Article Text

Download PDFPDF
P2X receptor-mediated visceral hyperalgesia in a rat model of chronic visceral hypersensitivity
  1. G-Y Xu1,
  2. M Shenoy1,
  3. J H Winston1,
  4. S Mittal1,
  5. P J Pasricha2
  1. 1
    Division of Gastroenterology and Hepatology, Department of Internal Medicine, University of Texas Medical Branch, Galveston, TX, USA
  2. 2
    Division of Gastroenterology and Hepatology, Stanford University School of Medicine, Stanford, California, USA
  1. Dr G Y Xu, Division of Gastroenterology and Hepatology, Department of Internal Medicine, University of Texas Medical Branch, Galveston, TX 77555, USA; gyxu{at}


Background: Irritable bowel syndrome (IBS) is a common gastrointestinal disorder characterised by abdominal pain and bloating in association with altered bowel movements. Its pathogenesis and the underlying molecular mechanisms of visceral hyperalgesia remain elusive. Recent studies of somatic and other visceral pain models suggest a role for purinergic signalling mediated by the P2X receptor (P2XR) family.

Aims: To examine the role of P2XR signalling in the pathogenesis in a rat model of IBS-like visceral hyperalgesia.

Methods: Visceral hypersensitivity was induced by colonic injection of 0.5% acetic acid (AA) in 10-day-old rats and experiments were conducted at 8 weeks of age. Dorsal root ganglion (DRG) neurons innervating the colon were labelled by injection of DiI (1,1′-dioleyl-3,3,3′,3-tetramethylindocarbocyanine methanesulfonate) fluorescence into the colon wall.

Results: Visceral hypersensitivity was reversed by TNP-ATP (2′-(or-3′)-O-(trinitrophenyl) ATP), a potent P2X1, P2X3 and P2X2/3 receptor antagonist. Rapid application of ATP (20 μM) induced a fast inactivating current in colon-specific DRG neurons from both control and AA-treated rats. There was a twofold increase in the peak ATP responses in neurons from AA-treated rats. These currents were sensitive to TNP-ATP (100 nM). Under current-clamped conditions, ATP evoked a larger membrane depolarisation in neurons from neonatal AA-treated rats than in controls. P2X3R protein expression was significantly enhanced in colon-specific DRGs 8 weeks after neonatal AA treatment.

Conclusions: These data suggest that the large enhancement of P2XR expression and function may contribute to the maintenance of visceral hypersensitivity, thus identifying a specific neurobiological target for the treatment of chronic visceral hyperalgesia.

Statistics from

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.

ATP is increasingly being recognised as a participant in diverse biological events including neurotransmission and pain,16 mediated by a variety of P2X receptors (P2XRs).79 Sensory neurons in dorsal root ganglion (DRG) express P2XRs, ligand-gated cation channels that are assembled from at least seven different subunits.1013 Spinal afferents predominantly display homomeric P2X3 and heteromeric P2X2/3 subtypes14 and this system has been shown to be important in several animal models of visceral injury including the urinary bladder and colon.1517 In visceral organs, where distension is a particularly painful stimulus, ATP may participate in mechanosensory nociception. Excessive stretch causes release of ATP from epithelial cells which then acts on P2X3 homomultimeric and P2X2/3 heteromultimeric receptors on intramural spinal nerves triggering pain signalling in urological organs (bladder and ureter)1517 and the colon.18 Further, this pathway appears to be amplified in inflammatory conditions.19 Although these data imply a role for P2XRs in inflammatory visceral pain, it is not known whether they also contribute to pain signalling in common “functional” bowel disorders such as irritable bowel syndrome (IBS). We addressed this question using a previously validated rat model of chronic functional visceral hyperalgesia. This model is developed by a transient noxious stimulus (dilute acetic acid (AA)) delivered to the colorectum of neonatal rats which results in persistent colonic sensory dysfunction in adults that is not accompanied by any evidence of inflammation or injury, hence mimicking a functional visceral pain disorder such as IBS.20 Our findings implicate an important role for ATP signalling in IBS-like visceral hyperalgesia and identify the P2XR as a potential molecular target for the treatment of this condition.


Induction of chronic visceral hyperalgesia

Experiments were performed on male Sprague–Dawley rats. Care and handling of these animals were approved by the Institutional Animal Care and Use Committee at the University of Texas Medical Branch and were in accordance with the guidelines of the International Association for the Study of Pain. Ten-day-old pups received an infusion of 0.2 ml of 0.5% AA solution in saline into the colon 2 cm from the anus.20 Controls received an equal volume of saline. Experiments were performed in these rats between 8 and 12 weeks of age.

Cell labelling

Colon-specific DRG neurons were labelled by injection of 1,1′-dioleyl-3,3,3′,3-tetramethylindocarbocyanine methanesulfonate (DiI; Invitrogen, Carlsbad, California, USA) into the colon wall as described previously.20 In brief, animals were anaesthetised with a cocktail of ketamine (80 mg/kg) and xylazine (5–10 mg/kg, intraperitoneally) After a midline laparotomy, DiI (25 mg in 0.5 ml methanol) was injected in 2 μl volumes at 10–15 sites on the exposed colon extending from the level of the bladder to about 6 cm in an oral direction. To prevent leakage and possible contamination of adjacent organs with the dye, the needle was left in place for 1 min and each injection site was washed with normal saline following each injection. The colon was gently swabbed prior to closing the abdomen. Animals were returned to their housing and given free access to drinking water and standard food pellets.

Dissociation of DRG neurons and patch clamp recording

Isolation of DRG neurons from adult Sprague–Dawley rats has been described previously.2123 Briefly, 8 weeks after neonatal AA or vehicle injection, rats were killed by cervical dislocation, followed by decapitation. DRGs (T13–L2 and L6–S2) were bilaterally dissected out and transferred to an ice-cold, oxygenated fresh dissecting solution, containing (in mM): 130 NaCl, 5 KCl, 2 KH2PO4, 1.5 CaCl2, 6 MgSO4, 10 glucose and 10 HEPES, pH 7.2 (osmolarity =  305 mOsm). After removal of the connective tissue, the ganglia were transferred to a 5 ml dissecting solution containing collagenase D (1.8–2.0 mg/ml; Roche, Indianapolis, Indiana, USA) and trypsin (1.2–1.5 mg/ml; Sigma, St Louis, Missouri, USA) and incubated for 1.5 h at 34.5°C. DRGs were taken from the enzyme solution, washed and transferred to 2 ml of the dissecting solution containing DNase (0.5 mg/ml; Sigma). A single cell suspension was subsequently obtained by repeated trituration through flame-polished glass pipettes. Cells were plated onto acid-cleaned glass coverslips.

Coverslips containing adherent DRG cells were put in a small recording chamber (0.5 ml volume) and attached to the stage of an inverting microscope (Olympus, Tokyo, Japan), fitted for both fluorescence and bright-field microscopy. DiI-labelled neurons were identified by a fluorescence microscope at room temperature with normal external solution containing (in mM): 130 NaCl, 5 KCl, 2 KH2PO4, 2.5 CaCl2, 1 MgCl2, 10 HEPES and 10 glucose, pH 7.4 with NaOH. Osmolarity, 295–300 mOsm. For the patch clamp recording experiments, cells were continuously superfused with normal external solution. Recording pipettes were pulled from borosilicate glass tubing using a horizontal puller (P-97, Sutter Instruments, Novato, California, USA) and typically had a resistance of 1.5–3.0 MΩ when filled with normal pipette solution. For perforated patch recording, 3 μl of a 50 mg/ml stock solution of amphotericin B (Calbiochem, La Jolla, Califonia, USA) in dimethylsulfoxide (DMSO; Sigma) was added to 0.5 ml of pipette solution to yield a final concentration of 200 μg/ml. The pipette tip was initially filled with amphotericin B-free solution, containing (in mM) 100 KMeSO3, 40 KCl and 10 HEPES, pH 7.25 adjusted with KOH (290 mOsm). The pipette was then back-filled with the amphotericin B/pipette solution before being used immediately to obtain a giga-ohm seal. Tip potentials were zeroed before membrane–pipette seals were formed. Perforation of the membrane patch, as revealed by the appearance of slow capacitance transients, occurred within 5–25 min and recordings were only made when access resistance fell to <15 MΩ. The membrane was clamped at −60 mV by a Dagan 3911 patch clamp amplifier (Dagan Corporation, Minneapolis, Minnesotta, USA). Capacitive transients were corrected using capacitive cancellation circuitry on the amplifier that yielded the whole-cell capacitance and access resistance. Up to 90% of the series resistance was compensated electronically. Considering the peak inward current amplitudes of <5 nA, the estimated voltage errors from the uncompensated series resistance would be <5 mV. The leak currents at −60 mV were always <20 pA and were not corrected. The currents were filtered at 2–5 kHz and sampled at 50 or 100 μs per point. ATP-evoked currents and voltage were recorded with a Dagan 3911 patch clamp amplifier; and data were acquired and stored on a Dell computer for later analysis using pCLAMP 9.2 (Axon Instruments, Sunnyvale, California, USA). All experiments were performed at room temperature (∼22°C).

Application of drugs

ATP (Sigma) and TNP-ATP (2′-(or-3′)-O-(trinitrophenyl) ATP, sodium salt, Molecular Probes, Eugene, Oregon, USA) were in the external solution and in solution in saline, respectively. ATP and TNP-ATP was pressure delivered to the recorded cells through two applicators to avoid limiting peak ATP response.6 For behavioural experiments, TNP-ATP (50 μmol/kg) was intraperitoneally injected 5 min before measurement of the nocifensive behavioural response to colorectal distension (CRD).

Western blotting and immunostaining

Protein extracts from pooled DRGs (T13–L2 and L6–S2) were prepared in SDS buffer: 50 mM Tris-HCl, 133 mM NaCl, 2% SDS, 1 mM dithiothreitol (DTT), 1 mM phenylmethylsulfonyl fluoride (PMSF), 1:100 dilution of protease inhibitor cocktail (Sigma), pH 8. A 25 μg aliquot of protein was fractionated on 10% polyacrylamide gels (Bio-Rad, Hercules, California, USA). Proteins were transferred to polyvinylidene difluoride (PVDF) membranes (Bio-Rad) at 25 V overnight at 4°C, blocked for 2 h in TBS (50 mM Tris-HCl, 133 mM NaCl, pH 7.4) containing a 5% dilution of Carnation non-fat milk powder and incubated with primary antibody (anti-P2X3 at 1:1000 or P2X1 at 1:200) for 2 h in TBS containing 1% milk at room temperature. After washing in TBST (0.5% Tween-20), membranes were incubated with horseradish peroxidase (HRP)-conjugated secondary antibodies (1:5000, Santa Cruz Biotechnology, Santa Cruz, California, USA) in TBS containing 1% milk for 1 h at room temperature. Bands were visualised using an enhanced chemiluminescence (ECL) kit (Amersham, Buckinghamshire, UK) and appropriate exposure to Kodak X-ray film. Membranes were subsequently stripped and re-probed for actin (anti-actin 1:5000, Chemicon, Temecula, California, USA). Films were scanned and band intensities measured using Optic Quant software (Packard Instruments, Meriden, Connecticut, USA). P2X3R and P2X1R protein expression was normalised to actin.

For immunofluorescence staining of P2X3R, 1–2 weeks after injection of DiI, rats were perfused transcardially with 150 ml of phosphate-buffered saline (PBS) followed by 400 ml of ice-cold 4% paraformaldehyde (PFA) in PBS. DRGs T13–S2 were removed and postfixed for 1 hour in PFA and cryoprotected overnight in 20% sucrose in PBS. Sections (10 μm) on plus slides were incubated sequentially with P2X3R antibody (Neuromics, Edina, Minnesota, USA, 1/200) and anti-rabbit Alexa Fluor 488 (Invitrogen; 1/200). Sections were viewed using a BX60 Olympus (Center Valley, Pennsylvania, USA) microscope equipped with filter cubes appropriate for DiI (rhodamine filter) and appropriate band pass and barrier filter for Alexa 488. Images were captured and analysed using Metaview software (Nikon, Melville, New York, USA). To ensure that a neuron was counted only once, serial sections were placed on consecutive slides with at least 50 μm between sections on the same slide.

Behavioural testing for nocifensive responses

Visceral hypersensitivity was measured 8 weeks after neonatal AA treatment by grading the behavioural response of rats to CRD as described previously.20 24 Briefly, under mild sedation (1% Brevital (25 mg/kg intraperitoneally)), a flexible latex balloon (5 cm) attached to a tygon tubing was inserted 8 cm into the descending colon and rectum via the anus and held in place by taping the tubing to the tail. Rats were placed in small Lucite cubicles and allowed to adapt for 30 min. CRD was performed by rapidly inflating the balloon to constant pressure using a sphygmomanometer. The balloon was inflated to 20, 40 and 60 mm Hg, for 20 s followed by 2 min rest. Behavioural response to CRD was measured by visual observation of the abdominal withdrawal reflex (AWR) by a blinded observer, and AWR scores were scored either 1 (normal behavior), 2 (contraction of abdominal muscles), 3 (lifting of abdominal wall) or 4 (body arching and lifting of pelvic structures).

Data analysis

No neuron with a resting membrane potential more depolarised than −40 mV was included in the data analysis. All data are expressed as mean (SEM). Statistical significance was determined by Student t test or analysis of variance (ANOVA) test, as appropriate. p Values <0.05 were considered significant.


Chronic functional colonic hyperalgesia is mediated by P2X3Rs

Visceral sensitivity was determined by measuring the AWR scores in response to CRD at 8–10 weeks of age. In accordance with our previously reported results,20 the AWR scores were higher in sensitised rats at 20, 40 and 60 mm Hg distension pressures compared with controls (fig 1A, *p<0.05). We then determined the change in responsiveness to CRD after administration of TNP-ATP, a potent P2X1, P2X3 and P2X2/3 receptor antagonist, 5 min before balloon distension. We used a relatively high concentration of this agent (50 μmol/kg in 0.5 ml, intraperitoneally), a dose similar to those selected by other investigators.25 In sensitised rats, TNP-ATP treatment caused a significant decline in the mean AWR scores to CRD at 20, 40 and 60 mm Hg when compared with vehicle injection (fig 1B). In control rats, TNP-ATP had no significant effects on the AWR scores (fig 1C), suggesting that this agent did not act as a non-specific analgesic and that P2XRs do not normally participate in the responses to CRD.

Figure 1 Effects of TNP-ATP (2′-(or-3′)-O-(trinitrophenyl) ATP) pretreatment on behavioural response to colorectal distension (CRD). Abdominal withdraw reflex (AWR) scores were used as a function of distension pressure (20, 40 and 60 mm Hg). (A) Neonatal acetic acid (AA) treatment significantly enhanced sensitivity to CRD when measured at 8 weeks of age. CON, saline-treated rats, n = 9; AA, acetic acid-treated rats, n = 8. *p<0.05. (B). TNP-ATP treatment significantly reduced AWR scores in neonatal AA-treated rats (*p<0.05, compared with normal saline (NS), n = 9 for both groups). (C) TNP-ATP has no effect on AWR scores in neonatal saline-treated control rats (n = 6 for both groups).

Chronic functional colonic hyperalgesia is associated with potentiation of ATP-activated currents of colon-specific neurons

To test the hypothesis that ATP signalling is enhanced in this model, currents evoked by ATP in single colon-specific DRG neurons were measured using perforated patch recording techniques. Since both thoracolumbar (TL) and lumbarsacral (LS) DRG neurons innervate the colon through splanchnic and pelvic nerves,2628 we divided the colon-specific DRG neurons into two groups: (a) TL group (including T13, L1 and L2 DRGs) and (b) LS group (including L6, S1 and S2 DRGs). DiI-labelled neurons were isolated 8 weeks after AA or saline infusion and were used for the study. At −60 mV holding potential, ATP (20 μM) evoked fast inactivating currents in the majority of both TL and LS neurons from control rats (fig 2A). Since the slow inactivating current was evoked in <10% of neurons and this figure did not change after neonatal AA treatment, only the fast inactivating currents were included in this study. The average peak current density obtained from control rats was 3.9 (1.1) pA/pF (n = 10) in TL DRGs and 6.0 (1.3) pA/pF, (n = 19) in LS DRGs, respectively. In sensitised animals, these values were increased by a factor of more than twofold in both TL (9 (0.08) pA/pF, n = 12) and LS (16.1 (0.08) pA/pF, n = 14) DRG neurons (*p<0.05, fig 2B,C).

Figure 2 Potentiation of ATP-induced currents in neonatal acetic acid (AA)-treated rats. (A) Examples of currents evoked by ATP application (20 μM) in thoracic/lumbar (TL, left) and lumbar/sacral (LS, right) dorsal root ganglion (DRG) neurons from control rats. The cell membrane was held at –60 mV. The solid line above each trace indicates the duration of ATP application. (B) Examples of currents evoked by ATP application in TL and LS neurons from rats 8 weeks after neonatal AA treatment. Under similar conditions, ATP evoked large currents in AA-treated animals. (C) The mean peak current (I) density, measured by dividing the current amplitude by the whole-cell membrane capacitance, in TL and LS neurons from AA-treated rats was 2.4- and 2.2-fold greater than those in control animals (CON), respectively (*p<0.05).

Since ATP is a non-selective purinergic agonist and activates more than one subunit of P2XRs,2 we then wished to identify the predominant subtypes of P2XR in colonic afferents in both TL and LS DRGs. We first used the potent P2X1, P2X3 and P2X2/3 receptor antagonist TNP-ATP to determine whether homomeric P2X2Rs were present in colon-specific DRG neurons.2 29 30 At 100 nM, TNP-ATP blocked all fast ATP currents in TL and LS DRG neurons from both control and AA-sensitised rats (fig 3), suggesting that they were mediated by one or more of these receptor subtypes.

Figure 3 Inhibition of ATP-evoked inward currents by TNP-ATP (2′-(or-3′)-O-(trinitrophenyl) ATP). Pretreatment with TNP-ATP (100 nM), a potent P2X1, P2X3 and P2X2/3 receptor antagonist, completely blocked the ATP-induced inward currents both in control (A, C) and in the acetic acid (AA)-treated group (B, D). The ATP- (20 µM) induced inward currents in the control and AA-treated group of rats (left) were completely blocked by TNP-ATP (middle). These currents were partially recovered after a 3 min wash (right).

Voltage dependence of ATP-activated currents in colon-specific neurons

We then determined the voltage dependence of fast ATP currents in colon-specific DRG neurons from control and sensitised rats. Currents in response to ATP were measured at different holding potentials (from −100 to 40 mV). The peak current versus voltage (I–V) curves were plotted. The ATP currents obtained from both control and AA-treated groups demonstrated an inward rectification and reversed at about +5 mV (fig 4), indicating that these currents were due to non-specific cationic channels and that neonatal AA treatment did not alter the permeation properties of P2XRs of colon-specific sensory neurons. The kinetic characteristics of the fast ATP responses were also studied. The fast ATP current was rapidly activating and inactivating. The rise time of the fast current was between 8 and 35 ms, consistent with our previous reports.6 21 The inactivation time ranged from 258 to 426 ms. Neonatal AA treatment did not alter the kinetic characteristics of fast ATP currents.

Figure 4 Current–voltage (I–V) relationships of ATP-evoked currents. (A,B) Examples of ATP-induced current traces in neurons from control (A, left) and acetic acid (AA)- (B, left) treated rats. The I–V relationships of ATP currents exhibit a steep voltage dependence in colon-specific DRG neurons from both control (A, right) and neonatal AA-treated (B, right) rats. ATP-evoked currents (I) were measured at different holding potentials. The reversal potentials of the currents did not change after AA treatment. Data were obtained from five different cells.

Chronic functional colonic hyperalgesia is associated with enhancement of ATP-induced membrane depolarisation

We next compared the effect of ATP (20 μM) on the membrane depolarisation of colon-specific DRG neurons from control and AA-treated rats under current clamp conditions (fig 5). Application of ATP to recorded neurons induced local depolarisation (fig 5A) or provoked action potential (AP, fig 5B). These responses were also completely blocked by TNP-ATP (100 nM, fig 5A, B). The percentage of ATP-responsive cells was similar in both control and sensitised rats (TL: CON, 83%, n = 12; AA, 86%, n = 14, p>0.05; LS: CON, 77%, n = 26; AA, 85%, n = 20, p>0.05, fig 5C). However, the number of APs induced by ATP application was different between control and neonatal AA-treated rats. Only one TL neuron (∼20%) and three LS neurons (∼48%) from control rats exhibited APs after ATP application, while ∼50% of TL neurons and 78% of LS neurons from AA-treated rats displayed APs after ATP application (fig 5D). Since these differences could have been due to upregulation of sodium channels, as has been reported in inflammatory models of pain,31 32 we minimised their contribution by using a depolarising prepulse to inactivate both TTX-sensitive and TTX-resistant sodium channels.33 34 Under these conditions, the application of ATP produced smaller depolarisations but they were still significantly higher in neonatal AA-treated rats than in controls (fig 6).

Figure 5 Membrane depolarisation evoked by application of ATP (20 μM). (A, B) Pretreatment with TNP-ATP (2′-(or-3′)-O-(trinitrophenyl) ATP; 100 nM) completely blocked both the ATP-induced membrane depolarisation in control (A) and the action potential (AP) in the acetic acid (AA)-treated group (B). (C) The percentages of ATP-responsive colon neurons from both thoracic/lumbar (TL) and lumbar/sacral (LS) dorsal root ganglia (DRGs) were not significantly different between control and neonatal AA-treated rats. (D) The percentage of colon neurons with AP after ATP application was significantly higher in AA-treated rats than in control rats. ATP application evoked AP only in one colon neuron of the TL DRG (∼20%) and in three LS DRG neurons (∼48%) from the control group. However, there were ∼50% of TL neurons and 78% of LS neurons from AA-treated rats discharged after ATP application (D). *p<0.05,
Figure 6 Neonatal acetic acid (AA) treatment increases ATP-evoked depolarisations in colon-specific dorsal root ganglion (DRG) neurons. (A) ATP application (20 μM) evoked one action potential (AP) in a colon-specific DRG neuron from neonatal AA-treated rats. Resting potential was −51 mV. (B) A prepulse depolarisation to −10 mV for a period of 8 s was used to inactive Na+ channels before the ATP application. With the prepulse, ATP application did not evoke an AP but produced a 43 mV depolarisation in the same neuron as in A. Solid lines under current traces indicate the period of ATP application (A) and the period of prepulse and ATP application (B). A bar graph showed a significant increase in ATP-induced membrane depolarisation in the AA-treated group compared with controls (C, *p<0.05).

Chronic functional colonic hyperalgesia is associated with upregulation of P2X3R expression in colonic DRG

As mentioned previously, the attenuation of nociceptive responses by TNP-ATP suggests the involvement of P2X1, P2X2/3, P2X3, but not P2X4 and P2X5, receptor subtypes. Although there is very limited P2X1R expression in DRG neurons,13 we performed western blot analysis on protein extracts from TL and LS DRGs and did not detect any changes in response to neonatal AA treatment (fig 7). In contrast, 8 weeks after colonic exposure to AA or saline, the expression of P2X3Rs in LS DRGs from AA-treated rats was significantly greater than that from control rats (fig 7C,D, p<0.05). A similar trend was noted in TL DRGs, but this was not statistically significant (fig 7A,B, p>0.05). We also examined P2X3R expression in DiI-labelled colon-specific neurons by immunohistochemistry (fig 8). While the total number of DiI-labelled DRG neurons did not change between control and AA-treated rats, the proportion expressing P2X3Rs was significantly higher in both TL (fig 8A,C) and LS DRG (fig 8B,D) neurons from AA-treated rats as compared with controls (TL, 5.5% vs 18.2%, p<0.02; LS, 27.8% (1.5%) vs 56.5% (8.8%), p<0.001, fig 8).

Figure 7 P2X1 and P2X3 receptor protein expression in thoracic/lumbar (TL) and lumbar/sacral (LS) dorsal root ganglia (DRGs) from 8-week-old rats treated with either saline (CON) or acetic acid (AA) as neonates. Western blots showing P2X1 and P2X3 receptor protein expression in TL (A) and LS (B) DRGs from CON and AA-treated rats. Actin was used as loading control for each group. (C, D) Bar graphs show that neonatal AA reatment significantly enhanced the P2X3 receptor expression in LS DRGs (D). n = 4 for each group. *p<0.05 compared with controls.
Figure 8 Neonatal acetic acid (AA) treatment enhanced the percentage of P2X3 receptor (P2X3R)-positive colon-specific dorsal root ganglion (DRG) neurons. (A, B) P2X3R immunofluorescence (green) in DiI (1,1′-dioleyl-3,3,3′,3-tetramethylindocarbocyanine methanesulfonate)-labelled (red) colon neurons in sections from thoracic/lumbar (TL; A) and lumbar/sacral (LS; B) DRGs. P2X3Rs are predominantly expressed in small and medium but not in large DRG neurons from control (CON) and AA-treated rats. Arrows designate neurons showing co-localisation between P2X3R and DiI; arrowheads mark DiI-labelled neurons in which P2X3R immunofluorescence was not detected. Bar = 100 μm. (C, D) Bar graphs showing that the percentage of colon-specific TL (C) and LS (D) DRG neurons expressing detectable P2X3Rs was significantly higher in neonatally AA-sensitised, adult rats compared with controls (*p<0.05).


Over the past several decades, convincing experimental evidence has accumulated to implicate a significant role for ATP and P2XRs in pain signalling in a variety of settings.6 14 3538 In visceral organs, where distension is a particularly painful stimulus, ATP has been postulated to mediate mechanosensory nociception. According to this paradigm, excessive stretch causes release of ATP from epithelial cells which then act on P2X3 homomultimeric and P2X2/3 heteromultimeric receptors on intramural spinal nerves, triggering pain signalling. This was first demonstrated in urological organs (bladder and ureter),1517 but subsequently also shown to apply to the colon.18 Further, this pathway appears to be amplified in inflammatory conditions, suggesting an important role in nociceptive sensitisation. Thus, Wynn et al19 used an in vitro rat model to study the effects of colitis (induced by 2,4,6-trinitrobenzene sulfonic acid (TNBS) 5–10 days prior to the experiments) on purinergic signalling and showed significant changes at several key steps: colitis enhanced the release of ATP in response to distension and the excitability of sensory nerves to ATP. Further, the number of DRG neurons responding to ATP was increased, accompanied by an increase in the number of those staining for the P2X3R.19 These results are consistent with those of other investigators studying this signalling system in inflammatory models of the urinary bladder and joints.39

The results of our study are significant because they provide further evidence that ATP–P2XR signalling may play an important role in “functional” visceral pain 40 41—that is, pain occurring in the absence of overt structural or inflammatory processes (fig 1). To prove this, we used a previously validated approach of mild colonic irritation in the neonatal period, which does not produce significant inflammation but results in visceral hypersensitivity that persists into adulthood.20 The present study adds P2XR to the list of key nociceptive molecules that participate in hypersensitivity in this model and underscores the fact that such visceral sensitisation is accompanied by long-lasting plasticity of sensory neurons.

A unique feature of our study is the systemic in vivo use of the potent P2X1, P2X3 and P2X2/3 receptor antagonist TNP-ATP in a model of colonic pain. This has previously been used in vivo in a mixed somatic–visceral model of writhing following peritoneal irritation in mice with an ED50 of 6.35 μM/kg25 and in other animal models.19 Since TNP-ATP has been reported to be broken down rapidly in vivo,42 we used a dose several fold higher than the ED50 in mice and were able to achieve a significant attenuation of the behavioural response to CRD in sensitised rats. No significant effect was seen in control rats, suggesting that this was not a non-specific analgesic effect. This also suggests that the role of the P2XR pathway in signalling colonic distension may not be as important in health as in the sensitised state. This is in keeping with the results of studies directly measuring nerve responses28 as well as others that have shown that there is a relatively minor contribution of ATP signalling to distension-induced nerve firing but that this is significantly enhanced in the presence of colitis.18 19

The most prominent change was the large enhancement of ATP-evoked current amplitudes (fig 2). The average peak current density of the fast responses in TL and LS DRG neurons from AA-treated rats was 2.4- and 2.2-fold larger than controls, respectively (fig 2C). Since ATP activates more than one subtype of P2XRs in control DRG neurons,2 it is of interest to identify the predominant subtypes of P2XR in both colon-specific TL and LS DRG neurons and to determine whether the same P2XR subtypes are expressed in DRGs after neonatal AA treatment. As compared with homomeric P2X2Rs, homomeric P2X1 and P2X3 and heteromeric P2X2/3 are more sensitive to TNP-ATP by a factor of ⩾500-fold; thus low concentrations (<1 μM) of TNP-ATP should block P2X1, P2X3 and P2X2/3 receptor-mediated responses but leave most homomeric P2X2 receptor-mediated responses intact.2 29 30 At 100 nM, TNP-ATP blocked all fast ATP currents in both TL and LS DRG neurons from both control and AA-sensitised rats (fig 3). Thus, the responses mediated by homomeric P2X2Rs, if present in our cells, would be small. The expression of P2X1R was not altered after neonatal AA treatment (fig 7). Although we cannot exclude a contribution from post-translational changes affecting this receptor, it is therefore unlikely that P2X1R is the major receptor type responsible for the potentiation of ATP-evoked responses in these experiments. In functional studies comparing the P2X3, P2X2/3 agonist α,β,me-ATP with ATP, we found that both produced similar responses, adding credence to our statement that P2X3R and P2X2/3R may be involved in chronic visceral hyperalgesia (unpublished observations). We can therefore tentatively conclude that homomeric P2X3R and heteromeric P2X2/3R play the most important role in the sensitisation of the response of colon-specific DRGs to ATP.

Possible mechanisms for the potentiation of ATP currents include an increase in single-channel conductance and channel opening probability, and/or upregulation of P2XR expression. An increase in the opening probability of ATP channels is not likely because the kinetic properties of ATP currents remain unchanged after neonatal AA treatment. The current–voltage (IV) curves of ATP-evoked currents show inward rectification in control and sensitised rats (fig 4). Furthermore, the IV curves show that ATP currents reversed at near +5 mV in control cells, and that AA treatment did not change the reversal potentials of ATP responses. Therefore, neonatal AA treatment had no significant effect on the permeation properties of P2XRs. A very likely mechanism contributing to the increase in receptor function is an increase in expression of P2X3Rs. The percentage of P2X3R-positive colon-specific DRG cells was significantly higher in both TL and LS DRG neurons 8 weeks after neonatal AA treatment (fig 8). Although P2X3R is normally found predominantly in small nociceptive sensory neurons that bind the lectin IB4, it is possible that the upregulation of P2X3R expression may be due to a higher proportion of calcitonin gene-related peptide (CGRP)-labelled small nociceptive neurons after neonatal AA treatment since Wynn et al19 have provided evidence suggesting that, after TNBS-induced colitis, the percentage of CGRP-labelled neurons that express P2X3R increases significantly. However, further experiments are needed to examine this. The western blotting analysis showed that, in contrast to LS DRGs, there was no significant increase in P2X3R protein expression in TL DRGs (fig 7). This may reflect the fact that only a relatively small number of all the DRG neurons are derived from the colon. However, in general, our findings suggest that LS afferents develop a more robust P2XR (in terms of both currents and changes in expression) to chronic sensitisation than TL neurons. These findings are in contrast to those of Brierly et al which showed that colonic serosal neurons of TL origin express P2X3R and respond much more vigorously than LS neurons.28 This may be explained by differences in species (mice vs rats) or methodology, but could also reflect the fact that neonatal sensitisation may preferentially affect LS nerves.

Our data also add to the literature on the differences between TL and LS pathways. The distal colon has dual afferent innervation, supplied by lumbar splanchnic nerves, originating in TL DRGs, as well as pelvic nerves, originating in LS DRGs. Although mechanosensitive nerves are present in both pathways, LS afferents appear to be more sensitive and responsive to circular stretch in health.28 Further, there is evidence to suggest differences in the role these two pathways play in pathological conditions. Thus, in normal rats, the response to CRD is eliminated by a LS rhizotomy 43 but not by cutting the splanchnic nerves.44 However, after the induction of colitis, CRD responsiveness is partially restored despite an LS rhizotomy,43 suggesting that the TL pathway becomes more important in disease states. Corroborative evidence for this hypothesis comes from other rodent studies 45 46 as well as in humans. Thus, CRD elicits referred pain in sacral dermatomes in healthy volunteers, a pattern that changes in patients with IBS with expansion into the TL dermatomes.47 Thus a model is emerging in which acute colorectal pain in health may be subserved by LS innervations, with TL afferents participating more vigorously only in sensitised states. Although our experiments were not designed to test this hypothesis, our results are generally compatible with it.

In conclusion, our data demonstrate that chronic functional visceral hyperalgesia is associated with potentiation of ATP-evoked responses and an enhanced expression of P2X3Rs in colon-specific sensory neurons. These data identify a specific molecular mechanism for functional colonic hypersensitivity that may pave the way for novel therapeutic strategies for IBS and related disorders.



  • See Commentary, p 1193

  • Competing interests: None.

  • Ethics approval: Care and handling of the animals were approved by the Institutional Animal Care and Use Committee at the University of Texas Medical Branch and were in accordance with the guidelines of the International Association for the Study of Pain.

Linked Articles