Hyperactivity of the Lateral Septum Leads to Hypersensitivity in Susceptible Mice

Everyone needs social interactions, not only humans but also most species. Positive social interaction feedback is a reward, promoting more social behaviors and acting as a stress buffer. On the contrary, negative social interactions such as school bullying may become a type of stress that leads to mental illnesses including anxiety and depression disorders. Chronic social defeat stress (CSDS) is a widely accepted social trauma model in rodents based on repeated social attacks, following which the defeated mice can be classified as susceptible (SUS) and resilient (RES) according to their social performance. SUS mice rather than RES mice show significant social avoidance, anhedonia, weight loss, and endocrine disruption [1]. At present, the CSDS model is commonly used in the study of anxiety and depression, but its social effects are rarely studied. Li and colleagues’ work sheds light on how social defeats affect responses to social information, especially social reward [2].

Recent studies have suggested that social avoidance caused by chronic social failure is controlled by the suppression of mesolimbic dopaminergic reward centers: the ventral tegmental area (VTA) and the nucleus accumbens (NAc) [3, 4]. Besides reward deficiency, the mice being attacked have fear-enhancing phenotypes associated with the activity of the oxytocin pathway in the lateral septum (LS) [5]. Similarly, the activation of GABAergic neurons in the LS by acute stress causes hyperalgesia and anxiety-like behaviors [6]. However, most studies have focused on interactions between social defeat stress and mental abnormalities, such as depression-like behaviors. It is not clear how social trauma affects social rewards at the cellular and circuitry levels. Recently published in Nature, Li et al. revealed that a population of LS neurotensin (NT^LS) neurons are hyperactivated following CSDS, gating the susceptibility to the stress and occluding social reward effects.

Based on the social interaction ratio, mice that underwent the CSDS were divided into RES and SUS groups, with SUS mice exhibiting social avoidance even toward non-threatening juvenile mice. At the same time, RES mice were more willing to take social interactions with juveniles which were recognized as rewards to them. Taking the learning and memory process into consideration, Li et al. found no deficits in non-social cognitive tests, supporting the conclusion that CSDS induces a specific disruption in social rewards but not in general reward processing. To characterize the neuronal types mediating the susceptibility, Li et al. performed ex vivo high-throughput screening of Fos with iDISCO⁺ and ClearMap assay on samples collected post-exposure to CSDS and found over 94% colocalization between Nt and Fos in SUS mice, while there was very limited overlap with Crhr2⁺, Oxtr⁺, Sst⁺, and Drd3⁺ neurons which were all functionally expressed in the LS. Moreover, NT neurons were specifically enriched in the lateral-ventral part of the LS, where the Fos accumulated. These findings confirmed that NT neurons that accumulated in the LS were highly activated in SUS mice, implying a role of NT^LS neurons in gating social reward and social avoidance (Fig. 1).

Fig. 1.

NTLS neurons in socially defeated mice gate social reward processing. Li et al. showed that NTLS neurons in SUS mice were hyperactivated during the juvenile-paired session. Similar phenotypes were induced in RES mice by activating NTLS neurons with hM3Dq. In parallel, inhibition of NTLS neurons with hM4Di in SUS mice reversed the decreased social interaction that resulted from CSDS.

The LS has been demonstrated to be a nexus for mood and motivation, where NT neurons are enriched to modulate neuroendocrine activity and feeding behaviors [7]. However, it is not well-understood whether NT^LS neurons respond to CSDS and how they control the social reward processing. Li et al. aimed to determine the association of NT^LS hyperactivation with CSDS susceptibility in vivo. Combining fiber photometry with Nt-Cre transgenic mice, Li et al. examined the real-time NT^LS neuronal activity of CSDS mice during their social interactions with juveniles, finding significantly higher fluorescent Ca²⁺ activity in SUS mice rather than RES or control (CTRL) mice. Notably, analogous elevated neuronal activity was observed when unstressed CTRL mice were experiencing tail pinch and social attack, either in female or male mice. These findings suggest that SUS mice may take juveniles as social threats as a result of the NT^LS neuronal hyperactivation.

Next, to establish a causal relationship between the activity of NT^LS neurons and changes in social behaviors after CSDS, Li et al. bidirectionally manipulated NT^LS neuronal activity using the designer receptors exclusively activated by designer drugs (DREADDs) strategy. As demonstrated above, higher activity of NT^LS neurons was related to the social reward deficits in SUS mice. In order to mimic the hyperactivation state, Li et al. activated NT^LS neurons with hM3Dq tools in RES mice following CSDS. As expected, SUS-like phenotypes were induced in hM3Dq-RES mice, represented by reduced interaction durations in multiple behavioral tests. In parallel, inhibition of NT^LS neurons with hM4Di in SUS mice reversed the decreased social interaction that resulted from CSDS. Moreover, manipulation of NT^LS neurons had no significant effects on the general preference test without a social context. Consistent effects were found in mice of both sexes, verifying the modulation of NT^LS neurons in social behavioral changes induced by CSDS. To further assess the actions of NT^LS neurons in responding to stress, Li et al. replaced CSDS with non-social stressors such as chronic restraint stress, finding that anxiety-like behaviors rather than social interactions were modified. Overall, Li et al. showed that NT^LS neuronal activity could be a bridge between social stress and social reward.

The LS provides widespread innervations to those brain regions regulating distinct behaviors. Li et al. took advantage of viral tracing and chemogenetic neuro-control to dissect the NT^LS downstream targets mediating social behaviors. Among multiple regions identified, they mainly focused on the NAc, the anterior hypothalamic nucleus (AHN), the lateral hypothalamus, and the periaqueductal grey (PAG), based on their well-known social and reward involvement. With little overlap visualized on the application of triple stains to the NAc, AHN, and PAG, they demonstrated that LS neurons separately project to these regions, which was further confirmed by retrograde viral tracing. Li et al. also examined the functions of these NT^LS circuits with optogenetics, showing that activating the NT^LS→AHN or NT^LS→NAc circuit, rather than the NT^LS→PAG circuit reduces the duration of social investigation, while none of them had significant effects on social avoidance.

Previously, it had been shown that acute stress can activate LS GABAergic neurons to induce anxiety-like behaviors [6]. Therefore, LS hyperactivation in the social defeat model is not an entirely new concept. Nonetheless, Li and colleagues’ study is remarkable as it reveals that NT^LS neurons in SUS mice are hyperactivated in response to CSDS, resulting in mistakenly perceiving juveniles as threats rather than rewards. Their study provides theoretical support for the clinical use of radiological examination to screen individuals with high susceptibility to social stress. It is also significant in offering potential targets for the medical diagnosis and treatment of social deficits accompanying post-traumatic stress disorder.

Meanwhile, several intriguing questions are raised. First, the NT is a neuropeptide known to modulate GABAergic neurons by increasing GABA release [8]. Yet, the GABA and NT signaling works in different modes: GABA is released at the basal activity level, while NT is recruited once neurons are over-activated and act like a brake to suppress over-reaction [7]. In the future, it will be necessary to dissect the actions of GABA and NT separately. This could be addressed by respectively blocking GABA or NT signaling with pharmacology or CRISPR/Cas9 techniques, followed by behavioral analysis to determine the individual effects of GABA and NT on the performance of mice, including susceptibility to social stress, perception of social reward, and anxiety-like behaviors. Second, according to cleared whole-brain Fos mapping, the pedunculopontine nucleus (PPN) and the mediodorsal nucleus of the thalamus (MD) are also highly activated in SUS mice. Since the PPN and MD play roles in reward and decision-making [9], it will be interesting to figure out whether the hyperactivity of these brain regions contributes to the deficits in social reward processing. Last, we human beings share a response to social stress similar to rodents; they can also be divided into SUS and RES subpopulations [10]. It is important to figure out why resilient people can minimize the adverse effects of social stress and how we can find ways to take precautions. Unfortunately, the reason why NT^LS neurons are not hyperactivated in RES mice has not been examined. It will be interesting to investigate whether NT^LS neurons in RES mice are suppressed by upstream inhibitory input or just blunted by an elevated threshold. However, either hypothesis needs further evidence from neural circuit mapping and electrophysiological assays.

In conclusion, this study broadens our understanding of the LS neural circuits underlying social reward deficits after social defeat stress, providing new insights for the clinical treatment of post-traumatic stress disorder and related psychiatric diseases.