Calorie restriction partially attenuates sickness behavior induced by viral mimetic poly I:C

Calorie restriction (CR) has been shown to extend the mean and maximum lifespan in both preclinical and clinical settings. We have previously demonstrated that CR attenuates lipopolysaccharide (LPS)-induced fever and sickness behavior. CR also leads to reductions in pro-inflammatory and increases in anti-inflammatory profiles. LPS is a bacterial mimetic; however, few studies have explored this phenomenon utilizing a viral mimetic, such as polyinosinic:polycytidylic acid (poly I:C). Dose-dependently, poly I:C induced an increase in core body temperature (T b ), with the largest dose (5000 µ g/kg) resulting in a 1.62 ◦ C ( ± 0.23 ◦ C) T b increase at 7 h post-injection in ad libitum mice and was associated with reduced home-cage locomotor activity. We then investigated the effect of 50% CR for 28 days to attenuate fever and sickness behavior induced by a poly I:C (5000 µ g/kg) viral immune challenge. CR resulted in the partial attenuation of fever and sickness behavior measures post-poly I:C. The freely fed, control mice demonstrated a 2.02 ◦ C ( ± 0.22 ◦ C) increase in T b at 7 h post-injection compared to the CR poly I:C group which demonstrated an increase in T b of 0.94 ◦ C ( ± 0.27 ◦ C). Locomotor patterns post-injection were different, CR mice displayed a reduction in activity during the light phase, and the control group displayed a reduction during the dark phase. CR moderately attenuated the neuroinflammatory response with a reduction in microglial density in the ventromedial nucleus of the hypothalamus. The fever and sickness behavior attenuation seen after CR may be driven by similar anti-inflammatory processes as after LPS; however, further investigation is required.


Introduction
Calorie restriction (CR), a dietary regimen with reduced calories, protein, vitamins and minerals whilst avoiding malnutrition [1] has been shown to extend the mean and maximum lifespan in a variety of animals and to delay the onset of immunosenescence and age-related diseases [2][3][4][5].CR dose-dependently attenuates lipopolysaccharide (LPS)-induced fever and sickness behavior, with a 50% CR for 28 days attenuating LPS-induced febrile response, reduced locomotor activity, and anorexia [6].Sickness behavior is a collection of brain-controlled symptoms that develop in sick individuals and animals that typically include anorexia, fatigue, reduction in locomotor activity, social withdrawal, malaise, hyperalgesia, sleep disruptions and fever, all of which are mediated via pro-inflammatory cytokines [7,8].This effect was dose-dependent; a 25% restriction or a 50% CR for shorter periods only resulted in a partial attenuation of sickness behavior [9].CR does not appear to permanently impact the animals' ability to develop a febrile response, indicating an intact thermoregulatory system [10].Evidence suggests that CR can exert potent anti-inflammatory effects and modulate the distribution of immune cells during a steady state and their activation during infections [11,12].However, the majority of the literature has focused on immune response changes in CR animals following a bacterial challenge with LPS; a Gram-negative bacterial mimetic.There is limited literature investigating whether an immune response to a viral immune challenge, such as from the viral mimetic polyinosinic:polycytidylic acid (poly I:C), would be similar to that observed after a bacterial challenge.
The effects of CR on sickness responses appear to be due to a reduction in pro-inflammatory cytokine expression in the hypothalamus coupled with increased anti-inflammatory cytokine profiles [6].Suppression of cytokine signaling 3 (SOCS3) and interleukin (IL)− 10; both anti-inflammatory agents, were increased, whereas, cyclooxygenase-2 (COX-2) and leptin; which both demonstrate pro-inflammatory effects and are associated with fever generation [13,14], were reduced in CR mice post-LPS injection [6].The anti-inflammatory agent, SOCS3, inhibits the production of IL-6, a crucial component for COX-2 production [15,16], as well as accentuating the anti-inflammatory effects of IL-10 [17].Moreover, in the periphery, CR rats demonstrated increases in plasma corticosterone (known to have anti-inflammatory properties [18]) and decreases in serum IL-6 post-LPS [9].
Infection with LPS activates the immune system via a specific pathway, thus, it is unclear whether the above findings can be universally applied to all pathogen-induced sickness behavior [19].Different pathogens have different molecular patterns recognized by different toll-like receptors (TLR) on the surface of macrophages [20,21].For instance, TLR4 is the only TLR to recognize the LPS molecular pattern and initiates a unique chain of responses [19], including the myeloid differentiation primary response 88 (MyD88) dependent and Toll/IL-1 receptor domain-containing adaptor inducing interferon-β (TRIF) dependent pathways [22].MyD88 and TRIF are responsible for the secretion of pro-inflammatory cytokines such as IL-1, IL-6, and tumor necrosis factor-alpha (TNF-α); ultimately initiating sickness behavior [23].
Whilst TLR4 activation is exclusively driven by the MyD88 and TRIF pathways [24], other TLR pathways, which are activated in response to other stimuli, utilize different features of the immune system to initiate an inflammatory response [22].TLR3 can be activated by the viral mimetic, polyinosinic:polycytidylic acid (poly I:C), a synthetic double-stranded RNA, that is used extensively to assess in vivo immune response as it mimics the replication intermediates present in cells infected with RNA viruses [20,23].Systemically administered poly I:C, induces the classical features of sickness behavior including fever/hypothermia, malaise, hypoactivity, anhedonia, disruption of circadian rhythms, and hypophagia [17,20].Furthermore, the use of poly I:C is safer and improves reproducibility when compared to the use of live viruses.Unlike TLR4, TLR3 responds specifically to double-stranded RNA, a by-product of viral DNA transcription and a replication intermediate of single-stranded RNA viruses [25].Interestingly, TLR3 differs from other pathways as it does not engage the MyD88-dependent pathway; instead solely activating the TRIF pathway [24].The TRIF pathway initiates similar cellular reactions as the MyD88-dependent pathway; both result in pro-inflammatory cytokine expression, leading to the proposition that similar cytokine profiles are also expressed post-TRIF activation [23].In consideration that bacterial and viral infections activate the immune system via different mechanisms, it is valuable to discern their respective effects after CR on sickness behavior.Additionally, to also distinguish whether changes observed post-LPS/bacterial challenge are similar to viral infections.
Calorie restriction has been demonstrated to exert positive effects upon life span and certain age-related diseases, however, it is uncertain whether the suppressed responses to immune challenges such as LPS are adaptive and beneficial.Sickness behaviors, such as fever following a bacterial challenge, are adaptive allowing proper acute proinflammatory responses to fight off the infection which is not seen in CR rodents.Thus, it is important to assess whether CR reduces levels of sickness behavior following a viral infection with poly I:C.The present study aimed to assess the potential for CR to attenuate sickness behavior induced by a TLR3 agonist, poly I:C.A dose-response experiment was conducted to assess the optimal dose of poly I:C to induce fever and sickness behavior.We then characterized fever and sickness behavior induced by the optimal dose of poly I:C in CR and control mice.We hypothesized that CR would attenuate poly I:C-induced sickness behavior, including fever, reduced locomotor activity, weight loss and anorexia and this would be associated with a reduced neuroinflammatory profile.

Animals
Sixty-three male C57BL/6 J mice (aged 8-14 weeks) were obtained from Animal Resources Centre (Canning Vale, WA, AUS) and allowed to acclimate at the facility for one week.Mice were housed individually in 30 × 16 × 12 cm polypropylene basin cages with sawdust and shredded paper provided as bedding and cardboard rolls supplied for environmental enrichment.They were maintained on a 12:12 light-dark cycle (lights on: 0700-1900) and at an ambient temperature of 30 ± 2 • C, which is within the thermoneutral zone for this species [26,27].Mice were fed standard rodent chow (Barastoc, Melbourne, VIC, AUS) and provided water ad libitum (AL) during acclimation and surgery recovery periods.All experiments were conducted at La Trobe Animal Research and Teaching facility and according to La Trobe University Animal Ethics Committee approved protocols, Approval Number AEC17-30.

Biotelemetry surgery and determination of sickness behavior
Following acclimation, a biotelemetry device (G 2 E-mitter®, STARR Life Sciences Corp, Oakmont, PA, USA; 1.1 g; 15.5 ×6.5 mm) was surgically implanted under aseptic conditions into the peritoneal cavity of all mice.Prior to surgery, mice were subcutaneously administered carprofen (5 mg/kg); a non-steroidal anti-inflammatory agent, and bupivacaine (2 mg/kg); a local anesthetic, at the site of the incision.Mice were anesthetized with isoflurane (induction: 5% isoflurane, flow rate 1 L/minute; maintenance: 2.5-3% isoflurane, flow rate 0.5 L/minute).The incision site was shaved and sterilized three times with 4% chlorhexidine and once with 0.5% chlorhexidine in ethanol (70%).A 1-2 cm incision was made in the skin and peritoneal muscle wall and the device was inserted.The muscular incision was closed using non-absorbable 5-0 sutures (Silkam®, B. Braun, Bella Vista, NSW, AUS) and the skin was sealed using 7 mm wound clips.Mice were placed on a heating pad to maintain T b during surgery and post-surgery until conscious.Carprofen was administered for two days post-surgery and mice were allowed a minimum of seven days to recover before proceeding with the experiment.
To measure T b and locomotor activity each mouse cage was placed on a telemetry receiver (ER-4000, STARR Life Sciences Corp., Oakmont, PA, USA).The implanted biotelemetry emitter device generated a continuous unique frequency signal proportional to a mouse's T b ( ± 0.1 • C).The receiver sampled this frequency continuously, which was decoded by VitalView® software (STARR Life Sciences Corp., Oakmont, PA, USA) at one-minute intervals that converted frequencies to core temperature by using predetermined calibration values.Locomotor activity was also determined by the receivers, which were equipped with a matrix of antennae, scanned in sequential order to locate the position and orientation of the biotelemetry device.VitalView® software recorded the number of matrices crossed by the mouse in one-minute intervals.

Poly I:C dose-response
To assess the most effective dose to elicit a 1 • C change in T b , we assessed five different poly I:C doses.Briefly, high molecular weight poly I:C (InvivoGen, San Diego, CA, USA) was resuspended in pyrogen-free saline and heated to 70 • C to improve solubility and then allowed to cool for 1 h at room temperature (RT) to ensure proper annealing of double-stranded RNA.Following the surgery recovery period, 31 mice were assigned into five groups, four poly I:C groups (500, 1000, 2000, and 5000 μg/kg poly I:C dissolved in sterile saline) and a control group (sterile saline).On the day of injection, mice were weighed approximately one hour after lights-on (0800 h) and then returned to their home cage.Two to three hours after lights-on (0900 and 1000 h), poly I:C (either 500, 1000, 2000 or 5000 μg/kg) or saline was administered intraperitoneally (i.p.) and mice were returned to their home cage for measurement of T b and locomotor activity for 24 h.All injections were in a volume of 150 μL/25 g.

Dietary regimens
The CR group's food intake was calculated weekly based on the average daily food intake of the AL group for three consecutive days.Food intake was determined to the nearest 0.1 g by providing a set amount of food each day and weighing the remaining food 24 h later.Food for all groups was provided daily approximately 30 min before lights-out (i.e., 1830 h) and dietary manipulations continued for 28 days prior to poly I:C injection.Dietary manipulations were maintained whilst poly I:C-induced sickness behavior and fever data was collected.Water was provided ad libitum to all groups.The dietary composition of the AL and CR dietary regimens has been published previously [28].

Poly I:C effects after 28 days of CR
On the 29th day of dietary manipulation, mice (and food provided to AL animals) were weighed 30 min after lights-on (0730 h) and returned to their home cage.Two to three hours after lights-on (0900 and 1000 h), 5000 μg/kg poly I:C (dose determined by dose-response study) or sterile saline was administered intraperitoneally, and mice were returned to their home cage for measurement of T b and locomotor activity for 24 h.All injections were in a volume of 150 μL/25 g.

Microglial immunohistochemistry
At the end of telemetry recording, 24 h after either poly I:C or saline injections for both dose-response and CR experiments, the animals were deeply anesthetized with Lethabarb (approximately 150 mg/kg pentobarbitone sodium i.p.).The mice were then transcardially perfused with phosphate-buffered saline (PBS: 4 • C, pH 7.4), followed by 4% paraformaldehyde in PBS (4 • C, pH 7.4).All experiments took place between 0900 and 1300 h to limit the potential effect of circadian rhythms on any parameters measured.After 2 h of post-fixation in the same fixative, the brains were removed and placed in 30% sucrose in PBS (4 • C) for approximately 48 h.Brains were blocked into forebrain and hindbrain sections with a coronal cut at the caudal border of the mammillary bodies (approx.− 3.52 mm bregma) using a mouse brain matrix (Cat No. RMBS-200 C, Kent Scientific Corporation, Torrington, CT, USA), and rapidly frozen in isopentane cooled on dry ice and stored at − 80 • C. We proceeded to cut the forebrains into 30 μM coronal sections using a cryostat.Sections were cut in a one-in-six series and stored at − 20 • C until use.
Briefly, one of the one-in-six series of sections from each animal was used.At least three mice from each treatment group were processed at the same time in batches.To remove endogenous peroxidase activity, brain sections were immersed in 1% hydrogen peroxidase (H 2 O 2 ) in 0.05 M PBS (15 min at RT), followed by a 0.3% Triton X-100 in PBS for 15 min at RT.The sections were incubated in the primary antibody (Iba-1, 1:2500, rabbit, RT, overnight [approx.16 h], Wako Pure Chemical Industries, Ltd., Osaka, Japan).This was followed by a secondary antibody, at RT, for 90 min (biotinylated goat anti-rabbit IgG; 1:200, Vector Laboratories, Burlingame, CA, USA).Afterwards, sections were incubated in avidin-biotin horseradish peroxidase (HRP) complex (ABC; min; 1:200; Vector Laboratories), followed by 0.25 mg/mL 3,3′-diaminobenzidine tetrahydrochloride (DAB; Sigma-Aldrich) and 0.012% H 2 O 2 in 0.1 M PBS to visualize the HRP activity, seen as amber staining.We stopped the reaction when the contrast between specific cellular and non-specific background labeling was optimal (approx.10 min).Sections were mounted onto gelatin-coated slides, air-dried, dehydrated in a series of alcohols, cleared in xylene and coverslipped.
An experimenter blinded to treatment conditions assessed the sections for numbers of cells and morphology Iba-1 labeling as described [34], using photomicrograph images imported into image analysis software ImageJ (National Institutes of Health, Bethesda, MD, USA).Briefly, we took all photomicrograph images from a Nikon 90i upright microscope (Nikon, Melville, NY, USA) with a 20 times objective lens using a Nikon DS-Fi1 digital camera and NIS Elements Advanced Research software (Nikon).Images were taken at 2560 × 1920-pixel density.Using Image J, each image was converted to 16-bit for analysis.For each region, we selected a sub-region of interest, identified according to the Paxinos adult mouse atlas [35], and analyzed three-four sections, 180 µm apart as our sampled result for each animal.Paraventricular nucleus of the hypothalamus (PVN): three sections between -0.58 and − 1.06 mm bregma; arcuate nucleus of the hypothalamus (ARC): four sections between -1.58 and − 2.06 mm bregma; lateral hypothalamus (LH): four sections between -1.70 and − 2.54 mm bregma; and the ventromedial nucleus of the hypothalamus (VMH): four sections between -1.06 and − 2.06 mm bregma).We took the summed counts of the Iba-1 positive cells from 3 to 4 sections.All positively immunolabeled cells with a clearly visible nucleus were considered as Iba-1 positive.For morphology, all cells were categorized as ameboid (0primary processes), intermediate (2 -4 primary processes) or ramified (five or more primary processes) as described previously [34].

Statistical analysis
Data was analyzed using GraphPad Prism Version 9 (GraphPad Software, San Diego, CA, USA).To determine differences in T b for both the dose-response and CR experiments T b data for the 4 days immediately preceding the injection was collapsed into hourly means to use as a baseline measure.Changes from baseline were calculated by comparing each hour following injection with the corresponding hourly baseline value.A similar process was conducted for locomotor activity data; however, instead of hourly averages, averages for the light and dark cycles were calculated.T b data was analyzed using a mixed-design ANOVA, with drug and diet as the between-subjects factor and time as the within-subjects factor.Locomotor activity data was analyzed using a mixed-design ANOVA using immune challenge and diet as betweensubjects factors and time as a within-subjects factor.Percentage changes in body weight were also analyzed using a two-way ANOVA.The percentage of food intake was calculated by comparing the individual animal's average food intake over a 4-day period prior to poly I:C challenge to their food intake 24 h after poly I:C challenge and analyzed using two-way ANOVA.For Iba-1 positive cell analysis, we analyzed summed counts and morphology for each region using a two-way ANOVA with diet and immune challenge as between-subject factors.Where significant main or interaction effects were observed, post hoc analysis was performed using Tukey's multiple comparisons test.

Poly I:C injections affect core T b
Poly I:C induced a significant dose-dependent increase in core T b over time (significant interaction between dose and time (F (32,208) = 4.99, p < 0.001; Fig. 1a).Poly I:C induced a mild T b increase in the 500 µg/kg (1.15 ± 0.19 • C) and 1000 µg/kg (0.83 ± 0.17 Both the saline and 500 µg/kg groups were no different compared to baseline; however, the 1000 µg/kg group demonstrated a higher T b compared to baseline at 7-and 8-hours post-injection (p = 0.019 and 0.043 respectively).The 2000 µg/kg group also demonstrated a higher T b compared to baseline at two-time points, 3-and 7-hours post-injection (p = 0.044 and 0.036 respectively) and the 5000 µg/kg group differed from baseline at 3-, 4-, 7-, and 8-hours post-injection (range p = 0.036-0.004).

Locomotor activity is attenuated during the dark phase following poly I:C administration
All poly I:C concentrations and saline mice demonstrated minimal reductions in locomotor activity during the light phase post-injection (F [8,52] = 2.45, p = 0.044; Fig. 1b).However, during the dark phase, the saline and 500 µg/kg groups demonstrated a moderate reduction in locomotor activity, whereas reductions were larger in the 1000, 2000 and 5000 µg/kg groups.When compared to the baseline, no group demonstrated a difference compared to baseline during the light phase; however, during the dark phase, the saline (p = .025),500 µg/kg (p = 0.008), 1000 µg/kg (p = 0.011), and 5000 µg/kg (p = 0.001) groups all demonstrated a significant reduction in locomotor activity from baseline.We have demonstrated that a 5000 µg/kg dose of poly I:C has a pronounced effect on both core T b and locomotor activity, thus was used to examine sickness behavior in CR and control (AL) mice.

Neuroinflammatory response to poly I:C is partially attenuated in CR mice
We assessed four regions of the hypothalamus known to be involved in sickness responses.Calorie restriction decreased the number of microglia in the arcuate nucleus of the hypothalamus compared to AL mice (significant interaction between immune challenge x diet: F[1,7] = 6.433, p = 0.038; Fig. 4b), however, there were no other significant main effects (immune challenge main effect:: F[1,7] = 0.183, p = 0.682; diet main effect:: F[1,7] = 0.338, p = 0.579) or post hoc differences observed in the arcuate nucleus or in other brain regions assessed.
To characterize the hypothalamic microglia, we assessed microglial morphology.We found indications of increased susceptibility of the central effects of a poly I:C immune challenge in the arcuate nucleus of the hypothalamus microglia in CR compared to AL mice (significant main effect of diet: F [1,5] = 7.71, p = 0.038; Fig. 5b, f, j).These CR mice displayed an increase in the percentage of microglia classified as ramified.In the lateral hypothalamic region, we saw a similar pattern with fewer intermediate microglia and more ramified microglia (intermediate: significant main effect of diet: F [1,7] = 10.75, p = 0.013; Fig. 5h and ramified: significant main effect of diet: F [1,7] = 27.99,p = 0.001; Fig. 5l).

Discussion
The present study characterized poly I:C-induced sickness behavior with the primary aim of determining whether CR could attenuate poly I: C-induced sickness behavior.In a dose-response experiment, increasing  * denotes a significant difference from both saline groups at p < 0.05, * * denotes a significant difference from both saline groups at p < 0.01, * ** denotes a significant difference from both saline groups at p < 0.001, # denotes a significant difference from the CR poly I:C group at p < 0.05, ## denotes a significant difference from the CR poly I:C group at p < 0.01.doses of poly I:C induced a dose-dependent increase in core T b , with the highest dose of poly I:C (5000 µg/kg) inducing the largest rise in core T b, with a peak T b at 7 h post-injection (1.62 ± 0.23 • C).A similar pattern was observed with a change in locomotor activity; all poly I:C groups demonstrated a reduction in activity during the dark phase postinjection.Utilizing the highest dose of poly I:C tested (5000 µg/kg), we determined that CR partially attenuates poly I:C-induced sickness behavior, with a reduced increase in T b , locomotor activity and weight loss compared to AL mice.We also demonstrate a mild attenuated neuroinflammatory response to an immune challenge with poly I:C in the CR mice compared to AL mice.
The CR experiment, utilizing the highest dose of poly I:C tested in the dose-dependent experiment (5000 µg/kg), determined that CR partially attenuated poly I:C-induced sickness behavior.The AL poly I:C group demonstrated the largest increase in T b , peaking at 7 h post-injection (2.02 ± 0.22 • C), similar to the dose-response experiment.In comparison, the CR poly I:C group only demonstrated an intermediate response, with a peak T b increase of 0.94 ± 0.27 • C also at 7 h post-injection.These results may be somewhat confounded by the timing of the poly I:C fever peak; 7 h post-injection is also the same time when the CR animals demonstrate CR-induced anticipatory behavior, which is a wellknown outcome of the entrainment process initiated via a set meal time [6,36,37].However, in this experiment baseline T b was averaged over the previous 4 days for hourly means, indicating the increase of T b is genuine and the effect of CR-induced anticipatory behavior was taken into account.The CR animals demonstrated a reduction in locomotor activity during this timeframe (discussed below), it is still possible that the increase in T b seen in the CR animals is somewhat masked by an activity-associated increase in T b during the same timeframe as locomotor activity can act as a thermoregulatory effector [38].
In contrast, while we show that CR can impair sickness behavior in response to a poly I:C challenge, there are studies showing increased susceptibility of CR animals to infection by intact pathogens such as influenza A virus and nematode worm (Heligmosomoides bakeri) infection despite similar or enhanced immune system function [39][40][41].
Thus, inferring that CR animals may respond differently to viral and bacterial infections, via different mechanistic pathways, and thus lead to different effects on sickness behavior.CR mice intranasally infected with influenza A decreased survival, increased virus titers, and reduced natural killer cell activity, responsible for recognition of virally infected cells, in the lungs and spleen [39,42].The researchers postulate that the decreased survivability and increased susceptibility to influenza infection in CR mice may not have the necessary energy reserves to meet the demand associated with this particular viral infection.To test this hypothesis, the research group followed with a study that showed that the commencement of 2 weeks AL feeding after CR increased body weight, improved survivability and mitigated the decline of natural killer cell function after influenza infection [43].Although the previously mentioned studies show negative effects of CR in response to infection, these outcomes depend on the type of infectious agent and should be tempered with additional research.As CR was shown to decrease mortality rate, and reduce organ damage and inflammatory response after sepsis and salmonella infections [44,45] and we show that CR attenuates sickness behavior in response to the viral mimetic poly I:C.
It is noteworthy that the pattern of locomotor activity changes postpoly I:C between AL and CR mice is opposite.During the light phase immediately post-injection, the CR poly I:C group was the only group to demonstrate a reduction, which could be due to CR animals having an attenuated sickness behavioral response with activity returning to baseline faster than AL animals.More specifically to the AL groups, the light phase presents as a confounder of a floor effect for rodents, for which it is their resting phase and they demonstrate already very limited levels of activity [46].During the dark phase post-injection this pattern reversed, with the CR poly I:C group returning to baseline activity levels, and the AL poly I:C group demonstrating a large reduction in activity; a pattern which has been demonstrated previously in CR mice administered LPS [6].To our knowledge, there has been only one other study that has investigated poly I:C / sickness behavior after a form of CR and found the opposite result, where 9 days of intermittent (alternate day) fasting exaggerated reductions of locomotor activity and freezing in animals challenged with 12000 µg/kg poly I:C but not in AL + poly I:C mice [47].They suggest that the inability of poly I:C alone to significantly increase plasma TNF-α might be the explaining factor for the absence of sickness behavior in AL + poly I:C mice.However, Zenz et al. used a much larger dose of poly I:C than our current study (12000 µg/kg vs 5000 µg/kg), utilized a different form of calorie restriction and did not detail differences between light and dark phase activity levels, which makes direct comparison difficult.
Interestingly, given the T b and locomotor activity results, the CR and AL poly I:C groups demonstrated similar reductions in body weight postinjection.Further, these two groups ate a similar amount of food following poly I:C administration.However, it must be noted that the CR group ate all their allocated food immediately post-injection and given this was a similar amount compared to the AL poly I:C group it is hard to determine whether or not this variable was affected by poly I:C in the CR group.The drive to eat after an immune challenge in CR animals has been shown previously [6], which may be driven by increased hypothalamic levels of the orexigenic peptide, neuropeptide Y (NPY) [6,48].NPY is known to exert anti-inflammatory effects [49]; therefore, may be playing a similar role post-poly I:C in these CR animals to partially attenuate sickness behavior, however, further experiments are required.
Although not measured in the current study, a previous study measured both peripheral and central cytokine levels after poly I:C, giving some insight into mechanisms of poly I:C-induced febrile response.Interestingly, peripheral levels of IL-6, interferon (IFN)β, IL-1β, and TNF-α all peaked at 3 h post-injection, albeit using a much higher dose of poly I:C compared to the current study, 12,000 µg/kg [50].Hypothalamic and hippocampal levels of IL-6 and IL-1β peaked early, between 3-and 6-hours post-poly I:C.However, TNF-α and INFβ both peaked later, between 15 and 24 h post-poly I:C, of which the authors suggested was related to initiating and maintaining the hypothermic response seen during the dark phase after the large dose of poly I:C [50].The AL poly I:C group from the current study demonstrated a limited hypothermic response at the beginning of the dark phase (data not shown), therefore a similar cytokine profile in these mice may be evident if investigated.
Interestingly, another poly I:C and sickness behavior investigation showed that a central IL-1 receptor antagonist (IL-1ra) attenuated, but did not abolish both core T b increases and hypothalamic expression of IL-6 after a 750 µg/kg dose of poly I:C [51].Interestingly, this reduction in T b and central IL-6 was still accompanied by a significant reduction in body weight and food intake in the rats who received poly I:C and IL-1ra [51].This suggests that poly I:C induced reductions in body weight and food intake are not IL-1 dependent processes like they are after LPS [52].Additionally, Fortier and colleagues illustrated that central IL-1ra administration after poly I:C did not affect TNF-α increases [51], suggesting that increased TNF-α may be responsible for mediating the anorexic effects of poly I:C.Similarities can be drawn between these findings and our current study, whereby CR was able to diminish, but not abolish, poly I:C induced T b increases; however, both body weight and food intake changes were similar to the AL-fed mice.This suggests that CR may inhibit IL-1-dependent mechanisms after poly I:C, but not TNF-α.
The abovementioned pro-inflammatory cytokines trigger a cascade of signaling which ultimately results in increased COX-2 and subsequently prostaglandin E 2 (PGE 2 ), a critical pathway for fever generation [53,54].Hypothalamic expression of COX-2 is upregulated after poly I:C [51] and cerebrospinal fluid levels of PGE 2 are also elevated after poly I: C [55].So it isn't surprising that when a specific COX-2 inhibitor, celecoxib, is co-administered with poly I:C that there is a complete attenuation of febrile response [55].Similar to our previous investigations of CR and LPS-induced sickness behavior, CR may work to inhibit the poly I:C-induced production of COX-2 and mPGES-1, a PGE 2 synthase enzyme which is functionally coupled with COX-2 [56], both of which were reduced, but not completely diminished, in CR mice after LPS [6].
Alongside the cascade of pro-inflammatory cytokine and COX-2 signaling and thus the activation of PGE 2 and subsequent generation of fever, previous literature has demonstrated that administration of poly I:C can induce increases in microglial activation gene expression markers and microgliosis post-poly I:C [57,58], consistent with the mild microglial activation we see in our current study.The precise cellular mechanisms underlying the poly I:C-induced microgliosis, however, remains unclear.Town and colleagues have shown microgliosis after poly I:C was TLR3 dependent [58], whilst Zho et al. have demonstrated that microglial activation mainly occurs in a TLR3 independent pathway (MyD88 dependent pathway) but also TLR3 independent pathways [57].One possible explanation for the differences in cellular pathways activated by this stimulus is the time post-poly I:C administration.Zhu and colleagues investigated the MyD88 pathway 6 h after central injection of poly I:C whilst Town et al. demonstrated morphological evidence of microglial activation 24 h after injection of poly I:C; similar to our current study, suggesting that the initial microgliosis response is activated via a TLR3 independent pathway whilst chronic microglia activation by this stimulus is TLR3 dependent.
The current study provides the first comprehensive assessment of the effect of CR on poly I:C-induced sickness behavior.In a dose-response study, we found poly I:C induced a dose-dependent increase in core T b and a reduction in locomotor activity.CR partially attenuated poly I:Cinduced sickness behavior which was characterized by an intermediate increase in T b and reduction in locomotor activity and a comparable decrease in food intake and body weight and a mild attenuation of microgliosis in the hypothalamus compared to the AL poly I:C group.In combination with the previous literature, these results suggest that CR leads to various alterations in poly I:C-induced inflammatory pathways, however, further elucidation of these exact changes is needed.
• C) groups during the first-hour post-injection.Maximal increases in T b for the 2000 µg/kg and 5000 µg/kg groups occurred at 3 (0.86 ± 0.13 • C) and 7 h (1.62 ± 0.23 • C) post-injection respectively.The saline group demonstrated a significantly lower change in T b compared to the 500 µg/kg group at 5 h post-injection (p = 0.019) and the 1000 µg/kg group at 7 h post-injection (p = 0.004).Both the 2000 and 5000 µg/kg groups demonstrated a larger change in T b compared to the saline group during hours 3-, 5-, 6-, 7-, and 8-hours post-injection (range p = 0.038 to < 0.001).
Core T b is attenuated following poly I:C administration in CR compared to AL mice (F (24,224) = 6.15, p < 0.001; Fig. 3a).Poly I:C induced a T b increase in the AL poly I:C animals, which peaked at 7 h post-injection (2.02 • C ± 0.22 • C) and to a lesser extent in the CR poly I: C group, also peaking at 7 h post-injection (0.94 • C ± 0.27 • C).Both control and CR saline groups demonstrated small spikes in T b in the first hour post-Poly I:C administration; however, returned to baseline for the remaining time post-injection.Significant effects were seen from 2 to 8 h post-injection.At 2 h post-injection the AL poly I:C group differed from all other groups (range p = 0.031-0.006).During hours 3-5 h postinjection the AL poly I:C group demonstrated a larger change in T b compared to both saline groups (range p = 0.027 to < 0.001) and then differed from all other groups at 6 h post-injection (range p = 0.034 to < 0.001).During 7-and 8-hours post-injection the AL poly I:C group demonstrated the largest change in T b compared to all other groups (range p = 0.048 to < 0.001).The CR poly I:C mice demonstrated an intermediate increase in T b ; compared to both saline groups (range p = 0.030 to < 0.001); however, lower compared to the AL poly I:C group at 7-and 8-hours post-injection (p = 0.004 and 0.048 respectively).Both the saline groups were no different compared to baseline; however, the AL poly I:C group demonstrated a higher T b compared to baseline during all hour's post-injection (range p = 0.046 to p = 0.001).Conversely, the CR poly I:C group only differed from the baseline at 8 h post-injection (p = 0.021).

Fig. 1 .
Fig. 1. a) Change in T b ( • C) from baseline for 8 h post-injection.b) Change in locomotor activity from baseline during light and dark phases post-injection.Data are mean ± SEM. * denotes a significant difference from the saline group at p < 0.05, * * denotes significant difference from the saline group at p < 0.01, * ** denotes significant difference from the saline group at p < 0.001, # denotes significant difference from baseline at p < 0.05, ## denotes a significant difference from baseline at p < 0.01.

3. 2 . 3 .
Locomotor activity is reduced post-poly I:C administration Administration of Poly I:C induced a reduction in absolute locomotor activity in both the AL and CR mice; however, in different patterns (F[6, 56] = 7.16, p < 0.001; Fig. 3b).The CR mice demonstrated a large reduction during light phase (− 7.74 ± 2.07); however, this pattern then reversed during the dark phase post-injection, with the AL poly I:C mice (− 6.95 ± 1.22) demonstrating a reduction and the CR group returning to near baseline levels.During the light phase the CR poly I:C group demonstrated reduced locomotor activity compared to all other groups (range p = 0.039-0.008),whereas during the dark phase this pattern reversed, with the AL poly I:C group demonstrating a reduced activity compared to all other groups (range p = 0.020-0.005).The CR poly I:C group differed from baseline during the light phase (p = 0.019) and the AL poly I:C group differed during the dark phase (p = 0.002).

Fig. 3 .
Fig. 3. a) Change in T b ( • C) from baseline for 8 h post-injection.b) Change in locomotor activity from baseline during light and dark phases post-injection.Data are mean ± SEM.* denotes a significant difference from both saline groups at p < 0.05, * * denotes a significant difference from both saline groups at p < 0.01, * ** denotes a significant difference from both saline groups at p < 0.001, # denotes a significant difference from the CR poly I:C group at p < 0.05, ## denotes a significant difference from the CR poly I:C group at p < 0.01.

Fig. 5 .
Fig. 5. Classification of ionized calcium-binding adapter molecule-1 (Iba-1)-immunolabelled cells 24 h after an i.p. saline vehicle or poly I:C injection in control (AL) or 50% calorie restricted (CR) mice.a) percentage of ameboid Iba-1 cells in the paraventricular nucleus of the hypothalamus (PVN).b) percentage of ameboid Iba-1 cells in the arcuate nucleus of the hypothalamus (ARC).c) percentage of ameboid Iba-1 cells in the ventromedial nucleus of the hypothalamus (VMH).d) percentage of ameboid Iba-1 cells in the lateral hypothalamus (LH).e) percentage of intermediate Iba-1 cells in PVN.f) percentage of intermediate Iba-1 cells in ARC.g) percentage of intermediate Iba-1 cells in VMH.h) percentage of intermediate Iba-1 cells in LH. i) percentage of ramified Iba-1 cells in PVN.j) percentage of ramified Iba-1 cells in ARC.k) a percentage of ramified Iba-1 cells in VMH.l) percentage of ramified Iba-1 cells in LH. # main effect of diet.p < 0.05.