Involvement of the hypothalamus–raphe magnus–spinal defecation center axis in stress-induced defecation in rats

Abstract

Acute psychological stress stimulates colonic motility; however, the specific descending neural pathways linking the brain to the distal colon remain incompletely characterized. Here, we provide evidence that the hypothalamus‒raphe magnus‒spinal cord‒pelvic nerve pathway mediates stress-induced defecation in male Sprague–Dawley rats. Using urethane-anesthetized rats, we used viral tracing, pharmacological, and chemogenetic approaches to show that hypothalamic neurons projecting to the raphe magnus nucleus activate descending serotonergic neurons probably via oxytocinergic signaling, thereby stimulating colorectal motility through the pelvic nerve. In conscious rats, inhibiting this pathway suppresses stress-induced defecation. These findings highlight the role of the hypothalamus‒raphe magnus‒spinal defecation center axis in transmitting acute psychological stress to the colorectum.

The hypothalamic neurons projecting to the raphe magnus nucleus activate descending serotonergic neurons, thereby activating spinal defecation center. The pathway plays a role in transmitting acute psychological stress to the colorectum.

Introduction

Colorectal motility is regulated by the coordinated activity of the enteric nervous system and the central nervous system. Among central structures, the hypothalamic paraventricular nucleus (PVH) and dorsomedial hypothalamus (DMH) play key roles in the autonomic and emotional modulation of gastrointestinal function¹. Acute psychological stress enhances colonic motility through the activation of hypothalamic and brainstem nuclei, including the PVH, DMH, and locus coeruleus/subcoeruleus^2–8. These regions are thought to interact with descending pathways that influence spinal autonomic centers^9–11. In addition to these excitatory mechanisms, accumulating evidence indicates that oxytocin (OXT) within the hypothalamus exerts inhibitory effects on stress-induced activation of colonic motility in conscious rodents^12–14. These findings suggest that hypothalamic OXT neurons may play a counterregulatory role in balancing the excitatory pathways that cause stress-induced defecation and that the balance between excitatory and inhibitory hypothalamic inputs is critical for maintaining normal bowel function under stressful conditions.

Several lines of evidence indicate that the vagus nerve serves as a neural route that transmits stress signals from the brain to the colon^15–18. A recent study using a retrograde tracing method revealed that the distal colon (in addition to the proximal colon) receives vagal innervation in mice¹⁹, thereby emphasizing the important role of the vagus nerve in stress-induced defecation. However, the vagal pathway may not fully account for the mechanism of stress-induced defecation (at least in other animal species, including humans and rats) because the autonomic nerves from the lumbosacral spinal cord also project to the distal colon and the rectum^20,21. In accordance with this scenario, the crucial roles of the spinal defecation center in the lumbosacral spinal cord and the supraspinal defecation center in the brainstem in the regulation of motility in the distal colon and rectum have been documented^21–23. We have previously demonstrated that serotonergic neurons projecting from the raphe magnus nucleus to the lumbosacral spinal cord activate preganglionic neurons of the parasympathetic pelvic nerve and consequently increase colorectal motility^24,25. The involvement of descending serotonergic neurons located in the raphe magnus nucleus in the central regulatory mechanism of colorectal motility was further validated by experiments using the chemogenetic method, in which inhibition of the neurons suppressed colorectal motility induced by intracolonic noxious stimulus in anesthetized rats or defecation after the application of psychological stress in conscious rats²⁴. The aim of this study was to demonstrate that hypothalamic neurons enhance colorectal motility through activation of the raphe magnus. Our findings revealed that the hypothalamus-raphe-spinal defecation center axis is a neural pathway that transmits signals from the brain to the colorectum in response to acute stress.

Results

The medullary raphe region receives neural inputs related to psychological stress from the hypothalamic nuclei

We first assessed the hypothalamic regions, which may provide neural input to the raphe magnus nucleus in response to psychological stress. Neurons projecting to the raphe magnus nucleus were labeled with EGFP by using retrograde adeno-associated virus (AAV) vector injection into the raphe magnus nucleus (2.9 mm caudal to bregma, on the midline, 10.9 mm ventral to the bregma) (Fig. 1A). Some of the EGFP-labeled axons were observed in the raphe magnus and pallidus nuclei (Fig. 1B). The rats were subsequently exposed to water avoidance stress (WAS) for 2 h (Fig. 1C). Representative fluorescence images showing colocalization (yellow) of immunostained EGFP (green) and c-Fos (red) in the hypothalamus are shown in Fig. 1D, E. A schematic of these results in two rats per group is shown in Fig. S1, and a representative of these results is shown in Fig. 1F. AAV-labeled neurons were predominantly found in the paraventricular hypothalamic nucleus (PVH) and the dorsal part of the caudal hypothalamus, including the dorsomedial hypothalamus (DMH), dorsal hypothalamic area (DA) and lateral hypothalamic area (LH) (Fig. 1F). Compared with control rats, c-Fos expression in these EGFP-labeled neurons was significantly greater among WAS-exposed rats (Fig. 1G).

Fig. 1. Water avoidance stress activates hypothalamic neurons projecting to the medullary raphe region.

A Schematic outline of retrograde labeling from the raphe magnus nucleus by injection of AAVretro. B EGFP-expressing neurons in the medullary raphe region. py: pyramidal tract, RMg: raphe magnus nucleus, rRPa: rostral raphe pallidus nucleus. Scale bar: 200 μm. C Setup for the water avoidance stress (WAS) test. An observation cage with a central platform was filled with water to a level 1 cm below the top of the platform, and the rats were placed on the platform for 2 h. The control rats were subjected to the same procedures without water. EGFP-positive (green) and stress-induced c-Fos immunoreactive (red) cells in the PVH (D) and the DMH, DA and LH (E). The filled arrowheads indicate EGFP-positive neurons coexpressing c-Fos; the open arrowheads indicate EGFP-positive neurons without c-Fos expression. 3V, third ventricle; mt, mammillothalamic tract. Scale bars: 100 μm (main), 30 μm (inset). F Coronal images of hypothalamic sections, indicating distances caudal to the bregma. Distribution of EGFP-positive neurons with or without c-Fos immunoreactivity in control and stress-exposed rats. f, fornix. G Percentages of c-Fos-positive cells among EGFP-labeled populations in hypothalamus regions (n = 4 rats per group). The data are presented as the means ± SEMs. *P < 0.05.

Glutamatergic activation of the hypothalamic nuclei enhances colorectal motility

We also examined whether activation of the identified hypothalamic regions enhanced colorectal motility in rats anesthetized with urethane. Since glutamatergic neural inputs to the hypothalamus are pivotal for stress-induced responses^9,11,26, we injected the AMPA agonist (S)-AMPA to activate neurons in the hypothalamus (Fig. 2A). After recording, the injection site was marked by green microspheres (Fig. 2B). Colorectal motility was measured via the method shown in Fig. 2C. When (S)-AMPA was injected into the hypothalamus, a marked increase in colorectal motility was observed (Fig. 2D). Moreover, the arterial pressure and heart rate clearly increased (Fig. 2D). The results of (S)-AMPA injections into various hypothalamic sites are summarized in Fig. 2E, where darker red indicates stronger responses of colorectal motility. The strongest colorectal motor responses were observed when the drug was injected near the third ventricle, 2.1–2.4 mm caudal to the bregma, corresponding to the caudal PVH, DA and rostral DMH. In contrast, almost all the sites of (S)-AMPA injections, including sites that did not enhance colorectal motility, elicited increases in arterial pressure and heart rate (Fig. 2F–H). Diagrams showing the increases in blood pressure and heart rate in shades of red, corresponding to Fig. 2E, are shown in Fig. S2.

Fig. 2. Colorectal motility is enhanced by the glutamatergic activation of the hypothalamic nuclei.

A Schematic outline of (S)-AMPA injection into the hypothalamus. B Injection site marked with FluoSpheres. C In vivo recording using anesthetized rats. The colorectum was cannulated at the distal colon and anus and filled with saline solution. The intraluminal pressure and expelled fluid volume were measured. D Representative trace showing increased colorectal motility, arterial pressure and heart rate following (S)-AMPA injection. E Coronal sections of the hypothalamus corresponding to atlas plates (modified from Paxinos & Watson, 2007), showing distances caudal to the bregma. Red shading indicates the strength of the colorectal motility response to (S)-AMPA injection in 43 rats. Circles represent injection sites, and the adjacent numbers indicate the number of propulsive colorectal contractions. (See Fig. S1 for the arterial pressure and heart rate.) Changes in propulsive contractions of the colorectum (F), mean arterial pressure (MAP; G) and heart rate (HR; H) after (S)-AMPA injections. The data are grouped by the number of colorectal contractions [<10: crosses (n = 31), 10–25: filled triangles (n = 8), >25: filled circles (n = 4)]. Following this grouping, colorectal contractions significantly increased in the groups with >10 contractions. The MAP and HR increased regardless of group. **P < 0.01, ****P < 0.0001. Box-and-whisker plots show the data distribution: boxes represent the interquartile range (25th–75th percentile), the line inside the box indicates the median, whiskers represent the minimum and maximum, and the plus sign (+) indicates the mean.

To determine whether the colorectal motor response induced by (S)-AMPA injection into the hypothalamus is mediated by central CRH transmission, we examined the effect of intracerebroventricular (i.c.v.) preadministration of the CRH antagonist astressin on the colorectal motor response under urethane anesthesia (Fig. S3A). The injection sites of (S)-AMPA are shown in Supplementary Fig. 3B. When astressin was preadministered i.c.v., the injection of (S)-AMPA into the hypothalamus increased colorectal motility to an extent that was comparable to that observed with i.c.v. saline administration (Fig. S3C–E).

The Colorectal motor response to hypothalamic activation is mediated by descending serotonergic neurons and the pelvic nerve

To determine whether the enhanced colorectal motility in response to the activation of hypothalamic neurons is mediated by descending serotonergic neurons, the effects of the intrathecal administration of serotonergic receptor blockers into the lumbosacral spinal cord were examined under urethane anesthesia (Fig. 3A). Figure 3B shows the injection sites of the drug. When saline was preadministered into the lumbosacral spinal cord, the injection of (S)-AMPA into the hypothalamus increased colorectal motility (Fig. 3C, E). In contrast, prior intrathecal administration of ketanserin and dolasetron, which are 5-HT2 and 5-HT3 receptor antagonists, respectively, completely blocked the colorectal motor response induced by (S)-AMPA injection (Fig. 3D, E). Then, we determined the effects of surgical excision of the pelvic nerve (Fig. 3F). Figure 3G shows the injection sites of (S)-AMPA. In sham-operated control rats, injection of (S)-AMPA into the caudal PVH, DA or rostral DMH increased colorectal motility (Fig. 3H, J). In rats in which the pelvic nerve was cut, activation of these hypothalamic regions failed to increase colorectal motility (Fig. 3I, J).

Fig. 3. Spinal serotonergic transmission and the pelvic nerve mediates the colorectal motor response to hypothalamic activation.

A Schematic outline for (B–E). B (S)-AMPA injection sites following the intrathecal (i.t.) administration of saline or ketanserin/dolasetron are indicated by circles and triangles, respectively. Representative traces of colorectal intraluminal pressure and expelled volume after (S)-AMPA injection prior to i.t. administration of saline (C) or ketanserin/dolasetron (D). The traces shown are from different animals. E Summarized graph showing changes in propulsive colorectal contractions following (S)-AMPA injection after saline or ketanserin/dolasetron i.t. administration (n = 5 per group). **P < 0.01. Lines represent median values. F Schematic outline for (G–J). G (S)-AMPA injection sites in control rats or pelvic nerve-transected rats are indicated by circles and triangles, respectively. Representative traces of colorectal intraluminal pressure and expelled volume following (S)-AMPA injection in control (H) and pelvic nerve-transected (I) rats. J Summarized graph showing changes in propulsive colorectal contractions after (S)-AMPA injection in control and pelvic nerve-transected rats (n = 5 per group). P < 0.01. Lines represent median values.

Oxytocinergic transmission in the raphe magnus nucleus is involved in the colorectal motor response to hypothalamic activation

Next, we identified the neurotransmitter that activates the descending serotonergic pathway in the raphe magnus nucleus. We focused on glutamate and oxytocin, since they have been reported as neurotransmitters of hypothalamic neurons projecting to the medullary raphe region^9,11,27. In this series of experiments, antagonists were administered to the raphe magnus nucleus under a condition of increased colorectal motility by prior hypothalamic injection of (S)-AMPA in urethane-anesthetized rats (Fig. 4A). The injection sites of (S)-AMPA in the hypothalamus are shown in Fig. 4B. The injection sites of the antagonists were mainly located in the raphe magnus, with minimal spread into the rostral raphe pallidus and pyramidal tract being observed (Fig. 4C). Vehicle injection into the raphe magnus nucleus had no effect (Fig. 4D). An oxytocin receptor antagonist, L-368,899, suppressed the increase in motility (Fig. 4E). In contrast, a combined injection of the glutamate receptor antagonists AP5 and CNQX did not exert apparent suppressive effects on colorectal motility (Fig. 4F). Quantitative analysis revealed that the suppressive effect of the oxytocin receptor antagonist, but not the suppressive effect of glutamate receptor antagonist, was significant compared with that of the vehicle control (Fig. 4G).

Fig. 4. Oxytocinergic transmission in the raphe magnus nucleus is involved in the colorectal motor response to hypothalamic activation.

A Schematic outline for B–G. Injection sites of (S)-AMPA in the hypothalamus (B) and vehicle, L-368,899 or AP5/CNQX in the raphe magnus-centered medullary raphe region (C) are shown as circles, triangles and squares, respectively. Representative traces showing the effects of antagonists on colorectal intraluminal pressure and expelled volume. Vehicle (D), L-368,899 (E) or AP5/CNQX (F) was administered to the raphe magnus nucleus under a condition of increased colorectal motility by prior hypothalamic injection of (S)-AMPA. Traces are from different animals. G Summarized graph showing changes in propulsive colorectal contractions after raphe injection of vehicle (n = 4), L-368,899 (n = 7) or AP5/CNQX (n = 5). **P < 0.01; ns: not significant. Lines indicate median values. H Schematic outline for (I–K). Hypothalamic neural axons were labeled with AAV-hrGFP, and descending neurons were labeled with FluoroGold. I Fluorescence images showing hrGFP labeling and oxytocin (OXT) immunoreactivity in the PVH and DA. Scale bars: 100 μm. J Injection site of FluoroGold in the spinal cord. Scale bars: 200 μm. K Confocal images showing hrGFP-labeled (a) and OXT-positive (b) axon swellings (arrowheads) are adjacent to FluoroGold-labeled (c) and TPH2-positive (d) cell bodies in the raphe magnus nucleus. Scale bar: 10 μm.

To obtain morphological evidence for direct connections between hypothalamic oxytocin neurons and raphe serotonergic neurons projecting to the lumbosacral spinal cord, immunohistochemistry combined with tracer labeling was carried out (Fig. 4H). One week after injection of the AAV vector into the hypothalamus, hrGFP was expressed in many cells in the PVH, DA, and DMH (Fig. 4I). Their axonal fibers were clearly visualized and found to be abundantly distributed in the raphe magnus nucleus (Fig. 4K-a). Some of the hrGFP-labeled axons were immunostained with antibodies against oxytocin (Fig. 4K-a and -b). When a retrograde tracer (i.e., FluoroGold) was injected into the lumbosacral spinal cord, many spinal neurons were labeled with the tracer (Fig. 4J), and the tracer was detected in the raphe magnus nucleus (Fig. 4K-c). Some of the FluoroGold-positive neurons were immunostained with antibodies against TPH2, a marker of serotonergic neurons (Fig. 4K-d). Merged images revealed a close association between these axons and descending serotonergic neurons in the raphe magnus nucleus (Fig. 4K-e), indicating the presence of putative synapses formed by hypothalamus-derived axons on descending serotonergic neurons in the raphe magnus nucleus.

The activation of the hypothalamus→raphe pathway causes psychological stress-induced defecation

We examined whether activated hypothalamic neurons would increase colorectal motility through their direct projection to descending serotonergic neurons in the raphe magnus nucleus. To express the fluorescent protein mCherry and the inhibitory receptor hM4Di specifically in hypothalamic neurons projecting to the raphe magnus nucleus, we used the double virus vector infection method. A retrograde AAV vector for the expression of Cre recombinase was injected into the raphe magnus nucleus, and another AAV vector for Cre-dependent expression of mCherry and hM4Di was injected into the hypothalamic site, in which colorectal motility-regulating neurons are thought to be localized (Fig. 5A). The descending neurons in the raphe magnus nucleus were visualized via the injection of FluoroGold into the lumbosacral spinal cord (Fig. 5B). Three weeks after the injection of the AAV vectors and FluoroGold, mCherry was found to be expressed in the PVH, DA and DMH (Fig. 5C), and many spinal neurons were labeled with the tracer (Fig. 5D). mCherry-labeled axonal swellings were observed in the raphe magnus nucleus (Fig. 5E-a). Some TPH2-positive neurons in the raphe magnus nucleus were labeled with FluoroGold (Fig. 5E-b and -c), indicating that the neurons were descending neurons projecting to the lumbosacral spinal cord. Merged images revealed that the TPH2-positive descending neuron was adjacent to mCherry-labeled axonal swellings (Fig. 5E-d).

Fig. 5. The activation of the hypothalamus→raphe pathway causes psychological stress-induced defecation.

A Schematic outline of pathway-specific manipulation of hypothalamus→raphe projection neurons. B Schematic outline for (C–E). mCherry was specifically expressed in hypothalamic neurons projecting to the raphe magnus nucleus by the double virus vector infection method. The descending neurons in the raphe magnus nucleus were visualized via the injection of FluoroGold into the lumbosacral spinal cord. C Fluorescence of mCherry and immunofluorescence staining of OXT in the PVH and DA. In the inset, mCherry-expressing cell bodies are indicated by arrowheads; mCherry-expressing cell bodies expressing OXT are indicated by filled arrowheads. Scale bars: 100 μm (main) and 10 μm (inset). D Injection of FluoroGold into the spinal cord. Scale bars: 200 μm. E Confocal images showing mCherry-labeled (a) axon swellings (arrowheads) apposed to FluoroGold-labeled (c) and TPH2-positive (b) cell bodies in the raphe magnus nucleus. Scale bar: 10 μm. F Schematic outline for (G) to (N). G (S)-AMPA injection sites after intravenous administration of CNO in rats expressing mCherry and hM4Di-mCherry are indicated by circles and triangles, respectively. Regions where cell bodies are present in rats expressing mCherry (H) and hM4Di-mCherry (I). A line in the brain map encloses the region where mCherry fluorescence was observed in one individual. Effects of intravenous administration of CNO on increased colorectal motility due to glutamatergic activation of the hypothalamus in rats expressing mCherry (J) and hM4Di-mCherry (K). Representative examples obtained from different rats are shown. Summarized graph showing the numbers of propulsive contractions of the colorectum (L) and changes in MAP (M) and HR (N) after (S)-AMPA injections in rats expressing mCherry (n = 7) and hM4Di-mCherry (n = 5), respectively. *P < 0.05, **P < 0.01. Lines in the graph indicate median values. O Schematic outline for (P, Q). Summarized graph showing the number of fecal pellet outputs induced by WAS for 1 h in rats expressing mCherry (P; n = 12) or hM4Di-mCherry (Q; n = 10). The two connected dots show data for the vehicle and CNO i.p. in one individual. *P < 0.05.

Next, we examined whether the enhancement of colorectal motility induced by the injection of (S)-AMPA into the hypothalamus was suppressed by the inhibition of the hypothalamus→raphe pathway in urethane anesthetized rats (Fig. 5F). The sites at which (S)-AMPA was injected into the hypothalamus are shown in Fig. 5G. We confirmed that the double virus vector infection method successfully expressed mCherry or hM4Di-mCherry in the target neurons of the hypothalamus (Fig. 5H, I). In mCherry-expressing control rats, injection of (S)-AMPA into the hypothalamus after intravenous administration of clozapine N-oxide (CNO) increased colorectal motility and increased blood pressure and heart rate (Fig. 5J). In contrast, in hM4Di-expressing rats, CNO administration significantly inhibited the (S)-AMPA-induced increase in propulsive contractions and inhibited increases in blood pressure and heart rate (Fig. 5K). Figure 5L–N shows the quantitative data of propulsive contractions, changes in blood pressure and heart rate, respectively.

In freely behaving rats, we investigated whether stress-induced defecation was suppressed by inhibition of the hypothalamus→raphe pathway (Fig. 5O). In mCherry-expressing control rats, CNO administration did not change the number of feces excreted during the WAS test (Fig. 5P). Conversely, in hM4Di-expressing rats, CNO administration decreased the number of feces (Fig. 5Q).

Discussion

In the present study, we showed that hypothalamus→raphe projecting neurons provide excitatory input to descending serotonergic neurons, thereby causing stress-induced defecation. Our findings may provide important insights into the neural pathway underlying the brain‒gut axis as well as the pathophysiological mechanisms of stress-induced defecation disorders.

The present study revealed that neurons localized in the caudal PVH, DA and rostral DMH are involved in stress-induced defecation. Previous studies have shown that PVH and DMH neurons are activated in response to cold stress or social defeat stress^4,9,11,26. The activation of glutamatergic transmission in these hypothalamic nuclei causes increases in heart rate, blood pressure and body temperature⁹, thus indicating the critical role of the nuclei in sympathetically mediated responses to stress. As previously reported⁴, we found that WAS, which can elicit stress-induced defecation, increases c-Fos expression in the caudal PVH, DA and rostral DMH (Fig. 1). Furthermore, glutamatergic activation of these nuclei enhances colorectal motility (Fig. 2). Almost all of the tested hypothalamic sites elicited cardiovascular responses, such as increased arterial pressure and heart rate, which is consistent with the findings of previous studies^9,10,28,29. In contrast, only a limited subset of sites corresponding to the caudal PVH, DA and rostral DMH consistently induced robust increases in colorectal motility. These hypothalamic regions likely contain overlapping populations of neurons that contribute to both defecation and cardiovascular responses, whereas broader hypothalamic areas extending to the DMH mainly participate in sympathetic cardiovascular regulation.

Our findings indicate that the activation of specific hypothalamic regions enhances colorectal motility through the activation of descending serotonergic neurons in the raphe magnus nucleus, the release of serotonin into the lumbosacral spinal cord and the activation of the parasympathetic pelvic nerve. This study provided empirical evidence regarding the involvement of serotonergic neurons descending from the raphe magnus nucleus to the spinal cord in increased colorectal motility due to stimulation of the hypothalamus, as the pharmacological blockade of serotonin receptors in the lumbosacral spinal cord or pelvic nerve transection suppressed the motility responses of the colorectum (Fig. 3). Moreover, morphological findings revealed that nerve endings of hypothalamic neurons may form synapses with descending serotonergic neurons in the raphe magnus nucleus (Fig. 5). Interestingly, activation of the hypothalamus not only activates sympathetic nerves but also simultaneously facilitates parasympathetic pelvic nerve originating from the lumbosacral spinal cord. It is known that descending serotonergic neurons regulate inputs of pain signals³⁰. In fact, the intraluminal administration of capsaicin, a noxious stimulus, activates these descending neurons, resulting in enhanced colorectal motility^24,31. It is also known that acute stress suppresses pain sensitivity through activation of the descending pathway^32,33. Overall, it seems likely that stress-induced activation of descending serotonergic neurons functions primarily to suppress the input of painful signals at the level of the spinal cord and that enhanced colorectal motility is promoted as a byproduct of the primary reaction to stress. This may explain the intriguing observation that the parasympathetic pelvic nerve is activated simultaneously with sympathetic nerves during exposure to stress, leading to defecation. However, it should be noted that the pelvic nerve is not purely composed of parasympathetic fibers, as it also contains sympathetic fibers innervating pelvic organs²⁰. Nevertheless, we believe that the parasympathetic efferent component of the pelvic nerve plays a dominant role in mediating the defecatory response, due to the fact that activation of the hypothalamus facilitates (but does not inhibit) colorectal motility.

The present study revealed that the enhancement of colorectal motility in response to the activation of hypothalamic nuclei is due to the release of oxytocin into the raphe magnus nucleus. Oxytocinergic neurotransmission is known to regulate serotonergic neurons in the dorsal raphe nucleus³⁴. Here, we provide a novel finding that oxytocin also modulates serotonergic neurons projecting to the spinal cord. However, previous studies have reported that hypothalamic OXT neurons exert inhibitory effects on stress-induced activation of colonic motility in conscious rats^12–14. This apparent discrepancy may be due to the dual actions of OXT in the central nervous system. The inhibitory action of OXT seems to largely depend on the suppression of CRH neurons^12,13. Conversely, the excitatory action of OXT is exerted through direct activation of the raphe magnus, which primarily occurs for pain modulation. Aligning with this process, the activation of the oxytocin receptor in the raphe magnus nucleus has been observed to exert antinociceptive effects³⁵. Both of these actions are consistent as representing mechanisms for stress resistance; however, they exert opposite effects on colonic motility. The balance between these two mechanisms may be related to the complexity of defecation disorders during stress; specifically, diarrhea is a predominant disorder in some cases, whereas constipation is the predominant disorder in other cases. Urethane anesthesia is known to induce c-Fos activation in the brain³⁶. Therefore, differences in experimental conditions—especially under an awake condition or urethane anesthesia—may also contribute to the opposite effects.

Corticotropin-releasing hormone (CRH) acts on the PVH, leading to increased colonic motility⁶. The effects of CRH are reported to be mediated not only by the parasympathetic output of the vagus nerve^15,16 but also by that of the pelvic nerve^6,37. Furthermore, CRH neurotransmission has been demonstrated to occur locally within the PVH³⁸. Therefore, it is possible that the enhanced colorectal motility observed after glutaminergic activation of the PVH is due to the action of CRH. However, the CRH receptor antagonist had no effect on the colorectal motility induced by (S)-AMPA administration into the PVH (Fig. S3). These findings suggest that the PVH→raphe pathway, which is involved in stress-induced defecation, can be directly activated by glutaminergic transmission, with a minor (if any) contribution from CRH. Aligning with this scenario, the PVH (as well as the DMH) receives excitatory glutaminergic input from corticolimbic circuits and the amygdala, which process emotional and psychological stress signals^39,40. Similar to the regulatory circuits controlling sympathetic responses, excitatory projections from these regions are likely to form a neural circuit that causes stress-induced defecation.

Our findings demonstrate that signals from specific hypothalamic nuclei are transmitted to the colorectum via the pelvic nerve (rather than the vagus nerve). In support of this idea, several studies have demonstrated that the vagus nerve provides little innervation to the distal colon and the rectum⁴¹. However, there is also evidence for vagal innervation of the distal colon¹⁹. Therefore, it is possible that the vagus nerve serves as the neural pathway transmitting stress signals from the hypothalamus to the colorectum. A study involving pseudorabies virus (PRV) infection of the distal colon revealed that PRV-positive cells were present in caudal medullary sites, including the dorsal motor nucleus of the vagus nerve and the nucleus of the solitary tract, as well as in hypothalamic regions such as the PVH and LH^42,43. Importantly, when rats subjected to spinal cord transection were examined, PRV-positive cells in the hypothalamus were almost absent, although these cells were confirmed in the caudal medullary sites⁴². This finding is consistent with the idea that stress-related signals from the hypothalamus are transmitted to the colorectum by the descending spinal pathway. Our results do not necessarily rule out a modulatory role of the vagally mediated pathway; instead, they highlight the contribution of the hypothalamus-raphe-spinal axis.

Our findings suggest that the hypothalamus-raphe-spinal defecation center axis is related to stress-induced defecation in rats. However, it should be noted that the present study focused on acute stress conditions and did not involve chronic conditions. Hence, the mechanism by which prolonged activation of the axis affects colorectal motility remains to be determined. In particular, whether plastic changes in neural components within the axis contribute to persistent bowel dysfunction would be interesting to investigate, due to the fact that several lines of evidence indicate that the descending neurons from the brain to the lumbosacral defecation center plastically change depending on chronic changes in endogenous conditions^44–46. In addition, it should be noted that the present data were mainly obtained under urethane anesthesia. Anesthetic conditions can influence basal autonomic tone, neurotransmitter release, and stress-related neuronal activation. Some of the data, including those collected from DREADD experiments assessing stress-induced defecation, were obtained from conscious rats, thereby confirming that the hypothalamus-raphe pathway also contributes to defecatory responses under physiological conditions. Therefore, the effects of anesthesia should be considered when the present findings are compared with those obtained in conscious scenarios, as has been observed in previous studies^12–16.

In summary, this study provides evidence that the hypothalamus‒raphe‒spinal defecation center axis mediates stress-induced defecation in rats. Whether these findings are relevant to the neural basis of stress-sensitive functional bowel disorders warrants further investigation using experimental models of irritable bowel syndrome (IBS). Recent reports that stress-induced activation of the hypothalamus‒raphe‒spinal cord pathway modulates nociception³² provide support for the involvement of this axis in the pathogenesis of IBS, which is characterized by abdominal pain. Furthermore, this study provides new insights into the brain‒gut axis and suggests that the pelvic nerve serves as a dominant neural route to transmit signals from the brain to the distal colon and the rectum.

Methods

Animals

Male Sprague–Dawley rats (6–8 weeks old, 200–300 g) were purchased from Japan SLC (Shizuoka, Japan). The animals were housed in plastic cages at 22 °C with a 12 h:12 h light:dark cycle (lights on from 07:00 to 19:00) and given free access to laboratory chow (MF; Oriental Yeast) and water. Animals were monitored throughout experiments for signs of pain or distress. Humane endpoints were predefined, and animals were euthanized if severe distress was observed. The experimental procedures were performed in accordance with the guidelines for the care and use of laboratory animals approved by the Committee for Animal Research and Welfare of Gifu University (2023-174, 2023-190, 2023-193, 2024-179, 2024-180, 2024-196).

Experimental design

All experiments were conducted according to the timeline outlined below and corresponded to the respective Results section. Figure 1; after stereotaxic injection of AAV into the raphe magnus under ketamine/xylazine anesthesia, the rats were kept for ≥ 3weeks to allow expression of DREADD. The rats were then subjected to water avoidance stress (WAS) for 2 h (or control exposure for the same duration), following immediate perfusion fixation for c-Fos analysis. Figure 2; for in vivo colorectal motility recordings, rats were anesthetized with urethane, underwent colorectal and femoral artery cannulation, stabilized for 1 h, and then received hypothalamic (S)-AMPA. Further, colorectal motility and cardiovascular variables were quantified for 20 min before and after injection. Figure 3; in urethane-anesthetized recordings, serotonergic receptor antagonists (or saline) were administered intrathecally before hypothalamic (S)-AMPA, and the pelvic nerve (or sham) was transected before hypothalamic activation. Figure 4; antagonists (or saline) were injected into the raphe magnus under a condition of increased colorectal motility by prior hypothalamic injection of (S)-AMPA in urethane-anesthetized rats, and the effects were quantified over the initial 5 min after raphe injection. Anterograde AAV (hrGFP) was injected into the hypothalamus, and FluoroGold into the L6–S1 spinal cord for anatomical analysis, and tissues were collected after 1 week for immunofluorescence. Figure 5; after double AAV injections (raphe AAVretro-Cre and hypothalamic Cre-dependent hM4Di/mCherry) following the expression period of ≥3 weeks, CNO was administered intravenously before hypothalamic (S)-AMPA administration in urethane-anesthetized rats for physiological recordings. CNO (or vehicle) was intraperitoneally administered 5 min before a 1 h WAS session in freely behaving rats for defecation quantification. Animals were perfusion-fixed at the end of each recording or behavioral session to verify expression and injection sites. Animals were assigned to experimental groups using random allocation whenever applicable. For pharmacological experiments, the order of drug/vehicle administration was counterbalanced across animals.

Adeno-associated virus (AAV) vector Injection

The pAAV-CMV-EGFP-2A-Cre plasmid was constructed by using the pAAV-CMV-MCS backbone from the AAV Helper-Free System (Agilent Technologies). The EGFP-2A-Cre cassette was subcloned from the plasmid that was previously used to generate the orexin-Cre mouse line⁴⁷. The pAAV-CAG-FLEX-hM4Di-mCherry plasmid was generated by using the pAAV-CAG-FLEX backbone obtained from Addgene (plasmid #44361), and the hM4Di-mCherry cassette was subcloned from Addgene plasmid #44362. The pAAV-CAG-FLEX-mCherry plasmid was constructed in a previous study⁴⁸, and the pAAV-CMV-hrGFP plasmid was purchased from Agilent Technologies. All of the AAVs were packaged using the AAV Helper-Free system (Agilent Technologies).

To label neurons projecting to the raphe magnus nucleus in a retrograde manner, an AAV vector, AAVretro-CMV-EGFP-2A-Cre [1.6 × 10¹² gc/mL], was injected into the raphe magnus nucleus (Fig. 1A). To label the axons of neurons projecting from the hypothalamus, an AAV vector, AAV9-CMV-hrGFP [4.0 × 10¹² gc/mL], was injected into the hypothalamus (Fig. 4H). To express an inhibitory DREADD hM4Di and the fluorescent protein mCherry specifically in neurons of the hypothalamus projecting to the raphe magnus nucleus, a double virus infection method was used, in which two different AAV vectors were injected into two monosynaptically connected areas (Fig. 5A). AAV9-CAG-FLEX-hM4Di-mCherry [1.8 × 10¹² gc/mL] or AAV9-CAG-FLEX-mCherry(+ STP)-WPRE [2.5 × 10¹² gc/mL] was injected into the hypothalamus in combination with the injection of AAVretro-CMV-EGFP-2A-Cre [1.6 × 10¹² gc/mL] into the raphe magnus nucleus to selectively transduce hM4Di-mCherry or mCherry alone in neurons projecting from the hypothalamus to the raphe (hypothalamus→raphe projecting neurons). The rats were anesthetized with a mixture of ketamine hydrochloride (40 mg/kg) and xylazine (5 mg/kg) and placed in a stereotaxic instrument (SR-5R-HT; Narishige Scientific Laboratory). An injection cannula (Eicom) was connected to Teflon tubing filled with distilled water. A 1-µL Hamilton syringe was connected to the other end of the tubing. Then, a solution containing an AAV vector was aspirated from the cannula. The solution (400 nL/site) was injected into the raphe magnus nucleus (2.9 mm caudal to lambda, on the midline, 10.9 mm ventral to the bregma) or the bilateral hypothalamus (2.4 mm caudal to bregma, 0.2 mm lateral to the midline and 8.4 mm ventral to bregma) through the cannula using a manually operated syringe manipulator. Injection was performed at a constant rate of 100 nL/min (Narishige Scientific Laboratory). The injection volume was visually confirmed by the movement of the aqua oil interface along the graduated Teflon tube. The rats were used for in vivo recordings of colorectal motility and/or behavioral experiments at least three weeks after injections of the viral vectors into the brain.

Retrograde tracer injection

To retrogradely label neurons projecting to the lumbosacral spinal cord, a neural retrograde tracer, FluoroGold (4%, dissolved in saline; Fluorochrome), was injected into the spinal cord at the L6-S1 level. The rats were anesthetized with a mixture of ketamine hydrochloride (40 mg/kg) and xylazine (5 mg/kg) and the vertebral column was fixed using a pair of spinal adapters. After carefully removing the paraspinal muscles, the L1 dorsal spinous process was removed to expose the dura mater and lumbar spinal cord at L6 and S1. An injection cannula (Eicom) was connected to Teflon tubing and filled with FluoroGold. The solution (1 µL/site) was slowly ejected into the L6-S1 spinal cord parenchyma (bilateral, 0.6 mm lateral to the posterior median spinal vein, 1.0 mm ventral to the spinal cord surface) through the cannula using a manually operated syringe manipulator.

Water avoidance stress (WAS) test

The WAS was used to induce psychological stress. An observation cage (25 × 25 × 40 cm³) with a platform (cylinder; 9 cm diameter, 8 cm high) was filled with water to a height of 1 cm below the top of the platform. Prior to the experiment, both WAS and control rats were habituated to the observation cage and the water environment for several days to minimize novelty-induced stress. For the experiments examining c-Fos expression, the rats were placed on the platform for 2 h. For the control group, the rats were placed in the same cage without water during the test period, following the same handling and habituation procedures. In experiments in which hypothalamus→raphe projecting neurons were specifically inhibited by DREADD, rats expressing hM4Di-mCherry specifically in these neurons were placed on the platform for 1 h. Clozapine N-oxide (CNO; 3 mg/kg) was administered intraperitoneally 5 min before the WAS test began. For the control, rats expressing only mCherry were used. The number of fecal pellets expelled from each rat was counted until the end of the test.

Recording of colorectal motility in vivo

Colorectal motility was recorded using an in vivo experimental system modified from that previously described⁴⁹. Briefly, the rats were anesthetized with urethane (1–1.3 mg/kg intraperitoneally). A cannula was placed in the femoral artery and connected to a pressure transducer to measure changes in blood pressure. Heart rate (HR) was calculated from the obtained arterial pulse signals using LabChart (ADInstruments). Under anesthesia, a midline laparotomy was performed to expose the distal colon. The colorectum was cannulated at the distal colon and anus by using polyethylene tubing (PE-50, AS ONE, Osaka, Japan). The cannulas were secured by ligating the tubing together with the adjacent colonic wall to prevent slippage or leakage. The body wall was closed, and the distal colon cannula was protruding. The distal colon cannula was connected to a Mariotte bottle filled with saline solution that was warmed to 37 °C. Intraluminal pressure was measured by a pressure transducer connected to the anal cannula, and expelled fluid volume was measured by collecting fluid outlet via a one-way valve on a container connected to a force transducer. The baseline intraluminal pressure was maintained at 4–6 mmHg by adjusting the height of the Mariotte bottle. After surgery, the rats were maintained for 1 h to stabilize their colorectal motility and blood pressure. The invasive nature of the procedure for measuring colorectal motility in vivo itself may affect the activity of autonomic neural activation⁵⁰. Therefore, the physiological parameters were carefully monitored, and the experimental conditions were standardized to minimize this potential influence.

Drug administration

After surgery for colorectal motility measurement, a guide cannula (ID: 0.06 mm, OD: 0.15 mm; Eicom) was placed using a stereotaxic instrument (SR-5R-HT; Narishige Scientific Laboratory). The guide cannula was placed 1.5 mm dorsal to the target area, i.e., the hypothalamus or the raphe magnus nucleus. An injection cannula (Eicom) was cut to be long enough to allow the tip to protrude 1.5 mm from the tip of the guide cannula. For the injection of drugs, an injection cannula was connected to Teflon tubing and filled with either (S)-AMPA ((S)-α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid, an AMPA glutamate receptor agonist, 10 mM), L-368,899 hydrochloride (an oxytocin receptor antagonist, 10 mM), a mixture of AP5 (D(–)-2-amino-5-phosphonovaleric acid, an NMDA glutamate receptor antagonist, 10 mM) and CNQX (6-cyano-7-nitroquinoxaline-2,3-dione, an AMPA/kainate glutamate receptor antagonist, 10 mM), or dimethyl sulfoxide (DMSO, 10%) as a vehicle control. A 1-µL Hamilton syringe filled with mineral oil was connected to the other end of the Teflon tubing. The solution (100 nL/site) was slowly ejected through the cannula using a manually operated syringe manipulator. To identify the injection site, 0.2% FluoSpheres (Thermo Fisher) were injected (100 nL) at the same site with the same injection cannula after recording. Drug administration into the femoral vein and into the lumbosacral spinal cord was performed by methods described previously⁴⁹. The femoral vein was cannulated for intravenous injection of CNO (3 mg/kg). For the administration of ketanserin (100 nmol), dolasetron (100 nmol) or 0.9% saline to the lumbosacral spinal cord, a 30-gauge needle connected to a polyethylene tube was inserted between the L1 and L2 vertebrae, which correspond to the L6-S1 spinal cord, from the dorsal surface until a tail flick or hind limb twitching appeared.

Immunofluorescent histological analysis

At the end of the WAS tests or in vivo recordings of colorectal motility, transcardial perfusion fixation was performed with 4% paraformaldehyde in PB, and the brain and spinal cord of each rat were collected. After overnight fixation, the brain was soaked in 30% sucrose in PBS, frozen with O.C.T. compound medium (Sakura Finetech), and then sectioned at a thickness of 30 µm with a cryostat (CM 1850; Leica Biosystems). The cryosections were mounted on silane-coated slides immediately after sectioning and allowed to dry at room temperature for 30 min, followed by two 5 min washes with PBS. Then, the sections were blocked for 2 h with 5% normal donkey serum (Jackson ImmunoResearch) in PBS containing 0.5% Triton X-100. Following the blocking step, primary antibodies were diluted in 5% normal donkey serum and applied to the slides at 4 °C for 48 h. After 48 h, the primary antibody was removed, and the sections were washed three times (10 min each) with PBS. The appropriate secondary antibody was diluted in PBS and applied to the sections for 2 h at room temperature. After being washed three times with PBS to remove the secondary antibody, the slides were mounted with mounting medium (Prolong Glass; Thermo Fisher). The primary antibodies used were as follows: goat anti-GFP (1:200; ab6673, Abcam), rabbit anti-tryptophan hydroxylase 2 (TPH2) (1:1,000; PA1-778, Thermo Fisher), mouse anti-c-Fos (1:1000; ab208942), and rabbit anti-oxytocin (1:1000; 20068, ImmunoStar). The secondary antibodies used were as follows: Alexa 594-conjugated anti-mouse IgG (1:500; 715-585-151, Jackson ImmunoResearch), Alexa 488/594-conjugated anti-goat IgG (1:500; 705-545-147/711-585-152, Jackson ImmunoResearch), and Alexa 647-conjugated anti-rabbit IgG (1:500; A-31571, Thermo Fisher). The sections were scanned with a confocal laser scanning microscope (LSM 900; Carl Zeiss) and ZEN software (Carl Zeiss). For c-Fos quantification, every sixth coronal section (30 µm thick) was selected for staining, which encompassed the hypothalamic regions between approximately bregma −1.8 mm and −3.6 mm (corresponding to the PVH, DMH, and surrounding areas). Counting was bilaterally performed in each section, and the number of c-Fos-positive neurons was averaged across approximately 10–12 sections per animal.

Reagents

The compounds used were urethane (Tokyo Chemical Industry), ketamine hydrochloride (Daiichi Sankyo), xylazine hydrochloride (Elanco), (S)-AMPA (Sigma‒Aldrich), ketanserin tartrate (AdipoGen), dolasetron mesylate hydrate (Sigma‒Aldrich), L-368,899 (Tocris), AP5 (Sigma‒Aldrich), CNQX (Sigma‒Aldrich) and CNO (Cayman Chemical). Urethane and dolasetron were solubilized in distilled water. (S)-AMPA was dissolved in 0.9% saline (Otsuka Pharmaceutical Factory). Ketanserin, L-368,899, a mixture of AP5 and CNQX, and CNO were dissolved in DMSO (Nacalai Tesque) and diluted with saline (final concentration of DMSO was 10%).

Statistics and reproducibility

The experimental unit was a single animal for all in vivo recordings, histological analyses, and behavioral experiments. Propulsive contractions of the colorectum, mean arterial pressure (MAP) and heart rate (HR) were quantified by using data recorded for 20 min before and after (S)-AMPA injection into the hypothalamus. The effects of L-368,889 and AP5/CNQX injection into the raphe magnus nucleus on (S)-AMPA-induced colorectal motility were quantified using data collected during the initial 5 min after raphe injection. Colorectal contractions were counted when the intraluminal pressure increased by >5 mmHg above baseline with associated fluid expulsion. MAP was calculated from arterial pulse signals using LabChart (ADInstruments). Animals were excluded if post hoc histological verification showed incorrect injection/cannula placement, insufficient viral expression at the target site, or failure of spinal/vascular cannulation. These criteria were applied consistently across groups. For immunofluorescence quantification and fecal pellet counting, the assessors were blinded to the experimental conditions. Data analysis was performed using GraphPad Prism 10 (GraphPad Software). Statistical comparisons between two groups were conducted using the Mann–Whitney U test (Figs. 1G; 3E, J; 5L, M, P, Q, and S3E). Statistical analysis of propulsive contractions following the raphe injections was performed using two-way ANOVA with Bonferroni’s multiple comparisons test (Fig. 2E–G). All the statistical tests were two-sided, and P values < 0.05 were considered statistically significant. Further details on the statistical analyses are available in the supplementary statistics file.

Ethical approval

All experimental procedures conformed to the “Regulations for Animal Experiments at Gifu University” and were approved by the president of the university after review by the Committee for Animal Research and Welfare of Gifu University (2023-174, 2023-190, 2023-193, 2024-179, 2024-181, 2024-196). The regulations of Gifu University conform to the Japanese “Act on Welfare and Management of Animals” and “Standards Relating to the Care and Keeping and Reducing Pain of Laboratory Animals (Notice of the Ministry of the Environment No. 88 of 2006).” We have complied with all relevant ethical regulations for animal use.

Reporting summary

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

Author contributions

Conceptualization: N.Y., Ta.S., Y.S. Methodology: N.Y., To.S., A.M., H.Y., Y.H., Ta.S., and Y.S. Investigation: N.Y., To.S., A.M., H.Y., Y.H. Validation: N.Y., To.S., Ta.S., and Y.S. Visualization: N.Y., Ta.S., and Y.S. Funding acquisition: N.Y. and Y.S. Project administration: Ta.S. and Y.S. Supervision: Ta.S. and Y.S. Writing—original draft: N.Y. and Y.S. Writing—review & editing: N.Y., To.S., A.M., H.Y., Y.H., Ta.S., Y.S.

Peer review

Peer review information

Communications Biology thanks Yvette Taché and the other anonymous reviewer(s) for their contribution to the peer review of this work. Primary handling editor: Ophelia Bu. A peer review file is available.

Data availability

The numerical source data underlying all graphs are provided in Supplementary Data 1 and Supplementary Data 2. All other data that support the findings of this study are available from the corresponding author upon reasonable request.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s42003-026-09779-5.