Subcaste differences in neural activation suggest a prosocial role for oxytocin in eusocial naked mole-rats

    Subcaste differences in neural activation suggest a prosocial role for oxytocin in eusocial naked mole-rats Georgia A. Hathaway, Mariela Faykoo-Martinez, Diana E. Peragine, Skyler J. Mooney, Melissa M. Holmes PII: DOI: Reference:

S0018-506X(15)30219-1 doi: 10.1016/j.yhbeh.2015.12.001 YHBEH 3996

To appear in:

Hormones and Behavior

Received date: Revised date: Accepted date:

10 August 2015 24 November 2015 19 December 2015

Please cite this article as: Hathaway, Georgia A., Faykoo-Martinez, Mariela, Peragine, Diana E., Mooney, Skyler J., Holmes, Melissa M., Subcaste differences in neural activation suggest a prosocial role for oxytocin in eusocial naked mole-rats, Hormones and Behavior (2015), doi: 10.1016/j.yhbeh.2015.12.001

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

SC

RI

PT

Subcaste differences in neural activation suggest a prosocial role for oxytocin in eusocial naked mole-rats

Departments of Cell & Systems Biology and Ecology & Evolutionary Biology, University of

TE

b

Department of Psychology, University of Toronto Mississauga, Mississauga ON L5L 1C6

D

a

MA

NU

Georgia A. Hathawaya, Mariela Faykoo-Martineza, Diana E. Peraginea, Skyler J. Mooneya, & Melissa M. Holmesa,b,*

AC CE P

Toronto, Toronto ON M5S 3G5

* Corresponding Author: Melissa Holmes Department of Psychology University of Toronto Mississauga 3359 Mississauga Road Mississauga, ON L5L 1C6 905-828-3956 [email protected]

1

ACCEPTED MANUSCRIPT Abstract The neuropeptide oxytocin (OT) influences prosocial behavior(s), aggression, and stress

PT

responsiveness, and these diverse effects are regulated in a species- and context-specific manner.

RI

The naked mole-rat (Heterocephalus glaber) is a unique species with which to study context-

SC

dependent effects of OT, exhibiting a strict social hierarchy with behavioral specialization within the subordinate caste: soldiers are aggressive and defend colonies against unfamiliar conspecifics

NU

while workers are prosocial and contribute to in-colony behaviors such as pup care. To determine if OT is involved in subcaste-specific behaviors, we compared behavioral responses

MA

between workers and soldiers of both sexes during a modified resident/intruder paradigm, and

D

quantified activation of OT neurons in the hypothalamic paraventricular nucleus (PVN) and

TE

supraoptic nucleus (SON) using the immediate-early-gene marker c-fos co-localized with OT neurons. Resident workers and soldiers were age-matched with unfamiliar worker stimulus

AC CE P

animals as intruders, and encounters were videotaped and scored for aggressive behaviors. Colony-matched controls were left in their home colony for the duration of the encounters. Brains were extracted and cell counts were conducted for OT immunoreactive (ir), c-fos-ir, and percentage of OT-c-fos double-labeled cells. Results indicate that resident workers were less aggressive but showed greater OT neural activity than soldiers. Furthermore, a linear model including social treatment, cortisol, and subcaste revealed that subcaste was the only significant predictor of OT-c-fos double-labeled cells in the PVN. These data suggest that in naked molerats, OT promotes prosocial behaviors rather than aggression and that even within subordinates, status exerts robust effects on brain and behavior. Keywords: oxytocin, eusocial, naked mole-rat, c-fos, social status, sub-caste, prosocial, aggression 2

ACCEPTED MANUSCRIPT Introduction A wealth of research has demonstrated the important role of neuropeptides in mammalian

PT

social behavior. One such neuropeptide, oxytocin (OT), plays a particularly prominent role in diverse social behavior both within and between species. OT promotes pair bonding and/or

RI

affiliation in a variety of mammals, including prairie voles (Insel and Hulihan, 1995; Williams et

SC

al., 1994), meadow voles (Beery and Zucker, 2010), tamarins (Snowdon et al., 2010), chimpanzees (Crockford et al., 2013) and humans (Kosfeld et al., 2005). However, in addition to

NU

its role in affiliation, OT has also been linked to social dominance and aggression, and is associated with inter-male aggression in rats (Ebner et al., 2000) and mate-guarding in male and female prairie voles (Bales and Carter, 2003; Winslow et al., 1993). Importantly, the effects of

MA

OT on behavior are context-dependent, as OT may promote affiliation in one context but aggression in another.

D

Across species, differing levels of sociality and social organization correspond to differences in

TE

OT levels and OT receptor (OTR) distribution. For example, monogamous prairie voles, which exhibit greater social attachments to both mates and young than polygamous meadow voles, have

AC CE P

greater OT receptor binding in the nucleus accumbens (NAcc) and the bed nucleus of the stria terminalis (BST) (Insel and Shapiro, 1992). Furthermore, in comparison with the solitary Cape mole-rat, eusocial naked mole-rats exhibit greater OT receptor binding in these same areas (Kalamatianos et al., 2010). In regards to context, for pairs of hamsters that produce stable dominant-subordinate relationships, infusions of OT into the medial preoptic anterior hypothalamic continuum (MPOA-AH) increase flank-marking in dominant hamsters in a dosedependent way (Harmon et al., 2002). In dominant male squirrel monkeys, intracerebroventricular (i.c.v.) injections of OT increase sexual advances and aggressive behavior towards female conspecifics (Winslow and Insel, 1991). The same manipulation performed in subordinates increases affiliation, having no effect on other behaviors. One mammalian species in which social roles are highly relevant for survival and fitness is the eusocial naked mole-rat (Heterocephalus glaber). Naked mole-rats live in large subterranean groups numbering up to approximately 300 individuals (Brett, 1991; Lacey and Sherman, 1991). Each colony has a single breeding female (the queen) and 1-3 breeding males. All other animals, 3

ACCEPTED MANUSCRIPT called subordinates, are reproductively suppressed, do not generally exhibit sex differences (Holmes et al., 2007), and participate in foraging, nest and tunnel maintenance, pup care, and defense against intruders. Within the subordinate caste, further social stratification and divisions

PT

of labor exist: „soldiers‟ or larger, older mole-rats are observed to be more aggressive and defend the colony, whereas „workers‟ or younger, smaller individuals participate in cooperative

RI

behaviors such as pup care and colony maintenance (Jarvis et al., 1991; Lacey and Sherman,

SC

1991). Recent work from our lab (Mooney et al., 2015) supports the classification of these „subcastes‟ and further indicates that such behavioral specializations remain stable in the long

NU

term if colony demands are consistent.

The present study sought to determine whether differences in OT activity are associated with the

MA

distinct behavioral roles found in workers and soldiers. Specifically, we hypothesized that soldiers would be more aggressive and exhibit greater OT activation than workers in a context

D

where aggression serves the needs of the group. Therefore, we employed a modified resident/intruder paradigm and compared the behavioral responses to unfamiliar intruders in

TE

worker and soldier residents of both sexes, and assessed associated activation of OT neurons

AC CE P

using double-label immunohistochemistry for OT and the immediate early gene marker c-fos (Flanagan et al., 1993). We also measured circulating cortisol to evaluate the putative relationship between level of arousal and activation of OT neurons in the different subcastes.

Methods

Animals and Housing We used a total of 60 adult subordinate naked mole-rats from 3 different colonies: 20 experimental animals were tested as residents in a modified resident/intruder (RI) paradigm, paired with 20 stimulus intruder animals, and 20 un-manipulated animals served as in-colony controls. Representatives from each group came from each of the 3 colonies. RI and in-colony groups contained both workers and soldiers (N = 10 each per group) while stimulus intruder animals were all age-matched workers from a different colony than the respective resident. Naked mole-rat colonies were housed in large (65 cm L x 45 cm W x 23 cm H), medium (46 cm 4

ACCEPTED MANUSCRIPT L x 24 cm W x 15 cm H), and small (30 cm L x 18 cm W x 13 cm H) polycarbonate cages connected by tubes (25 cm L x 18 cm D) and kept on a 12:12 light/dark cycle at 28-30°C. Animals were fed ad libitum with a diet of sweet potato and wet 19% protein mash (Harlan

PT

Laboratories, Inc.). The age of the animals ranged from 13 months to 3 years of age and the animals weighed between 23 and 69 g (Table 1). All experimental procedures followed federal

RI

and institutional guidelines and were approved by the University Animal Care Committee.

SC

In this study, we selected soldiers and workers that had been prescreened for aggressive and nonaggressive behavior (respectively) toward unfamiliar conspecifics. We note here that labeling

NU

animals as “workers” may require additional confirmation, since we did not select animals based on in-colony working behavior but instead confirmed them as “non-soldiers”. In the case where

MA

an appropriate prescreened RI animal could not be age-matched with a stimulus animal from another colony (N=1), we pseudo-randomly selected animals on the basis of weight: animals

D

over 40 g were categorized as soldiers and those under 35 g as workers, a classification based on earlier findings (Mongillo et al., 2014). This same classification was used for the majority of in-

TE

colony animals (N=16). All stimulus intruder animals were classified as workers based on weight

AC CE P

(< 35 g). To further confirm that weight was a suitable method for subcaste classification in the current sample, we performed a t-test comparing workers and soldiers across treatment (RI versus in-colony) for animals that were prescreened for aggressive behavior. This demonstrated that subcaste differences in weight, favoring soldiers, were present in those animals classified using behavior alone (Table 1). While every effort was made to have equal representation of males and females in our groups, including sex-matching RI animals and stimulus intruders, it is often difficult to determine naked mole-rat sex on the basis of external observation alone (Peroulakis et al. 2002; Holmes et al., 2009). We confirmed sex at the time of dissection and discovered that our groups were not balanced across sex (e.g., only 2 males were present in the soldier in-colony group). Due to the lack of statistical power in our resulting sex groups and the fact that status effects outweigh the effects of sex in this species (Holmes et al., 2007) we did not include sex as a variable in our t-test and ANOVA analyses (see below).

Behavioral Testing and Scoring 5

ACCEPTED MANUSCRIPT Each RI worker or soldier was habituated in isolation for 5 min in one medium-sized cage of their home colony by sealing the connecting tube. Following habituation, the stimulus intruder animal was introduced to the same chamber as the RI animal and pairs were video-recorded for

PT

30 min. Animals remained in the chamber for an additional 1.5 h to provide the time required for c-fos accumulation. In-colony control animals were marked for identification and placed back in

RI

their home colony for 2 h. We used behavioral scoring software (Observer, Noldus Information

SC

Technology) to quantify the frequency and duration of behaviors for the first 10 min of resident/intruder videos, scoring one animal at a time from each pair. Behaviors were broadly

NU

categorized as interactive (including non-aggressive and aggressive acts) or non-interactive (such as climbing and digging). A list of all behaviors scored is provided in Table 2. When quantifying

MA

behavior, scorers were blind to the sex and subcaste of the animal and inter-rater reliability was 97%. We did not collect behavioral data for in-colony animals because no direct behavioral comparisons could be made between animals interacting freely with multiple colony members

TE

D

and those interacting with a single intruder.

AC CE P

Tissue Extraction and Immunohistochemistry Following the 2 h experimental window, all animals were deeply anesthetized with isoflurane and rapidly decapitated. Brains were then extracted, post-fixed in 4% paraformaldehyde for 4 h, and transferred to a solution of 20% sucrose kept at 4°C. Using a sliding microtome, brains were blocked at the olfactory bulbs and the cerebellum and sliced into four series of coronal sections 30 µm thick. These were then stored at -20°C until processing. Trunk blood was collected and kept on ice until centrifugation. Serum was stored at -20°C until processing. In order to identify active OT neurons, we performed double-label immunohistochemistry for OT and the immediate-early gene product c-fos (see Figure 1). The specificity of the OT antibody used in this study has been previously confirmed (Rosen et al., 2008). Briefly, a one in four series was stained first for c-fos and then for OT, with experimental groups yoked across each staining run. For c-fos, tissue sections were first incubated for 90 min in blocking solution [Phosphate Buffered Saline (PBS) + 4% normal goat serum (NGS) + 0.3% TritonX-100 + 0.9% H2O2] at room temperature. Then, sections were incubated in a primary antibody solution for c6

ACCEPTED MANUSCRIPT fos [1:500 rabbit anti-c-fos polyclonal antiserum (Santa Cruz Biotech, Inc., Dallas, TX)] in blocking solution (without H2O2) for 24 h at 4°C. Next, sections were incubated in a secondary antibody solution [1:200 goat anti-rabbit (Vector Laboratories, Burlingame, CA) in PBS + 2%

PT

NGS + 0.3% TritonX-100] for 90 min at room temperature, followed by incubation in ABC solution for 90 min. Finally, a 6 min 3,3‟-diaminobenzidine (DAB) reaction was used to produce

RI

a brown reaction product. For the c-fos protocol, tissue sections were rinsed in PBS 3 x 5 min

SC

before and between each subsequent step, as well as after the DAB reaction. Before staining tissue for OT, sections were incubated in sodium citrate (0.44 g in 30 ml dH2O)

NU

at 80°C for 30 min to reduce cross-reactivity between c-fos and OT antisera. For OT staining, tissue sections were first washed in 1% H2O2 for 10 min. Sections were then placed in blocking

MA

solution for 1 h [PBS + 10% NGS + 0.3% TritonX-100] at room temperature. Then, tissue sections were incubated in a primary antibody solution [1:1000 anti-OT polyclonal antiserum

D

(Millipore) + 2% NGS] for approximately 72 h at 4°C. Following incubation, sections were placed in a secondary antibody solution [1:500 goat anti-rabbit antibody + 2% NGS] for 1 h and

TE

then incubated for 1 h in ABC solution. We visualized OT using a Vector-SG reaction (Vector

AC CE P

Laboratories) for 6 min, which produces a blue/gray reaction product. For the OT protocol, tissue sections were rinsed in PBS with 0.1% pig gelatin and 0.3% TritonX-100, 3 x 5 min before and between each subsequent step, as well as after the Vector-SG reaction. Following these protocols, tissue was mounted onto gelatin-coated slides, dehydrated, and cover-slipped using Permount (Fisher Scientific).

Cell Counts The total number of OT-immunoreactive (ir), c-fos-ir, and the number of OT-c-fos doublelabeled cells (Fig. 1) were quantified for the paraventricular nucleus (PVN) and supraoptic nucleus (SON) in the hypothalamus. We also quantified the total number of c-fos-ir in the medial amygdala (MeA), a region that does not contain OT-ir cell bodies. Stereological analyses were conducted bilaterally using Stereologer software (Stereology Resource Center, Inc.) by an observer blind to experimental treatment. Outlines of the PVN, SON and MeA were traced in each section (~3 sections per animal) and the number of OT-ir and c-fos-ir cells was quantified 7

ACCEPTED MANUSCRIPT using a rare event protocol for OT-ir and a dissector protocol for c-fos-ir. Numbers of OT-c-fosir cells were recorded manually while performing counts for OT-ir. For the dissector protocol, total numbers were counted on 3-dimensional planes (50 µm frame area) by focusing through the

PT

z-axis, with guard zones in place to prevent the inclusion of split nuclei. The raw number for

SC

RI

each animal was multiplied by the sampling ratio to provide an estimate for the entire region.

Hormone Assays

NU

To measure cortisol, serum was diluted at 1:10 in buffer and processed using a Cayman Chemical kit (Cat. No. 500360) with a sensitivity of 80 pg/mL (intra-assay variation <14%).

MA

According to the manufacturer, this kit cross-reacts with prednisolone (22%), cortexolone (6.1%), cortisone (2.0%), corticosterone (1.3%), and <1.0% with all other tested steroids.

D

Absorbance was measured at 405 nm as per manufacturer instructions and all samples were run

TE

in duplicate with a Synergy-HT-Bio-Tek 96-well microplate reader. The averages of duplicate

AC CE P

readings were corrected by the dilution factor prior to statistical analysis.

Statistical Analyses

Total aggression scores for RI animals were calculated by summing the frequencies or durations of individual aggressive behaviors for each animal. We also calculated aggression as a percentage of total interactive behavior (total interactive = total aggressive + total nonaggressive) using the total aggression duration score (% aggressive, Table 2). Independent samples t-tests were conducted to compare RI workers and soldiers on the frequency and duration of aggression as well as the % aggressive score. We also performed t-tests on the duration of non-aggressive and non-interactive behaviors in order to determine whether increases in aggression were due to increased general activity. For analyses of cell counts, four animals were excluded from the OT single-label and five from the c-fos and double-label analyses due to poor tissue quality. We therefore conducted our analyses on the remaining RI and in-colony animals (N = 36 for OT single-label, N = 35 for c-fos and double-label). Because group differences were detected in the number of c-fos-ir cells in the PVN (see below), we also 8

ACCEPTED MANUSCRIPT calculated the percentage of OT cells that were double-labeled with c-fos for each animal (n double-labeled OT-c-fos neurons / n OT neurons x 100 = % OT-c-fos) and ran a 2-way caste by treatment ANOVA on each of three dependent variables: OT-ir, c-fos-ir, and % OT-c-fos-ir cell

PT

counts. Circulating cortisol was also analyzed using a 2-way caste by treatment ANOVA.

RI

For all aforementioned analyses, Eta squared (η2) was used as an estimate of effect size for main effects and interactions and Cohen‟s d was used as a measure of effect size for pairwise

SC

comparisons. Finally, a general linear model (GLM) was conducted to determine significant predictors of our main variable of interest, % OT-c-fos-ir. We included caste and treatment as

NU

categorical variables, cortisol as a continuous variable (grand-mean centered), and a treatment x cortisol interaction term. Behavior was not included in the model because we did not collect

MA

behavioral data for in-colony control animals. No other interaction terms were included due to the lack of relationship between any of the other predictors. All analyses were performed using R

Behavior

AC CE P

TE

D

statistical software and the threshold for statistical significance was set at p = 0.05.

Results

For aggression scores, t-tests revealed that soldiers were significantly more aggressive than workers (t(18) = -2.28, p = 0.035, Cohen‟s d = -1.02 for % aggressive; t(18) = -2.23, p = 0.038, Cohen‟s d = -0.99 for frequency; t(18) = -2.15, p = 0.045, Cohen‟s d = -0.96 for duration; Fig. 2A-C). For non-aggressive interaction scores, there were no significant differences between workers and soldiers (t(18) = -0.66, p = 0.52, Cohen‟s d = -0.30). Workers were significantly more active outside of interaction than soldiers (t(18) = 2.37, p = 0.030, Cohen‟s d = 1.06), demonstrating that the increased aggression in soldiers was not due to increases in overall activity. Immunohistochemistry

9

ACCEPTED MANUSCRIPT For percentage of double-labeled cells in the PVN, a significant main effect of caste (F(1, 32) = 6.10, p = 0.019, η2 = 0.16; Fig. 3A) showed that workers had greater % OT-c-fos-ir than soldiers. No main effect of treatment (F(1, 32) = 0.20, p = 0.66, η2 = 0.00), or caste by treatment

PT

interaction (F(1, 32) = 0.11, p = 0.74, η2 = 0.00), was found. We found the same main effect of caste when analyzing c-fos-ir counts (F(1, 31) = 4.90, p = 0.034, η2 = 0.13; Fig. 3B), where

RI

workers had greater number of c-fos-ir cells than soldiers. All other effects were non-significant

SC

(F(1, 32) = 0.46, p = 0.50, η2 = 0.01 for treatment; F(1, 32) = 1.42, p = 0.24, η2 = 0.038 for the interaction). For OT-ir single-label counts, no significant effects emerged (F(1, 32) = 0.13, p =

NU

0.72, η2 = 0.00 for caste; F(1, 32) = 0.058, p = 0.81, η2 = 0.00 for treatment; F(1, 32) = 0.073, p = 0.79, η2 = 0.00 for the interaction).

MA

For the SON, no significant effects were detected for % OT-c-fos-ir or OT-ir single labeled cells. For c-fos-ir, no significant effects were found (F(1, 32) = 0.13, p = 0.72, η2 = 0.00 for caste; F(1,

D

32) = 0.058, p = 0.81, η2 = 0.00 for treatment; F(1, 31) = 3.56, p = 0.069, η2 = 0.10 for the interaction). We also ran an ANOVA on c-fos-ir in the MeA and found no significant effects in

TE

this region (F(1, 35) = 0.65, p = 0.43, η2 = 0.018 for caste; F(1, 35) = 0.66, p = 0.42, η2 = 0.017

Cortisol

AC CE P

for treatment; F(1, 35) = 0.019, p = 0.89, η2 = 0.00 for the interaction).

A caste by treatment ANOVA revealed a significant main effect of treatment on circulating cortisol (F(1, 32) = 42.27, p < 0.0001, η2 = 0.56; Fig. 4), with RI animals having higher cortisol levels than in-colony controls. The effect of caste and the caste by treatment interaction were non-significant (F(1, 32) = 0.45, p = 0.51, η2 = 0.00 for caste; F(1, 32) = 0.035, p = 0.85, η2 = 0.00 for the interaction).

Linear Model Predicting OT-c-fos Immunoreactivity We conducted a general linear model (GLM) predicting % OT-c-fos-ir with caste, treatment, cortisol, and a treatment x cortisol interaction term as predictors. None of the predictors showed a significant effect on the outcome variable, save for caste, where being a worker significantly predicted greater % OT-c-fos-ir (t(1) = 2.35, p = 0.025, partial r2 = 0.077). The overall model 10

ACCEPTED MANUSCRIPT itself was non-significant (F(4, 31) = 2.02, p = 0.12, Multiple R2 = 0.21). In an attempt to determine if sex had any impact on activation of OT neurons, which we could not test in the ANOVAs due to unbalanced samples, we also ran a model adding sex as a predictor. In this

PT

model, caste was non-significant as a predictor (t(1) = 1.77, p = 0.087, partial r2 = 0.094) and the

SC

RI

overall model was also non-significant (F(5, 30) = 1.72, p = 0.16, Multiple R2 = 0.22).

Discussion

NU

These data demonstrate that within naked mole-rat subordinates, further subcaste differences exist between workers and soldiers for both behavior and neural activation.

MA

Specifically, as expected, workers were significantly less aggressive than soldiers in a modified resident/intruder paradigm. However, contradicting our initial hypothesis, workers showed

D

greater activation of OT neurons in the PVN than soldiers both in-colony and when faced with an

TE

intruder. No caste or treatment effects were seen on OT or c-fos immunoreactivity in the SON. Interestingly, we found that all animals in the resident/intruder paradigm, irrespective of

AC CE P

subcaste, exhibited higher circulating levels of cortisol than in-colony controls, suggesting that exposure to an intruder increases arousal in both workers and soldiers. The GLM suggests that the major contributor to differences in OT neuron activation in the PVN is the subcaste of the animal, and not social treatment, sex, or cortisol levels. We found that subcaste affected aggression in the RI animals, with soldiers exhibiting more aggression than workers. This was predicted given that most of our experimental animals were categorized based on a pre-screen for aggression and, importantly, aligns with other studies describing a soldier sub-caste: larger, older subordinates exhibit more aggression when a foreign mole-rat is introduced to an intact colony (O‟Riain and Jarvis, 1997) or when paired with an unfamiliar conspecific in a neutral setting (Mooney et al., 2015). However, our current data do not support a role for OT in the promotion of status-dependent aggression, since we found lower OT-c-fos-ir in animals that were significantly more aggressive in a resident/intruder paradigm. In addition to being more aggressive, soldiers perform fewer cooperative behaviors within the colony compared to workers, such as pup care and tunnel maintenance (Lacey and Sherman, 1991; Jarvis et al., 1991; Mooney et al., 2015). Because our results indicate that OT neurons are 11

ACCEPTED MANUSCRIPT less active in soldiers in both in-colony and RI treatment groups, we propose that subcaste differences in activation of OT neurons are associated with differences in prosocial behaviors and not aggression. Additional support for this idea comes from our recent demonstration that

PT

peripheral OT administration in naked mole-rat subordinates increases investigatory behavior of and time spent in proximity to a familiar conspecific (Mooney et al., 2014). Co-administration of

RI

peripheral OT with an OTR antagonist prevents these effects. While worker-soldier differences

SC

within subordinates or responses to unfamiliar animals were not explicitly examined, the results indicate that prosocial behaviors in subordinates may be controlled by the action of OT.

NU

Other species also exhibit status differences in OT and its associated neuronal and receptor activation. Squirrel monkeys of differing social status exhibit status-specific behaviors

MA

upon OT administration, i.e. aggression in dominants and submission in subordinates (Winslow and Insel, 1991). Subordinate, but not dominant, rats exhibit lower OTR mRNA expression in

D

the MeA after combined stressors and aggressive encounters (Timmer et al., 2011). Microinfusion of an OTR antagonist in this area promotes long-term establishment of the

TE

hierarchy. While data was not collected on activation of OT neurons in the PVN, it suggests a

AC CE P

role for OT activity in maintaining social standing. In zebra finches, where the close OT homolog mesotocin (MT) is expressed, MT neuron activation in the PVN is significantly predicted by the interaction of sex and dominance in that dominant females have greater activation of MT neurons (Kelly and Goodson, 2014). Similar effects exist in the mandarin vole such that dominant females exhibit greater OT-ir neurons in the PVN than dominant males and in males only, subordinates exhibit significantly less OT-ir neurons than dominants (Qiao et al., 2014). In contrast to our study, these studies suggest that in species with little to no reproductive skew, sex is a major determinant of OT neuron morphology in the PVN. As well, since dominance can be associated with both aggression and prosociality, greater OT may promote a broader social competency that leads to dominance within a group. In species such as the naked mole-rat, where social roles are more fixed (at least within an intact colony), it is likely more adaptive for OT to promote only certain behaviors within certain subcastes (such as prosociality in workers but not soldiers). Results from the ANOVA and GLM suggest that for OT neurons in the PVN, social status (in this case subcaste) is what drives differences in activation, regardless of social 12

ACCEPTED MANUSCRIPT treatment, sex, or cortisol levels. That c-fos-ir in the PVN followed a similar pattern to OT-c-fos double label but that no differences were found in OT-ir neuron number further suggests that these subcaste differences occur at the level of neural activation and not neural morphology. The

PT

current study did not find any differences in OT neuron number between soldiers and workers, indicating that subcaste differences in neural morphology are either non-existent or more subtle

RI

than differences between castes. In terms of OT, Mooney and Holmes (2013) found that

SC

subordinates have a greater number of OT cells in the PVN, but not the SON, than breeders, though the study did not further classify subordinates into workers and soldiers. Thus,

NU

differences in neural morphology may exist between soldiers and workers in other brain areas important for sociality such as those outlined above, and may be uncovered with additional

MA

volumetric and/or phenotypic analyses.

In contrast to the aggression and OT-c-fos data, in which differences were driven by

D

subcaste, social treatment was the main determinant of differences in cortisol with RI animals exhibiting higher levels than in-colony controls. In addition to this, cortisol did not significantly

TE

predict OT neuron activity in the GLM. The lack of relationship between OT and cortisol

AC CE P

appears to contradict other findings, which indicate that OT attenuates stress in a variety of paradigms (Neumann, 2008). Several studies demonstrate that OT acts in the PVN on corticotropin-releasing factor (CRF). In rats, OT administration attenuates the expression of CRF mRNA in the PVN as well as corticosterone release in response to restraint stress (Windle et al., 2004). While status and sex effects exist in naked mole-rat CRF1 and CRF2 receptor density, respectively (Beery et al., in press), measuring CRF release and circulating cortisol following OT manipulation will better reveal OT‟s anxiolytic actions in this species. For many mammalian species, social status is related to stress reactivity (DeVries et al., 2003). Interestingly, our cortisol results suggest otherwise and show a driving effect of treatment and not subcaste. Our results suggest that for workers and soldiers alike, being paired with an unfamiliar animal is stressful compared to being in the colony. In a naked mole-rat colony studied by Clarke and Faulkes (1997), urinary cortisol levels increased significantly upon removal of the queen for all animals regardless of caste. Thus, threats to the integrity of the colony, either the death of the queen or a foreign intruder, appear to modulate cortisol levels in all subordinates. Indeed, DeVries et al. (2003) state that status effects on stress vary depending 13

ACCEPTED MANUSCRIPT on the level of social stability and the species-specific behavioral profiles of dominant and subordinate animals. Because we did not see subcaste differences in cortisol in our study, it is possible that soldiers cannot be considered „dominant‟ in the same way that breeders can in

PT

stable naked mole-rat colonies. Furthermore, the fact that cortisol did not contribute significantly to OT-c-fos-ir in the GLM suggests that any relationship(s) between OT, stress, and subcaste are

RI

complex and may depend on other factors not considered in our study.

SC

In sum, our findings indicate that within naked mole-rat subordinates, further status differences exist in the activation of OT neurons and behavior toward an unfamiliar intruder.

NU

While soldiers were more aggressive toward intruders, workers exhibited greater OT neural activity regardless of social treatment. Importantly, subcaste was the only significant predictor of

MA

differences in neural activation, suggesting that social status within the colony plays a prominent role in shaping responses to social situations. Future work will need to directly test whether

D

manipulation of central OT has differential effects on the subcaste-specific behaviors of soldiers and workers. Ultimately, investigation into the neurobehavioral effects of OT in this species will

AC CE P

organization in mammals.

TE

offer insight into the neural mechanisms underlying social status and complex social

14

ACCEPTED MANUSCRIPT

Acknowledgments This work was funded by a NSERC Discovery Grant (402633) to MMH, a NSERC CGS

PT

master‟s award to GAH, a NSERC PGS doctoral award to DEP, and an Ontario Graduate Scholarship to SJM. We thank Alyssandra Chee-a-tow for assistance with microscopy and

AC CE P

TE

D

MA

NU

SC

RI

Gabriela Rivera for assistance with tissue collection and slicing, as well as video scoring.

15

ACCEPTED MANUSCRIPT

PT

References

Bales, K. L., Carter, C. S., 2003. Sex differences and developmental effects of oxytocin on

RI

aggression and social behavior in prairie voles (Microtus ochrogaster). Horm. Behav. 44,

SC

178-184.

Beery, A. K., Zucker, I., 2010. Oxytocin and same-sex social behavior in female meadow

NU

voles. Neuroscience 169, 665-673.

Beery, A.K., Bicks, L., Mooney, S.J., Goodwin, N.L, & Holmes, M.M. Sex, social status, and

MA

CRF receptor densities in naked mole-rats. J. Comp. Neurol., in press. Brett, R. A., 1991. The ecology of naked mole-rat colonies: burrowing, food, and limiting

D

factors. In: Sherman, P. W., Jarvis, J. U. M., Alexander, R. D. (Eds.), The Biology of the

TE

Naked Mole-Rat. Princeton University Press, Princeton, NJ, pp. 137-184.

AC CE P

Clarke, F. M., & Faulkes, C. G., 1997. Dominance and queen succession in captive colonies of the eusocial naked mole–rat, Heterocephalus glaber. Proc. Roy. Soc. B: Biol. Sci. 264, 993-1000.

Crockford, C., Wittig, R. M., Langergraber, K., Ziegler, T. E., Zuberbühler, K., Deschner, T., 2013. Urinary oxytocin and social bonding in related and unrelated wild chimpanzees. Proc. Roy. Soc. B: Biol. Sci. 280, 2012-2765. DeVries, A. C., Glasper, E. R., & Detillion, C. E., 2003. Social modulation of stress responses. Physiol. Behav. 79, 399-407. Ebner, K., Wotjak, C. T., Landgraf, R., Engelmann, M., 2000. A single social defeat experience selectively stimulates the release of oxytocin, but not vasopressin, within the septal brain area of male rats. Brain Res. 872, 87-92. Flanagan, L., Pfaus, J., Pfaff, D., McEwen, B., 1993. Induction of Fos immunoreactivity in oxytocin neurons after sexual activity in female rats. Neuroendocrinology 58, 352-358. 16

ACCEPTED MANUSCRIPT Harmon, A., Moore, T., Huhman, K., Albers, H., 2002. Social experience and social context alter the behavioral response to centrally administered oxytocin in female Syrian hamsters.

PT

Neuroscience 109, 767-772. Holmes, M. M., Goldman, B. D., Goldman, S. L., Seney, M. L., Forger, N. G., 2009.

RI

Neuroendocrinology and sexual differentiation in eusocial mammals. Front.

SC

Neuroendocrinol. 30, 519-533.

Holmes, M. M., Rosen, G. J., Jordan, C. L., de Vries, G. J., Goldman, B. D., Forger, N. G., 2007.

NU

Social control of brain morphology in a eusocial mammal. Proc. Natl. Acad. Sci. U. S. A. 104, 10548-10552.

MA

Insel, T. R., Hulihan, T. J., 1995. A gender-specific mechanism for pair bonding: oxytocin and partner preference formation in monogamous voles. Behav. Neurosci. 109, 782-789.

D

Insel, T. R., Shapiro, L. E., 1992. Oxytocin receptor distribution reflects social organization in

TE

monogamous and polygamous voles. Proc. Natl. Acad. Sci. U. S. A. 89, 5981-5985.

AC CE P

Jarvis. J. U. M., O‟Riain, M. J. McDaid, E., 1991. Growth and factors affecting body size in naked mole-rats. In: Sherman, P. W., Jarvis, J. U. M., Alexander, R. D. (Eds.), The Biology of the Naked Mole-Rat. Princeton University Press, Princeton, NJ, pp. 358-383. Kalamatianos, T., Faulkes, C. G., Oosthuizen, M. K., Poorun, R., Bennett, N. C., Coen, C. W., 2010. Telencephalic binding sites for oxytocin and social organization: A comparative study of eusocial naked mole‐ rats and solitary cape mole‐ rats. J. Comp. Neurol. 518, 1792-1813. Kelly, A. M., & Goodson, J. L., 2014. Personality is tightly coupled to vasopressin-oxytocin neuron activity in a gregarious finch. Front. Behav. Neurosci. 8. doi:10.3389/fnbeh.2014.00055 Kosfeld, M., Heinrichs, M., Zak, P. J., Fischbacher, U., Fehr, E., 2005. Oxytocin increases trust in humans. Nature 435, 673-676.

17

ACCEPTED MANUSCRIPT Lacey, E., Sherman, P., 1991. Social organization of naked mole-rat colonies: evidence for divisions of labor. In: Sherman, P. W., Jarvis, J. U. M., Alexander, R. D. (Eds.), The

PT

Biology of the Naked Mole-Rat. Princeton University Press, Princeton, NJ, pp. 275-336. Mongillo, D. L., Kosyachkova, E. A., Nguyen, T. M., Holmes, M. M., 2014. Differential effects

RI

of chronic fluoxetine on the behavior of dominant and subordinate naked mole-rats. Behav.

SC

Brain Res. 258, 119-126.

Mooney, S. J., Holmes, M. M., 2013. Social condition and oxytocin neuron number in the

NU

hypothalamus of naked mole-rats (Heterocephalus glaber). Neuroscience. 230, 56-61. Mooney, S. J., Douglas, N. R., Holmes, M. M., 2014. Peripheral administration of oxytocin

MA

increases social affiliation in the naked mole-rat (Heterocephalus glaber). Horm. Behav. 65, 380-385.

D

Mooney, S. J., Filice, D. C. S., Douglas, N. R., & Holmes, M. M., 2015. Task specialization and

TE

task-switching in eusocial mammals. Anim. Behav. 109, 227-233.

AC CE P

Neumann, I. D., 2008. Brain oxytocin: a key regulator of emotional and social behaviours in both females and males. J. Neuroendocrinol. 20, 858-865. O‟Riain, M., Jarvis, J., 1997. Colony member recognition and xenophobia in the naked mole-rat. Anim. Behav. 53, 487-498. Peroulakis, M. E., Goldman, B., Forger, N. G., 2002. Perineal muscles and motoneurons are sexually monomorphic in the naked mole‐ rat (Heterocephalus glaber). J. Neurobiol. 51, 33-42. Qiao, X., Yan, Y., Wu, R., Tai, F., Hao, P., Cao, Y., & Wang, J., 2014. Sociality and oxytocin and vasopressin in the brain of male and female dominant and subordinate mandarin voles. J. Comp. Physiol. A. 200, 149-159. Rosen, G.J., de Vries, G.J., Goldman, S.L., Goldman, B.D., & Forger, N.G., 2008. Distribution of oxytocin in the brain of a eusocial rodent. Neurosci. 155, 809-817. 18

ACCEPTED MANUSCRIPT

Snowdon, C. T., Pieper, B. A., Boe, C. Y., Cronin, K. A., Kurian, A. V., Ziegler, T. E., 2010.

PT

Variation in oxytocin is related to variation in affiliative behavior in monogamous, pairbonded tamarins. Horm. Behav. 58, 614-618.

RI

Timmer, M., Cordero, M. I., Sevelinges, Y., & Sandi, C., 2011. Evidence for a role of oxytocin

SC

receptors in the long-term establishment of dominance hierarchies. Neuropsychopharmacol. 36, 2349-2356.

NU

Williams, J. R., Insel, T. R., Harbaugh, C. R., Carter, C. S., 1994. Oxytocin administered centrally facilitates formation of a partner preference in female prairie voles (Microtus

MA

ochrogaster). J. Neuroendocrinol. 6, 247-250.

Windle, R. J., Kershaw, Y. M., Shanks, N., Wood, S. A., Lightman, S. L., Ingram, C. D., 2004.

D

Oxytocin attenuates stress-induced c-fos mRNA expression in specific forebrain regions

AC CE P

2974-2982.

TE

associated with modulation of hypothalamo–pituitary–adrenal activity. J. Neurosci. 24,

Winslow, J. T., Insel, T. R., 1991. Social status in pairs of male squirrel monkeys determines the behavioral response to central oxytocin administration. J. Neurosci. 11, 2032-2038. Winslow, J. T., Shapiro, L., Carter, C. S., Insel, T. R., 1993. Oxytocin and complex social behavior: species comparisons. Psychopharmacol. Bull. 29, 409-414.

19

ACCEPTED MANUSCRIPT

PT

Figure Captions

Figure 1. Photomicrographs of stained coronal sections of the naked mole-rat brain showing (A)

RI

double label OT and c-fos immunoreactivity in the paraventricular nucleus (PVN) and (B) OT or c-fos single-labeled and OT-c-fos double-labeled cells in the PVN. Black arrow points to the

SC

third ventricle in (A) and to a double labeled OT-c-fos-ir cell in (B). The black arrowhead in (B) points to a single labeled OT-ir cell that is adjacent to and partially overlapping with a single

NU

labeled c-fos-ir cell. Scale bar, shown in B, = 50 µm for (A) and 20 µm for (B).

MA

Figure 2. Mean (+/- SEM) aggression scores for resident Workers (white bars) and Soldiers (grey bars), expressed as (A) % aggressive behavior, (B) frequency, and (C) duration. * Denotes a

D

statistically significant difference between Workers and Soldiers (p < 0.05). Figure 3. Mean (+/- SEM) immunoreactivity for RI animals (grey bars) and in-colony controls

TE

(white bars) broken down by subcaste. % OT-c-fos-ir (A) and raw number of c-fos-ir (B) cells

AC CE P

are shown. * Denotes a statistically significant main effect of subcaste (p < 0.05). Figure 4. Mean (+/- SEM) circulating cortisol in ng/mL for RI animals (grey bars) and in-colony controls (white bars) broken down by subcaste. *** Denotes a statistically significant main effect of treatment (p < 0.001).

20

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC CE P

TE

D

MA

Figure 1

21

AC CE P

Figure 2

TE

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

22

TE AC CE P

Figure 3

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

23

AC CE P

Figure 4

TE

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

24

ACCEPTED MANUSCRIPT

Table 1. Breakdown of experimental groups: stimulus animals, residents, and controls. *Prescreened resident soldiers weighed significantly more than prescreened workers (t(21) = -6.27, p < 0.0001, Cohen‟s d = -2.62). W =

grams. SD = standard deviation. Residents

S

N

20

10

Sex, N female

9F

Age, years (SD)

1.59 (0.47)

Weight, g (SD)*

30.6 (3.28)

S

W

10

10

10

5F

2F

8F

2.51 (0.45)

1.87 (0.53)

2.15 (0.47)

1.47 (0.46)

53.10 (11.26)

31.20 (3.22)

47.90 (9.01)

26.70 (3.89)

MA

3F

AC CE P

TE

D

Controls

W

SC

--

NU

Sub-Caste

RI

Stimulus Animals

PT

workers, S = soldiers. F = number of females (number of males calculated by subtracting from N in each group). g =

25

ACCEPTED MANUSCRIPT

Table 2. Behaviors scored in resident-intruder interactions, classified according to Interactive (including Aggressive/Non-Aggressive) or Non-Interactive behavior. * Used to create the Percent Aggressive Score: Sum of Aggressive / (Sum of Aggressive + Sum of Non-Aggressive). **Fall/Wrestle was counted in reverse, so that each

PT

animal‟s Total Aggressive score included the other NMR‟s score in order to accurately reflect aggression. NMR =

Classification Interactive Aggressive*

Behavior

Description

NU

SC

Open Mouth Gaping Biting Incisor Fencing Shoving Batting Fall/Wrestle** Jerking Away Being Sniffed Huddling Sniff Appraisal Under Pass Pass Over Avoidance

AC CE P

TE

D

MA

Non-Aggressive*

Non-Interactive

RI

naked mole-rat.

Active Inactive Climbing Digging Self-Nuzzling Self-Grooming Environmental Maintenance Gnawing

26

Territorial stretching open of mouth Use of teeth on skin Jaws locking for a period of time Pushing away with head Forelimb fighting One NMR on its back Defensive body movement NMR is still while being sniffed Making skin contact Sniffing head, neck, thorax Body passes under other NMR Body passes over other NMR Actively moving away from other NMR In transit; default behavior No engagement in any activity Using limbs to climb cage wall Displace bedding using fore/hind limbs Coprophagy Repeated moving of limbs to clean body Displace or move nesting material Teeth scraping against surface of cage

ACCEPTED MANUSCRIPT

Highlights

PT

We compared behavior and oxytocin neural activity between naked mole-rat sub-castes.

RI

Soldiers were more aggressive than workers during a resident/intruder paradigm.

SC

Workers exhibited greater oxytocin neural activity than soldiers in this paradigm.

AC CE P

TE

D

MA

NU

Sub-caste is a significant predictor of oxytocin neural activity.

27