Prostaglandin signaling drives peripheral inflammation-induced reduction of hypothalamic oxytocin-positive neurons

Yibing Li; Hao Xu; Hongzhi Guo; Mingrui Zhai; Li Cao; Lei Xiao

doi:10.1007/s00018-026-06248-4

. 2026 May 12;83(1):271. doi: 10.1007/s00018-026-06248-4

Prostaglandin signaling drives peripheral inflammation-induced reduction of hypothalamic oxytocin-positive neurons

Yibing Li ¹, Hao Xu ¹, Hongzhi Guo ¹, Mingrui Zhai ², Li Cao ^3,^✉, Lei Xiao ^1,^2,^3,^✉

PMCID: PMC13338091 PMID: 42120781

Abstract

Oxytocin (OXT) is a pleiotropic neuropeptide with diverse physiological functions, including anti-inflammatory effects. Endogenous OXT, primarily produced by neurons in hypothalamic paraventricular nucleus (PVN) and supraoptic nucleus (SON), is known to be reduced in various pathological states, particularly during peripheral inflammatory infections. However, the mechanisms by which peripheral inflammation leads to reduced OXT signaling remain poorly understood. In a mouse model of lipopolysaccharide (LPS)-induced chronic inflammation, we observed a selective reduction in the number of magnocellular (Magno) OXT-immunopositive neurons in the PVN, with no significant changes in PVN parvocellular (Parvo) or SON OXT-immunopositive neurons. Electrophysiological recordings revealed hyperexcitability of PVN OXT Magno neurons after LPS treatment, whereas Parvo neurons showed reduced activity. Microglial activation was preferentially localized to Magno neurons-dominant PVN subregions after LPS treatment. Single-cell transcriptomic analysis indicated higher expression of the Ptger4 gene in Magno OXT neurons, and bulk RNA sequencing of PVN and SON tissues highlighted enrichment of prostaglandin-related pathways following LPS challenge. Pharmacological inhibition and genetic knockdown experiments confirmed that prostaglandin E2 (PGE2)-EP4 signaling mediates the reduction of PVN OXT-immunopositive neurons and drives microglial phagocytosis of Magno OXT neurons under inflammatory conditions. Thus, these results not only identify the specific impact of peripheral inflammation on PVN Magno OXT neurons but also uncover the involvement of prostaglandin signaling in this process.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00018-026-06248-4.

Keywords: Oxytocin, Microglia, Prostaglandin signaling, Peripheral inflammation

Introduction

Oxytocin (OXT) is a critical neuropeptide involved in a wide range of physiological processes, including the modulation of inflammation [1]. It exerts broad anti-inflammatory effects through cell-type-specific mechanisms in diverse pathological settings [2–4]. For example, in lipopolysaccharide (LPS)-induced systemic inflammation, OXT acts on immune cells to suppress the release of pro-inflammatory cytokines such as TNF-α and IL-1β [5]. In food allergy-induced intestinal inflammation, OXT targets intestinal epithelial cells and inhibits the production of cytokines, including IL-25 and IL-33, thereby alleviating intestinal inflammation [6]. In a rat model of myocardial infarction, OXT modulates neutrophil and macrophage activity, limiting their infiltration into infarcted areas and enhancing the secretion of TGF-β, which helps resolve cardiac inflammation [7]. Moreover, in several neurological disorders, OXT has been shown to attenuate neuroinflammation and support neuronal survival [3, 4]. These findings suggest that maintaining systemic OXT levels could serve as a protective strategy against immune-related diseases.

Endogenous OXT is primarily synthesized and released by neurons located in the paraventricular (PVN) and supraoptic nuclei (SON) of the hypothalamus [1], and some studies have also reported that OXT can be produced in the periphery, such as the gastrointestinal tract, dorsal root ganglia, and aortic renal ganglia [8–11]. Hypothalamic OXT neurons are broadly classified into magnocellular (Magno) and parvocellular (Parvo) subtypes. Magno OXT neurons are distributed in both PVN and SON, while Parvo OXT neurons are only distributed in the PVN [1]. Magno OXT neurons project to the posterior pituitary and release OXT into the bloodstream, whereas Parvo neurons mainly innervate central regions and mediate OXT signaling within the brain. In several inflammation-related disorders, such as sepsis [4, 12], Parkinson’s disease [13], diabetes [14], and ischemic [3], both plasma and brain OXT levels have been reported to be decreased. A reduction in the number of hypothalamic OXT-immunopositive neurons has also been observed in human patients and animal disease models [13–17]. Notably, in mice prenatally exposed to valproic acid, PVN Parvo OXT neurons appear to be particularly vulnerable [15]. Nevertheless, the mechanisms responsible for the reduction of hypothalamic OXT neurons under pathological conditions, including peripheral inflammation, remain poorly understood.

In this study, we demonstrate that peripheral inflammation specifically induces the reversible reduction of Magno OXT-immunopositive neurons in the PVN. By integrating electrophysiological recording, transcriptomic analyses, pharmacological inhibition, and RNA interference (RNAi), we find that the reduction of PVN OXT-immunopositive neurons is mediated by prostaglandin signaling and microglia phagocytosis. Our findings highlight a potential therapeutic strategy of preserving endogenous OXT signaling to mitigate diseases associated with peripheral inflammation.

Results

The reduction in the number of PVN OXT-immunopositive neurons in the LPS-induced chronic inflammation model

We adopted the four daily repeated injections of 1 mg/kg LPS to establish a chronic peripheral inflammatory challenge in male mice [18] (Fig. 1A). Compared with the saline (SAL)-injected controls, LPS administration resulted in a significant reduction in body weight and elevated plasma TNF-α level (Fig. 1B). Furthermore, the mRNA levels of inflammatory cytokines, including IL-1β and TNF-α, were elevated in the PVN of LPS-treated mice (Fig. 1B). We used the immunostaining method to label the OXT+ neurons in the hypothalamus, and found a significant decrease in the number of OXT+ neurons within the PVN following LPS treatment (87.84% ± 3.26 of control, p = 0.0125, unpaired t-test) (Fig. 1C and D). Consistently, PVN OXT mRNA expression was also decreased in the LPS-treated mice (Fig. 1D). In addition to PVN, OXT+ neurons in the SON were also analyzed, but we did not observe a significant change in SON OXT+ neurons (Fig. 1E). We next examined whether LPS specifically affected OXT neurons by analyzing vasopressin (AVP)-immunopositive neurons, another major neurohypophyseal cell type that shares high sequence homology with OXT neurons [19]. Neither the number of PVN AVP+ neurons nor the level of AVP mRNA was significantly changed after LPS treatment (Fig. 1F). Previous studies have reported the sex differences in both the oxytocin system and neuroimmune interactions [20, 21]. We investigated the influence of chronic inflammation on PVN OXT neurons in female mice and found that LPS treatment similarly reduced the number of OXT+ neurons in the PVN of female mice (Fig. S1A-B), whereas the numbers of SON OXT+ neurons and PVN AVP+ neurons remained unchanged (Fig. S1C-D).

We further established a sepsis model with a single high-dose LPS injection (15 mg/kg) to investigate acute systemic inflammation (Fig. S2A). Notably, this acute challenge also significantly reduced the number of OXT+ neurons in the PVN (Fig. S2B-D), while no significant changes were detected in SON OXT+ neurons or PVN AVP+ neurons (Fig. S2E-F). To explore LPS dose dependence, we measured changes in the number of OXT+ neurons using a single low-dose LPS injection (1 mg/kg), and no significant effect was observed (Fig. S2G-H). These findings suggest that the reduction in PVN OXT+ neurons occurs only when inflammation reaches a certain threshold.

All of these data indicate that peripheral LPS-induced inflammation primarily leads to the reduction of PVN OXT+ neurons, but has no significant effect on PVN AVP+ neurons and SON OXT+ neurons in both male and female mice.

A selective reduction of PVN Magno, but not Parvo, OXT-immunopositive neurons in the LPS-induced chronic inflammation model

Hypothalamic OXT neurons are traditionally divided into magnocellular (Magno) and parvocellular (Parvo) subtypes based on their axonal projections to the posterior pituitary [1], with reported subtype-specific alterations under pathological conditions [15]. To determine whether peripheral neuroinflammation induced by LPS differentially affects these subtypes in the PVN, we intravenously administered FluoroGold (FG) to selectively label Magno OXT neurons [15, 22] (Fig. 2A). In line with earlier reports [15, 23], Magno OXT+ neurons were predominantly localized in the anterior and medial portions of the PVN, whereas Parvo neurons were mainly situated in the posterior region (Fig. 2B, C). LPS treatment induced a marked reduction of OXT+ neurons specifically in the anterior PVN compared with the SAL-injected controls (Fig. 2D, Fig. S2C), suggesting a preferential reduction of Magno OXT+ neurons. This was further confirmed by FG tracing, which revealed a significant decrease in the number of FG+, OXT+ (Magno) neurons, but not FG-, OXT+ (Parvo) neurons, following LPS administration (Fig. 2E).

We further utilized OXT-Cre; Ai3 transgenic mice, which efficiently label OXT+ neurons with eYFP fluorescence signal in the PVN (Fig. 2A and F). In this transgenic mouse line, eYFP expression is driven by OXT promoter-dependent Cre activity [24]. Consistent with the decrease in OXT+ neurons, we observed a significant reduction in eYFP-labeled cells in LPS-treated mice compared with controls (Fig. 2F). The reduction in OXT+ neurons may reflect either a reduced neuropeptide synthetic capacity or actual neuronal loss [15]. We first assessed the total number of PVN neurons in LPS-treated mice by Nissl staining and found that LPS treatment did not alter the total number of PVN neurons (Fig. S3A-B). Furthermore, we evaluated PVN OXT+ neurons at day 10 post-LPS treatment (Fig. 2G). Both serum TNF-α levels and the number of PVN OXT+ neurons returned to baseline by 10 days post-LPS treatment (Fig. 2H and Fig. S3C). Different from OXT+ neuronal change, serum OXT levels were increased at 12 h after LPS challenge but decreased at 10 days after LPS treatment (Fig. S3D). Together, these findings indicate that LPS treatment does not induce the death of PVN Magno OXT neurons, but rather reduces the OXT synthesis in these neurons.

PVN Magno and Parvo OXT neurons exhibit opposite activity changes in LPS challenge

Peripheral inflammation is known to alter the activity of OXT neurons and OXT levels [4, 25, 26]. Given that peripheral LPS challenge differentially affects PVN Magno and Parvo OXT neurons (Fig. 2E), we asked whether their electrophysiological properties are also differentially modulated. We used the AAV-OXT-Venus virus to label PVN OXT neurons [27], and then treated mice with either SAL or LPS (Fig. 3A). The OXT-Venus virus efficiently labeled PVN OXT neurons (Fig. 3A-B). Using whole-cell patch-clamp recording, we distinguished Magno OXT neurons from Parvo subtype based on the presence of a transient outward rectification (Fig. 3C) [23, 27, 28]. In PVN Magno OXT-Venus+ neurons, LPS treatment increased both spontaneous activity and neuronal excitability (Fig. 3D-F), which was consistent with the elevation of serum OXT levels at 12 h post-LPS treatment (Fig. S3D). This was accompanied by a depolarized resting membrane potential and elevated input resistance (Fig. 3E). In contrast, LPS challenge suppressed the spontaneous activity of Parvo OXT-Venus+ neurons and showed a tendency to reduce their excitability (Fig. 3G-I). While the resting membrane potential of Parvo neurons was unaltered, their input resistance was significantly decreased following LPS administration (Fig. 3H). Together, these results demonstrate that peripheral inflammatory challenge exerts opposite effects on the electrophysiological properties of PVN Magno and Parvo OXT neurons, and PVN Magno OXT neuronal hyperexcitability may enhance OXT transport and reduce OXT synthesis, leading to the loss of OXT+ neurons.

Fig. 3 — LPS induces opposite changes in the activity of PVN Magno and Parvo OXT neurons. (A) Top: Experimental procedure. Bottom: AAV-OXT-Venus viral labeling of PVN OXT neurons. n = 3 mice. (B) Example image showing colocalization of Venus+ neurons with OXT+ neurons. (C) Electrophysiological procedure and examples for distinguishing Magno and Parvo OXT neurons. (D) Example traces of spontaneous activity (Top) and spikes evoked by 90 pA current injection (Bottom) of PVN Magno OXT-Venus+ neurons in different groups. (E) Left: Summary of spontaneous firing frequency of Magno OXT-Venus+ neurons. Middle and Right: Summaries of resting membrane potential and input resistance of Magno OXT neurons. n = 28 and 28 neurons from 6 mice and 6 mice for SAL and LPS-treated groups. (F) Spike number evoked by current injection from 0 pA to 90 pA in Magno OXT-Venus+ neurons. n = 28 and 28 neurons for SAL and LPS-treated groups. (G)Same as (D), but for Parvo OXT neurons. (H) Same as (E), but for Parvo OXT-Venus+ neurons. n = 17 and 20 neurons from 6 mice and 6 mice for SAL and LPS-treated groups. (I) Same as (F), but for Parvo OXT-Venus+ neurons. * p < 0.05, ** p < 0.01, **** p < 0.0001, two-way ANOVA test with Sidak multiple comparisons test or unpaired t-test

Microglia are mainly activated in the anterior part of PVN to phagocyte Magno OXT neurons

Peripheral inflammation is known to activate brain-resident innate immune cells-microglia to subsequently influence neuronal function and even phagocytose neuronal structure [29]. Given that LPS challenge specifically reduced the number of OXT+ neurons in the PVN but not in the SON (Fig. 1C-E), we examined whether microglial activation differed between these regions. We found that LPS treatment significantly increased the density of Iba-1 + microglia in the PVN, with no significant change observed in the SON (Fig. 4A-B and E). Since LPS treatment specifically reduced Magno OXT+ neurons in the rostral PVN (Fig. 2B-E), we further analyzed microglial distribution along the rostral-caudal axis of the PVN. Consistent with the regional reduction of OXT+ neurons, microglia were more densely clustered in the rostral subregion of the PVN after LPS administration (Fig. 4B). In inflammatory contexts, microglia can transition from a surveillance state to a phagocytic state, potentially engulfing stressed but still viable neurons-such as those in a hyperexcitable state [30]. To evaluate whether microglia exhibited increased phagocytic activity toward OXT neurons, we performed three-dimensional reconstructions of microglia and OXT+ elements at the anterior and posterior parts of PVN. These analyses revealed that LPS treatment increased the volume of OXT neuronal material contained within microglia compared to controls at the anterior part of PVN (Fig. 4C), but not at the posterior part of PVN (Fig. S4A). Microglial phagocytic function in the PVN was confirmed by the elevation of CD68 expression (Fig. 4D). At the same time, we also reconstructed the AVP+ neuron elements and found that LPS treatment did not change the volume of AVP neurons within microglia (Fig. S4B). Additionally, we examined another glial cell-astrocyte, and found that the number of PVN astrocytes was not affected by LPS (Fig. 4F).

Fig. 4 — LPS treatment triggers microglia-mediated phagocytosis of rostral PVN OXT+ neurons. (A) Example images showing OXT+ neurons and Iba-1 + cell staining in the rostral (Left) and caudal (Right) PVN in different conditions. (B) Statistical distribution of Iba-1 + cell density along the rostral-caudal axis of the PVN. n = 4 and 5 mice for SAL and LPS-treated groups. (C) Left: Example 3D reconstruction of microglia and OXT+ neurons in the PVN rostral region in different conditions. Right: The volume of OXT neuronal material contained within microglia. n = 11 and 11 cells from 3 and 3 mice for SAL and LPS-treated groups. (D) Left: Representative images of CD68 and Iba-1 co-staining; Right: Quantitative analysis of the proportion of CD68-positive microglia, n = 5 and 6 mice for SAL and LPS-treated groups. (E) Left: Example images showing OXT+ neurons and Iba-1 + cell staining in the SON. Right: The density of Iba-1 + cells in the SON. n = 4 and 4 mice for SAL and LPS-treated groups. (F) Left: Example images showing OXT and GFAP+ cell staining in the PVN. Right: The density of GFAP+ cells in the PVN. n = 5 and 4 mice for SAL and LPS-treated groups. (G) Example images showing OXT+ neurons and Iba-1 + cell staining in the PVN treated with different conditions. (H) Statistical results of relative PVN OXT+ neuronal number in different conditions. n = 6, 6, 5, 6 mice for SAL & Con, SAL & PLX5622, LPS & Con and LPS & PLX5622 groups. * p < 0.05, ** p < 0.01, **** p < 0.0001, two-way ANOVA test with Tukey’s multiple comparisons test or unpaired t-test

To test whether microglia were responsible for PVN OXT+ neuronal reduction, we administered the CSF1R inhibitor PLX5622 for 7 days prior to LPS to deplete microglia [31]. PLX5622 effectively removed PVN microglia in both SAL and LPS groups (Fig. 4G and Fig. S5A) and prevented the LPS-induced reduction in PVN OXT+ neurons (Fig. 4G-H). In addition, we also recorded the spontaneous activity of PVN Magno OXT neurons, and observed that depleting microglia with PLX5622 also abolished the LPS-induced overexcitation of these neurons (Fig. S5B). Together, these results demonstrate that peripheral inflammation activates microglia in the rostral PVN, which in turn may hyperactivate Magno OXT neurons and phagocytose their structures.

Prostaglandin signals are elevated in the PVN under LPS challenge

Given that peripheral LPS challenge leads to microglia-mediated overexcitation and phagocytosis of PVN Magno OXT neurons, we sought to identify the mechanism underlying the differential changes of Magno and Parvo subtypes. By re-analyzing a published single-cell dataset of PVN OXT neurons [22], we compared the expression profiles of microglia- and inflammation-related genes between these two populations. This revealed significantly higher expression of Ptger4 (encoding prostaglandin E receptor 4, EP4 receptor) in the PVN Magno OXT neurons compared to Parvo neurons (Fig. 5A).

Fig. 5 — The prostaglandin signal pathway mediates the LPS treatment-induced effects on PVN OXT signals. (A) Differentially expressed genes (DEGs) between PVN Magno and Parvo OXT neurons. Blue and red dots with labeled gene names indicate the top 15 DEGs expressed in Parvo and Magno neurons. (B) Relative content of EP4 mRNA levels in the PVN in SAL and LPS-treated groups. n = 8 and 9 mice for SAL and LPS-treated groups. Solid circles represent data from male mice, and open circles represent data from female mice. (C) Same as (B), but for COX2 mRNA. (D) Same as (B), but for Ptges mRNA. (E) Relative content of PGE2 protein concentrations in the PVN. (F) Correlation analysis between COX2 and OXT mRNA levels in the PVN region. n = 17 mice, R indicates Pearson correlation coefficient and * indicates p < 0.05, two-tailed t-test. (G) Same as (F), but between EP4 and OXT mRNA. (H) Left: Heatmap comparison of prostaglandin metabolic pathway gene changes induced by LPS in the PVN and SON regions. Right: Comparison of gene expression changes in the PGE2 metabolic pathway. Each dot corresponds to one gene. (I) Left: Same as (H), but for phagocytosis. Right: Comparison of gene expression changes in phagocytosis, genes from the heatmap on the left. * p < 0.05, ** p < 0.01, *** p < 0.001, unpaired t-test

The cellular sources of prostaglandin E2 (PGE2) in the brain include epithelial cells and glial cells [32]. Reactive microglia synthesize and release PGE2, which can modulate neuronal function through the EP4 receptor [32, 33]. The biosynthesis of PGE2 under pathological conditions is primarily mediated by the enzymes COX2 and Ptges. Since intracerebral administration of PGE2 has been reported to selectively increase plasma OXT but not the AVP levels [34], we assessed the impact of LPS on the prostaglandin signals in the PVN. LPS treatment significantly elevated mRNA levels of the EP4 receptor and COX2 (Fig. 5B-C), with a non-significant increasing trend observed for Ptges mRNA (Fig. 5D). However, these transcriptional changes did not lead to a significant increase in PVN PGE2 concentration (Fig. 5E), which may be due to its rapid in vivo metabolism [35]. Given that LPS reduced OXT mRNA in the PVN (Fig. 1D), we further examined the relationship between OXT and prostaglandin signaling. Although EP4 mRNA levels did not correlate with OXT mRNA, a significant negative correlation was found between COX2 and OXT mRNA (Fig. 5F-G).

Given that LPS treatment selectively reduced the number of OXT+ neurons in the PVN but not in the SON (Fig. 1C-E), we performed bulk RNA-sequencing to compare the influence of LPS on prostaglandin and phagocytosis-related signal pathways between these two regions. Obvious inter-group differences were observed in both PVN and SON between the control and LPS groups (Fig. S6A-B). Differential expression analysis identified 804 upregulated and 45 downregulated genes in the PVN, and 867 upregulated and 59 downregulated genes in the SON following LPS challenge (Fig. S6C). GO enrichment analysis indicated that while both regions were significantly enriched in immune and neuroinflammatory pathways, only the PVN showed specific enrichment for the prostaglandin biosynthetic process (Fig. S6D-E). To further explore the regional differences, we directly compared the expression of genes involved in prostaglandin metabolism and phagocytosis. The mRNA of key enzymes involved in PGE2 synthesis, such as Ptgs2 (COX2), Ptges, Ptgs1, and Ptger4, were strongly upregulated in the PVN (Fig. 5H). Similarly, phagocytosis-related genes were induced to a greater extent in the PVN following LPS treatment (Fig. 5I). Collectively, these results demonstrate that the prostaglandin signaling pathway may mediate the specific reduction of PVN Magno OXT+ neurons following peripheral LPS challenge.

Inhibition of prostaglandin signals alleviates LPS-induced reduction of PVN OXT-immunopositive neurons

To directly test whether microglial phagocytosis of PVN Magno OXT neurons depends on PGE2 signaling after LPS challenge, we pharmacologically inhibited two key nodes of the pathway—COX2 and EP4 receptors, which were also significantly upregulated by the LPS (Fig. 5B-C). Celecoxib (CELE), a selective COX2 inhibitor, was administered intraperitoneally 1 h before LPS treatment. While LPS challenge markedly increased the density of PVN Iba-1 + cells, this effect was significantly suppressed by 40 mg/kg CELE pretreatment [36, 37] (Fig. 6A, C, and D). Furthermore, the LPS-induced increase in the proportion of COX2-expressing microglia was also reversed by CELE (Fig. 6C, E). Most importantly, CELE treatment substantially attenuated both the reduction of PVN OXT+ neurons and microglia-mediated phagocytosis of OXT neurons induced by LPS (Fig. 6F and Fig. S7A-B).

Fig. 6 — Inhibiting prostaglandin signaling alleviates LPS-induced reduction of PVN OXT+ neurons. (A) and (B) Schematic diagram of CELE and ONOAE administration protocols, respectively. (C) Example images showing the colocalization of Iba-1 + cells and COX2 signal in the PVN region of mice in different groups. (D) Quantitative analysis of PVN microglial density in different groups. n = 5, 5, 7, 6 mice for SAL & Con, SAL & CELE, LPS & Con, and LPS & CELE groups. (E) Proportion of PVN microglia expressing COX2 in different groups. n = 3, 3, 4, 4 mice for SAL & Con, SAL & CELE, LPS & Con, and LPS & CELE groups. (F) Left: Example images of OXT+ neurons and Iba-1 + cells in the PVN region in different groups. Right: Relative numbers of OXT+ neurons in the PVN region. n = 5, 5, 7, 6 mice for SAL & Con, SAL & CELE, LPS & Con, and LPS & CELE groups. (G) Same as (F), but for the EP4 antagonist ONOAE. n = 9, 9, 9, 8 mice for SAL & Con, SAL & ONOAE, LPS & Con, and LPS & ONOAE groups. * p < 0.05, ** p < 0.01, **** p < 0.0001, two-way ANOVA test with Tukey’s multiple comparisons test

To specifically block PGE2 signaling in the brain, we intracerebroventricularly administered ONO-AE3-208 (ONOAE), a selective EP4 receptor antagonist that does not cross the blood-brain barrier [38], 1 h prior to LPS challenge (Fig. 6B). Consistent with the effects of COX2 inhibition, central EP4 antagonism by ONOAE significantly attenuated the LPS-induced increase in the density of PVN Iba-1 + cells (Fig. S7C) and reduction of PVN OXT+ neurons (Fig. 6G). Furthermore, ONOAE treatment suppressed the phagocytosis of OXT neurons by reactive microglia (Fig. S7D-E). We also directly recorded the activity of PVN Magno OXT neurons, and consistent with PGE2-induced elevation of plasma OXT levels [34], PGE2 application significantly increased the spontaneous activity of PVN Magno OXT neurons (Fig. S8A-B), and did not change neuronal resting potential and input resistance (Fig. S8C). However, pre-treating with the ONOAE blocked the effect of PGE2 on PVN Magno OXT neurons (Fig. S8D-F). Together, these results confirm that prostaglandin signaling is involved in the changes of PVN Magno OXT neurons following peripheral LPS challenge.

Knockdown of EP4 receptors in PVN OXT neurons alleviates the LPS-induced reduction of OXT-immunopositive neurons

In addition to PVN OXT neurons, microglia also express EP4 receptors, which have been shown to modulate their LPS reactivity and phagocytic capacity [39–41]. To investigate the changes in EP4 expression levels in OXT neurons under peripheral inflammation, we compared the proportion of OXT+ neurons that express EP4 between the SAL and LPS groups, and observed that the percentage of EP4-positive OXT neurons was significantly increased in the LPS-treated group relative to the SAL group (Fig. S9). We investigated the specific role of EP4 expressed in PVN OXT neurons by knocking down EP4 receptors in these neurons using OXT-Cre mice together with a Cre-dependent EP4 RNA interference virus (Fig. 7A). Compared with the LPS & NC group, the LPS & EP4i group exhibited a significant reduction in the proportion of EP4-positive OXT neurons (Fig. 7B-C), validating the knockdown efficiency. Compared with the SAL & NC group, the LPS & NC group showed a marked reduction in OXT+ neurons, which was restored in the LPS & EP4i group (Fig. 7D-E). Although LPS did not significantly reduce serum OXT levels, OXT levels in the LPS & EP4i group were higher than those in both the SAL & NC and the LPS & NC groups (Fig. 7F). To investigate the biological significance of rescuing OXT+ neurons, we evaluated mouse sickness behaviors, including core body temperature and locomotion [42]. Compared with the SAL & NC group, the LPS & NC group exhibited decreased body temperature and significantly reduced locomotor activity. Notably, after EP4 knockdown in PVN OXT neurons, LPS no longer reduced body temperature or locomotor activity (Fig. 7G-H). Furthermore, we examined the effect of specific EP4 knockdown on microglial phagocytic capacity. Compared with the LPS & NC group, the LPS & EP4i group exhibited a significant reduction in microglia-specific phagocytosis (Fig. 7I-J). Together, these results indicate that EP4 receptors expressed in PVN OXT neurons are necessary for the LPS-induced reduction of PVN OXT+ neurons and microglial phagocytosis.

Fig. 7 — Specific knockdown of EP4 receptors in PVN OXT neurons alleviates the LPS-induced reduction of OXT+ neurons and mouse sickness behaviors. (A) Schematic diagram of the experimental procedure. (B) Representative immunofluorescence images showing co-staining of EP4 and OXT. (C) Quantitative analysis of the proportion of EP4-positive OXT+ neurons. n = 3 and 3 mice for each group. (D) Representative images showing EGFP expression and OXT immunostaining in the PVN. (E) Relative number of OXT+ neurons in the PVN among groups. n = 3, 3, 4 mice for SAL & NC, LPS & NC, LPS & EP4i groups. (F)-(H) Same as (E), but for serum OXT levels (F), rectal temperature (G), and locomotion (H). (I) Example images showing 3D reconstruction of microglia and OXT neurons in the PVN. (J) The volume of OXT neuronal material contained within microglia. n = 10 and 10 cells from 3 and 3 mice for SAL and LPS-treated groups. * p < 0.05, ** p < 0.01, *** p < 0.001, unpaired t-test or one-way ANOVA with Tukey’s multiple comparisons test

Discussion

OXT serves as a critical neuromodulator in a wide range of physiological functions [1, 2]. Although decreased OXT levels and a reduction in OXT+ neurons have been frequently documented in various disease states [3, 4, 12–14], the mechanisms responsible for the decline in hypothalamic OXT signaling remain poorly understood. Our study reveals a selective reversible reduction of PVN Magno OXT-immunopositive neurons following LPS-induced peripheral inflammatory challenge. Furthermore, we demonstrate that the PGE2–EP4 signaling pathway mediates the reduction of OXT+ neurons by modulating neuronal activity and microglial phagocytosis. Based on our findings, we propose a model in which peripheral inflammation increases PGE2 synthesis in the PVN, leading to hyperexcitation of Magno OXT neurons via EP4 receptor activation. Concurrently, LPS-induced microglial activation disrupts the structural integrity of these hyperexcitable neurons through phagocytosis, thereby impairing neuronal function and ultimately reducing OXT synthesis capacity (Fig. 8).

Fig. 8 — Schematic diagram illustrating the mechanism by which LPS treatment reduces PVN Magno OXT-immunopositive neurons. The PGE2–EP4 signaling pathway mediates the LPS-induced reduction of OXT+ neurons by modulating neuronal activity and microglial phagocytosis

Previous studies have shown that both brain and plasma OXT levels are reduced in sepsis [4, 43]. To align with these findings, we used relatively severe models of systemic inflammatory stress (four injections of 1 mg/kg LPS and a single high dose of 15 mg/kg LPS) [18, 44]. We further observed that a single low-dose LPS had no significant effect on PVN OXT+ neurons (Fig. S2G-H). These findings indicate that the reduction of PVN Magno OXT+ neurons occurs only when inflammation reaches a certain threshold. It has been reported that seven consecutive days of 0.5 mg/kg LPS injection (a regimen of chronic low-grade inflammation) reduced OXT levels in the anterior cingulate cortex [45]. The possible influence of chronic low-grade inflammation on hypothalamic OXT signaling remains to be further investigated.

In our study, although LPS reduced the number of PVN OXT+ neurons at 12 h post-injection, serum OXT levels were elevated at this time point. Our electrophysiological recordings showed that LPS significantly excited PVN Magno OXT neurons (Fig. 3D-F), which primarily release OXT into the blood circulation; accordingly, serum OXT levels were increased in our study (Fig. S3D). However, at 10 days post-LPS injection, serum OXT levels were lower, even though the number of OXT+ neurons had recovered (Fig. 2G-H; Fig. S3D). In the later phase of LPS treatment, OXT release from PVN Magno neurons may be reduced, consequently leading to decreased serum OXT levels (Fig. S3D). Consistent with our observation, in septic patients, plasma OXT levels were increased at 6 h but decreased to 50% at 48 h after onset [43]. Knocking down EP4 in PVN OXT neurons protected against the reduction of Magno OXT+ neurons and also alleviated sickness behaviors (Fig. 7). Moreover, peripheral OXT levels are closely associated with many functions, such as feeding behavior, reproductive system function, and cardiovascular function [46, 47]. Whether decreased serum OXT levels at the late stage are associated with long-term physiological functions and behavioral changes remains to be experimentally investigated.

Our study reveals the selective reduction of PVN Magno OXT+ neurons in LPS-induced peripheral inflammatory challenge (Fig. 2D-E), contrasting with the predominant loss of Parvo OXT+ neurons reported in social dysfunction models [15]. Magno OXT neurons constitute the primary source of circulating OXT, and consistent with the specific change of PVN Magno OXT+ neurons observed in our experiments, the dysfunction of plasma OXT signals has been observed in both humans and animal models under peripheral inflammatory conditions [12, 14, 26]. In comparison, Parvo OXT neurons-enriched in autism spectrum disorder (ASD)-related risk genes—exhibit reduced activity upon exposure to developmental ASD risk factors [15, 22]. Through analysis of a published single-cell dataset [22] and our bulk sequencing data, we identified involvement of the prostaglandin signaling pathway in the selective reduction of PVN Magno OXT+ neurons following LPS challenge (Fig. 5). This conclusion is further supported by pharmacological experiments inhibiting PGE2 synthesis, blocking the EP4 receptors, and knocking down of EP4 receptors in PVN OXT neurons (Fig. 6 and 7). In line with recent studies highlighting the PGE2–EP4 axis in microglia–neuron interactions [32, 33], we propose that targeting this pathway could represent a promising strategy for preserving OXT signaling in peripheral inflammation–related diseases.

In addition to neuron-type selectivity, our study reveals region-specific reduction of Magno OXT+ neurons, with the PVN showing greater reduction than the SON under inflammatory challenge (Fig. 1 and Fig. S2). This regional difference may be attributable to microglia–neuron interactions, as LPS induced microglial activation in the PVN but not in the SON (Fig. 4), and triggered a more pronounced upregulation of the PGE2 synthesis pathway in the PVN (Fig. 5). The heightened inflammatory sensitivity of the PVN may be facilitated by its anatomical proximity to the third ventricle—a region with a relatively permeable blood–brain barrier that may allow peripheral inflammatory signals to more readily access local microglia [48].

Hypothalamic OXT release into the bloodstream is known to regulate immune system development, surveillance, and homeostasis [49]. Our findings demonstrate that immune dysfunction can in turn impair hypothalamic OXT neurons-specifically, that peripheral inflammation activates the brain’s resident immune cells, microglia, which mediate this process. Microglia participate in physiological processes — including synaptic pruning, axonal remodeling, and the clearance of apoptotic cells through direct contact with neurons, playing an essential role in maintaining the fine structure of neural circuits [50]. However, under neuroinflammatory conditions, microglial phagocytic activity can become dysregulated, influencing neuronal function and survival. For instance, in the early stages of Alzheimer’s disease, microglia mediate synapse loss via the complement cascade [51]. Following cerebral ischemia, microglia in mice exhibit prolonged contact with synapses and participate in the remodeling of damaged neural circuits [52]. In this study, we found that the phagocytic activity of microglia was enhanced and specifically impaired the structure of Magno OXT neurons (Fig. 4C-D). Depleting microglia with the CSF1R inhibitor PLX5622 not only prevented the reduction of OXT+ neurons in the PVN but also normalized their hyperexcited activity (Fig. 4G-H and Fig. S5B). Notably, although AVP neurons in the PVN share certain functional overlaps with OXT neurons, we observed no reduction in AVP+ neuronal numbers or microglial phagocytosis toward them (Fig. 1F and Fig. S4B). This selectivity is consistent with reports showing differential changes in plasma OXT and AVP in experimental sepsis, and that gestational stress selectively diminishes PVN OXT-but not AVP-neurons via microglia-dependent mechanisms during development [53]. However, in demonstrating the role of microglia in the effects on OXT neurons, the use of PLX5622 resulted in systemic depletion of microglia and potentially induced alterations in certain monocyte populations [54]. To definitively distinguish between central and peripheral effects, we propose that future studies using microglia‑specific genetic approaches (e.g., Tmem119‑CreERT2 mice) will be required [55].

We propose a positive feedback mechanism in which PGE2 signaling via EP4 receptors on Magno OXT neurons constantly enhances their excitability. As a Gs-coupled receptor, EP4 activation initiates the cAMP/PKA signaling cascade—a pathway known to support sustained neuronal firing [56]. Using electrophysiological approaches, we confirmed that PGE2 increases the excitability of Magno OXT neurons through EP4 (Fig. S8). This elevated neuronal activity recruits nearby microglia and results in increased PGE2, which then reactivates neuronal EP4 receptors, further amplifying excitability. We propose that PGE2 leads to persistent overexcitation of OXT neurons, which in turn promotes microglial phagocytosis of OXT neurons and causes depletion of OXT synthesis. Under peripheral inflammation, knockdown of EP4 receptors in OXT neurons significantly reduces microglial phagocytosis of OXT neurons (Fig. 7G) and increases serum OXT levels (Fig. 7F). Intraperitoneal administration of celecoxib to inhibit PGE2 synthesis reduced microglial activation and its COX2 expression following LPS treatment, and restored the decreased number of OXT+ neurons (Fig. 6A-F). However, it should be noted that in addition to activated microglia, epithelial cells and glial cells can also produce PGE2 [32]. Therefore, the specific effect of microglia-derived PGE2 in the PVN requires further investigation. Furthermore, non-neuronal cells, particularly microglia, may contribute to the overall increase in EP4 expression [39–41]. Thus, cell‑type‑specific approaches are warranted to dissect their respective roles in this process.

Given the short half-life, rapid degradation, and limited blood-brain barrier penetration of exogenous oxytocin, preserving endogenous OXT signaling represents a promising alternative for mitigating inflammation [1]. Our study identifies PGE2 signaling as a key mechanism underlying the reduction of Magno OXT+ neurons during peripheral inflammation. In addition, we further complemented the impact of OXT neurons on sickness behavior under peripheral inflammation. Previous studies have reported that LPS induces sickness behaviors in mice, including reduced locomotor activity and decreased body temperature [42]. Although during peripheral inflammation, we observed a reduction of only a subset of OXT+ neurons, this reduction showed a functional correlation with sickness behavior (Fig. 7G-H). This finding establishes a rationale for developing targeted interventions aimed at protecting endogenous OXT neurons as a therapeutic strategy.

Materials and methods

Mouse strains

All animal procedures in this study were approved by the Animal Care and Use Committee of Fudan University. Male mice aged 8–12 weeks were used for all the experiments unless otherwise indicated in the figure legends. Animals were kept in grouped housing within a controlled barrier environment, with a 12 h/12 h light–dark schedule (lights on from 08:00 to 20:00). Temperature was maintained at 24 ± 2 °C and relative humidity at 50 ± 5%. Mouse lines used in this study include: C57BL/6 mice were obtained from Jiangsu Huachuang Sino Pharmaceutical Technology Co., Ltd; B6.129 S-OXT^{1.1(cre)Dolsn}/J (OXT-Cre, #024234, Jackson Laboratory) mice were used to genetically target oxytocin expressing neurons; B6.Cg-Gt(ROSA)26Sor^{tm3(CAG−EYFP)Hze}/J (Ai3, #007903) reporter strain was crossed with OXT-Cre mice, enabling specific labeling of OXT+ cells. Mouse genotyping was conducted following standard procedures on the Jackson Laboratory website.

LPS-induced peripheral inflammation models

A low dose of LPS (1 mg/kg, L4130, Sigma-Aldrich) was used to establish a milder peripheral inflammation model [57]. Mice were randomly assigned to the saline (SAL) group and the LPS group. The LPS group received intraperitoneal injections of 1 mg/kg LPS daily for four consecutive days, while the SAL group received injections of an equal volume of saline. Animals’ body weights were recorded daily. 12 h or 10 days after the final injection, mice were euthanized, and brain tissue was harvested for subsequent experiments. A high dose of LPS (15 mg/kg) was used to establish a sepsis model [58]. Mice were randomly intraperitoneally injected with SAL and LPS. After 24 h, mice were euthanized, and brain tissue was harvested.

Tail vein injection

For selective labeling of PVN Magno OXT neurons, 40 µL of 1% FluoroGold (FG; Fluorochrome, cat#FC10000) was injected intravenously into the tail vein using a 1-mL syringe and 25-G needle [22]. Peripheral inflammation was induced five days post-injection to allow for retrograde tracer transport.

In vivo stereotaxic intracranial injection

Mice (5–6 weeks old) were anesthetized with 1–2% isoflurane and placed in a stereotaxic apparatus (E07370-005, RWD). After drilling a cranial window, 200 nL of virus was bilaterally injected into the PVN (AP: -0.01 mm, ML: ±0.03 mm, DV: -0.46 mm & -0.49 mm) with a flow rate of 2–3 nL/s using a microsyringe pump controller (NanoJect III, Drummond Scientific Company). The needle remained in place for 10 min before removal. Mice were used for the following experiments at least 3 weeks post-injection.

The following AAV viruses were used in this study: AAV2/9-OXT-Venus (3.00 × 10¹² vector genomes/mL, a gift from Dr. Valery Grinevich); AAV2/9-hEF1a-DIO-MasterRNAi155(EP4)-eGFP-WPRE-pA (1.05 × 10¹² vector genomes/mL); AAV2/9-hEF1a-DIO-MasterRNAi155(NC)-EGFP-WPRE-pA (1.05 × 10¹² vector genomes/mL).

Cannula implantation

For intraventricular (ICV) drug administration, a guide cannula was stereotactically implanted bilaterally above the lateral ventricle (AP: +0.04 mm; ML: ±0.10 mm; DV: −0.25 mm) using the same surgical procedure as described for viral injections. Stainless dummies were inserted into the guide cannula to prevent clogging. At least 1 week after surgery, mice received ICV infusions of either vehicle (5% DMSO in saline) or 4 mM ONO-AE3-208 (ONOAE; MCE, HY-50901) [38]. The injection cannula, which extended 0.25 mm beyond the guide cannula tip, was used to deliver 1 µL of solution per side over 3 min. The cannula was left in place for an additional 5 min after infusion to allow for drug diffusion. LPS was administered systemically 1 h after the ICV microinjection.

Sickness behavior

Body temperature change and fatigue are often used to evaluate sickness behaviors [32, 42]. Fatigue was measured by the open field test 10 h after the final LPS injection. The mice were placed in the center of the plexiglass box for free exploration and were recorded using an infrared monitoring camera for 15 min. Total distance was calculated to assess locomotion. After 2 h, body temperature was measured using an anal temperature detector (KW Company, FT3400). Following body temperature measurement, the mice were euthanized.

Peripheral blood serum extraction and ELISA

Mice were anesthetized with isoflurane, and the orbital artery was ruptured to collect orbital blood into an EP tube. The blood sample was allowed to stand for 15 min, followed by freezing on ice for 30 min. It was then centrifuged at 3,000 rpm for 15 min at 4 °C, and the supernatant was collected as serum. Serum oxytocin or TNF-α concentrations were determined using a commercial ELISA kit (JM-02559M1, JINGMEI, China; JLW10484, JIANGLAI, China), with all procedures adhering strictly to the manufacturer’s instructions.

Immunohistochemistry, imaging and analyses

Mice were deeply anesthetized with isoflurane and transcardially perfused with 4% paraformaldehyde (PFA) dissolved in 0.1 M phosphate-buffered saline (PBS). The brains were extracted and post-fixed in the 4% PFA solution at 4 °C for 24 h. For sectioning, 40 μm coronal sections were obtained using one of two methods: (1) directly using a vibratome (Leica VT1200), or (2) after cryoprotection in 20% and 30% sucrose solutions (24 h each) and subsequent OCT embedding, using a cryostat (Leica CM1950). Sections containing the PVN were systematically divided into three equal sets to minimize batch effects. One randomly selected set was processed for immunofluorescence staining. Briefly, sections were permeabilized with 0.2% Triton X-100 in PBS for 1 h at room temperature (RT) and then blocked with immunostaining blocking solution (Azure Sky, #P0260) for 30 min at RT. Subsequently, sections were incubated with primary antibodies diluted in a dedicated antibody dilution buffer (Azure Sky, #P0103) for 24 h at 4 °C. The following primary antibodies were used: rabbit anti-OXT (1:4000, Abcam ab212193), rabbit anti-AVP (1:4000, ImmunoStar 20069), guinea pig anti-Iba1 (1:1000, Oasis Biofarm OB-PG049-02), rabbit anti-COX2 (1:400, Proteintech 27308-1-AP), chicken anti-GFAP (1:5000, Abcam AB4674), mouse anti-EP4 (1:50, Santa Cruz sc-55596), and mouse anti-CD68 (1:1000, Abcam ab201340). After incubation, sections were washed three times for 10 min each in PBS and then incubated with appropriate secondary antibodies (Goat anti-guinea pig Alexa Fluor 488, Goat anti-rabbit Alexa Fluor 594/647, Goat anti-chicken Alexa Fluor 488; all from Life Technologies, used at 1:1000 in PBS) for 2 h at RT in the dark. Following another three 10-min washes in PBS, the sections were mounted on glass slides, air-dried, and coverslipped with an anti-fading mounting medium composed of glycerol and TBS (9:1) containing Hoechst 33,342 (1:1000, ThermoFisher Scientific, #H3570).

Tissue sections were initially scanned using an Olympus VS120 slide scanning microscope for overview imaging. High-resolution confocal images were then acquired with a Nikon A1 confocal laser scanning microscope equipped with a 20× objective. Z-stack images were collected at 2-µm intervals for cell counting.

For cell counting, six consecutive sections encompassing the PVN or SON, as defined by a standard brain atlas, were selected from each animal. Cell counts and colocalization analyses were performed using ImageJ software (Fiji, NIH). To minimize batch-to-batch variation in OXT and AVP neuron quantification across different experimental cohorts, the raw counts from each batch were normalized.

Three-dimensional (3D) image reconstruction

Images for cellular reconstruction were acquired at 0.5-µm z-intervals using a Nikon A1 confocal laser scanning microscope with a 40× objective. Microglial engulfment of OXT neurons or AVP neurons was analyzed in IMARIS 10.0.0 by generating 3D surface renderings of microglia. A consistent threshold was applied across samples to ensure accurate reconstruction of fine microglial processes. Following microglial reconstruction, the “Mask” function in IMARIS was used to isolate and reconstruct OXT or AVP neuronal volumes contained within each microglial cell. The “Detail” function was then employed to quantify the volumes of both the microglia and the engulfed OXT or AVP fragments. To account for potential confounding effects of microglial proliferation on phagocytic metrics, the relative volume of OXT or AVP neurons inside microglia (i.e., the ratio of engulfed OXT volume to total microglial volume) was used as a normalized measure of phagocytic activity. For each mouse, three randomly selected images, each containing at least six microglial cells, were reconstructed. Three mice were analyzed per experimental group.

In vitro electrophysiological recording

Brain slices were prepared according to established methods [23, 28, 59]. Briefly, mice were deeply anesthetized with isoflurane and transcardially perfused with ice-cold, oxygenated (95% O₂/5% CO₂) artificial cerebrospinal fluid (ACSF, in mM): 127 NaCl, 2.5 KCl, 25 NaHCO₃, 1.25 NaH₂PO₄, 2 CaCl₂, 1 MgCl₂, and 25 glucose (osmolarity ~ 310 mOsm/L). The brain was quickly removed and placed in a sectioning chamber containing ice-cold ACSF (perfused with 95% O₂/5% CO₂). Coronal slices (~ 250 μm thick) were prepared using a Vibratome 1000PLUS microtome. Slices containing the PVN region were collected and subsequently incubated in oxygenated ACSF at 34 °C for ~ 25 min, followed by equilibration at room temperature for ~ 1 h before recording.

For electrophysiological recordings, brain slices were transferred to a recording chamber and continuously perfused with oxygenated ACSF at 1.5–2 mL/min. Chamber temperature was maintained at ~ 30 °C using a feedback in-line heater (TC-324 C; Warner Instruments). Fluorescently labeled PVN OXT neurons were visualized using infrared differential interference contrast (IR-DIC) microscopy and identified based on the fluorescent signal.

Current clamp recordings were established using glass microelectrodes (3–5 MΩ) filled with an intracellular solution containing (in mM): 135 K-gluconate, 4 KCl, 10 HEPES, 10 mM Na-phosphocreatine, 4 mM MgATP, 0.4 mM Na₂GTP, and 1 mM EGTA, with pH maintained at 7.2–7.3 (osmolarity approximately 295 mOsm/L). After membrane rupture, each neuron was allowed to stabilize for approximately 2 min before data acquisition to avoid puncture-induced artifacts. Signals were amplified using a Multiclamp 700B amplifier, digitized at 20 kHz, and low-pass filtered at 4 kHz via pCLAMP software (Molecular Devices). Three sequential protocols were applied to characterize intrinsic properties: spontaneous activity (protocol 1), current-step response (protocol 2), and spike latency test (protocol 3) [23]. For protocol 1, action potentials were recorded for 2 min under zero-current conditions to determine the basal firing rate. For protocol 2, current was injected in 10 pA increments from − 100 pA to + 100 pA (250 ms duration per step). The resting membrane potential (RMP) was measured during the 0 pA step. The input resistance (Rin) was calculated from the steady-state voltage response to − 100 pA current. For protocol 3, a two-step protocol was used to distinguish Magno from Parvo subtypes. Following a − 100 pA hyperpolarizing step (300 ms), depolarizing currents from − 35 pA upward (5 pA increments, 300 ms) were applied. Spike latency was defined as the interval between hyperpolarization offset and the first action potential across sweeps. As previously established [23], Magno neurons exhibit longer spike latencies than Parvo neurons.

During pharmacological experiments, the following compounds were bath-applied at specified concentrations: prostaglandin E2 (PGE2, 2 µM; MCE) and the EP4 receptor antagonist ONO-AE3-208 (5 µM; MCE).

RNA extraction and reverse transcription-quantitative PCR (RT-qPCR)

Total RNA was extracted from homogenized tissue samples using ice-cold TRIzol reagent (DP424, Tiangen) following the manufacturer’s protocol. After purification, the RNA was stored at -80 °C until use. cDNA was synthesized from 1000 ng of total RNA per sample using the FastKing-RT SuperMix kit (KR118, Tiangen).

Quantitative real-time PCR (qPCR) was performed using the SuperReal PreMix Plus (SYBR Green, Tiangen) on a QuantStudio 3 thermocycler (Thermo Fisher Scientific). Reaction specificity was verified by melt curve analysis. Relative gene expression was normalized to β-actin as an internal control and analyzed using the QuantStudio Design & Analysis Software.

The gene-specific primers used were:

5′-CTGTGCTGGACCTGGATATGCG-3′ (forward)
5′-AGCTCGTCCGCGCAGCAGATG-3′ (reverse) for OXT;
5′-TACGCTCTCTGCTTGCTTCC-3′ (forward)
5′-ACTGTCTCAGCTCCATGTCG-3′ (reverse) for AVP;
5′-CAGGATGCAGAAGGAGATTAC-3′ (forward)
5′- AACGCAGCTCAGTAACAGTC-3′ (reverse) for Actin;
5′-CAACCAACAAGTGATATTCTCCATG-3′ (forward)
5′-GATCCACACTCTCCAGCTGCA-3′(reverse) for IL-1β;
5′-CATCTTCTCAAAATTCGAGTGACAA-3′ (forward)
5′-TGGGAGTAGACAAGGTACAACCC-3′ (reverse) for TNF-α;
5′-GGGCCATGGAGTGGACTTAAA-3′ (forward)
5′-ACTCTGTTGTGCTCCCGAAG-3′ (reverse) for COX2;
5′-TCTCACTCTCAGTCCCGGTG-3′ (forward)
5′-GGGGTTGGCAAAAGCCTTC-3′ (reverse) for Ptges;
5′-TAGCCTCTCTGGCTTTCCAA-3′ (forward)
5′-TACCTCCAACCTCAGCCATC-3′ (reverse) for Ptger4;

RNA sequencing and analyses

Tissue samples from the PVN and SON were microdissected under a stereomicroscope and rapidly placed in liquid nitrogen for freezing. Total RNA was extracted from the sample using the RNeasy mini kit (Cat. No. 74804, Qiagen). Library preparation and sequencing were conducted by Shanghai Biotechnology Corporation. Briefly, poly(A) + RNA was purified using poly(dT) magnetic beads and used for sequencing library construction. The resulting libraries were quantified using a Qubit^® 2.0 Fluorometer (Life Technologies), and insert size distribution was assessed with an Agilent 2100 Bioanalyzer (Agilent Technologies) to determine molar concentration. Cluster generation was performed on the cBot system at 10 pM, followed by paired-end sequencing on an Illumina NovaSeq 6000 platform.

Raw sequencing reads were preprocessed to remove ribosomal RNA reads, adapter sequences, short inserts, and other low-quality reads. Cleaned reads were aligned to the mouse reference genome (GRCm39) using Hisat2 (v2.0.4). Transcript assembly and quantification were performed with StringTie (v1.3.4) using reference annotations, and FPKM values were calculated for known genes. Differential expression analysis was carried out using edgeR, with significance thresholds set at |log₂FC| ≥ 1 and FDR ≤ 0.05.

Data analysis

Imaging data were processed using ImageJ (FIJI distribution, NIH). All imaging data for cell reconstruction were completed using IMARIS software. Electrophysiological data were initially processed with pClamp (version 10.7, Molecular Devices), prior to statistical analysis and visualization using MATLAB (R2023a, MathWorks) and GraphPad Prism (v.9.5.1, GraphPad Software). All data are presented as means ± SEM. Normality was assessed using the D’Agostino and Pearson omnibus normality test. For comparisons between two groups, normally distributed data were analyzed using a two-tailed unpaired or paired Student’s t-test, while non-normally distributed data were evaluated with the Mann–Whitney or Wilcoxon signed-rank test, as appropriate. For comparisons across four groups involving two independent variables, two-way ANOVA was applied to normally distributed data, followed by an appropriate multiple comparisons test; non-parametric alternatives were used when normality assumptions were not met. A p-value of less than 0.05 was considered statistically significant.

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary Material 1^{(6.2MB, docx)}

Acknowledgements

This work was supported by grants from the Noncommunicable Chronic Diseases-National Science and Technology Major Project (2023ZD0506800), National Natural Science Foundation of China (32371058, 81970727, 31900738), Open Research Fund of Basic Medicine College (JCKFKT-MS-010), Shanghai Municipal Science and Technology Major Project (2018SHZDZX01), ZJ Lab, and Shanghai Center for Brain Science and Brain-Inspired Technology.

Author Contributions

L.X., L.C., and Y.L. designed the experiments. Y.L. carried out all the molecular experiments and analysis with the help from H.X. and M.Z., Y.L. performed electrophysiology and data analyses with the help from H.X, Y.L. and Z.G. conducted transcriptome analysis. L.X. and Y.L. wrote the paper, with contributions from all authors. Correspondence should be addressed to L.X. (leixiao@fudan.edu.cn).

Data availability

The authors confirm that the data supporting the findings of this study are available within the article and its Supplementary material.

Declarations

Ethics approval

All mice experiments complied with the requirements of the Laboratory Animal Management Committee of Fudan University regarding the breeding and management of laboratory animals as well as their usage. All efforts were made to minimize animal suffering and the number of animals used.

Competing interests

The authors have no relevant financial or nonfinancial interests to disclose.

Footnotes

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Contributor Information

Li Cao, Email: caoli@smmu.edu.cn.

Lei Xiao, Email: leixiao@fudan.edu.cn.

References

1.Cui X, Xiao L (2025) Complexity of the hypothalamic oxytocin system and its involvement in brain functions and diseases. Neurosci Bull 41(7):1267–1288. 10.1007/s12264-025-01424-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
2.Jiang J et al (2023) Intertwined associations between oxytocin, immune system and major depressive disorder. Biomed Pharmacother 163:114852. 10.1016/j.biopha.2023.114852 [DOI] [PMC free article] [PubMed] [Google Scholar]
3.Yang M et al (2023) Oxytocin improves intracerebral hemorrhage outcomes by suppressing neuronal pyroptosis and mitochondrial fission. Stroke 54(7):1888–1900. 10.1161/strokeaha.123.043391 [DOI] [PubMed] [Google Scholar]
4.Jiang J et al (2023) Oxytocin alleviates cognitive and memory impairments by decreasing hippocampal microglial activation and synaptic defects via oxtr/erk/stat3 pathway in a mouse model of sepsis-associated encephalopathy. Brain Behav Immun 114:195–213. 10.1016/j.bbi.2023.08.023 [DOI] [PubMed] [Google Scholar]
5.Clodi M et al (2008) Oxytocin alleviates the neuroendocrine and cytokine response to bacterial endotoxin in healthy men. Am J Physiol Endocrinol Metab 295(3):E686–E691. 10.1152/ajpendo.90263.2008 [DOI] [PubMed] [Google Scholar]
6.Yu Y, Li J, Liu C (2022) Oxytocin suppresses epithelial cell-derived cytokines production and alleviates intestinal inflammation in food allergy. Biochem Pharmacol 195:114867. 10.1016/j.bcp.2021.114867 [DOI] [PubMed] [Google Scholar]
7.Jankowski M et al (2010) Anti-inflammatory effect of oxytocin in rat myocardial infarction. Basic Res Cardiol 105(2):205–218. 10.1007/s00395-009-0076-5 [DOI] [PubMed] [Google Scholar]
8.Danhof HA et al (2023) Microbial stimulation of oxytocin release from the intestinal epithelium via secretin signaling. Gut Microbes 15(2):2256043. 10.1080/19490976.2023.2256043 [DOI] [PMC free article] [PubMed] [Google Scholar]
9.Dayanithi G et al (2018) Vasopressin and oxytocin in sensory neurones: Expression, exocytotic release and regulation by lactation. Sci Rep 8(1):13084. 10.1038/s41598-018-31361-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
10.Li E et al (2024) Control of lipolysis by a population of oxytocinergic sympathetic neurons. Nature 625(7993):175–180. 10.1038/s41586-023-06830-x [DOI] [PMC free article] [PubMed] [Google Scholar]
11.Ohlsson B et al (2006) Oxytocin is expressed throughout the human gastrointestinal tract. Regul Pept 135(1–2):7–11. 10.1016/j.regpep.2006.03.008 [DOI] [PubMed] [Google Scholar]
12.Oliveira-Pelegrin GR et al (2013) Oxytocin affects nitric oxide and cytokine production by sepsis-sensitized macrophages. Neuroimmunomodulation 20(2):65–71. 10.1159/000345044 [DOI] [PubMed] [Google Scholar]
13.Wang Y et al (2024) Oxytocin protects nigrostriatal dopamine signal via activating gabaergic circuit in the mptp-induced parkinson’s disease model. Adv Sci (Weinh) 11(36):e2310244. 10.1002/advs.202310244 [DOI] [PMC free article] [PubMed] [Google Scholar]
14.Lei Y et al (2024) Region-specific transcriptomic responses to obesity and diabetes in macaque hypothalamus. Cell Metab 36(2):438–453. e6 [DOI] [PubMed] [Google Scholar]
15.Tsurutani M et al (2024) Selective vulnerability of parvocellular oxytocin neurons in social dysfunction. Nat Commun 15(1):8661. 10.1038/s41467-024-53092-w [DOI] [PMC free article] [PubMed] [Google Scholar]
16.Purba JS, Hofman MA, Swaab DF (1994) Decreased number of oxytocin-immunoreactive neurons in the paraventricular nucleus of the hypothalamus in parkinson’s disease. Neurology 44(1):84–89. 10.1212/wnl.44.1.84 [DOI] [PubMed] [Google Scholar]
17.Purba JS et al (1993) Decreased number of oxytocin neurons in the paraventricular nucleus of the human hypothalamus in aids. Brain 116(Pt 4):795–809. 10.1093/brain/116.4.795 [DOI] [PubMed] [Google Scholar]
18.Kokona D et al (2018) Colony-stimulating factor 1 receptor inhibition prevents disruption of the blood-retina barrier during chronic inflammation. J Neuroinflammation 15(1):340. 10.1186/s12974-018-1373-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
19.Jurek B, Neumann ID (2018) The oxytocin receptor: From intracellular signaling to behavior. Physiol Rev 98(3):1805–1908. 10.1152/physrev.00031.2017 [DOI] [PubMed] [Google Scholar]
20.Dumais KM, Veenema AH (2016) Vasopressin and oxytocin receptor systems in the brain: Sex differences and sex-specific regulation of social behavior. Front Neuroendocrinol 40:1–23. 10.1016/j.yfrne.2015.04.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
21.Pramanik S et al (2024) Microglia signaling in health and disease - implications in sex-specific brain development and plasticity. Neurosci Biobehav Rev 165:105834. 10.1016/j.neubiorev.2024.105834 [DOI] [PubMed] [Google Scholar]
22.Lewis EM et al (2020) Parallel social information processing circuits are differentially impacted in autism. Neuron 108(4):659–675. e6 [DOI] [PMC free article] [PubMed] [Google Scholar]
23.Chen S et al (2022) Morpho-electric properties and diversity of oxytocin neurons in paraventricular nucleus of hypothalamus in female and male mice. J Neurosci 42(14):2885–2904. 10.1523/jneurosci.2494-21.2022 [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Madisen L et al (2010) A robust and high-throughput cre reporting and characterization system for the whole mouse brain. Nat Neurosci 13(1):133–140. 10.1038/nn.2467 [DOI] [PMC free article] [PubMed] [Google Scholar]
25.Matsunaga W et al (2000) Lps-induced fos expression in oxytocin and vasopressin neurons of the rat hypothalamus. Brain Res 858(1):9–18. 10.1016/s0006-8993(99)02418-x [DOI] [PubMed] [Google Scholar]
26.da Costa LHA et al (2021) Microglial activation modulates neuroendocrine secretion during experimental sepsis. Mol Neurobiol 58(5):2133–2144. 10.1007/s12035-020-02241-5 [DOI] [PubMed] [Google Scholar]
27.Eliava M et al (2016) A new population of parvocellular oxytocin neurons controlling magnocellular neuron activity and inflammatory pain processing. Neuron 89(6):1291–1304. 10.1016/j.neuron.2016.01.041 [DOI] [PMC free article] [PubMed] [Google Scholar]
28. Xiao L, Priest M, Nasenbeny J, Lu T, Kozorovitskiy Y. Biased Oxytocinergic Modulation of Midbrain Dopamine Systems. Neuron, 95(2): 368-384.e5 10.1016/j.neuron.2017.06.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
29.Yanuck SF (2019) Microglial phagocytosis of neurons: Diminishing neuronal loss in traumatic, infectious, inflammatory, and autoimmune cns disorders. Front Psychiatry 10:712. 10.3389/fpsyt.2019.00712 [DOI] [PMC free article] [PubMed] [Google Scholar]
30.Brown GC, Neher JJ (2014) Microglial phagocytosis of live neurons. Nat Rev Neurosci 15(4):209–216. 10.1038/nrn3710 [DOI] [PubMed] [Google Scholar]
31.Wu ZH et al (2025) Astrocyte-microglia crosstalk in the hippocampus mediates cognitive impairments induced by chronic intermittent hypoxia. Neurobiol Dis 216:107117. 10.1016/j.nbd.2025.107117 [DOI] [PubMed] [Google Scholar]
32.Osterhout JA et al (2022) A preoptic neuronal population controls fever and appetite during sickness. Nature 606(7916):937–944. 10.1038/s41586-022-04793-z [DOI] [PMC free article] [PubMed] [Google Scholar]
33.Liu Y et al (2024) Ng2 glia protect against prion neurotoxicity by inhibiting microglia-to-neuron prostaglandin e2 signaling. Nat Neurosci 27(8):1534–1544. 10.1038/s41593-024-01663-x [DOI] [PMC free article] [PubMed] [Google Scholar]
34.Summy-Long JY et al (2002) Nitric oxide, interleukin and prostaglandin interactions affecting the magnocellular system. Brain Res 940(1–2):10–20. 10.1016/s0006-8993(02)02556-8 [DOI] [PubMed] [Google Scholar]
35.Bygdeman M (2003) Pharmacokinetics of prostaglandins. Best Pract Res Clin Obstet Gynaecol 17(5):707–716. 10.1016/s1521-6934(03)00043-9 [DOI] [PubMed] [Google Scholar]
36.Jiao Y et al (2023) Exosomal pge2 from m2 macrophages inhibits neutrophil recruitment and net formation through lipid mediator class switching in sepsis. J Biomed Sci 30(1):62. 10.1186/s12929-023-00957-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
37.Palma BD et al (2009) Prostaglandin involvement in hyperthermia induced by sleep deprivation: A pharmacological and autoradiographic study. Life Sci, 84(9–10): pp. 278 – 81.10.1016/j.lfs.2008.12.009 [DOI] [PubMed]
38.Ohinata K et al (2006) Activation of prostaglandin e receptor ep4 subtype suppresses food intake in mice. Prostaglandins Other Lipid Mediat 81(1–2):31–36. 10.1016/j.prostaglandins.2006.06.008 [DOI] [PubMed] [Google Scholar]
39.Bonfill-Teixidor E et al (2017) Differential expression of e-type prostanoid receptors 2 and 4 in microglia stimulated with lipopolysaccharide. J Neuroinflammation 14(1):3. 10.1186/s12974-016-0780-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
40.Niraula A et al (2023) Prostaglandin pge2 receptor ep4 regulates microglial phagocytosis and increases susceptibility to diet-induced obesity. Diabetes 72(2):233–244. 10.2337/db21-1072 [DOI] [PMC free article] [PubMed] [Google Scholar]
41.Woodling NS et al (2014) Suppression of alzheimer-associated inflammation by microglial prostaglandin-e2 ep4 receptor signaling. J Neurosci 34(17):5882–5894. 10.1523/jneurosci.0410-14.2014 [DOI] [PMC free article] [PubMed] [Google Scholar]
42.Zhang J et al (2024) Projections from subfornical organ to bed nucleus of the stria terminalis modulate inflammation-induced anxiety-like behaviors in mice. Sci Adv 10(48):eadp9413. 10.1126/sciadv.adp9413 [DOI] [PMC free article] [PubMed] [Google Scholar]
43.Yadollahzadeh M et al (2023) Variable levels of oxytocin during sepsis: The role of oxytocin in sepsis pathophysiology. J Intensive Care Med 38(11):997–1002. 10.1177/08850666231177255 [DOI] [PubMed] [Google Scholar]
44.Li P et al (2014) Protective effects of the active fraction from the tuber of scirpus yagara in mouse endotoxin shock model. J Ethnopharmacol 158 Pt A 331–337. 10.1016/j.jep.2014.10.032 [DOI] [PubMed]
45.Shi CN et al (2024) Oxytocin attenuates neuroinflammation-induced anxiety through restoration of excitation and inhibition balance in the anterior cingulate cortex in mice. J Affect Disord 362:341–355. 10.1016/j.jad.2024.05.144 [DOI] [PubMed] [Google Scholar]
46.Asker M et al (2023) Peripherally restricted oxytocin is sufficient to reduce food intake and motivation, while cns entry is required for locomotor and taste avoidance effects. Diabetes Obes Metab 25(3):856–877. 10.1111/dom.14937 [DOI] [PMC free article] [PubMed] [Google Scholar]
47.Dunietz GL et al (2024) Oxytocin and women’s health in midlife. J Endocrinol 262(1). 10.1530/joe-23-0396 [DOI] [PMC free article] [PubMed]
48.Dantzer R et al (2008) From inflammation to sickness and depression: When the immune system subjugates the brain. Nat Rev Neurosci 9(1):46–56. 10.1038/nrn2297 [DOI] [PMC free article] [PubMed] [Google Scholar]
49.Li T et al (2016) Approaches mediating oxytocin regulation of the immune system. Front Immunol 7:693. 10.3389/fimmu.2016.00693 [DOI] [PMC free article] [PubMed] [Google Scholar]
50.Hu Y, Tao W (2024) Current perspectives on microglia-neuron communication in the central nervous system: Direct and indirect modes of interaction. J Adv Res 66:251–265. 10.1016/j.jare.2024.01.006 [DOI] [PMC free article] [PubMed] [Google Scholar]
51.Hong S et al (2016) Complement and microglia mediate early synapse loss in alzheimer mouse models. Science 352(6286):712–716. 10.1126/science.aad8373 [DOI] [PMC free article] [PubMed] [Google Scholar]
52.Wake H et al (2009) Resting microglia directly monitor the functional state of synapses in vivo and determine the fate of ischemic terminals. J Neurosci 29(13):3974–3980. 10.1523/jneurosci.4363-08.2009 [DOI] [PMC free article] [PubMed] [Google Scholar]
53.Rosin JM et al (2021) A subpopulation of embryonic microglia respond to maternal stress and influence nearby neural progenitors. Dev Cell 56(9):1326–1345. 10.1016/j.devcel.2021.03.018 [DOI] [PubMed] [Google Scholar]
54.Spiteri AG et al (2022) Plx5622 reduces disease severity in lethal cns infection by off-target inhibition of peripheral inflammatory monocyte production. Front Immunol 13:851556. 10.3389/fimmu.2022.851556 [DOI] [PMC free article] [PubMed] [Google Scholar]
55.Faust TE et al (2023) A comparative analysis of microglial inducible cre lines. Cell Rep 42(9):113031. 10.1016/j.celrep.2023.113031 [DOI] [PMC free article] [PubMed] [Google Scholar]
56.Cheng H et al (2021) Role of prostaglandin e2 in tissue repair and regeneration. Theranostics 11(18):8836–8854. 10.7150/thno.63396 [DOI] [PMC free article] [PubMed] [Google Scholar]
57.Beier EE et al (2017) Alternative microglial activation is associated with cessation of progressive dopamine neuron loss in mice systemically administered lipopolysaccharide. Neurobiol Dis 108:115–127. 10.1016/j.nbd.2017.08.009 [DOI] [PMC free article] [PubMed] [Google Scholar]
58.Yan C et al (2023) Carvacrol protects mice against lps-induced sepsis and attenuates inflammatory response in macrophages by modulating the erk1/2 pathway. Sci Rep 13(1):12809. 10.1038/s41598-023-39665-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
59.Cui X et al (2023) A putative loop connection between vta dopamine neurons and nucleus accumbens encodes positive valence to compensate for hunger. Prog Neurobiol 229:102503. 10.1016/j.pneurobio.2023.102503 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Material 1^{(6.2MB, docx)}

Data Availability Statement

The authors confirm that the data supporting the findings of this study are available within the article and its Supplementary material.

[CR1] 1.Cui X, Xiao L (2025) Complexity of the hypothalamic oxytocin system and its involvement in brain functions and diseases. Neurosci Bull 41(7):1267–1288. 10.1007/s12264-025-01424-1 [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR2] 2.Jiang J et al (2023) Intertwined associations between oxytocin, immune system and major depressive disorder. Biomed Pharmacother 163:114852. 10.1016/j.biopha.2023.114852 [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR3] 3.Yang M et al (2023) Oxytocin improves intracerebral hemorrhage outcomes by suppressing neuronal pyroptosis and mitochondrial fission. Stroke 54(7):1888–1900. 10.1161/strokeaha.123.043391 [DOI] [PubMed] [Google Scholar]

[CR4] 4.Jiang J et al (2023) Oxytocin alleviates cognitive and memory impairments by decreasing hippocampal microglial activation and synaptic defects via oxtr/erk/stat3 pathway in a mouse model of sepsis-associated encephalopathy. Brain Behav Immun 114:195–213. 10.1016/j.bbi.2023.08.023 [DOI] [PubMed] [Google Scholar]

[CR5] 5.Clodi M et al (2008) Oxytocin alleviates the neuroendocrine and cytokine response to bacterial endotoxin in healthy men. Am J Physiol Endocrinol Metab 295(3):E686–E691. 10.1152/ajpendo.90263.2008 [DOI] [PubMed] [Google Scholar]

[CR6] 6.Yu Y, Li J, Liu C (2022) Oxytocin suppresses epithelial cell-derived cytokines production and alleviates intestinal inflammation in food allergy. Biochem Pharmacol 195:114867. 10.1016/j.bcp.2021.114867 [DOI] [PubMed] [Google Scholar]

[CR7] 7.Jankowski M et al (2010) Anti-inflammatory effect of oxytocin in rat myocardial infarction. Basic Res Cardiol 105(2):205–218. 10.1007/s00395-009-0076-5 [DOI] [PubMed] [Google Scholar]

[CR8] 8.Danhof HA et al (2023) Microbial stimulation of oxytocin release from the intestinal epithelium via secretin signaling. Gut Microbes 15(2):2256043. 10.1080/19490976.2023.2256043 [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR9] 9.Dayanithi G et al (2018) Vasopressin and oxytocin in sensory neurones: Expression, exocytotic release and regulation by lactation. Sci Rep 8(1):13084. 10.1038/s41598-018-31361-1 [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR10] 10.Li E et al (2024) Control of lipolysis by a population of oxytocinergic sympathetic neurons. Nature 625(7993):175–180. 10.1038/s41586-023-06830-x [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR11] 11.Ohlsson B et al (2006) Oxytocin is expressed throughout the human gastrointestinal tract. Regul Pept 135(1–2):7–11. 10.1016/j.regpep.2006.03.008 [DOI] [PubMed] [Google Scholar]

[CR12] 12.Oliveira-Pelegrin GR et al (2013) Oxytocin affects nitric oxide and cytokine production by sepsis-sensitized macrophages. Neuroimmunomodulation 20(2):65–71. 10.1159/000345044 [DOI] [PubMed] [Google Scholar]

[CR13] 13.Wang Y et al (2024) Oxytocin protects nigrostriatal dopamine signal via activating gabaergic circuit in the mptp-induced parkinson’s disease model. Adv Sci (Weinh) 11(36):e2310244. 10.1002/advs.202310244 [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR14] 14.Lei Y et al (2024) Region-specific transcriptomic responses to obesity and diabetes in macaque hypothalamus. Cell Metab 36(2):438–453. e6 [DOI] [PubMed] [Google Scholar]

[CR15] 15.Tsurutani M et al (2024) Selective vulnerability of parvocellular oxytocin neurons in social dysfunction. Nat Commun 15(1):8661. 10.1038/s41467-024-53092-w [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR16] 16.Purba JS, Hofman MA, Swaab DF (1994) Decreased number of oxytocin-immunoreactive neurons in the paraventricular nucleus of the hypothalamus in parkinson’s disease. Neurology 44(1):84–89. 10.1212/wnl.44.1.84 [DOI] [PubMed] [Google Scholar]

[CR17] 17.Purba JS et al (1993) Decreased number of oxytocin neurons in the paraventricular nucleus of the human hypothalamus in aids. Brain 116(Pt 4):795–809. 10.1093/brain/116.4.795 [DOI] [PubMed] [Google Scholar]

[CR18] 18.Kokona D et al (2018) Colony-stimulating factor 1 receptor inhibition prevents disruption of the blood-retina barrier during chronic inflammation. J Neuroinflammation 15(1):340. 10.1186/s12974-018-1373-4 [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR19] 19.Jurek B, Neumann ID (2018) The oxytocin receptor: From intracellular signaling to behavior. Physiol Rev 98(3):1805–1908. 10.1152/physrev.00031.2017 [DOI] [PubMed] [Google Scholar]

[CR20] 20.Dumais KM, Veenema AH (2016) Vasopressin and oxytocin receptor systems in the brain: Sex differences and sex-specific regulation of social behavior. Front Neuroendocrinol 40:1–23. 10.1016/j.yfrne.2015.04.003 [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR21] 21.Pramanik S et al (2024) Microglia signaling in health and disease - implications in sex-specific brain development and plasticity. Neurosci Biobehav Rev 165:105834. 10.1016/j.neubiorev.2024.105834 [DOI] [PubMed] [Google Scholar]

[CR22] 22.Lewis EM et al (2020) Parallel social information processing circuits are differentially impacted in autism. Neuron 108(4):659–675. e6 [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR23] 23.Chen S et al (2022) Morpho-electric properties and diversity of oxytocin neurons in paraventricular nucleus of hypothalamus in female and male mice. J Neurosci 42(14):2885–2904. 10.1523/jneurosci.2494-21.2022 [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR24] 24.Madisen L et al (2010) A robust and high-throughput cre reporting and characterization system for the whole mouse brain. Nat Neurosci 13(1):133–140. 10.1038/nn.2467 [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR25] 25.Matsunaga W et al (2000) Lps-induced fos expression in oxytocin and vasopressin neurons of the rat hypothalamus. Brain Res 858(1):9–18. 10.1016/s0006-8993(99)02418-x [DOI] [PubMed] [Google Scholar]

[CR26] 26.da Costa LHA et al (2021) Microglial activation modulates neuroendocrine secretion during experimental sepsis. Mol Neurobiol 58(5):2133–2144. 10.1007/s12035-020-02241-5 [DOI] [PubMed] [Google Scholar]

[CR27] 27.Eliava M et al (2016) A new population of parvocellular oxytocin neurons controlling magnocellular neuron activity and inflammatory pain processing. Neuron 89(6):1291–1304. 10.1016/j.neuron.2016.01.041 [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR28] 28. Xiao L, Priest M, Nasenbeny J, Lu T, Kozorovitskiy Y. Biased Oxytocinergic Modulation of Midbrain Dopamine Systems. Neuron, 95(2): 368-384.e5 10.1016/j.neuron.2017.06.003 [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR29] 29.Yanuck SF (2019) Microglial phagocytosis of neurons: Diminishing neuronal loss in traumatic, infectious, inflammatory, and autoimmune cns disorders. Front Psychiatry 10:712. 10.3389/fpsyt.2019.00712 [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR30] 30.Brown GC, Neher JJ (2014) Microglial phagocytosis of live neurons. Nat Rev Neurosci 15(4):209–216. 10.1038/nrn3710 [DOI] [PubMed] [Google Scholar]

[CR31] 31.Wu ZH et al (2025) Astrocyte-microglia crosstalk in the hippocampus mediates cognitive impairments induced by chronic intermittent hypoxia. Neurobiol Dis 216:107117. 10.1016/j.nbd.2025.107117 [DOI] [PubMed] [Google Scholar]

[CR32] 32.Osterhout JA et al (2022) A preoptic neuronal population controls fever and appetite during sickness. Nature 606(7916):937–944. 10.1038/s41586-022-04793-z [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR33] 33.Liu Y et al (2024) Ng2 glia protect against prion neurotoxicity by inhibiting microglia-to-neuron prostaglandin e2 signaling. Nat Neurosci 27(8):1534–1544. 10.1038/s41593-024-01663-x [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR34] 34.Summy-Long JY et al (2002) Nitric oxide, interleukin and prostaglandin interactions affecting the magnocellular system. Brain Res 940(1–2):10–20. 10.1016/s0006-8993(02)02556-8 [DOI] [PubMed] [Google Scholar]

[CR35] 35.Bygdeman M (2003) Pharmacokinetics of prostaglandins. Best Pract Res Clin Obstet Gynaecol 17(5):707–716. 10.1016/s1521-6934(03)00043-9 [DOI] [PubMed] [Google Scholar]

[CR36] 36.Jiao Y et al (2023) Exosomal pge2 from m2 macrophages inhibits neutrophil recruitment and net formation through lipid mediator class switching in sepsis. J Biomed Sci 30(1):62. 10.1186/s12929-023-00957-9 [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR37] 37.Palma BD et al (2009) Prostaglandin involvement in hyperthermia induced by sleep deprivation: A pharmacological and autoradiographic study. Life Sci, 84(9–10): pp. 278 – 81.10.1016/j.lfs.2008.12.009 [DOI] [PubMed]

[CR38] 38.Ohinata K et al (2006) Activation of prostaglandin e receptor ep4 subtype suppresses food intake in mice. Prostaglandins Other Lipid Mediat 81(1–2):31–36. 10.1016/j.prostaglandins.2006.06.008 [DOI] [PubMed] [Google Scholar]

[CR39] 39.Bonfill-Teixidor E et al (2017) Differential expression of e-type prostanoid receptors 2 and 4 in microglia stimulated with lipopolysaccharide. J Neuroinflammation 14(1):3. 10.1186/s12974-016-0780-7 [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR40] 40.Niraula A et al (2023) Prostaglandin pge2 receptor ep4 regulates microglial phagocytosis and increases susceptibility to diet-induced obesity. Diabetes 72(2):233–244. 10.2337/db21-1072 [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR41] 41.Woodling NS et al (2014) Suppression of alzheimer-associated inflammation by microglial prostaglandin-e2 ep4 receptor signaling. J Neurosci 34(17):5882–5894. 10.1523/jneurosci.0410-14.2014 [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR42] 42.Zhang J et al (2024) Projections from subfornical organ to bed nucleus of the stria terminalis modulate inflammation-induced anxiety-like behaviors in mice. Sci Adv 10(48):eadp9413. 10.1126/sciadv.adp9413 [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR43] 43.Yadollahzadeh M et al (2023) Variable levels of oxytocin during sepsis: The role of oxytocin in sepsis pathophysiology. J Intensive Care Med 38(11):997–1002. 10.1177/08850666231177255 [DOI] [PubMed] [Google Scholar]

[CR44] 44.Li P et al (2014) Protective effects of the active fraction from the tuber of scirpus yagara in mouse endotoxin shock model. J Ethnopharmacol 158 Pt A 331–337. 10.1016/j.jep.2014.10.032 [DOI] [PubMed]

[CR45] 45.Shi CN et al (2024) Oxytocin attenuates neuroinflammation-induced anxiety through restoration of excitation and inhibition balance in the anterior cingulate cortex in mice. J Affect Disord 362:341–355. 10.1016/j.jad.2024.05.144 [DOI] [PubMed] [Google Scholar]

[CR46] 46.Asker M et al (2023) Peripherally restricted oxytocin is sufficient to reduce food intake and motivation, while cns entry is required for locomotor and taste avoidance effects. Diabetes Obes Metab 25(3):856–877. 10.1111/dom.14937 [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR47] 47.Dunietz GL et al (2024) Oxytocin and women’s health in midlife. J Endocrinol 262(1). 10.1530/joe-23-0396 [DOI] [PMC free article] [PubMed]

[CR48] 48.Dantzer R et al (2008) From inflammation to sickness and depression: When the immune system subjugates the brain. Nat Rev Neurosci 9(1):46–56. 10.1038/nrn2297 [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR49] 49.Li T et al (2016) Approaches mediating oxytocin regulation of the immune system. Front Immunol 7:693. 10.3389/fimmu.2016.00693 [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR50] 50.Hu Y, Tao W (2024) Current perspectives on microglia-neuron communication in the central nervous system: Direct and indirect modes of interaction. J Adv Res 66:251–265. 10.1016/j.jare.2024.01.006 [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR51] 51.Hong S et al (2016) Complement and microglia mediate early synapse loss in alzheimer mouse models. Science 352(6286):712–716. 10.1126/science.aad8373 [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR52] 52.Wake H et al (2009) Resting microglia directly monitor the functional state of synapses in vivo and determine the fate of ischemic terminals. J Neurosci 29(13):3974–3980. 10.1523/jneurosci.4363-08.2009 [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR53] 53.Rosin JM et al (2021) A subpopulation of embryonic microglia respond to maternal stress and influence nearby neural progenitors. Dev Cell 56(9):1326–1345. 10.1016/j.devcel.2021.03.018 [DOI] [PubMed] [Google Scholar]

[CR54] 54.Spiteri AG et al (2022) Plx5622 reduces disease severity in lethal cns infection by off-target inhibition of peripheral inflammatory monocyte production. Front Immunol 13:851556. 10.3389/fimmu.2022.851556 [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR55] 55.Faust TE et al (2023) A comparative analysis of microglial inducible cre lines. Cell Rep 42(9):113031. 10.1016/j.celrep.2023.113031 [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR56] 56.Cheng H et al (2021) Role of prostaglandin e2 in tissue repair and regeneration. Theranostics 11(18):8836–8854. 10.7150/thno.63396 [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR57] 57.Beier EE et al (2017) Alternative microglial activation is associated with cessation of progressive dopamine neuron loss in mice systemically administered lipopolysaccharide. Neurobiol Dis 108:115–127. 10.1016/j.nbd.2017.08.009 [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR58] 58.Yan C et al (2023) Carvacrol protects mice against lps-induced sepsis and attenuates inflammatory response in macrophages by modulating the erk1/2 pathway. Sci Rep 13(1):12809. 10.1038/s41598-023-39665-7 [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR59] 59.Cui X et al (2023) A putative loop connection between vta dopamine neurons and nucleus accumbens encodes positive valence to compensate for hunger. Prog Neurobiol 229:102503. 10.1016/j.pneurobio.2023.102503 [DOI] [PubMed] [Google Scholar]

PERMALINK

Prostaglandin signaling drives peripheral inflammation-induced reduction of hypothalamic oxytocin-positive neurons

Yibing Li

Hao Xu

Hongzhi Guo

Mingrui Zhai

Li Cao

Lei Xiao

Abstract

Supplementary Information

Introduction

Results

The reduction in the number of PVN OXT-immunopositive neurons in the LPS-induced chronic inflammation model

Fig. 1.

A selective reduction of PVN Magno, but not Parvo, OXT-immunopositive neurons in the LPS-induced chronic inflammation model

Fig. 2.

PVN Magno and Parvo OXT neurons exhibit opposite activity changes in LPS challenge

Fig. 3.

Microglia are mainly activated in the anterior part of PVN to phagocyte Magno OXT neurons

Fig. 4.

Prostaglandin signals are elevated in the PVN under LPS challenge

Fig. 5.

Inhibition of prostaglandin signals alleviates LPS-induced reduction of PVN OXT-immunopositive neurons

Fig. 6.

Knockdown of EP4 receptors in PVN OXT neurons alleviates the LPS-induced reduction of OXT-immunopositive neurons

Fig. 7.

Discussion

Fig. 8.

Materials and methods

Mouse strains

LPS-induced peripheral inflammation models

Tail vein injection

In vivo stereotaxic intracranial injection

Cannula implantation

Sickness behavior

Peripheral blood serum extraction and ELISA

Immunohistochemistry, imaging and analyses

Three-dimensional (3D) image reconstruction

In vitro electrophysiological recording

RNA extraction and reverse transcription-quantitative PCR (RT-qPCR)

RNA sequencing and analyses

Data analysis

Supplementary Information

Acknowledgements

Author Contributions

Data availability

Declarations

Ethics approval

Competing interests

Footnotes

Contributor Information

References

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases