Interactions Between Commensal Bacteria and Enteric Neurons, via FPR1 Induction of ROS, Increase Gastrointestinal Motility in Mice

Abstract

Background & Aims:

Reduced gastrointestinal (GI) motility is a feature of disorders associated with intestinal dysbiosis and loss of beneficial microbes. It is not clear how consumption of beneficial commensal microbes, marketed as probiotics, affects the enteric nervous system (ENS). We studied the effects of the widely used probiotic and the commensal Lactobacillus rhamnosus GG (LGG) on ENS and GI motility in mice.

Methods:

Conventional and germ-free C57B6 mice were gavaged with LGG and intestinal tissues were collected; changes in the enteric neuronal subtypes were assessed by real-time PCR, immunoblots and immunostaining. Production of reactive oxygen species (ROS) in the jejunal myenteric plexi and phosphorylation (p) of mitogen-activated protein kinase 1 (MAPK1) in the enteric ganglia were assessed by immunoblots and immunostaining. Fluorescence in situ hybridization was performed on jejunal cryosections with probes to detect formyl peptide receptor 1 (FPR1). GI motility in conventional mice was assessed after daily gavage of LGG for 1 week.

Results:

Feeding of LGG to mice stimulated myenteric production of ROS, increased levels of phosphorylated MAPK1, and increased expression of choline acetyl transferase by neurons (P<.001). These effects were not observed in mice given N-acetyl cysteine (a ROS inhibitor) or LGGΩSpaC (an adhesion-mutant strain of LGG) or FPR1-knockout mice. Gavage of mice with LGG for 1 week significantly increased stool frequency, reduced total GI transit time, and increased contractions of ileal circular muscle strips in ex vivo experiments (P<.05).

Conclusions:

Using mouse models, we found that LGG-mediated signaling in the ENS requires bacterial adhesion, redox mechanisms, and FPR1. This pathway might be activated to increase GI motility in patients.

Graphical Abstract

LGG induced ROS via FPR1 activates phosphorylation of p44/42 MAPK/Erk 1/2) in the enteric neurons in sub mucosal and myenteric plexi neurons leading to neuronal differentiation that favors cholinergic neurons (ChAT) and enhances GI motility. NAC, the ROS inhibitor, and the pilin defective mutant LGGΩSpaC, abolishes the effects of LGG, and fails to induce ROS and downstream Erk 1/2 phosphorylation and subsequent improvements in GI motility.

Introduction

Physiological gastrointestinal motility is orchestrated by the enteric nervous system (ENS), which is organized into two main plexi -the sub mucosal (Meissner’s) and the myenteric (Auerbach’s) plexi ¹. The peristaltic reflex, comprised of sequential contractions and relaxations, are mediated by intrinsic primary afferent neurons (IPANs) that have cell bodies residing in sub mucosal and myenteric plexi with cytoplasmic projections extending into the mucosa ². Upon stimulation by luminal bacterial by-products such as N-formyl peptides, IPANs release specific neuropeptides, which in turn regulate immune functions as well as pain perception ³. In addition, enteroendocrine cells and mucosal glial cell populations can also be stimulated by nutrients and bacterial by-products ⁴. Though initial ENS development occurs during the fetal phase, the terminal differentiation of enteric neurons is established during the post-natal development period, and is closely linked to the presence of microbiota ⁵. Germ free (GF) mice exhibit significant reduction in myenteric plexi densities of jejunum and ileum and decrease in the amplitude of muscle contractions in the GI tract ⁶, which resemble genetic mouse models where specific transcription factors for ENS development are knocked out ⁷.

Clinically, reduced gastrointestinal (GI) motility is associated with numerous human disorders including diabetes, Parkinson’s disease (PD), autism and constipation-predominant irritable bowel syndrome (IBSC). Commensal gut microbiota is essential for normal excitability of gut sensory neurons and myenteric neurons ⁵, and the aim of this study is to elucidate the signaling mechanisms in the ENS that underlie probiotic-induced changes in GI motility. Though several studies have examined the impact of probiotics on gut motility in general, only few studies have focused on their mechanisms of action involving the modulation of neuronal excitability by short chain fatty acids like butyrate, activation of sensory neurons via bacterial polysaccharides ⁸, signaling to ENS via micro vesicles ⁹ and modulation of calcium-dependent potassium channels ¹⁰. In germ free animals, various probiotic bacteria have been shown to regulate gut motility ¹¹ and interestingly different strains of the same probiotic bacteria have been shown to exert region-specific differences in gut motility ^{12, 13}. Limited, yet encouraging, evidence exists for Lactobacillus rhamnosus GG (LGG) in attaching in vivo to colonic mucosae for more than a week after discontinuation of administration, highlighting it potential therapeutic application in regulating gut motility¹⁴.

Previously, we demonstrated that the human gut commensal and widely used probiotic Lactobacillus rhamnosus GG (LGG) has anti-inflammatory effects ¹⁵, and also mediates redox dependent signaling in the intestinal epithelia, with effects on epithelial proliferation, migration, barrier function and restitution ^16–18. Here we examined the effects of LGG on the enteric nervous system (ENS) and GI motility using conventional and germ free (GF) mice. Acute contact with LGG (after 2h oral gavage) markedly stimulated the production of reactive oxygen species (ROS) in the myenteric plexi, which induced significant phosphorylation of p44/42 mitogen activated protein kinase (Erk 1/2) in the myenteric neurons in the ENS, leading to an upregulation of enteric neuropeptides in both GF and conventional mice. However, these effects were abrogated when an adhesion defective strain of LGG viz, LGGΩSpaC was used, and in animals co-treated with LGG and the antioxidant N-acetyl cysteine (NAC). These findings suggest that LGG-mediated signaling effects in enteric neurons are both adhesion and redox-dependent. Studies using the formyl peptide receptor 1 and 2 knockout mice (FPR1 KO & FPR2 KO) revealed that effects of LGG on enteric neuronal signaling were largely mediated by FPR1, expressed on enteric neuronal cells. Finally, daily gavage of LGG in conventional mice for at least 1 week resulted in significantly increased stool frequency, reduced whole gut transit time and enhanced ileal muscle strip contractions ex vivo. Together, these data demonstrate a novel mechanism for LGG-mediated augmentation of gut motility via FPR1.

Methods

Mice.

Eight- to 12-week-old C57B6 male mice (Jackson Laboratory, Strain No. 000664) were used for all experiments. After experimental procedures, mice were euthanized with CO2. All murine experimental procedures were reviewed and approved by the Institutional Animal Care and Use Committee at Emory University and were performed according to the Emory guidelines for the ethical treatment of animals.

Bacterial strains and growth conditions.

Bacterial strains used in this study include Lactobacillus rhamnosus GG (ATCC 53103) and an isogenic strain GG-spaC mutant (ΩspaC::Eryr, CMPG10102) ¹⁹. For routine culture, lactobacilli were grown in de Man-Rogosa-Sharpe (MRS) at 37°C under static, microaerophilic conditions. After overnight culture, the bacterial strains were centrifuged at 3,000 × g for 5 min, and the resulting pellets were resuspended in Hanks’ balanced salt solution (HBSS). This process was repeated twice, and the cultures were then diluted to 1X 10¹⁰ cfu/ml before gavage to the mice ¹⁶.

Antibodies and Chemicals.

Phospho-Erk (p44/42 mitogen activated protein kinase), Total Erk 1/2, and phospho-CEBP-β from Cell Signaling (Danvers, MA), Peripherin (Millipore, MA), choline acetyl transferase (Millipore, MA), SERT (Abcam), Hand-2 (Invitrogen), N-acetyl cysteine and fMLF (Sigma), ROSstar™ 550 (Hydrocyanine-3 probe) from LI-COR ^{20, 21}.

Western blotting.

The myenteric plexi from ileum/jejunum/colon from mice were flash frozen and the lysates were immunoblotted for pErk 1/2, Total Erk 1/2, GAPDH, choline acetyl transferase (ChAT) , SERT, Hand-2 and Peripherin.

ROS assay by confocal imaging of the longitudinal muscle myenteric plexus (LMMP).

Mice were fasted for 16 h and intraperitoneally (i.p.) injected with 100 μM ROSstar550, 15 min before oral gavage of either of HBSS or LGG or LGGΩSpaC or at 1 × 10¹⁰ CFU/ml ¹⁶. A separate group of LGG gavaged mice were also gavaged with N-acetyl cysteine (@ 150 mg/kg body weight). The mice were sacrificed 2h post gavage, and LMMP was micro dissected from the jejunum, and mounted on a glass slide with Vectashield. The fluorescent images were immediately captured by confocal microscopy from 5 random fields on each slide. For quantitation of ROS, fluorescence intensity of the images was measured using the ImageJ software. Immunostaining. The jejunal cryosections were stained for p-Erk1/2, total Erk 1/2, p-CEBP-β & ChAT along with nuclear stain DAPI, and confocal imaging was done, followed by arbitrary fluorescence quantification using Image J.

CLARITY-immunostaining.

The whole mouse ileum was processed for CLARITY staining ²². The cleared tissue was incubated with ChAT (1: 300), pErk 1/2 (1:100), Peripherin (1: 800) antibodies over night at 4°C, followed by incubation with respective Alexa Fluor secondary antibodies and DAPI at room temperature. The tissue was mounted on glass slides in 80% glycerol and imaged using confocal microscopy. The z-stacks were captured from the gut lumen toward the mucosal layer to visualize the pErk staining in the submucosal and myenteric plexi.

Arbitrary fluorescence quantification.

Staining intensity in at least 8-10 myenteric ganglia per mouse per experimental condition (HBSS, LGG, LGG SpaC, LGG + NAC) was assessed in a blinded fashion using image J software after the composite image was split into respective color channels (blue-DAPI, green-pErk 1/2 or Total Erk 1/2 or ChAT), and the intensity of green fluorescence in the myenteric ganglia was calculated and plotted as relative arbitrary intensity units (RFU). The intensity from n = 3 to 5 mice per group per experimental condition was averaged to plot the RFU graphs for the respective experiments.

Real time PCR (qPCR).

The myenteric plexi /ileum from mice were flash frozen, and RNA was extracted using RNAesy Kit (Qiagen). cDNA was synthesized from RNA using the iscript supermix kit (Bio Rad), and SYBR green reaction mix (BioRad) was used for q-PCR with primers for Peripherin, choline acetyl transferase (ChAT), Tyrosine hydroxylase (TH), Neuropeptide Y (NPY), nNOS (neuronal nitric oxide synthase), SERT (serotonin reuptake transporter), FPR1, FPR2 ²³ and enteric neuronal transcription factor Hand-2.

Fluorescence In situ hybridization (FISH).

The probes for mouse FPR1 and Peripherin were purchased from ACD Bio. FISH was performed using the RNAScope fluorescent kit and imaged using confocal microscopy.

Gastrointestinal Motility studies.

(A) Stool frequency (number of stool pellets extruded per mouse per hour)- was measured in mice receiving a daily gavage of HBSS, LGG at day 6 and day 13 ²⁴. (B) Total GI transit time: Mice were gavaged with red carmine dye and the time needed to extrude the first red pellet was used an index of total GI transit time ²⁵. Isometric muscle recording was done on ileal circular muscle strips as described previously ^{24, 26}.

Statistical analysis.

The data were analyzed using Graph Pad Prism 5 software (GraphPad, La Jolla, CA. Data are represented as Mean +/− SEM. One-way analysis of variance (ANOVA) was used to compare data in experiments involving 4 treatment groups viz HBSS, LGG, LGG SpaC and LGG + NAC, followed by Tukey’s post hoc comparison test. Student’s t test was used to compare data from experiments involving 2 treatment groups. Differences were considered significant at P < .05. Data are represented as Mean +/− SEM in all figures.

Results

LGG-induced reactive oxygen species (ROS) triggers p44/42 MAPK (Erk 1/2) phosphorylation in the enteric ganglia of germ free (GF) mice.

Our laboratory has previously shown that LGG stimulates ROS generation in intestinal epithelia, triggering the activation of extracellular regulated kinase viz p44/42 mitogen activated protein kinase (MAPK/Erk 1/2) ¹⁶. Hence, in the current study we assessed if oral administration of LGG to GF mice could induce ROS and Erk 1/2 phosphorylation in the myenteric ganglia, in addition to the epithelial effects of LGG that we have previously observed. Interestingly, GF mice receiving LGG by oral gavage (2h) showed increased ROS generation as detected with the superoxide specific dye Hydro Cy3 as we have shown in the past ^{17, 18, 27, 28}. Increased fluorescence was observed in the cell bodies and fibers of the longitudinal muscle myenteric plexi (LMMP) of LGG-treated mice as compared to mice gavaged with Hanks balanced salt solution (HBSS, Figures 1A & 1B). As Erk 1/2 MAPK can be induced by redox signaling ¹⁶, we next sought to assess the activation status of this pathway. Indeed, 2 hour contact of LGG induced phosphorylation of Erk 1/2 in the cells of the jejunal myenteric ganglia of GF mice as observed by immunostaining and confocal microscopy (Figures 1C & 1D). These data reveal that LGG mediates acute ROS production and subsequent Erk 1/2 phosphorylation in the enteric ganglia.

Figure 1. Lactobacillus rhamnosus GG (LGG) induces reactive oxygen species (ROS) and triggers p44/42 MAPK (Erk 1/2) phosphorylation in the enteric ganglia of GF mice.

GF mice were intra peritoneally injected with hydro-Cy-3 (20 uM) for 15 minutes before they were gavaged with Hank’s balanced salt solution (HBSS) or LGG @ 10¹⁰ colony forming units (cfu). After 2h, mice were sacrificed and the longitudinal muscle myenteric plexi (LMMP) from the jejunum was micro dissected and imaged by confocal microscopy. A separate piece of jejunum was cryosectioned and immunostained with antibodies for phospho-Erk 1/2 and detected with AF 488/green (pErk 1/2) and AF 594/red (Peripherin), along with DAPI as the nuclear stain. (A) Confocal images of LMMP showing ROS generation (B) Number of ROS-positive cells per ganglion in GF mice treated with HBSS and LGG. Representative confocal sections of jejunum showing (C) phospho-Erk ½ (green) and peripherin (red) in the myenteric ganglia, and (D) Graphical representation of the intensity of p-Erk (green) fluorescence in the myenteric ganglia as assessed by Image J. Magnification 40X, Scale bar 50 μm, Mean +/− SEM, n =3 per group, * P < 0.05 by t-test.

LGG induces enteric neuronal remodeling in GF mice.

Having demonstrated the effects of LGG on the ROS-Erk 1/2 signaling in the myenteric ganglia of GF mice, we next assessed the impact of these signaling events on enteric neuronal differentiation by real time PCR, Western blotting and immunostaining. We found that GF mice receiving LGG demonstrated a significant increase in Hand-2 (enteric neuron transcriptional factor that regulates enteric neuronal differentiation), choline acetyl transferase (ChAT) and serotonin transporter (SERT) by real time PCR (Figure 2A) and Western blotting (Figure 2B & 2B*). Further, we validated the increase in ChAT neuron immunoreactivity in myenteric ganglia by immunostaining on cryosection (Figures 2C & 2C*). These data suggest that LGG can induce enteric neuronal differentiation and ROS-Erk 1/2 signaling in GF conditions despite the underlying aberrant ENS phenotype.

Figure 2. Lactobacillus rhamnosus GG (LGG) induces enteric neuronal remodeling in GF mice.

GF mice were gavaged with HBSS or LGG (@ 10¹⁰ cfu) for 2 h and sacrificed. RNA was extracted from the ileum and real time PCR (q-PCR) was done to assess the changes in peripherin, Hand-2 and neuronal subtypes like ChAT, SERT, NPY, nNOS and TH. Immunostaining for ChAT was performed on cryofixed sections, and green fluorescence in the myenteric ganglia (arrows) representing ChAT immunoreactivity assessed by Image J. (A) Gene expression of Peripherin, Hand-2, and various neuronal subtypes viz ChAT, SERT, NPY, nNOS and TH. (B & B*) Representative Western blot for Hand-2, ChAT SERT and the graphical representation (C) Representative images of ChAT (green) and peripherin (red) immunostaining in the myenteric ganglia (arrows), DAPI (blue). (C*) Graphical representation of ChAT (green) staining intensity in arbitrary fluorescence units using Image J (n = 4 per group). Mean +/− SEM, n = 3 per group, * P < 0.05, ** P < 0.01, *** P < 0.001 by t-test. Magnification 40X, Scale bar B-50 μm & C-100 μm.

LGG-induced p44/42 MAPK (Erk 1/2) phosphorylation in the enteric ganglia of conventional mice is contact and redox-dependent.

GF mice exhibit several developmental abnormalities, prompting us to determine if LGG induced neural changes are observed in animals with a normal microbiota – a necessary precondition for successful probiotic approaches. In addition, we also tested an adhesion-mutant strain of LGG viz LGGΩSpaC (a mutant strain of LGG that fails to attach to epithelia due to a lack of a pilin subunit) to explore if LGG-mediated effects are contact dependent. Previous work has shown that this mutant strain is unable to induce redox signaling in gut epithelia ¹⁶. Further, to confirm the role of redox signaling on the downstream effects on the ENS, we also tested the effects of LGG along with the antioxidant N-acetyl L-cysteine (NAC) (gavaged @ 150mg/kg body weight). Similar to GF mice, conventional mice receiving LGG produced more ROS than the mice receiving HBSS (Figure 3A). Further, the conventional mice gavaged with the adhesion mutant LGGΩSpaC, and also the mice co administered wild type LGG and NAC exhibited significantly reduced ROS (Figure 3A). Concomitant with the ROS generation pattern, we found that conventional mice receiving LGG exhibited enhanced phosphorylation of Erk 1/2 in the enteric ganglia than the buffer controls (in addition to the Erk 1/2 phosphorylation in the epithelium as we previously reported¹⁵) and is attenuated in mice receiving the adhesion mutant LGGΩSpaC or treated with both LGG and NAC (Figures 3B). Indeed, activation of Erk is clearly seen in sub mucosal (outlined with white circles) and myenteric ganglia (arrows) Figure 3B). Further co-labeling studies on isolated longitudinal muscle myenteric plexi (LMMP) also demonstrated significant Erk 1/2 phosphorylation in the myenteric neurons and neuronal processes from LGG-treated mice (pErk positive and Peripherin positive cells) (Figure 3C). In addition, Z-stack reconstructions of multiple confocal images from clarity immunostaining unambiguously showed LGG induced activation in the epithelial (as we previously published) as well as both the cell bodies of the submucosal (Figure 3C*) and myenteric plexi (Supplementary Video 1). We observed that the total Erk 1/2 levels were not significantly different in the enteric ganglia among these treatment groups (Figures 3D & 3D*). We further validated the phosphorylation of Erk 1/2 by Western blotting (Figure 3E & 3E*) and observed LGG-induced enhanced phosphorylation of Erk 1/2 in the jejunum, ileum and colon. These data demonstrate that even in the presence of normal microbiota in conventional mice, LGG can induce ROS -Erk 1/2 signaling in enteric neurons similar to that in GF conditions, and these effects were both redox and adhesion-dependent.

Figure 3: LGG-induced redox signaling and Erk 1/2 phosphorylation in the enteric neurons is contact/adhesion-dependent.

Conventional mice received intra peritoneal injections of hydro-Cy-3 15 minutes before gavage with HBSS or LGG (@ 10¹⁰ cfu), LGGΩSpaC or LGG along with NAC (@ 150 mg/kg body weight). After 2h, mice were sacrificed, and the jejunal LMMP micro dissected for confocal imaging. The jejunal cryosections were immunostained for pErk 1/2 or total Erk 1/2 along with DAPI (nuclear stain, blue). (A) Confocal images of LMMP showing ROS production in cell bodies (arrows), (B) Representative cryosections of jejunum showing pErk (green) immunostaining, (C) p-Erk 1/2 (green) fluorescence in LMMP & (C*) submucosal plexi from mice treated with HBSS or LGG for 2 h. (D & D* ) Total Erk 1/2 (green) in the myenteric ganglia (arrows), (E & E*) Representative Western blot for pErk/Total Erk 1/2 and the graphical representation. Mean +/− SEM from 3 to 4 independent experiments (n =3 - 4 mice per group per experiment), ANOVA followed by Tukey’s test, *P < 0.05, ** P < 0.01, *** P < 0.001. Magnification 40X. Scale bar 50 μm.

LGG induced enteric neuronal remodeling in conventional mice is adhesion and redox-dependent.

We next investigated if ROS-pErk signaling could induce neuronal differentiation in conventional mice similar to that observed in GF conditions. We observed that 2 hour administration of LGG (but not the adhesion mutant LGGΩSpaC or LGG+NAC) resulted in a significant induction of Hand-2, ChAT neurons and SERT transporter by real time PCR (Figure 4A) and Western blotting (Figure 4B & 4B*), suggesting that LGG-induced neuronal remodeling is adhesion and redox-dependent. We further validated the increase in LGG-induced ChAT neuron expression by immunostaining in jejunal myenteric plexi cryosections (Figures 4C) and jejunal LMMP (Figures 4D) and CLARITY immunostaining (Figure 4E). The CLARITY method permits evaluation of the entire plexus in situ, addressing the bias secondary of sectioning representation. In addition, LGGΩSpaC and LGG + NAC treated mice did not show significant upregulation of ChAT by staining (Figures 4C & 4D). These data suggest that ROS-pErk 1/2 signaling axis mediates LGG-induced neuronal remodeling in the ENS. It has been demonstrated that Erk 1/2 activation leads to downstream phosphorylation of CEBP-β. In neurons, p-CEBP-β is known to play a key role in the transcription of cholinergic neurons like ChAT ²⁹. In addition to Erk 1/2 activation, we also found that LGG treatment increases the phosphorylation of CEBP-β in conventional (Supplementary Figures 1A & 1A*) and GF mice (Supplementary Figures 1B & 1B*). We also observed increased p-CEBP-β staining in the LMMP from LGG treated mice (Supplementary Figure 1C), which exactly resembled that of Erk 1/2 activation and was sensitive to ROS inhibition. Thus, we conclude that LGG- induced neuronal differentiation in conventional mice is redox and adhesion-dependent similar to that observed in GF conditions and involves Erk 1/2 and p-CEBP-β phosphorylation in enteric ganglia.

Figure 4: LGG-induced ENS remodeling in conventional mice is sensitive to inhibition by the ROS inhibitor N-acetyl L-cysteine (NAC).

Conventional mice were gavaged with HBSS, LGG @ 10¹⁰ cfu/ml) or LGGΩSpaC (@ 10¹⁰ cfu) or LGG along with NAC (@ 150 mg/kg body weight). After 2h, the mice were sacrificed, RNA extracted from the ileum and q-PCR was performed to assess the changes in Hand-2, ChAT neurons and SERT transporter. Immunostaining for ChAT/Peripherin was performed on cryofixed sections and green fluorescence in the myenteric ganglia representing ChAT was assessed by Image J. Further LMMP were peeled from jejunum, fixed and immunostained for ChAT/DAPI. (A) q-PCR (B & B*) Representative Western blot for Hand-2, ChAT, SERT and Peripherin and the graphical representation (C) ChAT staining of jejunal cryosections with myenteric ganglia (arrows) (D) ChAT staining in jejunal LMMP and (E) Representative 3D confocal images of ChAT (green)/Peripherin (red) immunostaining in whole mouse ileum by CLARITY immunostaining. Corresponding staining data in B, C & D are also represented by histograms. Magnification 40x. Scale bar 50 um. Mean +/− SEM, n = 4 per group, ANOVA followed by Tukey’s test, * P < 0.05, *** P < 0.001.

Formyl peptide receptor (FPR1) is critical to LGG-mediated effects of ROS activation and pErk 1/2 phosphorylation in the enteric ganglia.

We next investigated if bacterial products like formylated peptides (N-formyl-methionyl-leucyl-phenylalanine (fMLF)) could also induce effects similar to LGG. Past work from our laboratory has demonstrated that epithelial redox signaling can be induced by bacterial products via interaction with the epithelial-expressed formyl peptide receptors ^{16, 27}. We observed that conventional mice gavaged with fMLF (100 nM) demonstrated remarkably enhanced ROS generation (Figure 5A) and Erk 1/2 phosphorylation in enteric ganglia (Figure 5B). We next assessed the role formyl peptide receptors 1 and 2 (FPR1 & FPR2) in mediating the effects of bacterial peptides like fMLF. Compared to WT mice, formyl peptide receptor 1 (FPR1) knockout mice (FPR1 KO) exhibited a significant reduction in LGG-induced ROS production (Figure 5C). Concomitantly, Erk 1/2 phosphorylation in the enteric ganglia of FPR1 KO mice was significantly reduced as compared to WT (Figure 5D), indicating that LGG-induced ROS and p-Erk 1/2 are largely dependent on FPR1 signaling. We found that FPR2 knockout mice (FPR2 KO) responded to LGG with enhanced ROS production (Figure 5C) and Erk 1/2 activation (Figure 5D) comparable to that in conventional mice. We next assessed if FPR1-mediated ROS-Erk 1/2 signaling is critical to the downstream effects on neuronal differentiation. We observed that FPR1 KO mice did not exhibit upregulation of ChAT and SERT mRNA in response to LGG administration as detected by real time PCR (Figure 5E). These data suggest that induction of ROS & Erk 1/2 phosphorylation in the myenteric ganglia are FPR1-dependent and essential for neuronal differentiation.

Figure 5. Formyl peptide receptor (FPR1) is critical to the LGG-mediated effects of ROS activation and pErk 1/2 phosphorylation in the enteric ganglia.

Conventional mice were gavaged with HBSS or fMLF (0.5 uM/100 gm body weight). In a separate experiment, WT, FPR1 and FPR2 KO mice were gavaged with HBSS or LGG (@ 10¹⁰ cfu/ml). After 2h, mice were sacrificed and the jejunal LMMP was micro dissected and imaged for ROS. The jejunum was cryo sectioned and immunostained for pErk 1/2 and total Erk 1/2 along with DAPI (nuclear stain, blue). RNA was extracted from the ileum of WT, FPR1 and FPR2 KO mice and q-PCR was performed with primers for Peripherin, ChAT and SERT. Immunostaining for ChAT was performed by CLARITY on whole ileum. Confocal images of LMMP showing ROS production (arrows) in conventional mice gavaged HBSS or fMLF in (A), and representative sections of jejunum showing pErk (green) and Total Erk 1/2 (green) in the myenteric ganglia (arrows) in (B) Confocal images of ROS production in WT FPR1 KO and FPR2 KO mice shown in C with the corresponding histogram. Representative sections of jejunum showing pErk (green) staining in the myenteric ganglia (arrows) in WT, FPR1 KO and FPR2 KO mice in D, with the histogram showing intensity of p-Erk 1/2 (green) fluorescence. Fold change in ChAT and SERT mRNA shown in (E). Mean +/− SEM, n = 4 per group, ANOVA followed by Tukey’s test, ** P < 0.01, *** P < 0.001. Magnification 40x, Scale bar 50 μm.

Localization of FPR1 on enteric neurons by qPCR and fluorescence in situ hybridization (FISH).

It has been shown that dorsal root ganglia and vomeronasal neurons express FPR receptors and modulate several immune and chemosensory functions²³. However, it is not known if enteric nervous system (ENS) expressed FPRs. RNA extraction and qPCR performed on the micro dissected myenteric plexi (isolated from jejunum, ileum and colon) from WT mice demonstrated that FPR1 & 2 are indeed expressed in the enteric neurons (Figure 6A). Further, in situ hybridization using RNA probes demonstrated that, in addition to its presence on the epithelium as we have previously demonstrated ¹⁸, FPR1 is also expressed in the peripherin-positive neuronal cells of the myenteric plexi (Figures 6B & 6C, ganglion zoomed for clarity in 6C), but not in samples obtained from FPR1 KO mice used as negative control (Supplementary Figure 2). This novel finding of the presence of FPR1 on enteric neurons reveals the significance of neuronal processes projecting into the mucosa (as depicted by CLARITY staining in Figures 6D & 6E and Supplementary Video 2) in sensing luminal bacteria, as well as its key role in LGG-induced signaling in enteric neurons as FPR1 KO mice failed to induce ROS-pErk 1/2 signaling in the myenteric plexi.

Figure 6. Localization of FPR1 receptors on enteric neuronal cells in mouse intestine by q PCR and Fluorescence in situ hybridization (FISH).

RNA was extracted from the myenteric plexi of jejunum, ileum and colon of mice and q-PCR performed using murine FPR1, FPR2 primers, and normalized with peripherin. Jejunal cryosections were subjected to FISH using FPR1 RNA probes along with pan neuronal marker peripherin probes. (A) Relative expression of FPR 1 & 2 in jejunum, ileum & colon (B) Localization of FPR1 on epithelial cells and enteric ganglia (white arrows), (C) Zoomed image of myenteric ganglia showing localization of FPR1 on peripherin positive neurons (D) CLARITY staining for peripherin (green) and DAPI on whole ileum and (E) 3D CLARITY image created from Z stack of image in D showing neuronal processes (green) extending into the gut lumen. Magnification 40x, Scale bar 50 um.

Improved GI motility in mice receiving daily gavage of LGG.

Having observed that LGG exerts acute effects on enteric neuronal signaling pathways, we next assessed the effects of daily single administration of LGG on GI motility in conventional mice. We observed that conventional mice receiving a single daily gavage of LGG for at least 1 week exhibited improved motility as evident from an increase in stool frequency (Figure 7A) and reduction in whole gut transit time (Figure 7B), without significant changes in stool wet weight and water content (Figure 7C & 7D), indicating the absence of diarrheal phenotype. Further, isometric muscle recording on ex vivo ileal circular muscle strips incubated with NAME (NO inhibitor) demonstrated a significant increase in EFS-induced contractions in LGG-treated mice (Figures 7E & 7F). These observations collectively indicate that LGG favors EFS-induced contractions in the presence of L-NAME and improves GI motility without diarrheal phenotype. We also observed similar effects on stool frequency and whole gut transit time (Supplementary Figures 3A & 3B) when LGG was administered daily for 2 weeks.

Figure 7: Daily LGG gavage improves gastrointestinal motility in Conventional mice.

Conventional mice were gavaged daily with HBSS, LGG, or LGGΩSpaC @ 10¹⁰ cfu/ml for 1 or 2 weeks, and gastrointestinal motility was assessed by stool frequency (number of stool pellets per hour per mouse), total gastrointestinal (GI) transit time (time for expulsion of the first red stool pellet after gavage of the red carmine dye) stool wet and dry weights, and stool water content. A separate group of conventional mice were gavaged daily with HBSS or LGG @ 10¹⁰ cfu/ml for 2 weeks, and GI motility was assessed by isometric muscle recording from distal ileal circular muscle strips after incubation with L-Nitro-Arginine Methyl Ester (L-NAME) followed by electrical field stimulation (EFS) at 24 V, 10 Hz, 0.3 milliseconds for 20 seconds. The y axis represents force, millinewton (mN); x axis represents time. Contraction was expressed as a percentage change from baseline muscle tone. (A) Stool frequency (B) GI transit time (in minutes) C) Stool wet weight and (D) Stool water content after 1 week of LGG gavage, (E) Representative tracings from circular ileal muscle strips of mice gavaged with HBSS or LGG for 2 weeks, (F) Percent contraction. Mean +/− SEM, ANOVA followed by Tukey’s test or t test, n = 5, * P < 0.05.

Discussion

Our laboratory has previously demonstrated that gut commensals can induce physiological levels of ROS that can exert significant redox-dependent regulatory effects on host immune function ³⁰, intracellular signaling, and cytoskeletal dynamics ^{21, 28, 31}. Bacterially-induced redox signaling occurs in the epithelia by the action of formyl peptide receptors (FPRs) and the epithelial NADPH oxidase, Nox1, in a manner highly analogous to the classical respiratory burst characteristic of professional phagocytic cells, which utilize Fpr1 and Nox2. Rapid and highly localized oxidation of a subset of redox-sensitive regulatory enzymes results in activation of effector pathways including activation of ERK/MAPK. In the current study, we show that the natural commensal and probiotic Lactobacillus rhamnosus GG (LGG) exerts a major influence on the enteric nervous system (ENS) in both GF and conventional mice via contact and redox-dependent pathways involving FPR1.

Acute effects of LGG in germ free and conventional mice included an upregulation of the enteric neuronal transcription factor Hand-2 that promotes neuronal differentiation, followed by an up regulation of neuronal ChAT neurons and SERT transporter. Further, our investigations show that mutant form of LGG, viz, LGGΩSpaC, which lacks the pilin sub unit for mucosal adhesion, and LGG administered along with ROS inhibitor, NAC, has significantly reduced capacity to induce ROS, Erk 1/2 phosphorylation, and up regulation of Hand-2 and specific enteric neuronal populations. Furthermore, we demonstrate that daily LGG administration for a minimum of 1 week resulted in improved gastrointestinal motility function in conventional mice as observed from enhanced stool frequency and reduction in GI transit time (which is abrogated in presence of adhesion mutant). Ex vivo studies on ileal circular muscle strips by isometric muscle recording revealed that LGG enhances contractile responses, which corroborates our findings that LGG increases ChAT neurons via adhesion and redox dependent mechanisms. Thus, our data collectively demonstrate that microbial signals are transmitted to the enteric neurons mostly via ROS, since ROS inhibition by NAC can abolish the downstream signals of Erk 1/2 phosphorylation and changes in neurochemical coding and motility.

Hypothetically, perception of mucosal proximal bacteria such as LGG (and not the adhesion defective mutant), likely involves local diffusion of formylated peptides across a mucus barrier for perception by epithelial, enteroendocrine cells (EEC), glia or transepithelial processes of the enteric neurons ⁴. We have previously demonstrated a crucial role for FPR1 in regulating epithelial proliferation and motility ^{18, 27}. Our novel finding here is the discovery of FPR 1 and 2 in myenteric plexi, specifically identification of FPR1 expression on the enteric neurons by in situ hybridization, and the critical role of FPR1 in the regulation of enteric neuronal signaling and gut motility.

FPR expression has been documented in CNS microglia (where they also express cytokine receptors and modulate several CNS immune functions) and on murine olfactory sensory neurons where they mediate microbial chemo sensing ²³. Spasmogenic properties of fMLF and formyl peptide receptors (FPRs) has been previously reported in guinea pig ileum and found to enhance the upper GI transit in mice ³². Using FPR1 KO and FPR2 KO mice, we demonstrate that FPR1 is critical for LGG-mediated effects on GI motility via ROS induction and Erk 1/2 phosphorylation in the myenteric ganglia, while FPR2 was dispensable. In addition, we also observed that FPR1 KO mice showed a trend for higher stool wet weight and water content resembling a diarrheal phenotype (Supplementary Figure 4A & 4B), implying a role for FPR1 in anti-secretory and water absorptive functions (in addition to the motility functions) as supported by the findings that LGG also phosphorylates Erk 1/2 in the sub mucosal neurons (Supplementary Video 1) We observed that FPR2 KO mice exhibited GI motility similar to that observed in conventional mice (Supplementary Figure 5).

Thus, our study suggests the crucial role of FPR1 on transepithelial processes of neuronal cells positioned to function as informational conduits from the luminal environment to the submucosal and myenteric units of the ENS. FPR1 ligation results in activation of ROS likely via the action of NADPH oxidase and subsequent significant activation of ERK/MAPK in the enteric neuronal cells. ERK signaling in the neurons culminates in alteration of neural phenotypes in the plexi and eventuates in increased motor activity via increase in ChAT neurons, that mediate the contractile responses and facilitate peristaltic activity ³³. We found that LGG induces p-CEBP-β, which is an important transcription factor for cholinergic neurons. Selective loss of cholinergic enteric neurons and GI motility disorders has been reported in various pathologies like Parkinson’s disease/PD ^{34, 35}, multiple sclerosis ³⁶, diabetes ³⁷ and fecal incontinence leading to a poor quality of life. Hence our findings have clinical significance and applications in that LGG might be a promising probiotic in addressing constipation-like phenotype in these pathologies. Further, our current findings have significant implications as FPR1 activation on neuronal and non-neuronal cells can modulate key processes like neuronal differentiation ³⁸ and inflammatory signaling in neurodegenerative diseases ³⁹ that has direct impact on gut motility.

Further, it would be interesting to explore if LGG can normalize the post-natal enteric neuronal development in neonatal germ free mouse pups ⁴⁰. We speculate that RNA sequencing of the enteric ganglia (obtained from biopsies) from patients with diabetes, Parkinson’s disease or multiple sclerosis might help us capture an ENS-specific signature in delayed versus increased (faster) motility that could be corrected upon administration of probiotics like LGG. If successful, a personalized probiotic supplement could be tailored to match the disease pathologies of patients to correct the specific GI motility defects and improve their quality of life.

Taken together, our study reveals a dynamic dialogue between the epithelial and neuronal FPR1 and LGG via ROS/pErk 1/2 that ultimately regulates gastrointestinal motility, and the potential of the probiotic LGG-induced ROS in managing and treating gastrointestinal motility disorders associated with neurodegenerative diseases. It remains to be explored if LGG has effects on the central nervous system as probiotics have recently been shown to have psychoactive properties ^{41, 42}.