Mu opioid receptors in the medial preoptic area govern social play behavior in adolescent male rats

Abstract

Neural systems underlying important behaviors are usually highly conserved across species. The medial preoptic area (MPOA) has been demonstrated to play a crucial role in reward associated with affiliative, non-sexual, social communication in songbirds. However, the role of MPOA in affiliative, rewarding social behaviors (e.g., social play behavior) in mammals remains largely unknown. Here we applied our insights from songbirds to rats to determine whether opioids in the MPOA govern social play behavior in rats. Using an immediate early gene (i.e., Egr1, early growth response 1) expression approach, we identified increased numbers of Egr1-labeled cells in the MPOA after social play in adolescent male rats. We also demonstrated that cells expressing mu opioid receptors (MOR, gene name Oprm1) in the MPOA displayed increased Egr1 expression when adolescent rats were engaged in social play using double immunofluorescence labeling of MOR and Egr1. Furthermore, using short hairpin RNA (shRNA)-mediated gene silencing we revealed that knockdown of Oprm1 in the MPOA reduced the number of total play bouts and the frequency of pouncing. Last, RNA sequencing differential gene expression analysis identified genes involved in neuronal signaling with altered expression after Oprm1 knockdown, and identified Egr1 as potentially a key modulator for Oprm1 in the regulation of social play behavior. Altogether, these results demonstrate that the MPOA is involved in social play behavior in adolescent male rats and support the hypothesis that the MPOA is part of a conserved neural circuit across vertebrates in which opioids act to govern affiliative, intrinsically-rewarded social behaviors.

1. Introduction

A growing body of research using the unique communication properties of songbirds has revealed neural mechanisms that underlie reward associated with affiliative, vocal-social interactions in non-sexual contexts ^1–3. Specifically, studies demonstrate a crucial role for opioids that bind to mu opioid receptors (MOR, gene name Oprm1) in the medial preoptic area (MPOA) in affiliative communication and reward that differs from that observed from mate-directed communication ^1,2,4,5. In rodent models, the MPOA is well-known for its role in sexual and maternal reward ^6–8; however, the role of MPOA in reward beyond these contexts has not been well studied.

Neural systems that underlie important behaviors are often highly conserved across species. For instance, many emotional networks in the brain including seeking, fear, rage, care and play are shared by all vertebrate species ⁹. The mesolimbic reward system and social behavior network are quite similar across the five major vertebrate lineages: mammals, birds, reptiles, amphibians, and teleost fish ¹⁰. This suggests that studies of MPOA in songbirds may be uncovering a central nucleus that is part of a core, conserved neural circuit in which opioids act to initiate, reward, and maintain important social behaviors in non-sexual contexts. If true, then the role of MPOA in affiliative song identified in birds should extend to behaviors in mammals that share central features with affiliative song. One such behavior is social play, which like affiliative bird song, is a highly rewarding, spontaneous, nonsexual, affiliative behavior that is important for developing cognitive and social skills ¹¹.

Extensive research has indicated that opioids that bind to MORs stimulate and reward play in juvenile rats ^12,13; however, the MPOA is generally considered nonessential for social play. Although the MPOA was one of the first regions in which opioids were implicated in rat play ¹⁴, subsequent correlational studies did not find relationships between play behavior and either c-Fos expression or opioid binding in the MPOA ^15,16. Moreover, lesions to MPOA did not disrupt juvenile play behavior ^17–19. However, this is what would be predicted if the mechanisms that underlie affiliative, rewarding singing behavior in songbirds extend to social play in mammals. Specifically, starlings given MPOA lesions either do not show deficits in affiliative singing behavior or they sing more than controls ⁴. Furthermore, if lesions to the MPOA disinhibit play behavior (as they do affiliative singing behavior in birds), given that in the past play studies rats were already playing at high rates, this could have been masked by a ceiling effect. Finally some lesion data do implicate the MPOA in social play. Specifically, lesions to the MPOA in juvenile rats were found to disrupt adult sexual behavior, unless the juveniles were housed under social conditions that allowed them to play ^18,19. This suggests that juvenile play behavior induces activity in the MPOA and can modify MPOA-dependent socio-sexual behaviors later in life.

In this study, we applied our insights from songbirds to rats to determine whether opioids in the MPOA of adolescent rats govern social play behavior. First, we explored indirectly a possible role for the MPOA in the performance of social play behavior using Egr1 (early growth response 1, also known as Zif268 or ZENK in songbird studies), a marker for cellular responses, and further determined whether MOR cells in the MPOA may be involved in social play using double immunofluorescence labeling of MOR and Egr1. Second, we knocked down gene expression of Oprm1 in the MPOA via adeno-associated viral (AAV) vectors and short hairpin RNA (shRNA)-mediated gene silencing to evaluate whether decreases in Oprm1 expression in the MPOA modulate social play behavior. Lastly, we investigated the molecular mechanisms by which MORs in the MPOA regulate social play using RNA sequencing (RNA-Seq) approaches.

2. Materials and methods

2.1. Animals

SAS Sprague-Dawley (SD) rats purchased from Charles River (Wilmington, MA, USA) were used in this study. The subjects were maintained under a 12:12 hr light/dark cycle (lights off at 09:00 AM CST) and at a controlled temperature (~22°C) with ad libitum access to rodent chow and tap water. All procedures followed the guidelines of the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by the Animal Care and Use Committee of the University of Wisconsin. Adequate measures were taken to minimize pain and discomfort of animal subjects.

2.2. Egr1 expression induced by social play behavior

A total of 40 male and 10 female rats at ages of 31–32 days (PN31–32) upon arrival in our animal facility were used and housed in mixed-sex groups of five containing 4 males and 1 female. After a 3-day acclimation in our colony, rats were housed individually for 24 hrs. We chose the isolation paradigm as it has been used effectively in past studies on immediate early gene expression and social play ^20–22 to enhance the motivation to play. Following this isolation period, animals were placed in the test cage either in groups consisting of 4 males and 1 female (play group) or alone (no-play group). The backs of the animals were labeled with a black Sharpie marker to identify individuals. Four males were labeled with 0, I, II, ++, separately, with no labeling of female. With this easy-to-identify labeling, we were able to score from recorded videos social play behavior displayed by 5 animals that were simultaneously interacting. Measures of play behavior (i.e., the same play parameters as in 2.6 below) were video-recorded immediately for 20 min and analyzed using VLC media player 3.0.8 (https://www.videolan.org/index.html). We chose the 20-min time frame of recording because rats display most social play behavior within the first 10–20 min of reuniting ^13,23. After the test, animals were placed back into their separate cages for 90 min to ensure the peak expression of play-associated immediate early gene Egr1.

2.3. Preparation of brain tissue slices for double fluorescence immunohistochemistry

After 90 min of social play test, play and no-play male rats were anaesthetized with isoflurane and then transcardially perfused with 4% paraformaldehyde in 0.1 M phosphate buffer (PB; pH 7.4). Brains were post-fixed overnight in the same fixative and then cryoprotected with 0.1 M PB containing 30% sucrose at 4°C for two days. Brains were snap-frozen and 40-μm coronal sections were sliced on a cryostat and stored in cryoprotectant solution at −20°C until processing. MPOA sections including the rostral (Bregma −0.24 to −0.48 mm) and caudal (Bregma −0.48 to −0.72 mm) levels of MPOA were collected according to The Rat Brain in Stereotaxic Coordinates from Paxinos and Watson (2007) ²⁴.

2.4. Double fluorescence immunohistochemistry of Egr1 and MOR with tyramide signal amplification (TSA)

Past research investigated the relationship between expression for the immediate early gene c-fos/c-Fos and rat play behavior ^20,25. In pilot tests we also examined c-Fos labeled cells in the MPOA and confirmed that there was no relationship between play behavior and expression of this immediate early gene (data not reported here). In contrast, our pilot studies indicated that Egr1 expression did relate to play behavior. Thus we examined Egr1 labeling in this study. Double fluorescence labeling was carried out based on our recently published research ^26,27 at room temperature unless otherwise indicated. Brain sections were washed 5×5 min with 0.02 M PBS to remove cryoprotectant, and then incubated in 1.5% H₂O₂/50% methanol for 30 min to inhibit endogenous peroxidase activity and enhance the penetration of antibodies into tissue sections. They were rinsed 3×10 min in 0.3% Triton X-100/0.05% normal goat serum (NGS)/0.02 M PBS, incubated for 60 min in blocking solution (10% NGS/0.3% Triton X-100/0.02 M PBS), and incubated overnight at 4°C with rabbit anti-MOR antiserum (#24216, ImmunoStar, Hudson, WI, USA; diluted 1:10000) in primary antibody incubation solution (PAIS, 0.3% Triton X-100/1% NGS/1% blocking reagent/0.02 M PBS). After primary antibody incubation, sections were washed in wash buffer (TBST, 0.05% Tween-20/0.15 M NaCl/0.1 M Tris-HCl, pH 7.5) and incubated for 1 hr with HRP-conjugated goat anti-rabbit antiserum (#7074, Cell Signaling Technology, diluted 1:100), washed 3×10 min with TBST, then incubated for 10 min in Cy3-conjugated tyramide (TSA™ Plus Cyanine 3 kit, PerkinElmer, Waltham, MA; red for MOR labeling). Because the MOR and Egr1 antibodies are both raised in rabbit, we blocked cross-reactivity by heating sections in a citric acid buffer (10 mM, pH 6.0) at 98°C for 5 min. Heated sections were washed in TBST 3×5 min, blocked for 1 hr using the blocking solution as described above, and incubated with rabbit anti-Egr1 (15F7, #4153, Cell Signaling Technology, Beverly, MA, USA; diluted 1:1000) antiserum overnight at 4°C. Following the second primary incubation, sections were washed 3×10 min in TBST, incubated in HRP-conjugated goat anti-rabbit antiserum for 1 hr, washed again in TBST 3×10 min, and labeled by incubation in Alexa Fluor 488-conjugated tyramide (Alexa Fluor 488 TSA kit; Molecular Probes, Eugene, OR; green for Egr1 labeling). Sections were then washed 3×10 min in TBST, mounted on subbed slides, allowed to dry in a dark room for 24 hrs, and coverslipped using Serva DePeX (Crescent Chemical Company, Islandia, NY). The concentration of antibodies used was determined after pilot testing with serial dilutions. As controls, quenching of HRP activity prior to incubation with Cy3- or Alexa Fluor 488-conjugated tyramide, or omission of the primary antibodies or HRP-linked secondary antibodies, completely abolished corresponding specific labeling.

2.5. Short hairpin RNA (shRNA) design and microinjections of AAV vectors into the MPOA

Target adeno-associated viral (AAV) vectors (i.e., AAV5-GFP-U6-Oprm1-shRNA) were produced by Vector Biolabs (Philadelphia, PA, USA). The shRNA sequences directed toward rat Oprm1 were 5′-CACC GCCTGAATCCAGTTCTTTACGCTCGAGCGTAAAGAACTGGATTCAGGC-TTTTT −3′, targeting sequence of GCCTGAATCCAGTTCTTTACG of the rat Oprm1 gene. The designed shRNA targets the common regions of all 15 known transcript variants, and thus knocks down mRNAs of all the 15 transcript variants. AAV vectors with a scrambled oligonucleotide sequence (i.e., AAV5-GFP-U6-Scr-shRNA, Vector Biolabs, Cat. #7042) with no match with any known genes were used as a negative control. Transcription of the rat Oprm1 shRNA was driven by a U6 promoter, while the reporter gene, green fluorescent protein (GFP) was driven by a CMV promoter. Based on our recent work of gene silencing in mice ²⁸, we selected serotype 5 of AAV vectors as AAV5 vectors display high transduction efficiency. The titers of the packaged viruses were 1.7×10¹³ GC/ml. The efficacy of shRNA knockdown of Oprm1 mRNA was validated in vitro (knockdown of >91% at mRNA level, performed by the vendor). We followed our protocol for microinjection of AAV vectors in the mouse brain ²⁸. A total of 24 male rats were used. In brief, pups at ages of PN15, PN17, PN19 upon arrival with lactating mothers were used. After arriving at the animal facility, pups were left undisturbed and housed with their mothers until injected and weaned from the dam at PN21. Juvenile rats were securely placed on a stereotaxic apparatus (David Kopf Instruments, Tunjunga, CA, USA) and anesthetized with isoflurane at vaporizer setting of 2%. A midline anterior to posterior incision was made. Then, the Bregma was determined and cannula (33 gauge) tip was moved to the coordinates of target site bilaterally according to The Rat Brain in Stereotaxic Coordinates ²⁴. AP=−0.2 mm, ML=± 0.4 mm, and DV=8.0 mm. AAV vectors expressing either rat Oprm1 shRNA (i.e., AAV5-GFP-U6-Oprm1-shRNA) or Scr shRNA (i.e., AAV5-GFP-U6-Scr-shRNA) were injected bilaterally into the MPOA in a volume of 0.5 uL per side. After injection, the cannula was left in place for additional 5 min. The number of rats microinjected of AAV vectors were 22 (11/11 for Oprm1 shRNA/Scr shRNA) because 2 animals died during surgery.

2.6. Social play behavior test following knockdown of Oprm1 in the MPOA

Viral vector-injected and weaned juvenile rats (PN21) were housed in mixed-sex groups of five containing animals from all treatment groups (2 Oprm1 shRNA-injected male rats and 2 Scr shRNA-injected male controls, and 1 non-AAV-injected age-matched female play partner). The social play behavior paradigm employed here was very similar to our previous work ^29,30 and those from other research groups ^31,32, in which animals were housed in mixed-sex groups, and observations of social play took place in the home cage. Although for the Egr1 study we introduced short-term isolation to enhance the motivation to play (see 2.2), here we chose home-cage testing paradigm because it is more ecologically valid. The female play partner was used for validating sex differences in social play ^29,30. Social play behavior was scored on PN35–39 (2 weeks after AAV injection) for five consecutive days in the home cage. Every morning, 2 hrs before behavior testing, the backs of the animals were labeled with a black Sharpie marker to identify individuals as above (2.2). Two hrs before the play behavior test (at 8:00 AM), animals were moved into the testing room for habituation. The light/dark cycle in the testing room was set the same as in the housing room (12:12 hrs light/dark cycle, lights off at 9:00 AM). Behavioral testing was conducted during the dark phase of the light/dark cycle. In the testing room, low levels of red illumination were provided by four 25 W red incandescent lights during behavioral testing. Animals were video recorded for two 20-min trials per day over 5 days, with the first trial occurring 1 hr after lights-off and the second trial occurring 3 hrs after lights-off, for a total observation time of 200 min per animal. The frequency of each component of play behavior below was calculated over the entire observation time. Pouncing was scored when one rat lunged and put its forepaws on the dorsal side of another rat. Pinning was scored when one rat stood over another with its forepaws on the ventral side of the opposing rat. Wrestling/boxing was scored when two animals engaged in rolling and tumbling over each other or making jabbing movements at each other with the forepaws. Biting was scored when one rat grabbed another rat’s skin with its mouth (neck or tail). Chasing was scored when one rat chased another rat. Total play bouts were the sum of each component of social play behavior, including pouncing, pinning, wrestling, biting and chasing.

2.7. Tissue collection and RNA extraction for qPCR and RNA sequencing

Twenty-four hrs after the last behavior test, rats were lightly anesthetized with isoflurane and decapitated. Brains from Scr- and Oprm1-shRNA injected rats were dissected alternating between the two groups. Brains were flash frozen in isopentane on dry ice and stored at −80°C until being sectioned on a cryostat (Leica CM1850, Bannockburn, IL, USA). Sections of 200-micron thickness were cut and mounted on gelatin-coated glass slides. MPOA from Bregma −0.24 to −0.72 mm was dissected bilaterally using a micropunch technique ³³ under a dissecting microscope. Samples were subsequently stored at −80°C until RNA extraction. Total RNA was extracted with the Aurum Total RNA Fatty and Fibrous Tissue Kit (Bio-Rad, Hercules, CA, USA) as described in our previous work ³⁴. The integrity, purity and concentration of RNA were measured as reported ³⁴. Purified total RNA was stored at −80°C until processed.

2.8. Quantitative real-time PCR (qPCR)

To confirm downregulation of Oprm1 in the MPOA, qPCR assay was carried out as in our previous work ^34,35. Briefly, 100 ng of RNA was reverse transcribed to cDNA using a SuperScript III First-Strand Synthesis System (Invitrogen, Carlsbad, CA, USA). The cDNA was then amplified using a SsoFast EvaGreen Supermix kit in a BioRad CFX96 Touch Real-Time PCR Detection System (Bio-Rad, Hercules, CA, USA). Primer sequences for genes of interest and reference genes are as follows: 5’-ATCCAGTTCTTTACGCCTTCC-3’ (forward) and 5’-GATGTTCCCTAGTGTTCTGACG -3’ (reverse) for Oprm1, 5’-GGAGCGCACGATCTTCTTCA-3’ (forward) and 5’-AGGGTGTCGCCCTCGAA-3’(reverse) for GFP, 5’-AGACAGCCGCATCTTCTTGT-3’ (forward) and 5’-CTTGCCGTGGGTAGAGTCAT-3’ (reverse) for Gapdh that was adapted from Li et al. 2015 ³⁶, 5’-CACGAACACCCTGTGGATG-3’ (forward) and 5’-GGAATGTGAACCGTTTCTGC-3’ (reverse) for Ywhaz. The Oprm1 primers adapted from Halim et al. 2018 ³⁷ amplify all 15 transcript variants of the rat Oprm1 gene. A standard curve was constructed to estimate the empirical PCR reaction efficiency, and a dissociation curve analysis was performed to insure specificity of PCR products. To yield individual relative expression level values for genes Oprm1 and GFP, mean Ct values obtained from qPCR were transformed according to the Pfaffl method ^38,39. The resulting values represent Oprm1 or GFP mRNA expression normalized against two reference genes Gapdh and Ywhaz. These relative expression values were used for statistical analysis of gene expression.

2.9. RNA sequencing

Effects of Oprm1 downregulation in the MPOA on patterns of gene expression in the MPOA were examined using RNA-Seq. Construction and sequencing of stranded RNA libraries were conducted at the University of Wisconsin-Madison Biotechnology Center Gene Expression Center & DNA Sequencing Facility. Total RNA was verified for purity and integrity with NanoDrop One Spectrophotometer and Agilent 2100 Bioanalyzer, respectively. All samples met the Illumina sample input guidelines and then were prepared using the Illumina® TruSeq® Stranded mRNA Sample Preparation kit (Illumina Inc., San Diego, CA, USA). For each library preparation, mRNA was purified from 150 ng total RNA using poly-T oligo-attached magnetic beads. Subsequently, each poly-A enriched sample was fragmented using divalent cations. Fragmented RNA was synthesized into double-stranded cDNA using SuperScript II Reverse Transcriptase (Invitrogen, Carlsbad, CA, USA). Random primers were used for first strand cDNA synthesis followed by second strand cDNA synthesis using DNA Polymerase I. Double-stranded cDNA was treated with RNase H to remove mRNA and purified by paramagnetic beads (Agencourt AMPure XP beads, Beckman Coulter). The cDNA products were incubated with Klenow DNA Polymerase to add an ‘A’ base (Adenine) to the 3’ end of the blunt DNA fragments. DNA fragments were ligated to Illumina unique dual adapters with a single ‘T’ base (Thymine) overhang on the 3’end. Adapter ligated DNA was amplified in a Linker Mediated PCR reaction (LM-PCR) for 12 cycles using Phusion™ DNA Polymerase and then purified by paramagnetic beads. Six additional PCR cycles were added following low yields across the sample set. Quality and quantity of the finished libraries were assessed using an Agilent HS DNA chip (Agilent Technologies, Santa Clara, CA, USA) and Qubit® dsDNA HS Assay Kit (Invitrogen, Carlsbad, CA, USA), respectively. Libraries were standardized to 2 nM. Paired-end 2x150 bp sequencing was performed, using standard SBS chemistry (v3) on an Illumina NovaSeq6000 sequencer. Images were analyzed using the standard Illumina Pipeline, version 1.8.2. One sample from the knockdown group had extremely low raw reads compared to all other samples and was identified as an outlier using multiple tools, including unsupervised clustering with multidimensional scaling and a clustered image map as well as WGCNA analysis. Consequently, this sample was removed from data analysis. The data for the outlier sample is included as part of the GEO dataset. The final samples for RNA-Seq analysis were Scr shRNA control (n=7); Oprm1 shRNA knockdown (n=5). Expression was normalized using RSEM ⁴⁰. The average for primary reads was 65 672 025 with a range from 47 643 081 to 84 155 698. The average primary mapping rate was 87.73%.

2.10. Differential gene expression analysis

We conducted differential gene expression analysis of RNA-Seq data using the EdgeR Bioconductor Package, v. 3.9 ⁴¹. Using the standard setting of a minimum of 10 counts in at least five of the samples (the size of the smaller of the two groups) as the criterion for gene inclusion, a total of 13,125 genes were included in the analysis. Statistics, including false discovery rates, are provided in Supplementary Table S1.

2.11. Identification of genes associated with Oprm1 knockdown via machine learning

Using the machine learning tool, Weka ⁴², and the SVMAttributeEval approach in the Select attributes feature, we identified the top 50 genes that were found to be most useful in machine learning classification of the dataset when using the support vector machine approach.

2.12. Weighted gene co-expression network analysis (WGCNA) analysis

We used the WGCNA systems biology approach to construct gene co-expression networks and gene modules from gene expression datasets ⁴³. Prior to analysis, data were log₂ +1 transformed and genes with sparse expression were removed. WGCNA was run on 14,579 genes using R software. To generate a weighted network of genes (nodes) and their expression correlations (edges), correlations were raised to a soft thresholding power β of 12. We used unsupervised hierarchical clustering, a minimum module size of 30 genes, the signed mode, the deepSplit parameter set to 2, the mergeCutHeight parameter set to 0.15, and a threshold setting for merging modules of 0.25. Module eigengene values were also evaluated in terms of their relationships with the various behavioral traits. Some modules were exported as a Cytoscape network file and manually trimmed to consist of genes of interest and their gene-to-gene correlations prior to visualization with Cytoscape v3.7.2 ⁴⁴.

2.13. Enrichment analysis

Enrichment analysis was run on the gene set produced by differential analysis and on the gene modules produced by WGCNA. This method identifies genes that are over-represented in the gene set or modules that have been associated with particular functions, drug actions or other phenotypes. Enrichment tools included ToppCluster ⁴⁵, GeneOntology ⁴⁶, and EnrichR ⁴⁷.

2.14. Quantification of Double Fluorescence Labeling

All fluorescence-labeled images were captured sequentially using an inverted Zeiss LSM 710 Meta laser scanning confocal microscope (Zeiss; Oberkochen, Germany). All images in each sample area were acquired with a screen resolution of 1024×1024 pixels using a 40× objective lens. For quantitative analysis of single-labeling or colocalization, cell counting was carried out using 40× magnification photomicrographs throughout the rostrocaudal extent of the MPOA. The number of single-labeled cells characterized by clearly stained somata for MOR-expressing cells and nuclei for Egr1-immunoreactive cells was counted bilaterally in every other section in each animal. The counting area was a 460 μm × 540 μm unit, which covered the entire MPOA. Simultaneously, double-labeled cells were also enumerated based on the coincidence of red-labeled somata of MOR and green-labeled nuclei of Egr1 in the same cells. A total of 4–5 sections were measured in both the rostral and caudal MPOA of each animal. The counting was performed manually by an experienced investigator blind to the labeling conditions using the MetaMorph software (Molecular Devices, LLC, San Jose, CA, USA). All confocal images were transferred to Adobe Photoshop 6.0 (Adobe Systems; San Jose, CA), with adjustments of brightness and contrast. The number of single- or double-labeled cells from bilateral MPOA per rat was averaged. Two animals of each group were used for pilot experiments to determine optimal staining conditions. The measures from 8 animals per group were averaged to obtain the final mean±SEM.

2.15. Statistical analysis

Statistical analyses were performed using GraphPad Prism 8 software (GraphPad Software, San Diego, CA, USA). All values were expressed as mean±SEM. Data on Egr1 and MOR were analyzed using two-way analysis of variance (ANOVA) with manipulation (2 levels: play and no-play) × region (3 levels: rostral, caudal and entire MPOA) as between-subject factors. When the overall significant effects were detected, multiple group differences were assessed using Bonferroni’s post hoc test for pairwise comparisons of means. Two-group (e.g., no-play and play) comparisons in social play behavior and gene expression were analyzed using multiple unpaired t tests. Correlations were run to assess relationships between Egr1 expression (i.e., number of Egr1 immunoreactive cells) and social play behaviors in play animals. To ensure an unbiased observation of various components of social play behavior, behaviors were scored by an experienced investigator unaware of the experimental manipulations. Overall level of statistical significance was set at p<0.05.

3. Results

3.1. Social play behavior increased Egr1 expression in the MPOA

Prior to the analysis of Egr1 expression, we confirmed that all the rats in the play group did play across the testing session (Fig. 1). To investigate whether MPOA is involved in social play behavior in adolescent male rats, we examined the expression of the immediate early gene, Egr1, a marker for cellular responses following exhibition of social play. Immunofluorescent labeling of Egr1 showed that Egr1-immunopositive cells were sporadically expressed within the entire MPOA (Fig. 2A,C) in adolescent male rats who did not experience play, while social play induced robust Egr1 expression in both rostral (Fig. 2B) and caudal (Fig. 2D) MPOA. It appears that a greater number of Egr1-positive cells was induced in caudal MPOA than in rostral MPOA after social play (Fig. 2B,D), suggesting that the rostral and caudal regions may have different roles in social play. Additionally, Egr1-positive cells were centered on the sexually dimorphic nucleus of the caudal MPOA (Fig. 2D), where the most intense signal for MOR was located ⁴⁸.

Figure 1:

Social play behavior in adolescent male rats after isolation for 24 hrs and reunion with play partners. Note that play rats (N=10) displayed high frequency of pouncing, pinning, wrestling, chasing and total play bouts except for biting during the 20-min testing session.

Figure 2:

Immunofluorescence labeling of Egr1-expressing cells in the rostral (A,B) and caudal MPOA (C,D) of no-play (A,C) and play (B,D) rats. Note that in both rostral and caudal MPOA, social play largely increased Egr1 expression (B,D) when compared to the no-play adolescent male rats (A,C). Oval-shaped areas indicate that Egr1-positive cells were centered on the sexually dimorphic nucleus. Scale bar: 500 μm. Egr1, early growth response 1; MPOA, medial preoptic area.

Two-way ANOVAs revealed a significant effect of manipulation (play or no-play) [F(1,42)=49.99, p<0.0001] and a significant effect of region (rostral, caudal or entire MPOA) [F(2,42)=5.11, p=0.01] but no significant interaction between manipulation and region. Bonferroni’s multiple comparisons tests showed that social play enhanced the number of Egr1-expressing cells in rostral (Fig. 3A, p=0.0083), caudal (Fig. 3C, p<0.0001), and the entire (Fig. 3E, p=0.001) MPOA. Furthermore, more Egr1 cells were induced in caudal MPOA than in rostral MPOA (Fig. 3A,C; p=0.0071). However, comparisons of MOR-expressing cells revealed no significant effect of manipulation, region or interaction (Fig. 3A,C,E).

Figure 3:

Quantitative analysis of expression and colocalization of cells expressing immunoreactivity for Egr1 and MOR within the rostral (A,B), caudal (C,D) and entire (E,F) MPOA following social play. Note that the number of Egr1-expressing cells and cells double labeled with Egr1 and MOR was greatly increased in the rostral (A), caudal (C) and entire (E) MPOA of play animals, while no changes were observed in MOR-expressing cells (A,C,E), suggesting that MOR cells were responding to social play. Further, the proportion of double labeled cells divided by the MOR cells was also increased in the rostral (B), caudal (D) and entire (F) MPOA of play animals, with no changes when divided by the Egr1 cells (B,D,F). ** P < 0.01, *** P < 0.001, play group versus no-play group.

In analysis of the animals in the play group, no significant correlations were found between any of the components of social play behavior, including pouncing, pinning, wrestling, biting, chasing and total play bouts and Egr1 expression in the rostral, caudal and entire MPOA (all p values>0.25).

3.2. Social play activated MOR-expressing cells in the MPOA

To evaluate whether MORs in the MPOA participate in social play behavior, we examined the degree to which MOR-expressing cells related to social play with double fluorescence labeling of MOR and Egr1. Consistent with previous findings ^48,49, we observed abundant expression of MORs in the MPOA. Differing from previous observations showing that the majority of MOR immunoreactivities were fibers ⁵⁰, we found many MOR-immunoreactive neurons featured with clearly stained cell bodies (Fig. 4B), which allowed for reliable counting of double labeled cells. Double fluorescence labeling showed that a portion of MOR-immunoreactive cells co-expressed Egr1 after social play (Fig. 4D–F). In contrast, very few MOR-expressing cells were colocalized with Egr1 in no-play animals (Fig. 4A–C) due to the low baseline level of Egr1.

Figure 4:

Representative photomicrographs of double immunofluorescence labeling of cells expressing immunoreactivity for Egr1 (A,D) and MOR (B,E) in the MPOA of no-play (A-C) and play (D-F) rats. Note that a portion of MOR-immunoreactive cells co-expressed Egr1 after social play (F). Arrowheads indicate single labeled MOR-expressing cells; arrows indicate cells co-expressed Egr1 and MOR. Scale bar: 50 μm. Egr1, early growth response 1; MOR, opioid receptor, mu 1; MPOA, medial preoptic area.

Two-way ANOVAs revealed a significant effect of manipulation on the number of double labeled cells [F(1,42)=37.83, p<0.0001], but no significant effect of region or interaction between manipulation and region. Bonferroni’s multiple comparisons test revealed that the number of cells double labeled with Egr1 and MOR was greatly increased within the rostral (Fig. 3A, p=0.0048), caudal (Fig. 3C, p=0.0013) and entire (Fig. 3E, p=0.0039) MPOA following social play. Further analysis with two-way ANOVAs revealed a significant effect of manipulation on the proportion of double labeled cells divided by the MOR cells (the number of double labeled cells divided by the number of MOR-immunopositive cells) [F(1,42)=53.59, p<0.0001], but no significant effect of region or interaction between manipulation and region. Bonferroni’s multiple comparisons test revealed that the proportion of double labeled cells divided by the MOR cells was greatly increased within the rostral (Fig. 3B, p=0.0002), caudal (Fig. 3D, p=0.0007) and entire (Fig. 3F, p=0.0004) MPOA following social play. However, no changes in the proportion of double labeled cells were observed when divided by the Egr1 cells (the number of double labeled cells divided by the number of Egr1-immunopositive cells) (Fig. 3B,D,F).

3.3. Verification of the specificity and efficacy of Oprm1 knockdown in the MPOA

In our pilot study, fluorescent immunolabeling of viral vector-injected (i.e., Scr shRNA or Oprm1 shRNA AAV vectors) brain sections with anti-GFP antibodies (i.e., the protein product of reporter gene, GFP) showed that GFP-immunoreactive cells were expressed in the MPOA, with minimum or no spread to the neighboring and distant areas (Fig. 5A,B), indicating that delivery of viral vectors (i.e., injection site) was specific to the MPOA. To ensure an efficient transduction of shRNA into the MPOA neurons, we, furthermore, carried out quantitative analysis of mRNA levels of GFP by qPCR as described recently ²⁸. We adopted a Cq<32 cutoff value to ensure a successful injection and efficient transduction of the viral vectors in the MPOA. Only animals injected with viral vectors that displayed a Cq value <32 measured in qPCR assay were included for gene and behavioral data analysis (N=7 for Scr shRNA, N=6 for Oprm1 shRNA).

Figure 5:

Verification of the viral transduction (A,B) and knockdown of Oprm1 (C) in the MPOA of shRNA-injected rats. Immunofluorescence labeling with GFP that serves as a reporter to label the transduced cells and quantitative real-time PCR (qPCR) analysis of expression changes of Oprm1 in the MPOA of shRNA-injected rats were performed. Note that a large number of cells in the MPOA were identified to express GFP (A), suggesting a successful injection and efficient transduction of the viral vectors in the MPOA. High magnification of photomicrograph (B) shows typical GFP-immunolabeled cells (arrows). Scale bars: 200 μm in A, 50 μm in B. GFP, green fluorescent protein. ** P < 0.01, Scr shRNA group versus Oprm1 shRNA group. Scr shRNA, scrambled short hairpin RNA.

qPCR analysis identified that mRNA level of Oprm1 in the MPOA was significantly downregulated in Oprm1 shRNA-injected group [t(11)=3.46, p=0.0054] (by 19.6%) as compared to Scr shRNA-injected control group (Fig. 5C). To further assess the magnitude of the observed differences between groups, the effect sizes were calculated according to the Cohen’s term d. The value of Cohen’s d was 1.98. The 19.6% downregulation is very similar to our recent study, which has been demonstrated to have significant behavioral effects ²⁸.

3.4. Knockdown of Oprm1 in the MPOA decreased social play behavior in adolescent male rats

To determine how MORs in the MPOA govern social play behavior, we knocked down Oprm1 gene expression using shRNA-mediated gene silencing of MOR. Overall, Oprm1 shRNA-injected adolescent rats displayed less social play than that of Scr shRNA-injected control animals. Specifically, knockdown of Oprm1 reduced pouncing, one major component of initiation of play behavior. Statistical analysis revealed that decreases in Oprm1 gene expression in the MPOA reduced the frequency of pouncing (p=0.018) and total play bouts (p=0.046), but had little effect on pinning, wrestling, biting and chasing (Fig. 6).

Figure 6:

Effects of Oprm1 knockdown in the MPOA on social play. Note that decreases in Oprm1 gene expression reduced the frequency of pouncing and total play bouts, while had little effect on pinning, wrestling, biting and chasing. * P < 0.05, Scr shRNA group versus Oprm1 shRNA group. Scr shRNA, scrambled short hairpin RNA.

3.5. Differential gene expression between groups

We evaluated differential gene expression between Oprm1 knockdown and control groups and found 349 genes with a p-value less than 0.05, including the immediate early gene, Egr1, the glutamate receptor subunits Grina and Grik3, the somatostatin receptor SStr1, and the neuropeptide Avp (Supplementary Table S1). While none of these reached significance using a False Discovery Rate correction procedure, in prior studies ⁵¹ we have found qPCR validation of top differential expressing genes at this p value level and thus we expect the differences may be biologically meaningful. Enrichment analysis tools allow genes of interest (e.g., genes with significantly altered expression) to be evaluated against preexisting datasets to determine whether an over-representation occurs within a range of biological traits. Here, analyses revealed that these genes included an enrichment for the synapse, suggesting alterations in CNS function. When separately examining up and down regulated genes in EnrichR, we found that upregulated genes in rats treated with shRNA relative to controls match genes that are upregulated when Egr1 is missing in mice (adjusted p value=0.006), suggesting downregulation of Egr1 could underlie some of the expression changes. Further, in EnrichR upregulated genes with Oprm1 knockdown match genes downregulated when mice are exposed to morphine (adjusted p value=0.0008). For all expression results, see Supplementary Table S1. The results of differential gene expression have been uploaded to NCBI’s Gene Expression Omnibus (GEO) with accession number GSE148007.

3.6. Connections of brain expression profiles with treatment group

We used the Select attributes feature within the machine learning tool Weka to identify the top genes that allow for identification of the two groups. Among the top genes identified by SVMAttributeEval were Egr1, Dpf1, Ost4, Pde3b, Rasa12, Atf7, and Avpr1a.

3.7. WGCNA results

WGCNA identified 114 modules of genes (Supplementary Table S1). Module creation in WGCNA is unsupervised and independent of group or trait identification. Relationships between these modules and group and traits were then evaluated. A full listing of the modules, the member genes, and how those modules relate to traits is provided in Supplementary Table S1. The orangered4 module was the most significantly enriched for group (control v. knockdown) and Egr1 was within this module. Interestingly, 9 genes in this module were also part of the top 50 genes for attribute selection in Weka, including Egr1, Zdhhc24, Dpf1, Pbdc1, Fam129a, Rida, and Avpr1a. When just looking at module association with play behavior, the orange module was significantly linked to pouncing, wrestling, biting, and chasing. Within this module is the neuropeptide, Vip (vasoactive intestinal peptide), that has been previously linked with play behavior ⁵². Fig. 7 shows genes within the orange module that interact with Vip. Additional modules that were significantly connected to some aspect of play were: thistle, orangered3, and skyblue2. The Oprm1 gene was within the skyblue3 module.

Figure 7:

Coexpression network analysis using WGCNA. (A) Egr1 is connected to a large subset of genes. (B) Oprm1 is identified as the most closely linked gene to the module.

4. Discussion

Opioids in the MPOA of songbirds act to stimulate and maintain affiliative, non-sexual vocal-social interactions ^1,53,54. The present study extended research findings on this type of affiliative, rewarding social behavior from songbirds to rats. Using an immediate early gene (IEG) expression approach, we found that social play behaviors were associated with a robust increase in expression of Egr1 in the MPOA. We also determined that cells expressing MORs in the MPOA displayed increased numbers of cells colocalized with MOR and Egr1 when adolescent rats expressed social play using double immunofluorescence labeling of MOR and Egr1. Furthermore, using shRNA-mediated gene silencing we revealed that knockdown of Oprm1 in the MPOA suppressed social play behavior in male adolescent rats. Taken together, these findings for the first time offer strong support to the notion that opioids in the MPOA govern social play behavior. Moreover, results indicate that studies of the MPOA in songbirds are uncovering a central nucleus that is part of a core, conserved neural circuit in which opioids act to facilitate important social behaviors in contexts in which behavior is considered to be intrinsically rewarding.

4.1. Social play increases expression of immediate early gene Egr1 in the MPOA

Expression of IEGs such as c-Fos, Egr1, Arc and c-Jun is a powerful tool for delineating the neural circuitries that may mediate drug effects and behavioral responses ^55–57. Induction of IEGs is thought to reflect the functional cellular responses associated with the stimulus. In this study, we observed a robust increase in Egr1 expression in the MPOA following social play, suggesting the involvement of MPOA in play behavior. Most prior studies on social play that used the c-fos marker did not investigate the MPOA ^20,25 and the one study that included this region showed no c-Fos induction in the MPOA after performance of play ¹⁶. This apparent discrepancy may be explained by two factors. First, we used Egr1 as a marker for cellular responses instead of c-fos/c-Fos. Although the fos gene and its protein product c-Fos are the most commonly used marker for identifying cellular responses and neural circuitries, the absence of c-fos/c-Fos expression should not be necessarily interpreted as a lack of activation for a neural site in a functional circuitry ^58–60. It has been reported that the pattern of induction of the IEGs in certain brain regions in the same experimental context can be unique to the markers measured ^57,61,62. For instance, a low or undetectable signal for c-fos was expressed in the nucleus accumbens and the basolateral amygdala in response to immobilization stress, while Arc (activity regulated cytoskeletal associated protein) was highly induced ⁵⁷. Hence, use of multiple IEGs helps to compensate for missing changes of one particular IEG. Second, the inconsistent results may be attributed to the methodological differences as well. We adopted a robust test paradigm for social play. Animals were first socially isolated for 24 hrs to enhance the motivation to play ⁶³ and then group-housed (4 males and 1 female) for a 20 min-session of play. Thus, play behavior in this study occurred in group-housed animals, which differs from the pair-housed rats (i.e., experimental animal and play partner) tested in most previous studies ^20,25. It is still not clear whether exposure to multiple play partners enhances social play. Subsequent studies are needed to determine whether play behavior is potentiated after social isolation in group-housed rats as compared to pair-housed ones. We did not find Egr1 expression was correlated with any components of social play in the rostral, caudal and entire MPOA in play animals. One plausible explanation is the ceiling effect that there is no room for an increase in Egr1 expression due to the high level of play. It could also be that exposure to social partners alone (even in the absence of play) explains the increase in Egr1; however, the results of the shRNA manipulations demonstrate a causal role for Oprm1 in the MPOA specifically in play behavior.

4.2. Opioids in the MPOA facilitate social play

The Egr1 expression study suggests a role for the MPOA in social play. However, it is still not determined which signaling pathways within the MPOA were responding. Since opioids in songbirds regulate affiliative social behavior and modulate social play in juvenile rats, we predicted that opioid neurotransmission in the MPOA may play an important role in social play. We employed a double labeling technique, which allows for a detailed analysis of changes in brain MOR activity with high anatomical resolution to identify whether the MPOA was an important neural site of action for opioid modulation of this behavior. As predicted, relative to controls, in rats in the play group we indeed found an increase in the number of cells double-labeled with MOR and Egr1, which was accompanied by no changes in the number of MOR cells (Fig. 3A,C,E) , strongly suggesting that social play is accompanied by activation of MOR-containing cells in the rat MPOA. Consistent with this finding, the increased proportion of double cells when divided by MOR, but not Egr1 in play compared to the no-play animals (Fig. 3B,D,F), further supports an increase in activation of MOR-containing cells in the rat MPOA after social play. Because the increase in the percentage of double labeled MOR cells is larger in play compared to the no-play animals, it is believed that cellular activity in only a small portion of MOR-expressing cells may be sufficient to produce behavioral effects.

Despite the suggestive involvement of an opioid signaling pathway in the MPOA in social play, the IEG expression does not provide sufficient information on the causal nature of MOR responses. To determine the extent to which increased activity of MOR is facilitatory or inhibitory, we knocked down Oprm1 gene expression in the MPOA, and observed that decreases in Oprm1 suppressed social play. These findings contrast largely with those of previous lesion studies, which showed no reliable disruptive effects ^17,19. However this is identical to what has been observed in studies of affiliative birdsong in which siRNA downregulation of MOR in the MPOA suppressed singing behavior (our unpublished observations); whereas, lesions either increased or had no effects on song ⁴. These similarities further support the hypothesis that studies of affiliative birdsong and rat play reveal a conserved neural circuit that underlies intrinsically-rewarded social behaviors.

The present results demonstrate that knockdown of Oprm1 in the MPOA reduced the frequency of total play bouts and pouncing, one major component of initiation of play behavior, but had little effect on pinning, wrestling, biting and chasing. However, it is up to now not clear whether the reduction in pouncing but not other aspects of social play behavior is indicative of a selective involvement of MOR activity in the MPOA in play initiation. Because there is a trend towards a reduction in other components of social play after Oprm1 knockdown in the MPOA (see Fig. 6), it is likely that other aspects of social play behavior might have been affected as well if sufficient interference in MOR activity in the MPOA was achieved, leaving open the possibility of non-specific effects of MOR in the MPOA on social play behavior.

It is assumed that opioids exert their regulatory effects on social play primarily via altering rewarding aspects of play rather than play motivation ^12,64. Infusion of the opioid met-enkephalin into MPOA induces reward in rats ⁶⁵, and the MPOA has been implicated in sexual reward ^6,8. Thus we interpret our results as evidence that MPOA MOR contribute to the rewarding properties of play behavior, similar to what has also been observed for NAc ¹².

4.3. RNA-Seq highlights a central role in play for Oprm1 and Egr1 in MPOA

Several genes involved in neuronal signaling exhibited altered expression due to Oprm1 knockdown, and these included glutamate receptors, the somatostatin receptor, and the neuropeptide, vasopressin. Interestingly, the transcription factor and immediate early gene, Egr1, also showed significant downregulation with Oprm1 knockdown. Using machine learning approaches, Egr1 was also identified as one of the top genes that could help predict the knockdown versus control group. Additionally, WGCNA identified the orangered4 module as being most closely linked to group and Egr1 was in this module along with nine other genes that were identified by machine learning tools as being predictive of group, including Zdhhc24, Dpf1, Pbdc1, Fam129a, Rida, and Avpr1a. Unexpectedly, when up and down regulated genes were analyzed separately, a significant overlap was found between genes upregulated with Oprm1 knockdown and genes upregulated with Egr1 knockout mice. This finding suggests the possibility that Oprm1 knockdown leads to lower expression of the transcription factor, Egr1, and in turn the lower Egr1 produces additional expression changes. Together, the RNA-Seq findings suggest a complex interaction with Oprm1 and Egr1. Activation of Oprm1 usually leads to activation of inhibitory pathways ⁶⁶, while Egr1 expression is usually triggered by an elevation of excitability ⁶⁷. One scenario could involve downregulation of Oprm1 in GABA neurons whereby the loss of inhibition in turns leads to elevated GABA release and inhibition in downstream neurons (and thus decreased Egr1 expression). As seen in the double labeling studies, a subset of Oprm1-positive neurons shows elevated Egr1 with play behavior and the co-expression of Oprm1 and Egr1 is also borne out by recent single neuron RNA-Seq studies from the Allen Brain Institute (https://portal.brain-map.org/atlases-and-data/rnaseq). Whether or how decreases in Oprm1 affect Egr1 within the same neurons or in downstream neurons remains to be determined.

Conclusions

The present study for the first time provides evidence supporting a critical role for the MPOA in rat social play behavior. Given that social play in mammals is a highly rewarding, motivated, affiliative behavior ^11,68–70, and the MPOA is highly conserved (i.e., it is neuroanatomically, neurochemically and functionally similar) in birds and mammals ^71,72, we suggest that studies of MPOA, birdsong, and rat play are uncovering a central nucleus that is part of a core, conserved neural circuit across species in which opioids act to stimulate and maintain intrinsically-rewarded social behaviors.

Supplementary Material

Table S1

Table S1: Differential gene expression and WGCNA module analysis after Oprm1 knockdown in the MPOA