Article Text

other Versions

Original article
Nerve activity recordings in routine human intestinal biopsies
  1. Carla Cirillo,
  2. Jan Tack,
  3. Pieter Vanden Berghe
  1. Laboratory for Enteric Neuroscience (LENS), TARGID, KU Leuven, Leuven, Belgium
  1. Correspondence to Professor Pieter Vanden Berghe, Laboratory for Enteric Neuroscience (LENS), TARGID, KU Leuven, O&N 1 Herestraat 49 – box 701, Leuven 3000, Belgium; pieter.vandenberghe{at}


Background Most direct understanding of enteric nerve (patho)physiology has been obtained by electrode and imaging techniques in animal models and human surgical samples. Until now, neuronal activity recordings from a more accessible human tissue source have remained a true challenge.

Objectives To record nerve activity in human intestinal biopsies using imaging techniques.

Design Submucous plexus was isolated from duodenal biopsies. Enteric nerves were functionally and morphologically examined using calcium (Ca2+) imaging and immunohistochemistry. Exogenous application of high-K+ solution, the nicotinic cholinergic receptor agonist (1,1-dimethyl-4-phenylpiperazinium; DMPP) or serotonin (5-HT), and electrical stimulation of interganglionic fibre tracts were used to activate the neurons, and intracellular Ca2+ concentrations ([Ca2+]i) were monitored. Enteric ganglia were stained with neuronal and glial markers.

Results Using high-K+ solution, 146 neurons were identified in 70 ganglia (44 biopsies from 29 subjects). The exogenous application of DMPP or 5-HT caused a transient [Ca2+]i increase, respectively, in 68% and 63% of the neurons identified by high-K+. Electrical stimulation evoked responses in 57% of the neurons; these responses were totally or partly suppressed by tetrodotoxin or zero-Ca2+ solution, respectively. Immunohistochemical analysis showed both isolated neurons and ganglia interconnected by typical interganglionic fibre bundles. The average number of ganglia was 7.7±6.0 per biopsy and each ganglion contained on average 4.5±1.2 neurons.

Conclusion In this study, for the first time, live recordings were performed of nerve activity in intestinal biopsies. This novel approach is of key importance to study living neurons in both health and disease and to test newly developed compounds in an in-vitro human tissue model.

  • Achalasia
  • appetite
  • calcium imaging
  • dyspepsia
  • enteric nervous system
  • enteric neurones
  • enteric neurons
  • functional bowel disorder
  • functional dyspepsia
  • gastric emptying
  • gastroduodenal motility
  • gastroesophageal reflux disease
  • gastrointesinal endoscopy
  • gastrointestinal motility
  • human
  • imaging
  • immunohistochemistry
  • intestinal biopsies
  • intracellular signalling
  • irritable bowel syndrome
  • neurobiology
  • neurogastroenterology
  • neurophysiology
  • optical recording
  • visceral sensitivity

Statistics from

Significance of this study

What is already known about this subject?

  • The ENS is affected in patients with gastrointestinal and neurological disorders.

  • Optical imaging of nerve activity is a powerful technique to gain better knowledge of gastrointestinal pathophysiology.

  • Surgical samples, which were the sole source to study human enteric neurons until now, are limited in availability and are restricted to patients with severe diseases.

What are the new findings?

  • Live recordings of nerve activity can be successfully performed in small samples such as intestinal biopsies.

  • This novel approach allows characterising human enteric nerves in health and disease.

How might it impact on clinical practice in the foreseeable future?

  • We provide a novel tool to advance the understanding of ENS abnormalities in gastrointestinal diseases.

  • The use of human tissues is directly translational, as the efficacy of newly developed compounds can be verified in vitro even before clinical testing.

In recent years, the interest in using imaging techniques to record activity from the enteric nervous system (ENS) has increased substantially due to the need for validation and translation of results from animal models into humans. Most of what is known about enteric nerves so far is derived from animal studies.1–5 In order to translate these seminal research findings into clinically relevant knowledge, it is of key importance that, at some stage, human freshly dissected intestinal tissues can also be used successfully.6

In humans, several studies have examined the mechanisms underlying intestinal secretion and motility by recording nerve activity from myenteric and/or submucous preparations from surgical tissues.7–13 Optical recordings of neuronal activity using either voltage or calcium (Ca2+)-sensitive dyes have been shown to be a valid tool to record simultaneously from several neurons in animal tissues.1–3 14–16 In the past decade these imaging techniques were also used successfully in human surgical tissues to record structural aspects, as well as neuronal activity in the ENS.7 8 13–16 These reports have provided new remarkable insights into comprehending ENS (dys)function observed in several gut disorders, and have made an important step forward in translating animal experiments into human studies. Nonetheless, they suffer from the fact that tissue accessibility is restricted to patients with severe cases of intestinal diseases who therefore needed surgery. Moreover, the serious conditions leading to surgical intervention often complicated the interpretation of the results.

The most accessible and safest way to obtain human enteric neurons, specifically those of the submucous layer, is via mucosal biopsies, which are routinely taken for histological evaluation during endoscopic examination. Two pioneering studies have recently used submucous plexus, isolated from colonic mucosal biopsies, to characterise enteric neurons morphologically and to investigate the expression and aggregation of α-synuclein in Parkinson's disease patients, respectively.17 18 However, until now it remained a challenge to record reliably from human enteric nerves in these ‘easily accessible’ biopsies. Nonetheless, such an approach is necessary to advance further the current knowledge of human enteric neurobiology in general, and of submucous plexus pathophysiology in particular. Furthermore, this will also help to identify potential therapeutic targets and to test newly developed compounds at an early stage before true clinical use.

Given the need for such a live approach, we first aimed to verify whether intestinal biopsies taken during endoscopy would contain enough viable enteric nerves. Second, we explored whether these biopsies would be suitable to apply imaging techniques to record activity directly from submucous neurons. With some adaptations to the existing technique, we were able to measure neuronal Ca2+ transients elicited by selective receptor agonists or electrical stimulation.

Materials and methods


In this study we used biopsies from 48 subjects (17 men, mean age 54±19 years and 31 women, mean age 49±18 years) who were referred to our gastroenterology unit for clinical and endoscopic evaluation. The subjects included in the study were undergoing endoscopy for presumed functional disorders. No gastrointestinal lesions, whether inflammatory or neoplastic, were observed during the course of the endoscopy. The study protocol (ML7400) was approved by the Ethics Committee of Leuven University Hospital, Belgium. Written informed consent was obtained from each subject.

Endoscopy of the upper gastrointestinal tract

Two duodenal biopsies were taken by experienced endoscopists using standard biopsy forceps with needle (Micro-Tech single-use biopsy forceps with pin NBF01-11123180, diameter 2.3 mm, forceps opening width 6.7 mm; Micro-Tech, Nanjing, China). After removal, the biopsies were immediately immersed in ice-cold (4°C) Krebs solution (in mM: 120.9 NaCl, 5.9 KCl, 1.2 MgCl2, 2.5 CaCl2, 11.5 glucose, 14.4 NaHCO3 and 1.2 NaH2PO4) previously oxygenated (95% oxygen/5% carbon dioxide), and kept on ice for no more than 1 h until dissection. Biopsies were subsequently carefully stretched and pinned flat in a Sylgard-lined Petri dish (figure 1A) and dissected under a stereomicroscope while continuously perfused with oxygenated (95% oxygen/5% carbon dioxide) ice-cold Krebs solution. The inner submucous layer (figure 1B) was carefully removed from the mucosa using watchmaker's forceps. This fragile tissue was then gently stretched and pinned flat in a special recording chamber (our design) in which in and outflow volume could be tightly controlled. In the preliminary phase of the study, we evaluated the success rate of having sufficient submucous plexus in biopsies taken with forceps with needle compared with biopsies taken with forceps without needle. Similar success rates were compared between biopsies taken from ‘folded’ and ‘distended’ areas of the duodenum. We found that the best submucous plexus preparations were obtained from biopsies taken with forceps with needle from the ‘folded’ areas of the duodenum.

Figure 1

Duodenal mucosal biopsies contain submucous plexus and ganglia. (A) Duodenal mucosal biopsies (∼5 mm2) were obtained from subjects undergoing gastrointestinal endoscopy. Scale bar: 2 mm. (B) Picture showing a typical submucous layer carefully dissected from the mucosa using watchmakers' forceps. Scale bar: 1 mm. (C) Differential interference contrast image of an identified ganglion. The dotted line indicates the edges of the ganglion. Scale bar: 20 μm. (D) Second harmonic (SH) image of the same ganglion in (C) visualising the surrounding collagen matrix (red). Scale bar: 20 μm. (E) Immunostaining of the ganglion (green, Hu C/D) embedded in the collagen matrix (red, collagen SH). Scale bar: 50 μm. (F) Three-dimensional image of ganglia in the submucous plexus stained with Hu C/D (green) and NF200 (red). Scale bar: 50 μm.

Visualisation of human enteric ganglia

The human submucous plexus ganglia were easily recognised under differential interference contrast (DIC) optics because of the high degree of interference generated by the heavily coiled interconnecting fibre tracts (figure 1C). The typical structure of collagen matrix in which the ganglia are embedded is readily visible under DIC but also using second harmonic (SH) imaging (figure 1D). SH images were generated on a Zeiss LSM510 (Jena, Germany) multiphoton microscope with a pulsed MaiTai laser (Spectraphysics, Newbury, UK) by sending 800 nm light onto the sample, SH generated by collagen were filtered (417/60 nm) and captured on a non-descanned detector (Cell Imaging Core, KU Leuven, Belgium).

Freshly dissected submucous preparation loading

After removal, submucous layers were loaded with the fluorescent Ca2+ indicator Fluo-4 AM (Molecular Probes, Invitrogen, Merelbeke, Belgium) to measure the relative changes in cytosolic Ca2+ concentration. To identify optimal loading conditions, we varied Fluo-4 (1, 5, 10 μM) and Cremophor EL (surfactant agent; Fluka Chemika, Buchs, Switzerland; 0–0.01%) concentrations as well as loading time (30, 60, 120 min). All experimental procedures were carried out at room temperature. As the oxygen supply is critical, special care was taken to perfuse the submucous preparations continuously during Fluo-4 loading.

Ca2+ imaging in human submucous plexus

After loading, the recording chamber was mounted onto an upright Zeiss Examiner microscope equipped with a 20× (NA 1) water dipping lens and coupled to a monochromator (Poly V) and cooled CCD camera (Imago QE) both from TILL Photonics (Gräfelfing, Germany). A gravity-fed perfusion system ensured continuous and constant perfusion (1 ml/min) of the preparation with 95% oxygen/5% carbon dioxide-gassed Krebs solution (at room temperature) and excess solution was removed via a peristaltic suction pump, which kept the experimental volume constant (3 ml). Fluo-4 was excited at 475 nm, and its fluorescence emission was collected at 525/50 nm. Images (640×512 pixels2) were acquired at 2 Hz.

Fibre tracts were electrically stimulated by a platinum electrode (diameter 25 μm), which was guided by a mechanical micromanipulator (Narishige, London, UK) and placed on a fibre tract 0.3–0.7 mm away from the ganglion of interest. Trains (2 s, 20 Hz) of 300 μs electrical pulses (Grass Instruments, Quincy, Massachusetts, USA) were used. To confirm the neuronal origin of the signal, we conducted experiments in the presence of tetrodotoxin (1 μM, Sigma, Bornem, Belgium). To investigate the primary source of Ca2+ in the submucous neurons we also tested the electrical stimulation in the absence of extracellular Ca2+ (2 mM EGTA). To study the effect of post-synaptic stimuli, neuronal receptor agonists, such as nicotinic cholinergic receptor agonist 1,1-dimethyl-4-phenylpiperazinium (DMPP, 10 μM, Fluka Chemika) and serotonin (5-HT; 10 μM, Sigma) were applied via a local perfusion pipette for 20 s. Neurons were only included in the Ca2+ analysis when they displayed a sharply increasing Ca2+ response to high-K+ perfusion.

Images were collected using TILLVision software (TILL Photonics) and stored on a personal computer. Further analysis was done using custom-written macros in IGOR PRO (Wavemetrics, Lake Oswego, Oregon, USA). To remove drift and movement artefacts due to perfusion, the image stack was registered to the first image. Regions of interest were drawn over each cell, fluorescence intensity was normalised to the basal fluorescence at the onset of the recording for each region of interest (ΔF/F0), and peaks were analysed. A peak was considered if the signal rose above baseline plus five times the intrinsic noise level. The percentage of responsive cells and the maximum intracellular Ca2+ concentrations ([Ca2+]i) peak amplitude and duration (at t50%) were determined.


After Ca2+ imaging, all submucous preparations were fixed for 30 min at room temperature in phosphate buffered saline (PBS) with 4% paraformaldehyde. After fixation, the samples were rinsed 3×10 min with PBS and incubated for 2 h at room temperature in blocking buffer containing PBS, 1% Triton X-100, 2% goat serum and 2% donkey serum. Incubation with primary antibodies chicken anti-neurofilament 200 kD (NF200, 1:500; Abcam, Cambridge, UK), mouse anti-panneuronal HuC/D (1:200; Molecular Probes, Invitrogen) and rabbit anti-S100 (1:500; Dako, Glostrup, Denmark) diluted in blocking buffer was carried out overnight at 4°C. Following incubation with primary antibodies, submucous specimens were then incubated for 2 h with goat anti-chicken Alexa fluor 594 (1:1000; Molecular Probes, Invitrogen), donkey anti-mouse Alexa fluor 488 (1:1000; Molecular Probes, Invitrogen) and donkey anti-rabbit AMCA (1:500; Vector Labs Ltd., Peterborough, UK). Control experiments were performed by omitting the primary antibodies. After 3×10 min washes, submucous specimens were mounted on a microscope slide in Citifluor (Citifluor Ltd., Leicester, UK). Immunohistochemical stainings were visualised under an epifluorescence microscope (BX 41 Olympus, Belgium) with U-MNUAUV, U-MWIBA3 and U-MWIY2 filtercubes for visualising blue, green and red probes, respectively. Images were recorded using Cell^F software on an XM10 (Olympus) camera. Pictures were adjusted for contrast and brightness before overlay and quantification. Each preparation was entirely screened to count ganglia and neurons. Confocal images were recorded using a Zeiss LSM510 Meta confocal microscope (Cell Imaging Core, KU Leuven, Belgium).

Data and statistical analysis

The proportions of neurons responding to different stimuli were determined relative to all neurons identified by high K+ depolarisation and compared with χ2 tests. All results are presented as mean±SEM. The amplitudes and durations of the responses in the presence of tetrodotoxin or absence of Ca2+ were compared with those registered after electrical stimulus application. The n values include successful recordings in which neurons were responsive to high-K+ solution. Responses to electrical stimulation or agonists (DMPP and 5-HT) include all neurons that had an initial response to high-K+ solution. Gender effects were compared with a Student's t test and age effects with analysis of variance.


Recognition of submucous ganglia in human biopsies

After careful dissection of the biopsies (average size 5±1 mm2, figure 1A,B) and using DIC optics, submucous ganglia were relatively easily recognised in the connective tissue of the preparation (figure 1C). No vital stain was necessary due to the high contrast generated by the strongly coiled fibre tracts. As an alternative technique we also used SH imaging, a label-free method that is capable of visualising the surrounding collagen matrix directly (figure 1D). This non-linear imaging technique can either be applied to live tissues or after fixation and in combination with immunohistochemistry (figure 1E).

Quantification of enteric ganglia and neurons

The human submucous plexus architecture is quite typical as it comprises both isolated neurons and small ganglia, which connect in three or four directions to each other by fasciculated fibre bundles (figure 1F). Immunohistochemical analysis of 72 biopsies from 48 subjects showed the presence of 7.7±6.0 ganglia per biopsy (579 ganglia in total, see also table 1). We also identified a total of 2607 neurons (Hu C/D positive) in all labelled ganglia, which are on average 4.5±1.2 neurons per ganglion (figure 2). Moreover, we identified 184 isolated neurons in total with an average of 2.1±1.8 per biopsy. Besides neurons, S100 immunostaining also revealed glial cells inside the ganglia while some stained glial cells were also located in the broad fibre bundles extending from the ganglia (figure 2).

Table 1

Quantitative characteristics of duodenal biopsies

Figure 2

Ganglia in the human submucous plexus are characterised by neurons and glial cells. Composite of 3 by 3 images representing neurons and glial cells, as well as a merged image. Submucous ganglia were immunostained for the pan-neuronal marker Hu C/D (green, examples of Hu C/D positive neurons are indicated by asterisks) and the glial marker S100 (blue, examples of S100 positive glia cells are indicated by arrows). Merge of blue and green images. Note that glial cells entangle neurons. The images were taken at different planes (2 μm apart and recorded with one Airy unit pinholes). All scale bars: 20 μm.

Effect of age and gender on the number of submucous enteric neurons

In all biopsies used for immunohistochemical analysis, we checked whether gender and age had an effect on the number of ganglia per biopsy and of neurons per ganglion, as identified by positive Hu C/D staining. In general, we found no statistically significant difference between men and women (figure 3A). In particular, the analysis revealed that the number of ganglia per biopsy was 6.9±4.4 in men (189 in 17 men) and 8.3±6.6 in women (390 in 31 women; p>0.05) and the average number of neurons per ganglion was 4.6±1.6 in men and 4.4±1.2 in women (p>0.05; figure 3A). As for the number of neurons per biopsy, we recognised an average of 32.6±21.3 in men and 37.0±29.7 in women (p>0.05).

Figure 3

Number of submucous neurons per ganglion in relation to age and gender of the subjects. (A) The average number of neurons/ganglion subdivided based on subject gender (men: left and women: right). No significant differences were found (p>0.05). (B) The graph shows the average number of neurons/ganglion in relation to the age of the subjects. Subjects were grouped in decades19 and no significant differences were found between the groups (p>0.05).

We also analysed the effect of age on the number of neurons per ganglion and found no statistically significant correlation between age and the number of neurons identified (figure 3B). In detail, we virtually grouped the subjects in decades19: less than 30 years (11 subjects), 31–40 years (eight subjects), 41–50 years (seven subjects), 51–60 years (eight subjects), over 60 years (14 subjects) and found that the average number of neurons per ganglion was 5.2±2.0 (<30 years), 4.7±1.2 (31–40 years), 4.9±1.8 (41–50 years), 4.1±1.7 (51–60 years), 4.2±0.9 (>60 years; p>0.05; figure 3B).

Optimisation of the loading protocol and neuroimaging

First, we optimised the loading protocol (see the Materials and methods section) to obtain uniform cytosolic staining in submucous neurons and in fibre strands. We found the best result was obtained by loading the submucous plexus in oxygenated Krebs with 1 μM Fluo-4 AM and 0.01% Cremophor EL surfactant agent for 20 min at room temperature.

Using this protocol, human submucous ganglia could be easily recognised based on their Fluo-4 signal (figure 4A, left) because the morphology of ganglia and fibre tracts matched the appearance of the submucous plexus in the immunohistochemical analysis (figure 4A, middle).

Figure 4

Depolarisation and receptor agonists cause transient intracellular calcium (Ca2+) concentration ([Ca2+]i) responses in submucous neurons. (A, left) Fluo-4 signal from a typical ganglion. The numbers 1 and 2 indicate the neurons of which Ca2+ responses were plotted in C. (A, middle) Fluorescence image of the ganglion shown in A. Note that the neurons 1 and 2 express the neuronal marker Hu C/D (green), while only neuron 2 also expresses NF200. (A, right) Recordings of the responses of neuron 1 (green arrow) and 2 (red arrowhead) to high-K+ showing the typical fast upstroke in neurons to high-K+. (B, left) Fluo-4 signal and (C) recordings from submucous neurons to 1,1-dimethyl-4-phenylpiperazinium (DMPP) (C, left) and serotonin (5-HT) (C, right); the numbers 1 (green arrow) and 2 (red arrowhead) refer to the neurons analysed. (B, right) Fluorescence image of the ganglion analysed. Note that neuron 1 and 2 express the neuronal marker Hu C/D (green). (D, left) Number of neurons responding to DMPP and 5-HT. Each bar shows the percentage of responding neurons (identified by high-K+ depolarisation). (D, right) Relative fluorescence (Δ F/F0 in %) recorded from neurons when stimulated with DMPP and 5-HT. All scale bars: 20 μm.

Ca2+ imaging was successfully performed in 70 ganglia from 44 biopsies from 29 subjects. A total of 146 neurons responsive to high-K+ solution (peak amplitude 9±1%; duration at t50% 19.3±1.17 s) was recorded (figure 4A, right).

Response of enteric neurons to selective nicotinic agonist DMPP and 5-HT

To investigate whether receptor agonists were also able to activate the neurons, we applied DMPP and 5-HT to activate nicotinic and serotonergic receptors in the human submucous plexus (figure 4B). DMPP (10 μM, 20 s) induced a significant [Ca2+]i rise, indicating that 63.5% of the neurons identified by high-K+ depolarisation received cholinergic input (n=47 neurons, 17 ganglia, 11 subjects; figure 4D, left). The mean maximum amplitude was 8±1% and the response lasted 25.0±2.6 s at t50% (figure 4C left; figure 4D right).

As acetylcholine is not the only neurotransmitter in the ENS of the small bowel,16 we also tested the effect of another neurotransmitter that mediates fast excitatory post-synaptic potentials especially in the submucous plexus. Local application of 5-HT (10 μM, 20 s) generated a transient increase in [Ca2+]i (5±1%; duration at t50% 31.3±19.2 s) in 68.9% of the neurons identified by high-K+ (n=51 neurons, 19 ganglia, 11 subjects; figure 4B,C right, figure 4D).

Response of enteric neurons to electrical stimulation

In this protocol one fibre tract per ganglion was randomly selected and electrically stimulated, which elicited a Ca2+ increase in the fibre tract and in a subset of neurons (figure 5A, left). Trains of electrical pulses activated 57.1% of the high-K+ identified neurons (n=42 neurons, 16 ganglia, 12 subjects; figure 5C, left). The Ca2+ transient had a peak amplitude of 3±1% and lasted 19.8±3.1 s at t50% (figure 5B, up; figure 5C, right). It is likely that only a subset of the total viable neurons responded to electrical stimulation because we stimulated only one of the fibre tracts interconnecting the ganglia.

Figure 5

Electrical stimulation (ES; 2 s, 20 Hz) of interconnecting fibre tracts elicits transient calcium (Ca2+) responses that are blocked by tetrodotoxin (TTX; 1 μM) and removal of extracellular calcium (0 mM Ca2+). (A, left) Fluo-4 fluorescence recording from a typical ganglion. The numbers 1 and 2 indicate the neurons of which Ca2+ responses were plotted in B. (A, right) Fluo-4 fluorescence image of the ganglion analysed. Note that neurons 1 and 2 express the neuronal marker Hu C/D (green). (B, up) Recordings of the responses of neuron 1 (green arrow) and 2 (red arrowhead) to electrical stimulation. (B, down left) Electrical stimulation after superfusion with tetrodotoxin (5 min) did not cause any Ca2+ transients. (B, down right) Electrical stimulation after application of a 0 mM Ca2+ solution for 10 min, the electrical stimulation-induced Ca2+ transient (2 s, 20 Hz) was significantly inhibited. (C, left) Number of neurons responding to electrical stimulation in the presence of 0 mM Ca2+-Krebs solution or tetrodotoxin. Each bar shows the % of responding neurons (identified by high-K+ depolarisation). (C, right) Relative fluorescence (Δ F/F0 in %) recorded from neurons when stimulated with electrical stimulation in the presence of 0 mM Ca2+-Krebs solution or tetrodotoxin. All scale bars: 20 μm.

In all neurons recorded, the response to electrical stimulation was totally blocked by tetrodotoxin (1 μM), confirming that the signal was of neuronal origin (figure 5B, down left, figure 5C). Moreover, the electrically evoked Ca2+ response in submucous neurons was largely dependent on the influx of extracellular Ca2+ as 0 mM Ca2+ Krebs solution caused the total suppression of the response in 86.7% of neurons, which was reversible upon washout (figure 5B down right, figure 5C).


The results presented in this study clearly show that live nerve recordings can be performed in intestinal biopsies taken during routine endoscopy. To our knowledge, this is the first report using such small and ‘easily accessible’ samples to examine neuronal activity in the human gut. The ENS has the unique ability to control several intestinal functions (secretion, motility, inflammatory processes) independently from the central nervous system,20 therefore, it is pivotal to investigate the pathophysiological features of such complex system in living subjects. Until now, enteric nerve activity in humans could only be studied in the limited available number of surgical samples.6–13 Unfortunately, this approach suffers from the fact that tissue accessibility is restricted to patients with severe cases of intestinal diseases who therefore needed surgery. More recently, the problem related to the availability of surgical samples has been bypassed using colonic mucosal biopsies, taken during routine endoscopy, to obtain submucous plexus preparations suitable for immunohistochemical characterisation of enteric neurons.17 18 These seminal studies let us hypothesise that it would be possible also to explore neuronal functions in such freshly dissected samples.

Both myenteric and submucous neurons are affected in patients with inflammatory or functional intestinal diseases, leading to altered secretion and motility.21–24 Recent evidence also demonstrates that many neurological disorders, such as Parkinson's disease, not only affect the central nervous system but also cause neurodegeneration within the ENS.25 26 In patients with this type of disorder, neuronal morphological abnormalities can even be detected in the submucous plexus.

In this paper, we present the successful development of such a method that allows measuring enteric nerve activity in the submucous plexus from human intestinal biopsies. First of all, we set up a suitable protocol to load and visualise ganglia and neurons in these preparations. Due to the architecture of the submucous ganglia, normal DIC microscopy was sufficient to recognise them from the background. This was due, on the one hand, to the heavily wound up nerve bundles but also, on the other hand, to the collagen surrounding the ganglia that generates sufficient interference contrast. Indeed, we confirmed that collagen was the main source of interference contrast as the SH27 images generated by collagen perfectly matched the DIC images. Within the ganglia, neurons were easily detectable by their Fluo-4 fluorescence. These neurons proved to be viable, as they responded with a strong [Ca2+]i increase to high-K+ depolarisation. Electrical stimulation also evoked a Ca2+ response in 57.1% of the neurons identified by high-K+ application. The electrical stimulus used in this study is known to elicit slow excitatory post-synaptic potentials that are highly likely to be superimposed by action potentials. Apart from this synaptic communication, the responsive population also includes neurons that were antidromically activated by electrical stimulation. In either case, neuronal action potential firing was essential as all responses were sensitive to tetrodotoxin. In the absence of extracellular Ca2+ only small responses were still present in very few neurons. We also tested whether the neurons could be activated by receptor stimulation using DMPP and 5-HT, and we found that Ca2+ responses were reliably elicited in over 60% of the neurons by either of the agonists, which, again, confirms the neuronal origin of the signal. Compared with neuronal Ca2+ responses recorded in neurons from experimental animals, the amplitude of the responses we measured was relatively low.28 29 This is mainly due to the fact that the submucous preparations have a relatively high autofluorescent background, which cannot be easily corrected for because it is not at all uniform. A confocal approach would definitely generate higher quality images, but we show that, even with an easily accessible wide-field fluorescence microscope, reliable nerve signals can be recorded in these preparations. The success rate we obtained was approximately 60%. We estimated that approximately half the failures were due to technical reasons, a fraction that was a lot higher during the preliminary phase. In the other half (20% of the total trials) the lack of response most likely reflected the biological state of the tissue, as no obvious technical problem, such as problematic dissection or bad Fluo-4 loading, was identified. The latter is obviously of great interest and requires further investigation to relate the absence of a Ca2+ signal to any biological or pathological condition.

To confirm further that the recorded responses were actually generated by neurons, all the submucous samples were processed by immunohistochemistry. The immunostaining showed both isolated neurons and ganglia interconnected by broad fibres typical of the complex tridimensional architecture of the submucous plexus in the human gut. The immunohistochemical analysis allowed us also to perform a more detailed quantitative analysis of ganglia and neurons in the submucous plexus, as reported in detail in table 1. To our knowledge, this is the first study reporting a characterisation of the submucous plexus in the upper gastrointestinal tract using human intestinal biopsies. Indeed, until now, the number of submucous ganglia has been studied only in the colon and the rectum17 18 30–32 or in surgical duodenal specimens.19 The number of neurons per ganglia (approximately five) we found is in good agreement with what was found in the submucous plexus from colonic biopsies.17 However, there is no accordance in the number of ganglia per biopsy and, as a consequence, the total number of neurons per biopsy.

Finally, we also investigated the possible correlation between the total and the average number of neurons per ganglion or per biopsy, and the gender and age of the subjects studied. Interestingly, our imunohistochemical data revealed that the number of neurons and ganglia was not correlated with age and/or gender. Although this is in agreement with a recent study reporting no age-dependent differences in neuron numbers in the human submucous plexus,19 it contrasts with the age-dependency of colonic myenteric neurons described by Bernard et al.33 The absence of an age and gender factor in our study is an important advantage as it proves that the technique is applicable to subjects from a wide age range.

Although further investigations are needed to correlate our data with other parameters such as patient pathology, neuron subpopulations, etc, the simplicity and reproducibility of the method could turn this approach into a ‘routine’ technique to detect nerve activity, which can be combined with morphological and immunohistochemical characterisation of the active nerves.

To summarise, we confirmed that enough neurons are present in human intestinal biopsies and, second, we demonstrated that is possible to perform optical recordings successfully from individual neurons. Finally, we were able to localise the recorded ganglia after fixation and to perform a matched immunostaining.

Our study is likely to have a significant impact on the field of gastroenterology, as it provides a novel tool to further the understanding of ENS abnormalities in either inflammatory or functional or even neurological disorder-related intestinal diseases, especially in those with unknown aetiology.25 26 34–40 This tool will also help to identify new pharmacological targets for the treatment of these pathologies.


The authors would like to thank the members of LENS for their critical comments and skilled technical assistance. Confocal and SH imaging was performed in the Cell Imaging Core (CIC, KU Leuven, Belgium).


View Abstract


  • Funding This work was funded by Methusalem (BOF, KU Leuven, JT) and FWO (G.0501.10, PVB). This work was funded only by academic funding bodies (University and Scientific Council, Flanders, Belgium).

  • Competing interests None.

  • Patient consent Obtained.

  • Ethics approval The study protocol (ML7400) was approved by the Ethics Committee of Leuven University Hospital, Leuven, Belgium.

  • Provenance and peer review Not commissioned; internally peer reviewed.

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.