Article Text

This article has a correction. Please see:

Download PDFPDF

TNFα is a key mediator of the pronociceptive effects of mucosal supernatant from human ulcerative colitis on colonic DRG neurons
  1. Charles Ibeakanma,
  2. Stephen Vanner
  1. Gastrointestinal Diseases Research Unit, Kingston General Hospital, Queen's University, Kingston, Ontario, Canada
  1. Correspondence to Dr Stephen Vanner, 76 Stuart St, GIDRU, Kingston General Hospital, Kingston, Ontario, Canada K7L 2V7; vanners{at}


Objectives Abdominal pain is a serious cause of morbidity in patients with inflammatory bowel disease. To better understand the mechanisms and potentially identify new targets for treatment, the effects of inflammatory supernatant from colonic biopsies of patients with active ulcerative colitis (UC) on mouse colonic nociceptive dorsal root ganglia neurons were examined.

Methods Acutely dissociated dorsal root ganglia neurons innervating the mouse colon were incubated in supernatants obtained from colonic biopsies from patients with UC. Whole-cell patch clamp recordings were obtained to examine the effects on neuronal excitability. The role of tumour necrosis factor α (TNFα) was studied using TNFα receptor (TNFR) knockout mice and comparing supernatant and TNFα actions.

Results UC supernatants significantly decreased the rheobase and increased action potential discharge, indicating increased neuronal excitability. Human biopsies exhibited high levels of TNFα, and mouse colonic neurons only exhibited TNFR1 mRNA. Incubation with TNFα recapitulated the supernatant effects on neuronal excitability, and supernatant and TNFα actions were almost completely blocked in TNFR knockout mice. In voltage clamp studies, transient IA and IK currents were suppressed and Nav 1.8 currents were enhanced by TNFα and UC supernatant, suggesting that multiple underlying mechanisms contributed to the enhanced excitability.

Conclusions UC supernatants enhance neuronal excitability of sensory dorsal root ganglia neurons innervating the colon. TNFα is a key mediator which acts at neuronal TNFR1 to modulate Kv and Nav currents. Together these data provide a number of potential new targets for pain management in UC.

  • Abdominal pain
  • electrophysiology
  • inflammatory bowel disease TNFα
  • ulcerative colitis

Statistics from


Inflammatory bowel disease (IBD), including Crohn's disease (CD) and ulcerative colitis (UC), is a common disorder that affects >1 million North Americans.1–3 This chronic inflammatory condition typically has a relapsing and remitting course characterised by bloody diarrhoea, fatigue and abdominal pain. These symptoms are a major cause of morbidity and can severely limit quality of life. Current treatment of abdominal pain is largely limited to narcotic analgesics such as morphine, but these have important side effects, including altered cognition, drowsiness, decreased energy levels, and nausea and vomiting. Furthermore, in patients with active colitis, they can precipitate life-threatening toxic megacolon.4 As a result, there is a great need to better understand the mechanisms underlying pain in patients with IBD and to identify new therapeutic targets.

Chemical and mechanical stimuli in the colon are detected by axon terminals of dorsal root ganglia (DRG) neurons which relay signals to the CNS.5 6 During colitis, such as a UC flare-up, a large number of proinflammatory mediators are released within the intestinal wall and the serum, which can elicit pain by activating specialised receptors on the Aδ and C-fibre nociceptors.7 8 Assays of inflamed mucosa in UC and CD have increased protein and mRNA of a number of cytokines, including interleukin-1β (IL-1β), IL-6, IL-8 and tumour necrosis factor α (TNFα).9–11 There is evidence that many of these mediators can sensitise channels, such as the transient receptor potential (TRP) channels, resulting in enhanced neurotransmission produced by mechanical, chemical and other stimuli.12–15 Furthermore, they can also modulate Nav and Kv ion channels which underlie the generation of the action potentials,16–19 thus also increasing pain signalling from a broad range of sensory stimuli. Recent evidence suggests that antinociceptive mediators are also released,20 21 which may be dependent on the inflammation's type (eg, IBD vs infectious) and progression (eg, acute vs chronic). Therefore, the net effect of the inflammatory milieu on neuronal excitability may depend upon the balance of these factors.

Among the pronociceptive mediators, TNFα is thought to play an important role. When applied subcutaneously or perineurially, TNFα lowers the mechanical activation threshold in nociceptors, rapidly evokes ongoing activity in C-fibres and elicits mechanical allodynia and/or hyperalgesia.6 22 However, it is unknown whether colonic DRG neurons are affected similarly, and whether any effects on neuronal excitability make a significant contribution within the wider inflammatory cytokine milieu of IBD.

The ability to obtain a representative supernatant from mucosal biopsies from patients with active UC allows the study of the net effect of this inflammatory milieu on nociceptive neurons. We found that these supernatants markedly increased the excitability of mouse colonic nociceptive neurons. Using combined in vitro electrophysiological and molecular techniques we sought to determine the role that TNFα may play in these actions and to examine voltage-gated ion channels which underlie these changes in excitability.

Materials and methods

Human biopsy samples

A total of 136 colonic tissue samples were obtained from 17 patients (eight UC and nine controls) undergoing colonoscopy at the Hotel Dieu Hospital and Kingston General Hospital, Kingston, Ontario, Canada. The protocol was approved by the Queen's University Human Ethics Committee; all patients gave informed consent. UC diagnosis was based upon endoscopic and histological criteria. Control biopsies were from patients undergoing colon cancer screening; none exhibited endoscopic signs of inflammation. In patients with active UC, eight adjacent biopsies were collected from the region of inflamed colon. The samples were weighed and transferred into well plates of 250 μl of RPMI with 10% fetal calf serum (FCS), penicillin/streptomycin and gentamicin/amphotericin B, and incubated overnight at 37°C, under 95% O2/5% CO2. Subsequently the supernatants were collected and stored at −80°C for further use.

Retrograde labelling and isolation of colonic-projecting neurons

All animal protocols were approved by the Queen's University Animal Care Committee. Adult mice (CD1) 5–6 weeks old from Charles River Laboratories (Montreal, QC, Canada) of either sex were used for all studies. Colonic-projecting DRG neurons were labelled using Fast Blue (Cedarlane Laboratories, Homby, ON, Canada) 7 days in advance as previously described.23 For DRG isolation, mice were anaesthetised with isoflurane and killed by cervical transection. The spinal cord was removed and DRG from T9 to T13 were dissected, as these levels receive nociceptive input from the colon.24 25 The DRG neurons were acutely dissociated as previously described.26 27 Dispersed neurons were suspended in 500 μl of RPMI medium containing 10% fetal bovine serum (FBS), 100 U/ml penicillin, 0.1 mg/ml streptomycin and 2 mM glutamine, and plated on Pure-Col- (Inamed Biomaterials, Fremont, California, USA) coated coverslips (60 μl/ml). To this was added 250 μl of supernatant from either patients with UC or controls, or a human TNFα solution (final concentration: 100 ng/ml) (Sigma Aldrich, St Louis, Missouri, USA) or RPMI medium alone, for 24 h in a humidified incubator at 37°C, 95% air and 5% CO2 before electrophysiological studies.

Electrophysiological recordings

Current or voltage patch clamp experiments were performed on Fast Blue-labelled neurons. Small neurons (≤40 pF capacitance) were studied since they are thought to be nociceptors.1 26 28 29 Signals were amplified by an Axopatch 200B amplifier filtered at 5 kHz and digitised at 20 kHz with a Digidata 1322A A/D converter and analysed using Clampex 10.0 (all from Axon Instruments, San Jose, California, USA).


For perforated-patch current clamp recordings with amphotericin B (Sigma Aldrich), the following solutions were used (mM): (external), 140 NaCl, 5 KCl, 1 MgCl2, 2 CaCl2, 10 HEPES and 10 d-glucose, pH adjusted to 7.4 with NaOH; (pipette), 110 K-gluconate, 30 KCl, 10 HEPES, 1 MgCl2 and 2 CaCl2 with pH adjusted to 7.25 using KOH. Perforated-patch voltage clamp recordings of potassium currents were obtained in sodium-free bath solution: (mM) 140 N-methyl-d-glucamine (NMDG), 4 KCl, 1.8 HEPES, 1 d-glucose, 1 CaCl2, 1 MgCl2, pH 7.4 adjusted with HCl. For whole-cell voltage clamp recording of sodium currents, the patch pipette solution contained (mM): 10 NaCl, 110 CsCl, 3 MgCl2, 10 HEPES, 0.6 Na-GTP, 3 ATP and 10 EGTA, pH adjusted to 7.3 with CsOH. Bath solution contained (mM): 55 NaCl, 90 TEA-Cl, 1 CaCl2, 2 MgCl2, 0.1 CdCl2, 10 HEPES, 5 glucose, and 5 4-aminopyridine (4-AP), pH adjusted to 7.4 with NaOH. The liquid junction potential was compensated for by adjusting the zero current potential.

Potassium current recordings

IA and IK currents were separated as previously described.27 30 Briefly, total potassium current was elicited using 500 ms voltage steps from −90 to +50 mV in 10 mV increments from a holding potential of −100 mV. IK was then elicited using the same pulse protocol, but from a holding potential of −60 mV to inactivate IA. The amplitude of the IK current was measured at 450 ms at each voltage step (see figure 1). IA was isolated by subtracting IK from the total potassium current. The amplitude of the IA current was measured at the peak of the transient component at each voltage step (see figure 1). These voltage protocols may result in a small crossover contamination of the two currents, but previous studies in our laboratory suggest that this contamination would not significantly alter the interpretation of the results.27 Activation and inactivation curves were analysed as previously described.27 30

Figure 1

Supernatant from tissue from patients with ulcerative colitis (UC) induces hyperexcitability of mouse nociceptive dorsal root ganglia (DRG) neurons. (A) Representative perforated current clamp recordings in response to a 500 ms current injection at the rheobase (left) and twice the rheobase (right) from neurons incubated in UC supernatant, control supernatant and control media. (B) Summary data illustrating the effects of incubation of DRG neurons with supernatant from tissues from patients with UC (black bar), control patient tissues (grey bar) and control media (white bar). Data represent the mean±SEM. *p <0.05, **p <0.001, one-way analysis of variance (ANOVA) with Bonferroni multiple comparison test.

Nav current recordings

Membrane currents were recorded using the whole-cell patch clamp technique and activation and inactivation curves analysed as previously described.27 30 Cells were held at a holding potential of −120 or −100 mV and then stepped from −80 to +35 mV in 10 mV intervals for 200 ms each. Sixty to seventy per cent series resistance compensation was applied. Only cells with adequate voltage clamp and space clamp31 and with gradual Na+ current activation were used. Tetrodotoxin (TTX; 1 μM) was used to isolate TTX-resistant (TTX-R) Nav1.8 currents. TTX-R Nav1.9 currents should have an activation threshold ∼20 mV more negative than the threshold for TTX-R Nav1.8 currents32; however, none was seen.

ELISA measurement of TNFα content of supernatants

Human TNFα ELISA kits (Medicorp, Montréal, QC, Canada) were used to assay supernatants and standards simultaneously as per the manufacturer's instructions. Polyclonal goat antihuman TNFα antibodies were used as capturing antibodies, and biotinylated polyclonal rabbit antihuman TNFα antibodies as detecting antibody. Streptavidin–horseradish peroxidase (HRP) and stabilised chromogen were added as colour indicators. Optical densities of plates were read at 450 nm in a Titertek Multiskan Plus photometer (Titertek Instruments, Huntsville, Alabama, USA) directly after colour reactions were stopped with acid. All steps were performed at room temperature and samples were assayed in duplicate. Results are expressed as pg/mg tissue wet weight.

TNFα receptor mRNA amplification

Total RNA was isolated from laser-captured Fast Blue-labelled colonic DRG neurons using the TRIzol reagent (Invitrogen, Carlsbad, California, USA) according to the manufacturer's instructions. Reverse transcription of 0.8 μg of RNA was performed using Expand RT (Roche, Mannheim, Germany) and oligo(dT12–18) (Invitrogen). A 2 μl aliquot of each reverse transcript served as a template for PCR amplification to estimate the mRNA expression of TNF receptor 1 (TNFR1) and TNFR2 using Platinum Taq DNA Polymerase (Invitrogen). Primers used for analysis were: TNFR1 sense 5′-GGA TTG TCA CGG TGC CGT TGA AG-3′, and TNFR1 antisense 5′-TGA CAA GGA CAC GGT GTG TGG C-3; and TNFR2 sense 5′-AAC GGG CCA GAC CTC GGG T-3′, and TNFR2 antisense 5′-AGA GCT CCA GGC ACA AGG GC-3′.

Amplification conditions were 3 min at 94°C, followed by 35 cycles of 30 s at 94°C, 1 min at 55°C and 1  at 72°C. After 35 cycles of reaction, which preliminary studies showed was within the linear range for these primers and samples, 5 ml of each amplified fragment and 1 μl of loading buffer were subjected to 1.5% agarose gel (Bioshop, Burlington, ON, Canada) electrophoresis separation and stained with ethidium bromide (Fisher, Ottawa, ON, Canada). Experiments were repeated five times.


All data are expressed as the mean±SEM. Treatment effects were analysed by Student t test using Prism (GraphPad, San Diego, California, USA). One- or two-way analysis of variation (ANOVA) with Bonferoni correction was applied for comparison of multiple groups. Fitting of activation/inactivation data was done using the Boltzmann equation fit function in Origin 6.0 (OriginLab, Northampton, Massachusetts, USA). Mean voltages of half activation (Vh) and slope factors (k) were obtained from individual Boltzmann curve fits.


Patient characteristics

Mucosal biopsies were obtained from regions of colonic inflammation in eight patients with active UC (see table 1).

Table 1

Patient characteristics

Control biopsies were obtained from nine patients undergoing colonoscopy for colon cancer screening. None of the patients had complaints of diarrhoea and the mucosa had a normal endoscopic appearance.

UC supernatant induces hyperexcitability of mouse nociceptive DRG neurons

Fast Blue-labelled mouse DRG neurons exhibited increased excitability following overnight incubation in UC supernatant (representative traces shown in figure 2A). The rheobase decreased by 51.7 and 55.6% in neurons incubated in the UC supernatant (n=15) compared with those incubated in control supernatant (n=14, p <0.05) and control media (n=15, p <0.05), respectively. The number of action potentials elicited at twice the rheobase was also increased by 115.4 and 87.5% compared with control supernatant (n=14, p<0.001) and control media (n=15, p<0.001), respectively. There was no difference in the mean input resistance or the resting membrane potential between neurons incubated in UC supernatant and control supernatant or control media. There also was no difference in excitability between neurons incubated in control supernatant and those incubated in control media alone (figure 2B).

Figure 2

Mouse colonic dorsal root ganglia (DRG) neurons express tumour necrosis factor α receptor 1 (TNFR1) and TNFα is markedly elevated in supernatants from patients with ulcerative colitis (UC). (A) In whole DRG neurons, expression of both TNFR1 and TNFR2 was confirmed by reverse transcription–PCR (RT–PCR) and agarose gel (1.2%) electrophoresis. (B) However, expression of TNFR1, but not TNFR2, mRNA was confirmed in laser-captured, Fast Blue-labelled colonic DRG neurons by RT–PCR. (C) Evaluation of TNFα level in supernatants from patients with UC (n=6) and control patients (n=6) using ELISA techniques. Summary bar graph demonstrating that UC supernatant contained significantly higher levels of TNFα (black bar) compared with control patient supernatant (white bar). Values are the means±SEM of three independent experiments. **p <0.005.

TNFR expression in mouse colonic DRG neurons and TNFα concentration in human biopsy supernatants

To demonstrate that TNFRs are expressed on the T9–T13 level mouse colonic DRG neurons electrophysiologically recorded from, we profiled the expression of TNFR1 (figure 3A, left gel) and TNFR2 (figure 3A, right gel) in whole DRGs by reverse transcription–PCR (RT–PCR) and confirmed the presence of both transcripts. However, using retrograde Fast Blue labelling and laser capture microdissection, we found transcripts for TNFR1 alone (figure 3B). The absence of signal in the ‘no template’ samples shows that there is no genomic or other DNA contamination. Supernatants obtained from patients with UC were found by ELISA to contain >6-fold higher TNFα levels (41.7±9.5 pg/mg tissue) compared with control patient supernatant (6.5±3.4 pg/mg tissue; p <0.007; n=6) (figure 3C). In a separate series of experiments, we measured TNFα in ng/ml in the supernatant using the human TNFα (Sigma) employed in the electrophysiological studies as the standard for the ELISA measurements and found that the supernatant TNFα values ranged from 0.4 to 1.6 ng/ml (mean=0.75 ± 0.18 ng/ml).

Figure 3

Exogenous tumour necrosis factor α (TNFα) also increases excitability of nociceptive dorsal root ganglia (DRG) neurons. Data illustrating that overnight incubation with 100 ng/ml exogenous TNFα (black bar) induced hyperexcitability of DRG neurons, causing a significant reduction in the rheobase with significant increases in the number of action potentials at twice the rheobase and action potential half-width compared with neurons incubated in control media (white bar). The membrane potential did not differ between the two conditions. Data represent the mean±SEM. *p<0.05.

TNFα increases neuronal excitability of mouse colonic DRG neurons

To examine the role of TNFα in the action of UC supernatant, we incubated mouse colonic DRG neurons in TNFα (100 ng/ml). As we found with UC supernatant-incubated neurons, incubation with TNFα (n=15) resulted in a significant decrease in the rheobase (p <0.03) compared with control media (n=14) (figure 4). Similarly, TNFα resulted in a >70% increase in the number of action potentials at twice the rheobase (p <0.01) versus controls. Neither the resting membrane potential (TNFα, −48.2±1.7 mV; control, −51.04±1.6 mV) nor the input resistance were significantly different between the groups. Thus, TNFα also increases colonic DRG neuronal excitability. To ensure that the 100 ng/ml employed in the electrophysiological studies was physiological, we repeated the electrophysiological studies using a TNFα value (1 ng/ml) comparable with that measured in the supernatants (n=9 cells) and compared this with controls (n=8 cells). We found that the rheobase was also markedly decreased (TNFα mean=53.3±8.0 vs control mean=95.0±8.0 pA; p=0.0035) and there was a significant increase in action potential discharge at twice the rheobase (TNFα mean=4.2 ± 0.6 vs control mean=2.0 ± 0.7; p=0.03).

Figure 4

The effects of ulcerative colitis (UC) supernatant and exogenous tumour necrosis factor α (TNFα) on neuronal excitability are markedly attenuated in TNFα receptor (TNFR) knockout mice. Data shown illustrate reversal of effects of (A) UC supernatant and (B) TNFα on excitability of dorsal root ganglia (DRG) neurons isolated from TNFR knockout (black bars) and wild-type mice (white bars). Data represent the mean±SEM. *p <0.05. Numbers in parentheses indicate numbers of neurons tested.

The effects of UC supernatant on neuronal excitability are markedly attenuated in TNFR knockout mice

In order to examine the contribution of TNFα to the overall effect of UC supernatant on DRG neuronal excitability, we studied the effects of UC supernatant and exogenous TNFα on neurons isolated from wild-type and TNFR knockout mice (figure 5). We confirmed the presence or deletion of TNFR1 and TNFR2 in the wild-type and knockout mice, respectively, by RT–PCR and agarose gel (1.5%) electrophoresis. β-Actin was used as loading control and was detected in both wild-type and knockout cDNA to confirm cDNA integrity. The absence of signal in the ‘no template’ and wild-type DRG RNA samples confirmed that there was no genomic DNA or other contamination (data not shown).

Figure 5

Tumour necrosis factor α (TNFα) and ulcerative colitis (UC) supernatant suppress KV+ currents. (A) Representative voltage clamp trace of KV+ currents isolated while holding at −100 mV (total currents) or −60 mV (IK), using a 500 ms voltage step in 10 mV intervals. Subtraction of IK from the total current yielded IA (−100–(−60)). Arrowheads indicate the points at which current amplitudes were measured and analysed for the corresponding voltage steps. Current–voltage relationship of the isolated IA (left) and IK (right) currents illustrate significant suppression of both currents in TNFα-(B) and UC supernatant- (C) treated neurons. (D) Effects of exogenous TNFα (left) or UC supernatant (right) on the steady-state inactivation properties of isolated KV+ currents. The steady-state inactivation curves for both IA (left) and IK (right) currents in neurons incubated with TNFα or supernatant were obtained by plotting the normalised test current amplitudes against conditioning prepulse potential and were fitted using the Boltzmann function. Incubation with TNFα or UC supernantant resulted in a hyperpolarising (left) shift of the steady-state inactivation curve for both currents. Data represent the mean±SEM. *p <0.05, **p <0.001, ***p <0.0001, two-way analysis of variance (ANOVA) with Bonferroni post-test.

In control experiments using media only, we confirmed there were no differences in the intrinsic excitability of the wild-type (n=14 cells) compared with the TNFR knockout neurons (n=13 cells). (Wild-type mean rheobase=84.3±9.8 pA vs TNFR knockout mean=76.9±6.3 pA; mean action potentials at twice the rheobase in wild-type mice=2.1 ± 0.5 vs TNFR knockout mean=1.8 ± 0.5.) After TNFα incubation, the rheobase in the knockout mice DRG neurons was 52% higher (n=17, p=0.003), and the mean number of action potentials at twice the rheobase was 44% lower (p=0.04), than in wild-type (figure 5A).

The effects of UC supernatant on neuronal excitability (see figure 2) were next examined in the TNFR knockout mice to assess what contribution TNFα made to the overall supernatant effect (figure 5B). The rheobase in neurons incubated in the UC supernatant (n=17) was 45% higher (figure 5) than in wild-type neurons (n=17; p=0.05). The number of action potentials at twice the rheobase was decreased by 46% compared with the wild type (p=0.03).

TNFα and UC supernatants modulate Kv and Nav currents

Voltage-gated Na+ and K+ channels are fundamental in determining neuronal excitability and we and others have shown previously that inflammation can modulate these channels in DRG neurons.1 18 23 33 Acute application of TNFα has been shown to modulate Nav 1.8 currents in somatic DRG neurons,34 but the effect of overnight incubation of TNFα and UC supernatant on these and Kv currents is unknown.

Thus, to understand the ionic mechanisms underlying the TNFα-induced hyperexcitability of mouse nociceptive DRG neurons, we isolated voltage-gated potassium (Kv) currents based on their inactivation properties (figure 1A) and compared current densities between control and those incubated in TNFα (100 ng/ml) or UC supernatant and control supernatant. TNFα incubation caused a significant decrease of IA (p=0.01) and IK (p=0.0003) current densities, respectively, compared with control neurons (figure 1B). Similarly, UC supernatant incubation caused a significant decrease of IA (p=0.007) and IK (p=0.006) current densities, respectively, compared with control supernatant (figure 1C). Incubation with TNFα or UC supernatant did not alter the voltage dependency of activation of either IA or IK currents (data not shown). However, there was a hyperpolarising shift of the steady-state inactivation curve for both currents with TNFα and of the IA current with the UC supernatant. The Vh values for the steady-state inactivation of IA in TNFα-treated and control neurons were −91.6±0.7 and −80.1 ± 0.5 mV; p=0.0001, with k values of −10.04±0.5 and −10.2±0.5, respectively; whereas the Vh values of IK in TNFα-treated and control neurons were −72.1±0.4 and −65.9±0.3 mV; p = 0.001, with k values of 8.8±0.3 and 9.6±0.3, respectively. The Vh values for the steady-state inactivation of IA in the UC supernatant and control supernatant neurons were −86.4±0.7 and −76.2±0.8 mV; p 0.001; with k values of −12.3±0.5 and −11±0.7, respectively.

Nav 1.8 currents were isolated by including TTX (1 μM) in the bath solution to block TTX-sensitive currents. As illustrated in figure 6, overnight incubation of colonic DRG neurons with exogenous TNFα (n=8 cells for TNFα and controls) or UC supernatant (n=9 for UC supernatant and n=10 for control supernatant) enhanced the TTX-R Nav1.8 currents. The current–voltage relationship demonstrated activation of the currents at approximnately −42 mV, with a peak close to −10 mV (p=0.001 for TNFα and p=0.009 for UC supernatant). TNFα and UC supernatant did not affect either the voltage dependency of activation or steady-state inactivation of the currents (data not shown).

Figure 6

Tumour necrosis factor α (TNFα) and ulcerative colitis (UC) supernatant enhances tetrodotoxin-resistant (TTX-R) Nav 1.8 currents. (A) Representative voltage clamp trace of TTX-R Nav 1.8 currents isolated using 1 μM TTX from control (left trace) and TNFα- (right trace) incubated neurons. Cells were held at a holding potential of −100 mV and stepped from −80 to +35 mV in 10 mV intervals for a duration of 200 ms. (B) Current–voltage relationship for isolated TTX-R Nav 1.8 currents showing significant augmentation of currents in neurons incubated in TNFα (left) or UC supernatant (right) as opposed to control media or control supernatant, respectively. Data represent the mean±SEM. *p <0.05, **p <0.001, two-way analysis of variance (ANOVA) with Bonferroni post-test.


The inflammatory milieu surrounding the axon terminals of nociceptive DRG neurons in the intestine contains a wide repertoire of inflammatory mediators, many of which can either directly activate or sensitise these neurons7 8 13 14 19 and, in some cases, have opposing antinociceptive actions.19–21 35 The exact nature of these mediators depends partly upon the type of inflammation (eg, UC vs CD vs infectious colitis). Given this variability, the first aim of this study was to determine the net effect of the inflammatory milieu on the excitability of nociceptive neurons innervating the inflamed region of the colon in UC. UC is a chronic inflammatory disease of the colonic mucosa and submucosa, and this inflamed tissue is readily accessible through pinch biopsies obtained during endoscopic procedures.

We incubated neurons in supernatant containing secreted inflammatory mediators from the mucosa and submucosa for ∼24 h to mimic the inflammatory milieu in human active UC (ie, as opposed to acute application for minutes) and found that this supernatant induced hyperexcitability in colonic nociceptive neurons. Since enhanced neuronal excitability can mediate increased nociceptive signalling,36 37 our results suggest that the inflammatory cytokine profile from human patients with UC can play an important nociceptive role in these patients. The second aim of this study was to identify a key mediator underlying this response. We examined whether TNFα was important because it was known to activate DRG neurons directly, and clinical treatments currently exist to target its actions in humans.

Several lines of evidence suggested that TNFα was central to the neuronal excitability changes evoked by the human supernatants in our study. We observed high levels of TNFα in the supernatant from patients with UC, and confirmed the presence of TNFR1 transcript on labelled DRG neurons.38–42 To ensure colonic-projecting neurons were used in the RT–PCR studies, we used laser capture microdissection to isolate Fast Blue-labelled DRG neurons. Previous studies34 of unidentified DRG neurons demonstrated that acute TNFα application could modulate Na+ channels in DRG neurons, but the effects on neuronal excitability and colonic neurons per se were unknown. Although we did not directly compare supernatant and TNFα effects, we found that incubation of TNFα (24 h) for the same duration resulted in very similar actions on neuronal excitability, that is decreased rheobase and increased action potential discharge compared with controls. Moreover, when these findings were examined in TNFR knockout and wild-type mice, the decreases in supernatant and TNFα effects were similar in knockout animals. Together, these findings suggest that abrogation of the TNFα effect on nociceptive DRG neurons could have significant effects on pain in these patients with IBD.

Changes in the rheobase and rate of action potential discharge, as seen in the present study following supernatant and TNFα incubation, can result from modulation of one or more of the voltage-gated ion channels underlying action potential generation. In TNBS (2,4,6-trinitrobenzene sulfonic acid) colitis and ileitis animal models of IBD, neuronal hyperexcitability is associated with modulation of both Kv and Nav 1.8 currents.1 23 27 43 There is also evidence that acute application of individual inflammatory mediators, including TNFα, can modulate one or more of these channels in a similar fashion.13 14 18 36 37 We tested the effects of TNFα incubation and the UC supernatant using voltage clamp techniques and found significant reductions in transient IA and delayed rectifier IK currents and enhancement of the Nav 1.8 currents, both of which would lead to enhanced neuronal excitability. The intracellular signalling event(s) which mediate the TNFR responses remain to be fully elucidated, but post-translational and transcriptional changes in these ion channels have been implicated.44 45 We observed a leftward shift in the IA inactivation curve, suggesting a smaller proportion of available channels could be activated at a given voltage, and as a result could contribute significantly to the enhanced neuronal excitability observed in our study. This finding has been observed in a number of inflammatory models1 27 30 44 and, although the mechanism underlying this shift is unclear, post-translational modification, such as phosphorylation of these channels through protein kinases, has been proposed.44 45 Our previous molecular studies of the expression of Nav 1.8 channels during inflammation also suggested that post-translational mechanisms may underlie the observed increases in Nav 1.8 currents and we failed to demonstrate increased Nav1.8 transcript following 24 h TNFα incubation.43 Together, these findings suggest that post-translation modulation of channels is important, although they do not exclude that transcriptional changes may also occur during the evolution of inflammatory processes.1 43 44 Previous studies have implicated both p38 and ERK (extracellular signal-regulated kinase) signalling in such events,34 38 but detailed studies with selective antagonists of these intracellular pathways are needed to determine whether one or more of these pathways preferentially modulates specific ion channels.

In summary, we have shown that supernatants obtained from mucosal biopsies from patients with UC induced hyperexcitability of nociceptive colonic DRG neurons. Futhermore, we have found that TNFα is a major mediator of these events and that modulation of multiple ion channels is involved. Thus, the treatment of patients with UC with anti-TNFα therapy such as infliximab could decrease pain, at least in part, by directly inhibiting the action of TNFα on nociceptive neurons. It is possible that other mediators in the supernatant have acute actions on neuronal excitability which desensitised during the more sustained exposure examined in our study. Although we studied changes in ion channels in the soma to reflect what occurs in the nerve terminal within the inflamed intestine, there may be additional mechanisms restricted to the nerve terminal, such as immune secretion of antinociceptive factors,20 21 which could further modulate the changes we observed. This study focused on patients with UC given the relative ease of access to inflamed tissue in the distal colon; however, given the evidence that the balance of pronociceptive and antinociceptive factors may differ in patients depending on the nature of the pathophysiology of the inflammation,20 21 it will be important to determine whether important differences exist in patients with CD or other pathology such as infectious or ischaemic colitis. It will also be important to dissect the contribution of other pronociceptive factors, including those released by bacteria, for example lipopolysaccharides and lipoproteins, to determine their direct and indirect contribution to nociception in these patients.

Significance of this study

What is already known about this subject?

  • Abdominal pain is a serious cause of morbidity in patients with IBD.

  • The levels of a large number of proinflammatory cytokines are increased in assays of human IBD biopsies.

  • Individual cytokines sensitise nociceptive neurons in animal models by modulating ion channels.

What are the new findings?

  • Supernatant from biopsies of human UC induces hyperexcitability of nociceptive neurons innervating the colon.

  • Human TNFα acting on TNFα receptor1 is a key mediator of these actions,

  • UC supernatant and TNFα increase Nav 1.8 channels and suppress Kv channels through direct actions on the neurons.

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

  • Inhibitors of TNFα production or TNFα receptor 1 on neurons could directly modulate pain in patients with IBD.

  • Inhibitors of Nav 1.8 channels and/or activators of Kv channels could also provide novel approaches to treating IBD pain.


SV is supported by an operating grant from the Crohn's and Colitis Foundation of Canada (CCFC) and a CCFC research scientist award. CI is supported by SV's CCFC operating grant and a GIDRU CIHR Training Grant in Digestive Sciences. We are grateful to Dr Ian Spreadbury for his valuable comments on the manuscript and helpful discussions, and Iva Kosatka, Margaret O'Reilly and Anne Marie Crotty for their technical assistance.



  • Funding Crohn's and Colitis Foundation of Canada, and Canadian Institutes of Health Research.

  • Competing interests None.

  • Ethics approval This study was conducted with the approval of the Queen's University Human Ethics Committee.

  • Provenance and peer review Not commissioned; externally 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.

Linked Articles

  • Correction
    BMJ Publishing Group Ltd and British Society of Gastroenterology