Gentle stroking stimuli induce affiliative responsiveness to humans in male rats

Shota Okabe; Yuki Takayanagi; Masahide Yoshida; Tatsushi Onaka

doi:10.1038/s41598-020-66078-7

. 2020 Jun 4;10:9135. doi: 10.1038/s41598-020-66078-7

Gentle stroking stimuli induce affiliative responsiveness to humans in male rats

Shota Okabe ¹, Yuki Takayanagi ¹, Masahide Yoshida ¹, Tatsushi Onaka ^1,^✉

PMCID: PMC7272613 PMID: 32499488

Abstract

Gentle tactile stimuli have been shown to play an important role in the establishment and maintenance of affiliative social interactions. Oxytocin has also been shown to have similar actions. We investigated the effects of gentle stroking on affiliative relationships between humans and rats and the effects of gentle stroking on activation of oxytocin neurons. Male rats received 5-min stroking stimuli from an experimenter every other day for 4 weeks between 3 and 6 weeks of age (S3–6 group), for 4 weeks between 7 and 10 weeks of age (S7–10 group), or for 8 weeks between 3 and 10 weeks of age (S3–10 group). Control rats did not receive stroking stimuli. Rats in the S7–10 and S3–10 groups emitted 50-kHz calls, an index of positive emotion, more frequently during stroking stimuli. Rats in the S3–6, S7–10, and S3–10 groups showed affiliative behaviors toward the experimenter. Oxytocin neurons in the hypothalamic paraventricular nucleus of rats in the S3–6, S7–10, and S3–10 groups were activated following stroking stimuli. These findings revealed that post-weaning repeated stroking stimuli induce an affiliative relationship between rats and humans and activation of oxytocin neurons.

Subject terms: Emotion, Social behaviour

Introduction

Social mammals develop affiliative relationships with intraspecific individuals. Affiliative relationships between mothers and infants^1,2 and between pairs of males and females in monogamous species^3,4 have been intensively investigated. Social animals sometimes establish affiliative relationships not only with intraspecific animals but also with interspecific animals. Sheep⁵, lambs⁶, and dogs⁷ have been shown to be able to establish affiliative relationships with humans. Other animals such as rats⁸, foxes⁹, dolphins¹⁰, cattle¹¹, and horses¹² have also been reported to show positive reactions toward humans. Affiliative interactions with animals have been suggested to reduce anxiety and to have positive effects on the health conditions of humans^13,14. In addition, human-animal relationships improve the welfare of animals^15,16. However, little is known about the neural mechanisms underlying the development and maintenance of human-animal affiliative relationships.

Rats emit several types of ultrasonic vocalizations (USVs). In aversive situations, such as exposure to predators, rats emit USVs with a frequency range of 18–32 kHz, referred to as “22-kHz calls”^17–21. On the other hand, in appetitive situations, such as social play, rats elicit vocalizations with higher frequency (more than 35 kHz), referred to as “50-kHz calls”^20–22. Thus, 22-kHz and 50-kHz calls are generally considered to be indicators of negative and positive emotional states, respectively²⁰. Fifty-kHz calls have been further divided into frequency-modulated calls (FM calls) and flat calls (Non-FM calls). FM calls are further divided into several subtypes, including complex, upward, step up, and trill calls^20,23. FM 50-kHz calls have been shown to be emitted during intensively affective situations such as social play and mating behavior²⁴. In contrast, non-FM 50-kHz calls have been shown to be emitted during non-social situations such as feeding behavior²⁵. These behavioral characteristics are useful for clarifying affiliative relationships between humans and rats in experimental situations. Thus, we used rats as good model animals for investigating heterogeneous affiliative relationships.

Tactile stimulation has been suggested to play a fundamental role in the establishment and maintenance of intimate relationships between individuals. Positive physical contacts such as hugging and massage by partners have been shown to reduce cortisol release and heart rate increase in response to stressful stimuli in humans²⁶. In primates, social grooming has been suggested to facilitate affiliative bonds between individuals^27,28. Not only conspecific grooming but also interspecific gentle stroking has been suggested to induce anxiolytic actions in behavioral and neuroendocrine systems in lambs⁶. We have also shown that massage-like gentle stroking stimuli induce 50-kHz calls in rats²⁹. However, it was not known whether stroking stimuli induce an affiliative relationship between rats and humans. The amount and time period of stroking stimuli required to induce an affiliative response were also unknown.

Oxytocin is mainly synthesized in the hypothalamic paraventricular nucleus (PVN), supraoptic nucleus (SON), and bed nucleus of the stria terminalis (BNST). Oxytocin has been shown to facilitate affiliative social behaviors^30–33 and to induce anti-stress or anxiolytic actions^34,35. Several studies revealed that massage or non-noxious tactile stimulation increases oxytocin release in rats^29,36, dogs³⁷, and humans³⁸. Massage-like hand movement of a baby induces an increase in plasma oxytocin concentrations in the mother³⁹. We have also demonstrated that stroking stimuli activate oxytocin neurons, particularly in the caudal PVN²⁹. From these findings, we considered that stroking stimuli activate oxytocin neurons and contribute to an affiliative relationship between humans and rats.

The present study aimed to clarify the effects of stroking stimuli on affiliative relationships between humans and rats and to determine the appropriate timing of stroking stimuli to create an affiliative relationship between humans and rats. For that purpose, male rats at the age of 3 weeks were assigned to four experimental groups with different periods of stroking, and then they received several behavioral tests for examination of anxiety-related behavior and affiliative behaviors toward an experimenter. Plasma testosterone concentrations, which affect social behaviors⁴⁰, start to increase steeply at postnatal 7 weeks of age and reach adult concentrations at 9–10 weeks of age⁴¹. Thus, in the present study, stroking stimuli were applied during the postnatal periods of 3–6, 3–10, and 7–10 weeks. Effects of stroking stimuli on activation of oxytocin neurons were also investigated by examining the expression of c-Fos protein, an index of neuronal activation, in the hypothalamus and BNST.

Materials and Methods

Subjects

Animal experiments were carried out after receiving approval from the Animal Experiment Committee of Jichi Medical University and were in accordance with the Institutional Regulations for Animal Experiments and Fundamental Guidelines for Proper Conduct of Animal Experiments and Related Activities in Academic Research Institutions under the jurisdiction of the Ministry of Education, Culture, Sports, Science and Technology.

Forty-four male rats of the Lewis strain were used in the present study. These animals were produced by mating of pairs of sires and dams obtained from an animal supplier (LEW/ CrlCrlj, Charles River Laboratories Japan, Inc., Kanagawa, Japan) and were weaned at the age of three weeks. After weaning, male rats were housed in pairs and used in the present study. The animals were maintained under a 12: 12 h light/dark cycle (lights on at 7:30 am) at 22 ± 2 °C and 55 ± 15% relative humidity. Food and water were available ad libitum.

Groups and general experimental design

Stroking stimuli were applied by an experimenter (S.O.) with the hand of the experimenter over the back of each animal at a speed of 5–10 cm/sec for 5 min every other day, as described previously²⁹. Rats were assigned to four groups: S3–6 group, S7–10 group, S3–10 group, and N3–10 group. Rats in the S3–6 group received 5-min stroking stimuli from the experimenter for 4 weeks between 3 and 6 weeks of age (n = 12). Animals in the S7–10 group received stroking stimuli for 4 weeks between 7 and 10 weeks of age (n = 10). Animals in the S3–10 group received stroking stimuli for 8 weeks between 3 and 10 weeks of age (n = 12). Control rats in the N3–10 group did not receive stroking stimuli (n = 10). At 10 weeks of age, all of the rats were housed individually and then behavioral tests were conducted. Time schedules of treatments and behavioral tests are shown in Fig. 1.

Ultrasonic vocalizations

General protocols for recording

Ultrasonic vocalizations (USVs) were recorded by use of a microphone (Type 4158 N, Aco, Tokyo, Japan) designed for measurements of sound pressure levels of 20–100,000 Hz sounds. The microphone was placed at a distance of approximately 20 cm from a rat and was connected to a sensor amplifier (SR-2200, Ono Sokki, Kanagawa, Japan). Acoustic data were recorded at a sampling rate of 200 kHz and with 16 bits by an Avisoft RECORDER (Version 4.2, Avisoft Bioacoustics, Berlin, Germany) and analyzed by a software program of Avisoft SASLab Pro (Version 5.2, Avisoft Bioacoustics). Spectrograms were generated with a fast Fourier transform length of 256 points and an overlap of 75%.

Recording of USVs

In the S3–6, S7–10, and S3–10 groups, USVs were recorded during a 5-min stroking session once each week during the developmental period (Fig. 1). Each rat on the experimenter’s lap was given massage-like stroking stimuli from the hand of the experimenter (S.O.). The numbers of 50-kHz calls (frequencies between 35 and 100 kHz) and 22-kHz calls (frequencies between 18 or more and less than 32 kHz) were manually counted.

During the test period, USVs in the basal condition when rats were kept in their individual home cages were recorded for 5 min more than 25 min after setting microphones above mesh lids of their home cages. USVs in a holding condition were recorded during a 5-min period when each rat was placed on the experimenter (S.O.)’s lap. USVs in a stroking condition were then recorded during the next 5-min stroking period. The numbers of 50-kHz calls and 22-kHz calls were manually counted. Fifty-kHz calls were divided into two types: FM and Non-FM calls. FM calls were further divided into 10 types: complex, upward, downward, split, step up, step down, multi-step, trill, inverted-U, and composite²³. Non-FM calls were divided into two types: flat and short²³. Fifty-kHz calls were manually classified into 12 types according to the shape of a spectrogram plot of each syllable. The numbers of these types of USVs were counted in basal, holding, and stroking conditions.

Behavioral tests

Open field test

Anxiety-related behavior was analyzed by an open field test. Each animal was placed in an open field (W60 × D60 × H40 cm) and kept there for 10 min. Total movement distance was recorded as an index of locomotion, and total time spent staying in the center area, which was 36% of the total open field arena, and total time spent immobile were automatically measured by using a behavior analyzing program, Time OFCR1 (O’Hara & CO., LTD, Tokyo, Japan).

Social preference test

Preference to the experimenter (S.O.)’s hand compared with a novel object was determined by a social preference test. The experimenter (S.O.)’s hand was placed at one corner of a test box (W60 × D60 × H40 cm), and a novel plastic bottle (W6× D6 × H14 cm) was placed at the diagonally opposite corner of the test box. Each animal was introduced into the test box at a corner other than the corners where the hand and the bottle were placed. The time spent staying in a square (20 × 20 cm) where the hand or the bottle was placed was measured automatically during a 10-min test period by using an analyzing program (Time OFCR1). Preference score was calculated as [(time spent staying in the hand area) – (time spent staying in the bottle area)], and preference scores were compared among the four groups.

Following behavior test

Proactive approach behavior toward the experimenter (S.O.)’s hand (“following” behavior) was examined in a test box (W60 × D60 × H40 cm). The experimenter (S.O.)’s hand was placed at one corner of the test box. Each animal was placed at the corner diagonally opposite to the hand. The hand was moved from one corner to the next corner along the wall over a period of 7 seconds and then kept at the corner for 3 seconds. This movement was repeated for 5 min. The time spent following the experimenter (S.O.)’s hand within a distance of approximately 3 cm was measured manually. In addition, the number of 50-kHz calls during the following behavior was recorded and counted. The experiment and analysis were conducted by experimenters (S.O. and Y.T.) blind to treatments of the animals.

Conditional place preference (CPP) test

The reward value of stroking stimuli was analyzed by a CPP test. The test apparatus consisted of a black box (W40 × D30 × H30 cm), a white box (W40 × D30 × H30 cm) and a boundary wall (30 × 30 cm) with a door hole (8 × 8 cm) through which rats could enter either box. Both boxes had ceilings in which video cameras were installed. The black box had black walls with a meshed floor, and its luminance was 10 lux. The white box had white walls with a grid floor, the grids being arranged in parallel with the boundary wall between the black and white boxes, and there was no illumination. On day 1, each rat was placed in the white box and habituated to the CPP test box for 10 min with the boundary door open so that the rat could enter either box. On day 2, each rat was placed in the black box with the door open for 10 min and then the time spent staying in each box was measured automatically by using software for behavior analysis (ImageJ LD1, O’Hara and Co.), which was produced on the basis of ImageJ program (National Institutes of Health, U.S.A.), to identify initial preference (pre-conditioning). The box in which each rat stayed for a longer time was defined as the initially preferred box (I.P. box) and the other box was defined as the initially non-preferred box (I.N.P. box). Locomotion distance in each box was also measured. Conditioning was performed from day 3 to day 8 with the door connecting the two boxes being closed. On day 3, rats were introduced into the I.P. box and kept there for 5 min. On day 4, rats were placed in the I.N.P. box and given stroking stimuli by the experimenter (S.O.) for 5 min. The set of 2-day conditioning procedures was repeated another 2 times from day 5 until day 8. On day 9, rats were placed in the I.P. box with the door open and the time spent in each box and distance of movements in each box were measured by using an analyzing program. In addition, [(time spent staying in the I.N.P. box during the post-conditioning)/(time spent staying in the I.N.P. box during the pre-conditioning)], [(locomotion distance in the I.N.P. box during the post-conditioning)/(locomotion distance in the I.N.P. box during the pre-conditioning)], and [(locomotion distance in the I.P. box during the post-conditioning)/(locomotion distance in the I.P. box during the pre-conditioning)] were used as preference ratios and were compared among the four groups.

Immunohistochemical detection of c-Fos protein in oxytocin-immunoreactive (-ir) neurons

The rats in each group were assigned to one of two conditions: a stroking condition and a non-touch control condition. In the stroking condition, rats received stroking stimuli for 5 min and then they were placed back in their individual home cages. In the non-touch condition, rats were kept in their individual home cages. Ninety minutes after termination of stroking stimuli, the animals were anesthetized with Avertin (tribromoethanol; 200 mg/kg, i.p.) and perfused transcardially with heparinized saline (20 U/mL) followed by 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4) for 15 min. The brains were immediately removed from the skulls, post-fixed in 4% paraformaldehyde overnight, placed in 30% sucrose in 0.1 M phosphate buffer until they sank, and frozen in dry ice. The hypothalamic part of each frozen brain was sectioned coronally at 30 µm and processed for immunochemical detection of c-Fos protein and oxytocin as described previously²⁹. Sections containing the dorsal zone of the medial parvicellular part of the caudal PVN (3 sections, from 1.92 mm to 2.16 mm posterior to the bregma), rostral PVN (6 sections, from 1.2 mm to 1.80 mm posterior to the bregma), supraoptic nuclei (SON, 7 sections, from 0. 6 mm to 1.32 mm posterior to the bregma) or BNST (4 sections, from 0.72 mm to 1.08 mm posterior to the bregma) were examined at an interval of 120 µm in each rat. The sum of the numbers of c-Fos-ir neurons, oxytocin-ir neurons, and double-positive neurons were counted in each brain region, the boundary of which was determined according to a brain atlas⁴². The percentage of oxytocin-ir neurons expressing c-Fos protein was determined.

Statistical analysis

Statistical analyses were performed using JMP 13.0.0 software (SAS Institute, NC, USA), Prism 6 (GraphPad Software Inc., CA, USA), R and SPSS Statistics 22 (IBM, IL, USA). The numbers of USVs at 3, 4, 5, 6, 7, 8, 9, and 10 weeks in the three stroking groups were analyzed separately for within-group comparisons by using Friedman’s test followed by the post-hoc Dunn’s test (Fig. 2C–H). Data for the open field test (Fig. 3A), preference score of the social preference test (Fig. 3B right), following behavior test (Fig. 3C), and number of USVs during the following behavior test (Fig. 3C) were analyzed by the Kruskal-Wallis test followed by the post-hoc Steel-Dwass test. Times spent in the bottle and hand areas in the social preference test (Fig. 3B left) were analyzed by repeated measures two-way analysis of variance (ANOVA) (group × area) followed by post-hoc Holm’s test. Data for the numbers of 50-kHz and 22-kHz calls in basal, holding, and stroking conditions (Fig. 4) were also analyzed by repeated measures two-way ANOVA (group × condition) followed by post-hoc Holm’s test. The numbers of vocalizations in FM and non-FM and those of each syllable (Fig. 5 and supplementary Fig. 1) were analyzed by repeated measures two-way ANOVA (syllable pattern × condition) followed by post-hoc Holm’s test.

Ultrasonic vocalizations (USVs) during stroking stimuli. (**A,B**) Examples of spectrograms of 50-kHz calls (A) and 22-kHz calls (B). Time courses of the numbers of vocalizations during stroking in the S3–6 group (C, 50-kHz; D, 22-kHz), S7–10 group (E, 50-kHz; F, 22-kHz), and S3–10 group (G, 50-kHz; H, 22-kHz). In the S3–6 group, the number of 50-kHz calls did not significantly change. In the S7–10 group, the number of 50-kHz calls at the age of 9 weeks was significantly larger than that at the age of 7 weeks. In the S3–10 group, the numbers of 50-kHz calls at 7, 8, 9, and 10 weeks of age were significantly increased compared to the number at the age of 5 weeks. In addition, the number of 50-kHz calls at 8 weeks of age was significantly larger than that at 4 weeks of age. The numbers of 22-kHz calls during stroking did not significantly change in any of the three groups. ^†P < 0.01; *P < 0.05, post-hoc Dunn’s test.

Results of the open field test, social preference test, and following behavior test. (A) Total distance of movements (left panel), total time spent staying in the center area (middle panel), and total time spent immobile (right panel) in an open field test. There were no significant differences in the total center time and immobile time among the groups. (B) Difference between durations of staying in the hand area and the bottle area in a social preference test. The time spent staying in the hand area was significantly longer than that in the bottle area in the S3–6, S7–10 and S3–10 groups. Preference scores [(time spent staying in the hand area) – (time spent staying in the bottle area)] in the S3–6 and S3–10 groups were significantly higher than that in N3–10 group. (C) Durations of following behavior and number of 50-kHz calls during a following behavior test. Durations of following behavior (left panel) were significantly longer in the S3–6, S7–10, and S3–10 groups than in the N3–10 group. The numbers of 50-kHz calls (right panel) were significantly larger in the S3–6 and S3–10 groups than in the N3–10 group. ^†P < 0.01; *P < 0.05, a versus c, P < 0.01, b versus c, P < 0.05. Post-hoc Steel-Dwass test (A, right graph of B, and C) and repeated measures two-way ANOVA (group × area) followed by post-hoc Holm’s test (left graph of B) were performed. Error bars denote standard error of the mean.

Numbers of 50-kHz (A) and 22-kHz (B) calls. The number of 50-kHz calls in the stroking condition was significantly larger than that in the basal or holding condition. The number of 22-kHz calls did not significantly change. ^†P < 0.01; *P < 0.05. Repeated measures two-way ANOVA (group × condition) followed by post-hoc Holm’s test was performed. Error bars denote standard error of the mean.

Numbers of FM and non-FM calls in the N3–10 (A), S3–6 (B), S7–10 (C), and S3–10 (D) groups. In the S3–6 group, the number of calls in the stroking condition was significantly larger than that in the basal or holding condition. In the S7–10 group, the number of calls in the stroking condition was significantly larger than that in the basal condition. In the S3–10 group, the number of calls in the stroking condition was significantly larger than that in the basal or holding condition. In addition, the number of calls in the holding condition was significantly larger than that in the basal condition. ^†P < 0.01, *P < 0.05. A comparison among conditions was conducted by repeated measures two-way ANOVA (syllable pattern × condition) followed by post-hoc Holm’s test. Error bars denote standard error of the mean.

Data for CPP (Fig. 6) were analyzed by repeated measures two-way ANOVA (group × timing) followed by post-hoc Holm’s test or the Kruskal-Wallis test followed by the post-hoc Steel-Dwass test. Results of immunohistochemical experiments (Fig. 7 and supplementary Fig. 2) were analyzed by two-way factorial ANOVA (group × stimuli) followed by post-hoc Holm’s test. In post hoc multiple comparisons, every possible comparison was conducted and P values were adjusted by Holm’s method. Correlations between percentages of oxytocin-ir neurons expressing c-Fos protein in the caudal PVN and behavioral data (results of the open field test, social preference test and following behavior test, results for 50-kHz calls during the stroking condition, and results of the CPP test) and correlations between the number of 50-kHz calls during the stroking condition and other behavioral data were analyzed by Spearman’s correlation test. Data of all 4 groups were collapsed into a single analysis for correlation analysis. P < 0.05 was considered statistically significant.

Results

Developmental changes of ultrasonic vocalizations

In the developmental period, the number of 50-kHz calls rats emitted during stroking stimuli was small and did not significantly increase between the ages of 3 weeks and 6 weeks (Fig. 2C) in the S3–6 group. In the S7–10 group, the number of 50-kHz calls during stroking stimuli increased and the number of 50-kHz calls at the age of 9 weeks was significantly increased compared to that at the age of 7 weeks (Fig. 2E, P = 0.0289, Friedman’s test, P = 0.0437, post-hoc Dunn’s test). In the S3–10 group, the number of 50-kHz calls gradually increased and the numbers of calls at 7, 8, 9, and 10 weeks of age were significantly increased compared to the number of calls at 5 weeks of age (Fig. 2G, P < 0.0001, Friedman’s test, 5 weeks versus 7 weeks, P = 0.0018; 5 weeks versus 8 weeks, P = 0.0013; 5 weeks versus 9 weeks, P = 0.0036; 5 weeks versus 10 weeks, P = 0.0021, post-hoc Dunn’s test). In addition, the number of 50-kHz calls at 8 weeks of age was significantly larger than that at 4 weeks of age (Fig. 2G, P < 0.0001, Friedman’s test, P = 0.0434, post-hoc Dunn’s test).

The number of 22-kHz calls during stroking did not significantly change in any of the three groups (Fig. 2D,F,H).