The materials and methods have been previously reported in Tivey et al. [20].
All animal work was carried out in accordance with the UK Animals (Scientific Procedures) Act 1986 following ethical approval by the Roslin Institute’s Animal Welfare and Ethical Review Body (AWERB study number B026) and carried out in the Roslin Institute’s Bioscience and Veterinary Services Facility. The authors complied with the ARRIVE guidelines.
Subjects
Across two replicates, 64 juvenile Wistar rats (mean weights ± SEM: Replicate 1: Females 55.12 ± 1.92g, Males 60.1 ± 2.14g; Replicate 2: Females 46.81 ± 1.91 g, Males 52.73 ± 1.61g) were sampled (32 per replicate; Charles River, Kent, UK). Each replicate was split evenly between males and females. Rats arrived at the Roslin Institute Bioscience and Veterinary Services facility at 23-24 days of age. The rats were derived from four different litters: four female and four male rats from the same litter were used (four litters in total). Treatment (Control or Tickled) was randomly assigned to each rat, balancing for body weight and littermates so that average weights for each treatment group and for each sex were not different. There was an equal number of animals from each litter in both treatment groups.[20]
Rats were housed in same-sex pairs, with each cage containing a tickled and control rat (tickled rats were marked with a black pen mark on their tail, control rats had no mark, treatment was pseudo-randomly assigned balancing for body weight [60] and litter; no littermates were housed in the same cage. Standard clear plastic cages (46 x 25 x 21 cm) with a wire lid were used. Each cage contained aspen chip bedding, one shredded paper nest and one aspen chew stick (Nepco, Warrensburgh, USA). Food (14 % protein rodent maintenance diet, Envigo, UK) and water were available ad libitum and the room temperature, humidity and light intensity was held stable at 18 - 23 oC, 40 - 60 % and 25 lux respectively. The cages were pseudorandomly arranged in a cage rack to account for differing lux levels through the height of the rack and balanced for sex and litter. The rats were held on a reversed 12-hour light: 12-hour light dark cycle (lights on: 00:00, lights off: 12:00) and were tested in the tickling test arena during the dark phase. Body weight (g) was recorded daily following testing between 16:00 - 18:00 in the dark phase. The rats were checked daily (by laboratory personnel at 08:00, during the light phase) and nitrile gloves were worn when handling the animals. To minimise handling stress rats were picked up gently by holding them behind their forelegs and then cupping them with both hands. Following arrival at the animal facility, rat body weight (g) was taken daily and rats were acclimatised to their new surroundings for five days before they were habituated to the tickling test arena (a Perspex open box,60 (length) x 60 (width) x 25 (height) cm that was lined with LabMat; LabLogic Systems Ltd., England). [20]
Experimental Design
Rat sample size was determined using a power equation using variance and mean values from previous data ([18]; calculated sample size: 16 per group). The order in which the cages and cage mates were tested each day was pseudo-randomised to account for time of day, sex, treatment and lux levels of the cages in the cage rack.[20]
All testing (weighing, habituation and tickling) was carried out in the home cage room. The area used for all testing was enclosed by a thick, plastic curtain, at the opposite end of the room to the cage rack. Testing was carried out in the dark phase under red light; only one experimenter (female) carried out the testing and was the only person present during testing. The rats were brought to a bench in the enclosed area used for testing in their home cage. One rat was tested while the second cage mate remained in the home cage. The home cage was placed on the other side of the curtain from the testing apparatus so that no USVs were detected by the rat in the home cage (the curtain blocked USV transmission, measured using the ultrasonic microphone). Animals were given a five-day habituation period during which they were placed alone in the centre of the testing arena for a total of five min per day. An immobile right hand (wearing a nitrile glove covered with a white cotton glove) was placed in the arena to habituate the rats to the glove and the researcher. [20]
After the habituation phase, on day six, the rats began ten days of behavioural testing. Each animal was tested in the arena for two min. per day for ten days. One cage mate (randomised order) was taken from the home cage and placed in the test arena. The rat was tested for two min., timed with a stopwatch and both video and sound recordings were recorded digitally using a video camera placed above the testing arena (Panasonic HD HC-V10) and an ultrasound microphone suspended about 30 cm above the testing arena (Pettersson M500-384 USB Ultrasound microphone, Pettersson Electronik, Sweden). The testing arena was cleaned with 70% ethanol gel and allowed to dry between the testing of each rat and testing was carried out in the first three hours of the dark phase. [20]
For the tickled group, rats were placed in the testing arena, and a hand (wearing a white cotton glove) was placed motionless on one wall of the testing arena (the wall position and placement of the hand was randomised each time) for the first 15 sec. of testing. Following these 15 sec. of release (i.e., where the experimenter’s hand was motionless on the side of the arena and the rat received no contact with the hand), the rat was tickled for 15 sec. by making rapid finger movements on the nape of the neck. If the rat turned its body around to rear up at the hand or rolled onto its back during tickling, it was also tickled on its ventral side; the rats were not manually flipped and pinned as described elsewhere (e.g. [1]). Although we will retain the terms tickling throughout the manuscript, this approach to tickling has been described as ‘playful handling’ in Bombail et al. [2]. The 15 sec. bouts of tickling and release were alternated during the two min. of testing. For the control group, rats were placed in the arena for two min., with a hand (wearing a white cotton glove) resting motionless on one wall of the testing arena (the wall position and placement of the hand was randomised each time). If the rat reared at the hand, the hand was gently moved away to one side and then replaced back in position. Following testing, the rat was gently picked up (as described previously) and body weight was measured before it was returned to its home cage. The other cage mate was then removed (following cleaning of the testing arena described above) and placed in the testing arena and the behavioural testing was repeated. The cotton gloves were only worn by the hand in the testing arena during testing and each rat was exposed to its own individual cotton glove to avoid any potential olfactory bias. [20]
On the final day of testing (day 10) rats were removed from their home cage between 60 - 90 min. after they were tested in the testing arena. For those 60 - 90 min. between the final testing being carried out and being euthanised, the rats were returned in their home cage to the cage rack and remained in the home room. Rats were then weighed and euthanized with an intraperitoneal injection of sodium pentobarbitone (Pentoject; 20% w/v 200 mg/ml) carried out by experienced laboratory personnel. The order in which the rats were euthanised was pseudo-randomised and balanced for treatment group and sex. Following confirmation of death, the rats were decapitated and the brains were removed from the skull and prepared for paraformaldehyde (PFA) fixation. [20]
Measurements
Ultrasonic Vocalisation Analysis (USVs)
An ultrasound microphone (M500-384, Pettersson Elektronik, Sweden) and Audacity software (Version 2.1.3, Pennsylvania, United States of America) were used to record the spectrograms of USVs. To establish changes in calling frequency over time, the total number of 50 kHz USVs and 22 kHz USVs emitted during the two min. testing on the first, middle and final day (days 1, 5 and 10) were manually counted from spectrograms produced in Ultravox XT 3.2 (Noldus Information Technology, Wageningen, Netherlands). Call events were visually categorised using call parameters for 50 kHz USVs and 22 kHz USVs (see Table 1 for definitions). [20]
Testing Arena Behaviour
Video camera (Panasonic HD HC-V10) footage was obtained during the two min. testing phase on each day of testing. In order to analyse the effect of tickling on scampering and hand approaches, behaviours were analysed on days 1, 5 and 10 of testing. Using video footage, the total number of scampers and approaches made towards the hand were manually counted on days 1, 5 and 10 (see Table 1 for definitions). The experimenter who was blinded to sex, but not to treatment, due to being able to see whether individual rats were tickled or not in the digital recording. [20]
Immunohistochemical procedures
Following dissection, brains were immediately immersed in 4% PFA in 1 x phosphate buffered saline (PBS) at 4 oC for five days. The brains were then transferred into 15 % sucrose in 4% PFA for a further 24 - 48 h. at 4 oC. Once sunk, brains were transferred into 30 % sucrose in 1 x PBS for 24 - 48 h. incubation at 4 oC or until the brain had sunk to the bottom.
The brains were dissected into three parts: 1) the main brain (including cerebral cortex, diencephalon and mesencephalon), 2) the olfactory bulb and 3) the cerebellum, brainstem and spinal cord. Samples were snap frozen on dry ice, wrapped in foil and stored in a labelled small plastic bag at -70 oC until they were sectioned using a freezing microtome. The main brain was sectioned coronally into 1 x PBS (alternate sections taken for oxytocin and vasopressin double labelling) in a petri dish at 50 µm on a freezing microtome. Following sectioning, the brain sections were transferred into cryoprotectant and stored at -20 oC.
Fos and oxytocin or arginine vasopressin double immunohistochemistry
Double labelling was performed using the free-floating sections. To visualise active neural populations, rabbit anti-Fos antibody (K-25 Santa Cruz sc253) was used to stain Fos expressing neurons and rabbit anti-oxytocin polyclonal IgG (Immunostar; 20068) was used to stain oxytocinergic cell bodies. Rabbit anti-vasopressin polyclonal IgG (Chemicon; AB1565) was used to stain oxytocinergic cell bodies. On a shaking platform set to 95 rpm, sections were washed four times 15 min. in 1x PBS with 0.2% triton (PBS-T), rinsed for 5 min. in 1x PBS, endogenous peroxidases were quenched in 0.3% hydrogen peroxide for 20 min., rinsed for 5 min. in 1x PBS, incubated with 0.3% hydrogen peroxide for a further 20 min., washed three times 10 min. in 0.2% PBS-T, immersed in 0.2% PBS-T and 5% normal goat serum (NGS) for 60 min. to block non-specific binding, and incubated in rabbit anti-Fos antibody at 1:1000 diluted with 0.2% PBS-T and 3% NGS for 3.5 days at 4°C. [61]
Sections were then washed four times 10 min. in 0.2% PBS-T and then incubated in biotinylated anti-rabbit secondary antibody (Vectastain elite ABC rabbit kit; Vector; PK6101). Sections were incubated for 60 min. on a shaking platform at room temperature. Sections were then washed for 30 min. in 0.2% PBS-T and then incubated in Avidin / Biotinylated horseradish peroxidase conjugate (Vectastain elite ABC rabbit kit; Vector; PK6101) for 60 min. on a shaking platform at room temperature. Sections were then washed for 30 min. in 0.2% PBS-T, rinsed in 1 x PBS for 5 min., rinsed in 0.2M Sodium acetate buffer, antibody-antigen complexes were visualised using nickel II sulphate, ammonium chloride and 0.2M sodium acetate buffer solution mixed with 25 mg/ml DAB solution and 30% hydrogen peroxide. Sections were monitored under a light microscope the reaction halted by immersing sections in 0.2M sodium acetate buffer once positive staining was seen. Sections were rinsed for 25 min. in 1x PBS and endogenous peroxidases were quenched by incubating sections for 20 min. in 0.3% hydrogen peroxide. Sections were washed for 20 min. in 1x PBS-T before being incubated overnight at 4°C in rabbit anti-oxytocin polyclonal IgG at 1:10,000 diluted with 0.2% PBS-T and 5% NGS. The second set of sections for visualising vasopressin were incubated overnight at 4°C in rabbit anti-vasopressin polyclonal IgG at 1:4,000. [61]
Sections were then washed for 30 min. in 0.2% PBS-T and processed using biotinylated anti-rabbit secondary antibody and Avidin / Biotinylated horseradish peroxidase conjugate as previously described. Sections were washed for 30 min. in 0.2% PBS-T, rinsed in 1 x PBS for 5 min., oxytocin and vasopressin positive cells were visualised using DAB in 1x PBS and 30% hydrogen peroxide without nickel enhancement. Once positive staining was seen the reaction was halted by 25 min. washes in 1 x PBS before the sections were stored in 1 x PBS at 4°C until float-mounted onto gelatinised chrome-alum subbed glass microscope slides. [61]
Slides were allowed to dry before they were dehydrated and coverslipped. Sections were dehydrated through an increasing series of alcohol as follows: 70% ethanol 5 min., 95% ethanol 5 min., two times 99% ethanol 5 min., xylene/99% ethanl 5 min., three times Xylene 5 min. Slides were coverslipped using xylene-based mounting medium (Pirtex). [61]
Image acquisition and analysis
Brain regions of interest (ROI) were analysed in the following order: medial parvocellular PVN, posterior magnocellular PVN, dorsal parvocellular PVN, and SON. The ROI boundaries were defined using Paxinos and Watson’s stereotaxic rat brain atlas [62] and referring to Swanson and Kuypers [42]. A light microscope (Nikon E600 Brightfield Microscope) was used; the region was identified with a 4 x objective, an image was taken at the centre of the region at 10x magnification using Zen 2 software for image capture. This was done for each ROI (see Table 2) for both left and right hemispheres and across four representative sections per rat. [61]
Cell counting was performed blinded to sex, and treatment, by one observer using ImageJ 1.48v software (W. Rashband, National Institutes of Health, USA). Each image was opened in ImageJ and the Brightness and Contrast function was set to auto levels. The multi-point tool was used to manually count Fos-ir positive cells which were identified as having a dark grey/ black nucleus, and double-labelled cells which had a black nucleus and brown cell body. Any ambiguity in whether cells were positive or not (e.g. faint staining, or dark staining where it was not possible to distinguish between a black nucleus and brown cell body), led to the cells not being counted as positive. [61]
The total number of Fos-ir positive nuclei and total number of Fos-ir and oxytocin positive double-labelled cells were quantified separately. Counts were made in the posterior magnocellular, medial parvocellular and dorso-parvocellular subregions of the PVN [42,63] (Fig. 1 and 2) and the SON (Fig. 1 and 3). Total number of Fos-ir positive nuclei and total number of Fos-ir and vasopressin positive double-labelled cells were also quantified separately. Quantification was performed in the posterior magnocellular and medial parvocellular subregions of the PVN [42,63] and the SON. [61]
Statistical Analysis
Statistical analysis was carried out in R Studio and R (v 4.0.3, The R Foundation for Statistical Computing Platform (2020). Model adequacy was verified by examination of residuals [64] via the DHARMa package [65]. [61]
Linear models (using the glmmTMB package; [66]) were fitted to the 50 kHz USV, 22 kHz USVs, scampering and hand approach behaviour data. The models included sex, treatment and day as fixed effects, as well as the interaction between them, with effects reported through ANOVA comparisons [67] to compare the differences between group means rather than the linear relationships between variables. Post hoc analyses were carried out to investigate pairwise comparisons using the Tukey method [68], with statistical significance based on p < 0.05 threshold level. [61]
For the neural analyses, model adequacy was verified by examination of residuals [64] via the DHARMa package [65]. Generalised linear mixed models (GLMMs) in the glmmTMB package [66] were used to analyse the effect of treatment (tickled or non-tickled control), sex and the interaction of treatment and sex on Fos-ir immunoreactivity in oxytocinergic and vasopressinergic neural populations in the PVN (each region of the PVN was analysed separately). Dependent on model fitting and overdispersion, family links were set to either poisson or negative binomial distributed errors (‘nbinom2’ in the MASS package; [69]) with default transformations. All models included batch and cage as random effects and were nested (batch followed by cage). This was done to account for the variation from the non-independence of rats from the same cage and potential batch effects. For all models, both sex and treatment were included as fixed effects, as well as the interaction between them. A regression was carried out to test the linear relationships between Fos-ir and the fixed effects. Statistical significance was based on p < 0.05 threshold level and adjusted for multiple comparisons using the Tukey method [68]. All models included an offset of the total number of Fos-ir positive nuclei, to account for this uncontrolled variable in each sampling event, effectively calculating proportions of double-labelled cells/ total Fos-ir positive cells. [61]
All graphs were generated in R Studio using ggplot2 [70]. The data in the graphs are presented as the estimated marginal means ± 95% confidence intervals (CI). [61]