Social transmission of inflammation in mice

The ability to detect and respond to sickness in others promotes survival. Here we show that mouse dams respond to immune challenged pups by mirroring their inflammatory response. Dams with pups subjected to immune challenge displayed a marked induction of inflammatory mediators in both the brain and the periphery, accompanied by an increase in maternal behaviors and corticosterone levels. This social transmission of inflammation did not require physical contact, and it contributed to the stress hormone response in the dams. In adult dyads, interaction with an immune challenged cagemate did not elicit robust inflammatory signaling but induced an increased responsiveness to a subsequent immune challenge. The identification of social transmission of inflammation, or inflammatory responsiveness, may open new avenues for research on social behavior, just like the description of similar phenomena such as observational fear and transmitted pain has done.


Introduction
In many species, individuals react to the physiological and affective state of conspecifics.In certain cases, there is social transmission or transfer of the affective state to the interacting partner, leading to a form of affective mirroring.For example, social interaction with somebody who is stressed or in pain can lead to transmission of the stress or the pain-related state to the naïve partner, or "bystander" (Smith et al., 2021;Smith et al., 2016;Sterley et al., 2019;Olsson and Phelps, 2007).Such transmission occurs both in humans and rodents and might prepare the bystander for coming challenges.
Immune challenge is a strong threat, both for the affected individual and bystanders.For the latter, it could be important to detect the sickness to be able to either avoid or support the affected individual.In these cases, kinship often drives the interaction in the supportive direction (Zajdel et al., 2019;Kopec et al., 2019).While it is established that bystanders can detect sickness in others and adapt their behavior accordingly (Regenbogen et al., 2017;Hennessy et al., 2014); it is not known if there is a component of transmission or mirroring of the inflammatory state.Such a transfer or mirroring could potentially prime the immune function of the bystander and possibly also affect the hormonal status and behavior.
Here, we investigated how exposure to sick conspecifics affects behavior and the levels of inflammatory mediators in pup-dam dyads and adult female mice.Our results suggest that inflammation can be socially transmitted.

Animals
C57Bl6J female wild type mice from Janvier Labs (Le Genest-Saint-Isle, France) were used in all experiments.Pregnant females were housed individually in standard individually ventilated cages until pups were born.In all cases, this was the first pregnancy of the dam.Experiments were performed postnatal day (PND) 8-9 when dams were 13-14 weeks old.The number of pups per litter ranged from 4 to 9. For adult experiments, female mice were similarly 13-14 weeks old.Mice were kept under controlled conditions of humidity (40-60 %) and temperature (22 ± 1 • C) on a 12 h/12 h light/dark cycle (lights on at 7 A.M.) with food and water available ad libitum.All experiments described in the manuscript were performed using standard individual cages made of clear polysulfone plastic (194 mm x 130 mm x 381 mm) (NextGen500 Allentown).All animal experiments were approved by the Animal Care and Use Committee at Linköping University.

Experimental design
Dams with litters were moved to a room with the temperature set at 26-27 • C at least 30 min before the experiment.In a first set of experiments, pups PND 8-9 were given an intraperitoneal (i.p.) injection of lipopolysaccharide (LPS, 40 µg/kg; E. coli O111:B4, Sigma Aldrich; dissolved in sterile saline) or vehicle (saline).All pups in each litter were injected with the same substance (LPS or saline).The LPS dose was chosen based on our previous study (Zajdel et al., 2019) which induced a strong systemic immune response in the pups.Immediately after injection, the pups were placed back into their home cage and allowed interaction with their dam for 3 h.During that time, maternal behavior was recorded and scored.Subsequently, dams were subjected to a pup retrieval test or sacrificed for molecular and histological analysis.
A second set of experiments was performed in a similar manner to that described above, except that after injections, pups were placed separated from the dam in a new clean cage next to hers.Each cage was covered by a plastic cover with holes.The dam cage was separated from the pup cage by about 5 cm.The dam and their pups could see, smell, and hear each other but they could not have physical contact.An additional control group was included where dam and pups were left together and undisturbed (e.g.no injection).After 3 h, dams were sacrificed and samples from brain, liver, mammary gland and serum were taken for quantitative PCR, ELISA, and immunohistochemistry. Finally, in a separate set of experiments, pup litters were injected with LPS or saline, as previously described, and dams were injected with indomethacin (10 mg/kg, i.p.; Sigma-Aldrich) or vehicle (NaOH 4 % in saline, i.p.), and pups and dams were then allowed to interact for 3 h.
To avoid any possible auditory, visual or olfactory signaling from the sick pups to the control groups, the two experiments (dams exposed to LPS-injected pups and dams exposed to saline-injected pups, respectively) were conducted on different days, with the order of the treatments balanced between the replicates.
To examine social transmission of inflammation between adult dyads, two conspecific female mice were housed together for at least 2 weeks.Female pairs were moved to a room with the temperature set at 22 ± 1 • C at least 30 min before the start of the experiment.In a first experiment, we copied the design from the experiments involving pups and dams.Thus, one of the mice in each dyad was injected with 40 µg/kg of LPS (or saline) and placed back in the cage with the non-injected "bystander".After 3 h of exposure to the immune challenged cagemate, the bystander was sacrificed for molecular analysis.In a second set of experiments, conducted to investigate if exposure to a sick cagemate increased the immune responsiveness of the bystander, one of the females was injected with 100 µg/kg LPS (or saline) and placed back with the bystander for 4 h.After that, the bystander was immune challenged by an i.p. injection of 10 µg/kg LPS (or saline) and sacrificed 2 h later for molecular analysis.The dose of 100 µg/kg was used to induce a stronger immune response than in the previous set of experiments but without compromising the general condition of the animal.The time of exposure was increased to 4 h (+2h after LPS injection to the bystander) since a higher dose of LPS has a more prolonged effect.The dose of 10 µg/kg was used to trigger a mild immune response in the bystander, avoiding ceiling effects in the expression of inflammatory mediators.

Behavioral tests and analysis
Dam and pup interaction was recorded for 3 h after LPS or saline injections.During this time, the standard cage lid was replaced with a transparent cover with holes to allow recording and scoring of maternal care.Behavioral scoring was performed during 40-60 min, 80-100 min, and 120-140 min after injection.The dam behaviors evaluated were pup-directed behaviors (pup-nursing, pup-grooming and time spent in the nest), as well as non-pup directed behaviors, aimed at self-care and maintenance (self-grooming and digging).Nursing behavior was defined as the time that the dam was over the pups, relatively immobile in an arch posture supported by rigid fore-or hind-limbs, with pups attached to the nipples.Pup grooming was defined as the time the dam spent touching any part of the pup's body with her tongue, nose or forepaws, or stretching her nose in the direction of the pups.Time spent in the nest was defined as the time the dam spent physically inside the nest in an active or passive behavior.Digging was defined as a series of fast alternating movements of the forepaws scraping back bedding material resulting in accumulation of bedding in a pile under the abdomen of the animal.The behavior was scored as self-grooming when the dam was performing a sequence of cleaning steps by licking and scratching its own fur.
Pup retrieval behavior was tested after 3 h of exposure to pups that had been injected with LPS or saline.Pups were separated from the dams and positioned at the opposite side of the nest.The time to retrieve each pup was scored.Dams that did not retrieve any pup or took more than five minutes to retrieve the first pup were excluded.In the present experiment, 3 out of 15 dams exposed to saline-injected pups and 5 out of 16 dams exposed to LPS-injected pups did not retrieve their pups.

Quantitative PCR
Dams were sacrificed by asphyxiation with CO 2 followed by cervical dislocation.The hypothalamus, hippocampus, nucleus accumbens, liver and mammary glands were dissected, immersed in RNAlater (Qiagen; Hilden, Germany) and stored at − 20 • C for later analysis.RNA was extracted using RNeasy Lipid Tissue Kit (Qiagen) according to the manufacturer's instructions.The RNA concentrations and quality were measured with a NanoDrop spectrophotometer (ThermoFisher Scientific).Only RNA samples with A260/A280 and A260/A230 ratios > 1.8 were included in the experiment.cDNA synthesis was performed with High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems).For quantitative PCR, TaqMan Gene Expression Master Mix (Applied Biosystems) was used together with the following TaqMan gene expression assays: Cyclooxygenase-2 (Ptgs2;Mm00478374_m1), Interleukin-1β (Il1b; Mm01336189_m1), tumor necrosis factor (Tnf; Mm00443258_m1), interleukin-6 (Il6; Mm00446190_m1), C-X-C motif chemokine ligand 10 (Cxcl10; Mm00445235_m1), and C-C Motif Chemokine Ligand 2 (Ccl2; Mm00441242_m1).As endogenous control, Glyceraldehyde 3-phosphate dehydrogenase (Gapdh; Mm99999915_g1) was used.Reactions were performed in a Real-Time 7900 Fast apparatus (Applied Biosystems).Relative quantification was carried out using the 2 ΔΔCT method.Gene expression changes are expressed as fold change versus the respective control group.In the adult experiment, cytokine index was calculated as follows: first the fold change of each inflammatory mediator was normalized to the control group (non-challenged mice exposed to a saline injected cagemate).Then, the average of the normalized values of all mediators was calculated for each animal, obtaining a single value reflecting the strength of the immune response.

Corticosterone assay
Trunk-blood was collected in standard tubes.Blood samples were kept at 22 • C for 30 min, and then centrifuged at 5000 rpm at 4 • C for 30 min to collect serum.Steroids were extracted from 10 μl of serum with the Steroid Displacement Reagent (Enzo Life Sciences) and corticosterone levels were measured by an enzyme immunoassay, according to the manufacturer's protocol (Corticosterone Kit; Enzo).The minimal detectable concentration of corticosterone was 27 pg/ml.Optical densities were read at 405 nm.The values were then calculated using a curve fit, ranging from 32 to 20000 pg/ml.The intra-assay coefficient of variation (CV) was 4,02-5,8 % and inter-assay CV was 5,13 %.
Sections were analyzed on a Nikon 80i microscope equipped with epifluorescence, using 10X objective (for quantification) or 20X objective (for representative figures).Quantification of cyclooxygenase-2 (Cox-2) and lipocalin positive cells was performed in 10X field sections.A total of four sections per mouse were used for quantification.Two sections were taken from the hypothalamic area and two of the striatal area.No significant differences in the quantification of Cox-2 and lipocalin were found between these brain regions.Therefore, they were averaged.Captured images were processed in Adobe Photoshop (Adobe Systems) with adjustments of brightness and contrast.

Magnetic Luminex Screening assay
Serum levels of CCL2, TNF, IL6 were measured by a customized Magnetic Luminex Screening Assay according to according to the manufacturer protocols (R&D systems).Briefly, plates were incubated in an orbital shaker with 50 μL/well of different diluted samples and concentration of standard chemokines provided with 50 μL of mouse, magnetic, premixed, microparticle cocktail with antibodies specific for each cytokine and chemokine.After washings using magnetic bead separator (HydroFlex™), plates were incubated with premixed biotinantibodies cocktail specific for each cytokine and chemokine.Antibody cytokine and chemokine complexes were stained with Streptavidin-PE.Plates were read using Luminex FLEXMAP 3D w/ xPONENT 4.3 and results were obtained as median fluorescence intensity (MFI).A standard curve was generated for each cytokine and chemokine to convert MFI into corresponding relative concentrations.Cytokines and chemokines levels are expressed as pg/ml.The intraassay coefficient of variation (CV) was 4,48-7,05 %.

Statistical analysis
Data analysis was carried out in Graphpad Prism v. 6.01.Results are presented as mean ± SEM.When data was not normally distributed, as indicated by a Shapiro-Wilk test, nonparametric tests were used and when the majority of data was normally distributed, or the numbers of observations were high, parametric tests were used.Statistical comparisons of two groups were done using Mann-Whitney test or t-test.When more than two groups were compared, the Kruskal-Wallis test was used.When having two categorical variables and normally distributed data, 2-way ANOVA was performed, followed by Tukey post-hoc analysis multiple comparisons tests.Pearson correlation coefficients were used to examine the relationship between inflammatory gene expression in the hypothalamus and corticosterone levels in dams exposed to sick pups.P < 0.05 was considered statistically significant.

Results and discussion
To investigate if inflammation, or inflammatory responsiveness, can be socially transmitted we used mice injected with LPS as a model.First, we focused on how dams react to sick offspring, due to their tight bond and the importance of their interaction for the survival of the pups.We induced systemic inflammation in 8-9 days old mouse pups by intraperitoneal injection of LPS (40 µg/kg; Fig. 1A).This procedure increased pup-directed support behaviors (Fig. 1B-D and Suppl.Fig. 1) and corticosterone levels (Fig. 1E) in the dams, showing that the dams detected and responded to the challenged state of the pups.After 3 h of interaction with the immune challenged pups, the levels of inflammatory transcripts in the hypothalamus of the dams were examined.Surprisingly, we found a strong induction of inflammatory genes such as Ptgs2 (encoding cyclooxygenase-2), Il1b, Tnf, Il6, Ccl2 and Cxcl10 (Fig. 1F).Furthermore, we detected a strong induction of PTGS2/ cyclooxygenase-2 and Lipocalin-2 protein in brain endothelial cells (Fig. 1G-L).In several replicates of the experiment, we observed that the inflammatory response induced by sick pups and the fraction of responsive dams varied within and between experiments (Fig. 1M-O).We were unable to identify the factors underlying this variability, but we never detected an inflammatory state in a dam with saline-injected pups, indicating that the effect was specific to dams with sick pups.Inductions of cytokine and chemokine mRNAs were also detected in other brain regions (hippocampus and nucleus accumbens; Fig. 2A, B) and peripheral organs (liver and mammary glands; Fig. 2C, D) of dams exposed to sick pups.In addition, the levels of circulating cytokines and chemokines were also increased in the serum of dams exposed to sick offspring (Fig. 2E), showing that the immune response in the dams was systemic.
We next investigated the relation between the transmitted inflammatory response and the elevation of corticosterone in the dams with immune challenged pups.Corticosterone levels strongly correlated with cytokine levels in the hypothalamus (Suppl.Fig. 2).To get an indication of whether the transferred inflammation mediated the stress hormone release in the dams exposed to immune challenged pups (in the same cage), we inhibited inflammatory signaling in the dams with the prostaglandin synthesis inhibitor indomethacin, which previously has been shown to inhibit inflammation-induced corticosterone release (Elander et al., 2009).Indeed, indomethacin attenuated the increase in corticosterone levels induced by exposure to immune challenged pups (Fig. 1P) indicating a causal role of the inflammatory signaling for the stress hormone response.
To investigate if the inflammatory signaling in the dams required physical contact with the pups, we moved the pups to a separate cage next to the cage with the dam, injected them, and kept them there for the entire experiment (Fig. 3A).Despite the lack of physical contact between pups and dam, we found an induction of inflammatory genes in the hypothalamus of the dams (Fig. 3B, C), although this induction was milder than the one in the experiment allowing direct contact.Thus, the transferred inflammatory signaling was not due to LPS excretion from the pups, or any mechanism involving direct physical contact between pups and dam.We found no induction of inflammatory genes in dams with separated pups given saline (Fig. 3B), suggesting that stress induced by the separation is not sufficient for such an induction.Future studies blocking the transmitted signals (for example using filter top cages) or different sensory modalities will be required to elucidate the route of transmission, but olfactory cues are likely candidates since they have been shown to be important for avoidance of sick animals (Kavaliers et al., 2022) and social transfer of stress and hyperalgesia (Smith et al., 2021;Smith et al., 2016;Sterley et al., 2019).In the latter case, the signal was shown to spread between cages (Smith et al., 2016), similar to what we found in the present study.However, we cannot exclude other routes of transmission such as visual cues (Olsson and Phelps, 2007;Keum and Shin, 2019) or vocalizations (Lecca et al., 2023).
Next, we investigated if a similar transfer of inflammation takes place between adult mice (Fig. 4A).In dyads of females, we injected one of the mice (the "cagemate") with LPS to induce sickness, and assessed the cytokine mRNA levels in the hypothalamus of the other mouse (the "bystander").We observed no consistent induction of inflammatory genes in the bystander mice (Suppl.Fig. 3.) indicating that the social transfer of inflammatory signaling, as observed between pups and dams, is not present in all ages and kinships, at least at this dose and timepoint.However, when bystander mice were given a mild immune challenge (LPS, i.p., 10 µg/kg) after a somewhat longer (4 h) exposure to a cagemate subjected to a stronger immune challenge (LPS, i.p., 100 µg/kg), we observed a mildly potentiated induction of inflammatory mediators in the hypothalamus, as indicated by an elevated cytokine index (normalized mean fold-change of all inflammatory mediators measured; Fig. 4B).This elevation was driven by an increased induction of Il1b and a trend towards increase of some of the other mediators (Fig. 4C-G).Similar potentiation of inflammatory signaling has been observed previously in animals interacting with sick conspecifics (Hamasato et al., 2014) and after acute stress (Fonken et al., 2018).Furthermore, a potentiation of the immune response has been seen in humans shown photographs depicting symptoms of infectious disease (Schaller et al., 2010).The potentiation in adults can be seen as a "less strong" form of transmission and is conceptually similar to "transmitted pain", in which a mouse in pain induces increased sensitivity to pain (Smith et al., 2021;Smith et al., 2016), but not spontaneous pain, in bystanders.It is unclear why the "strong form" of transmission, i.e. induction of inflammatory signaling in an unchallenged bystander, took place between pups and dams but not between adult females.Perhaps it reflects an adaptation in the dam to the necessity of close physical contact with sick pups, or a mechanism for supporting them by mediators through the milk (Sanidad et al., 2022).
For stress, fear and pain, the social transfer has been explained in terms of an adoption of the sensory-affective state of a partner in distress.Such mechanisms could also be at play for the transfer of inflammatory signaling that we observe here.However, this inflammatory mirroring is different from the other examples of transfer in the sense that it not only involves activity in the neurons underpinning sensoryaffective states but also a strong and broad induction of inflammatory mediators in the brain and periphery.Such induction is unlikely to be a generic stress effect since separation of the dams from the pups, which is stressful for the dams, was not enough to trigger any inflammatory response.Furthermore, although some strong and persistent stress paradigms can induce low-grade inflammatory signaling (Dantzer, 2018;Biltz et al., 2022), the magnitudes of such immune activations are much lower than the ones we observe here.Our findings extend previous studies showing that disgusting stimuli induces TNFα in saliva in humans (Anja Juran et al., 2022;Stevenson, 2012) and that exposure to sick conspecifics induces TNF in the olfactory bulb of rats (Hamasato et al., 2017).
Our findings are in line with studies showing that both humans and rodents can detect sickness in conspecifics and adjust behavior accordingly.The increased interaction between inflamed pups and their dams resonates well with previous studies showing that while inflammation elicits social avoidance in unrelated individuals, it can increase the motivation for supportive interactions between individuals with strong bonds (Muscatell and Inagaki, 2021).We studied pups at an age when they are highly dependent on the dam and a tight bidirectional bond is formed.We did not examine if inflammation in pups can be transmitted to sires, virgins or unrelated dams, so it is unclear if the transmission requires strong bonds.In addition, future studies will tell if social We conclude that dams exposed to immune challenged pups respond with a strong systemic induction of inflammatory genes.The induction contributed to the stress hormone response in the dam and did not rely on physical contact between dams and pups.These observations show that systemic inflammation can be socially transmitted.We believe our findings will open new avenues for research, just like the identification similar concepts such as observational fear and transmitted pain has done.

Fig. 1 .
Fig. 1.Exposure to immune challenged pups elicited a central inflammatory response in dams.Dams exposed to immune challenged pups (A) displayed increased maternal-related behaviors at the expense of other behaviors (B-D; n = 8-13) and showed increased plasma levels of corticosterone (E; n = 9-10).They also displayed an induction of cytokine and chemokine mRNA in the hypothalamus (F; n = 8-10) and an induction of PTGS2 and Lipocalin-2 protein in the brain endothelium (G-L; quantification in G, J, n = 6; and representative pictures from the immunohistochemistry in H, I, K, L).The induction of inflammatory genes in the hypothalamus varied between individuals (M, N; n = 34-41) and experimental batches (O).The corticosterone release induced by exposure to sick pups was attenuated by the cyclooxygenase inhibitor indomethacin (P).Indo.; indomethacin, Lcn2; lipocalin-2, LPS; Lipopolysaccharide, NaCl; Saline, Veh; vehicle.Scale bar = 50 µm.Data are expressed as mean ± SEM.Gene expression changes are expressed as fold change versus the control group.Statistical analysis was performed with Mann-Whitney test, or Two-way ANOVA followed by Tukey posthoc test, with *P < 0.05, **P < 0.01, ***P < 0.001.

Fig. 2 .
Fig. 2. The inflammatory response elicited by exposure to immune challenged pups is systemic.The induction of cytokines after exposure to sick pups was generalized to other brain regions, such as hippocampus (A; n = 8-10) and nucleus accumbens (B; n = 8-10).A strong induction of cytokines and chemokine mRNA was also found in the liver (C; n = 5-6), mammary gland (D; n = 7-12) and serum (C; n = 11-14) of dams exposed to sick pups.Data are expressed as mean ± SEM.Gene expression changes are expressed as fold change versus the control group.Statistical analysis was performed with Mann-Whitney test with *P < 0.05, **P < 0.01, ***P < 0.001.