Chronic morphine administration enhances nociceptive sensitivity and local cytokine production after incision

While opioids are a mainstay of treatment for chronic pain and are second only to NSAIDS terms of the number of prescriptions written for common chronic pain syndromes [1], problems with long term use are becoming better recognized. Tolerance and dependence on these drugs have been studied for some time. More recently, the phenomenon of opioid-induced hyperalgesia (OIH) has come under study (see ([2‐6]) for recent reviews). This form of hyperalgesia has been noted in human populations maintained long term with methadone [7‐11], treated with morphine for back pain [12], and infused with the potent short-acting opioid remifentanil in experimental pain laboratories [13‐18]. Based largely on studies in laboratory animals, several mechanisms have been suggested including sensitization of peripheral nociceptors [19‐21], enhanced production and central release of excitatory amino acid (EAA) and peptide neurotransmitters [22, 23], diminished uptake of EAA neurotransmitters in the spinal cord [24], and enhanced descending facilitation of spinal nociceptive neurotransmission from the rostral ventromedial medulla (RVM) [25]. Many of these responses represent mechanistic counterbalances to the acute effects of opioid receptor activation.

Clinically, post-incisional pain has been observed to be more severe in chronically opioid consuming patients [26]. These patients consume several times the normal doses of opioids in the perioperative period and do not always achieve the same level of comfort as previously opioid naïve patients [26, 27]. The situation is sufficiently widely recognized that guidelines for the management of these patients have been provided [28]. Several groups have studied the phenomenon of opioid-exaggerated post-incisional pain in laboratory models. Our group has studied the problem in a rat model of incisional pain in which the animals were treated chronically with morphine prior to the incisional being made [29]. In this model it was observed that opioid treated animals displayed mechanical allodynia in excess of saline treated ones with the difference not appearing until 2 days after incision, reaching maximum at 3 days and persisting for at least 7 days. Celerier et al. observed that the administration of high dose opioids at the time of incision generated hind paw peri-incisional sensitization in excess of what was observed when using inhalational anesthesia alone [30]. Again, the opioid-induced excess hyperalgesia was prominent for up to 7 days following incision. The mechanisms responsible for the opioid-induced excess in nociceptive sensitization have not been fully explored.

Recently we observed that the acute systemic administration of opioids like morphine can substantially reduce the abundance of specific cytokines accumulating in the peri-incisional tissue [31]. Several of these cytokines including interleukin-1β (IL-1β), IL-6, tumor necrosis factor-α (TNFα) have been observed to support enhanced nociceptive sensitivity in various rodent models [32‐34]. Both keratinocytes and infiltrating neutrophils were observed to be producing these cytokines near the wound edges [31]. Because it appears that the chronic effect of opioids is sometimes to cause effects paradoxically opposite to the acute effects of these drugs, we hypothesized that mice chronically treated with opioids would display exaggerated levels of skin cytokine production after hind paw incisions. This would explain some portion of the exaggerated nociceptive sensitization seen in these animals. Such observations would be novel and might suggest approaches to better controlling pain in vulnerable clinical populations.

Opioid-induced hyperalgesia (OIH) is a state of paradoxically enhanced pain sensitivity observed in both humans and animals after chronic exposure to opioids including morphine, oxycodone, fentanyl, heroin and others. To-date most explorations of this phenomenon's mechanism have focused on alterations in functional elements within the central nervous system and to a lesser degree neuroplastic changes involving primary afferent sensory neurons [2]. Our studies were focused further in the periphery and involved opioid effects on skin cytokine levels both before and after hindpaw incision. Peripheral cytokine production had not prior to this time been implicated in the mechanism of OIH. Our results using the C57BL/6 strain of mice reproduced those we had previously reported for rats in that morphine treatment both lead to enhanced nociceptive sensitivity in unperturbed animals, and significantly enhanced the sensitization of hindpaws for several days after incision [29]. More importantly, we went on in the present studies to show that the peri-incisional expression of cytokines previously demonstrated to be inhibited by morphine given acutely before incision [31] were moderately to strongly enhanced in the tissues of mice chronically pretreated with morphine. These cytokines included IL-1β, IL-6, TNFα, G-CSF and KC. Of these, only IL-6 was not elevated in the previously morphine treated mice compared to saline treated mice 72 hours after incision, the time point at which the difference in mechanical sensitization was maximal. The broad spectrum cytokine inhibitor pentoxifylline reduced both the excess cytokines and excess sensitization measured in the morphine treated animals. The specific site of action of pentoxifylline was not determined, though pentoxifylline has been noted to reduce cytokine production by keratinocytes in culture [39] as well as after ultraviolet light burns and psoriatic skin [40, 41]. Finally, we demonstrated that the initiation of morphine treatment at the time of incision did sensitize the hind paws of mice. However, there did not appear to be any interaction of morphine administration with nociceptive sensitization in this context. Furthermore, morphine administered under these circumstances did not lead to any significant differences in wound area cytokine levels. Taken as a whole, our data suggest that chronic morphine treatment prior to incision leads to excess production of cytokines thus enhancing nociceptive sensitization after incision. Administration of morphine beginning at the time of incision does sensitize mice, but this is not in excess of what is seen with morphine treatment of incision naïve mice, and there is little enhancement of cytokine production. Thus cytokine mediation of exaggerated post-incisional nociception is more relevant to the treatment of chronically opioid consuming patients than otherwise opioid naïve patients being treated for postoperative pain.

A range of mechanisms causing OIH have been proposed and recently reviewed [2]. The principal mechanisms currently considered responsible for OIH include those leading to enhanced function or activity of afferent fibers, second order or projection neurons and descending facilitory fibers from the brainstem. For example, chronic morphine treatment leads to the up-regulation of production and release of neurotransmitters by primary afferent nerve fibers [42, 43]. Once released, glutamate reuptake systems in the spinal cord function less efficiently by virtue of lower levels of expression of neuronal and glial glutamate reuptake transporters [24]. Also, it appears that for a given level of primary neurotransmitter like glutamate or substance P that more nociception related behavior is generated in mice having been treated for several days with morphine. Finally, descending facilitation from the RVM is enhanced in chronically morphine treated animals constituting an independent pathway for facilitation of nociceptive signal transmission [25]. In our own previous studies we had implicated beta-2 adrenergic receptors located on peripheral nerve terminals as being linked to OIH [21]. It should be recognized that one key feature separating our observations of cytokine participation in post-incisional OIH versus most other observations is that this peripheral cytokine mechanism only seems to be operative when tissue has been damaged. We failed to find morphine-induced differences in skin cytokine levels prior to incision and thus cannot invoke these mediators as supporting hyperalgesia observed in uninjured mice. This pattern of expression distinguishes our observations from those of others who found upregulation of IL-1β, IL-6 and TNFα in spinal cord microglia after morphine treatment even without additional trauma [44, 45].

Our previous studies used immunohistochemical techniques to demonstrate that both infiltrating neutrophils and peri-incisional keratinocytes produce the cytokines we measured as elevated after incision [31]. However, as was the case in our previous experiments using acute morphine administration, the degree of peri-incisional neutrophil infiltration did not correlate with peri-incisional cytokine levels. Importantly, chronic morphine treatment was not seen to alter wound area neutrophil infiltration. Thus a simple lessening of neutrophil migration does not seem to be the explanation for the changes in inflammatory mediators observed in our experiments. The function of the infiltrating cells was not, however, directly assessed. Systemic opioid administration has been noted to alter immune system functioning in a number of ways [46].

Future studies addressing this issue will need to address at least two possible mechanisms. The first mechanism involves the participation of peripheral afferent neurons in controlling keratinocyte cytokine production. The second focuses on opioid actions on the keratinocytes themselves. Addressing neural mechanism first, chronic opioid exposure up-regulates the production of SP and CGRP at the level of both mRNA and protein [22, 42, 43]. Though commonly thought of participating only in augmenting nociceptive signal transmission, these neuropeptides act on keratinocytes to enhance the production of IL-1β, IL-6, TNFα and other cytokines [47]. Thus the enhanced availability and release of these neuropeptides after incision might explain the augmented levels of cytokines in the incisional wounds of morphine treated mice. Second, opioid receptors are found on keratinocytes, and keratinocytes produce opioid peptides [48, 49]. These receptors have clear roles in modulating keratinocyte differentiation, wound healing, and inflammatory responses [50]. While little investigation has focused directly on opioid modulation of keratinocyte cytokine production, it could be hypothesized that chronic exposure of keratinocytes to opioids could alter their cytokine producing capacity thus leading to enhanced cytokine levels after incision. This possibility could be addressed directly.

Our studies suggest that some of the cytokines analyzed may be of special functional interest. The cytokines most likely to be involved in supporting OIH in our incisional model include IL-1β, G-CSF and KC. For all three of these cytokines peri-incisional skin levels were augmented in morphine treated mice after incision to levels above those observed in the opioid naïve animals. Furthermore, the cytokine production inhibitor pentoxifylline reduced the skin concentration of all three of these cytokines by >50% in the morphine treated animals three days after incision at the same time as pentoxifylline reversed the excess sensitization of the hind paws. A variable degree of information is available from the literature to assist in interpreting these observations. With respect to IL-1β, detailed work has been provided showing that sub-picogram quantities of subcutaneous IL-1β cause mechanical allodynia. This allodynia is probably mediated by activation of central metabotropic glutamate receptors [51, 52]. Thus, excess levels of this cytokine might reasonably be expected to enhance nociception. For KC, there appears to be little information concerning the direct effects of KC on nociceptive thresholds. On the other hand, investigators have linked this cytokine to the recruitment of opioid peptide expressing neutrophils in inflamed skin. Blockade of KC activity reduced this peripherally mediated anti-nociception [53]. However, given the lack of effect of chronic morphine treatment on neutrophil recruitment in the incisional pain model, it is unclear if the previously reported KC mediated chemoattractant effects are important to our observations. Lastly, there is little work describing possible effects of G-CSF on nociceptive thresholds in control animals or in models of tissue injury. Of note, however, is that the subcutaneous administration of G-CSF in humans can cause a psoriasis-like inflammatory dermatitis. Analysis of G-CSF induced lesions revealed evidence of inflammation and elevated levels of cytokines including IL-8, IL-12 and TNFα [54]. Though we followed levels of several mediators, it is possible that additional cytokine mediators participate in the chronic morphine administration post-incision exacerbation of sensitization as only a small subset of all known cytokines were measured in our studies.

Excessive and difficult to treat pain after surgery has been noted for patients chronically consuming opioids. For example, de-Leon Casasola demonstrated in a series of clinical investigations that chronic opioid consuming patients require approximately three times greater doses of epidural opioids to control pain after surgical procedures, and that these patients tend to have more serious pain for a prolonged period after their operations [27, 55, 56]. Looking at required doses of parenteral morphine, Rapp et al. came to many of the same conclusions, namely that about three times as much opioid was required to control pain in chronic opioid consuming patients [26]. Importantly, it was also observed that opioid side effects were greater and the degree of pain relief was poorer despite the enhanced doses. More recently it was shown that chronic back pain sufferers develop measurable hyperalgesia after only one month of opioid administration [12]. Reviews are available further characterizing the problem and making suggestions for postoperative management of pain in chronically opioid consuming patients [28, 57]. However, very little work has been done specifically directed at evaluating analgesic techniques of potential benefit to the chronic opioid consuming patient.

Along with exacerbated post-incisional nociception comes excess cytokine production in animals administered morphine on a chronic basis. The excess incisional area cytokine production may explain the enhanced nociception in laboratory animals and by extension may explain some amount of the excess surgical pain suffered by humans chronically consuming opioids. This work provides a rational basis on which to design therapies aimed at reducing OIH as it impacts postoperative pain. Our work suggests that strategies designed to reduce the production of cytokines globally or perhaps reduce the production of several specific cytokines may be of benefit to opioid consuming patients. Our results to-date do not shed light on the required opioid doses or treatment durations which might be most closely associated with excess post-incisional cytokine production. In addition, different opioids such as morphine and buprenorphine have differing abilities to alter other aspects of immune function. Thus different opioids might have differing intrinsic abilities to control wound area inflammatory mediators [46]. As enthusiasm for the use of opioids for the treatment of chronic pain grows, the number of patients using opioids prior to surgical procedures is likely to rise. Therefore, the need for better postoperative pain management techniques applicable to this population is of rapidly increasing importance.