Preclinical studies of opioids and opioid antagonists on gastrointestinal function


Beverley Greenwood-Van Meerveld, PhD, Gastrointestinal Research, Oklahoma Center for Neuroscience, VA Medical Center, 151, 921 NE 13th St, Oklahoma City, OK 72104, USA.
Tel.: +405 270 0501 ext. 3547; Fax: +405 290 1719;


Opioid receptors in the gastrointestinal (GI) tract mediate the effects of endogenous opioid peptides and exogenously administered opioid analgesics, on a variety of physiological functions associated with motility, secretion and visceral pain. The studies reviewed or reported here describe a range of in vivo activities of opioid receptor antagonists upon GI function in rodents, focusing on µ receptors. Naloxone, and the peripherally acting µ-opioid receptor antagonists alvimopan and methylnaltrexone, reverse morphine-induced inhibition of GI transit in mice and rats, and morphine- or loperamide-induced inhibition of castor oil-induced diarrhoea in mice. At doses producing maximal reversal of morphine-induced effects upon GI transit, only the central nervous system (CNS) penetrant antagonist naloxone was able to reverse morphine-induced analgesia. Both central and peripheral opioid antagonists may affect GI function and/or visceromotor sensitivity in the absence of exogenous opioid analgesics, suggesting a constitutive role for endogenous opioid peptides in the control of GI physiology. Furthermore, in contrast to naloxone, alvimopan does not produce hypersensitivity to the visceromotor response induced by nociceptive levels of colorectal distension in a rodent model of postinflammatory colonic hypersensitivity, suggesting that in the periphery endogenous μ-opioid receptor-mediated mechanisms do not regulate colonic sensitivity. The data support the hypothesis that peripherally acting opioid antagonists may be able to selectively block opioid receptors in the GI tract, thereby preserving normal GI physiology, while not blocking the effects of endogenous opioid peptides or exogenous opioid analgesics in the CNS. These findings suggest that the primary sites of action of µ-opioid agonists with respect to inhibition of GI function are in the periphery, whereas analgesic activity resides primarily in the CNS.

Opioid receptors

Endogenous opioids and opioid analgesic drugs play a major role in the control of gastrointestinal (GI) function and dysfunction, including gut motility, epithelial transport and visceral sensitivity (for review, see 1,2). Endogenous opioid peptides (enkephalins, β-endorphin and dynorphins) are located in specific sites of the brain, spinal cord, autonomic ganglia and within the GI tract at the level of the enteric nervous system.

Three distinct opioid receptors (µ, κ and δ) are expressed throughout the central nervous system (CNS) and in peripheral regions, including the GI tract. The receptors consist of seven helical regions in the cell membrane and an intracellular region that couples G-proteins. They bind alkaloid opiates of plant origin (morphine), endogenous opioid peptides (enkephalins, endorphins, dynorphins, endomorphins, casomorphins, hermorphins) and amphibian skin opioids (dermorphins, deltorphins). The three opioid receptor types have been cloned in the mouse, rat and man. Peripheral opioid receptors (µ, κ and δ) are involved in enteroenteric and myenteric reflexes that provide functional regulation of the GI tract. The distribution and in vitro function of these receptors is discussed extensively elsewhere in this supplement (Sternini et al.3; Wood & Galligan4; and Sanger & Tuladhar5).

Recently, a 17-amino acid neuropeptide, nociceptin/orphanin FQ (N/OFQ) has been identified as the endogenous ligand for the opioid receptor-like 1 (ORL1) receptor (also referred to as the OP4 receptor − a fourth member of the classical µ, κ and δ opioid receptor family). Although ORL1 belongs to the opioid receptor family, it does not bind classical opioids, and the ORL1-N/OFQ system has pharmacological actions distinct from the opioid receptor system. Nociceptin, which is the endogenous peptide agonist for ORL1, has been found in the GI tract of the pig,6 rat7 and guinea pig.8 The ORL1 receptor is expressed in neurones, where it reduces activation of adenylyl cyclase and Ca2+ channels, while activating K+ channels in a manner similar to opioids.9 There is evidence that central ORL1 receptors are involved in a variety of functions within the GI tract (nociception, food intake and GI motility). However, the role in the GI tract of nociceptin receptors, like that of the classical opioid receptors, remains largely unclear.

Role of opioids in the Gastrointestinal tract

In vivo animal studies have revealed that endogenous opioids released peripherally can modulate GI motor and secretory functions, and the effects depend on the nature of the subclass of receptor involved. Agonist and antagonist drugs that possess selectivity for the individual subtypes of opioid receptors have been used in vivo and in vitro to determine the effects of opioids on GI functions. Exogenous opioid receptor ligands, with different affinities for the opioid receptor subtypes, have been used effectively to modify and normalize altered gut functions.

The µ-opioid receptor agonists morphine and, to a greater extent, diphenoxylate and loperamide, have been shown in vivo to slow GI transit by their effects on neurones innervating the circular and longitudinal muscle of the intestine. Trendelenburg demonstrated in 191710 that morphine inhibits peristalsis in an isolated preparation of guinea-pig small intestine. Since then, many in vivo animal studies have confirmed that morphine and related opioids delay GI transit. The effects of opioids in delaying intestinal transit is species-dependent and therefore, interspecies differences must be taken into account when evaluating the effects of opioid receptor agonists on GI motility. A delay in GI transit occurs through an inhibition of propulsive motility (rat) or a stimulation of non-propulsive or segmental patterns of motility in dog and man (for review see 11). The ‘Trendelenburg’ preparation was used further in the classical animal experiments demonstrating that morphine, fentanyl and met-enkephalin cause inhibition of both longitudinal and circular muscle.10

Later studies, using a vascularly perfused intestinal segment, demonstrated that morphine, dermorphin, D-Ala2-D-Met5-enkephalin, FK 33-824 and dynorphin reduced the frequency of peristaltic waves and the maximal ejection pressure (see paper by Sanger & Tuladhar.5). The in vivo effects of opioid agonists in delaying GI transit are due most probably to an inhibition of the release of acetylcholine, as well as to the release of non-adrenergic–non-cholinergic (NANC) neurotransmitters from enteric nerves.4 These effects are mediated, at least in part, via µ-opioid receptors present on circular muscle motor neurones.12 In the rat ileum, selective agonists of µ (PLO17) and κ (U-50,488H) receptors inhibit neurotransmitter release along the ascending excitatory reflex pathway.13

A review of the literature reveals that both µ and δ receptors are involved in the effects of opioids on motility. Delta receptors do not regulate the activity of myenteric excitatory motor neurones, as selective δ-receptor agonists (DPDPE) or antagonists (ICI174864) are ineffective.14 However, in the circular muscle of the guinea pig and human colon, NANC inhibitory responses are reduced by activation of δ receptors.15,16 Kappa opioid agonists have been found in vivo to slow GI transit in the guinea pig and mouse, but not in the rat. Nociceptin has the opposite effect to morphine on GI transit, and accelerates large intestinal transit in the rat through an opioid independent mechanism.17

Although morphine is a powerful antidiarrhoeal agent, the abuse potential of morphine limits its usefulness in the treatment of diarrhoea. In animal models of diarrhoea, peripherally acting µ-opioid receptor agonists such as diphenoxylate and loperamide inhibit diarrhoea, not only through a disruptive effect on intestinal propulsion, but also directly though an antisecretory effect at the level of the epithelial cell itself (for review see 18). Opioid pathways are also implicated in the adaptation response to stress-induced mucosal pathophysiology, including colonic ion secretion and permeability. Specifically, in the rat, acute stress induces adaptive responses via endogenous opioid pathways that promote restoration of normal mucosal physiology.19

Opioids have also been implicated in the effects of intestinal inflammation. In a mouse model of chronic intestinal inflammation, the peripheral effects of opioid receptor agonists and antagonists suggest that there is a greater effect of systemic opioids in slowing GI transit during chronic inflammation, which decreases intestinal secretion and inhibits intestinal permeability. Although the mechanism of this effect requires further investigation, recent studies have shown an increase in the expression of µ-, κ- and δ-opioid receptors during intestinal inflammation.20–23

Role of opioids in visceral pain

Opioid ligands have been employed to normalize altered visceral sensitivity and, as such, may have a potential clinical use in patients with abdominal pain associated with irritable bowel syndrome (IBS). In a rat model of visceral nociception, central administration of µ (morphine) or δ (DPDPE) receptor agonists inhibited visceromotor behavioural responses (VMR) to nociceptive levels of colorectal distension (CRD).24 Recent studies suggest that electrical stimulation of the rostroventral medulla (RVM) modulates nociceptive reflex responses.25 When the colon of the rat was made hypersensitive by inflammation, central administration of morphine into the RVM attenuated the response to CRD.26 Interestingly, this inhibitory effect of morphine was enhanced in the presence of a CCK-B, but not a CCK-A receptor antagonist, suggesting that visceral hyperalgesia can be modulated via morphine and CCK in the rat RVM.

Kappa receptor agonists

Kappa receptor agonists have been shown to increase pain threshold in animal models of visceral nociception.24,27–30 Moreover, κ-receptor agonists reverse peritoneal irritation-induced ileus and visceral pain in rats (for review see 31–33). Other reports suggest that peripheral, but not central, administration of the κ-receptor agonist U-50,488H inhibits the VMR response to noxious CRD24,28,29 and that inflammation of the colon does not change its potency.30

In support of a peripheral mechanism for κ-receptor agonists, recent studies found that κ-receptor agonists such as U-50,488H and fedotozine, but not µ- and δ-receptor agonists, appear to exert their antinociceptive effect at the level of nerve endings of mechanosensitive pelvic nerve afferent fibres.34 Additional studies by Gebhart and colleagues35–37 investigating pelvic nerve afferent fibre responses to luminal colonic distension, indicate that the principal effects of κ-receptor agonists are produced at or near the terminals of pelvic nerve afferent fibres in the GI tract. These sites of action need not be on the pelvic nerve terminals themselves, but could be associated with neurones of the enteric or intrinsic nervous system of the gut that interact with sensory afferent nerve terminals in the pelvic nerve. Based on these preclinical in vivo data, it was suggested that peripheral receptor agonists could be useful in the management of chronic visceral pain. In clinical trials, fedotozine has been shown to increase the threshold of perception to colonic distension and to reduce symptoms of IBS.38

A peripherally acting κ agonist, asimadoline (EMD-61,753), has a high affinity and selectivity for the κ-opioid receptor39 but has a reduced ability to cross the blood–brain barrier.40 Asimadoline has been shown to reduce sensation to gastric and colonic distension41,42 in the rat and in man.43 However, more recent data have suggested that the antinociceptive activity of κ-opioid receptor agonists such as U-50,488H and EMD-61,753 may be caused in part by the blockade of sodium currents, which is a non-opioid receptor activity of the compounds. Such an activity has been demonstrated in colonic sensory neurones suggesting that, in addition to their opioid receptor actions, these agents may also act in a manner that resembles local anaesthetics.36,44

Opioid receptor antagonists

Pharmacological significance of selective modulation of peripheral opioid receptors

At present, there are over 300 synthetic analogues of endogenous opioid peptides that are structurally designed to act as selective agonists or as antagonists. Novel opioid agonists and antagonists are useful tools to define the role of activation of opioid receptors in the gut, and they offer a potential new class of therapeutics for GI dysfunction. Morphine and other opioids are potent analgesics that work by stimulating µ-opioid receptors in the brain. However, µ-opioid receptors are also located in the periphery, including the wall of the GI tract, and stimulation of the GI receptors results in the common unwanted effects of opioid bowel dysfunction, including severe constipation, hard stools, straining, incomplete evacuation, bloating, abdominal distension and increased gastroesophageal reflux.1

Recent research in the field of gastroenterology has concentrated on the design of opioid molecules that do not pass the blood–brain barrier and thus have selectivity for the peripheral opioid receptors. Opioid receptor antagonists having limited brain penetration after systemic or central administration may be useful in identifying the sites of action within the gut where opioids mediate GI function. They may also be useful therapeutically in reversing unwanted side effects of opiate analgesics such as constipation, while preserving centrally mediated analgesia.

Preclinical and clinical studies have been conducted with two peripherally acting opioid antagonists, methylnaltrexone45–54 and alvimopan,55–63 to test the hypothesis that antagonism of the effects of opioids in the gut can be accomplished without compromising analgesia. Both of these compounds are potent and reversible inhibitors of binding to cloned human µ-opioid receptors in vitro, with alvimopan demonstrating greater affinity and selectivity than methylnaltrexone or the centrally acting compound naloxone.58,59 Similar results were observed when comparing the binding and functional activity of alvimopan to that of naloxone at µ- and δ-opioid receptors in rodents, and to κ-opioid receptors in the guinea pig.55,56

In the in vivo studies, all three agents antagonized morphine-induced inhibition of GI transit in mice57,60,61 as well as morphine-induced inhibition of castor oil-induced diarrhoea in mice.61,62 In general, alvimopan was more potent in antagonizing the effects of morphine than methylnaltrexone or naloxone. In studies examining the effects of the antagonists on colonic transit, alvimopan antagonized the slowing of colorectal transit produced by morphine with an ED50 of 0.41 mg kg−1 p.o. when administered at 20 min prior to bead insertion.62 Oral doses of 1.0 and 3.0 mg kg−1 also antagonized the slowing of bead expulsion produced by morphine when alvimopan was administered at 30 min prior to bead insertion. Moreover, the 3.0 mg kg−1 dose antagonized morphine's effects with a 6 h pretreatment (not shown).

An example of the effect of alvimopan upon morphine-induced slowing of GI transit is shown in Fig. 1. In the rat, alvimopan at oral doses of 0.3 and 1.0 mg kg−1 partially antagonized the slowing of small intestinal transit of 113Sn-labelled microspheres produced by morphine (Fig. 1A; alvimopan alone had no effect at oral doses up to 3.0 mg kg−1, data not shown). Furthermore, although in vitro studies of transit through the guinea-pig colon have shown the inhibitory effects of the α2 adrenoceptor agonist clonidine to be naloxone sensitive,5 doses of alvimopan up to 3.0 mg kg−1 p.o. in rats failed to affect p-aminoclonidine-induced (10 µg kg−1 i.p.) slowing of transit (Fig. 1B). Further work is required to establish whether this difference is due to species, CNS vs. peripheral actions of the agents tested or in vivo vs. in vitro factors.

Figure 1.

Dose-dependent antagonism of morphine-induced inhibition of small intestinal transit by alvimopan in rats. Transit of 113Sn-labelled microspheres was measured 1 h following their intraduodenal administration via an in-dwelling cannula. Animals were administered either morphine [MOR; (A)] or para-aminoclonidine [PAC; (B)] 15 min prior to microsphere administration. The effect of alvimopan (ALV) upon the resultant slowing of transit was assessed by administering a range of doses as indicated, 45–60 min prior to the administration of the microspheres. n = 6–9/group. Significantly different from vehicle + MOR, Dunnett's test, *P < 0.05, **P < 0.01.

The potential for CNS activity of the peripherally acting agents has also been assessed. In morphine-dependent mice56,57,61,63 and rats57 alvimopan selectively precipitated the peripheral withdrawal sign of diarrhoea, whereas naloxone administration resulted in both diarrhoea and centrally mediated jumping behaviour. These data suggest that peripherally acting opioid antagonists should be able to selectively antagonize the GI effects of opiate analgesics, without compromising their analgesic activity in the CNS. Convincing evidence for this has been provided from studies in mice where alvimopan antagonized morphine-induced analgesia only at doses far in excess of the doses that antagonized the effects of opioid agonists on measures of GI function.57,62,63

Lack of effect of alvimopan upon visceral sensitivity

Opioids mediate their analgesic effects by activating µ-opioid receptors in the CNS, and the studies discussed above show no evidence for liability of peripherally acting µ-opioid receptor antagonists to affect CNS mediated analgesia. However, it has been suggested64 that extrinsic afferent nerves innervating the GI tract may also play a role in the μ-opioid receptor-medicated analgesia. Therefore, in order to assess the potential for a role of endogenous opioids in setting GI sensory tone, we examined whether peripheral opioid antagonism with alvimopan has any affect on colonic nociception in a rodent model. Moreover, because the peripheral analgesic effects of opioids on sensory neurones are best observed following inflammation,63 the effect of alvimopan was examined in a rodent model of postinflammatory colonic hypersensitivity. In both models, alvimopan alone had no effect on the VMR induced by noxious CRD (Fig. 2), suggesting that in the periphery endogenous µ-opioid receptor-mediated mechanisms do not regulate colonic sensitivity. However, animals that received naloxone displayed an increased sensitivity to CRD, an effect that was evident across the whole range of distending pressures. Indeed, in the postinflammatory group, evidence for increased sensitivity in the absence of distention was also observed (Fig. 2B). Therefore, there does appear to be an ongoing opioid-driven tone in the GI sensory system, but the data presented here suggest that this is not mediated in the periphery.

Figure 2.

Effects of µ-opioid receptor antagonists upon visceromotor responses in conscious rats. VMR to CRD in rats with normosensitive colons (A) and in rats with postinflammatory colonic hypersensitivity (B). Postinflammatory colonic hypersensitivity was induced by intracolonic administration of trinitro-benzenesulphonic acid (w/v) (TNBS, 50 mg kg−1, 0.5 ml, 25% EtOH), 30 days prior to testing. Animals received either a 60 min predose of alvimopan (3 mg kg−1, p.o.), or vehicle (3 ml kg−1 20% Cremophor EL, p.o.) or a 30 min predose of naloxone (20 mg kg−1 i.p.) or vehicle (3 ml kg−1 20% Cremophor EL, i.p). Data are expressed as mean ± SEM from six experiments for each group. (A) In animals with normosensitive colons, naloxone caused a statistically significant increase in the VMR (***P < 0.001 vs. vehicle control), while alvimopan had no significant effect. (B) VMRs to CRD were greater in rats with postinflammatory colonic hypersensitivity. Treatment with naloxone induced a further increase (***P < 0.001 vs. vehicle control) in the VMR, but alvimopan had no significant effect.

Conclusions and future directions

Although the actions of opioids in the gut have been well documented, major gaps remain in our understanding of the precise mechanisms underlying these effects and of the potential role of opioid systems in GI disorders. Opioid analgesic use is associated commonly with GI side effects and, interestingly, the expression of opioid receptors in the GI tract is increased under conditions of inflammation.20,22,23,65 These observations suggest a role for opioid systems in both normal GI function and pathophysiology. Selective opioid receptor agonists and antagonists are available for use both in in vitro and in vivo studies to identify the receptor types involved in functional responses. Coupled with findings from molecular and immunohistochemical localization studies, these agonists and antagonists are powerful tools for identifying the precise sites of action within the gut where opioids affect GI function. Compounds with activity limited to the periphery may help to provide an improved understanding of the relative contribution of central and peripheral sites of action in specific GI functions. Indeed, this property may also be of benefit clinically, where peripherally acting opioid antagonists may reverse unwanted peripheral side effects associated with opioid use while preserving centrally mediated analgesia.

A comparison of alvimopan and methylnaltrexone demonstrates that both compounds are competitive opioid receptor antagonists and act in the periphery. However, alvimopan demonstrates greater potency and affinity for the µ-opioid receptor in vitro, as well as greater potency and a longer duration of action in vivo for antagonizing morphine's effects in the gut.57–63 The greater potency of alvimopan in vivo compared with methylnaltrexone may be related to differences in oral absorption as well as differences in potency between the two compounds at opioid receptor level. Alvimopan antagonized morphine-induced analgesia in mice and rats only at doses far in excess of the doses that antagonized the effects of opioid agonists on measures of GI function.57,62,63 Collectively, the preclinical data on alvimopan are consistent with the clinical reports on the ability of alvimopan to antagonize agonist-induced slowing of GI transit66,67 in the absence of effects on analgesia or other centrally mediated responses.55,67,68

There are alternative and complementary avenues of research that may yield insights into mechanism of action of µ-opioid receptor antagonists in GI function. For example, the greater potency of alvimopan to antagonize agonist-induced inhibition of diarrhoea, compared with its effects on transit,61 suggests an antisecretory component to its actions in the gut. The absence of significant alterations in GI transit by alvimopan in the absence of agonist at single oral doses up to 200 mg kg−1, suggest a lack of effect on the regulation of GI transit by endogenous opioids in the normal physiological state. However, there is much still to learn about their precise role.

Further investigation of the dose–response relationships and the temporal aspects of antagonism produced by peripherally acting antagonists under conditions of inflammation is warranted. At present it is unclear whether these compounds might exhibit a greater inhibition of endogenous opioids under conditions where they are upregulated, such as stress, trauma or inflammation. It is hoped that research in this area will increase so that these important questions are answered in the near future.