Systemic Administration of the TLR7/8 Agonist Resiquimod (R848) to Mice Is Associated with Transient, In Vivo-Detectable Brain Swelling

Simple Summary Physiological changes to the body can affect the brain. For example, inflammation that originates in the body can be sensed by the brain. In this study, we used an agent called resiquimod (R848) to stimulate inflammation in the periphery and measured structural and metabolic brain responses in anesthetized mice using in vivo magnetic resonance imaging. Relative to baseline prior to drug administration, a high dose of R848 caused sickness behaviors and volume expansion in several cortical regions at 3 h that were no longer evident at 24 h. Transient volume expansion in response to peripheral immune stimulation is consistent with brain swelling. Abstract Peripheral administration of the E. coli endotoxin lipopolysaccharide (LPS) to rats promotes secretion of pro-inflammatory cytokines and in previous studies was associated with transient enlargement of cortical volumes. Here, resiquimod (R848) was administered to mice to stimulate peripheral immune activation, and the effects on brain volumes and neurometabolites determined. After baseline scans, 24 male, wild-type C57BL mice were triaged into three groups including R848 at low (50 μg) and high (100 μg) doses and saline controls. Animals were scanned again at 3 h and 24 h following treatment. Sickness indices of elevated temperature and body weight loss were observed in all R848 animals. Animals that received 50 μg R848 exhibited decreases in hippocampal N-acetylaspartate and phosphocreatine at the 3 h time point that returned to baseline levels at 24 h. Animals that received the 100 μg R848 dose demonstrated transient, localized, volume expansion (~5%) detectable at 3 h in motor, somatosensory, and olfactory cortices; and pons. A metabolic response evident at the lower dose and a volumetric change at the higher dose suggests a temporal evolution of the effect wherein the neurochemical change is demonstrable earlier than neurostructural change. Transient volume expansion in response to peripheral immune stimulation corresponds with previous results and is consistent with brain swelling that may reflect CNS edema.

A number of ligands have been designed to boost (agonists) or block (antagonists) inherent signal transduction at TLR7/8 receptors. They are of interest in acute applications, such as adjuvants to vaccines [17,18], and in chronic protocols, such as immunotherapies for cancer [19,20]. There is also emerging evidence that TLR7/8 receptors may be an appropriate target for the treatment of alcohol-use disorders (AUD) e.g., [21][22][23]. One TLR7/8 agonist, imiquimod, has FDA approval [24]. In preclinical studies, topical imiquimod (i.e., Aldara cream) was associated with systemic pro-inflammatory cytokine production, enlargement of spleen and draining lymph nodes [25], induction of CNS chemokine and interferon signaling, infiltration of immune cells into the CNS [26,27], and increased Iba-1 and GFAP staining indicating micro-and astro-gliosis [28]. Imiquimod given at 50 µg intraperitoneal (i.p.) increased depressive symptoms in mice 2 h after treatment, possibly via elevated IFNα expression [29].
A single injection of the E. coli endotoxin lipopolysaccharide (LPS, i.p.), which targets TLR4 [36,37], similarly induces acute (i.e., 2 h) pro-inflammatory cytokine signaling [38,39] and sickness responses [40]. In response to acute LPS-induced peripheral inflammation in rats, in vivo magnetic resonance imaging (MRI) and spectroscopy (MRS) revealed transient expansion of posterior midline CNS tissue volumes [41]-which may represent acute brain edema e.g., [42,43]-and transient elevations in striatal glutamine [41]. The current study, therefore, tested the hypothesis that peripheral TLR7/8 stimulation of mice with R848 would similarly result in MR-detectable changes to the brain including modulation of the glutamate/glutamine balance and brain swelling consistent with edema. To the best of our knowledge, only a single study has previously evaluated the in vivo effects of R848 using MRI but focused on cardiac, not brain, tissue [44].

Ethics Statement
All experimental procedures were conducted in accordance with the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The Institutional Animal Care and Use Committees at SRI International and Stanford University approved all procedures.

Animals and Drugs
A total of 24 male C57Bl mice (Jackson Laboratories, Sacramento, CA, USA), arriving at 6 weeks of age (24.88 ± 1.7 g), were used in this study. Only male mice were included to minimize sex effects [45][46][47] and to expand on previous work [15]. Mice were housed three per cage, maintained in a pathogen-free facility on a 12-h light/dark cycle, and had ad libitum access to regular chow and water.

Magnetic Resonance (MR) Scanning Procedures and Data Analysis
Each mouse was scanned three times with one scan per day over three consecutive days. The initial scan on day 1 was a baseline (scan 1). On day 2, mice were injected with either saline or R848 and scanned 3 h later (scan 2). On day 3, each animal underwent a final scan (scan 3), 24 h after the saline or R848 injection. These scan time points-3 h and 24 h after injection-were chosen because the literature reports optimal stimulation of the immune system (e.g., IFN-α and IFN-γ secretion) at 3 h and resolution typically by 24 h [49,50]. For each scan, mice were anesthetized with isoflurane (3% for induction; 0.5-2% for maintenance during the scan); body weight and temperature were acquired and recorded; then animals were placed on the cradle base with built-in water circulation for body temperature control. Temperature and respiration were monitored throughout each scan (∼1 h). All animals received subcutaneous saline (0.5 mL) for hydration at the end of the scan.

MR Spectroscopy
Point RESolved Spectroscopy (PRESS) data were obtained from two voxels [bilateral hippocampus (6.0 × 1.5 × 2.5 mm 3 ) and right dorsal striatum (3.0 × 3.0 × 2.5 mm 3 ) ( Figure 1a)] with TR = 2500 ms, TE = 16 ms, 760 points, spectral width 13.3177 ppm, receiver bandwidth 4000 Hz, and with VAPOR (variable power and optimized relaxation delays) outer volume water suppression (bandwidth 200 Hz). Water suppressed excitations (256) were at 2.35 ppm (the frequency of glutamate) and an additional unsuppressed acquisition (32 excitations) was acquired at 4.7 ppm for water scaling.  MR data were collected on a Bruker 70/16 US AVANCE III 7.0T system (Karlsruhe, Germany) with 380 mT/m gradient strength on each (X, Y, and Z) axis, slew rate of 3420 T/m/s, 16 cm bore size using a Bruker mouse head volume coil (23 mm) and ParaVision 6.1 software. A gradient-recalled echo (GRE) localizer scan was used to position the Metabolite concentrations were quantified with LCModel with eddy-current correction and water scaling (which accounts for CSF volume in the voxel) using a basis set of 21 metabolites. The analysis was run with the "water reference" option that provided reasonably meaningful absolute metabolite concentrations based on the unsuppressed water content of the voxel. The analysis window was 0.2-4.0 ppm. Data were pre-processed with zero-order phasing, referencing, and residual water line removal. Data were fit to a linear combination of a number of metabolites in a simulated basis set designed for Bruker 7T MRS data acquisition at TE = 16 ms provided by Stephen Provencher [sp@lcmodel.CA] [51,52] containing alanine, creatine (Cr), phosphocreatine (PCr), glutamine (Gln), glutamate (Glu), glycerophosphorylcholine (GPC), phosphorylcholine (PCh), glutathione (GSH), inositol (Ins), lactate, N-acetylaspartate (NAA), N-acetylaspartylglutamate (NAAG), scyllo-inositol, taurine, and several lipids and macromolecules ( Figure 1b). Only metabolite concentrations derived from fitted spectra consistently within average Cramér-Rao bounds <15% were considered (Table 1).
Structural MRI data preprocessing of each image included removal of noise [53] and inhomogeneity correction via ANTS 2.1.0 [54]. Each image was skull stripped by aligning a template to the scan via symmetric diffeomorphic registration [55], and the resulting deformation map was applied to the brain mask of the template. Image inhomogeneity correction was repeated on skull-stripped images. The structural template was segmented into CSF, gray matter, and white matter by fitting the histogram of image intensities with three Gaussians yielding the probability of the three tissue types on a voxel-by-voxel basis. A CSF mask was constructed comprising all voxels in which the probability of CSF was the largest of the three values (i.e., CSF volume, Table 2). In parallel, the template was parcellated into 42 regions-of-interest (ROIs) by first registering the Allen Reference Atlas (ARA) [56] to it and then projecting the CSF mask onto the ARA parcellation map. Finally, tissue segmentation and ARA parcellation were transformed to the space of each image using the prior deformation map and resampled in the template space by rigidly aligning the bias-corrected skull-stripped image to the template. Volume of the three tissue types within each ROI was calculated by Computational Morphometry Toolkit (CMTK) [57]. Table 2. Statistical results for regions of interest delineated by the Allen Reference Atlas.

Sickness Responses
A three-group by three-time point MANOVA on body weight (expressed as percent of baseline) was significant (F 4,42 = 11.9, p < 0.0001, Figure 2a). Body weight did not distinguish the three groups at scan (all normalized to zero) or scan 2 (F 2,24 = 0.6, p = 0.57); the three-group ANOVA was significant at scan 3 (F 2,24 = 20.2, p < 0.0001). At 24 h following injection, the 50 µg R848 group had lost 5.5% of their baseline body weight; the 100 µg R848 group lost 2.8% of their baseline body weight; the control group was unaffected.

Correlations between Brain and Response Measures
Non-parametric Spearman ρ evaluated relationships among variables. In the 50 µg R848-treated group, relationships between changes in hippocampal metabolite levels, body weight, temperature, and ROI volumes were not forthcoming. For ROI volumes, only the 100 µg R848-treated group was considered. The percent change in body weight between scan 1 and scan 3 correlated with percent change in volume between scan 1 and scan 2 for somatosensory cortex (r = −0.79, p = 0.02) and olfactory cortex (r = −0.73, p = 0.04) ( Figure 6).
Non-parametric Spearman  evaluated relationships among variables. In the 50 μg R848-treated group, relationships between changes in hippocampal metabolite levels, body weight, temperature, and ROI volumes were not forthcoming. For ROI volumes, only the 100 μg R848-treated group was considered. The percent change in body weight between scan 1 and scan 3 correlated with percent change in volume between scan 1 and scan 2 for somatosensory cortex (r = −0.79, p = 0.02) and olfactory cortex (r = −0.73, p = 0.04) ( Figure 6). Figure 6. Correlations between changes in body weight and changes in regional brain volumes (i.e., somatosensory and olfactory cortices) between scans 1 and 2 in the 100 μg R848-treated group.

Discussion
To the best of our knowledge, this is the first in vivo imaging study of the brain in response to peripheral R848 administration. The body weight loss observed at 24 h in the both 50 μg and 100 μg R848-treated mice is consistent with "sickness behavior" and previous observations, reported at doses as low as 10 μg [15,21]. Regarding temperature, previous studies suggest transient increases in temperature (e.g., at 2 h following R848) taken during the day, although there is some indication that body temperature in response to R848 may decrease at night [15,21].
Consistent with our previous study in rats in response to TLR4 activation with LPS, the current study in mice demonstrates transient volume expansion in cortical regions (including motor, somatosensory, and olfactory cortices) and pons. The time course of in vivo volume changes is consistent with the temporal profile of molecular changes, demonstrating acute inflammatory responses at early intervals, but virtually no residual responses [15,58]. Acknowledging that the regional distribution of TLR expression may depend on the type (e.g., hypoxia vs. concussion) and length of exposure to insult [59,60], the localization herein of the R848 response, to cortical regions, is compatible with evidence for TLR7 expression to neocortex of the developing mouse brain [61] and prominent cortical distribution in the adult mouse brain [16,62]. This pattern is in contrast to the LPS response in rats, which was localized to splenial, retrosplenial, and peri-callosal hippocampal regions, but comports with a predominately hippocampal distribution of TLR4 [63][64][65].
Transient metabolic changes observed at the 50 μg but not 100 μg dose may represent a continuous CNS response to the degree of peripheral inflammation. In other words, as there is evidence for a pro-inflammatory response to R848 to be concentration-dependent e.g., [66], lower concentrations of R848 with a lower pro-inflammatory response may only stimulate metabolic changes while higher doses associated with more pronounced proinflammation may result in brain swelling. Transient changes in the levels of NAA and PCr more likely reflect altered energy utilization than osmotic imbalance [67,68], which may manifest before volumetric changes evolve.
In summary, the current study recapitulates in vivo changes in body weight and temperature in response to peripherally administered R848 and expands on previous findings Figure 6. Correlations between changes in body weight and changes in regional brain volumes (i.e., somatosensory and olfactory cortices) between scans 1 and 2 in the 100 µg R848-treated group.

Discussion
To the best of our knowledge, this is the first in vivo imaging study of the brain in response to peripheral R848 administration. The body weight loss observed at 24 h in the both 50 µg and 100 µg R848-treated mice is consistent with "sickness behavior" and previous observations, reported at doses as low as 10 µg [15,21]. Regarding temperature, previous studies suggest transient increases in temperature (e.g., at 2 h following R848) taken during the day, although there is some indication that body temperature in response to R848 may decrease at night [15,21].
Consistent with our previous study in rats in response to TLR4 activation with LPS, the current study in mice demonstrates transient volume expansion in cortical regions (including motor, somatosensory, and olfactory cortices) and pons. The time course of in vivo volume changes is consistent with the temporal profile of molecular changes, demonstrating acute inflammatory responses at early intervals, but virtually no residual responses [15,58]. Acknowledging that the regional distribution of TLR expression may depend on the type (e.g., hypoxia vs. concussion) and length of exposure to insult [59,60], the localization herein of the R848 response, to cortical regions, is compatible with evidence for TLR7 expression to neocortex of the developing mouse brain [61] and prominent cortical distribution in the adult mouse brain [16,62]. This pattern is in contrast to the LPS response in rats, which was localized to splenial, retrosplenial, and peri-callosal hippocampal regions, but comports with a predominately hippocampal distribution of TLR4 [63][64][65].
Transient metabolic changes observed at the 50 µg but not 100 µg dose may represent a continuous CNS response to the degree of peripheral inflammation. In other words, as there is evidence for a pro-inflammatory response to R848 to be concentration-dependent e.g., [66], lower concentrations of R848 with a lower pro-inflammatory response may only stimulate metabolic changes while higher doses associated with more pronounced proinflammation may result in brain swelling. Transient changes in the levels of NAA and PCr more likely reflect altered energy utilization than osmotic imbalance [67,68], which may manifest before volumetric changes evolve.
In summary, the current study recapitulates in vivo changes in body weight and temperature in response to peripherally administered R848 and expands on previous findings by demonstrating an in vivo CNS response to R848. The brief metabolic changes at the lower dose may reflect energy imbalance and the transient volumetric changes (i.e., swelling) at the higher dose may represent brain edema in response to peripheral immune stimulation. Remaining to be elucidated is the mechanism of pro-inflammatory signal transduction from the periphery to the brain which may include stimulation of the vagus nerve; activation of cerebral endothelial cells; or direct access to brain parenchyma via circumventricular organs.