A Cellulose-Rich Diet Disrupts Gut Homeostasis and Leads to Anxiety through the Gut-Brain Axis

Abstract

It is widely said that a healthy intestinal environment plays an essential role in better mental condition. One known dietary nutrient that maintains the intestinal environment is dietary fiber. A recent study showed that maintaining the intestinal environment with dietary fiber alleviated symptoms of psychiatric disorders in animals. However, such effects have only been reported with soluble fiber, which is highly fermentable and promotes short-chain fatty acid (SCFA) production, and not with insoluble fiber. Therefore, we aimed to verify whether insoluble fiber, such as cellulose, can alter emotion via changes in the gut. We divided mice into two groups and fed either a standard diet (SD, which contains both insoluble and soluble dietary fibers) or a cellulose-rich diet (CRD, which contains cellulose alone as the dietary fibers). We found that CRD-fed mice display increased anxiety-like behavior. CRD-fed animals also showed decreased intestinal SCFA levels along with increased intestinal permeability, dysmotility, and hypersensitivity. This behavioral and physiological effect of CRD has been completely abolished in vagotomized mice, indicating the direct link between intestinal environment exacerbation to the emotion through the gut-brain axis. Additionally, we found that amygdalar dopamine signaling has been modified in CRD-fed animals, and the opioid antagonist abolished this dopaminergic modification as well as CRD-induced anxiety. Altogether, our findings indicate that consumption of cellulose alone as the dietary fiber may evoke intestinal abnormalities, which fire the vagus nerve, then the opioidergic system, and amygdalar dopamine upregulation, resulting in the enhancement of anxiety.

In recent years, the number of people suffering from daily stress and anxiety, as well as mental disease, has been increasing. In 2019, WHO reported that 970 million people were suffering from mental disease worldwide, which is more than 10% global population.¹ Among the numerous diseases, mental disorders were the second most prevalent disease of Years Lived with Disability worldwide in 2019,² showing that mental illness is a long-term affliction for many people. However, the diagnosis and treatment of mental illnesses are fraught with difficulties. The diagnosis for psychiatric disorders has unresolved issues like distinguishing risk from disorder, which makes it difficult to select appropriate treatment.³ Further, the response rate of both psychotherapies and pharmacotherapies for psychiatric disorders is ≤ 50%, suggesting insufficient development of effective treatments.⁴ These issues call for elucidation of the pathophysiology and pathogenesis of psychiatric disorders and further development of therapeutic methods.

Recently, the relationship between the intestinal environment and psychiatric disorders has been attracting attention.⁵ The intestinal environment of depressed patients is known to deteriorate,⁶ whereas germ-free mice display resistance to depression.⁷ These reports strongly suggest that maintaining a healthy intestinal environment and microbiota may improve mood and emotion. Even clinical trials in depressed patients have shown that probiotic intake alleviates anxiety along with gut microbiota modification.⁸ It has also been shown that the administration of probiotics alters the GABAergic nervous system, resulting in an anxiolytic effect in naive mice.⁹

Although probiotics are the most supported “intestinal health regulators,” dietary fiber also plays an important role in maintaining a healthy intestine.¹⁰ Dietary fiber is indigestible and digested by intestinal bacteria, promoting the production of short-chain fatty acids (SCFAs). SCFAs are known to play an important role in maintaining intestinal homeostasis, such as suppressing inflammation, intestinal immunity and various diseases.¹¹ However, these beneficial effects vary widely among different types of dietary fiber. Dietary fiber is classified into two categories based on its solubility in water: insoluble and soluble fiber. We previously showed that the SCFA concentrations in the cecum contents were increased in soluble fiber (digestion resistant dextrin)-fed mice, compared with those of insoluble fiber (cellulose)-fed animals.¹² Soluble fiber is fermentable and has a stronger intestine-regulating function; therefore, it is used as a prebiotic.¹³ Several past studies suggest that soluble dietary fiber alleviates psychiatric diseases by improving the intestinal environment. Thus, improvement of the intestinal environment by soluble fiber treatment relieves symptoms of depression,¹⁴ schizophrenia,¹⁵ and diabetes-induced anxiety¹⁶ in rodents. However, the effects of insoluble fibers such as cellulose on the intestinal environment and emotions have not been examined at all.

Cellulose is the main constituent of plants and, therefore, the most common dietary fiber consumed daily. Since cellulose is less fermentable than soluble fiber yet has high water retention properties, it can increase the bulk of the stool and physically stimulate the intestines to improve peristalsis (bowel movements),¹⁷ even in chronic constipation patients.¹⁸ The high-cellulose diet was also reported to have a protective effect against dextran sodium sulfate-induced colitis by modulating lipid metabolism and gut microbiota, suggesting that cellulose is effective in maintaining intestinal homeostasis.¹⁹

We therefore aimed to verify whether insoluble fiber can alter emotion via changes in the gut. Strikingly, we found that long-term ingestion of cellulose-rich diet (CRD) led to the rise of anxiety-like behavior. Such emotional and neuronal modification might be due to intestinal hypomotility and hypersensitivity caused by SCFA decrease by CRD exposure, which was ameliorated by either vagotomy or inhibition of opioid receptors. We further found that chronic CRD exposure increases the dopamine level in the amygdala. Thus, our results suggest that CRD-induced intestinal changes evoked amygdalar dopamine signaling abnormalities via the vagus nerve and then the opioidergic system, leading to the anxiogenic effect.

Results

Long-Term Ingestion of Cellulose-Rich Diet Enhanced the Anxiety-like Behavior in Mice

We first examined whether the cellulose-rich diet will affect the behavior, particularly the anxiety-like behavior. Mice were divided into two groups and fed either a standard diet (SD: contains 2.8% of both insoluble and soluble dietary fibers) or cellulose-rich food (CRD, contains 5% cellulose). Then, we assessed the anxiety-like behavior using the open field test, elevated plus maze test, and marble burying test (Figures 1 and S1). We found that CRD-fed mice display an increase in anxiety-like behavior in the marble burying test (Figure 1C,D) without the change in locomotor activity (Figure S1). On the other hand, there were no significant differences in anxiety-like behavior in the open field test and elevated plus maze test (Figure S1). Moreover, the level of corticosterone, a typical stress hormone, tended to be elevated but was low in the urine of 16-week CRD-fed animals (Figure 1E,F). These results indicate that the long-term ingestion of CRD time-dependently enhances the anxiety level in mice.

Figure 1.

CRD-fed mice displayed the increased anxiety level and corticosterone level modification. (A,B) Overall scheme of the experiments. Mice were divided into two groups and were fed either a standard diet (SD) or a cellulose-rich diet (CRD). Behavioral tests were done at the 2–4th, 6–8th, and 16–18th week of the feeding period with a one-week intervals for each testing. Individual groups of animals are prepared and used for other testing and sample collection in the 8th and 16th weeks, respectively. Serum, feces, and urine for corticosterone measurements were sampled from mice exposed to CRD for 16 weeks. Mice were housed in the metabolic cage for three days, and then feces and urine samples were collected 24 h on the third day (see panel B). (C) CRD exposure transiently increased the number of buried marbles on the marble burying test. (D) Example of marble burying was observed in the 16th week. CRD animals buried more marbles than SD groups. (E,F) Corticosterone concentrations in serum and urine. While CRD tendeds to increase serum corticosterone, it significantly decreased urine corticosterone. Data are expressed as means ± SEM of 10–31 mice/group for behavioral tests and 5–9 mice/group for corticosterone measurement. Statistical significance is indicated by *p < 0.05 according to unpaired t test or Mann–Whitney test.

CRD Causes Worsening of the Intestinal Environment, Including Increased Intestinal Permeability and Dysmotility

Because chronic CRD exposure did not reverse but rather exacerbated the anxiety, we hypothesized that chronic CRD would also worsen the gut environment. Hence, we evaluated what intestinal modification CRD may cause. The first sign of intestinal modification we found in CRD-fed animals was a significant increase in cecum pH (Figure 2B). This could be explained by the further measurement of intestinal short chain fatty acid (SCFA) and lactic acid contents, which has been remarkably decreased by CRD consumption (Figure 2C,D). These results suggest that long-term CRD consumption may cause a decrease in SCFA production in the intestine, which may lead to higher intestinal pH, one of the hallmarks of a worsened intestinal environment.¹¹

Figure 2.

CRD consumption evoked the remarkable deterioration of the intestinal environment. (A) Overall scheme of the experiments. Individual groups of animals were prepared and used for other tests and sample collection in the 8th and 16th week, respectively. (B–D) Increase in cecum pH (B) which may be caused by a significant decrease of SCFA and lactic acid in cecum content (C,D), has been observed in animals that consumed CRD for 8 weeks or 16 weeks. (E) Intestinal permeability was tested by oral administration of FITC-dextran at the 16th week of ingestion. The serum FITC-dextran concentration appeared higher in CRD animals, indicating that chronic CRD may disrupt the intestinal barrier. (F,G) Intestinal transit ratio in the charcoal meal test (F), together with the defecation frequency (the number of fecal pieces in 24 h) and total stool weight (G), represent intestinal dysmotility in CRD-fed animals. (H) Length of the large intestine did not differ between SD- and CRD-fed animals. Data are expressed as means ± SEM of 5–11 mice/group. Statistical significance is indicated by *p < 0.05, **p < 0.01, ***p < 0.005, ****p < 0.001 according to unpaired t test or Mann–Whitney test.

One of the most known roles of intestinal SCFA is the maintenance of intestinal barrier and motility.^20,21 We thus focused on the physiological changes in the intestinal tract and evaluated the intestinal permeability and motility. The FITC-dextran test revealed that chronic CRD consumption would induce intestinal hyperpermeability (Figure 2E). CRD consumption also induced intestinal dysmotility, resulting in a decrease in defecation frequently and stool weight without the change in body weight and food intake (Figures 2F,G, and S2). However, chronic CRD had no significant effect on the expression of genes related to the intestinal barrier (Figure S3). In addition, we confirmed the presence of intestinal inflammation by colorectal shortening and cytokine expression. We found no significant differences in the length of the large intestine (Figure 2H) and the mRNA expression of cytokines (Figure S4). Although no severe molecular difference or inflammation has been observed, these results indicate that CRD may induce the deterioration of the intestinal environment, including dysbiosis (decreased SCFA), dysmotility of the intestine, and intestinal barrier disruption.

CRD Causes Intestinal Hypersensitivity by Upregulation of TRPA1

We then assessed the intestinal responsiveness to noxious stimuli to investigate the possible effects of the biochemical and physiological changes in CRD-fed mice. Notably, CRD-fed mice displayed clear hypersensitivity to nociceptive stimuli by allyl Isothiocyanate (AITC, Figure 3A) and capsaicin (Figure 3B). We further found upregulation of intestinal transient receptor potential ankyrin 1 (TRPA1) (Figure 3C) and sodium glucose co-transporter1 (SGLT1) in CRD-fed group (Figure 3D), while TRP vanilloid 1 (TRPV1) expression had no change (Figure S5). AITC is the activator of TRPA1, and SGLT1 is known to fire the vagus nerve together with TRPA1 and TRPV1.²² These results, therefore, indicate that CRD consumption leads to the upregulation of TRPA1 and SGLT1, which may enhance the responsiveness to noxious stimuli, resulting in intestinal hypersensitivity and further vagus nerve firing.

Figure 3.

Chronic CRD exposure evoked intestinal hypersensitivity, possibly through increased TRPA1. (A,B) Hypersensitivity was examined using intestinal stimulation by rectally administered AITC (A) and capsaicin (B) during the 16th week of the ingestion period. Licking of the abdomen and squashing of the lower abdomen against the floor was taken as the pain-related behavioral score. CRD-exposed animals displayed increased hypersensitivity in both tests. (C,D) mRNA expression level of TRPA1 (C) and SGLT1 (D) was already increased in the intestine tissues of mice after 8 weeks of CRD exposure. Data are expressed as means ± SEM of 5–11 mice/group. Statistical significance is indicated by *p < 0.05, **p < 0.01 according to two-way ANOVA with Uncorrected Fisher’s LSD or with Sidak’s multiple comparisons test, unpaired t test, or Mann–Whitney test.

Vagal Signaling Mediates CRD-Induced Anxiety

We could find that chronic CRD will exacerbate both anxiety-like symptoms and the intestinal environment. Then, the next question will be whether CRD directly or indirectly modulated the brain to have an anxiogenic effect. Since the gut-to-brain transmission mediated by the vagus nerve has been suggested to modulate anxiety,²³ we conducted vagotomy, the transection of the hepatic vagal branch (Figure 4A). If CRD-evoked anxiety is diminished by vagotomy, it demonstrates that CRD does not directly modulate the anxiety-related brain function but rather indirectly modifies it, possibly through the gut-brain axis. Consequently, vagotomized animals did not display any sign of CRD-induced anxiety (Figure 4B) with no change in motor ability (Figure 4C). Importantly, vagotomized CRD-mice that displayed improvement in anxiety also showed a significant increase in cecum pH, which may be led by decreased SCFAs (Figure S6).

Figure 4.

CRD-induced anxiety has been suppressed by vagotomy. (A) Overall experimental scheme for the vagotomy. At least 4-weeks of recovery time was taken after the surgery. Then mice were fed with a standard diet (SD) or cellulose-rich food (CRD) for another 8 weeks, and behavioral tests were performed and sacrificed to collect the tissue. (B,C) CRD-evoked anxiety-like behavior in the marble burying test was suppressed by vagotomy (B) with no effect on locomotive activity (C). Data are expressed as means ± SEM of 7–8 mice/group. Statistical significance is indicated by *p < 0.05, **p < 0.01 according to one-way ANOVA with uncorrected Fisher’s LSD.

These results show that the gut-brain axis, composed of the vagus nerve, is essential for the CRD-induced enhancement of anxiety-like behavior. It also indicates that the anxiogenic effect of CRD occurs through the modification of the intestinal environment rather than direct modulation of the brain.

CRD-Induced Intestinal Hypersensitivity may Enhance Anxiety and the Amygdala through Opioidergic Systems

A recent study showed that gut-to-brain vagal transmission would activate the endogenous opioidergic systems in the brain.^24,25 It is also known that noxious stimuli or chronic hypersensitivity will activate opioidergic systems by the tonic release of endogenous opioids such as enkephalin.²⁶ Thus, we hypothesized that intestinal hypersensitivity evoked by chronic CRD may also cause the activation of the opioidergic systems, leading to the modification of brain function. To determine the contribution of opioid receptors, we injected the opioid receptor antagonist naloxone into mice and examined whether the blockage of opioid signaling could inhibit CRD-induced anxiety. As expected, naloxone injection suppressed the CRD-induced anxiety (Figure 5B) with no effect on the locomotor activity (Figure 5C). These results suggest that CRD consumption evoked intestinal hypersensitivity to excessive vagus nerve firing, which may activate the opioidergic system, which may lead to enhanced anxiety.

Figure 5.

An opioid receptor antagonist has suppressed CRD-induced anxiety. (A) Overall scheme of the experiments. (B,C) Effect of opioidergic system inhibition on anxiety-like behavior (B) and locomotive activity (C) was assessed in animals after 16 weeks of exposure to SD or CRD. Anxiety-like behavior was measured by a marble burying test. The locomotive activity was assessed by an open field test immediately after the marble burying test. Pretreatment of naloxone (1 mg kg^–1) could suppress the CRD-evoked anxiety with no effect on the locomotive activity. Data are expressed as means ± SEM of 5 mice/group. Statistical significance is indicated by ***p < 0.005 according to two-way ANOVA with Sidak’s multiple comparisons test.

CRD Alters the Dopaminergic System in the Amygdala, Which is Reversed by Vagotomy

To understand which molecular mechanisms underlie CRD-induced anxiety, we measured the monoamine level in various brain regions that are known to be involved in the regulation and expression of anxiety. We found that 16-week CRD-exposure significantly increased the dopamine level in amygdala (Figure 6B), while no significant differences were seen in other brain regions or for other monoamines (Figures 6A and S7A,B). We then measured the mRNA expression levels of dopamine receptors and transporters in the amygdala. Interestingly, CRD-fed mice displayed a significantly lower mRNA expression of dopamine D2 receptors (D2R) compared to the SD-fed mice in the eighth week of the ingestion period (Figure 6C), whereas there were no significant differences in the 16th week (Figure 6D). Notably, vagotomy reversed the CRD-induced decrease in D2R expression (Figure 6E), suggesting that vagal signaling is involved in changes in the amygdalar dopaminergic system. However, vagotomy had no significant effect on CRD-induced increases in amygdalar dopamine levels (Figure 6F).

Figure 6.

Chronic CRD modified dopamine signaling in amygdala. (A,B) Dopamine (DA), noradrenaline (NE), and serotonin (5-HT) level in the amygdala was measured by HPLC and compared between SD- and CRD-fed mice at a feeding period of 8 weeks (A) and 16 weeks (B). (C,D) The amygdalar mRNA expression level of dopamine receptors (D1R and D2R), as well as dopamine transporter (DAT), was determined in the 8th (C) and 16th (D) week after the exposure to either SD or CRD. (E) Vagotomy had no significant effect on the amygdalar dopamine level. (F) Decrease of amygdalar D2R mRNA expression level in CRD animals has been reversed by vagotomy. Data are expressed as means ± SEM of 5–10 mice/group. Statistical significance is indicated by *p < 0.05, **p < 0.01 according to one-way ANOVA with Uncorrected Fisher’s LSD, unpaired t test, or Mann–Whitney test.

As recent studies also suggest that neuroinflammation might contribute to emotional change,²⁷ we also focused on amygdalar inflammation. To evaluate the inflammation, we measured the mRNA expression levels of pro- and anti-inflammatory cytokines in the amygdala. Surprisingly, the mRNA expression levels of amygdalar TNF-α and IL-6 were decreased in the CRD-fed group (Figure S8A). Flow cytometric analysis of bone marrow also showed a decrease in the monocyte population (Figure S8B), which may suggest that CRD exposure decreased monocyte production in the bone marrow to decrease the inflammatory cytokine levels in the amygdala.

Discussion

In the present study, long-term ingestion of CRD induced chemical and physiological intestinal changes, such as decreased SCFA, intestinal hypersensitivity, and increased intestinal permeability. We also found that chronic CRD modulated the opioidergic system via the vagus nerve and then evoked anxiety with amygdalar dopaminergic abnormality.

We could observe CRD-induced anxiety, particularly in the marble burying test, but not in the open field test and the elevated plus-maze test. This could be explained by the neuronal mechanisms underlying respective behavioral tests used to assess anxiety. While repeated administration of serotonin reuptake inhibitors can suppress marble-burying behavior in mice,²⁸ it is ineffective for anxiety-like behaviors in the open field test and the elevated plus maze test.²⁹ The marble-burying behavior is also considered a model for obsessive-compulsive disorder (OCD). However, it is unlikely that CRD-fed mice showed OCD-like behavior. First, CRD consumption did not affect hippocampal serotonin levels, which is considered to be the etiology of OCD. Second, naloxone administration suppressed CRD-induced anxiety, while it is reported to worsen OCD symptoms in patients.³⁰ Altogether, CRD-induced behavioral modification that is found only in the marble-burying test does not fit either general anxiety symptoms or OCD symptoms. More detailed assessments are therefore required to identify which specific symptom of human pathological situation this CRD-induced behavioral modification will be.

We also found serum corticosterone concentration tended to be elevated, while the urinary cortisol concentration was significantly decreased in anxiety-evoking CRD-fed mice. We hypothesize that different corticosterone responses between serum and urine are due to which kind of stresses are reflected. Serum corticosterone often reflects acute stress, including rapid stress during blood collection.³¹ On the other hand, urinary corticosterone reflects the chronic stress. It has been reported that repeated surges of cortisol by chronic stress lead to cortisol dysfunction, like depletion of cortisol or hypersensitivity of the negative feedback system.³² Such cortisol decrease by HPA axis negative feedback has been reported in patients with chronic pain such as lumbago³³ or myogenous facial pain.³⁴ Based on these findings, chronic pain or hypersensitivity may enhance the negative feedback of the HPA axis and suppress corticosterone secretion.³⁵ Therefore, the decreased urinary corticosterone levels may indicate increased chronic stress, such as gut hypersensitivity, which would evoke dysfunction of the HPA-axis in CRD-fed mice. We also expect that this corticosterone reduction might cause CRD-induced intestinal dysmotility, as corticosterone secretion has been suggested to cause increased peristalsis.³⁶

As the possible molecular modulator of CRD-evoked anxiety, we propose the alteration of dopamine signaling in the amygdala that is observed in CRD animals. The amygdala is known for its regulatory role for anxiety and fear, which could be demonstrated by excessive activation of the amygdala in patients with psychiatric disorders with symptoms of excessive anxiety, such as depression, anxiety disorders, and PTSD.³⁷ Notably, microinjection of D1R and D2R agonists or antagonists into the amygdala produces anxiogenic or anxiolytic effects, respectively.³⁸ Moreover, chronic dopamine agonist exposure decreases D2R levels in dopaminergic neurons.³⁹ In the study, we found that no other monoamines but the release of DA and D2R mRNA expression in the amygdala are modified in CRD animals. We therefore hypothesize that modification of the amygdalar dopaminergic system by CRD exposure might be one of the causes of increased anxiety-like behavior. However, as dopaminergic activity was not consistent between weeks 8 and 16 of the dietary intervention, a more detailed study is required to clarify the involvement of modified amygdalar dopaminergic signaling to behavioral modification under the chronic CRD state.

In addition to brain monoamine, a recent study suggested that brain inflammation may also be involved in emotion. Thus, the increase of inflammatory cytokines IL-6 and TNF-α in the amygdala is associated with the induction of anxiety-like behavior,²⁷ and suppression of amygdalar TNF-α ameliorates anxiety-like behavior.⁴⁰ We therefore confirmed whether anxiety-evoking chronic CRD exposure caused the inflammation in the brain. Surprisingly, CRD decreased the mRNA expression of IL-6 and TNF-α in the amygdala. Interestingly, CRD-fed mice displayed a decrease of CD11b positive monocyte population in the bone marrow and spleen. Monocytes have been reported to circulate through blood vessels and release TNF-α in the brain, involving the synaptic turnover.⁴¹ Thus, reduced amygdalar TNF-α may be due to the decreased monocyte level in the bone marrow and spleen by long-term CRD exposure. Further, we conclude that behavioral modification by chronic CRD is not due to brain inflammation.

Our study clarified that the worsening of intestinal environment as the major cause of neuronal change and following enhanced anxiety. Long-term CRD evoked decreased SCFA concentrations which caused increase of cecum pH, reflecting the deterioration of intestinal environment. CRD consumption also increased intestinal permeability along with decrease of intestinal motility and defecation frequency, suggesting the global exacerbation of intestinal condition. It is known that SCFA promotes serotonin production in the intestinal tract to maintain normal peristalsis.⁴² Butyrate, one of the SCFAs, increases ChAT immunoreactive neurons in intestinal tissue, enhancing colonic contractions due to cholinergic innervation.²¹ SCFAs also enhance tight junctions and maintain the intestinal barrier,²⁰ which is often lost in several psychiatric diseases.⁴³ Interestingly, vagotomized mice showed no anxiety or amygdalar dopamine increase, despite of gut condition aggravation like decreased SCFA as well as increased cecum pH. This result also indicates that CRD itself does not have any direct anxiogenic effect; it will affect the central nervous system through activation of the vagus nerve. All in all, SCFA downregulation by CRD exposure may cause adverse physiological effects on the gut, e.g., increased intestinal permeability and dysmotility, leading to negative psychiatric effects through the gut-brain axis mediated by the vagus nerve.

Another interesting finding is that CRD-exposed animals displayed intestinal hypersensitivity to the noxious stimuli given by capsaicin or AITC. We also found a particular increase in TRPA1, one of the most known pain-related cation channels, which explains the revelation of intestinal hypersensitivity after chronic CRD exposure. In general, TRP family proteins are receptors that respond to temperature, pH, osmotic pressure, chemical stimuli, and mechanical stimuli in peripheral nerves and transmit these stimuli to the central nervous system by activating nociceptors.⁴⁴ TRPA1 has been shown to respond strongly to mechanical and chemical stimuli such as AITC. TRPA1 is also characteristically expressed in enterochromaffin cells (ECs) which activate the vagus nerve via serotonin secretion.⁴⁵ These findings suggest that CRD-induced increased expression of TRPA1 induces intestinal hypersensitivity, evoking excessive vagal signaling.

Interestingly, we also found increased capsaicin-induced pain-related behavior in CRD-fed mice even though there is no change in TRPV1 expression. This can also be explained by the enhancement of TRPA1 since TRPA1 is known to coexpress with TRPV1 and cross-talk each other. For example, TRPA1 activation by mustard oil sensitizes TRPV1 and enhances the capsaicin-induced TRPV1 activity.⁴⁶ TRPV1/TRPA1 double-positive sensory neurons show drastically increased TRPV1-mediated currents induced by capsaicin by glutamate exposure.⁴⁷ Further, the facilitation of TRPV1-mediated currents in TRPA1-positive dorsal root ganglia neurons produces heat hyperalgesia in mice.⁴⁷ These results suggest that increasing TRPA1 expression can possibly enhance capsaicin-evoked TRPV1 activity including hypersensitivity, just as we have observed in CRD animals.

There are three possible reasons for the increased expression of TRPA1. First, it is suggested that TRPA1 activation promotes serotonin production in ECs, which in turn stimulates peristalsis.⁴⁵ As CRD exposure caused decreased peristalsis, TRPA1 might be upregulated to restore the peristalsis to the normal state. Second, it is known that inflammation causes the increase of TRPA1 and number of TRPV1/TRPA1-responsive neurons in a mouse model of caerulein-induced pancreatitis,⁴⁸ suggesting that inflammation regulates TRPA1 expression and activity. As ileum IL-1β tended to increase by CRD exposure, that slight enhancement of inflammation may have triggered the increase of TRPA1. Finally, past study showed the contribution of reactive oxygen species to TRPA1 upregulation through nuclear factor-erythroid 2-related factor 2 (NRF2) activation,⁴⁹ which could be decreased by SCFA.⁵⁰ Therefore, the decrease in SCFA concentrations due to CRD consumption may increase oxidative stress, which may upregulate the TRPA1 expression by NRF2 activity. As CRD consumption also upregulated SGLT1, which is reported to be increased in H₂O₂-treated ECs,⁵¹ we expect that the contribution of oxidative stress might be the most plausible theory.

Past reports demonstrated that TRPA1 and SGLT1 might stimulate vagus signaling,^22,48 which may link to the dopaminergic system in the brain. The vagus nerve is known to project to nucleus tractus solitarii, and is suggested to affect amygdalar activity indirectly via other brain regions, such as ventral tegmental area (VTA) and locus ceruleus.⁵² Recent studies showed that vagus signaling first activate enkephalin-mediated endogenous opioidergic systems,^24,25 then the dopamine neuron in VTA,⁵³ which of both has been inhibited by vagotomy. In the present study, we could confirm the contribution of the endogenous opioid system to CRD-evoked anxiety-like behavior by acute naloxone administration. The endogenous opioid system indirectly promotes dopaminergic activation at the amygdala through inhibition of the GABAergic system in the VTA.⁵⁴ Further, as part of the gut-brain axis, it is said that the vagal afferent projected to NTS can activate dopaminergic neurons in VTA through the noradrenergic pathway.⁵⁵ We therefore propose the following hypothetic scheme: CRD-induced TRPA1 upregulation first causes increased transmission of nociceptive stimuli to the brain through the vagus nerve, which may activate the endogenous opioid system, which leads to activation of dopaminergic neurons at VTA to increase dopamine release at the amygdala in CRD-fed mice (Figure 7). However, our result still do not explain why “specifically” amygdala will be chosen to increase dopamine, since the dopamine level in the nucleus accumbens (the major projection region of dopamine neurons from VTA) was not increased in CRD-fed mice (Figure S7C). To clarify such specific regulation of dopaminergic signaling (possibly by CRD) will be the future outlook.

Figure 7.

Possible mechanism of CRD-induced anxiety. Chronic consumption of cellulose-rich food (CRD) will decrease intestinal SCFAs. It caused gut disability, such as reduced motility, increased intestinal permeability, and upregulation of TRPA1 and SGLT1. These physiological modifications resulted in intestinal hypersensitivity, possibly overstimulating the vagal transmission. Such vagal transmission from the gut projects to the nucleus tractus solitarii (NTS) to activate endogenous opioidergic systems.²⁵ The activation of the opioidergic system may suppress the GABAergic neuron in the ventral tegmental area (VTA) to increase the dopamine levels in the amygdala (Amyg), which possibly lead to the characteristic anxiety. The figure was created with BioRender.com.

We also can not exclude the possible contribution of intestinal microbiota to the present findings. A past study showed that chronic consumption of AIN-93G modified gut microbiota in mice.⁵⁶ Thus, AIN-93G increased the presence of Allobaculum to be the major genus (43.4%) of mice. This result suggests two important aspects: (1) AIN-93G leads to a significant loss of microbiome diversity, and (2) AIN-93G increases particular microbiota Allobaculum, which is frequently observed as the characteristic of aged mice. Since CRD (AIN-93M) has a similar composition as AIN-93G, it is possible that our model exhibits similar changes in the gut microbiota and affects the gut-brain axis to modify emotion. Altogether, other molecular mechanisms or pathways may exist to lead to CRD-evoked anxiety, which requires further detailed examination.

In summary, our novel findings suggest that CRD consumption causes (1) intestinal deterioration and hypersensitivity, which leads to (2) the overactivation of the opioidergic system mediated by the vagus nerve, possibly resulting in (3) enhanced anxiety and amygdalar dopaminergic abnormality. A number of recent evidence support that dietary fiber consumption benefits health by reducing the risk of various diseases such as metabolic disorders, cardiovascular diseases, and colorectal cancer.⁵⁶ Although the proportions of each nutrient are different in MF and AIN-93 M (Tables 1, 2, 3), there is no doubt that dietary fiber has the most significant effect on the intestine.¹⁷ The present study shows for the first time that that particular type of dietary fiber, in accurate the insoluble dietary fiber, may cause emotional problems by aggravating the gut environment. Our study confirms that diet affects the risk of developing psychiatric disorders, suggesting the importance of further detailed examination of diets to clarify, e.g., whether particular food/nutrition has anxiogenic or anxiolytic effects, which in the end might enable the prevention and treatment of psychiatric disorders through dietary intervention.

Table 1. Nutritional Composition of the Diets (%)^a.

nutrient	MF (standard diet)		AIN-93 M (cellulose rich food)
protein	crude protein [g]	23.2	casein [g]	14.0000
			L-cysteine [g]	0.1800
carbohydrate	crude carbohydrate [g]	54.7	corn starch [g]	46.5692
			α-corn starch [g]	15.5000
			sucrose [g]	10.0000
fat	crude fat [g]	4.9	soybean oil [g]	4.0000
dietary fiber	crude dietary fiber [g]	3.3	cellulose [g]	5.0000
			other variable fibers [g]	0
mineral	crud ash [g]	5.9	AIN93 mineral Mix [g]	3.5000
vitamin			AIN93 Vitamin Mix [g]	1.0000
other	water [g]	8.1	choline bitartrate [g]	0.2500
			t-butylhydroquinone [g]	0.0008

^aCreated with reference to the website of Oriental Yeast Co., Ltd. and CLEA Japan, Inc. (MF: https://www.oyc.co.jp/bio/LAD-equipment/LAD/ingredient.html, AIN-93M: https://www.clea-japan.com/company/outline.html) *Detailed composition is described in Table 2. ** Detailed composition is described in Table 3.

Table 2. Detailed Nutrient Composition of MF (Amount Contained in 100 g)^a.

mineral		vitamin		amino acid
calcium [g]	1.04	vitamin A* [IU]	1638	isoleucine [g]	0.92
phosphorus [g]	0.81	vitamin D3 [IU]	111	leucine [g]	1.77
magnesium [g]	0.24	vitamin E [mg]	8.9	lysine [g]	1.27
natrium [g]	0.21	vitamin K3** [mg]	0.04	methionine [g]	0.43
kalium [g]	0.99	vitamin B1 [mg]	1.92	cystine [g]	0.36
iron [mg]	11.1	vitamin B2 [mg]	1.01	phenylalanine [g]	1.06
copper [mg]	0.74	vitamin C [mg]	5	tyrosine [g]	0.74
zinc [mg]	5	vitamin B6 [mg]	0.87	threonine [g]	0.89
manganese [mg]	5.2	vitamin B12 [μg]	4.6	tryptophan [g]	0.3
		inositol [mg]	467.3	valine [g]	1.11
		biotin [μg]	30.1	arginine[g]	1.47
		pantothenic acid [mg]	2.14	histidine [g]	0.62
		niacin [mg]	10.22	alanine [g]	1.19
		choline [g]	0.18	aspartic acid [g]	2.12
		folic acid [mg]	0.16	glutamic acid [g]	3.94
				glycine [g]	1.15
				proline [g]	1.28
				serine [g]	1.1

^aCreated with reference to the website of Oriental Yeast Co. (https://www.oyc.co.jp/bio/LAD-equipment/LAD/ingredient.html).

Table 3. Detailed Nutrient Composition of AIN-93M [%]^a.

AIN93 mineral mix		AIN93 vitamin mix
CaCO3	35.7000	sucrose	97.3474
sucrose	20.9782	vitamin E (50%)	1.5000
KH2PO4	25.0000	nicotinic acid	0.3000
NaCl	7.4000	D-Pantothenic acid Ca	0.1600
K3C6H507·H2O	4.6600	D-Biotin (100%)	0.0020
K2SO4	2.8000	vitamin B2 (Over 98%)	0.0600
MgO	2.4000	vitamin B6	0.0700
FeC6H5O7·XH2O	0.6060	vitamin B1	0.0600
ZnCO3	0.1650	vitamin A (325,000 IU/g)	0.1231
Na2SiO3·9H2O	0.1450	vitamin D3 (100,000 IU/g)	0.1000
MnCO3	0.0630	folic acid	0.0200
CuCO3Cu (OH)2	0.0324	vitamin B12 (0.1%)	0.2500
CrK (SO4)2·12H2O	0.0275	vitamin K1	0.0075
H3BO3	0.0082
NaF	0.0064
NiCO3·2Ni (OH)2·4H2O	0.0032
LiCl	0.0017
Na2SeO4	0.0010
KIO3	0.0010
(NH4)6Mo7O24·4H2O	0.0008
NaVO3	0.0007

^aCreated with reference to the website of CLEA Japan, Inc. (https://www.clea-japan.com/company/outline.html).

Methods and Materials

Animals

Male ICR mice (Tokyo Laboratory Animals, Japan) aging 8-weeks-old at the beginning of the experiments were used. All mice were kept in a room maintained at 22 ± 2 °C, with humidity of 60 ± 5%, with a 12-h light (100–150 lx): 12-h dark cycle (light on between 8:00 and 20:00). Food and water were available ad libitum except during behavioral observations. A total of two cohorts are prepared: MF (Oriental Yeast Co., Ltd., Tokyo, Japan)-fed standard diet (SD) group and AIN-93 M (CLEA Japan, Inc., Tokyo, Japan)-fed cellulose-rich food (CRD) group. We defined AIN-93 M as a “high cellulose diet” because AIN-93 M contains only 5% cellulose as dietary fiber, while MF contains 3.3% crude fiber, both soluble and insoluble. Note that total calories of SD and CRD are the same (3.58 kcal/gram). Tables 1–3 show their detailed compositions. The duration of the dietary intervention was up to 18 weeks, with major behavioral tests and dissection at weeks 8 and 16. Experiments were conducted in accordance with permission from the Committee for Animal Experimentation of the School of Science and Engineering at Waseda University (permission #A23-093), and in accordance with the Law (No. 105) and Notification (No. 6) of the Japanese Government.

Drugs

Most of the reagents and drugs have been purchased from Fujifilm Wako Pure Chemicals, Osaka, Japan, and antibodies have been purchased from Beckton Dickinson, NJ, USA, otherwise stated in the methods.

Behavioral Tests

Behavioral tests to assess anxiety (marble burying test, open field test, elevated plus maze) and intestinal hypersensitivity (AITC- or Capsaicin-induced hypersensitivity) were conducted. The respective tests were done at the 2–4th, 6–8th, and 16–18th weeks of the feeding period, with a one-week interval for each test. All experiments were conducted from 17:00–20:00 due to avoid changes in behavior due to the time of day, and data was assessed by automated data collection with software or individual observer in a blind manner.

Marble Burying Test

The marble burying test was conducted with slight modifications to previous studies⁵⁷ at the 4, 8, and 16th week of the feeding period. Individual mice were placed in a cage (24 × 37 × 17 cm) containing 20 marbles of various colors with a diameter of 1.5 cm arranged in a 4 × 5 grid, each on top of 5 cm of white softwood bedding, then left undisturbed for 15 min. A marble with more than two-thirds of its height covered with bedding was considered buried marble. Digital photographs were obtained at uniform angles and distances for each test cage, and an independent observer counted buried marbles.

Open Field Test

Open field test was conducted at 2, 6, and 17th weeks of the feeding period. The mouse was individually placed in the corner of a cage (24 × 37 × 17 cm) and allowed to move freely for 5 min. The locus of its activity was recorded by video camera and analyzed using the video tracking software ANY-maze (Stoelting, USA). The number of center-zone entry, center-zone staying time, traveled distance was measured and assessed.

Elevated Plus Maze Test

Elevated plus maze test was conducted in 3, 7, and 18th weeks of the feeding period. The plus maze consisted of two 29 × 7 cm opened arms, two 29 × 7 cm closed arms, and a 7 cm square center zone. The mouse was placed at its center and allowed to move freely for 5 min. Its activity locus was recorded by video camera and analyzed using the video tracking software ANY-maze (Stoelting, USA). The staying time in the open and closed arm, the distance traveled in the open and closed arm, and the number of entries in the open and closed arm were measured and assessed.

AITC-Induced Hypersensitivity

Allyl isothiocyanate (AITC)-induced hypersensitivity was assessed using the methods described previously.⁵⁸ Mice were habituated in the 15 cm × 10 cm × 8 cm observation cage for 20 min. AITC (2% vol vol^–1 in rape seed oil; Nacalai Tesque Inc., Kyoto, Japan) was administered in 0.05 mL intrarectally, and their behavior was then observed for 15 min. Pain-related behaviors were determined as follows: licking of the abdomen, stretching the abdomen, squashing of the lower abdomen against the floor, and immobility. Behaviors were assessed in 5-s intervals, and scores were calculated as follows: 2 points for pain-related behavior for at least 3 out of 5 s, 1 point for less than 3 out of 5 s, and 0 points for never performing the behavior during the 5 s. Immobility was measured from the point at which it stopped for more than 5 s.

Capsaicin-Induced Hypersensitivity

Capsaicin-induced hypersensitivity was assessed using the methods described previously.⁵⁹ Mice were habituated in the 15 cm × 10 cm × 8 cm observation cage for 20 min. Capsaicin (0.1 wt % vol-1 in 4% Tween80/0.9% saline) was administered in 0.1 mL intrarectally, and their behavior was then observed for 15 min. Pain-related behaviors were assessed and scored as described above in the AITC-induced hypersensitivity test.

Naloxone Administration

Naloxone (1 mg kg^–1) was dissolved in saline and injected i.p. into mice 20 min prior to the marble burying test. The test was conducted as a crossover trial with a 4-day recovery period. The effect of naloxone administration on locomotor activity was assessed by an open-field test immediately after the marble burying test.

Tissue Collection

Tissue collection has been done 3–5 days after the final marble burying tests. Briefly, mice were euthanized by rapid decapitation, and then the brain, spleen, bone, and whole gut (between pylorus to rectum) were quickly harvested on ice. The brain was separated into the prefrontal cortex (PFC), amygdala, striatum, nucleus accumbens, and hippocampus for further use in monoamine measurement and RT-PCR. Spleen and bone were processed for flow cytometry. The gut has been separated into cecum, jejunum, ileum, and colon for further use on short-chain fatty acid measurement and RT-PCR.

Brain Monoamine Levels Measurement

Brain monoamine levels were measured by high-performance liquid chromatography (HPLC) as previously reported.⁶⁰ Samples were first processed with 0.2 M perchloric acid solution (containing 100 M EDTA-2Na) together with isoproterenol (Sigma-Aldrich, USA) as an internal standard, followed by ultrasonic homogenizing and centrifugation at 15,000 rpm for 10 min to collect supernatant containing monoamines. Samples and the standards were then purified through a 0.45 μm filter (EMD Millipore, USA) and injected into an HPLC coupled with an electrochemical detector (HTEC-510, Eicom Co., Kyoto, Japan). Monoamine levels of samples were measured based on the measurement result of the standard. The mobile phase was 0.1 M acetate-citrate buffer (containing 5 mg L^–1 EDTA·2 Na, 190 mg L^–1 1-octanesulfonic acid sodium salt, and 15% methanol). The flow rate was 500 μL min^–1, the applied column temperature was 25 °C, and the voltage was + 750 mV for Ag/AgCl, respectively. EPC-300 software (version 2.5.10, Eicom) was used for data analysis.

RNA Extraction

mRNA expression levels were measured by real-time RT-PCR, using the protocol described in the previous study.⁴⁰ The brain tissue was collected directly in a tube containing 500 μL of Trizol (Ambion, USA) and then homogenized using Tissuelyser II (Qiagen, Germany). Then 500 μL of chloroform was added and centrifuged at 15,000 rpm for 10 min to collect supernatant. The one-third supernatant volume of CIA solution (chloroform: isoamyl alcohol = 49:1) was added and centrifuged at 11,500 rpm for 10 min. The supernatant was then transferred to another tube, 100 μL of 3 M sodium acetate and 100 μL of isopropanol were added and incubated at room temperature for 20 min. The sample was centrifugated at 11,500 rpm for 20 min, followed by pellet washing with 80% ethanol, then dissolved into 20 μL of DEPC (NIPPON GENE Co., Ltd., Tokyo, Japan) for further real-time RT-PCR.

Real-Time RT-PCR

The RNA samples were diluted with DEPC to achieve a 50 mg/mL concentration of RNAs using a spectrophotometer (NanoDrop, Thermo Fisher Scientific, USA). The adjusted samples were subjected to real-time RT-PCR using the One Step SYBR RT-PCR Kit (Takara Corporation, Tokyo, Japan) and the PIKO REAL 96 Real-Time PCR System (Thermo Fisher Scientific, USA). The primer sequences used to amplify each gene and the RT-PCR settings are shown in Supplementary Table S1–2. The relative expression of each target gene was normalized to that of 18s rRNA, and the data were analyzed by the ΔΔCt method.

Vagotomy

Vagotomy was conducted with slight modifications to the previous study.⁶¹ Vagotomy was performed under combined triadic anesthesia (Medetomidine Hydrochloride (domitor, ZENOAQ, Japan), Butorphanol Tartrate (vetorphale, Meiji, Japan), midazolam (Sandoz Corporation, Japan)). A laparotomy incision was made on the ventral midline and the abdominal muscle wall was opened with a second incision. The gastrohepatic ligament was severed using fine forceps, and the stomach was gently retracted, revealing the descending ventral esophagus and the ventral subdiaphragmatic vagal trunk. The hepatic branch of this vagal trunk was then transected using fine forceps. The Sham group was produced with the same procedure except for vagal trunk transection. Successful vagotomy was confirmed by response to CCK-8 (8 μg kg^–1, i.p.). Mice with lower food intake than the average of the CCK-8-treated control group were considered false-vagotomized and removed from the experimental group. The results of the CCK test are shown in Figure S9 and Table S4–5.

Cecal pH and Short-Chain Fatty Acid (SCFA) Measurement

For cecal pH measurement, the probe of the pH meter (Euthech Instruments, USA) was inserted directly into a dissected cecum and then waited until the value stopped fluctuating. Then, cecum content was rapidly collected on ice and processed for gas chromatography coupled with mass spectrometry (GC-MS) to measure the cecum SCFA levels, using the protocol previously described,^18,62 using model 7890B or 5977B instruments (Agilent Technologies, Inc., Santa Clara, CA, USA). 50 μL of sulfuric acid, 200 μL of chloroform, and diethyl ether were added to approximately 50 mg of cecum contents. Then 100 μL of trimethylsilylation reagent (TMSI-H; GL Science Inc., Tokyo, Japan) was added to 300 μL of supernatant containing SCFAs to convert target compounds into volatile and thermally stable derivatives. The sample was then incubated at 60 °C for 30 min and on ice for 10 min. It was centrifuged at 14,000 rpm at room temperature for 30 s, and 1 μL of its supernatant was subjected to GC-MS. A standard mixture containing acetic acid, propionic acid, lactic acid, and butyric acid was also subjected to GC-MS after operating as in the sample to generate the calibration curve. The capillary column was InertCap Pure-WAX (30 m × 0.25 mm, df = 0.5 μm) (GL Sciences, Japan), and helium gas was used as the carrier gas during the measurement to increase the initial temperature from 80 °C to a final temperature of 200 °C.

Charcoal Meal Test

Charcoal meal test was conducted with slight modifications from the previous study.⁵⁹ Mice were fasted about 16 h before the test. Vermilion Indian ink was orally administered to them at a dose of 10 mL/kg body weight. Mice were sacrificed 10 min after administration, and their duodenum and small intestine (between the pylorus and the colon) were removed and spread on the aluminum plate immediately. The transit ratio was determined as the distance of ink transport divided by the total length of the intestinal tract.

Intestinal Permeability Test

The intestinal permeability test was conducted with slight modifications from the previous study.⁵⁹ 200 μL of FITC-dextran (MW 4000) solution (50 mg/mL; Merck KGaA, Darmstadt, Germany) was orally administered to mice. Serum samples were collected from the orbital plexus under anesthesia with isoflurane, incubated at room temperature for 1 h, centrifuged at 3000 rpm for 20 min, and the supernatant was collected. Following to the 5-fold sample dilution in the Milli-Q water, plasma FITC concentration was measured as fluorescence intensity using BioTek Synergy H1 (Agilent Technologies Japan, Ltd., Tokyo, Japan).

Feces and Urine Collection

Mice were housed in the metabolic cage (Natsume Seisakusho Co., Ltd., Tokyo, Japan) for three days, and then feces and urine samples were collected 24 h a third day. A urine sample was centrifuged at 1000 g, and the supernatant was stored at −80 °C until use to measure corticosterone levels. Defecation frequency was defined as the number of fecal pieces in 24 h on the third day, and fecal mass was measured at that time.

Corticosterone ELISA

Urine and serum corticosterone levels were measured using the ELISA kit (Funakoshi Co., Ltd., Tokyo, Japan). Corticosterone measurements were performed according to the instructions provided in the ELISA kit. In brief, a urine sample was collected using a metabolic cage as described above and diluted 20-fold in the buffer. Serum samples were collected from the orbital plexus under anesthesia with isoflurane, incubated at room temperature for 1 h, centrifuged at 3000 rpm for 20 min, and the supernatant was stored at −80 °C until use. Plasma samples were diluted 100-fold in the buffer and processed according to the instructions of the kit. The corticosterone concentration was measured as absorbance using BioTek Synergy H1 (Agilent Technologies Japan, Ltd., Tokyo, Japan).

Flow Cytometry

Spleen and bone marrow were used for the flow cytometry. Single-cell suspensions were prepared from freshly harvested spleen and bone marrow using the following procedure: First, spleens were mushed using a plunger and passed through a 70 μm filter with 0.5 mL 1xACK buffer. Bone marrow is flushed out from the femur and tibia by 1 mL FACS buffer. The cell suspension was hemolyzed with 1xACK buffer, then centrifuged at 1500 rpm for 10 min, and the supernatant was discarded. After another washing of cell suspension and then further nonspecific Fc binding with antimouse CD16/32 (Bio X Cell, NH, USA), cell surface markers were stained with the following antibodies (Supplementary Table 3): antimouse CD19-BV510 (GK1.5), CD5-BV421 (53–6.7), NK1.1-APC (PK136), CD11b-PerCP-Cy5.5 (M1/70), Ly6C-FITC (AL-21), Ly6G-eF450 (1A8, Thermo Fisher Scientific), I-A/I-E-APC-Cy7 (M5/114.15.2, Biolegend). Flow cytometry was performed on CytoFLEX S (Beckman Coulter) and analyzed using CytExpert software (Beckman Coulter). The gating strategy is shown in Figure S3C. Only the cell population displayed significant differences between the experimental cohort chosen and analyzed.

Statistical Analyses

All data were analyzed using GraphPad Prism version 10.2.0 (GraphPad Software, USA). First, we confirmed whether the data followed a normal distribution using the D’Agostino-Pearson test/Kolmogorov–Smirnov test and whether the data were equal variances using the F value test/Bartlett’s test. The significance between the two independent groups was assessed by parametric analysis using unpaired t tests or by nonparametric analysis Mann–Whitney tests. Significance between three independent groups was assessed by parametric analysis using One-way ANOVA with uncorrected Fisher’s LSD, or One-way ANOVA with Tukey’s multiple comparisons test. Nonparametric analysis Kruskal–Wallis tests with Dunn multiple comparison tests were used for abnormally distributed data. Two-way ANOVA with Uncorrected Fisher’s LSD or with Sidak’s multiple comparisons test was used for data on two factors.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsptsci.4c00270.

• Anxiety-like behavior in open file test and elevated plus maze test; monoamine levels in the hippocampus, prefrontal cortex, nucleus accumbens, and striatum; the monocyte population in bone marrow and splenocytes; cecum pH and SCFA concentration in vagotomized mice; mRNA expression of tight junction, cytokine and TRPV1 genes in intestinal tissue; primer sequences for RT-PCR; RT-PCR setting; and antibodies used in flow cytometry (DOCX)

Author Contributions

K.I., C.N., and A.H. contributed to the study’s conception and design with the help of S.S. K.I. mainly conducted the experiments with the help of H.H. for the intestinal permeability test, flow cytometry and its analysis, Y.K. for charcoal meal, H.S. for GC–MS and HPLC, and A.H. for vagotomy and behavior tests. K.I., A.H., C.N., and Y.K. dissected and sampled the mice. K.I. prepared the manuscript with the help of C.N. and S.S. All authors contributed to the manuscript revision and read and approved the submitted version.

The authors declare no competing financial interest.