 30, 60 and 90 dpi, thus we evaluated the profile of expression of cytokines and CC-chemokines, that could be involved in cell migration to the CNS, at 20 dpi (before neuroinflammation onset) and at 45 dpi (in the presence of neuroinflammation). We observed increased intracerebral expression of the proinflammatory cytokines IFNγ and TNF, and of the CC-chemokines, MIP1α/CCL3, MIP1β/CCL4, RANTES/CCL5 and MCP-1/CCL2 at 20 dpi, preceding neuroinflammation. At 45 dpi, when neuroinflammation is present, although the expression of IFNγ, MIP1β/CCL4 and MCP-1/CCL2 were reduced, the intracerebral expression of TNF and other CC-chemokines remained upregulated. Similar results were described in a model of toxoplasmic encephalitis in BALB/c mice, during the acute (10 dpi) and chronic (30 dpi) phases of the infection. In this model, increase in intracerebral mRNA of MIP1α/CCL3, MIP1β/CCL4, RANTES/CCL5 and MCP-1/CCL2 was dependent on IFNγ expression, as leukocyte recruitment to brain tissue was impaired in IFNγ-deficient mice [81]. Astrocytes, microglial cells and inflammatory leukocytes can produce intracerebral chemokines, that might act as facilitators of the recruitment, adherence and transendothelial migration of leukocytes through the BBB to the brain parenchyma, thus influencing the composition of the inflammatory infiltrate [81]. On the other hand, the peripheral inflammation can cause changes in cytokine levels in the brain through several mechanisms and, therefore, control inflammatory cell invasion of the CNS. Macrophages residing the CNS can be activated through the vascular endothelium [82] or through the circumventricular organs of brain regions devoid of BBB [83]. Also, circulating cytokines can be actively transported across the BBB [82], and may, therefore, activate glial cells to produce cytokines. Thus, leakage of cytokines of systemic plasma, together with the intracerebral cytokines and chemokines, could contribute to the maintenance of neuroinflammation. In our study, most of the assessed inflammatory cytokines (IL-12, IL-6, TNF, IFNγ and CCL2/MCP-1) showed high serum levels in the early chronic infection (at 30 dpi), with peak at 60 dpi and control at 90 dpi, except for IFNγ and CCL2/MCP-1, that persisted elevated. IL-10 serum levels, however, tended to increase only at 60 dpi. Thus, as the specific immune response is established, it may contribute to control parasite growth, leading to decrease the number of parasite cysts in the CNS, and contributing to reduce the stimulus to production of pro-inflammatory cytokines [84]. Previous kinetics study (7 to 70 dpi) in ME-49-infected mice has shown increased IL-12 and IFNγ levels in the acute phase with a decrease in the chronic phase of T. gondii infection, but without returning to baseline levels [85]. Molecules of T. gondii stimulate innate Toll-like receptors, which lead to the production of IL-12 that with TNF synergistically act to induce IFNγ production, as part of a robust Th1 immune response, crucial to establish an efficient antiparasitic response in the acute phase [86]. In addition, in chronic infection IFNγ is pivotal to control multiplication and dissemination of the parasite within the brain [81]. Indeed, depletion of this cytokine in the chronic phase leads to reactivation of the infection and inflammatory foci in the CNS [87]. Therefore, continuous systemic Th1 immune response, crucially IFNγ, is necessary to control T. gondii parasitism in the brain in the chronic phase of infection [84, 88]. IL-6 is traditionally described as a pro-inflammatory cytokine. However, it has been shown that increased IL-6 levels may regulate the production of IL-12 and IFNγ, resulting in an anti-inflammatory signal [89]. On the other hand, IL-6 plays a protective role in chronic infection, as IL-6-deficient ME-49-infected mice show high numbers of cyst and mortality, with severe toxoplasmic encephalitis with areas of necrosis [90]. IL-10 downregulates the expression of IL-12 and Th1 cytokines, but not the CC-chemokine CCL2/MCP-1 [91]. The anti-inflammatory response triggered by IL-10 may act favoring tissue repair but also contributing to the maintenance of cysts in the CNS [92]. The expression of the CC-chemokine CCL2/MCP-1 can be induced by inflammatory stimuli, such as TNF [93]. Moreover, intracerebral CCL2/MCP-1, acting via CCR2, plays a role in activating microbicidal mechanisms that control T. gondii parasitism in the CNS [93]. Altogether, our data suggest that the C57BL/6 mice model infected with the ME-49 strain triggered an efficient effector immune response, involving cytokines and CC-chemokines, that contributes to parasite control and reduction in the number of cysts in the CNS. Further, neuroinflammation, which may lead to behavioral changes, could be maintained through positive regulation of the intracerebral and peripheral pro-inflammatory cytokines and CC-chemokines, attracting inflammatory cells. A variety of neuroendocrine, neurochemical and behavioral changes are proposed to be consequence of peripheral immune activation, through the release of pro-inflammatory cytokines [11]. It has been shown that anxiety and depression can be induced by the systemic administration of IFNα in patients with hepatitis C [26]. Even so, mild peripheral inflammation in non-infected humans leads to impaired spatial memory [94]. In C57BL/6 mice infected with the ANKA Plasmodium berghei strain, which causes cerebral malaria, and cognitive deficits can be triggered by the early migration of mononuclear cells to the brain, facilitating the increase of chemokine levels in the brain [95]. The transient systemic inflammation induced by intraperitoneal administration of LPS to a mouse model of chronic neurodegenerative disease, can exacerbate inflammation in the CNS and accelerate disease progression as impaired motor coordination and cognitive function [96]. In absence of neuroinflammation, the role of TNF levels in serum in inducing depression-like behavior was demonstrated in a murine model of experimental chronic Chagas disease [24]. Thus, we suggest that the increased levels of cytokines and CC-chemokines in plasma and their leakage in the CNS may sustain neuroinflammation, which may result in the observed behavioral changes, anxiety, depression, and hyperactivity in chronic T. gondii infection.

The disruption of the BBB resulting from the establishment of a vascular or cerebral pathology will result in the leakage of serum-derived components to the CNS, which can lead to brain dysfunction affecting the thinking processes, mood and behavior, and generate psychiatric disorders [97]. A hypothesis has been proposed that the high rate of psychiatric diseases is associated with the breakdown of BBB thus assuming a relationship of peripheral inflammatory processes in psychiatric disorders [11]. In the last decades, data have sustained that disruption of the BBB allows the extravasation of pro-inflammatory cytokines and immune cells that can activate the CNS resident cells and, therefore, induce neurodegeneration, underpinning behavioral alteration [11]. Here, our data show that in the three timepoints BBB integrity was impaired and the relative brain weight was raised, suggestive of brain edema, supporting the leakage of blood-born molecules into the CNS putatively contributing to neuroinflammation. Similar data were obtained in chronically ME-49-infected Swiss Webster [98]. Inflammatory processes in the CNS may contribute to determine the severity and prognosis of neurological and cognitive disorders and can both cause and result from BBB dysfunction [99]. Truly, BBB impairment is associated with brain pathophysiology in several neurological disorders in humans and experimental mice models, including traumatic brain injury [100], stroke [101], epilepsy [102], autoimmune encephalitis, schizophrenia [103], Alzheimer disease [102] and depression [103, 104]. Similarly, psychiatric disorders are related to changes in levels of pro-inflammatory cytokines in serum [105–107]. Patients with bipolar disorder have elevated peripheral levels of IL-6 and TNF [105]. Increased IL-1β, IL-10 and TNF levels are present in patients with depression [106]. Likewise, IL-12, IL-6, TNF and IFNγ peripherical levels are enhanced in patients with schizophrenia [107], and IL-6 and IL-10 levels elevated in children with ADHD [108, 109]. In acute and chronic T. cruzi infection, C57BL/6 mice are refractory to neuroinflammation and upregulation of cytokine expression in the CNS but show depressive-like behavior. This behavioral change was associated with increased TNF levels in serum and abolished by anti-TNF antibody [24], reinforcing that in an infectious situation peripheral blood cytokine may overflow into the CNS and contribute to behavioral alterations.

Altogether, our data indicate that persistence of T. gondii cysts in the CNS may stimulate intracerebral cytokine and CC-chemokine production, contributing to recruit inflammatory cells, thus sustaining neuroinflammation and BBB disruption, which may allow the leakage of inflammatory mediators into the brain tissue. Hence, in chronic toxoplasmosis the systemic and brain-born inflammatory milieu may contribute to behavioral changes, as anxiety, depression, and hyperactivity (S1 Graphical abstract). Therefore, multifactorial components shall be considered when proposing a therapeutic approach to hamper progression or to reverse mental disorders associated with chronic T. gondii infection.

Supporting information

Acknowledgments

We are indebted to Joyce Caroline Sá de Lima for the excellent technical assistance and the Multi-user Research Facility of Flow Cytometry—Multiparametric Analysis, Oswaldo Cruz Institute/Fiocruz, Rio de Janeiro, Brazil for sample acquisition.