Posterior Amygdala Regulates Sexual and Aggressive Behaviors in Male Mice

Male PA cells that project to MPN and VMHvl are largely distinct

Since MPN and VMHvl contribute to different social behaviors, we hypothesized that PA neurons that project to MPN and VMHvl likely non-identical and play differentially roles in male sexual and aggressive behaviors. To test this, we first injected a retrograde tracer into the MPN or VMHvl of male mice and found that labeled cells distribute differently within PA. MPN-projecting PA cells (PA^MPN) are abundant in the anterior dorsomedial part (Fig. 1a, b, e), whereas VMHvl projectors (PA^VMHvl) concentrate in ventrolateral PA throughout the anterior-posterior axis (Fig. 1c–e). Additionally, both PA^MPN and PA^VMHvl cells preferentially overlap with Esr1⁺ cells: nearly all PA^MPN cells express Esr1 (95.5 ± 1.1%, n = 3 animals) while approximately half of PA^VMHvl cells contain Esr1 (55.6 ± 2.5%, n = 3 animals; Fig. 1f).

To compare PA^MPN and PA^VMHvl cells in the same animals, we injected Alexa647 conjugated CTB into the VMHvl and Alexa555-CTB into the MPN and found the PA^VMHvl and PA^MPN cells are largely distinct at the single cell level. Among all PA cells, 16.7 ± 1.5% are retrogradely labeled from MPN, 22.2 ± 3.0% from VMHvl and only 3.3 ± 0.7% are labeled from both regions (n = 3 animals). The dual-labeled cells are largely located in the middle zone where PA^MPN and PA^VMHvl cells converge (Fig. 1g–j). To further visualize MPN-projecting PA^Esr1+ (PA^Esr1+MPN) and VMHvl-projecting PA^Esr1+ (PA^Esr1+VMHvl) cells, we injected a retrograde herpes simplex virus (HSV) encoding a Cre-dependent red fluorescent protein (HSV-hEf1α-LS1L-mCherry) into MPN and HSV-hEf1α-LS1L-EYFP into VMHvl of Esr1–2A-Cre male mice (Fig. 1k). Consistent with CTB labeling, PA^Esr1+MPN and PA^Esr1+VMHvl cells show largely distinct spatial distribution (Fig. 1l–o) and overlap minimally at the single cell level (PA^Esr1+MPN/total labeled cells: 54.3 ± 0.2%; PA^Esr1+VMHvl/total: 39.6 ± 0.8%; Overlap/total: 6.1 ± 0.6%; PA^Esr1+MPN/ PA^Esr1+: 21.5 ± 1.7%; PA^Esr1+VMHvl/ PA^Esr1+: 15.6 ± 0.6%; Overlap/ PA^Esr1+: 2.4 ± 0.6%; Predicted Overlap by chance: 3.4 ± 0.4%; paired t-test between Predicted Overlap and Observed overlap: p = 0.016; n = 3; Fig. 1n,o). Additionally, we quantified all retrogradely labeled Esr1⁺ cells in the amygdala and found that PA contains the largest fraction of labeled Esr1⁺ cells in both MPN- and VMHvl-targeted animals, pointing to PA as a major source of inputs to the medial hypothalamus (Extended Data Fig. 1).

Male PA^Esr1+ cells provide monosynaptic excitatory inputs to MPN^Esr1+ and VMHvl^Esr1+ cells

Consistent with previous reports^28,34, we found that PA consists of mainly excitatory neurons (92.3%, n = 2 mice) using Vglut1-ires-Cre × Ai6 (a ZsGreen reporter mouse) and Vgat-ires-Cre × Ai6 male mice (Extended Data Fig. 2a–c). Histological analysis revealed that approximately 50% of PA cells express Esr1 and over 95% of PA^Esr1+ cells are glutamatergic (Extended Data Fig. 2c).

We next examined whether PA^Esr1+ cells provide direct synaptic inputs to MPN and VMHvl cells, especially the social behavior relevant Esr1⁺ cells^5,7,29,30,35. We injected AAV-Ef1α-DIO-ChR2-EYFP into the PA and AAV-hSyn-DIO-mCherry into MPN or VMHvl of Esr1–2A-Cre male mice (Fig. 2a). Three weeks later, we performed whole-cell voltage-clamp recordings from mCherry⁺ cells in MPN or VMHvl while stimulating PA terminals using 0.5-ms blue light pulses (Fig. 2b). We observed optogenetically-evoked excitatory currents (oEPSCs) as well as inhibitory postsynaptic currents (oIPSCs) in the majority of recorded cells in MPN (53.1%) (Fig. 2b–e). oEPSCs featured a short latency (2.24 ± 0.18 ms; n = 25 cells) while oIPSCs displayed significantly longer latencies (8.58 ± 1.21 ms; n = 19 cells), suggesting that the oEPSCs are likely monosynaptic while oIPSCs are polysynaptic (Fig. 2e). Consistent with this hypothesis, bath application of CNQX, an AMPA receptor blocker, abolished both oEPSCs and oIPSCs (Fig. 2f–h).

Similar to the Esr1⁺ cells in MPN, two thirds of VMHvl^Esr1+ cells showed short-latency oEPSCs (2.06 ± 0.20 ms; n = 23 cells) followed by long-latency oIPSCs (10.43 ± 1.40 ms; n = 20 cells) upon 0.5 ms light stimulation (Fig. 2i–l). CNQX application abolished both oEPSCs and oIPSCs (Fig. 2m–o). Altogether, these results suggest that PA^Esr1+ cells provide direct excitatory and indirect inhibitory inputs to both MPN^Esr1+ and VMHvl^Esr1+ cells.

Male PA subregions that project to MPN and VMHvl display differential molecular phenotypes

Given the spatial segregation between PA^MPN and PA^VMHvl cells, we wondered whether these two populations also differ molecularly. As a first clue, a higher percentage of PA^MPN cells overlap with Esr1 than that of PA^VMHvl cells (Fig. 1f). To address this question more systematically, we injected red and green retrobeads into MPN and VMHvl of C57BL/6N male mice, respectively, and used later capture microdissection to collect fluorescence-intense PA subregions as well as the adjacent basomedial amygdala posterior part (BMAp), and then performed RNA-sequencing (Fig. 3a). Principal component analysis of RNAseq results revealed that samples from the same PA subregion clustered together and apart from other regions (Fig. 3b). We found 694 genes, including Esr1, enriched in PA compared to BMAp (Fig. 3c, Supplementary Table 2). Additionally, 172 genes were found at a higher level (fold change > 1.5) in the MPN-projecting subregion of PA compared to the VMHvl-projecting subregion, and 215 genes showed the opposite pattern (Fig. 3e, Supplementary Table 2). Then, we performed in situ hybridization to visualize mRNA expression of several candidate genes enriched in PA³⁶. Consistent with the RNAseq data, Pde11a, Prokr2 and Zic2 were expressed at high levels in the entire PA but not in the neighboring BMAp (Fig. 3d). Chrna7, Npy2r and Nnat predominantly localized to the MPN-projecting dorsomedial part of PA whereas Etv1, Calb2 and Dlk1 were largely confined to the VMHvl-projecting ventrolateral part of PA (Fig. 3f). Double in situ hybridization of Chrna7 and Dlk revealed their largely non-overlapping expression in PA (Dlk⁺: 27.5 ± 2.5 cells/section; Chrna7⁺: 71.8 ± 2.8 cells/section; Chrna7⁺Dlk⁺: 2.0 ± 0.3 cells/section; n = 4 sections from 2 animals). Lastly, we combined retrograde tracing and in situ hybridization and confirmed that over 90% of PA^MPN cells expressed Chrna7 whereas over 80% of PA^VMHvl cells expressed Dlk1 (Fig. 3g–i). Altogether, these data suggest that two largely distinct PA subpopulations with different gene expression and hypothalamic projection patterns exist in male mice.

PA^Esr1+MPN and PA^Esr1+VMHvl cells respond differentially during male sexual behaviors and aggression

To understand whether Esr1 marks the social behavior relevant population in PA, we performed double staining of c-Fos and Esr1 in animals that experienced mating and fighting. 74% and 67% of mating and fighting induced c-Fos⁺ cells contained Esr1, respectively, significantly higher than chance (~50%) (Extended Data Fig. 3a–e). Additionally, c-Fos⁺ cells induced by male mating and fighting distributed to different subregions of PA: while mating-induced c-Fos was concentrated in the anterior half of PA, fighting-induced c-Fos distributed evenly along the anterior-posterior axis of PA (Extended Data Fig. 3f). As mating and fighting induced c-Fos expression patterns matched with MPN and VMHvl retrograde labeling patterns in the PA, respectively, we next investigated the potential differential responses of PA^Esr1+MPN and PA^Esr1+VMHvl cells in male sexual and aggressive behaviors.

We performed fiber photometric recording in Esr1–2A-Cre male mice by injecting HSV-hEf1α-LSL-GCaMP6f into the MPN or VMHvl and implanted an optic fiber above PA (Fig. 4a, b, Extended Data Fig. 4)^7,37,38. Histological analysis revealed that majority of GCaMP6f⁺ cells overlapped with Esr1 (93.5 ± 1.0% cells, n = 3 animals; Fig. 4c). Consistent with the fact that a higher percentage of PA^MPN cells express Esr1, we observed a higher number of GCaMP6f⁺ cells in PA in animals with virus injected into MPN than in VMHvl (PA^Esr1+MPN: 310 ± 31 cells/animal; PA^Esr1+VMHvl: 171 ± 13 cells/animal, n = 5 animals each).

We found that PA^Esr1+MPN and PA^Esr1+VMHvl cells showed complementary response patterns during male-male and male-female encounters (Fig. 4d–z). PA^Esr1+MPN neurons showed a greater increase in Ca²⁺ upon female introduction than male introduction as well as higher responses during female investigation than male investigation (Fig. 4f–i, z). By contrast, PA^Esr1+VMHvl cells showed higher responses to males than females, during both entrance and investigation (Fig. 4q–t, z). Complementary responses in PA^Esr1+MPN and PA^Esr1+VMHvl cells persisted throughout the consummatory phase of social behaviors: PA^Esr1+MPN cells showed stronger signals during male mating than attacking, while PA^Esr1+VMHvl cells showed significantly higher responses during attacking than mating (Fig. 4j–n, u–z). We noticed that Ca²⁺ increases of PA^Esr1+MPN cells often preceded mounting onset, which was defined as the moment when the male first clasped its forelimbs around the female’s body (Fig. 4j). To understand the precise temporal relationship between mounting initiation and PA^Esr1+MPN cell activity, we tracked the position of the male and identified a subset of trials when the male initiated mounting after a short period (> 2 s) of quiescence. We reasoned that the onset of locomotion in those trials likely indicates the beginning of a mounting attempt. In those trials, we found that Ca²⁺ increased simultaneously or slightly before elevations in velocity (Extended Data Fig. 5a–e). Importantly, increases in PA^Esr1+MPN Ca²⁺ do not simply reflect movement initiation as no increases were observed when recorded males initiated movements without subsequent mounting attempts (Extended Data Fig. 5a–e). PA^Esr1+MPN cells continued to increase activity as sexual behaviors advanced (Fig. 4j–m, z). In contrast, PA^Esr1+VMHvl cells showed the highest response during fighting against male intruders (0.18 ± 0.03, n = 5 animals) and only increased activity moderately during the advanced stages of sexual behaviors (Fig. 4u–z). In contrast to the PA^Esr1+MPN cells, PA^Esr1+VMHvl cells showed no consistent increase in activity when the male initiated mounting (Extended Data Fig. 5f–j). Investigation of a plastic object elicited no detectable change in the activity of either PA^Esr1+MPN or PA^Esr1+VMHvl cells (Fig. 4z, Extended Data Fig. 6a–d). In three control animals that expressed GFP in PA^Esr1+ cells, we observed little change in fluorescence during any social behaviors (Extended Data Fig. 6e–q). Altogether, these data show that male PA preferentially sends mating-related signals to MPN and aggression-related signals to VMHvl.

Inhibitions of male PA^Esr1+MPN and PA^Esr1+VMHvl cells cause largely distinct deficits in social behaviors

Next, we investigated whether PA^Esr1+MPN and PA^Esr1+VMHvl cells are necessary for male sexual and aggressive behaviors. We injected a retrograde HSV expressing Cre-dependent Flippase (Flp) into either MPN or VMHvl bilaterally and AAV-Ef1α-fDIO-hM4Di-mCherry³⁹ into the PA bilaterally of Esr1–2A-Cre male mice (Fig. 5a, b). As expected, the vast majority of hM4Di-mCherry expressing cells were positive for Esr1 (86.9 ± 0.6%, n = 3 animals; Fig. 5c) and located in PA (88.0 ± 2.4% in PA^Esr1+MPN and 91.3 ± 3.8% in PA^Esr1+VMHvl, n = 8 animals each; Extended Data Fig. 4b).

Three weeks after virus injection, we screened for animals that showed consistent sexual and aggressive behaviors before injecting CNO and saline on separate days. In males expressing hM4Di in PA^Esr1+MPN cells, CNO nearly abolished sexual behaviors but did not affect aggressive behaviors (Fig. 5d–m). On saline-injected day, all tested males showed sexual behaviors towards a receptive female and achieved deep-thrust within the 10-minutes test period. Strikingly, on CNO-injected days, 5/8 males did not attempt to mount and none achieved deep-thrust (Fig. 5e–j). In contrast, aggressive behaviors, including attack duration and latency to attack, was not affected (Fig. 5l, m). Investigatory behaviors towards female increased, possibly as a result of the decrease in sexual behaviors, while male investigation duration remained unchanged (Fig. 5d, k). In a separate experiment, we examined changes in male sexual behaviors after hM4Di-mediated inactivation of an unselected population of PA^Esr1+ cells and observed similarly striking impairments in sexual behaviors (Extended Data Fig. 7).

Inactivating the male PA^Esr1+VMHvl cells caused different behavioral deficits. While CNO injection did not reduce mounting attempts or shallow-thrust, it decreased the duration of deep-thrust and increased its latency (Fig. 5o–t), consistent with a reported role of VMHvl in intromission¹³. Furthermore, PA^Esr1+VMHvl inactivation caused a significant decrease in aggression, as indicated by a significant diminution in attack duration (Fig. 5v, w). Investigations towards males and females were unchanged after PA^Esr1+VMHvl inactivation (Fig. 5n, u). Altogether, PA^Esr1+MPN cells are indispensable for all aspects of male sexual behaviors while PA^Esr1+VMHvl cells mainly mediate male aggression but also play a minor role in advanced stages of sexual behaviors.

Pharmacogenetic activations of male PA^Esr1+MPN and PA^Esr1+VMHvl neurons promote sexual and aggressive behaviors respectively

To test whether PA^Esr1+MPN and PA^Esr1+VMHvl cells are sufficient to drive male sexual and aggressive behaviors respectively, we bilaterally injected HSV-hEf1α-LS1L-Flp into either MPN or VMHvl and AAV-Ef1α-fDIO-hM3Dq-mCherry into bilateral PA in Esr1–2A-Cre male mice (Fig. 6a, b). Histological analysis revealed that vast majority of hM3Dq-mCherry expressing cells were positive for Esr1 (89.1 ± 7.0%, n = 3 animals; Fig. 6c) and confined in PA (88.2 ± 2.8% in PA^Esr1+MPN, n = 8 and 94.0 ± 2.0% in PA^Esr1+VMHvl, n = 6 animals; Extended Data Fig. 4c). All tested males were naïve to minimize spontaneous aggression and sexual behaviors.

During testing, we sequentially introduced a non-receptive adult female and a group-housed nonaggressive male into the home cage of the test male mice. In males expressing hM3Dq in PA^Esr1+MPN cells, i. p. injection of low dose of CNO (0.1 mg/kg) promoted female-directed sexual behaviors without altering aggressive or investigatory behaviors towards either males or females (Fig. 6d–q). On saline-injected day, 4/8 males showed mounting attempts and only 2/8 achieved brief shallow thrust. In contrast, after CNO injection, all males attempted to mount, often repeatedly, and all achieved shallow thrust despite the uncooperative behaviors of the non-receptive females (Fig. 6e–h). Interestingly, when 0.5 mg/kg CNO was administrated, which induces a higher level of neural activation than 0.1 mg/kg CNO⁴⁰, we observed a mixture of sexual and aggressive behaviors towards females: all males showed sexual behaviors and half of those animals also attacked females after several mounting attempts (Extended Data Fig. 8a–g). Importantly, the aggressive behavior appears to be specifically directed towards females as male-directed aggression did not change, suggesting that PA^Esr1+MPN cells do not promote natural inter-male aggression (Extended Data Fig. 8a, h–m).

By contrast, activating PA^Esr1+VMHvl cells did not affect sexual or investigatory behaviors towards either males or females at either CNO concentration (Fig. 6r–ee and Extended Data Fig. 8n–z). However, administering CNO but not saline promoted aggression in a dose-dependent manner. While only 2 out of 8 males attacked males after saline injection, 5 out of 8 attacked males and 2 attacked females with 0.1 mg/kg CNO (Extended Data Fig. 8s, t, y, z). At 0.5 mg/kg CNO, all animals attacked males and 6/8 also attacked females (Fig. 6w, x, dd, ee), strongly suggesting that PA^Esr1+VMHvl cells promote aggressive behaviors.

We also activated PA^Esr1+VMHvl and PA^Esr1+MPN cells optogenetically (20 Hz, 20 ms, 0.5–7.5 mW). While activation of PA^Esr1+VMHvl cells elicited stimulation-locked attack, we did not observe any increase in sexual behaviors upon optogenetic activation of PA^Esr1+MPN cells (Extended Data Fig. 9). In fact, these animals showed very little sexual behaviors during the entire testing period, suggesting that strong and synchronous exogenous stimulation of PA^Esr1+MPN cells is insufficient to promote sexual behaviors. Taken together, activation of PA^Esr1+VMHvl neurons can elicit aggression while enhancing the natural responses of PA^Esr1+MPN cells promotes male sexual behaviors towards females.

Connectivity of male PA^Esr1+MPN and PA^Esr1+VMHvl neurons

To achieve a comprehensive understanding of the connectivity of PA^Esr1+MPN and PA^Esr1+VMHvl cells, we injected retrograde HSV-hEf1α-LS1L-Flp into either MPN or VMHvl and AAV-hSyn-Con/Fon-EYFP in the PA of Esr1–2A-Cre male mice (Fig. 7a, b). As expected, we found that PA^Esr1+MPN neurons project densely to MPN while PA^Esr1+VMHvl neurons project strongly to VMHvl (Fig. 7c–e). PA^Esr1+MPN cells largely spare VMHvl while some fibers in MPN were observed from PA^Esr1+VMHvl cells, indicating that PA^Esr1+VMHvl may influence the MPN activity to some extent. This difference may also be because PA axons route through stria termanlis to hypothalamus and thus VMHvl-targeting fibers need to course through MPN while MPN-targeting axons do not pass VMHvl²⁶ (Fig. 7c–e). Additionally, PA^Esr1+MPN and PA^Esr1+VMHvl neurons send collaterals to several regions inside and outside of the hypothalamus, but with clear distinctions. Within the hypothalamus, the other major target of PA^Esr1+MPN neurons is the anteroventral periventricular nucleus (AVPV) whereas ventral premammillary nucleus (PMv) is the other main region receiving inputs from PA^Esr1+VMHvl neurons (Fig. 7c). Outside of the hypothalamus, PA^Esr1+MPN but not PA^Esr1+VMHvl cells project to the lateral septum ventral part (LSv) (Fig. 7c–e). Finally, both PA^Esr1+MPN and PA^Esr1+VMHvl cells project densely to the MeApd and BNSTpr, and moderately to the ventral subiculum (vSub), suggesting possible convergence of information in those areas. Altogether, the anterior biased PA^Esr1+MPN neurons show an anterior-biased projection pattern (AVPV and MPN) whereas the PA^Esr1+VMHvl neurons show a posterior-biased projection pattern (VMHvl and PMv).

We next used projection-specific monosynaptic rabies tracing to map inputs to PA^Esr1+MPN and PA^Esr1+VMHvl cells⁴¹. We injected HSV-hEf1α-LS1L-Flp into either MPN or VMHvl, and AAVs expressing Flp-dependent TVA receptors and a rabies glycoprotein into the PA of Esr1–2A-Cre male mice. After 2 weeks, EnvA-pseudotyped, glycoprotein-deleted rabies virus (EnvA-RVΔG-eGFP) was injected into PA (Fig. 7f). Seven days later, we examined inputs (eGFP⁺ cells) to PA^Esr1+MPN and PA^Esr1+VMHvl starter cells (eGFP⁺mCherry⁺ cells) across the entire brain (Fig. 7g–l). Hippocampus represents the largest input to both PA^Esr1+MPN and PA^Esr1+VMHvl cells. Approximately half of all retrogradely labeled cells, regardless of the PA starter cell population, were located in hippocampus, including approximately 30% in CA1 and 20% in vSub (Fig. 7h, j, l). Several other regions including medial septum (MS), paraventricular thalamus (PVT), posterolateral (CoApl) and posteromedial part of cortical amygdala (CoApm) and BMAp also provide moderate inputs to both PA^Esr1+MPN and PA^Esr1+VMHvl cells. Interestingly, PMv and MeApd, two regions that are known to project to PA appear to target the PA^Esr1+MPN subpopulation preferentially^26,42 whereas PA^Esr1+VMHvl cells receive relatively strong inputs from regions processing volatile olfactory information, including piriform cortex (Pir) and piriform-amygdalar area (PAA) (Fig. 7h, j, l)⁴³. Altogether, PA^Esr1+MPN and PA^Esr1+VMHvl are both well positioned to integrate olfactory, contextual, experiential and emotional state related information and serves as an important player in the male sexual and aggressive circuit respectively.

PA neurons play important roles in male sexual and aggressive behaviors

Nearly three decades ago, Canteras and colleagues performed classical anterograde tracing from PA, revealing its dense projection to the medial hypothalamus, and accordingly proposed a role for PA in social behaviors²⁶. Despite this early insight, the functional role of PA in male social behaviors has not been examined for decades⁴⁴. Recently, Zha et.al. reported that vesicular glutamate transporter 1 (Vglut1) expressing cells in the PA that project to the VMHvl can bi-directionally modulates male aggression⁴⁴. This result is largely consistent with our findings showing a key role of PA^Esr1+VMHvl cells in male aggression. However, PA^Esr1+VMHvl cells are likely to be more aggression-specific than the PA^Vglut1+VMHvl cells as PA^Esr1+VMHvl but not PA^Vglut1+VMHvl cells showed a consistent Ca²⁺ increase during male investigation and attack⁴⁴. Indeed, while over 95% of Esr1+ cells express Vglut1, less than half of the Vglut1⁺ cells express Esr1 (Extended Data Fig. 2).

Our results further demonstrate an indispensable role of Esr1⁺ neurons in PA in male sexual behaviors: inhibiting PA^Esr1+ cells or PA^{Esr1+ MPN} cells abolishes virtually all aspects of sexual behaviors. The central role of the PA in sexual behavior is particularly striking considering the relatively inconsistent behavioral deficit caused by inhibiting other major upstream regions of MPN, including MeA, BNSTpr and VMHvl. Killing the aromatase-expressing cells or acutely inhibiting GABAergic cells (the major population) within MeA does not decrease sexual behaviors in male mice^14,17. Lesioning BNST causes only minor impairments in sexual behaviors such as increased latency to ejaculate^45,46 although a recent study showed that inactivation or ablation of aromatase neurons in BNSTpr reduced intromission and ejaculation²³. Killing VMHvl progesterone (PR) expressing cells slightly decrease the number of intromission⁸. In contrast, inactivation of PA^Esr1+MPN cells compromised the entire sequence of the male sexual behaviors - from mounting initiation to ejaculation.

In contrast to clear behavior changes elicited by optogenetic activation of PA^Esr1+VMHvl cells, optogenetic stimulation of PA^Esr1+MPN cells failed to elicit sexual behaviors. This could be because mating is a multi-stage sequential process and MPN contains cells with diverse response patterns during sexual behaviors: some MPN cells are excited during pursuit of females and mounting initiation but inhibited during intromission while others are activated only during advanced stages of sexual behaviors⁴⁷. Non selective, synchronous activation of PA^Esr1+MPN afferents likely recruit MPN cells with diverse functions and thus fail to drive a coherent behavioral motor program. Additionally, unlike VMHvl cells which are mainly glutamatergic, MPN cells are primarily GABAergic and form extensive local synapses⁴⁸. Thus, activation of functionally diverse MPN cells by PA may lead to a general suppression of activity and consequently a lack of sexual behaviors altogether. Unlike optogenetic activation that directly evokes action potentials, hM3Dq activation increases the excitability of cells by engaging endogenous Gq signaling. Importantly, hM3Dq mediated neural activation is dependent on ligand concentration. A previous study using mice expressing hM3Dq in the cortical pyramidal cells found that 0.1 mg/kg CNO induced an increase in locomotion while 0.5 mg/kg CNO elicited seizures⁴⁰, suggesting that CNO can be titrated to elevate neural activity within a physiological range. It is therefore likely that the endogenous synaptically-driven spiking pattern of PA neurons, which in turn contribute to the diverse response patterns of MPN cells, is essential for generating complex behavioral sequences like mating.

PA enriched genes and their relevance to social behaviors

Mating- and aggression-related PA subregions show clear differences in gene expression patterns. Differentially expressed genes include transcription factors (Etv1, Zic2), G-protein coupled receptors (Npy2r, Prokr2), ion channels (Chrna7), calcium binding proteins (Calb2) amongst others. One of the PA enriched genes, Zic2, has been implicated previously in aggression: mice with hypomorphic mutations in Zic2 decreased inter-male aggression and social dominance⁴⁹ . Interestingly, the authors also noted a high abundance of Zic2 in PA and a decreased cell density in the area in mutant mice. Furthermore, while wild-type mice show strong acetylcholine esterase (AChE) staining in PA, AChE levels are low in mutant animals⁴⁹. The strong AChE staining in PA is consistent with our finding that Chrna7 expresses highly in the PA. Our retrograde tracing revealed that PA receives a moderate input from MS, a region containing abundant cholinergic cells, suggesting PA as a potential site for cholinergic modulation of aggression⁵⁰.

Sexually dimorphic PA

What is the function of PA in females? Anatomical studies in rodents revealed that PA exists in both males and females with comparable volume, morphology and cell density²⁸. However, several lines of evidences suggest that PA is likely to play sexually dimorphic roles in social behaviors. First, MPN serves differential functions between males and females. While MPN is the central region for male sexual behavior, its primary function in females is to mediate maternal behaviors^1–5,29. Second, VMHvl function is sexually dimorphic. While VMHvl is key to male aggression and only plays a minor role in male sexual behaviors, it is indispensable for both female sexual behaviors and aggression^6,8,30. In females, VMHvl appears to contain a molecularly distinct compartment specialized for female sexual behaviors that is absent in males³⁰. Third, PA shows sexually dimorphic projection pattern. While both male and female PA cells project to MPN, VMHvl, BNSTpr and MeApd, female PA appears to lack a strong projection to AVPV²⁸. Taken together, we speculate that female PA could be relevant for various social behaviors, including maternal, sexual and aggressive behaviors. The organizations of those behavior-relevant populations and their downstream circuits are likely to differ from those in males. Future studies focusing on the PA in females will help elucidate the sexual dimorphism of social behavior circuits.

The placement of PA in the social behavior circuits in males

In 2000, Larry Swanson proposed a basic wiring diagram for controlling motivated behaviors based on extensive developmental and neuroanatomical evidence³⁴. In his model, forebrain exerts “top-down” voluntary control on various motivated behaviors through a set of triple descending projections to the “behavioral control column” in the upper brainstem³⁴. The rostral part of the behavior control column is for ingestive and social behaviors, and contains several hypothalamic nuclei, including MPN and VMHvl. The caudal part of the column is for general exploratory and foraging behaviors and contains mammillary nucleus, substantia nigra (SNr) and ventral tegmental area. In his view, the overall pattern of motivational behavior circuitry originating in the cortical map, with its regionalization into functionally distinct areas, and then involving sequentially the striatum and pallidum, followed by the brainstem. The cortical projection is primarily glutamatergic whereas the striatal and pallidal projections are primarily GABAergic. A familiar example of the triple projection is the isocortical‐dorsal striatopallidal system in which (1) cortical regions send direct glutamatergic projections to the SNr and collateral and topographic projections to dorsal striatum; (2) dorsal striatum sends GABAergic projection to the SNr with a collateral projection to the pallidum; (3) pallidum then sends GABAergic projection to the SNr with a collateral to the thalamus, which in turn projects to cortex (Extended Data Fig. 10). Importantly, in this grand scheme, each and all of the anatomical structures in the cerebral hemisphere belong to either prototypical cortex, striatum or pallidum.

According to this view, we propose that PA along with the heavily interconnected vSub constitute the “social cortex” and mediate top-down regulation of the hypothalamic behavior control column. Similar to the classical isocortical‐dorsal striatopallidal system, PA/vSub, MeA and BNST provide a triple descending projection to the column (Extended Data Figure 10). First, PA/vSub (cortex) sends excitatory projection directly to the column (MPN and VMHvl) as well as collaterals to the MeA (rostrocaudal striatum)³⁴. MeA then sends inhibitory projection to the same part of the behavior column (MPN and VMHvl) as well as collaterals to the BNST (rostrocaudal pallidum). Lastly, BNST provides disinhibitory projection to the column. In this model, PA/vSub occupies a similar position in social behavior circuit as the motor cortex in voluntary motor circuit and determines the timing of the behavior initiation. Indeed, PA cells increase activity when the animal makes the first move towards the female to initiate mount. Suppressing PA activation blocks the initiation of sexual behaviors altogether.

The discovery of PA as a key node for social behaviors supports the existence of a universal diagram underlying cortical control of movement. PA is potentially a specialized “cortex” to control a set of stereotypical movements constituting innate social behaviors. This study offers a novel framework for understanding how social behaviors are organized and initiated at the circuit level and opens new avenues for studying the neural generation of innate behaviors.

Mice

All procedures were approved by NYULMC IACUC in compliance with the NIH guidelines for Care and Use of Laboratory Animals. Adult mice 12–38 weeks of age were used for all studies. Mice were housed in 18–23°C with 40–60 % humidity under a 12 h light-dark cycle (10 p.m. to 10 a.m. light), with food and water available ad libitum. Animals used for CTB retrograde tracing and RNA-Seq experiments were 14–16 week old wild-type C57BL/6N male purchased from Charles River. Male Esr1–2A-Cre (14–38 weeks, Jackson Stock No. 017911), male Vgat-ires-Cre × Ai6 (12–16 weeks, Jackson stock No. 016962 and 007906)⁵², male Vglut1-ires-Cre × Ai6 (12–16 weeks, Jackson stock No. 023527 and 007906) ⁵³ were used for functional, recording and histological experiments. Stimulus animals were adult C57BL/6N female and male or BALB/c male mice (> 8 weeks) purchased from Charles River. Stimulus animals were group housed. After surgery, all test animals were singly housed. All the experiments were performed during the dark cycle of the animals.

Viruses

AAV2-CAG-FLEX-GFP, AAV2-hSyn-DIO-mCherry, AAV1-Ef1α-DIO-hM4Di-mCherry, AAV5-hSyn-Con/Fon-EYFP, AAV2-Ef1α-DIO-ChR2-EYFP, AAVDJ-Ef1α-fDIO-EYFP and AAVDJ-hSyn-Con/Fon-hChR2(H134R)-EYFP were purchased from the University of North Carolina vector core. AAV5-CAG-FLEX(FRT)-TC and AAV8-CAG-FLEX(FRT)-G were purchased from the vector core of Stanford University. AAVDJ-Ef1α-fDIO-hM4Di-mCherry and AAVDJ-Ef1α-fDIO-hM3Dq-mCherry were made in Byungkook Lim lab. HSV-hEf1α-LS1L-EYFP, HSV-hEf1α-LS1L-Flp, HSV-hEf1α-LS1L-mCherry and HSV-hEf1α-LSL-GCaMP6f was purchased from Massachusetts General Hospital Neuroscience Center vector core. EnvA-G-Deleted Rabies-EGFP was purchased from Salk Institute vector core. All AAV titers ranged from 3.00 × 10¹² to 4.30 × 10¹³ genomic copies/mL. All HSV titers were 1.00 × 10⁹ genomic copies/mL. Rabies virus titers were > 3.00 × 10⁷ transforming units/mL.

Stereotactic surgery

Mice (12–20 weeks) were anesthetized with isoflurane (1–1.5%) and placed in a stereotaxic apparatus (Kopf Instruments Model 1900). Viruses and tracers were delivered into the targeted brain regions as described previously²⁹. All animals included in the final analysis have correct injection sites as verified by histology.

For single retrograde tracing experiments, 20 nL of Alexa555 conjugated CTB (Thermo Fisher, C22843) was injected into the unilateral MPN (AP: −0.31 mm, ML: ±0.31 mm, DV: −4.95 mm) or VMHvl (AP: −1.45 mm, ML: ±0.655 mm, DV: −5.655 mm). 3/4 and 3/6 animals that showed correct unilateral targeting in MPN and VMHvl, respectively, were included in the final analysis.

For dual retrograde CTB tracing experiments, 20 nL of Alexa555-CTB and Alexa647-CTB (Thermo Fisher, C22843 and C34778) was injected into the MPN and VMHvl, respectively. Brains were harvested 10 days later. 3/5 animals with correct targeting in both MPN and VMHvl were included.

For microdissection of distinct subregions in the PA, 30 nL of Red and Green retrobeads (Lumafluor Item, R170, G170) were injected into unilateral MPN and VMHvl, respectively. Brains were harvested 7 days later. 3/5 animals with correct targeting in both MPN and VMHvl were included in the final analysis.

For dual-retrograde tracing, HSV-hEf1α-LS1L-mCherry and HSV-hEf1α-LS1L-EYFP were injected into ipsilateral MPN and VMHvl, respectively. Brains were harvested two weeks later. 3/7 animals that showed correct unilateral targeting in both MPN and VMHvl were included in the final analysis.

For investigating the outputs of PA^Esr1+MPN and PA^Esr1+VMHvl projectors, 100 nL(130 nL) of HSV- hEf1α-LS1L-Flp was injected into the MPN (VMHvl) and simultaneously 100 nL of AAV5-hSyn-Con/Fon-EYFP was injected into the ipsilateral PA (AP: −2.30 mm, ML: ±2.40 mm, DV: −4.80 mm). Brains were harvested two weeks later. 3/5 animals that showed correct unilateral targeting in the MPN and correct virus expression in the PA and 3/5 animals that showed correct unilateral targeting in the VMHvl and correct unilateral virus expression in the PA were included in the final analysis.

For investigating the inputs of PA^Esr1+MPN and PA^Esr1+VMHvl cells, 100 nL (130 nL) of HSV- hEf1α-LS1L-Flp was injected into the MPN (VMHvl) and simultaneously 180 nL 1:1 AAV5-CAG-FLEX(FRT)-TC and AAV8-CAG-FLEX(FRT)-G were injected into the ipsilateral PA. After two weeks, 600 nL EnvA-G-Deleted Rabies-eGFP was injected into the PA of the same side. One week later, the brain was harvested for histological analysis. 3/5 animals that showed correct unilateral targeting in the MPN and correct virus expression in the PA and 3/6 animals that showed correct unilateral targeting in the VMHvl and correct virus expression in the PA were included in the final analysis.

For the slice recording experiment, 100 nL AAV2-Ef1α-DIO-ChR2-EYFP was injected into the PA and simultaneously 100 nL AAV2-hSyn-DIO-mCherry was injected into the MPN or VMHvl. The brains were used for recording three weeks later. 10/12 animals that showed correct unilateral targeting in MPN, VMHvl and PA were included in the final analysis.

For fiber photometric recording of the PA^Esr1+MPN (PA^Esr1+VMHvl) population, 100 nL (130 nL) HSV-hEf1α-LSL-GCaMP6f was injected into MPN (VMHvl). Control animals were injected with 100 nL of AAV2-CAG-FLEX-GFP unilaterally injected into the PA. Then, a custom made optic fiber assembly (Thorlabs, FR400URT, CF440) was inserted ∼150 μm above the dorsal boundary of the PA and secured using dental cement (C&B Metabond, S380). All recordings start two weeks after virus injection. 7/8 animals that showed correct unilateral targeting in the MPN and correct expression of GCaMP6f and fiber placement in the PA were included in the final analysis. 5/8 animals that showed correct unilateral targeting in the VMHvl and correct GCaMP6f expression and fiber placemen in the PA were included in the final analysis.

For pharmacogenetic inhibition of PA^Esr1+ cells, 100 nL of AAV1-hSyn-DIO-hM4Di-mCherry was injected bilaterally into the PA. Control animals were injected with 100 nL of AAV2-hSyn-DIO-mCherry into the PA bilaterally. 5/9 test animals that showed correct bilateral targeting of PA were included in the final analysis.

For pharmacogenetic inhibition of PA^Esr1+MPN or PA^Esr1+VMHvl, 100 nL (130 nL) of HSV-hEf1α-LS1L-Flp was injected into the MPN (VMHvl) and simultaneously 100 nL of AAVDJ-Ef1α-fDIO-hM4Di-mCherry was injected into the PA bilaterally. Control animals were injected with 100 nL (130 nL) of HSV-hEf1α-LS1L-Flp into the MPN (VMHvl) bilaterally and 100 nL of AAVDJ-Ef1α-fDIO-EYFP into the PA bilaterally. 8/10 test animals that showed correct bilateral targeting in the MPN and correct bilateral virus expression in the PA were included in the final analysis. 8/10 test animals that showed correct bilateral targeting in the VMHvl and correct bilateral virus expression in the PA were included.

For pharmacogenetic activation of PA^Esr1+MPN or PA^Esr1+VMHvl cells, 100 nL (130 nL) of HSV-hEf1α-LS1L-Flp was injected into the MPN (VMHvl) and simultaneously 120 nL of AAVDJ-Ef1α-fDIO-hM3Dq-mCherry was injected into the PA bilaterally. Control animals were injected with 100 nL (130 nL) of HSV-hEf1α-LS1L-Flp into the MPN (VMHvl) bilaterally and 100 nL of AAVDJ-Ef1α-fDIO-EYFP into the PA bilaterally. 8/11 test animals that showed correct bilateral targeting in the MPN and correct bilateral virus expression and fiber placement in the PA were included in the final analysis. 8/14 test animals that showed correct bilateral targeting in the VMHvl and correct bilateral virus expression and fiber placement in the PA were included in the final analysis.

For optogenetic stimulation of PA^Esr1+MPN or PA^Esr1+VMHvl cells, 100 nL (130 nL) of HSV-LS1L-Flp was bilaterally injected into the MPN (VMHvl) and simultaneously 200 nL of AAVDJ-hSyn-Con/Fon-hChR2(H134R)-EYFP was injected into the PA bilaterally. Control animals were injected with 130 nL of HSV-hEf1α-LS1L-Flp bilaterally injected into the VMHvl and simultaneously 150 nL of AAV5-hSyn-Con/Fon-EYFP into the PA bilaterally. Then, a 200-µm optic fiber assembly (Thorlabs, FT200EMT, CFLC230) was implanted 200 μm above each side of the PA during the same surgery. 5/10 test animals that showed correct bilateral targeting in the MPN, correct bilateral expression of ChR2-EYFP in the PA and correct bilateral fiber placements in the PA were included in the final analysis. 5/16 test animals that showed correct bilateral targeting in the VMHvl, correct bilateral expression of ChR2-EYFP in the PA (> 100 positive cells in all sections) and bilateral fiber placements in the PA were included in the final analysis.

Immunohistochemistry

Immunofluorescence staining proceeded as previously described^29,30. Briefly, mice were perfused transcardially with 1 × PBS followed by 4% paraformaldehyde (PFA) in 0.1 M PBS. Brains were extracted, post-fixed in 4% PFA for 2–3 hours at 4°C followed by 48 hours in 15% sucrose, embedded in OCT mounting medium, frozen on dry ice, cut to 50 μm thick sections using a cryostat (Leica Biosystems) and collected in PBS using 12 well plate. For antibody staining, brain sections were washed with PBS three times and blocked in PBS-T (0.3% Triton X-100 in 1 × PBS) with 10% normal donkey serum for 1 hour at room temperature. Sections were then incubated in primary antibody diluted in blocking solution at 4°C for 24–72 hours. We stained for Esr1 (rabbit anti-Esr1; 1:1000; Santa Cruz, sc-542, Lot #F1715), Fos (goat anti-Fos; 1:1000; Santa Cruz, sc52-g), or GFP (rabbit anti-GFP; 1:1000; Thermo Fisher, A11122)^29,30. Sections were then washed with PBS-T three times and incubated in secondary antibodies diluted in blocking solution (donkey anti-rabbit, donkey anti-goat, Alexa Fluor 488 or 594, 1:500; Thermo Fisher, A11055, R37118, R37119) with DAPI (1:20000; Thermo Fisher, D1306) or NeuroTrace 435/455 Blue Fluorescent Nissl Stain (1:200; Thermo Fisher, N21479) for 2–3 hours at room temperature. Sections were then washed with PBS for three times, mounted on Superfrost slides (Fisher scientific, 12–550-15) and coverslipped for imaging on a confocal microscope (Zeiss LSM 510 or 700 microscope) and/or a virtual slide scanner (Olympus, VS120).

Fluorescent In situ hybridization

Extracted brains were frozen on dry ice and 12-μm coronal brain sections were collected using a cryostat (Leica Biosystems). For all genes except Pde11a, we performed RNAscope labeling following the manufacturer protocol (ACD; Advanced Cell Diagnostics)³⁶. For Alexa555-CTB labeled tissues, incubation time for protease solution IV was shortened from 30 min to 30 s to preserve Alexa555-CTB expression. For Pde11a, RNA probes were generated from the pSPT19 Vector (Sigma) using primers reported on www.brain-map.org. Specifically, pSPT19 plasmid was cut with EcoRI and the antisense strand was transcribed using a T7 RNA polymerase with digoxigenin-labeled UTP (DIG RNA Labeling Kit; Sigma-Aldrich, 11175025910). The probe was then hybridized with the sample at 56 °C for 16 h, detected with horseradish peroxidase-conjugated anti-DIG antibody (1:500; Sigma-Aldrich, 11207733910)⁵⁴, amplified with biotin-conjugated tyramide (PerkinElmer, EL748001KT) and visualized with Alexa 488-conjugated streptavidin (Thermo Fisher, S32354)³⁰.

Cell counting and axon terminal quantification

To analyze labeled cells, 20 × confocal or epifluorescence images were acquired and cells were counted manually using ImageJ. Cells that were not entirely contained within a given region of interest (ROI) were excluded from analyses. For counting DAPI, Nissl, CTB⁺, Fos⁺ and Esr1⁺ cells, 20 × fluorescent confocal images were acquired. For counting GCaMP6⁺, EYFP⁺ and mCherry⁺ in dual retrograde labeling, pharmacogenetics and fiber photometry experiments, 20 × fluorescent images were acquired with a virtual slide scanner (Olympus, VS120). For Figure 1k, PA in each brain section was divided to four quadrants along the long and short axes. Then, labeled cells in each quadrant were counted and the sum from all sections (6–7 sections, 50 µm/section, ranging from Bregma −2.15 mm to −2.90 mm) was calculated and averaged across animals. For dual retrograde labeling, all mCherry⁺ and EYFP⁺ neurons in the amygdala were counted. For monosynaptic rabies tracing, all GFP⁺ cells throughout the brain were counted except those in the primary infection site. Figure 7c and 7h includes areas that contribute to >1% of total inputs to PA^Esr1+MPN or PA^Esr1+VMHvl cells. To quantify the density of the projections of PA^Esr1+MPN and PA^Esr1+VMHvl cells, we selected a boxed area (200 × 200 μm for AVPV, VMHvl, PMv and LSv, 300 × 400 μm for MPN, 250 × 250 μm for MeApd, 250 × 400 μm for BNSTpr, 300 × 200 μm for vSub) in each region containing fibers and calculated the average pixel intensity as F_raw using ImageJ and Photoshop. On the same image, a boxed area of the same size but in a brain region containing no fiber terminals was selected for calculating the background intensity (F_background). F_signal was then calculated as F_raw minus F_background. For each animal, F_signal was normalized by the maximum F_signal across all the analyzed regions. The normalized F_signal was then used for calculating the average terminal field intensity across animals.

In vitro electrophysiological recordings

Esr1–2A-Cre male adult mice were injected with 140 nL AAV2-Ef1α-DIO-ChR2-EYFP into the PA and 100 nL AAV2-hSyn-DIO-mCherry into MPN and VMHvl. Three weeks after virus injection, acute coronal brain slices of MPN and VMHvl (275 μm in thickness) were obtained using standard methods as described previously²⁹. Briefly, isoflurane-anesthetized mice were perfused with ice-cold artificial cerebrospinal fluid (ACSF) containing (in mM) 125 NaCl, 2.5 KCl, 1.25 NaH₂PO₄, 25 NaHCO₃, 1 MgCl₂, 2 CaCl₂ and 11 glucose. Slices were collected in ice-cold choline-based cutting solution (in mM: 110 choline chloride, 25.0 NaHCO₃, 2.5 KCl, 7 MgCl₂, 0.5 CaCl₂, 1.25 NaH₂PO₄, 25.0 glucose, 11.6 ascorbic acid, and 3.1 pyruvic acid) using a Leica VT1200s vibratome, incubated for 10–20 min in ACSF at 34°C and then at room temperature until use. Individual slices were transferred to a recording chamber mounted on an upright microscope (Slicescope Pro 6000, Scientifica) and continuously perfused with ACSF warmed to 32–34°C (TC-344C; Warner Instruments). All solutions were constantly bubbled with 95% O₂/5% CO₂. mCherry-labeled Esr1⁺ neurons within either MPN or VMHvl were identified with an Olympus 40 × water-immersion objective using infrared differential interference contrast optics and epifluorescence. Standard whole-cell voltage-clamp recordings were performed using glass electrodes (2–4 MΩ) containing (in mM) 135 CsMeSO₃, 10 HEPES, 1 EGTA, 3.3 QX-314 (Cl- salt), 4 Mg-ATP, 0.3 Na-GTP, and 8 Na₂-Phosphocreatine (pH 7.3 adjusted with CsOH). Series resistance did not exceed 20 MΩ. Membrane currents were amplified and low-pass filtered at 3 kHz using a Multiclamp 700B amplifier (Molecular Devices), digitized at 10 kHz and acquired using National Instruments acquisition boards (PCIe-6351 and PCI-6713) and a custom version of ScanImage (https://github.com/bernardosabatini/SabalabAcq) written in MATLAB (Mathworks) and analyzed offline using Igor Pro (Wavemetrics). To activate ChR2-expressing axons, brief pulses of full field illumination (1 ms duration; 10 mW⋅mm⁻² under the objective) were delivered onto the recorded cell with a blue LED light (pE-300white; CoolLED) at 30 s intervals. Optogenetically-evoked EPSCs and IPSCs (oESPSs and oIPSCs) were recorded by holding the membrane potential of recorded neurons at −70 and 0 mV (corrected for a ~8 mV liquid junction potential), respectively. 10 μM 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX; Tocris, 190) was bath applied to confirm that oEPSC is mediated by AMPA receptor and the oIPSC is polysynaptic and dependent on the oEPSC.

Behavioral analysis

Animal behaviors in all experiments were video recorded from both the side and top of the cage using two synchronized cameras (Basler, acA640–100 gm) and a commercial video acquisition software (StreamPix 5, Norpix) in a semi-dark room with infrared illumination at a frame rate of 25 frames/s. Behavioral annotation and tracking were performed on a frame-by-frame basis using custom software written in MATLAB (https://pdollar.github.io/toolbox/)⁶. For male sexual behaviors, “Attempt mount” was defined as when the male extends his forelimbs to grasp and mount the female’s flanks. “Shallowly thrust” was defined as when the male poses his forelegs over the female’s back and with his hindlimbs on the ground accompanying shallow pelvic thrusts. The “Shallowly thrust” onset was defined as the moment when the male begins to rapidly thrust against the female’s rear. “Deeply thrust” was defined as a deep rhythmic thrust following “Shallow-thrust”. The onset of “Deeply thrust” was defined as the time when the male performs the first deep-thrust, presumably inserting its penis into the female’s vaginal part. “Ejaculate” was detected when the male stops deep thrusting for seconds while continuously clutching onto the female and then slumps to the side of the female. The putative ejaculation event was confirmed by the presence of vaginal copulatory plug after the test. For aggression, attack was defined by a suite of actions initiated by the resident toward the male intruder, which included lunges, bites, tumbling, and fast locomotion episodes between such behaviors. Sniffing was defined by nose contact to the foreign body. Mouse behavior was manually annotated frame-by-frame by an experimenter not blind to the group assignment of the animals. Attack events in a randomly selected set of videos were annotated using Deeplabcut⁵⁵ and Simba⁵⁶ and good agreement between human and machine annotations was observed. During annotation, the neural responses were not available to the experimenter.

Fiber photometry

The test animals were screened for both aggressive behaviors and sexual behaviors before the surgery and randomly assigned to experimental and control groups. Fiber photometry was performed as described previously^10,29,30. Briefly, a 390-Hz sinusoidal blue LED light (30 μW) (LED light: M470F1; LED driver: LEDD1B; both from Thorlabs) were bandpass filtered (passing band: 472 ± 15 nm, FF02–472/30–25, Semrock) and delivered to the brain to excite GCaMP6. The emission lights traveled back through the same optic fiber, bandpass filtered (passing bands: 535 ± 25 nm, FF01–535/505, Semrock), passed through an adjustable zooming lens (Thorlab, SM1NR01 and Edmund optics, #62–561), detected by a Femtowatt Silicon Photoreceiver (Newport, 2151) and recorded using a real-time processor (RZ5, TDT). The envelope of the 390-Hz signals reflected the intensity of the GCaMP6 and was extracted in real time using a custom TDT OpenEx program. The signal was low pass filtered with a cut-off frequency of 5 Hz. During recording, a sexually receptive female C57BL/6N mouse (‘intruder’) was placed in the homecage of the experimental male mouse (‘resident’) until the resident reached ejaculation. After the session with a female intruder, a group housed non-aggressive male BALB/c or C57BL/6N mouse was placed in the homecage for 10 minutes and then an object was introduced into the home cage of the recorded male for 5 minutes. Each stimulus was presented with three minutes in between. To analyze the recording data, the MATLAB function “msbackadj” with a moving window of 25% of the total recording duration was first applied to obtain the instantaneous baseline signal. The instantaneous ΔF/F was calculated as (F_raw –F_baseline)/F_baseline. The peri-event histogram (PETH) of a given behavior was constructed by aligning the ΔF/F signal to the onset of the behavior. The ΔF/F values during ‘Attempted-mount’ and ‘Shallow-thrust’ was calculated based on trials that were not followed by ‘Deep-thrust’. The peak ΔF/F for each behavior episode were calculated as the maximum ΔF/F of the PETHs during each behavioral episodes minus the average ΔF/F in the duration-matched period prior to the behavior onset. The resulted values of all behavior episodes of the same behavior were then averaged for each animal and plotted in Figure 4z. The location of the animals were tracked using a custom software (https://github.com/pdollar/toolbox)^6,57. The velocity of the animal was calculated as the displacement of the animal location between two adjacent frames. The change points in velocity were determined using Matlab function “findchangepts”. Only change points after a period (>2s) of immobility (velocity < 2pixels/frame) were included for analysis. We separated the change points based on whether a mounting attempt occurred within 3s or not and calculate the average Ca²⁺ signal change between −1–0s and 0–1s of the change points followed by or not followed by sexual behaviors.

Pharmacogenetic neural silencing

The test animals were screened for both aggressive behaviors and sexual behaviors before the surgery and randomly assigned to experimental and control groups. For PA^Esr1+ inactivation experiments, we screened the sexual performance of the animals before and two weeks after surgery to ensure the high sexual performance. During screening, a receptive C57BL/6N female (predetermined using a highly sexually experienced male) was introduced into the home cage of the test mouse and only animals achieved intromission during the 10 minutes test periods were included in the final test. Three weeks after the injection, mice were intraperitoneally injected with saline and CNO (5 mg/kg, Sigma, C0832) on interleaved day. 30 min after saline or CNO injection, a sexually receptive female C57BL/6N mouse was placed in the homecage of the test mouse and test mouse was allowed to freely interact with the female for 10 minutes or until the male showed 15 times of attempted mounting or 10 times of intromission, whichever is shorter. For PA^Esr1+MPN and PA^Esr1+VMHvl inactivation experiments, males are screened for both aggressive behaviors and sexual behaviors both before and two weeks after the surgery. During aggression screening, a group-housed BALB/c male mouse was introduced into the home cage of the test mouse and only animals showed minimum 5 attacks during the 10 minutes testing period were included in the final test. Three weeks after the surgery, the sexual and aggressive behaviors were tested after CNO and saline injection. During each test, the test mouse encountered a receptive female first for maximally 10 minutes (see above) and a non-aggressive BALB/c male mouse for 10 minutes after the female was removed for three minutes.

Pharmacogenetic neural activation

For PA^Esr1+MPN and PA^Esr1+VMHvl activation experiments, test animals were naïve (no prior sexual experience or tested for aggression) and group housed before the virus injection and singly housed after surgery. Animals were assigned to test and control groups randomly. Three weeks after the surgery, test males were intraperitoneally injected with saline and CNO (0.1 or 0.5 mg/kg, Sigma, C0832) on separate days. 30 min after saline or CNO injection, a non-receptive female and then a non-aggressive BALB/c male mouse was introduced to the home cage of the test animal, each for 10 minutes with 2 minutes in between. Note that no test animal achieved deep thrust in this experiment as the female is non-receptive.

Intersectional optogenetic behavioral experiments

Test animals were naïve (no prior sexual experience or tested for aggression) and group housed before the virus injection and singly housed after surgery. Animals were assigned to test and control groups randomly. After three weeks of viral incubation, on the test days, two 200-μm multimode optical fibers (Thorlabs, FT200EMT) were connected with the bilateral implanted ceramic ferrules (Thorlabs, CFLC230–10) using a match sleeve (Thorlabs, ADAL1). Randomly selected group-housed adult BALB/c male, castrated BALB/c male, C57BL/6N female mice were introduced into the home cage of the subject. Blue light (473 nm, Shanghai Dream Laser) was delivered through the fiber bilaterally in 20 ms pulses at 20 Hz with the intensity ranging from 0.5 to 7.5 mW for 60 s using OpenEx (TDT). The light intensity was measured at the fiber end during pulsing using an optical power meter (Thorlabs, PM100D). A sham stimulation period (0 mW) for 60–180 s was interleaved with the real light stimulation period as an internal control. All behavioral tests were repeated at least once to ensure the reproducibility of any light-induced behavioral changes in the same session.

Subpopulation specific input and output mapping

To investigate downstream targets of PA^Esr1+MPN and PA^Esr1+VMHvl cells, the brain was harvested for histological analysis after two weeks of viral incubation. Every third section (50-μm thickness) was collected and immunohistochemistry was performed to amplify eGFP signals in neuronal fibers in downstream areas.

To map the inputs to PA^Esr1+MPN and PA^Esr1+VMHvl cells, we induced cell-type-specific TRIO (tracing the relationship between input and output) method⁴¹. One week after rabies injection, the brain was harvested for histological analysis. Every third section (50-μm thickness) was quantitated and all GFP-positive input cells were counted excluding the site of EnvA-G-Deleted Rabies-eGFP injection.

LCM RNA-Seq

Red and Green retrobeads (Lumafluor Item, R170 and G170) were injected into the ipsilateral MPN and VMHvl in C57BL/6N mice, respectively. One week later, mice were deeply anesthetized with isoflurane and brains were harvested and flash frozen in chilled 2-methylbutane (approximately −80 °C) within 5 min and stored at −80 °C until sectioning. 20-μm coronal sections were obtained using a cryostat and mounted on PEM membrane slides (Leica microsystems, 11505189). The slides were pretreated with 100% ethanol for 1 min and air dried. PA^MPN and PA^VMHvl subpopulations were laser-dissected based on the retrograde labeling using a LMD6000 fluorescence microscope system (Leica microsytems). The RNA samples and cDNA libraries were prepared with a PicoPure RNA isolation kit (ThermoFisher, KIT0204) and Clontech SMARTer Stranded Total RNA-Seq Kit (Cat# 635006) as previously described ³⁰. The cDNA libraries were sequenced with Illumina HiSeq 4000 using high output mode to achieve greater depth of coverage. Sequencing reads were aligned to the mouse genomes (build mm10/GRCm38) using the splice-aware STAR aligner⁵⁸ and then the counts for each gene were generated using featureCounts^59,60. Differential gene expression analysis between PA and BMAp or PA^MPN and PA^VMHvl populations was performed using the limma R package⁶¹ with normalized gene reads through an open-source software iDEP (http://ge-lab.org/idep/)⁶². We selected whole PA region enriched genes with a threshold of false discovery rate (FDR) < 0.05 and fold-change > 1.5 between PA and BMAp populations, and subpopulation-specific genes in the PA with FDR < 0.07 and fold-change > 1.5 between PA^MPN and PA^VMHvl populations.

Statistics

All statistical analyses for histological, in vivo and in vitro recording and behavioral experiments were performed using MATLAB2018a (Mathworks) or Prism7 software (GraphPad). Except for repeated measure two-way ANOVA, all data sets were tested for normality with Kolmogorov-Smirnov test. For one-way ANOVA, Brown-Forsythe test was used to test equality of variances. If the data set passed the normality test, parametric tests (unpaired t-test, paired t-test, repeated measure and ordinary one-way ANOVA with Geisser-Greenhouse correction and Turkey’s multiple comparison post-hoc test, ordinary two-way ANOVA with Bonferroni’s multiple comparison post-hoc test) were used, otherwise, non-parametric tests (Mann Whitney test, Wilcoxon matched-pairs signed rank test, Friedman test with Dunn’s multiple comparison post-hoc test) were used. Following repeated measure two-way ANOVA, post-hoc Bonferroni’s multiple comparison test was performed. Differential gene expression between brain regions was analyzed by two-tailed t-test with Benjamini-Hochberg false discovery rate correction. All significant statistical results were indicated on the figures with the following conventions: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. All error bars represent ± s.e.m. No statistical methods were used to pre-determine sample sizes but our sample sizes are similar to those reported in previous publications^{6,7,9,14,29,30}. Additional details of statistics can be found in Supplementary Table 1.

Life Sciences Reporting Summary

Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Data availability

We have uploaded the RNAseq data from this manuscript to the Gene Expression Omnibus (GEO) under accession number GSE 151798. All other data that support the findings of this study are available from the corresponding authors upon request.

Code availability

MATLAB code ScanImage is available at https://github.com/bernardosabatini/SabalabAcq. MATLAB codes for behavioral annotation and tracking are available from https://github.com/pdollar/toolbox. Custom TDT OpenEx programs and MATLAB codes for fiber photometry data analysis will be provided upon request to the corresponding authors.

1605430_Supp_Tab1

1605430_Supp_Tab2

• 1.Coolen LM & Wood RI Testosterone stimulation of the medial preoptic area and medial amygdala in the control of male hamster sexual behavior: redundancy without amplification. Behav Brain Res 98, 143–153, doi: 10.1016/s0166-4328(98)00063-1 (1999). [DOI] [PubMed] [Google Scholar]
• 2.Arendash GW & Gorski RA Effects of discrete lesions of the sexually dimorphic nucleus of the preoptic area or other medial preoptic regions on the sexual behavior of male rats. Brain Res Bull 10, 147–154, doi: 10.1016/0361-9230(83)90086-2 (1983). [DOI] [PubMed] [Google Scholar]
• 3.Klaric JS & Hendricks SE Effects of two-stage lesions of the medial preoptic area on sexual behavior of male rats. Physiol Behav 37, 539–542, doi: 10.1016/0031-9384(86)90281-7 (1986). [DOI] [PubMed] [Google Scholar]
• 4.Malsbury CW Facilitation of male rat copulatory behavior by electrical stimulation of the medial preoptic area. Physiol Behav 7, 797–805, doi: 10.1016/0031-9384(71)90042-4 (1971). [DOI] [PubMed] [Google Scholar]
• 5.Wei YC et al. Medial preoptic area in mice is capable of mediating sexually dimorphic behaviors regardless of gender. Nat Commun 9, 279, doi: 10.1038/s41467-017-02648-0 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 6.Lin D et al. Functional identification of an aggression locus in the mouse hypothalamus. Nature 470, 221–226, doi: 10.1038/nature09736 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 7.Lee H et al. Scalable control of mounting and attack by Esr1+ neurons in the ventromedial hypothalamus. Nature 509, 627–632, doi: 10.1038/nature13169 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 8.Yang CF et al. Sexually dimorphic neurons in the ventromedial hypothalamus govern mating in both sexes and aggression in males. Cell 153, 896–909, doi: 10.1016/j.cell.2013.04.017 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 9.Yang T et al. Social Control of Hypothalamus-Mediated Male Aggression. Neuron, doi: 10.1016/j.neuron.2017.06.046 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 10.Falkner AL, Grosenick L, Davidson TJ, Deisseroth K & Lin D Hypothalamic control of male aggression-seeking behavior. Nature neuroscience 19, 596 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 11.Wang L et al. Hypothalamic control of conspecific self-defense. Cell reports 26, 1747–1758. e1745 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 12.Stagkourakis S et al. A neural network for intermale aggression to establish social hierarchy. Nat Neurosci 21, 834–842, doi: 10.1038/s41593-018-0153-x (2018). [DOI] [PubMed] [Google Scholar]
• 13.Bergan JF, Ben-Shaul Y & Dulac C Sex-specific processing of social cues in the medial amygdala. Elife 3, e02743, doi: 10.7554/eLife.02743 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 14.Hong W, Kim DW & Anderson DJ Antagonistic control of social versus repetitive self-grooming behaviors by separable amygdala neuronal subsets. Cell 158, 1348–1361, doi: 10.1016/j.cell.2014.07.049 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 15.Lehman MN, Winans SS & Powers JB Medial nucleus of the amygdala mediates chemosensory control of male hamster sexual behavior. Science 210, 557–560, doi: 10.1126/science.7423209 (1980). [DOI] [PubMed] [Google Scholar]
• 16.Li Y et al. Neuronal Representation of Social Information in the Medial Amygdala of Awake Behaving Mice. Cell 171, 1176–1190 e1117, doi: 10.1016/j.cell.2017.10.015 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 17.Unger EK et al. Medial amygdalar aromatase neurons regulate aggression in both sexes. Cell Rep 10, 453–462, doi: 10.1016/j.celrep.2014.12.040 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 18.Miller SM, Marcotulli D, Shen A & Zweifel LS Divergent medial amygdala projections regulate approach-avoidance conflict behavior. Nat Neurosci 22, 565–575, doi: 10.1038/s41593-019-0337-z (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 19.Dulac C & Wagner S Genetic analysis of brain circuits underlying pheromone signaling. Annu Rev Genet 40, 449–467, doi: 10.1146/annurev.genet.39.073003.093937 (2006). [DOI] [PubMed] [Google Scholar]
• 20.Coolen LM, Peters HJ & Veening JG Distribution of Fos immunoreactivity following mating versus anogenital investigation in the male rat brain. Neuroscience 77, 1151–1161, doi: 10.1016/s0306-4522(96)00542-8 (1997). [DOI] [PubMed] [Google Scholar]
• 21.Kim Y et al. Mapping social behavior-induced brain activation at cellular resolution in the mouse. Cell Rep 10, 292–305, doi: 10.1016/j.celrep.2014.12.014 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 22.Dong HW & Swanson LW Projections from bed nuclei of the stria terminalis, posterior division: implications for cerebral hemisphere regulation of defensive and reproductive behaviors. J Comp Neurol 471, 396–433, doi: 10.1002/cne.20002 (2004). [DOI] [PubMed] [Google Scholar]
• 23.Bayless DW et al. Limbic Neurons Shape Sex Recognition and Social Behavior in Sexually Naive Males. Cell 176, 1190–1205 e1120, doi: 10.1016/j.cell.2018.12.041 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 24.Brody J, DeFeudis P & DeFeudis FJN Effects of micro-injections of L-glutamate into the hypothalamus on attack and flight behaviour in cats. Nature 224, 1330–1330 (1969). [DOI] [PubMed] [Google Scholar]
• 25.Dominguez JM, Gil M & Hull EM Preoptic glutamate facilitates male sexual behavior. J Neurosci 26, 1699–1703, doi: 10.1523/JNEUROSCI.4176-05.2006 (2006). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 26.Canteras NS, Simerly RB & Swanson LW Connections of the posterior nucleus of the amygdala. J Comp Neurol 324, 143–179, doi: 10.1002/cne.903240203 (1992). [DOI] [PubMed] [Google Scholar]
• 27.Watson C, Paxinos G & Puelles L The mouse nervous system. 1st edn, 140–72 (Elsevier Academic Press, 2012). [Google Scholar]
• 28.Simerly RB Wired for reproduction: organization and development of sexually dimorphic circuits in the mammalian forebrain. Annu Rev Neurosci 25, 507–536, doi: 10.1146/annurev.neuro.25.112701.142745 (2002). [DOI] [PubMed] [Google Scholar]
• 29.Fang YY, Yamaguchi T, Song SC, Tritsch NX & Lin D A Hypothalamic Midbrain Pathway Essential for Driving Maternal Behaviors. Neuron 98, 192–207 e110, doi: 10.1016/j.neuron.2018.02.019 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 30.Hashikawa K et al. Esr1(+) cells in the ventromedial hypothalamus control female aggression. Nat Neurosci 20, 1580–1590, doi: 10.1038/nn.4644 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 31.Simerly RB, Chang C, Muramatsu M & Swanson LW Distribution of androgen and estrogen receptor mRNA-containing cells in the rat brain: an in situ hybridization study. J Comp Neurol 294, 76–95, doi: 10.1002/cne.902940107 (1990). [DOI] [PubMed] [Google Scholar]
• 32.Ervin KS et al. Estrogen involvement in social behavior in rodents: Rapid and long-term actions. Horm Behav 74, 53–76, doi: 10.1016/j.yhbeh.2015.05.023 (2015). [DOI] [PubMed] [Google Scholar]
• 33.Kollack-Walker S & Newman SW Mating and agonistic behavior produce different patterns of Fos immunolabeling in the male Syrian hamster brain. Neuroscience 66, 721–736, doi: 10.1016/0306-4522(94)00563-k (1995). [DOI] [PubMed] [Google Scholar]
• 34.Swanson LW Cerebral hemisphere regulation of motivated behavior. Brain Res 886, 113–164, doi: 10.1016/s0006-8993(00)02905-x (2000). [DOI] [PubMed] [Google Scholar]
• 35.Wang L et al. Hypothalamic Control of Conspecific Self-Defense. Cell Rep 26, 1747–1758 e1745, doi: 10.1016/j.celrep.2019.01.078 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 36.Wang F et al. RNAscope: a novel in situ RNA analysis platform for formalin-fixed, paraffin-embedded tissues. J Mol Diagn 14, 22–29, doi: 10.1016/j.jmoldx.2011.08.002 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 37.Gunaydin LA et al. Natural neural projection dynamics underlying social behavior. Cell 157, 1535–1551, doi: 10.1016/j.cell.2014.05.017 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 38.Cui G et al. Concurrent activation of striatal direct and indirect pathways during action initiation. Nature 494, 238–242, doi:nature11846 [pii] 10.1038/nature11846 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 39.Knowland D et al. Distinct Ventral Pallidal Neural Populations Mediate Separate Symptoms of Depression. Cell 170, 284–297 e218, doi: 10.1016/j.cell.2017.06.015 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 40.Alexander GM et al. Remote control of neuronal activity in transgenic mice expressing evolved G protein-coupled receptors. Neuron 63, 27–39, doi: 10.1016/j.neuron.2009.06.014 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 41.Schwarz LA et al. Viral-genetic tracing of the input-output organization of a central noradrenaline circuit. Nature 524, 88–92, doi: 10.1038/nature14600 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 42.Canteras NS, Simerly RB & Swanson LW Organization of projections from the medial nucleus of the amygdala: a PHAL study in the rat. J Comp Neurol 360, 213–245, doi: 10.1002/cne.903600203 (1995). [DOI] [PubMed] [Google Scholar]
• 43.Petrovich GD, Canteras NS & Swanson LW Combinatorial amygdalar inputs to hippocampal domains and hypothalamic behavior systems. Brain Res Brain Res Rev 38, 247–289 (2001). [DOI] [PubMed] [Google Scholar]
• 44.Zha X et al. VMHvl-Projecting Vglut1+ Neurons in the Posterior Amygdala Gate Territorial Aggression. Cell Rep 31, 107517, doi: 10.1016/j.celrep.2020.03.081 (2020). [DOI] [PubMed] [Google Scholar]
• 45.Emery DE & Sachs BD Copulatory behavior in male rats with lesions in the bed nucleus of the stria terminalis. Physiol Behav 17, 803–806, doi: 10.1016/0031-9384(76)90044-5 (1976). [DOI] [PubMed] [Google Scholar]
• 46.Liu YC, Salamone JD & Sachs BD Lesions in medial preoptic area and bed nucleus of stria terminalis: differential effects on copulatory behavior and noncontact erection in male rats. J Neurosci 17, 5245–5253 (1997). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 47.Shimura T, Yamamoto T & Shimokochi M The medial preoptic area is involved in both sexual arousal and performance in male rats: re-evaluation of neuron activity in freely moving animals. Brain Res 640, 215–222, doi: 10.1016/0006-8993(94)91875-9 (1994). [DOI] [PubMed] [Google Scholar]
• 48.Hoffman NW, Kim YI, Gorski RA & Dudek FE Homogeneity of intracellular electrophysiological properties in different neuronal subtypes in medial preoptic slices containing the sexually dimorphic nucleus of the rat. J Comp Neurol 345, 396–408, doi: 10.1002/cne.903450306 (1994). [DOI] [PubMed] [Google Scholar]
• 49.Hatayama M et al. Zic2 hypomorphic mutant mice as a schizophrenia model and ZIC2 mutations identified in schizophrenia patients. Sci Rep 1, 16, doi: 10.1038/srep00016 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 50.Lewis AS et al. Bidirectional Regulation of Aggression in Mice by Hippocampal Alpha-7 Nicotinic Acetylcholine Receptors. Neuropsychopharmacology 43, 1267–1275, doi: 10.1038/npp.2017.276 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 51.Paxinos G & Franklin KBJ The mouse brain in stereotaxic coordinates. 3rd edn, (Elsevier Academic Press, 2008). [Google Scholar]
• 52.Vong L et al. Leptin action on GABAergic neurons prevents obesity and reduces inhibitory tone to POMC neurons. Neuron 71, 142–154, doi: 10.1016/j.neuron.2011.05.028 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 53.Madisen L et al. A robust and high-throughput Cre reporting and characterization system for the whole mouse brain. Nat Neurosci 13, 133–140, doi: 10.1038/nn.2467 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 54.Yamaguchi T, Danjo T, Pastan I, Hikida T & Nakanishi S Distinct roles of segregated transmission of the septo-habenular pathway in anxiety and fear. Neuron 78, 537–544, doi: 10.1016/j.neuron.2013.02.035 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 55.Mathis A et al. DeepLabCut: markerless pose estimation of user-defined body parts with deep learning. Nat Neurosci 21, 1281–1289, doi: 10.1038/s41593-018-0209-y (2018). [DOI] [PubMed] [Google Scholar]
• 56.Nilsson SR et al. Simple Behavioral Analysis (SimBA) – an open source toolkit for computer classification of complex social behaviors in experimental animals. 2020.2004.2019.049452, doi: 10.1101/2020.04.19.049452 %J bioRxiv (2020). [DOI] [Google Scholar]
• 57.Wang L, Chen IZ & Lin D Collateral pathways from the ventromedial hypothalamus mediate defensive behaviors. Neuron 85, 1344–1358, doi: 10.1016/j.neuron.2014.12.025 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 58.Dobin A et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29, 15–21, doi: 10.1093/bioinformatics/bts635 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 59.Cimmino L et al. Restoration of TET2 Function Blocks Aberrant Self-Renewal and Leukemia Progression. Cell 170, 1079–1095 e1020, doi: 10.1016/j.cell.2017.07.032 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 60.Liao Y, Smyth GK & Shi W featureCounts: an efficient general purpose program for assigning sequence reads to genomic features. Bioinformatics 30, 923–930, doi: 10.1093/bioinformatics/btt656 (2014). [DOI] [PubMed] [Google Scholar]
• 61.Ritchie ME et al. limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res 43, e47, doi: 10.1093/nar/gkv007 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 62.Ge SX, Son EW & Yao R iDEP: an integrated web application for differential expression and pathway analysis of RNA-Seq data. BMC Bioinformatics 19, 534, doi: 10.1186/s12859-018-2486-6 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]

Supplementary Materials

Data Availability Statement

We have uploaded the RNAseq data from this manuscript to the Gene Expression Omnibus (GEO) under accession number GSE 151798. All other data that support the findings of this study are available from the corresponding authors upon request.