Attenuation of Colitis-Induced Visceral Hypersensitivity and Pain by Selective Silencing of TRPV1-Expressing Fibers in Rat Colon

Abstract

Background

Transient receptor potential vanilloid 1 (TRPV1) cation channels, expressed on nociceptors, are well established as key contributors to abdominal pain in inflammatory bowel disease (IBD). Previous attempts at blocking these channels have been riddled with side effects. Here, we propose a novel treatment strategy, utilizing the large pore of TRPV1 channels as a drug delivery system to selectively inhibit visceral nociceptors.

Methods

We induced colitis in rats using intrarectal dinitrobenzene sulfonic acid. Visceral hypersensitivity, spontaneous pain, and responsiveness of the hind paws to noxious heat stimuli were examined before and after the intrarectal application of membrane-impermeable sodium channel blocker (QX-314) alone or together with TRPV1 channel activators or blockers.

Results

Intrarectal co-application of QX-314 with TRPV1 channel activator capsaicin significantly inhibited colitis-induced gut hypersensitivity. Furthermore, in the model of colitis, but not in naïve rats, QX-314 alone was sufficient to reverse gut hypersensitivity. The blockade of TRPV1 channels prevented this effect of QX-314. Finally, applying QX-314 alone to the inflamed gut inhibited colitis-induced ongoing pain.

Conclusions

Selective silencing of gut nociceptors by a membrane-impermeable sodium channel blocker entering via exogenously or endogenously activated TRPV1 channels diminishes IBD-induced gut hypersensitivity. The lack of effect on naïve rats suggests a selective analgesic effect in the inflamed gut. Our results suggest that in the colitis model, TRPV1 channels are tonically active. Furthermore, our results emphasize the role of TRPV1-expressing nociceptive fibers in colitis-induced pain. These findings provide proof of concept for using charged activity blockers for the blockade of IBD-associated abdominal pain.

Introduction

Abdominal pain is the presenting symptom in up to 70% of patients with inflammatory bowel disease (IBD), and at least one-third of patients have persistent pain despite optimal anti-inflammatory treatment.¹ IBD-induced pain is associated with higher anxiety levels, depression, social dysfunction, and work disability, thus devastatingly affecting patients’ lives.^2,3 Available analgetic treatments are nonspecific, with variable effectiveness and considerable side effects, including addiction and increased morbidity.^4,5

Recent studies have made substantial progress in understanding gut innervation by nociceptive neurons and proposed potential molecular and cellular mechanisms underlying IBD-induced pain.^6-9 Several lines of evidence demonstrate the role of the transient receptor potential vanilloid 1 (TRPV1) channel in IBD-induced pain in human and animal models of colitis.^9-15 This large pore cation nonselective transducer channel is expressed by both somatic nociceptive neurons and nociceptive peptidergic neurons innervating the gut.¹⁵ The activation of these channels by various proinflammatory mediators contributes to inflammatory pain.^16,17 In animal models of IBD, systemic blockade of TRPV1 reduces the intensity of visceral hypersensitivity, implying that abnormal activity of TRPV1 contributes to IBD-induced visceral pain.¹⁰ Nevertheless, selective blockers of TRPV1 have failed in human studies due to life-threatening side effects, specifically hyperthermia.¹⁸

In this study, we aimed to explore the effect of selective inhibition of TRPV1-expressing nociceptive fibers in IBD-induced pain. We utilized a selective silencing approach we previously developed to produce pain-selective anesthesia and to examine the role of TRPV1-expressing fibers in somatic pain and itch.^19,20 In this approach, we used the membrane impermeant sodium channel blocker QX-314 and shuttled it into nociceptive neurons through TRPV1 channels activated by capsaicin. We showed that the co-application of capsaicin and QX-314 led to selective blockade of nociceptive activity and inhibited pain without affecting other sensory or motor modalities.¹⁹ Furthermore, we and others showed that tonically open TRPV1 channels, following endogenous activation by either pruritogens or proinflammatory mediators, were sufficient to allow entry of QX-314 into neurons, allowing to dissect the role of these neurons in pain and itch.^20,21

Here, we show that in dinitrobenzene sulfonic acid (DNBS)–induced colitis in rats, silencing TRPV1-expressing neurons by intrarectal co-application of capsaicin with QX-314 reverses gut hypersensitivity. These results suggest that TRPV1-expressing nociceptive neurons innervating the gut play a role in colitis-induced pain. We further show that intrarectal administration of QX-314 alone in colitis conditions was sufficient to abolish colitis-induced visceral hypersensitivity and ongoing pain. Notably, the application of QX-314 alone in naïve animals did not affect the normal gut sensitivity, suggesting a selective effect of QX-314 on colitis-induced pain. These results propose that in an IBD model, TRPV1 channels are tonically active, allowing entry of QX-314 into nociceptive fibers innervating the gut. Indeed, the pharmacological blockade of TRPV1 channels prevented the effect of QX-314 alone on colitis-induced gut hypersensitivity. Our results demonstrate that TRPV1 channels activated during IBD may serve as a natural drug delivery system, laying the foundation for developing this platform as an effective and selective treatment for IBD-induced abdominal pain.

Methods

Animals

All of the procedures performed in this study were in full accordance with the guidelines of the Hebrew University Animal Care Committee and were approved and ratified by the Committee. Procedures were performed on male and female adult (250-274 g) Sprague Dawley rats. The rats were housed in groups of 4 under a 12-hour light/dark cycle. The room temperature and humidity remained stable during the entire experiment, and food and water were regularly provided. Prior to all experiments described, the animals went through a sufficient period of habituation to the testing apparatus used and the experimental environment.

Colitis Induction

Following acclimation to the animal facility, animals were anesthetized in an isoflurane chamber with a continuous flow of 4% isoflurane. Once sedated, a nose cone with a continuous flow of 1% to 2% isoflurane was applied to maintain anesthesia. The animals’ hind legs were taped to the surface of a plastic container and placed at a 45° angle by raising the container edge. A total of 30 mg in 250 μL of DNBS solution (see below in "Drugs") was administered transanally using a flexible tube inserted 6 cm proximal to the anal canal. The animals were left at a 45° angle for 90 seconds to avoid expelling of the DNBS solution. Animals were then allowed to recover and monitored twice daily for 3 days to assess for clinical signs of inflammation and weight loss. All experiments were conducted on day 4 after colitis induction, following which the animals were humanely sacrificed and the ano-colon carefully dissected, removed, and examined to ensure colitis induction (Supplementary Figure 1).

Intrarectal Drug Delivery

All treatments were administered via enema in isoflurane-sedated animals, as described previously. During administration, the animals were placed at a 45° angle for 90 seconds to reduce the excretion of the drug/vehicle. Animals were allowed to recover for 10 minutes after awakening from anesthesia until fully mobile.

Drugs

DNBS (Sigma-Aldrich; 556971) was prepared as follows: 240 mg of DNBS was dissolved in 1 mL of 100% ethanol and 1 mL of DDW, and each animal was administered 250 μL (30 mg) of the solution. Capsaicin (Sigma-Aldrich; 21750) was prepared by dissolving 5 mg of capsaicin in 0.5 mL Tween (Sigma-Aldrich; P9416), 0.5 mL ethanol, and 4 mL DDW, and each animal was administered 1 mL of solution. The 2% lidocaine and 2% QX-314 (N-ethyl-lidocaine) solutions were prepared by dissolving 80 mg of either lidocaine or QX-314 in 4 mL DDW, and each animal was administered 1 mL of solution. Capsazepine (Sigma-Aldrich; P9416) was prepared by dissolving 5, 10, 12.5, or 15 mg of capsazepine in 2.5 mL Tween (Sigma-Aldrich; P9416) and 2.5 mL ethanol, and each animal received 1 mL of solution. The vehicle was prepared using the same solvent solution used for the active components in the experimental groups without the addition of the active component.

Assessment of Visceral Hypersensitivity

Colorectal distention by inflatable balloon was used to assess visceral hypersensitivity. Three radial holes were carved into the terminal end of a flexible tube. Then, a 6-cm balloon was constructed from a piece of latex (Durex) and was tied to the end of the flexible tube using dental floss (Dr. Collins). The other end of the tube was connected to a barometer with a tridirectional valve, and a 10 mL syringe was used to manually fill the balloon and maintain the required pressure. The balloon lacked inherent compliance at volumes of up to 10 mL, used in all experiments.

Animals were anesthetized, and the balloon was lubricated (Sion Biotext) and inserted into the distal colon, 0.5 cm from the anus. Fabric tape (Life) was used to secure the balloon to the tail, and the animals were allowed to recover for 10 minutes until fully mobile and alert. After that, the balloon was inflated to a pressure of either 40 mm Hg for measuring capsaicin- or colitis-induced hypersensitivity or 60 mm Hg in naïve rats (Figure 1A). The pressure was kept constant over the duration of 8 minutes, and abdominal withdrawal reflexes (AWRs) were visually counted and used as a measure of visceral hypersensitivity.²²

AWRs were defined as characteristic involuntary responses that involve sudden kyphosis, cervical extension, and occasional vocalization.²³ AWRs were quantified visually during the experiment, and all experiments were filmed and recorded for post hoc data validation.

Assessment of Noxious Thermal Threshold

The response to noxious pain stimuli was assessed using the Hargreaves heat latency test (Ugo-Basile). The rats were placed on a plexiglass surface, and a laser beam calibrated to heat the paw to 52 °C was directed toward the plantar aspect of the hind paw. Paw withdrawal latency, the time to a behavioral withdrawal response, was measured.^24,25

Assessment of Ongoing Pain

To test ongoing pain behavior without evocation, facial characteristic analysis (pain grimace scale [PGS]) was performed prior and following inflammation with and without treatments. Rats were habituated to the room and chambers for 3 days prior to experiments. Afterward, animals were placed in a plexiglass chamber, and headshots were taken randomly 3 times over 20 minutes prior to inducing DNBS colitis. Four days after inducing colitis, animals were placed in the plexiglass chambers and imaged randomly 3 times over 20 minutes. Then, each animal was given a dose of QX-314 (20 mg/mL) and left to rest for 10 minutes. Following the treatment, 3 random images were recorded over the span of 20 minutes. Images were coded by blinded testers, and photos were randomized and scored (with an intensity rate of 0, 1, or 2) by the following criteria: orbital tightening, nose/cheek bulge, ear position, and whisker change. In each case, 0 indicated complete absence of the characteristic, 1 indicated moderate presence, and 2 indicated severe presentation. The PGS was calculated by averaging the scores given for each parameter for each photo, and averaging the score for a single animal over the specified time period.²⁶

Statistical Analysis

Statistical analyses were performed using Prism 7 (GraphPad Software). Power analysis was performed based on previous reports of changes in gut hypersensitivity in IBD models in rodents²⁷ to predetermine sample size at a minimum of 5 rats in each experiment group. Rats were randomly allocated to groups in all experiments. The normality of data distribution was assessed using the Shapiro-Wilk test. In the experiments described in Figure 2, data were not normally distributed, and in these experiments, nonparametric tests were used. The Kruskal-Wallis test with post hoc Dunn’s test was used to compare more than 2 independent samples. The Mann-Whitney test was used for comparing 2 independent samples, and the Wilcoxon matched pairs signed rank test was used for paired values. For the normally distributed data, unpaired t test, paired t test, and ordinary and repeated-measures 2-way analysis of variance were used when appropriate. Actual P values are presented for each dataset. The criterion for statistical significance was P < .05. Boxplots presented in all figures depict the mean (solid line), 25th and 75th percentiles, and 1.5 SD. Boxplots of not normally distributed data (Figure 2) also depict the median (dotted line). In some experiments (PGS, Figure 3), both males and females were used. To assess the presence of the main effects of sex or interactions of sex with experimental manipulations, these datasets were analyzed using sex as a factor. In no cases were any main effects of sex or interactions observed, so data were collapsed. In some experiments, only male rats were used.

Results

Co-Application of Capsaicin With QX-314 Into an Inflamed Colon Diminishes Colitis-Induced Visceral Hypersensitivity

We examined the role of TRPV1-expressing nociceptive neurons in colitis-induced visceral hypersensitivity by monitoring the changes in visceral responsiveness to pressure following the silencing of these neurons. We used the AWR to estimate visceral hypersensitivity²⁷ (Figure 1A). We measured the number of AWRs evoked in various intracolonic pressures, in naïve conditions, and in DNBS-induced colitis conditions (4 days after treatment with DNBS).²⁸ In naïve animals, we did not detect any response when applying intracolonic pressure of 40 mm Hg (Figure 1B, before) (see Methods). Applying the same pressure to the animals 4 days after DNBS led to substantial gut hypersensitivity (Figure 1B, before treatment). All animals with increased gut sensitivity also showed prominent colonic inflammation (Supplementary Figure 1, see Methods). In these conditions, intrarectal application of capsaicin together with QX-314 produced a significant decrease in visceral hypersensitivity such that it reverted to the levels measured before inducing colitis (Figure 1B). Application of capsaicin alone to the inflamed gut led to substantial sensitivity to colorectal distention at 40 mm Hg; therefore, these experiments were terminated (data not shown). Application of the vehicle (see Methods) did not affect colitis-induced visceral hypersensitivity (see Figure 2B, Supplementary Figure 2C). These results show that silencing of TRPV1-expressing neurons inhibits colitis-induced visceral hypersensitivity, suggesting a prominent role of these fibers in visceral pain.

Application of QX-314 Alone Significantly Affects Visceral Sensitivity in Inflamed but not in Naïve Rats

Because QX-314 is a permanently charged compound, its penetration through the cell membrane is limited.²⁹ We and others have previously demonstrated that injection of QX-314 alone, either subcutaneously or perisciatically to naïve animals, in doses up to 2%, does not produce analgesia.^19,30,31 Similarly, intrarectal application of 2% QX-314 alone to naïve animals did not change their responsiveness to intracolonic pressure, whereas application of 2% lidocaine fully inhibited it (Figure 2A, Supplementary Figure 2A, B). Notably, to evoke AWR in naïve animals, we used a higher pressure of 60 mm Hg because 40 mm Hg was insufficient to induce AWRs in naïve rats (see Figure 1B, before).

Surprisingly, applying 2% QX-314 alone in colitis conditions significantly reduced gut sensitivity (Figure 2B, Supplementary Figure 2D). The QX-314–induced analgesia was similar to that produced by applying 2% lidocaine or 2% QX-314 applied together with capsaicin (Figure 2B).

Application of QX-314 Alone Significantly Reduces Colitis-Induced Nonreflexive, Ongoing Pain

So far, we assessed colitis-induced pain by measuring changes in AWR evoked by applying visceral distension. This widely accepted approach provides a good estimation of changes in evoked visceral sensitivity in various conditions.^7,32 Other important and clinically relevant questions are whether DNBS-induced colitis leads to nonevoked ongoing pain, and what is the effect of selective silencing of visceral nociceptors on ongoing pain. To answer these questions, we assessed the levels of colitis-induced ongoing pain by evaluating a rat pain grimace scale (PGS).²⁶ As expected, before inducing colitis, rats did not exhibit facial characteristics of ongoing pain (Figure 3, Supplementary Figure 3A). Four days after inducing colitis, the PGS score was significantly increased, as rats exhibited substantial orbital tightening, cheek bulging, ear flattening, and whisker stiffening (Supplementary Figure 3B), the characteristic markers of spontaneous pain.²⁶ Notably, the application of 2% QX-314 alone alleviated the signs of spontaneous pain, reversing the PGS score to the values measured before colitis induction (Figure 3, Supplementary Figure 3C).

Altogether, our results show that applying QX-314 alone is sufficient to reduce both colitis-induced evoked visceral hypersensitivity and ongoing pain. These results imply that in the inflamed gut, QX-314 does not require exogenic activation of TRPV1 channels to access the cytoplasm of nociceptive neurons.

Local Blockade of TRPV1 Channels Prevents the Analgesic Effect of QX-314

What could explain the analgesic effect of membrane impermeant QX-314 in an inflamed colon? DNBS-induced gut inflammation could potentially lead to an increase in mucosal and blood vessel permeability, allowing QX-314 to reach circulation and affect pain systemically. If the latter is the case, QX-314 applied to the inflamed guts should have a general inhibitory effect on the responses to noxious stimuli. However, this is less likely considering the low membrane permeability of QX-314. Indeed, in colitis conditions, we did not detect any effect of QX-314 applied intrarectally on the thermal pain threshold measured from the rats’ paws (Figure 4; see Methods and Supplementary Movie 1).

Alternatively, QX-314 could enter nociceptive neurons innervating the gut via inflammation-mediated tonically active TRPV1 channels. The role of tonically active TRPV1 channels in inflammatory conditions, but not in IBD models, was previously suggested.^21,33 Moreover, it has been shown that this activation of TRPV1 channels is sufficient to allow the entry of QX-314 into neurons, leading to the inhibition of their activity.^20,21,33 To examine the hypothesis that in the colitis model as well, QX-314 accesses nociceptive neurons via tonically active TRPV1 channels, we pharmacologically inhibited TRPV1 channels and studied the effect of QX-314 on gut sensitivity in these conditions. To that end, we used TRPV1 selective antagonist capsazepine.³⁴ Because capsazepine has not been used before to inhibit colorectal TRPV1 channels, we first determined the dose of capsazepine sufficient to inhibit gut TRPV1 channel activation (Figure 5A). We administered naïve rats with subsequently increasing doses of intrarectal capsazepine followed by 1 mg/mL capsaicin. As expected, capsazepine dose dependently reduced capsaicin-evoked gut hypersensitivity (Figure 5A). We chose the dose of capsazepine, which fully prevented capsaicin-induced gut hypersensitivity (3 mg/mL) (Figure 5A), and applied capsazepine to the DNBS-treated animals (Figure 5B). In these conditions, the co-application of capsazepine and QX-314 prevented the analgesic effect of QX-314 (Figure 5B, Supplementary Figure 4A), which we previously described (see Figure 2B). The application of capsazepine alone did not affect colitis-induced visceral hypersensitivity (Figure 5B, Supplementary Figure 4B).

These data imply that TRPV1 channels are tonically active in DNBS-induced colitis and suggest that QX-314 affects gut hypersensitivity by entering nociceptive neurons, at least in part, via these tonically active TRPV1 channels.

Discussion

Despite the plethora of anti-inflammatory drugs available or in the pipeline for the treatment of IBD-associated inflammation, there is an unmet need for treating IBD-associated abdominal pain. In the current article, we advance the understanding of the molecular and cellular pathways involved in abdominal pain in a rodent model of IBD, with potential implications for drug development. We showed that local administration of a membrane impermeant sodium channel blocker QX-314 into the inflamed colon of a rat caused a decrease in visceral hypersensitivity and ongoing pain behavior. This effect of QX-314 was selective to the colon because we did not witness changes in sensitivity to noxious stimuli in the areas beyond the viscera. We further showed that the analgesic effect of QX-314 was specific to colitis and was not seen in naïve animals. Furthermore, we showed that the analgesic effect of QX-314 was prevented by co-administering QX-314 with capsazepine, a TRPV1 channel blocker. Taken together, our findings provide novel evidence proposing that TRPV1 channels in the inflamed gut may be tonically active, allowing the entry of charged and, therefore, membrane impermeant analgesic molecules selectively into nociceptive fibers.

Our results also suggest that abnormal activity of TRPV1-expressing nociceptive fibers is essential for generating colitis-induced pain, by demonstrating that selective silencing of TRPV1-expressing nociceptive neurons (by applying capsaicin and QX-314) produces a substantial reduction of colitis-induced pain. The abnormal activity of TRPV1-expressing nociceptive fibers may originate from inflammation-induced changes in the expression and functional properties of TRPV1 channels.^10,35 Previous studies demonstrated that colitis is associated with increased expression levels of TRPV1 in DRG neurons innervating the gut.^11,36 Importantly, increased expression of TRPV1 was also observed in human patients with ulcerative colitis.³⁷ Our results showing that QX-314 applied alone in colitis conditions leads to an analgesic effect, and that this effect is prevented by capsazepine, suggest that in these conditions, TRPV1 channels are tonically active. The increased expression of TRPV1 reported previously,¹⁰ and their colitis-induced spontaneous activity could contribute to the abnormal activation of TRPV1-expressing nociceptive neurons and, therefore, underlie IBD-induced pain. These results yet again emphasize the role of TRPV1 channels expressed on nociceptive neurons as both culprits and possible therapeutic targets in IBD-associated pain and inflammation. Based on these results, blockade of TRPV1 channels should alleviate IBD pain. Indeed, it has been demonstrated that chronic treatment with capsazepine decreases colitis-induced gut hypersensitivity.³⁸ However, we show that the acute application of capsazepine was not sufficient to alleviate colitis-induced hypersensitivity (Figure 4B, Supplementary Figure 4B). Nonetheless, it was sufficient to prevent the effect of QX-314 plausibly by minimizing the tonic opening of TRPV1 channels and limiting the entry of QX-314 into TRPV1-expressing nociceptive fibers.

Altogether, these results suggest that prolonged continuous inhibition of TRPV1 channels could be beneficial for the treatment of IBD-induced pain.¹⁰ Unfortunately, no current therapeutic options exist for the selective blockade of TRPV1 channels, as the available TRPV1 antagonists cause substantial side effects in humans.¹⁸ Here, instead of inhibiting TRPV1 channels to reduce colitis-induced hypersensitivity, we suggest silencing nociceptive neurons selectively. This approach affects colitis-induced hypersensitivity and pain plausibly by blocking the afferent activity of nociceptive neurons, as we demonstrated before in somatic nociceptive neurons.^19,20,39 Furthermore, this approach could also inhibit the efferent antidromic activity of nociceptive neurons, thereby diminishing neurogenic inflammation,^21,33 which could further decrease the inflammatory burden in IBD.

The lack of the acute analgesic effect of capsazepine may also suggest that the inflammation-mediated tonic activity of TRPV1 channels is not the only mechanism underlying the abnormal activity of nociceptive fibers innervating the gut. Inflammation-induced modulation of other channels beyond TRPV1, such as TRPA1 and Piezo2,^8,40-45 could also lead to abnormal activity of TRPV1-expressing nociceptive fibers, thus underlying colitis-induced pain. Nevertheless, the fact that these channels are expressed on TRPV1-containing neurons^8,41 suggests that our selective silencing approach would still block the activity of these neurons regardless of the molecular correlate of their abnormal activity.

Notably, although our capsaicin and QX-314 experiment showed a substantial reduction of colitis-induced visceral hypersensitivity and ongoing pain by a selective blockade of TRPV1-expressing fibers (Figure 1B), we cannot exclude that abnormal activity of other fibers innervating the gut also contributes to IBD-induced pain.

While in this study we used QX-314 as a proof of concept for selective silencing of visceral nociceptors, the use of QX-314 by humans is limited by its neurotoxicity,⁴⁶ and, therefore, not viable for a translational approach. Newly developed charged ion channel blockers³³ may prove to be safer to use in humans, especially if delivered locally to the gut. Further challenges will be to ensure a long-acting effect and to fine-tune these platforms to maximize their effectiveness.

Conclusions

In summary, our results suggest the central role of TRPV1-expressing nociceptive neurons and tonic activation of TRPV1 channels in IBD pain. Moreover, our findings provide a conceptual approach for selective inhibition of IBD pain by buttressing a basis for employing charged activity blockers to selectively and effectively block visceral pain.

Supplementary data

Supplementary data is available at Inflammatory Bowel Diseases online.

Funding

This work was supported by the Israel Science Foundation (grant agreement 1202/23), Israeli Cancer Research Foundation (Brause Initiative Application; grant agreement 22-402-QOL), Canadian Institutes of Health Research, the International Development Research Centre, the Israel Science Foundation, the Azrieli Foundation (grant agreement 2545/18), the Deutsch-Israelische Projectkooperation program of the Deutsche Forschungsgemeinschaft (grant agreement B.I. 1665/1-1ZI1172/12-1), and the Sessile and Seymour Alpert Chair in Pain Research.

Conflicts of Interest

The authors declare no competing interests.

Data Availability

All datasets generated and/or analyzed during the current study are available in the main text or upon request. Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact ([email protected]).

References

Author notes