Intestinal Epithelial Serotonin as a Novel Target for Treating Disorders of Gut-Brain Interaction and Mood

Abstract

Background & aims: Mood disorders and disorders of gut-brain interaction (DGBI) are highly prevalent, commonly comorbid, and lack fully effective therapies. Although selective serotonin reuptake inhibitors (SSRIs) are first-line pharmacological treatments for these disorders, they may impart adverse effects, including anxiety, anhedonia, dysmotility, and, in children exposed in utero, an increased risk of cognitive, mood, and gastrointestinal disorders. SSRIs act systemically to block the serotonin reuptake transporter and enhance serotonergic signaling in the brain, intestinal epithelium, and enteric neurons. Yet, the compartments that mediate the therapeutic and adverse effects of SSRIs are unknown, as is whether gestational SSRI exposure directly contributes to human DGBI development.

Methods: We used transgenic, surgical, and pharmacological approaches to study the effects of intestinal epithelial serotonin reuptake transporter or serotonin on mood and gastrointestinal function, as well as relevant communication pathways. We also conducted a prospective birth cohort study to assess effects of gestational SSRI exposure on DGBI development.

Results: Serotonin reuptake transporter ablation targeted to the intestinal epithelium promoted anxiolytic and antidepressive-like effects without causing adverse effects on the gastrointestinal tract or brain; conversely, epithelial serotonin synthesis inhibition increased anxiety and depression-like behaviors. Afferent vagal pathways were found to be conduits by which intestinal epithelial serotonin affects behavior. In utero SSRI exposure is a significant and specific risk factor for development of the DGBI, functional constipation, in the first year of life, irrespective of maternal depressive symptoms.

Conclusion: These findings provide fundamental insights into how the gastrointestinal tract modulates emotional behaviors, reveal a novel gut-targeted therapeutic approach for mood modulation, and suggest a new link in humans between in utero SSRI exposure and DGBI development.

Keywords: Disorders of Gut-Brain Interaction; In Utero SSRI; Mood; Vagal Afferent Signaling.

Figure 1.

Selective deletion of SERT from the intestinal epithelium throughout life promotes anxiolytic and antidepressive phenotypes. (A) Schematic of behavioral tests performed in Villin^Cre::SERT^fl/fl mice. (B) OFT: Total ambulatory distance, time in center, center entries, and number of vertical rearings. (C) NSFT: Latency to touch paper, bite and eat pellet, % mice not eating during the NSFT, and food consumption post-testing. (D) LDB: Time and entries into the illuminated side of the box. (E) EPM: Time and entries in open arms. (F) TST: Time spent immobile. Male (n = 32–57) and female (n = 34–53) mice. Student unpaired t test was used to compare groups and is represented as mean ± SEM. Analysis of variance with repeated measures was used to compare experimental groups across time. *P < .05, ns: non-significant. Data are represented as mean ± SEM in an aligned scatter plot.

Figure 2.

Inducible SERT deletion from the intestinal epithelium in adulthood promotes an anxiolytic phenotype. (A) Schematic of behavioral tests performed in Villin^Cre-ERT2::SERT^fl/fl-tamoxifen (TAM) mice. (B) OFT: Total ambulatory distance, time in center, center entries, and vertical rearings. (C) NSFT: Latency to touch paper, bite and eat pellet, % of mice not eating and food consumption post-testing. (D) LDB: Time and entries into the illuminated side of the box. (E) EPM: Time and entries in open arms. (F) TST: Time spent immobile. Males (n = 23–69) and females (n = 26–82). Student unpaired t test was used to compare groups, which are represented as mean ± SEM. *P < .05, **P < .01, ***P < .001. Data are represented as the mean ± SEM in an aligned scatter plot.

Figure 3.

Targeted deletion of SERT from the intestinal epithelium does not alter in vivo colonic motility. Colonic motility was observed in vivo (A, D) and through measurements of colonic migrating motor complexes (CMMCs, C, F) of Villin^Cre::SERT^fl/fl (A–C) and Tam-treated Villin^Cre-ERT2::SERT^fl/fl (D–F) mice. All parameters were measured from spatiotemporal heatmaps (representative: B and E) with black arrows indicating representative CMMCs. Data are from at least 4 (A, D; 30 mice/group), 3 (C; 10 mice/group), or 1 (F, at least 6 mice [males]/group) independent experiments. Data are represented as the mean ± SEM in an aligned scatter plot and analyzed with Student unpaired t test. **P < .01.

Figure 4.

Contrary to SERTKO, selective deletion of SERT from the intestinal epithelium does not affect serotonergic innervation in the ENS. (A–C) Whole-mount staining and quantitation of serotonergic axons (A, top; scale bars, 100 μm) and varicosities (A, bottom; scale bars, 20 μm), counterstained with antineuronal nuclear antibody type 1 (ANNA-1) in the colonic myenteric plexus of Villin^Cre::SERT^fl/fl (A and B) and Tam-injected Villin^Cre-ERT2::SERT^fl/fl mice (C). Whole-mount staining (D) and quantitation of serotonergic axons and varicosities in SERT^−/− mice (E). Scale bars, 20 μm. 5-HT varicosities were quantified by tracing around a cluster of ANNA-1 neurons that outline the ganglion (dotted blue line) and creating a threshold and subsequent binary mask to analyze positive punctate staining (black dots) that are >1.5 μm². Data are from at least 2 (B) and 1 (C and E) independent experiments with at least 5 mice/group. Data are represented as the mean ± SEM in an aligned scatter plot and analyzed with Student unpaired t test, which are represented as mean ± SEM. *P < .05.

Figure 5.

5-HT depletion from the intestinal epithelium promotes anxiogenic and depressive-like phenotypes. Vertical rearings from the OFT (A, E), latency to touch filter paper, touch food pellet and bite pellet from the NSFT (B, F), and time spent immobile in TST (C, G) of TPH1^−/− mice (A–C) and WT mice treated with LX1606 (E–G). Schematic of LX1606 administration protocol (D). Data are from at least 3 (A–C, at least 15 mice/group) and 1 (E–G, at least 10 mice [males]/group independent experiments). Data are represented as the mean ± SEM in an aligned scatter plot and analyzed with Student unpaired t test. *P < .05; **P < .01.

Figure 6.

Afferent vagal ablation normalizes anxiety- and depressive-like phenotypes in mice in which SERT is selectively deleted from the intestinal epithelium. (A) Schematic of SAP+/−CCK injections into nodose ganglia of Villin^Cre::SERT^fl/fl mice. (B) In situ hybridization of Phox2b and CCKa1 in nodose ganglion of mice injected with SAP+/−CCK. Scale bars, 50 μm. (C–G) Behavioral assessment of time and entries into the open arms in the EMP (C and D), latency to touch food pellet in the NSF represented as a Kaplan-Meier plot measurement over time (E and F) and sucrose preference (SPT) (G) of SAP+/−CCK injected Villin^Cre::SERT^fl/fl mice and controls. Data are from at least 1 independent experiment (C–F; at least 6 mice [males and females]/group), G; 6 mice [males]/group). Two-way analysis of variance was used to compare groups and treatment effect with multiple comparison tests using a Sidak correction. Data are represented as the mean ± SEM in an aligned scatter plot. *P < .05; **P < .0001.

Figure 7.

In utero SSRI/SNRI exposure is significantly associated with an increased risk of functional constipation. (A) Graphical narrative demonstrating recruitment groups and assessment of DGBI. (B) Frequencies of major DGBI classes in the first year of life. (C) Estimates, odds ratios, standard error, and P values of DGBI diagnoses. (D) Rate of functional constipation diagnoses over 0 to 12 months. The shaded regions represent the range of the 95% confidence interval.