CFA-treated mice induce hyperalgesia in healthy mice via an olfactory mechanism

Abstract

Social interactions with subjects experiencing pain can increase nociceptive sensitivity in observers, even without direct physical contact. In previous experiments, extended indirect exposure to soiled bedding from mice with alcohol withdrawal-related hyperalgesia enhanced nociception in their conspecifics. This finding suggested that olfactory cues could be sufficient for nociceptive hypersensitivity in otherwise untreated animals (also known as “bystanders”). The current study addressed this possibility using an inflammation-based hyperalgesia model and long- and short-term exposure paradigms in C57BL/6J mice. Adult male and female mice received intraplantar injection of complete Freund’s adjuvant (CFA) and were used as stimulus animals to otherwise naïve same-sex bystander mice (BS). Another group of untreated mice (OLF) was simultaneously exposed to the bedding of the stimulus mice. In the long-term, 15-day exposure paradigm, the presence of CFA mice or their bedding resulted in reduced von Frey threshold but not Hargreaves paw withdrawal latency in BS or OLF mice. In the short-term paradigm, 1-hr interaction with CFA conspecifics or 1-hr exposure to their bedding induced mechanical hypersensitivity in BS and OLF mice lasting for 3 hrs. Chemical ablation of the main olfactory epithelium prevented bedding-induced and stimulus mice-induced mechanical hypersensitivity. Gas chromatography-mass spectrometry (GC-MS) analysis of the volatile compounds in the bedding of experimental mice revealed that CFA-treated mice released an increased number of compounds indicative of disease states. These results demonstrate that CFA-induced inflammatory pain can modulate nociception in bystander mice via an olfactory mechanism involving dynamic changes in volatile compounds detectable in the rodent bedding.

Introduction

It is now widely recognized that social context can modulate nociceptive experience (Martin et al., 2014; Mogil, 2015). The effects of a social encounter on nociception can depend on a number of factors, including familiarity (Langford et al., 2006), genetics (D’Amato, 1997; D’Amato and Pavone, 1996), partner relationships (Osako et al., 2018), social hierarchy (Aghajani et al., 2013), pain modality (Langford et al., 2006; Pitcher et al., 2017), and past experience (Luo et al., 2020). Moreover, the presence of a conspecific in pain can enhance nociception in the tested subject (Smith et al., 2016).

The sensory channels through which social context leads to changes in nociception are only now beginning to be defined. Visual cues have been shown to alter nociceptive behavior in an observer animal, depending on the demonstrator’s pain state and the relationship between the interacting animals (Langford et al., 2006; Li et al., 2018; Pitcher et al., 2017). However, given the complexity of social communication, it is likely that multiple senses contribute to social modulation of pain. In rodents, olfactory communication plays an important role in social interactions, and rodents exhibit odor-induced changes in affective and behavioral responses in social encounters (Andraka et al., 2021; Takahashi et al., 2013). Indeed, odors of a stressed rat can produce an antinociceptive effect in a same-sex conspecific (Fanselow, 1985).

Our laboratory recently reported that housing animals in the same room as conspecifics with hyperalgesia associated with withdrawal from alcohol or morphine, or with persistent inflammation, is sufficient to induce hyperalgesia in otherwise experimentally naïve mice (termed “bystander mice”, Smith et al., 2016). An analogous phenomenon is seen in prairie voles (Walcott et al., 2018). Because animals in these studies were housed in separate cages, it is unlikely that the relevant cues were visual or auditory (e.g., ultrasonic vocalizations) in nature. Rather, olfactory cues seemed to be important, since exposure to soiled bedding from mice undergoing alcohol withdrawal was sufficient to induce nociceptive hypersensitivity (Smith et al., 2016). This latter finding indicated that odorants emitted by animals undergoing withdrawal from alcohol have the potential to modulate nociception. However, whether olfactory cues contribute to social modulation of pain more generally remains to be determined.

The present study was designed to determine whether hyperalgesic states induced by exposure to conspecifics subjected to persistent inflammation are mediated by olfactory cues, and whether the odorants mediating these olfactory responses modulate nociception by engaging the olfactory sensory system. We therefore tested whether exposure to bedding from C57BL/6J mice with inflammatory pain induced by Complete Freund’s Adjuvant (CFA) would induce hyperalgesia in otherwise untreated mice. Our first experiment tested the effects of long-term (15 days) exposure to bedding on nociceptive behaviors. We next tested effects of an acute bedding exposure, since it was recently demonstrated that a brief interaction with CFA-treated mice can also induce enhanced nociception in same-sex conspecifics (Smith et al., 2021). Then, to determine whether odorants from hyperalgesic mice engage olfactory circuits to modulate nociception, we tested the impact of chemical ablation of the olfactory epithelium on the response to soiled bedding. Finally, to identify possible constituents of the olfactory cue or cues, we used gas chromatography-mass spectrometry (GC-MS) to analyze the volatile compounds in bedding from CFA-treated and control mice to identify changes in the chemical profile following CFA treatment.

Methods

Animals and Housing

Adult female and male C57BL6/J mice, 8 weeks of age, were purchased from the Jackson Laboratory (Sacramento, CA). Upon arrival, mice were housed three per cage in standard shoebox cages (18.4 cm x 29.2 cm x 12.7 cm) in the Department of Comparative Medicine animal facility at Oregon Health & Science University, provided with 1/2” pelleted cellulose bedding (Biofresh^™, Patterson, NY), cotton nestlets and Enviro-Dri for nest building/thermoregulation, as well as access to food (LabDiet 5001; LabDiet) and water ad libitum. They were maintained on a 12-h light:dark cycle (lights on at 0600), with stable temperature and humidity (21 ± 1 °C, 32 ± 7%). Experiments were performed between 8:00 and 14:00. The light intensity in the colony cages ranged between 50 and 250 lux during the light phase, depending on position within the cage. Light intensity during the transfer experiments and buried food tests was between 150-250 lux during the light phase. Light intensity during testing was 250-350 lux.

After at least 4 d in the facility, mice were transferred to experimental rooms and singly caged for 7 d of habituation. During the entire experiment, animals were housed in shoebox cages with wire tops but without filter lids to allow olfactory communication among neighboring cages. These cages are placed on three-shelf racks (150 cm width x 61 cm depth x 188 cm height). A total of 150 female and 106 male mice were used in this study. Their body weights were between 20-30 g during experimentation. All protocols were approved by the Oregon Health & Science University Animal Care and Use Committee and performed within the National Institutes for Health Guidelines for the Care and Use of Laboratory Animals, as well as the Guidelines for the Care and Use of Mammals in Neuroscience and Behavioral Research.

Drug treatments

Localized inflammation.

Complete Freund’s adjuvant (CFA, 10 µL, 1 mg/mL heat-killed and dried Mycobacterium tuberculosis, Sigma-Aldrich, F5881) was used to induce an inflammatory response. Animals were anesthetized (3.5% isoflurane), and CFA or saline (10 µL) was injected into the plantar surface of either the left or right hind paw. Mice were then returned to their home cages to recover fully before being tested or interacting with other mice.

Ablation of olfactory epithelium.

Dichlobenil (2,6-Dichlorobenzonitrile, Sigma-Aldrich, D57558, 60 mg/mL in DMSO) was used to induce temporary olfactory dysfunction (Bergman et al., 2002; Mori et al., 2000). Dichlobenil (120 mg/kg) or vehicle (DMSO, 2 mL/kg) was injected i.p. 24 h before behavioral testing (Lazarini et al., 2012). Both dichlobenil- and vehicle-treated mice were provided with normal chow softened by water on the bedding to prevent food or water deprivation should they fail to locate these resources on the wire top due to olfactory dysfunction. Animals’ body weight and general behaviors were monitored throughout the experiment.

Von Frey mechanical nociceptive test

The von Frey testing was performed in transparent Plexiglas chambers (9 cm x 9 cm x 9 cm) placed on a metal grid sheet elevated 40 cm above the benchtop. The metal grid sheet has round holes of 3 mm diameter with 5 mm spacing (center to center), to allow access to the hind paws with von Frey filaments.

Prior to beginning testing, mice were acclimated in the chambers for 3 h each day for at least three days. On each testing day, mice were placed in the chamber for additional acclimation, until they stopped exploring, approximately 3 h. The “ascending stimulus” method was used (Deuis et al., 2017), with forces ranging from 0.04 g to 4 g. Each filament was applied to the center of the mouse’s hind paw plantar surface (area surrounded by the toe pads) up to five times. The application of the filament and the delivery of the force was indicated by the bending of the filament. Each application lasted for 6-8 s, with at least 5 s between applications. A positive nociceptive response was noted when the mouse exhibited at least one of the following behaviors: twitching, shaking, stomping, quick withdraw and licking of the hind paw. These behaviors were sometimes accompanied by a brief and sudden opening of the plantar and toes spreading out, as well as guarding/defensive behaviors afterwards. If three stimuli of a given force failed to evoke a response, the filament of the next higher force was then tested. If three out of five stimulations evoked nociceptive responses, then the force of that filament was considered the mechanical nociceptive threshold for that paw, and that paw was not tested further for that data point.

Hargreaves thermal nociceptive test

The glass platform of the Hargreaves apparatus (IITC Model 400 Heated Base and Series 8 software Model 390G) was maintained at around 32 °C throughout the habituation and testing period (Gaskill et al., 2012). The intensity of the radiant light source was kept at 5% in idle, and at 22% during stimulation.

Von Frey testing always preceded thermal testing to avoid potential sensory sensitization induced by repeated heat stimuli. In the days leading up to testing, mice were acclimated in the acrylic enclosures (10 cm x 10 cm x 12.7 cm) on the Hargreaves heated glass platform for 1 h. On testing days, after completing the von Frey testing, mice were placed in the Hargreaves apparatus until they stopped moving around, approximately 30 to 45 min. The stimulus was applied at the center of the hind paw plantar surface, in the area surrounded by the toe pads. A timer was started at the initiation of the light stimulation, and manually stopped when the mouse exhibited a nociceptive response as defined above. Both left and right hind paws of each mouse were tested four times with at least 30 s between stimulations, and the final three averaged to define response latency for that paw. The first set of the total four sets of latency data were excluded due to inconsistent responses to the initial stimulation.

Buried food test

To confirm the effect of dichlobenil on olfactory function, animals were tested for their ability to locate buried food. For 2-3 consecutive days before the test, singly housed male and female mice received a piece of palatable food (chocolate chip or dried cranberries, Bio-Serv) in their cages in addition to chow. The cages were checked the following day to verify that the food had been consumed. Female mice were tested with chocolate chips. Later, male mice were tested with dried cranberries, since chocolate chips were not available at that time.

All chow pellets were removed 20 h before testing. On baseline testing day, each mouse was tested alone in a clean cage with fresh bedding in a clean room, without the presence of other mice or food. The mouse was allowed 5 min to explore the testing cage, then briefly removed while the experimenter buried a piece of palatable food 1 cm below the surface of the clean bedding, and then returned to the cage. A filter top was placed to prevent the mouse from escaping. The mouse’s behavior was observed from 2 m away, and the latency for the mouse to find the food recorded. If the mouse did not find the food within 15 min, the latency was recorded as 15 min. Chow pellets were replaced after the test when the animal was returned to the home cage after the test.

Latency to finding the palatable food was assessed at 1, 2 and 8 days after administration of dichlobenil, and compared to baseline latency. To determine whether any failure to locate the palatable food was due to loss of motivation for the food rather than olfactory dysfunction, the buried treat was moved onto the surface of the bedding in the path of the test mouse at the end of the 15-min test. The mouse’s consummatory behavior when they came across the food was documented.

Prolonged bedding exposure

Female C57BL6/J mice were separated into two rooms for habituation. This experiment was performed using females since it has been shown that males develop mechanical hypersensitivity when cohoused with CFA-treated same sex conspecifics (Smith et al., 2016). In each room, mice were singly caged with wire tops without filter lids and cages placed on three-shelf racks (150 cm width x 61 cm depth x 188 cm height). These mice were divided into six treatment groups: CFA and SAL, mice receiving intra-plantar injection of CFA or saline; BS-CFA and BS-SAL, mice that were “bystanders” of either CFA or SAL mice housed in the same room with the corresponding treated animals; OLF-CFA and OLF-SAL, mice that exposed to the bedding from CFA or SAL mice and housed in different rooms from the above and each other. Cages of BS mice and mice receiving intra-plantar injections were placed next to each other alternately in their designated room. Similarly cages of OLF mice and cages containing the bedding were placed next to each other alternately in their room. There was no additional space between cages.

All six groups and a group of naïve females were tested for mechanical and thermal nociceptive responsiveness. Measurements were taken on 2 consecutive days before any treatment, and the average of the two days used as baseline (BL, Fig 1A). CFA was injected in the left hind paw of the CFA and SAL mice (marked as Tx), and CFA mice and their bystanders were tested on days 1, 3, 5, 7, 9, 11, 13, and 15 post-Tx. At the end of testing day 1, the thickness of the CFA-injected paw was measured as an indicator of the inflammation. Control experiments in which mice receiving saline intraplantar injections and their bystanders were tested were conducted separately on different days.

Figure 1.

Female C57BL6/J mice co-housed with CFA-treated conspecifics or exposed to their bedding for a prolonged period developed hypersensitivity to mechanical (von Frey), but not thermal (Hargreaves) stimulation. (A) Experimental timeline. Baseline (BL) thresholds of all groups were measured 1-2 days before their respective treatment or the initiation of prolonged treatment (Tx). Nociceptive tests (T) were performed on odd-numbered days post-Tx. Bedding from CFA- or SAL-treated mice was transferred every 24 h, before the testing day for the olfactory group. (B, C) Time course of von Frey thresholds of the left hind paw. Horizontal dashed lines indicate the combined baseline (BL) thresholds across treatment groups. (D, E) Time course of paw withdrawal (PW) latencies of the left hind paw. Dashed line marks the average baseline PW latency. Sample size: naïve, N = 5; CFA, N = 17; BS-CFA, N = 18; OLF-CFA, N = 16; SAL, N = 12; BS-SAL, N = 12; OLF-SAL, N = 16. Data were analyzed with 2-way ANOVA with repeated measures followed by Dunnett’s test to compare nociceptive responses on post-Tx days to BL within treatment groups. #, p < 0.05; ##, p < 0.01; ###, p < 0.001; ####, p < 0.0001 for CFA group. &, p < 0.05; &&&, p < 0.001 for BS-CFA group. ****, p < 0.0001 for OLF-CFA group in (B).

To expose OLF-CFA and OLF-SAL to bedding from CFA and SAL mice, clean and empty shoe box cages were placed adjacent to home cages of the OLF-CFA and OLF-SAL mice, and were used to hold soiled bedding during the exposure. Soiled bedding (50-60 g in total) was collected from 6 cages of CFA or SAL mice on testing days. Clean bedding was added to replace that removed and ensure consistency in bedding quantity. The soiled bedding from the 6 cages was evenly mixed and distributed into the empty shoe box cages starting on the first testing day of bystander groups (marked as Tx for OLF groups, Fig 1A). Every other day after that, newly collected soiled bedding was added to the existing bedding. Twenty-four h after every addition of soiled bedding, mice in OLF groups underwent nociceptive testing, i.e., OLF mice were tested on days 1, 3, 5, 7, 9, 11, 13, and 15 post-Tx.

Acute bedding exposure

We next determined whether 1 h exposure to either a conspecific or bedding from a conspecific subjected to CFA treatment was sufficient to induce behavioral hypersensitivity (experimental timeline in Fig 3A). Mice in the BS-CFA and BS-SAL groups were respectively allowed to interact with a CFA- or saline-treated conspecific for a period of 1 h. Mice in the OLF-CFA and OLF-SAL groups were respectively exposed to bedding from CFA- and saline-treated conspecifics for a period of 1 h. During the exposure, bedding was placed in the testing chamber but physically separated from mice with Plexiglas dividers. Mechanical threshold (von Frey) was determined prior to exposure (baseline, BL), immediately after interaction or bedding exposure (0 h), and at 1 and 3 h post-exposure.

Figure 3.

Exposure to soiled bedding from CFA-treated conspecifics induced mechanical nociceptive hypersensitivity comparable to that from direct interaction with CFA conspecifics. (A) Experimental design. (B, C) Time course of von Frey thresholds of male (B) and female (C) mice in BS-SAL, BS-CFA, and OLF-CFA groups. Measurements at 0 h were obtained immediately after social interaction or bedding exposure. 2-way ANOVA with repeated measures followed by a post-hoc pairwise comparison between treatment groups in males and females. *, p < 0.05; **, p < 0.01; ***, p < 0.001 comparing to BS-SAL group. Female sample size: BS-SAL, N = 6; BS-CFA, N = 6; OLF-CFA, N = 8. Male sample size: BS-SAL = 9; BS-CFA, N = 9; OLF-CFA, N = 9.

The effect of dichlobenil (Dich) or vehicle (VEH, DMSO) on nociceptive behaviors in BS and OLF groups was assessed immediately and at 1, 3, and 24 h. The experimenter was blinded to Dich/VEH treatment.

SAL and CFA Bedding Collection for GC-MS

Soiled bedding from male stimulus mice in the CFA and SAL treatment groups was collected before and 2 d after intra-plantar injection for each of three experimental cohorts. Clean bedding, which contained compounds existing in the bedding material and absorbed from the room atmosphere throughout the experiment, was collected the same time as the soiled bedding and used as a background control. All bedding was from the same batch to maximize the consistency of the samples. Three samples (4 g each) from the CFA-treated and SAL-treated groups were placed in 40 mL VOA vials with Teflon-lined septa (Restek Pure Chromatography, Bellefonte PA).

Headspace Solid Phase Microextraction (SPME) Gas Chromatography/Mass Spectrometry (GC/MS) Analysis

SPME GC/MS analysis was performed using a Leco Pegasus 4D GC × GC-TOFMS (Leco, St. Joseph, Michigan). The SPME assembly (Sigma-Aldrich, St. Louis, MO) used polydimethylsiloxane/divinylbenzene-coated fibers. Each SPME sampling event was carried out by inserting the fiber into a bedding sample vial for 30 min. The fiber was then inserted into the GC injector (235 °C, 10 min), which contained an SPME liner (Restek, Bellefonte, PA) for desorption. The GC oven temperature for each sample was programed as follows: held at the initial 45 °C for 3 min, then increased by 10 °C/min increment to reach and hold at 200 °C for 2 min, followed by 5 °C /min increment to 245 °C. The GC column (Agilent DB-VRX with 1.4 µm film thickness, 0.25 mm i.d., and 60 m length) interfaced directly into the MS source (200 °C). The MS detector was at 1650 V, 50 spectra/s, 35 to 500 atomic mass units. Samples from each cohort were analyzed on one day. The data were processed using the Leco ChromaTOF software with signal noise ratio (S/N) at 30. Compound identification was made by matching spectra to the NIST mass spectral library.

Data Analysis

Von Frey test.

Mechanical thresholds for the injected (left) paw are reported for experiments that included CFA- or SAL-injected animals. In the acute exposure experiments, in which there were no paw injections, thresholds for the two hindpaws were determined and averaged to define the threshold for that animal. Temporal smoothing was applied by averaging thresholds obtained on two consecutive testing days. Data were analyzed with two-way and three-way ANOVA with repeated measures, followed by a post-hoc Dunnett’s multiple comparisons test within treatment groups comparing to BL, or a post-hoc Sidak’s multiple comparisons test between treatment groups at individual time points.

Hargreaves test.

Paw withdrawal (PW) latency is reported for the left hindpaw. After temporal smoothing, data were analyzed using two-way ANOVA with repeated measures, followed by a post-hoc Dunnett’s multiple comparisons test within treatment groups comparing to BL, or a post-hoc Sidak’s multiple comparisons test between treatment groups.

Buried food test.

Latencies to finding the palatable treats before (baseline) and at different time points after administration of dichlobenil or vehicle were analyzed with two-way ANOVA with repeated measures on Time, followed by post-hoc Dunnett’s multiple comparisons test comparing to baseline, and Sidak’s multiple comparisons test between treatment groups.

GC-MS.

To semi-quantifiably analyze the volatile chemical composition of the bedding sample, peak area was used to indicate the relative abundance of the compound. Relative abundance (RA) was calculated by averaging the peak areas of the compound from the triplicate samples. Fold change (FC) after CFA or SAL injection was calculated for each compound using the formula:

$$FC=\frac{{RA}_{post-inj}^{soiled}-{RA}_{post-inj}^{clean}}{{RA}_{pre-inj}^{soiled}-{RA}_{pre-inj}^{clean}}$$

FC for each compound was compared to FC = 1 using a one-sample t-test.

Data are presented as mean ± SEM, and p < 0.05 was considered statistically significant. GraphPad Prism (v.9.4.1) was used to perform statistical analyses and to generate result figures. Detailed statistics are shown in Table 2. Experimental schematics were visualized with BioRender.

Table 2.

Statistical analyses for all results.

Figures	Test type	Stats
		Main effect	F (DFn, DFd)	p value	Post-hoc
Figure 1B	Two-way RM ANOVA	Time x Treatment	F (12, 208) = 5.326	P<0.0001	Dunnett’s
		Time	F (3.440, 178.9) = 12.24	P<0.0001
		Treatment	F (3, 52) = 23.75	P<0.0001
Figure 1C	Two-way RM ANOVA	Time x Treatment	F (12, 164) = 2.301	P=0.0098
		Time	F (3.198, 131.1) = 0.4249	P=0.7479
		Treatment	F (3, 41) = 6.902	P=0.0007
Figure 1D	Two-way RM ANOVA	Time x Treatment	F (12, 208) = 3.449	P=0.0001	Dunnett’s
		Time	F (3.132, 162.8) = 5.221	P=0.0015
		Treatment	F (3, 52) = 2.953	P=0.0410
Figure 1E	Two-way RM ANOVA	Time x Treatment	F (12, 164) = 0.9977	P=0.4535
		Time	F (3.510, 143.9) = 0.4697	P=0.7335
		Treatment	F (3, 41) = 3.898	P=0.0154
Figure 2A	Two-way RM ANOVA	Time x Paws	F (3.227, 51.64) = 17.32	P<0.0001	Sidak’s
		Time	F (3.182, 50.91) = 14.74	P<0.0001
		Paws	F (1.000, 16.00) = 80.45	P<0.0001
Figure 2B	Two-way RM ANOVA	Time x Paws	F (3.610, 61.37) = 2.027	P=0.1082
		Time	F (2.792, 47.46) = 15.75	P<0.0001
		Paws	F (1.000, 17.00) = 0.6699	P=0.4244
Figure 2C	Two-way RM ANOVA	Time x Paws	F (3.197, 51.15) = 7.705	P=0.0002	Sidak’s
		Time	F (2.646, 42.33) = 6.792	P=0.0012
		Paws	F (1.000, 16.00) = 18.79	P=0.0005
Figure 2D	Two-way RM ANOVA	Time x Paws	F (2.387, 40.59) = 2.890	P=0.0583
		Time	F (1.781, 30.28) = 2.551	P=0.1000
		Paws	F (1.000, 17.00) = 0.00057	P=0.9812
Figure 3B	Two-way RM ANOVA	Time x Treatment	F (6, 72) = 1.227	P=0.3029
		Time	F (2.531, 60.74) = 16.48	P<0.0001
		Treatment	F (2, 24) = 14.70	P<0.0001	Pairwise comparison
Figure 3C	Two-way RM ANOVA	Time x Treatment	F (6, 51) = 4.907	P=0.0005
		Time	F (2.508, 42.64) = 10.11	P<0.0001
		Treatment	F (2, 17) = 6.642	P=0.0074	Pairwise comparison
Figure 4A	Two-way RM ANOVA	Time x Treatment	F (3, 24) = 27.42	P<0.0001	Sidak’s, Dunnett’s
		Time	F (3, 24) = 29.89	P<0.0001
		Treatment	F (1, 8) = 44.32	P=0.0002
Figure 4B	Two-way RM ANOVA	Time x Treatment	F (3, 24) = 61.53	P<0.0001	Sidak’s, Dunnett’s
		Time	F (3, 24) = 55.51	P<0.0001
		Treatment	F (1, 8) = 161.6	P<0.0001
Figure 5B	Two-way RM ANOVA	Time x Bedding	F (4, 64) = 2.194	P=0.0795
VEH		Time	F (2.622, 41.96) = 0.9026	P=0.4367
		Bedding	F (1, 16) = 9.357	P=0.0075
Figure 5B	Two-way RM ANOVA	Time x Bedding	F (4, 52) = 0.2691	P=0.8966
Dich		Time	F (2.608, 33.91) = 1.418	P=0.2559
		Bedding	F (1, 13) = 0.0002187	P=0.9884
Figure 5C	Two-way RM ANOVA	Time x Treatment	F (4, 64) = 1.547	P=0.1994
VEH		Time	F (3.261, 52.18) = 6.463	P=0.0006
		Treatment	F (1, 16) = 30.47	P<0.0001
Figure 5C	Two-way RM ANOVA	Time x Treatment	F (4, 64) = 0.5764	P=0.6808
Dich		Time	F (2.185, 34.97) = 3.467	P=0.0386
		Treatment	F (1, 16) = 0.01579	P=0.9016
Figure 5D	Two-way RM ANOVA	Time x Bedding	F (4, 60) = 1.248	P=0.3006
VEH		Time	F (2.879, 43.19) = 9.725	P<0.0001
		Bedding	F (1, 15) = 10.46	P=0.0056
Figure 5D	Two-way RM ANOVA	Time x Bedding	F (4, 60) = 1.573	P=0.1931
Dich		Time	F (3.049, 45.73) = 6.197	P=0.0012
		Bedding	F (1, 15) = 0.002088	P=0.9642
Figure 6B	One-sample t-test			P values see Table 1
Figure 6C	One-sample t-test			P values see Table 1
Suppl. Figure 1	Paired t-test		t=15.01, df=16	P<0.0001
Suppl. Figure 3	Two-way RM ANOVA
A		L-R paw	F (1, 8) = 0.02148	P=0.8871
B		L-R paw	F (1, 5) = 0.2941	P=0.6109
C		L-R paw	F (1, 8) = 0.08889	P=0.7732
D		L-R paw	F (1, 5) = 0.1049	P=0.7591
E		L-R paw	F (1, 8) = 0.1188	P=0.7392
F		L-R paw	F (1, 7) = 0.2452	P=0.6357

Results

Prolonged exposure to CFA-treated mice and their bedding induced mechanical but not thermal hypersensitivity in otherwise naïve mice.

The first experiment determined whether 15-d exposure to bedding from CFA-treated conspecifics (OLF-CFA group) induced hyperalgesia. For comparison, whether “bystander” mice (BS-CFA) housed in the same room as the CFA-treated animals exhibited hyperalgesia was also determined (Fig 1A). CFA induced inflammatory edema in the injected (left) paw (Supplemental Fig 1), and caused both mechanical and thermal hypersensitivity in the injected paw that lasted for the duration of the experiment (Solid red circles in Figs 1B and 1D). Both OLF-CFA and BS-CFA mice developed lower mechanical thresholds (OLF-CFA: p < 0.0001 at all time points compared to BL. BS-CFA: p < 0.05 from day 9-15 compared to BL, Fig 1B). However, neither OLF-CFA nor BS-CFA mice showed statistically significant changes in their responses to thermal stimulation (In both groups, p > 0.09 for each time point compared to BL in post-hoc Dunnett’s test, Fig 1D).

Similar to naïve mice, vehicle control mice receiving intra-plantar saline injection did not show altered nociceptive behavior in either the von Frey or Hargreaves test (Figs 1C and 1E; p = 0.08 and p = 0.67, respectively, with one-way ANOVA repeated measure), and there were no changes in mechanical or thermal responses in the BS-SAL or OLF-SAL groups.

Though BS-CFA mice were housed individually, the cages were made with transparent material that would allow visual observation of neighboring animals’ nocifensive behaviors directed towards the injected (left) paw. To examine whether visual cues led to any lateral bias in nociceptive responding in the bystander group, we compared nociceptive thresholds of the left and right hind paws in BS-CFA and CFA groups. In the CFA-treated mice, mechanical and thermal thresholds were significantly reduced in the left compared to the right paw (Fig. 2A and 2C, p < 0.0001 and p < 0.001, respectively). By contrast, the BS-CFA mice showed similar nociceptive thresholds in the two hind paws (Fig. 2B and 2D). This observation indicated that any potential visual communication between CFA and BS-CFA mice in this setting did not result in lateral bias in nociceptive response.

Figure 2.

Mechanical and thermal hypersensitivity in left and right hind paws of CFA-treated (A and C) and BS-CFA (B and D) mice. CFA, N = 18; BS-CFA, N = 16. 2-way ANOVA repeated measures followed by a Sidak’s test between groups. ***, p < 0.001; ****, p < 0.0001 for main effect between left and right paws. &&&&, p < 0.0001 for main effect on Post-Tx days. For Sidak’s test, #, p < 0.05; ###, p < 0.001; ####, p < 0.0001 between left and right paws on individual Post-Tx days.

Acute exposure to CFA-treated mice or their bedding induced mechanical hypersensitivity in otherwise naïve mice

We next determined whether an acute bedding exposure would be sufficient to induce enhanced nociception. We examined the von Frey responses of three groups of male and female mice: OLF-CFA, mice who had 1-h exposure to bedding of CFA-treated conspecifics; BS-CFA, mice with 1-h interaction with the same CFA-treated conspecifics, and BS-SAL, control mice with 1-h interaction with SAL-treated conspecifics (Fig 3A). During the 1-h interaction with same-sex conspecifics, no fighting was observed in male mice. The von Frey thresholds were measured at 0, 1 and 3 h after bedding exposure or interaction (Figs 3B and C). In both male and female mice, OLF-CFA and BS-CFA groups exhibited significantly lower mechanical thresholds than BS-SAL controls (effect of treatment: male, p < 0.0001 for BS-CFA and p = 0.0035 for OLF-CFA; female, p = 0.024 for BS-CFA and p = 0.0037 for OLF-CFA compared to BS-SAL), and developed mechanical hyperalgesia compared to baseline. Mechanical hypersensitivity was seen in both hind paws in both male and female bystanders and bedding-exposed mice (Supplemental Fig 2C–F).

Olfactory ablation inhibited the acute hyperalgesic effect of soiled bedding from same-sex conspecifics.

Ablation of the olfactory epithelium (via dichlobenil) was used to determine the contribution of olfactory function to hyperalgesia. The efficacy and duration of olfactory ablation was verified using the buried food test. Compared to vehicle (VEH)-injected control mice, mice receiving dichlobenil (Dich) failed to retrieve the treats buried underneath the bedding within the 15-min cut-off for two days after Dich administration (Fig 4). When the same treat was placed on the surface of the bedding, each mouse located and consumed the treat within 1 min (data not shown). A retest on post-injection day 8 revealed that the ability to locate the treat had recovered to baseline levels. Based on these results, Dich or VEH was administered to OLF-CFA or BS-CFA mice 24 h before acute bedding exposure or interaction, and von Frey testing was conducted within 2 d after administration (experimental scheme in Fig 5A).

Figure 4.

Partial ablation of olfactory epithelium transiently prevented both male and female mice from finding buried treats. Testing before and at 1, 2 and 8 days after i.p. injection of dichlobenil or vehicle (N = 5 /sex/treatment). 2-way ANOVA with repeated measures, followed by a Dunnett’s Test for comparison with Day 0. Sidak’s multiple comparisons test was performed to compare latencies between treatment groups at the same time points. *, p < 0.05; ****, p < 0.0001 between VEH and Dich on individual days. ####, p < 0.0001 comparing to day 0 within Dich group.

Figure 5.

Olfactory ablation prevented mice in both OLF and BS groups from developing nociceptive hypersensitivity after acute exposure to CFA-bedding or interaction with CFA conspecifics. (A) Experimental timeline. (B – D) Time course of von Frey thresholds of olfaction-intact (VEH, left panel) or ablated (Dich, right panel) mice in male OLF (B), male BS (C), and female OLF (D) groups following bedding or social interaction associated with CFA- (solid red circle) or SAL-treated (blue open squares) conspecifics. Within VEH or Dich treatment groups, data were analyzed with 2-way ANOVA [Stimulus (CFA, SAL) x Time (BL, 0, 1, 3, and 24h)] (B) Male OLF mice: there was main effect of bedding (##, p = 0.0075) in VEH group; no significant difference was detected in Dich group. Sample size: VEH-CFA, N = 9; Dich-CFA, N = 9; VEH-SAL, N = 9; Dich -SAL, N = 6. (C) Male BS mice: main effect of bedding (####, p < 0.0001) and time (p = 0.0006) were detected in VEH group, with no interaction (p = 0.20). In Dich group, main effect of time (p = 0.039) but not treatment (p = 0.90). N = 9 in each treatment group. (D) Female OLF mice: main effect of bedding (##, p = 0.0056) and an effect of time (p < 0.0001), but no interaction (p = 0.30) in VEH group. A main effect of time (p = 0.0012) was detected in Dich group. Sample size: VEH-CFA, N = 8; Dich-CFA, N = 9; VEH-SAL, N = 9; Dich-SAL, N = 8.

VEH-treated mice exhibited significantly lower mechanical thresholds when exposed to bedding from the CFA-treated donors compared to SAL-treated donors (effect of donor group: p = 0.0075 for males, Fig 5B; p = 0.0056 for females, Fig 5D). By contrast, Dich-treated animals did not develop hyperalgesia when exposed to bedding from CFA-treated conspecifics (right panels in Fig 5B and C, main effect of bedding group p < 0.96 for both sexes). We also tested Dich-treated male bystander mice. There was no difference in mechanical nociception between Dich-treated BS-CFA and Dich-treated BS-SAL groups (p > 0.90, Fig 5C, right). By contrast, VEH-treated, olfaction-intact bystander mice showed significant hyperalgesia following a 1 h interaction with CFA-treated animals compared to interaction with SAL-treated animals (BS-CFA versus BS-SAL, p < 0.0001, Fig 5C, left).

These data confirm that a 1-h exposure to bedding from same-sex conspecifics with CFA-induced inflammatory pain induces enhanced mechanical hypersensitivity in both male and female mice. More importantly, they indicate that olfactory function is required for social enhancement of nociception produced by exposure of mice to conspecifics subjected to persistent inflammation.

CFA treatment altered the profile of volatile compounds in bedding

To identify the volatile odorant compounds in the bedding from CFA-treated animals that have the potential to convey information to conspecifics, we performed GC-MS analysis on bedding from CFA- and SAL-treated donors. GC-MS analysis revealed approximately 200 – 300 volatile compounds in each sample, which would include compounds secreted by mice and released from the bedding material, as well as other background chemicals in the experimental environment.

CFA or SAL injection did not result in appearance or disappearance of any compound in soiled bedding. There were 110 compounds found in the CFA group before and after injection, and 130 in the SAL group, with 101 of these seen in both groups (Fig 6A, Table 1). Of the shared 101 volatile compounds, 8 were consistently more abundant in clean bedding (2 to 9.8 fold, indicated by superscript c in Table 1). These likely existed in the environment, but declined in soiled bedding, possibly due to interaction with mice or their excretions. Another 31/101 compounds were found at equivalent levels in both clean and soiled bedding, suggesting that these arose from the bedding itself or the environment, and not from mice. After excluding compounds from the environment and SPME process, the remaining 61 compounds were more abundant in the soiled than clean bedding, both before and after intra-plantar injection, including 22 compounds detected exclusively in soiled bedding. This indicated that mice were likely the source of these 61 compounds, and we focused our analysis on these compounds.

Figure 6.

GC-MS analysis of volatile compounds in bedding from CFA- and SAL-treated animals (A) Number of detected compounds categorized based on association with CFA and SAL injection, and their relative abundance in soiled vs. clean bedding. (B and C) Fold change (FC) in relative abundance 2 days after stimulus mice received CFA (B) or SAL (C) injection compared to pre-injection for compounds that were richer in soiled bedding than clean bedding. Compounds with significant changes in relative abundance (p < 0.05) are highlighted in red, and labeled with numbers corresponding to the numbers of chemicals in Table 1.

Table 1. Volatile compounds that appeared in analysis results of soiled samples.

VOCs that appeared in analysis of soiled samples, with the retention time (R.T.) in minutes, GC-MS identification, molecular formula, and the mean fold change (FC) in abundance 2 days after mice receiving CFA or SAL treatment. The abundance of VOCs used to generate the mean FC was calculated by subtracting the abundance in clean bedding from that in soiled bedding. P value was calculated from three independent experiments. VOCs without mean FC listed for either CFA or SAL bedding are sometimes more abundant in the soiled bedding, but other times more in the cleaning bedding. Superscript c: VOCs appeared consistently lower in the soiled bedding than clean bedding; Superscript s: VOCs present in the environment or from the SPME fiber material.

#	R.T. (min)	GC-MS ID	Formula	CFA bedding		SAL bedding
				Mean FC	p value		Mean FC	p value
1	3.40	Argon c, s	Ar	2.55	0.1733	0.91	0.6395
2	3.48	Carbon dioxide s	CO2	1.62	0.3674	1.49	0.5305
3	3.85	Carbonyl sulfide s	COS
4	4.43	Acetaldehyde c	C2H4O	1.02	0.7798	1.49	0.3689
5	4.49	2-Butene c	C4H8	1.46	0.4724	0.94	0.8068
6	4.56	Butane c	C4H10	3.03	0.3575	1.31	0.1055
7	4.92	Methyl formate	C2H4O2
8	5.30	Trimethylamine	C3H9N	2.74	0.1858	1.97	0.078
9	6.00	Acetone	C3H6O	1.49	0.4602	0.66	0.1928
10	6.13	Furan	C4H4O
11	6.31	Cyclopropane, ethylidene-	C5H8	1.20	0.6321	1.10	0.7239
12	6.54	Dimethyl sulfide	C2H6S	1.17	0.3043	1.33	0.4853
13	6.75	Acetic acid, methyl ester	C3H6O2	1.27	0.5680	0.29	0.0065
14	7.00	Carbon disulfide	CS2	1.42	0.5451	0.91	0.8080
15	7.17	1-Propanol	C3H8O	1.04	0.8869	0.76	0.5276
16	7.32	Silanol, trimethyl- s	C3H10OSi
17	7.33	Cyclopentene	C5H8
18	7.38	Propanal, 2-methyl-	C4H8O
19	7.40	Methane, nitro-	CH3NO2	1.03	0.8327	1.14	0.5381
20	7.57	Acetic acid	C2H4O2
21	7.66	Methacrolein c	C4H6O	1.11	0.3318	2.04	0.4548
22	8.07	Acetic acid ethenyl ester	C4H6O2
23	8.14	Butanal c	C4H8O	1.18	0.1058	2.40	0.3835
24	8.27	2-Butanone	C4H8O	2.46	0.4003	0.69	0.1587
25	8.50	Furan, 2-methyl-	C5H6O	0.98	0.9656	0.94	0.8556
26	8.53	3-Buten-2-ol, 2-methyl-	C5H10O	1.70	0.4694	0.78	0.0352
27	8.62	Thiirane, methyl-	C3H6S	2.26	0.3133	0.98	0.9285
28	8.71	Trichloromethane	CHCl3	1.63	0.3319	1.14	0.6454
29	8.74	Furan, 2-methyl-	C5H6O	1.60	0.4111	0.93	0.7671
30	8.85	p-Dithiane-2,5-diol	C4H8O2S2	1.52	0.4593	1.23	0.4529
31	9.68	1-Butanol	C4H10O	1.02	0.9305	0.68	0.2958
32	9.75	Pentanal	C5H10O
33	9.85	Silanediol, dimethyl-	C2H8O2Si	1.65	0.4903	1.42	0.6481
34	9.88	2-Butanone, 3-methyl-	C5H10O	1.57	0.5792	0.84	0.3287
35	9.93	Propanoic acid	C3H6O2	1.83	0.2970	1.86	0.5994
36	10.11	Benzene	C6H6	1.88	0.4070	1.07	0.5565
37	10.24	3-Buten-2-one, 3-methyl-	C5H8O	1.34	0.1837	1.10	0.8261
38	10.43	Nitric acid, 1-methylethyl ester	C3H7NO3	0.64	0.2518	0.26	0.0048
39	10.53	2-Pentanone	C5H10O	2.49	0.4016	0.52	0.1143
40	10.53	Cyclopropane, butyl-	C7H14
41	10.65	2,3-Pentanedione	C5H8O2	0.90	0.0033	0.70	0.5057
42	10.72	Pentanal c	C5H10O	1.08	0.6368	3.40	0.3937
43	10.74	Heptane	C7H16
44	10.83	Methyl thiolacetate	C3H6OS	1.11	0.8663	0.97	0.9025
45	10.86	Furan, 2-ethyl-	C6H8O	1.16	0.8035	0.96	0.8292
46	11.03	Furan, 2,5-dimethyl-	C6H8O	0.79	0.5743	0.80	0.3091
47	11.57	Butanenitrile, 3-methyl-	C5H9N	1.37	0.2946	1.15	0.6914
48	11.86	Methyl Isobutyl Keton; 2-Hexanone	C6H12O	1.46	0.5448	0.80	0.5512
49	11.90	3-Penten-2-one	C5H8O	1.97	0.3181	1.02	0.9104
50	12.00	2-Butenal, 2-methyl-, (E)-	C5H8O	1.41	0.3646	0.93	0.4408
51	12.07	Disulfide, dimethyl	C2H6S2	1.94	0.3931	0.88	0.6265
52	12.19	2-Pentanone, 3-methyl-	C6H12O	1.10	0.7631	0.63	0.1208
53	12.45	2-Buten-1-ol, 3-methyl-	C5H10O	3.26	0.3621	0.82	0.0861
54	12.53	1-Penten-3-one, 2-methyl-	C6H10O	1.51	0.2715	0.53	0.2298
55	12.73	Toluene	C7H8
56	12.82	3-Hexanone	C6H12O	1.47	0.3847	0.72	0.1427
57	12.90	2-Octene	C8H16
58	12.96	2-Hexanone	C6H12O	1.91	0.4942	0.48	0.0459
59	12.98	Furan, 2-propyl-	C7H10O
60	13.09	Octane	C8H18
61	13.13	Furan, 2-ethyl-5-methyl-	C7H10O	0.83	0.4802	0.82	0.2960
62	13.17	Hexanal	C6H12O
63	13.26	2-Hexenal, 2-ethyl-; 2-Pentenal, 2,4,4-trimethyl-	C8H14O	2.27	0.1565	1.54	0.4405
64	13.30	Cyclotrisiloxane, hexamethyl- s	C6H18O3Si3
65	13.72	1,3-Octadiene	C8H14
66	14.03	Ethanone, 1-(2-methyl-1-cyclopenten-1-yl)-	C8H12O	1.24	0.3474	1.73	0.1574
67	14.06	Furfural	C5H4O2
68	14.18	Methyl ethyl disulfide	C3H8S2	2.07	0.2908	0.84	0.4565
69	14.22	2-Furanmethanol	C5H6O2
70	14.45	Cyclopentanone, 3-methyl-	C6H10O	1.73	0.3684	0.98	0.9203
71	14.49	Oxime-, methoxy-phenyl-_	C8H9NO2	2.14	0.4053	0.84	0.5524
72	14.68	2-Hexenal, 2-ethyl-	C8H14O	1.83	0.2781	1.00	0.9595
73	14.82	Ethylbenzene	C8H10
74	15.04	3-Heptanone	C7H14O
75	15.06	Benzene, 1,3-dimethyl-	C8H10
76	15.15	2-Heptanone	C7H14O	1.28	0.6905	0.42	0.0231
77	15.19	2-n-Butyl furan	C8H12O
78	15.37	Heptanal	C7H14O
79	15.61	o-Xylene	C8H10
80	15.67	Dimethyl sulfone	C2H6O2S	2.07	0.2336	1.09	0.5187
81	15.75	2(5H)-Furanone	C4H4O2	1.99	0.5306	3.86	0.4859
82	16.11	Butane, 1-isothiocyanato-	C5H9NS	1.21	0.4228	1.15	0.6293
83	16.41	3-Heptanone, 6-methyl-	C8H16O	1.18	0.6110	0.77	0.2113
84	16.48	2-Heptanone, 6-methyl-	C8H16O
85	16.58	Cyclotetrasiloxane, octamethyl- s	C8H24O4Si4	1.35	0.6116	0.64	0.0684
86	16.84	2-Furancarboxaldehyde, 5-methyl-	C6H6O2
87	16.97	Benzaldehyde	C7H6O	2.97	0.4334	0.72	0.1929
88	17.03	5-Hepten-2-one, 6-methyl-	C8H14O	0.96	0.8242	0.91	0.6225
89	17.14	6-Hepten-3-one, 4-methyl-	C8H14O	1.34	0.5015	0.50	0.0249
90	17.14	Furan, 2-pentyl-	C9H14O
91	17.17	Dimethyl trisulfide	C2H6S3	1.78	0.3757	0.85	0.5015
92	17.36	Octanal	C8H16O
93	17.64	Benzene, 1,2,3-trimethyl-	C9H12
94	17.76	Cyclotrisiloxane, hexamethyl- s	C6H18O3Si3	3.12	0.3708	1.14	0.6943
95	18.26	Sorbic acid vinyl ester	C8H10O2	1.01	0.9396	1.43	0.1162
96	18.28	6,8-Dioxabicyclo[3.2.1]octane, 7-ethyl-5-methyl-, (1R-exo)-	C9H16O2	1.27	0.4736	1.01	0.9031
97	18.38	7-Exo-ethyl-5-methyl-6,8-dioxabicyclo[3.2.1]oct-3-ene	C9H14O2	1.37	0.4388	1.15	0.3517
98	19.16	6,7-Dioxabicyclo[3.2.1]octane, 1-methyl-	C7H12O2	1.23	0.5155	1.14	0.4498
99	19.20	Nonanal c	C9H18O	1.90	0.0759	2.68	0.3098
100	19.42	Cyclopentasiloxane, decamethyl- s	C10H30O5Si5	1.32	0.5832	0.97	0.9296
101	20.15	Disulfide, methyl (methylthio)methyl	C3H8S3	1.95	0.2159	0.97	0.8956

Further analysis of the 61 compounds more abundant in soiled bedding revealed differences between the CFA and SAL groups. In the CFA-treated group, 55 of 61 (90.1%) compounds increased in abundance (Fig 6B), indicating faster accumulation than evaporation. By contrast, only 23 compounds showed an increase in abundance (37.7%) while 38 decreased in the SAL group (Fig 6C, CFA and saline groups significantly different, Fisher’s exact test, p < 0.0001), the latter indicating compounds that accumulated more slowly than they evaporated. For 6 of 38 compounds, this decrease was statistically significant: methyl ester acetic acid (Compound #13), 2-methyl-3-buten-2-ol (#26), 1-methylethyl ester nitric acid (#38), 2-hexanone (#58), 2-heptanone (#76) and 4-methyl-6-hepten-3-one (#89) (Fig 6C, highlighted in red). In comparison, most of these were not decreased, and in fact were moderately elevated following CFA injection (highlighted in orange in Fig 6B, also Fig 7A), although these increases were not statistically significant. One compound, 2,3-pentanedione, showed a statistically significant decrease in abundance following CFA treatment (Fig 6B, #41, highlighted in red).

Figure 7.

FC in abundance in post-CFA and post-SAL bedding for compounds significantly richer in soiled bedding (A) or in clean bedding (B). (A) Seven compounds that were richer in soiled bedding after CFA (1 compound) or SAL (6 compounds) injection. (B) A total of eight chemicals were richer in clean bedding and showed a modest decrease in abundance after both CFA and SAL injection.

Eight compounds that were less abundant in soiled compared to clean bedding were argon (#1), acetaldehyde (#4), 2-butene (#5), butane (#6), methacrolein (#21), butanal (#23), pentanal (#42), and nonanal (#99). Some of these 8 compounds showed further decrease in abundance in bedding 2 days after SAL or CFA injection (Fig 7B), but no changes were statistically significant.

Discussion

The present study expands the growing examples demonstrating social transfer of hyperalgesia between rodents (Martin et al., 2014; Mogil, 2015; Smith et al., 2016; 2021; Lu et al.,2018; Luo et al., 2020). While rodents housed in individually ventilated cages are unlikely to be influenced by animals in other cages, the social context of animals undergoing nociceptive tests is important. Specifically, here we showed that contact-free presence of soiled bedding from mice subjected to CFA-induced inflammation and hyperalgesia led to mechanical hypersensitivity in otherwise naïve C57BL/6J mice after both prolonged and acute exposures. Transient partial lesion of the main olfactory epithelium prevented this hypersensitivity in male and female mice, demonstrating that the pro-nociceptive effect required the main olfactory system. Chemical analysis of volatile compounds in the bedding revealed a general increase in the abundance of potential odor molecules in the CFA-treated group, whereas a large proportion of such compounds showed a decrease in the saline-treated group. Thus, behavioral changes in nociception induced by olfactory signals are unlikely to be explained by a change in abundance of one or two molecules.

In both long-term and short-term exposure paradigms, bedding from same-sex conspecifics treated with CFA was sufficient to produce mechanical hypersensitivity. However, while CFA-treated mice exhibited thermal hypersensitivity in Hargreaves test, neither “bystander” mice nor mice exposed to bedding from CFA-treated animals showed significant thermal hyperalgesia. Thermal hypersensitivity has been reported in bystanders using various thermal tests, such as tail immersion test, tail flick and hotplate tests (Smith et al., 2016; Smith et al., 2021; Mohammadi et al., 2020a; Mohammadi et al., 2020b), as well as Hargreaves test (Langford et al., 2006). In our study, the lack of thermal hyperalgesia in bystander mice using Hargreaves test could be a result of multiple factors, including social context during testing (e.g., dyads vs. individuals), the intensity of radiant heat stimulation, and other environmental factors that we may not understand fully.

In the present study, olfactory ablation using dichlobenil prevented CFA-bedding-induced nociceptive hypersensitivity. Dichlobenil damages the dorsomedial area of the main olfactory epithelium, an area that contains olfactory sensory neurons (Bergman et al., 2002; Mancuso et al., 1997; Vedin et al., 2004), but spares vomeronasal neurons (Denizet et al., 2017; John and Key, 2003; Piras et al., 2003). Because impairments of olfactory behavior are not necessarily closely matched morphological evidence of olfactory damage (Denizet et al., 2017; Genter et al., 1996; Lazarini et al., 2012; Lee et al., 2021; Yoon et al., 2005), we validated dichlobenil’s ability to impair olfactory function using the buried food test and tested the effects of bedding exposure within the 2-d window of confirmed olfactory impairment. We also confirmed that dichlobenil treatment did not reduce animals’ motivation to consume the treats. The fact that dichlobenil interfered with hypersensitivity induced by exposure to both CFA-treated animals and their bedding strongly argues that olfaction was the primary sensory channel mediating hypersensitivity in these experiments.

Our finding that ablation of the main olfactory epithelium prevented hyperalgesia induced by interaction with a CFA-treated conspecific is inconsistent with an earlier report (Langford et al., 2006) that olfactory lesion by ZnSO₄ did not affect hyperalgesia in mice interacting with a cage mate exhibiting writhing induced by acetic acid. However, mice with acetic acid-induced abdominal pain display prominent spontaneous nociceptive behaviors, by comparison with mice subjected to localized inflammation of a single extremity, and visual observation of spontaneous pain behaviors has often been reported to significantly impact the behavior of a bystander animal (Langford et al., 2006; Li et al., 2018; Pitcher et al., 2017). Moreover, the effect of a given social encounter may vary, depending on the pain modality (Pitcher et al., 2017). Viewed collectively, these findings suggest that the sensory channel(s) used to assess the health status of conspecifics could depend on particular pain behaviors.

Given olfactory cues present in bedding from CFA-treated animals were sufficient to induce tactile hypersensitivity, we used GC-MS analysis to semi-quantitatively analyze volatile compounds in bedding taken from cages housing the CFA- and SAL-treated animals, before and after treatment. Bedding content reflects the bedding material itself, the housing environment, and compounds released by the animals housed in that cage, and the balance between accumulation (presumably from the animals) and evaporation/degradation over time. That is, a net increase in abundance of a volatile compound indicates that more molecules of this compound were released into the bedding than evaporated, and a net decrease in abundance implies that any release of the compound did not make up for losses from evaporation. CFA and SAL treatment led to differential shifts in abundance of volatile compounds in bedding. While 90% of analyzed compounds were increased in abundance in the bedding after CFA treatment, more than 60% of the same compounds were decreased (some significantly) after SAL treatment.

Six volatile compounds were significantly decreased in abundance after injection of SAL but moderately increased after injection of CFA. Several of these compounds are known to be elevated in urine from animals in models of stress, aging and disease. For example, 2-heptanone was elevated in urine from stressed mice and rats (Gutierrez-Garcia et al., 2006; Schaefer et al., 2010), and exposure to 2-heptanone increased immobility in the forced-swim test (Gutierrez-Garcia et al., 2007), and increased activity in amygdala (Contreras et al., 2012). Interestingly, 4-methyl-6-hepten-3-one, significantly reduced post-injection in bedding from SAL-treated animals but elevated in bedding from CFA-treated animals in our study, has been found to be elevated in urine of aged mice and mice treated with bacterial endotoxin, but decreased in the urine of tumor-bearing mice (Hanai et al., 2012), which warrants further and more specific investigation of this compound as a social signal. It thus seems reasonable to conclude that compounds that were decreased in bedding from SAL-treated but not CFA-treated groups could provide a signal to conspecifics, and that activation of olfactory pathways by these compounds, individually or collectively, can contribute to social influences on pain-related behaviors. Though there has not been much research into rodent-sourced volatile compounds modulating pain behaviors, rodents release volatile alarm or appeasing pheromones (Kiyokawa, 2015; Junien et al., 2013; Morozov and Ito, 2018; Kiyokawa et al., 2023). We did not observe any of these known pheromones among the detected volatiles. This observation reciprocates a recent finding that social modulations of pain and fear diverge into different central mechanisms after receiving the respective social signals during an interaction (Smith et al., 2021).

Eight volatile compounds were more abundant in clean than soiled bedding. Most of these compounds are common in industrial manufacturing processes, and hence reasonable to exist in bedding materials. The fact that they decreased in the presence of the animals suggests that mice or mice excreta lead to degradation or increased evaporation of these molecules. However, since there was no difference between the CFA- and SAL-treated groups in the bedding content of these compounds, it is unlikely that the decrease over time contributed to the hyperalgesia seen in the CFA-treated group. It is nonetheless important to acknowledge that environmental factors could modulate nociception, or serve as a background or foundation for socially derived signals.

There were technical limitations in the present study that could be improved in future investigations. While many compounds were decreased in abundance in bedding from the SAL-treated animals compared to CFA-treated animals, relatively few compounds showed statistically significant changes in abundance in either group compared to a pre-injection baseline. This is likely due to the small sample size, a limitation of the manual sampling during GC-MS analysis. Future studies could increase SPME temperature to yield a larger VOC extraction for GC-MS analysis. However, such analysis may also present a somewhat different VOC profile from the olfactory signal that mice received at room temperature. Different types of SPME fibers to extract samples could also be considered. Furthermore, it must still be acknowledged that machine detection may not be as sensitive as the rodent olfactory system. For example, there are reports that various animal species can recognize human diseases via odorant signals (Williams and Pembroke, 1989; Church and Williams, 2001; Guest et al., 2021; Piqueret et al., 2022). During the COVID19 pandemic, it has been noted that canines are able to identify COVID-positive patients with above 80% accuracy after training, while researchers were not able to identify volatile chemical markers for COVID using GC-MS (Mendel et al., 2021). It would not be surprising if GC-MS analysis was unable to identify all relevant compounds detected by mice during a social interaction or bedding exposure.

It is also worth acknowledging that we did not confirm potential nociceptive effects of the GC-MS-identified compounds in behavioral tests. However, behavioral effects of olfactory cues are frequently contextual (Kaur et al., 2014; Bourne et al., 2013; Ryan et al., 2008), in which case the potential target compounds would only exert a pro-nociceptive effect in the specific social situations. The trigger for the observed bedding-induced pro-nociception is thus likely more complex than simple changes in abundance of a few compounds. A complex experimental design that takes the contextual nature of olfactory social cues into account would be needed to further narrow down or confirm the target compounds without false negative or false positive identification.

Social modulation of pain in humans is complex and nuanced (e.g., Saarijärv et al., 1990; Leonard and Cano, 2006; Borsook and MacDonald, 2010). A recent study in humans reported that a volatile compound emitted from human body could affect human aggression in a sex-dependent manner (Mishor et al., 2021). It seems plausible, therefore, that biological mechanisms, including those influenced by olfactory cues, could interact with other varied social factors to influence human pain behavior.

Conclusion

In current study, we demonstrated that exposure to mice with CFA-induced inflammatory pain or their soiled bedding enhanced nociception in naïve same-sex conspecifics, and that this pro-nociceptive effect was triggered by an olfactory sensory cue. We also identified volatile organic compounds that were increased in relative abundance in bedding of CFA-treated animals compared to controls. These compounds, combined or individually, could contribute to social modulation of pain via olfactory sensory activity.

Supplementary Material

Fig S2

Supplemental Figure 2. Time course of von Frey thresholds of both hindpaws in male and female mice in BS-SAL (A, B), BS-CFA (C, D), and OLF-CFA (E, F) groups. 2-way ANOVA repeated measures reveals that there is no main effect on paws in all groups. Female sample size: BS-SAL, N = 6; BS-CFA, N = 6; OLF-CFA, N = 8. Male sample size: BS-SAL = 9; BS-CFA, N = 9; OLF-CFA, N = 9.

Significance:

Social context can influence nociceptive sensitivity. Recent studies suggested involvement of olfaction in this influence. In agreement with this idea, the present study shows that the presence of mice with inflammatory pain produces nociceptive hypersensitivity in nearby conspecifics. This enhanced nociception occurs via olfactory cues present in the mouse bedding. Analysis of the bedding from mice with inflammatory pain identifies a number of compounds indicative of disease states. These findings demonstrate the importance of olfactory system in influencing pain states.