SEARCH

SEARCH BY CITATION

Keywords:

  • development;
  • Enteric nervous system;
  • microbiota;
  • postnatal;
  • small intestine

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Funding
  9. Disclosure
  10. Author Contribution
  11. References

Background

Normal gastrointestinal function depends on an intact and coordinated enteric nervous system (ENS). While the ENS is formed during fetal life, plasticity persists in the postnatal period during which the gastrointestinal tract is colonized by bacteria. We tested the hypothesis that colonization of the bowel by intestinal microbiota influences the postnatal development of the ENS.

Methods

The development of the ENS was studied in whole mount preparations of duodenum, jejunum, and ileum of specific pathogen-free (SPF), germ-free (GF), and altered Schaedler flora (ASF) NIH Swiss mice at postnatal day 3 (P3). The frequency and amplitude of circular muscle contractions were measured in intestinal segments using spatiotemporal mapping of video recorded spontaneous contractile activity with and without exposure to lidocaine and N-nitro-L-arginine (NOLA).

Key Results

Immunolabeling with antibodies to PGP9.5 revealed significant abnormalities in the myenteric plexi of GF jejunum and ileum, but not duodenum, characterized by a decrease in nerve density, a decrease in the number of neurons per ganglion, and an increase in the proportion of myenteric nitrergic neurons. Frequency of amplitude of muscle contractions were significantly decreased in the jejunum and ileum of GF mice and were unaffected by exposure to lidocaine, while NOLA enhanced contractile frequency in the GF jejunum and ileum.

Conclusions & Inferences

These findings suggest that early exposure to intestinal bacteria is essential for the postnatal development of the ENS in the mid to distal small intestine. Future studies are needed to investigate the mechanisms by which enteric microbiota interact with the developing ENS.

Key Messages
  • This study tested the hypothesis that the colonization of the intestine by microbiota influences the postnatal development of the enteric nervous system.
  • The enteric nervous system was studied in whole mount preparations of small intestine from germ-free and specific pathogen-free mice. Intestinal motility was measured using spatiotemporal mapping of video recorded spontaneous contractile motility.
  • The myenteric plexus of the jejunum and ileum of germ-free mice was structurally abnormal compared to specific pathogen-free mice, with a decrease in nerve density, a decrease in number of neuronal cell bodies per ganglion but a proportionate increase in nitrergic neurons.
  • The frequency and amplitude of muscle contractions was decreased in germ-free compared to specific pathogen-free jejunum and ileum.
  • The structural and functional abnormalities of the enteric nervous system in germ-free mice suggest that microbiota play a role in the adaptation of the enteric nervous system to the extrauterine environment.

Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Funding
  9. Disclosure
  10. Author Contribution
  11. References

The gastrointestinal (GI) tract is colonized in the first year of life by a dense mixture of microbiota.[1] This abundance of different bacterial species forms a unique relationship with its host, is critical for normal GI physiology and function, and has been shown to contribute to the breakdown of undigested or indigestible nutrients,[2, 3] synthesis of endogenous vitamins,[4] epithelial integrity and barrier function,[5] angiogenesis, and maturation of the mucosal immune system.[6, 7] Bacteria also interact with the enteric nervous system (ENS), where probiotic organisms have been shown to enhance the excitability of colonic enteric sensory neurons[8] while decreasing small intestinal motility.[9] The myenteric plexus has long been described as abnormal in adult germ-free (GF) rats[10] and these observations have recently been confirmed in GF mice.[11] A compelling scenario, in which intestinal bacteria are important for the proper functioning of the ENS, is emerging. The microbial influence on the normal development of the ENS in early life, however, remains unknown.

The ENS derives from the neural crest.[12] The neural crest-derived cells that migrate to the GI tract constitute a heterogeneous population that changes progressively as a function of developmental age, both while precursor cells are migrating and after they have reached the gut wall.[13] The development of the ENS does not stop at birth. Terminal differentiation of enteric neurons still occurs after birth, with terminally differentiated neurons co-existing with dividing neural precursor cells.[14] The maturation of enteric neurons into ganglia also persists in early postnatal life[15] with ongoing adaptation of the three-dimensional architecture of enteric nerve fibers within the first 2 weeks of life.[16] Similarly, while functional synapses and the two main classes of enteric neurons (Dogiel type I and II) can be distinguished electrophysiologically and morphologically at birth, major changes in these properties are ongoing from the postnatal period into adulthood.[17] It is reasonable to expect, therefore, that if the ENS is still being modified in postnatal life, it is still plastic and thus subject to both internal and external influences.[18]

In the current study, we test the hypothesis that intestinal bacteria are necessary for the early postnatal development of the ENS. We report that the myenteric plexus of the jejunum and ileum of GF mice is structurally abnormal on postnatal day 3 (P3)relative to specific pathogen-free (SPF) and altered Schaedler flora (ASF) colonized mice. These structural defects correspond with abnormalities in myenteric chemical coding and intestinal motility, suggesting that intestinal microbiota play an important role in the adaptation of the ENS to the extrauterine environment.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Funding
  9. Disclosure
  10. Author Contribution
  11. References

Animals

Timed pregnant NIH Swiss mice were purchased from Harlan Laboratories (Indianapolis, IN, USA) and maintained under SPF conditions. Timed pregnant GF and ASF colonized[19] gnotobiotic NIH Swiss mice were acquired from the Farncombe Axenic Gnotobiotic Unit (AGU) at McMaster University. ASF colonized offspring were derived from GF females, who were introduced to an ASF environment prior to conception. SPF and ASF colonized mice were housed in ultraclean conditions using ventilated racks; GF mice were housed in isolators in the AGU. The day a litter was found was designated as postnatal day 0 (P0). Litters from each condition were sacrificed on postnatal day 3 (P3) by exposure to CO2 gas. All mice were handled in a level II biological safety cabinet to prevent bacterial contamination. All experimental procedures were approved by the McMaster University Animal Research Ethics Board.

Tissue preparation and immunohistochemistry

To prepare whole mounts of duodenum, jejunum, and ileum, all tissues were collected immediately following sacrifice, cut open longitudinally and pinned serosal side down in individual Petri dishes. The tissues were fixed for 2 h at room temperature in a solution containing 4% phosphate-buffered formaldehyde (freshly prepared from paraformaldehyde; pH 7.4) then washed in phosphate buffered saline (PBS; pH 7.4). Laminar preparations of longitudinal muscle with adherent myenteric plexus (LMMP) were obtained by dissection of the gut wall.

For immunohistochemistry of whole mount preparations, tissues were permeabilized and blocked by incubation in PBS containing 0.4% Triton X-100 and either 4% normal horse or normal goat serum. Primary antibodies were applied overnight by incubation in 48 well plates at room temperature. Primary antibodies included rabbit antibodies to PGP9.5 (ubiquitin carboxyl-terminal hydrolase isozyme L1; dilution 1:500; AbD Serotec, Cedarlane, Burlington, ON, Canada), biotinylated mouse monoclonal antibodies to anti-human neuronal protein HuC/HuD (anti-HuC/D; dilution 1:50; Molecular Probes, Invitrogen Canada Inc., Burlington, ON, Canada), and rabbit antibodies to neuronal nitric oxide synthase (nNOS; dilution 1:500; Millipore, Cedarlane, Burlington, ON, Canada). Sites of antibody binding were detected by incubation for 2 h at room temperature with donkey anti-rabbit antibodies labeled with Alexa 594 (1:200; Molecular Probes), with streptavidin labeled with Alexa 488 (1:200; Molecular Probes) or with goat anti-rabbit antibodies labeled with Alexa 594 (1:200; Molecular Probes). In control preparations, no immunostaining was seen when primary antibodies were omitted (data not shown). Labeled tissue sections were mounted in Vectashield medium (Vector Laboratories Canada Inc., Burlington, ON, Canada) to minimize fading.

Image analysis

Immunolabeled tissue sections were viewed and photomicrographs obtained with a Retiga Imaging digital camera attached to a Leica DMRXA2 microscope (Leica Microsystems Inc., Concord, ON, Canada) and an Apple computer with Volocity software (Improvision Inc., Montreal, QC, Canada).

Nerve density in the myenteric plexi was determined by photographing at least one low power field in each of three regions of the whole mount preparation: proximal, middle, and distal. The area represented by PGP9.5 positive pixels (μm2) was quantified relative to the total area (μm2) per field using Volocity software (Improvision) and a standardized protocol. The average number of myenteric neurons per ganglion were quantified by manually counting the total number of HuC/D-positive neurons in each of 10 randomly selected myenteric ganglia across the proximal, middle, and distal regions of each whole mount preparation. Similarly, the number of nNOS-positive enteric neurons was manually counted and normalized to the total number of HuC/D-positive neurons in each of 10 ganglia chosen at random. Before analysis, all tissue preparations were coded and the investigator was blinded to their experimental condition.

Spatiotemporal mapping

Intestinal motility was measured using spatiotemporal mapping of video recordings, based on previously described protocols.[20, 21] The small intestine was removed and incubated in warm oxygenated Krebs solution (37 °C) consisting of (mmol/L): NaCl 119.8, KCL 57.7, NaHCO3 148.8, Na2HPO4·H2O 13.8, MgCl2·6H2O 11.9, CaCl2·2H2O 25.03, MgCl2·6H2O 11.9, and glucose 1.1. The small intestine was allowed to equilibrate for 30 min then separated into 3 cm segments of duodenum, jejunum, and ileum. The individual intestinal segments were then pinned at each end in an organ bath containing warm oxygenated Krebs solution, and spontaneous contractile activity was video recorded for 5 min using a camera mounted to a stereotaxic apparatus for fine positioning. Spatiotemporal maps were calculated using a custom Java algorithm developed by Dr. Sean Parsons at McMaster University.[21] Contractile frequencies and change in tissue diameter (amplitude) were derived from spatiotemporal map recordings of proximal, middle, and distal segments of each tissue and averaged.

Pharmacological agents

Lidocaine (100 μM; Sigma Aldrich, Oakville, ON, Canada) and N-nitro-L-arginine (NOLA; 200 μM; Sigma Aldrich) were introduced into the organ bath environment dissolved in the standard Krebs solution described above. Drug concentrations were selected as previously described.[21, 22] Lidocaine was used to block neural action potentials, while NOLA was used to inhibit nitric oxide production by the enzyme nitric oxide synthase (nNOS).

Statistical analysis

Data are presented as mean ± SD. Statistical analysis was performed using unpaired t-testing or one-way anova testing with post hoc Tukey's multiple comparison test where appropriate. A p < 0.05 was considered to be statistically significant.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Funding
  9. Disclosure
  10. Author Contribution
  11. References

The myenteric plexus develops abnormally in early postnatal GF mice

The myenteric plexi of SPF (n = 5), GF (n = 5) and ASF (n = 4) mice at P3 were examined in whole mount preparations of duodenum, jejunum, and ileum. The pattern of the myenteric plexus was visualized by immunolabeling with the pan-neuronal marker PGP9.5. The myenteric plexus in all small intestinal regions of SPF and ASF-colonized mice appeared in an organized lattice-like pattern, with even spacing between the ganglia and uniform thickness of the connecting fibers (Fig. 1A–C and G–I). While no prominent differences were observed between the duodenum of SPF, GF, and ASF mice, the GF jejunum and GF ileum appeared abnormal with unevenly spaced ganglia and thinner, less abundant connecting nerve fibers (Fig. 1D–F).

image

Figure 1. The myenteric plexus appears structurally abnormal in P3 germ-free (GF) mice. Myenteric nerves were visualized by immunolabeling with antibodies to PGP9.5 (red). (A–C) The myenteric plexus in the specific pathogen-free (SPF) duodenum, jejunum, and ileum appears organized in a lattice-like network, with even spacing between ganglia and uniform thickness of connective nerve fibers. (D) The myenteric plexus in the GF duodenum resembles that of SPF duodenum. (E and F) In GF mice, the myenteric plexus of the jejunum and ileum appears unorganized, with fewer ganglia and thinner connecting nerve fibers. (G–I) In altered Schaedler flora (ASF)-colonized mice, the structure appears similar to that observed in SPF-colonized animals. (J–L) The nerve densities of the myenteric plexus were quantified in whole mount preparations of SPF (n = 5), GF (n = 5), and ASF (n = 4) small intestine segments. The nerve densities, relative to SPF and ASF mice, were significantly decreased in the GF jejunum and ileum. Bar = 120 μm; *p < 0.05; ***p < 0.001; ns, no significance.

Download figure to PowerPoint

The nerve densities of PGP9.5 immunolabeled SPF, GF, and ASF duodenum, jejunum, and ileum were quantified using image analysis software and expressed as a percentage of total tissue area in μm2. In the duodenum (Fig. 1J), no significant differences were found between the nerve densities in SPF (28.8 ± 2.4), GF (30 ± 2.1), and ASF (28.4 ± 1.9) mice (anova, p = 0.45). In the jejunum (Fig. 1K), similar to our morphological observations, the nerve density was significantly decreased in GF relative to SPF (23.7 ± 1.6 vs 31.4 ± 2; p < 0.001) and to ASF animals (23.7 ± 1.6 vs 28.4 ± 1.7; p < 0.01). In the ileum (Fig. 1L), the nerve density was again significantly decreased in the GF relative to SPF (29.6 ± 1.4 vs 34.5 ± 1.5; p < 0.001) and to ASF mice (29.6 ± 1.4 vs 32.5 ± 1.1; p < 0.05).

Myenteric ganglia contain fewer neurons in early postnatal GF mice

Myenteric ganglia were visualized by immunolabeling neuronal cell bodies with antibodies to HuC/D (Fig. 2A–I). In the duodenum (Fig. 2J), the average number of myenteric neurons per ganglion in SPF (57.8 ± 18), GF (58 ± 23.4), and ASF mice (61 ± 11.6) was not significantly different (anova, p = 0.96). In the jejunum (Fig. 2K), the average number of myenteric neurons per ganglion was significantly decreased in the GF relative to SPF (32.4 ± 8.3 vs 52.6 ± 14.8; p < 0.01) and to ASF mice (32.4 ± 8.3 vs 51 ± 15.6; p < 0.01). Similarly, in the ileum (Fig. 2L), the average number of myenteric neurons per ganglion was significantly decreased in the GF relative to SPF (34 ± 7.9 vs 65.6 ± 12.7; p < 0.001) and ASF animals (34 ± 7.9 vs 64.1 ± 7.0; p < 0.01).

image

Figure 2. Myenteric ganglia contain fewer neurons and a higher proportion of nitrergic neurons in P3 germ-free (GF) mice. Total myenteric and nitrergic neurons were visualized by immunolabeling with antibodies to HuC/D (green) and nNOS (red), respectively, in specific pathogen-free (SPF; A–C), GF (D–F) and altered Schaedler flora (ASF; G–I) whole mount preparations of small intestine. (J–L) The average number of total HuC/D-positive myenteric neurons per ganglion in the GF jejunum and ileum were significantly decreased in comparison with SPF- and ASF-colonized mice, with no differences observed in the duodenum. (M–O) No significant differences in cell body size (μm2) were observed between SPF and GF gut regions. (P–R) The average number of nNOS-positive neurons per myenteric ganglion was significantly increased in the GF jejunum and ileum in comparison with SPF- and ASF-colonized mice, with no differences observed in the duodenum. SPF, n = 5; GF, n = 5; ASF, n = 4. Bar = 20 μm. *p < 0.05; **p < 0.01; ***p < 0.001; ns, no significance.

Download figure to PowerPoint

Cell body size (μm2) was measured using Volocity software. No differences were observed in the average cell body size of myenteric neurons in the SPF and GF duodenum (80.8 ± 4.1 vs 77.2 ± 4.1; p = 0.56; Fig. 2M), jejunum (78.5 ± 1.3 vs 82.3 ± 2.5; p = 0.22; Fig. 2N), and ileum (75.9 ± 4.9 vs 86.8 ± 3.1, p = 0.11; Fig. 2O).

The proportion of nitrergic myenteric neurons is increased in early postnatal GF mice

Myenteric nitrergic neurons were visualized by immunolabeling with antibodies to nNOS (Fig. 2A–I). The number of nNOS-immunoreactive neurons was manually counted and normalized to the total number of HuC/D-immunopositive neurons and expressed as a percentage of total neurons per ganglion. In the duodenum (Fig. 2P), the average proportions of nNOS-immunoreactive neurons per ganglion under SPF (25.5 ± 2.9), GF (26.9 ± 2.3) and ASF (25 ± 1.2) conditions were not found to be significantly different (anova, p = 0.45). In the jejunum (Fig. 2Q), the average proportion of nitrergic neurons per ganglion was significantly increased in the GF relative to SPF (31.2 ± 3.2 vs 24 ± 1.8; p < 0.01) and to ASF mice (31.2 ± 3.2 vs 23.8 ± 1.7; p < 0.01). In the ileum (Fig. 2R), the average proportion of nitrergic neurons per ganglion was significantly increased in the GF relative to SPF (36 ± 1.8 vs 23.3 ± 1.2; p < 0.001) and to ASF animals (36 ± 1.8 vs 24.7 ± 2.2; p < 0.001).

Intestinal motility is decreased in early postnatal GF mice

Duodenum, jejunum, and ileum were taken from SPF, GF, and ASF colonized mice at P3 and individually placed in an organ bath with warm Krebs solution. The frequency and amplitude of circular muscle contractions were measured using spatiotemporal mapping of video recorded spontaneous contractile activity. In the duodenum (Fig. 3A), the average frequency of circular muscular contractions per minute between SPF (20 ± 1.4), GF (20.2 ± 1.8), and ASF (19.7 ± 1.5) conditions was not found to be significantly different (anova, p = 0.91). Similarly, no significant difference (anova, p = 0.95) was seen between the average amplitude of circular muscle contractions in the duodenum (Fig. 3D) of SPF (1.77 ± 0.22), GF (1.81 ± 0.26), and ASF mice (1.78 ± 0.06). In the jejunum (Fig. 3B), however, the frequency of circular muscle contractions was significantly decreased in the GF compared with SPF (10.2 ± 1.6 vs 16.8 ± 1.3; p < 0.001) and ASF mice (10.2 ± 1.6 vs 16.3 ± 1.5; p < 0.001). The amplitude of circular muscle contractions (Fig. 3E) was also significantly decreased in the GF compared with SPF (1.1 ± 0.26 vs 1.73 ± 0.21; p < 0.01) and ASF jejunum (1.1 ± 0.26 vs 1.61 ± 0.06; p < 0.05). In the ileum (Fig. 3C), the frequency of circular muscle contractions was significantly decreased in the GF compared with SPF mice (10 ± 1 vs 14.2 ± 1.3; p < 0.01) mice but no significant difference was seen between GF and ASF animals (10 ± 01 vs 12.7 ± 2.1; p > 0.05). The amplitude of circular muscle contractions (Fig. 3F) was significantly decreased in the GF compared with SPF (0.96 ± 0.28 vs 1.72 ± 0.14; p < 0.001) and ASF ileum (0.96 ± 0.28 vs 1.45 ± 0.11; p < 0.05).

image

Figure 3. Small intestinal motility is abnormal in P3 germ-free (GF) mice. Average frequency and amplitude of circular muscle contractions were determined using spatiotemporal mapping of video recorded spontaneous contractile activity in segments of specific pathogen-free (SPF; n = 5), GF (n = 5) and altered Schaedler flora (ASF; n = 4) small intestine. (A–C) The frequency of muscle contractions was significantly decreased in the jejunum and ileum of GF mice in comparison with SPF- and ASF-colonized intestinal segments, with no differences observed in the duodenum. (D–F) The amplitude of muscle contractions was significantly decreased in the GF jejunum and ileum in comparison with SPF- and ASF-colonized intestinal segments, with no differences observed in the duodenum. *p < 0.05; **p < 0.01; ***p < 0.001; ns, no significance.

Download figure to PowerPoint

Intestinal motility is unaffected by lidocaine in GF mice

Neural activity was blocked by adding the sodium channel antagonist, lidocaine, to the organ bath environment (Fig. 4).The average frequency of muscle contractions per minute was significantly decreased in the lidocaine-treated relative to non-treated SPF duodenum (13.0 ± 2.9 vs 20.0 ± 1.8; p < 0.05; Fig. 4A), jejunum (7.8 ± 0.84 vs 16.8 ± 1.3; p < 0.001; Fig. 4B) and ileum (8.4 ± 1.8 vs 14.2 ± 1.3; p < 0.001; Fig. 4C). The average amplitude of circular muscle contractions was also significantly decreased in the lidocaine-treated relative to non-treated duodenum (1.2 ± 0.2 vs 1.8 ± 0.2; p < 0.05; Fig. 4A), jejunum (1.1 ± 0.1 vs 1.7 ± 0.2; p < 0.05; Fig. 4B) and ileum (1.1 ± 0.2 vs 1.7 ± 0.1; p > 0.05; Fig. 4C). The GF duodenum, in which we describe motility parameters similar to SPF conditions, also had a significant reduction in frequency of muscle contractions (12.4 ± 2.3 vs 20.2 ± 1.8; p < 0.05; Fig. 4A) and in amplitude of circular muscle contractions (1.2 ± 0.2 vs 1.8 ± 0.3; p < 0.05; Fig. 4D) with exposure to lidocaine. No reduction in the frequency of muscle contractions, however, was seen between lidocaine treated and non-treated GF jejunum (12.2 ± 0.86 vs 10.2 ± 1.6; p > 0.05; Fig. 4B) and ileum (11.2 ± 1.2 vs 10.0 ± 1.0; p > 0.05; Fig. 4C). No reduction in amplitude of circular muscle contractions was seen between lidocaine treated and non-treated jejunum (1.3 ± 0.2 vs 1.1 ± 0.3; p > 0.05; Fig. 4E) and ileum (1.1 ± 0.2 vs 1.0 ± 0.3; p > 0.05; Fig. 4F).

image

Figure 4. Small intestinal motility is unaffected by general neuronal blockade, but enhanced by specific nitrergic blockade in P3 germ-free (GF) mice. Lidocaine (‘L’; 100 μM) significantly inhibited both the frequency (A–C) and amplitude (D–F) of muscle contractions in the specific pathogen-free (SPF) duodenum, jejunum, ileum, and GF duodenum, but failed to influence either parameter in the GF jejunum or ileum. NOLA (‘N’) had no effect on the frequency of muscle contractions in SPF small intestine (A–C) but significantly increased the frequency of contraction in GF jejunum and ileum (B and C). SPF, n = 5; GF, n = 5. *p < 0.05; **p < 0.01; ***p < 0.001; ns, no significance.

Download figure to PowerPoint

Intestinal motility is enhanced by nitrergic blockade in GF mice

Nitric oxide production by nNOS was inhibited by adding NOLA to the organ bath environment. In the duodenum (Fig. 4A), no significant differences (anova, p = 0.61) were seen in the average frequency of muscle contractions between the NOLA-treated and non-treated SPF (19.5 ± 1.3 vs 20 ± 1.4) and GF mice (18.8 ± 2.2 vs 20.2 ± 1.8). In the jejunum (Fig. 4B), however, while there was no significant difference between the NOLA-treated and non-treated SPF mice (16.3 ± 1.7 vs 16.8 ± 1.3; p > 0.05), there was a significant increase in the average frequency of muscle contractions between the NOLA-treated and non-treated GF mice (15.3 ± 2.8 vs 10.2 ± 1.6; p < 0.01). Similarly, in the ileum (Fig. 4C), while there was no significant difference between the NOLA-treated and non-treated SPF animals (14.3 ± 2.6 vs 14.2 ± 1.3; p > 0.05), there was a significant increase in the frequency of muscle contractions between the NOLA-treated and non-treated GF mice (14.3 ± 2.2 vs 10.0 ± 1.0; p < 0.05).

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Funding
  9. Disclosure
  10. Author Contribution
  11. References

This study tested the hypothesis that the colonization of the GI tract by intestinal microbiota influences the postnatal development of the ENS. In P3 GF mice, we found that the myenteric plexus of the jejunum and ileum was abnormally patterned with an overall decrease in nerve density when compared with equivalent segments of small intestine from SPF-colonized animals. Myenteric ganglia in GF mice, furthermore, contained fewer neuronal cell bodies than those found in SPF-colonized small intestine with an increase in the proportion of inhibitory nitrergic neurons. Spatiotemporal mapping of small intestinal motility in early postnatal GF mice revealed decreased frequency and amplitude of muscle contractions that were unaffected by the general inhibition of neural activity but enhanced by the blockade of nitric oxide production. Colonization of mice with ASF flora was sufficient to produce a morphological and functional phenotype similar to that seen in SPF animals.

We overall found a gradient of changes in the SPF, ASF, and GF small intestine. In SPF and ASF-colonized mice, the myenteric plexus in the duodenum, jejunum, and ileum was organized into a lattice-like network, with increasing density of ganglia and connective fibers in a proximal to distal gut gradient, consistent with previous findings in conventional rodents.[16] In GF mice, the myenteric plexus of the jejunum and ileum were characterized by less overall organization, smaller ganglia, and less abundant connective nerve fibers between ganglia. These changes in ENS structure were specific to the jejunum and ileum with no changes observed in the duodenum, suggesting that the gradient of increasing microbial abundance in the gut[23] may be an important influence of microbial-enteric interactions. The number of constituent neuronal cell bodies within myenteric ganglia was also found to be decreased in the jejunum and ileum of GF mice.

The defects observed in the affected gut regions in P3 GF mice are similar to genetic mouse models in which specific factors known to be important in ENS developments are deleted. Deletion of Hand2, for example, results in defects in neurite outgrowth and a significant decrease in myenteric neuron numbers,[24] similar our observations in GF mice. Defects in enteric neuron aggregation, furthermore, have been described in mice in which the normal polysialylation of neural cell adhesion molecule by bone morphogenetic proteins has been blocked.[15] It is currently unclear how endogenous microbiota in the GI lumen might interact with the developing postnatal ENS to influence the process of ganglia formation, but from our observations, one possibility could include downstream effects on factors already known to be important in ENS formation. These downstream effects, while unknown at this time, seem to have lasting consequences, as the abnormalities in the ENS of GF animals persist into adulthood, with reductions in total neurons observed in 4-week old GF mice.[25] Earlier studies have shown that the myenteric plexus in the cecum of adult GF rodents is abnormal, with large and metabolically less active enteric neurons.[10] While we did not observe any differences in the size of myenteric neurons in the small intestine of GF mice, it is possible that the consequences of a GF environment on enteric postnatal development change over time.

The ENS has an important contribution to intestinal motility[26, 27] and abnormal motility patterns have previously been described in adult GF rodents.[28] We tested the hypothesis, therefore, that the morphological defects that we observed in GF mice would correspond to abnormalities in intestinal motility. Indeed, we found that the GF jejunum and ileum, in which significant defects in ENS patterning and nerve density were observed, also displayed reductions in parameters of intestinal motility. To further explore the extent to which abnormalities in neuronal function related to altered motility, we exposed SPF and GF small intestine to lidocaine, which blocks sodium channels and thus inhibits action potential formation in a manner similar to that of tetrodotoxin.[21] In these experiments, we found a significant inhibitory effect on SPF tissues, but not on GF jejunum or ileum. Our data suggest that the contribution of the intrinsic innervation of the gut to smooth muscle activity is dysfunctional in early postnatal GF mice, as inhibition of neurotransmission fails to influence the observable motility patterns in these animals. These findings are consistent with previous reports of delayed intestinal motility in GF rodents,[28] and provide novel insight into the early life onset of these defects.

While the structure of the myenteric plexus is subjected to dynamic change during the early weeks of life,[16] so too is the emergence of specific neurochemical phenotypes of myenteric neurons.[29, 30] In addition to the abnormal motility patterns in our early postnatal GF mice, we also observed an increased proportion of myenteric neurons immunolabeled with nNOS, an important neurotransmitter found in the inhibitory motor neurons that innervate both the circular and longitudinal muscle of the intestine.[31] Our finding of a proportional increase in nitrergic neurons might seem to be in contrast with recent work in which the total number of nNOS-immunoreactive neurons was found to be reduced in 4-week old GF mouse colon,[25] but there are several possibilities to account for the difference in observations. As discussed earlier, it is conceivable that there is a gradient of consequences from the lack of microbial-enteric interactions along the GI tract and that some of the downstream effects change between early postnatal life to adulthood; consequences in the small intestine, therefore, might be different than those manifest in the colon. In the study of 4-week old GF mouse colon, furthermore, the decrease in nNOS-immunoreactive neurons was expressed as an absolute number and it is unclear to what extent the decrease is related to their reported overall reduction in enteric neurons in the same specimens.[25] We had chosen to express the number of nNOS-immunoreactive cells as a proportion of myenteric neurons to account for the overall decrease in myenteric neurons per ganglion observed in the P3 GF mice. It should be noted that had we have chosen to quantify the number of nNOS-immunoreactive neurons as an absolute number, rather than normalizing to the number of HuC/D-immunoreactive neurons, we may also have found an absolute decrease in total nitrergic neurons between GF and SPF animals. Our finding of a proportionate increase, moreover, does not necessarily imply that the lack of microbiota causes an increase in nitrergic neurons but could also represent that the population of nitrergic neurons is relatively spared from the observed overall decrease in myenteric neurons. Either way, our finding of a proportionate increase in inhibitory nitrergic neurons seems to hold functional significance as the pharmacological blockage of nNOS with NOLA was shown to enhance the frequency of muscle contractions in the GF jejunum and ileum, in which the increased proportions of nNOS-immunoreactive neurons were observed. We recognize that our studies did not include an exploration of the possible influence of microbial-enteric interaction on the expression of other neurotransmitters, such as acetylcholine, and feel that this would be an appropriate area of further study. In general, however, our findings, are consistent with previous work in which the enteric microenviroment was found to have an influence on the chemical coding of the ENS.[9, 32, 33]

In conclusion, our observations demonstrate that a lack of exposure to intestinal bacteria early in life results in abnormal structural development of the ENS that is detectable at P3. These structural abnormalities involve changes in the proportional expression of inhibitory neurons within the myenteric plexus that may in part explain the abnormal motility patterns observed in these mice. These defects in the ENS do not appear to be lethal, as GF mice do survive to adulthood, but do so with abnormal physiology and growth.[34] Future studies are needed to investigate the mechanisms in which enteric microbiota interact with the developing ENS and influence the development of intestinal motility early in life.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Funding
  9. Disclosure
  10. Author Contribution
  11. References

The authors would like to thank the AGU staff at McMaster University for providing and maintaining GF animals, Dr. Sean Parsons at McMaster University for sharing customized software for analysis and Kal Mungovan for assistance with data analysis and manuscript preparation.

Funding

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Funding
  9. Disclosure
  10. Author Contribution
  11. References

This work was supported by grants from the Hamilton Health Sciences New Investigator Fund (NIF-11267; EMR) and from the McMaster Children's Hospital (EMR).

Author Contribution

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Funding
  9. Disclosure
  10. Author Contribution
  11. References

JC designed the research study, performed the research, analyzed the data, and wrote the manuscript; RB designed the research study, performed the research, and wrote the manuscript; EFV provided the germ-free mice and valuable advice; JDH assisted in experimental design and contributed essential tools; EMR designed the research study, analyzed the data, and wrote the manuscript.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Funding
  9. Disclosure
  10. Author Contribution
  11. References