Chronic Toxoplasma gondii infection and sleep‐wake alterations in mice

Abstract Aim Toxoplasma gondii (Tg) is an intracellular parasite infecting more than a third of the human population. Yet, the impact of Tg infection on sleep, a highly sensitive index of brain functions, remains unknown. We designed an experimental mouse model of chronic Tg infection to assess the effects on sleep‐wake states. Methods Mice were infected using cysts of the type II Prugniaud strain. We performed chronic sleep‐wake recordings and monitoring as well as EEG power spectral density analysis in order to assess the quantitative and qualitative changes of sleep‐wake states. Pharmacological approach was combined to evaluate the direct impact of the infection and inflammation caused by Tg. Results Infected mouse exhibited chronic sleep‐wake alterations over months, characterized by a marked increase (>20%) in time spent awake and in cortical EEG θ power density of all sleep‐wake states. Meanwhile, slow‐wave sleep decreased significantly. These effects were alleviated by an anti‐inflammatory treatment using corticosteroid dexamethasone. Conclusion We demonstrated for the first time the direct consequences of Tg infection on sleep‐wake states. The persistently increased wakefulness and reduced sleep fit with the parasite's strategy to enhance dissemination through host predation and are of significance in understanding the neurodegenerative and neuropsychiatric disorders reported in infected patients.

located in the central nervous system (CNS), eyes, and skeletal muscles. 2 Accumulated evidence indicates that this persistence of Tg may lead to physiological and behavioral consequences in infected rodents, resulting in reduced fear of Felidae their natural predators and even sexual attraction to their urine, [3][4][5][6][7][8][9] as well as decreased mechanisms of warning and anxiety, 10 and higher activity level. 11 Tg infection may also contribute to neurological and psychiatric symptoms in humans. Epidemiological studies suggested a possible link between Tg infection and schizophrenia 12,13 and other mood, behavioral, or neurological disorders. Several literature reviews confirmed the heterogeneity of designs and findings and failed to provide clear conclusions. Yet, these human studies have provided interesting clues, contradicting the current belief that chronic postnatal infection is asymptomatic in immune-competent subjects.
Surprisingly, little attention has been paid to sleep-wake cycle alteration. Yet, the large prevalence of sleep disorders in the general population closely matches that of toxoplasmosis, both estimated to be around 30%. 1,14 Moreover, sleep constitutes a highly sensitive index of brain functions, as sleep disorders are often found to be associated with mood, behavioral, and neurological disorders, even at their early stage. [15][16][17] Interestingly, Tg cysts are randomly distributed in all brain areas, but overlapping those that control the sleep-wake cycle. 18,19 Of note, neuromodulators known as sleep-wake regulators, such as melatonin and monoamines including norepinephrine, dopamine, and serotonin, are found to be markedly altered by Tg infection, indicating that it can affect vigilance states. 20 We therefore sought to establish a mouse model to unambiguously assess the consequences of chronic Tg infection on sleep and wakefulness in laboratory-controlled conditions, and thus, to highlight the impact of sleep alteration on behavioral, neurological, and neuropsychological functions in infected patients and to provide the required guidance for human studies.

| Animals
All experiments were conducted according to the European Ethics Community Directive for the use of research animals (2010/63/EC).
We followed the 3R principles to minimize the number of animals used, pain, and discomfort. The experimental protocol and procedures were approved by the local ethics committee for animal experimentation of Lyon 1 University (No. APAFIS#22363-2019100915103240).
Both experimental design and conduct of the study were performed strictly according to the ARRIVE guidelines. 23 Male, inbred CBA/Jrj mice were purchased from Janvier Labs, France with a certification as free of pathogenic viral, bacterial, and parasitic contaminants. Known to be susceptible to chronic infection by Tg, they were housed two per cage in our BSL2 animal facility (Hôpital de la Croix-Rousse, Lyon, France). Room temperature was maintained at 22°C (+/-2°C) with a 12 h light/dark cycle (light-on at 7am). Animals were monitored every day and handled regularly for habituation.

| Infection with Tg
All procedures manipulating infectious materials were performed in our BSL2 animal facility. Outbred female OF1 mice were also purchased from Janvier Labs, France. They were housed four per cage and were only used to multiply and perpetuate the Tg strain used for inoculation. We used Tg cysts of the canonical type II, avirulent PRU strain, representative of the European epidemiology. Considering its kystogenic profile, it allows to study the chronic phase of toxoplasmosis. We obtained the initial batch (Strain TgH 00001) from the French National Reference Center, as a brain homogenate of female Swiss Webster mice that had been infected intraperitoneally. The homogenate was diluted in sterile saline. OF1 mice were chronically infected by intraperitoneal injection of five cysts per mouse to maintain the Tg strain. On the day of inoculation, infected OF1 mice were euthanized under deep anesthesia. Their brains were removed, then homogenized in sterile saline. Using this homogenate, we orally infected 11-to 14-week-old CBA/Jrj mice by gavage (20 cysts) to mimic the natural route of infection and avoid possible peritoneal inflammation. 24 Control CBA/Jrj mice were inoculated in the same fashion with a brain homogenate prepared from healthy, non-infected OF1 mice. In order to study the chronic phase of infection, inoculated and control mice underwent surgery as described below between 18 and 22 weeks of age-at least eight weeks after inoculation-and with a body weight of 25-35 g. Infection was verified by the presence of blood-borne anti-Tg antibodies, using an ELISA kit (bioMérieux, France), between 4 and 5 weeks post-infection (pi).

| Surgery
Control and infected CBA/Jrj mice were anesthetized using isoflurane (oxygen flow rate 200 mL/min, 3%-4% initially, maintenance at 1%). We placed each mouse in a stereotaxic apparatus with a heating pad maintaining the animal rectal temperature at 37 ± 1°C. After shaving, the skin over the skull was rubbed with iodine, sectioned longitudinally, and reclined to expose the skull.
Four holes (Ø = 0.5 mm) were drilled in the skull at the following coordinates: frontal (1 mm lateral and anterior to bregma) and parietal (1 mm lateral to the midline at the midpoint between bregma and lambda) cortices of both left and right sides. These procedures preserve the integrity of all brain-protecting membranes (dura and pia, particularly). Four cortical electrodes (gold-plated, tinnedcopper wire, Ø = 0.4 mm; Filotex, Draveil, France) were placed into these holes for electroencephalogram (EEG). Three muscle electrodes (gold-plated fluorocarbon-coated stainless-steel wire, Ø = 0.03 mm; Cooner Wire, Chatworth, CA) were inserted into the neck for electromyogram (EMG) recordings. EEG was differentially recorded between two electrodes overlying frontal and parietal cortice of one side. These electrodes were priorly soldered to a multichannel electrical connector, and each was separately insulated with heat-shrinkable polyolefin/polyester tubing. Finally, the electrode assembly was anchored to the skull with Super-Bond (Sun Medical Co., Shiga, Japan) and embedded with dental cement.
The skin was sutured around implant. After subcutaneous injections of Carprofen (Rimadyl, 5 mg/kg) and sterile saline containing 5% glucose, animals were warmed until recovery. We kept each mouse under close observation for another six days: weight, behavior, eating, and drinking activities were monitored. Our implantation procedures allowed stable polygraphic recordings for more than six months.

| Polysomnographic data acquisition and analysis
After surgery, each mouse was housed individually in a transparent barrel (Ø 20 cm, height 30 cm) in a ventilated animal cabinet located in a sound-proof recording room with an ambient temperature of 22 ± 2°C, a 12 h light/dark cycle (light-on at 7am), food and water ad libitum. Mice were habituated to the recording cable connecting implant to a swivel connector for five days before starting polysomnographic (EEG and EMG) recordings.
Fronto-parietal EEG and EMG signals were amplified and filtered (bandwidth 0.3-100 Hz for EEG, 10-100 Hz for EMG). These signals were digitalized with CED 1401 converter (Cambridge Electronic Design, UK) with a sampling rate of 512 Hz using Spike2 software (Cambridge Electronic Design). We used Sleepscore software (ViewPoint, Lyon, France) to determine the vigilance states every 5-sec epoch: WK, SWS, and PS, according to previously described criteria. [25][26][27] All sleep-wake recordings were scored blind.
Total durations (±SEM) of each vigilance state were calculated for light (7am to 7 pm), dark (7 pm to 7am), and 24 h periods. For each vigilance state, the total number of episodes and the mean duration (±SEM) of one episode were calculated for each period.
All durations were expressed in minutes. In order to evaluate the contrast of sleep-wake states between light and dark phases and as a criterion of sleep-wake circadian rhythms, we used the light/dark ratio for sleep stages (dark/light for WK), obtained by 12 hours sleep-wake amounts before lights-off, divided by those after lights-off.
In addition to quantitative analyses, a qualitative approach consisted in assessing the possible difference of EEG power spectrum between infected and uninfected groups. EEG power spectra were computed every 5-sec epochs for the frequency range of 0.5-40 Hz using fast Fourier transform. For each vigilance state, the total power of densities was obtained over the frequency range of 0.5-40 Hz for each state of vigilance by adding all epochs of a given state. All power spectral densities were standardized at the different frequency ranges, which were expressed as percentage relative to the total power observed for each vigilance state.

| Experimental procedure
We performed three sets of independent experiments to minimize the possible variation in the effectiveness of inoculation. For all sets, we simultaneously recorded infected vs. uninfected control mice.
We used two infected vs. two uninfected control mice in the first, pilot batch, and two controls vs. six infected mice for each one of the two subsequent sets. In total, six control and 14 infected mice were subjected to the following experimental procedures.

| Spontaneous cortical EEG and sleep-wake recordings
After recovery from surgery and habituation to the recording cables, each mouse was recorded for two days. We repeated such recording sessions two to three times a week. Our recordings could last up to six months to evidence chronic alterations in sleep/wake cycle, as the pi period may be important to reveal possible differences between control and infected mice.

| Cortical EEG and sleep-wake cycle after pharmacological administration
Using pharmacological tools, we tentatively evaluated whether Tg infection causes sleep-wake effects by affecting surrounding tissue via inflammatory process or by direct Tg activity or excretion.
The anti-inflammatory corticosteroid DXM (dexamethasone sodium phosphate) was administered at a dose of 2.5 mg/kg/day (5 mg DXM per liter of drinking water) for seven consecutive days. 28 A longer period of administration has been shown to lead to a immunodepression which may reactivate Tg. 29 Sulfamethoxazole-Trimethoprim (SXT), an antimicrobial synergistic compound (Bactrim®) exhibiting activity against replicating forms of the parasite, is generally used against the effects of the acute phase of Tg infection. It was administered in drinking water respectively at a dose of 95 mg/kg/day and 19 mg/kg/day for 10 consecutive days. 28 Treatments were separated by a wash-out of 2 weeks ( Figure 1).

| Statistical analysis
We used GraphPad Prism 7.0 for all analyses. All experimental data were firstly examined for their value distribution using d'Agostino Pearson normality test. When the data were non-normally distributed, nonparametric Mann-Whitney tests were used. Two-way ANOVA and unpaired Student's t test were used when the data were normally distributed. Two-way ANOVA with repeated measures for both factors was used to evaluate the differences between control and infected mice with pi time. Drug effects were evaluated with pi time using the same procedure. These tests were followed by post hoc Holm-Sidak test for multiple pair-wise comparisons. All data are expressed as the mean ± SEM. All tests used were two-tailed with a significance level of alpha set at 0.05.

| General observations
Infected mice appeared to develop normally as compared to uninfected mice. No obvious abnormalities were detected in terms of general morphology, food intake, reactivity when being handled, or other behaviors under baseline conditions. Yet, infected mice appeared to execute more locomotive activities than uninfected mice as shown by our electromyogram (EMG) recordings ( Figure 2).
As reported previously, 30 a significant difference in body weight appeared at week 2 pi (−1.7 g weight gain vs +1.7 g in uninfected mice; n = 10 for both groups, P = 0.0159, nonparametric Mann-Whitney test). No difference in lifespan was observed.

| Increased wakefulness and decreased slowwave sleep in Tg-infected mice
In baseline conditions, Tg mice exhibited a circadian sleep-wake rhythm quite characteristic of CBA/Jrj mice, with increased activity ( Figure 2) and a larger amount of waking (WK) during the dark phase ( Figure 3). Compared to uninfected mice, the total amount of WK) in Tg mice was significantly increased over all light-dark phases. This effect was due to a significant increase in both mean episode  (Table 1).
Concomitantly, the total amount of slow-wave sleep (SWS) was decreased over all phases (P < 0.0001). The decreased SWS was characterized by a decrease in episode duration and an increase in episode number (Table 1) indicating sleep fragmentation in Tg mice. Such sleep fragmentation occurred more prominently during the light phase (the primary sleep period) than during the dark phase (when mice are more aroused) ( Table 1). No fragmentation was noted with regards to WK.
As revealed by hourly analysis of the sleep-wake amounts, the increase in WK and decrease in SWS were greater during the dark phase   . P value is calculated using two-way ANOVA with repeated measures to evaluate the differences between control and infected mice. All tests used were two-tailed with a significance level of alpha set at 0.05.

| Time course of sleep-wake changes induced by Tg infection
The

| Qualitative alterations in power spectral density of cortical EEG during sleep-wake states
The distribution of cortical EEG power density and their spectral morphology during all light-dark phases were quite similar between Tg and WT mice. Yet, we noticed a markedly increased power spectral density in the θ rhythm (6-9 Hz) for all sleep-wake states, notably WK, as well as a decrease in low δ rhythm (1-2 Hz) ( Figure 5). Of note, the differences seen during PS were less important than those seen during SWS and WK.

| Effects of pharmacological administrations on the sleep-wake changes caused by Tg infection
Since the major effect of Tg infection was enhanced WK, we attempted to identify its possible origin. Notably, such an effect could be due to a direct aggression on the brain tissue by Tg cyst or tachyzoite or to inflammatory reactions around the infected loci causing such longlasting WK enhancement. Data were analyzed globally over the whole course, that is, from the first to the last days of both treatments to take into account possible mechanisms of compensation. However, we observed the same effects on all periods (light, dark, and 24 h periods).

| A chronic murine model of Tg infection and experimental considerations
Sleep-wake impairment is often associated at an early stage with mood, behavioral, and neurological disorders. 15

| A long-lasting, marked and rarely observed enhancement of wakefulness
The main original finding of our study is that chronic Tg infection significantly alters the sleep-wake cycle in the mouse. These alterations consisted in a markedly increased amount of wakefulness, particularly during the dark period. SWS was concomitantly decreased at all periods. We found that the anti-Tg drug SXT had no clear impact on the sleep-wake effects after Tg inoculation, presumably because the direct impact of the parasite at this late chronic stage (ie, > month 3 pi) was already stable and irreversible. In contrast, the corticosteroid anti-inflammatory DXM reversed the increased WK quantitatively but not qualitatively. These pharmacological data seem to be quite conclusive for the involvement of an inflammatory process in the Tg-induced WK enhancement rather than that of the parasite itself.
Our results are in line with previous findings regarding Tg infection. First, sleep is an orchestrated neurochemical process on different brain networks 38 and involves many neurotransmitters and modulators, notably monoamines, whose balance was found to be modified after Tg infection. Indeed, Tg infection increases the level of dopamine but not that of serotonin and norepinephrine, 22 indirectly via the release of inflammatory cytokines, but mostly directly via the presence in its genome of genes coding for phenylalanine hydroxylase and tyrosine hydroxylase. 39 Figure 7).

| A putative strategy of Tg to promote its dissemination
This persistently increased WK and reduced sleep in mice with chronic Tg infection have not been previously reported but fit well with its suspected strategy by which it would promote its own dissemination by enhancing WK in order to facilitate encounters with their predators. 8,49 Increased WK was mostly noticed during dark phases when F I G U R E 7 Graphic abstract illustrating the main findings of the present study and our interpretations. When mice are infected chronically by Toxoplasma gondii, an intracellular parasite, they exhibit a marked increase in wakefulness and activities, thus promoting encounters with cats, its natural predators. It is likely that by this strategy to modify its hosts' sleep-wake and behavioral states, Toxoplasma gondii would enhance its own dissemination, thus explaining why it infects a wide range of animals and about 1/3 of the world population [Colour figure can be viewed at wileyonlinelibrary.com] increased activity is more likely to favor predation of nocturnal animals like mice. Reduced sleep duration at all phases would also contribute to keeping animals more opportune (Figure 7). Sleep and WK highly impact vigilance and attention processes. Although a lower attention and longer reaction time have been reported in Tg-infected animals 50,51 and man, 52-54 their links with WK alteration were not reported. The impact of this WK enhancement on vigilance and attention remains therefore to be addressed with specific behavioral tests.
Since the enhanced WK was stable over the whole periods, it would last for the host's entire lifespan, as a permanent outcome of infection. Intriguingly, they were not immediately at their strongest. Such progressive increase with the time elapsed pi was also reported in humans: the intensity of personality trait shift correlates with estimated duration of infection. 55 Manifestations of delayed sleep onset and restless legs syndromes that were observed in our patients with postnatal or congenital Tg infection tend to develop at least six to 10 years pi (Lyon cohort).
Considering sleep-wake cycle as a major brain function targeted by Tg to manipulate the behavior of infected hosts might also explain why our results are contrasted with other infection models that tend to report an increased sleep duration, suspected to be the consequence of increased interactions between inflammatory cytokines and brain sleep-regulating structures. This was the case for acute systemic infections, 56 including influenza, 57 but also in chronic infection. Herpes virus is reported to be associated with increased sleep duration and fragmentation. 58-60

| Medical significances and perspectives
We believe that the combination of parasite/host (Prugniaud/male clinically confirmed, would also promote serological or molecular detection of Tg in patients suffering from sleep disorders and prevention campaign against toxoplasmosis, the importance of which is sometimes debated. Such possible underestimation of risk associated with Tg infection may also concern the current worldwide debate on the cost-effectiveness of prenatal screening programs.
Altogether, we have established a novel, currently unique and insightful murine model of Tg infection suitable for basic and translational investigations. The use of this model has allowed us to demonstrate the Tg-induced sleep-wake alterations, notably a remarkable long-lasting enhancement of WK, with which Tg would likely promote its own propagation. Our study also demonstrates sleep-wake assessment as a potent and pertinent approach for long-term follow-up of infectious events, that is not always expected or sought.

ACK N OWLED G M ENTS
They would like to thank Dr Laurent Seugnet for critical reading of the manuscript and Mrs Grange for expert editorial assistance and English editing. We are grateful to the CRB Toxoplasma from the French National Reference Center for providing the Tg strain.

Graphical abstract was designed with contribution from Dr Florence
Persat (T. gondii picture) and designs by ilonitta/ Freepik.

CO N FLI C T O F I NTE R E S T
The authors declare no competing financial interests.