Distinct oxytocin signaling pathways synergistically mediate rescue-like behavior in mice

Abstract

Spontaneous rescue behavior enhances the well-being and survival of social animals, yet the neural mechanisms underlying the recognition and response to conspecifics in need remain unclear. Here, we report that observer mice experience distress when encountering anesthetized conspecifics, prompting spontaneous rescue-like behavior toward the unconscious mice. This behavior facilitates the earlier awakening of anesthetized mice while simultaneously alleviating stress in the helper mice. Our findings reveal that endogenous oxytocin (OXT) release from the hypothalamic paraventricular nucleus (PVN) to the oxytocin receptor (OXTR) in the central nucleus of the amygdala (CeA) regulates the emotional component of rescue-like behavior. In contrast, OXT release from the PVN to OXTR in the dorsal bed nucleus of the stria terminalis (dBNST) mediates the motor component of the behavior. Furthermore, we demonstrate that these two pathways exhibited distinct temporal dynamics and functional roles. The OXT^PVN-OXTR^CeA pathway is activated in a transient and intense manner, acting as a trigger for rescue-like behavior, whereas the OXT^PVN-OXTR^dBNST pathway responds in a sustained manner, ensuring the continuation of the behavior. These findings highlight the remarkable ability of rodents to engage in targeted helping behavior and suggest that distinct subcortical oxytocinergic pathways selectively and synergistically regulate the motor and emotional aspects of rescue-like behavior.

Keywords: BNST; amygdala; anesthesia; oxytocin; rescue-like behavior.

Fig. 1.

Mouse rescue-like behavioral paradigm. (A) Experimental schematic: showing observer mouse interacts with an anesthetized demonstrator, performing grooming and licking behaviors. LORR and RORR indicate the demonstrator’s anesthesia states. (B) Representative images of the observer’s body-grooming/licking (Left) and facial-licking (Middle: licking eyes, Right: licking mouth). (C) The ratio of the subtype of behaviors in male and female observers (n = 12 for each gender). (D) Time course of social licking (n = 12 for each gender). (E) Social licking time toward the cagemate, stranger, or dead mouse (n = 12 for cagemate and stranger groups; n = 9 for dead group). (F) Different ages of observers’ social licking time (n = 4 males + 4 females for each group). (G) Cumulative probability of demonstrators’ RORR latency. (H) Demonstrators’ RORR latency with and without observers’ rescue-like behavior (n = 8 males + 7 females for each group). (I) EEG recording schematic. (J) Time-frequency spectra of EEG in demonstrator with/without rescue-like behavior. (K) Empirical CDF of delta and gamma powers in demonstrators with/without rescue-like behavior (n = 5 for each group, all male). (L) Emotional discrimination test schematic. (M) Time spent on the side of the anesthetized or naïve mouse (n = 5 males + 6 females). (N) Schematic of divider social paradigm. (O) Plasma corticosterone levels in the divided-chamber test (n = 6 for saline + divider group, n = 7 for anesthetic + divider group, all male). (P and Q) Approach (P) and social licking time (Q) after saline or CP154526 injection (n = 6 males + 4 females for each group). Statistical analyses included paired-sample t test (M), independent-sample t test (F, H, P, and Q), two-way ANOVA with Tukey’s multiple comparison (E and O), and independent-sample Kolmogorov–Smirnov test (K). *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001; ns, no significance. Data are means ± SEM. ANOVA, analysis of variance; EEG, electroencephalogram; ELISA, enzyme-linked immunosorbent assay; CDF, cumulative distribution functions; LORR, loss of righting reflex; RORR, resumption of righting reflex.

Fig. 2.

OXT^PVN signaling mediates rescue-like behavior. (A–D) Approach (A and C) and social licking time (B and D) for WT and Oxt KO observers (n = 9 for each male group, n = 11 for each female group). (E) Schematic of fiber photometry recording of PVN^OXT neurons. (F and G) Averaged trace (F) and mean ΔF/F₀ (G) of OXT^PVN Ca²⁺ signals during the social approach to anesthetized or naïve mouse (n = 10 trials for each group). (H and I) Averaged trace (H) and mean ΔF/F₀ (I) of OXT^PVN Ca²⁺ signals perionset of social licking (n = 57 trials). (J) Schematic of PVN virus injection in the Oxt^Cre mice for chemogenetic inhibition. (K–N) Approach (K and M) and social licking time (L and N) for Gi− and Gi+ observers (n = 10 for each male group, n = 8 for each female group). Statistical analyses included paired-sample t test (G), independent-sample t test (A–D, I, and K–N). *P < 0.05, **P < 0.01, ****P < 0.0001. Data are means ± SEM. OXT, oxytocin; PVN, paraventricular nucleus of the hypothalamus.

Fig. 3.

OXT^PVN-OXTR^CeA projection regulates the emotional process of rescue-like behavior. (A) Schematic of fiber photometry recording of OXTR^CeA neurons. (B and C) Averaged trace (B) and mean ΔF/F₀ (C) of OXTR^CeA Ca²⁺ signals during the social approach to anesthetized or naïve mouse (n = 18 trials for naïve, n = 14 for the anesthetized). (D and E) Averaged trace (D) and mean ΔF/F₀ (E) of OXTR^CeA Ca²⁺ signals perionset of social licking (n = 75 trials). (F) Schematic of optogenetic inhibition of OXTR^CeA neurons. (G and H) Approach (G) and social licking time (H) with long-term optogenetic inhibition of OXTR^CeA Jaws or control (n = 6 males + 6 females for each group). (I) Schematic of optogenetic inhibition of OXT^PVN efferents in CeA. (J and K) Approach (J) and social licking time (K) with long-term optogenetic inhibition of PVN^OXT-CeA Jaws or control (n = 6 males + 5 females for each group). Statistical analyses included paired-sample t test (E) and independent-sample t test (C, G, H, J, and K). *P < 0.05, ****P < 0.0001. Data are means ± SEM. CeA, central nucleus of the amygdala.

Fig. 4.

The OXT^PVN-OXTR^CeA peptidergic pathway mediates stress-related emotional function in rescue-like behavior. (A) Schematic of fiber photometry recording of OXT-sensor. (B and C) Averaged trace (B) and mean ΔF/F₀ (C) of OXT-sensor in CeA during the social approach to the anesthetized or naïve mouse (n = 10 trials for each group). (D and E) Averaged trace (D) and mean ΔF/F₀ (E) of OXT-sensor in CeA perionset of social licking (n = 56 trials). (F) Schematic of virus injection into the PVN and CeA of Oxtr^f/f mice. (G and H) Approach (G) and social licking time (H) for control or Oxtr^PVN-CeA CKO mice (n = 8 males + 7 females for each group). (I) Plasma corticosterone level at 20 min of rescue-like behavior test for control or Oxtr^PVN-CeA CKO mice (n = 6 males for each group). (J and K) Emotional discrimination test schematic (J) and time on the anesthetized mouse side for control or Oxtr^PVN-CeA CKO mice (K) (n = 8 males + 7 females for each group). Statistical analyses included paired-sample t test (E) and independent-sample t test (B, G–I, and K). **P < 0.01, ***P < 0.001, ****P < 0.0001; ns, no significance. Data are means ± SEM.

Fig. 5.

OXT^PVN-OXTR^dBNST projection regulates licking-related motor function in rescue-like behavior. (A) Schematic of fiber photometry recording of OXTR^dBNST neurons. (B and C) Averaged trace (B) and mean ΔF/F₀ (C) of OXTR^dBNST Ca²⁺ signals during the social approach to the anesthetized or naïve mouse (n = 9 trials for naïve, n = 7 for anesthetized). (D and E) Averaged trace (D) and mean ΔF/F₀ (E) of OXTR^dBNST Ca²⁺ signals perionset of social licking (n = 58 trials). (F) Schematic of optogenetic activation of OXTR^dBNST neurons. (G and H) Licking behavior ratio in social (G) or nonsocial (H) context with activation of OXTR^dBNST ChR2 or control (n = 5 males + 5 females for each group). (I) Self-grooming ratio with activation onset in OXTR^dBNST ChR2 or control (n = 5 males + 5 females for each group). (J) Schematic of optogenetic inhibition of OXTR^dBNST neurons. (K) Interruption ratio of licking/grooming behavior with real-time inhibition of OXTR^dBNST Jaws or control (n = 5 males + 5 females for each group). (L) Schematic of optogenetic activation of OXT^PVN efferents in dBNST. (M) Licking behavior ratio with activation of OXT^PVN-dBNST ChR2 or control (n = 6 males + 4 females for each group). (N and O) Schematic (N) and behavioral results (O) of RTPP test for OXT^PVN-dBNST ChR2 or control mice (n = 6 males + 4 females for each group). (P) Schematic of optogenetic inhibition of OXT^PVN efferents in dBNST. (Q) Interruption ratio of licking/grooming behavior with inhibition of OXT^PVN-dBNST Jaws or control (n = 6 males + 4 females for each group). Statistical analyses included paired-sample t test (E), independent-sample t test (C, G–I, K, M, and Q), and two-way ANOVA with Tukey’s multiple comparison (O). **P < 0.01, ****P < 0.0001; ns, no significance. Data are means ± SEM. dBNST, dorsal bed nucleus of the stria terminalis. RTPP, real-time place preference.

Fig. 6.

OXT^PVN-OXTR^dBNST peptidergic pathway mediates motor but not emotional function in rescue-like behavior. (A) Schematic of fiber photometry recording of OXT-sensor. (B and C) Averaged trace (B) and mean ΔF/F₀ (C) of OXT-sensor in dBNST during the social approach to the anesthetized or naïve mouse (n = 10 trials for each group). (D and E) Averaged trace (D) and mean ΔF/F₀ (E) of OXT-sensor in dBNST perionset of social licking (n = 36 trials). (F) Schematic of virus injection into the PVN and dBNST of Oxtr^f/f mice. (G and H) Approach (G) and social licking time (H) for control or Oxtr^PVN-dBNST CKO mice (n = 5 males + 5 females for each group). (I and J) emotional discrimination test schematic (I) and time spent (J) on the naïve or anesthetized mouse side for control or Oxtr^PVN-dBNST CKO mice (n = 5 males + 5 females for each group). Statistical analyses included paired-sample t test (E), independent-sample t test (C, G, and H), and two-way ANOVA with Tukey’s multiple comparison (J). *P < 0.05, **P < 0.01; ns, no significance. Data are means ± SEM.

Fig. 7.

OXT^PVN-CeA and OXT^PVN-dBNST signaling pathways display distinct activity patterns during rescue-like behavior. (A) An example trace showing photometry recorded signal width and peak height in relation to the rescue-like behavior time in one licking/grooming bout. (B and C) The linear regression of OXTR^dBNST Ca2⁺ signal width (B) and peak height (C) with the behavior time. (D and E) The linear regression of dBNST OXT-sensor signal width (D) and peak height (E) with the behavior time. (F and G) The linear regression of OXTR^CeA Ca2⁺ signal width (F) and peak height (G) with the behavior time. (H and I) The linear regression of CeA OXT-sensor signal width (H) and peak height (I) with the behavior time. (J) Schematic depicting distinct subcortical OXT signaling pathways that independently control motor function of licking behavior and emotion component of rescue-like behavior. Statistical analyses used Pearson r correlation in B–I.