E nhanced thermal antinociceptive potency and anti-allodynic effects of morphine following spinal administration of endotoxin

Recently, an animal model of central inﬂammation characterized by widespread cutaneous hyperalgesia and allodynia following intracerebroventricular (i.c.v.) administration of lipopolysaccharide (LPS) was described. In the present study, we demonstrate that central administration of LPS via intrathecal (i.t.) injection produces bilateral tactile allodynia and thermal hyperalgesia in the rat. Also, the effects of morphine-induced antinociception were determined in this model. Here we demonstrate enhanced thermal antinociceptive potency of i.t. morphine in LPS-treated rats compared to controls. Intrathecal morphine was also effective in alleviating the tactile allodynia induced by LPS. Both the antinociceptive and anti-allodynic effects produced by i.t. morphine were completely antagonized by 2 1 pretreatment with subcutaneous naloxone (1 mg kg ). This study demonstrates the presence of both heat hyperalgesia and mechanical allodynia following central administration of LPS, and an increased antinociceptive potency of i.t. morphine in this model. 


. Introduction
the induction of peripheral inflammation, but few studies have examined pain mechanisms following the induction Prolonged tissue damage or nerve injury often leads to of central inflammation. chronic pain characterized by the genesis of hyperalgesia It is now well established that peripheral administration and / or allodynia, which persists after healing. Many of lipopolysaccharide (LPS; endotoxin) interacts with animal models have been developed in an attempt to various cell types to release cytokines IL-1, IL-6, and understand the etiologies of various pain syndromes fol-TNFa in the periphery and central nervous system, which lowing an inflammatory insult. These models include can in turn precipitate hyperalgesia [51][52][53][54]. Central application of various chemical irritants to induce peripher-administration of endotoxin can also initiate the production al inflammation (complete Freund's adjuvant, carrageenan, of cytokines [17,20,43] known to modulate nociception. mustard oil and formalin), all of which elicit spontaneous For example, intracerebroventricular (i.c.v.) administration pain and / or thermal hyperalgesia and tactile allodynia of IL-1b, IL-6 or TNFa elicited thermal hyperalgesia and [15]. The models described above have broadened our enhanced the response to wide dynamic range neurons in knowledge of the mechanisms underlying pain following the trigeminal nucleus caudalis to noxious stimulation in rats [22,34,35]. We and others have also reported cutaneous thermal hyperalgesia and tactile allodynia following i.c.v. administration of LPS [4,50]. In addition, spinal adminis- TNFa in the cerebrospinal fluid [48] as well as precipi-tating thermal hyperalgesia [30] and enhancing activity of 2 .2.2. Mechanical allodynia testing dorsal horn neurons [41].
Mechanical response thresholds were quantified by It has been well established that the antinociceptive determining the hind paw withdrawal response to von Frey potency of systemically administered opioids [26,27,32] filament stimulation according to the method described by and spinal opioids [24,46] is enhanced in peripheral Chaplan et al. [9]. In brief, animals were placed in a  inflammatory states. In the spinal cord, the complexity of Plexiglas box (21316327 cm) with a wire grid bottom changes induced by inflammation on opioid receptor through which the von Frey filaments were applied to the number, changes in neurochemical release and neuronal plantar surface of both hind paws. Filaments were applied firing patterns makes it difficult to ascertain the exact in either ascending or descending strength as necessary to mechanism of enhanced opioid sensitivity. It has been determine the filament closest to the threshold of response. demonstrated that all three opioid receptors are expressed The minimum stimulus intensity was 0.25 g and the in dorsal root ganglion neurons and are differentially maximum was 15 g. Based on the response pattern and the regulated by inflammation [1,2,12]. Moreover, in situ force of the final filament, the 50% response threshold (g) hybridization has shown that inflammation increases the was calculated. Animals were habituated to the testing percentage of dorsal root ganglion cells expressing mu-apparatus for at least 15 min on the day prior to testing and opioid receptors [25].
for 10 min prior to testing. In the current study, we characterize the nociceptive behaviors following spinal administration of LPS, by 2 .3. Drug administration schedule examining the occurrence of heat hyperalgesia and tactile allodynia using the hot plate test and von Frey hair Spinal administration of either vehicle or LPS (E. coli mechanical stimulation. We have also examined the effect lipopolysaccharide (Sigma stereotype 0111:B4) in 0.9% of priming with LPS to enhance the behavioral nociceptive sterile saline) was accomplished by injecting a 30-ml response to intrathecal (i.t.) administration of endotoxin, as volume into the subarachnoid space via a lumbar puncture priming was shown in an earlier study to augment the (between vertebrae L4 and L5), while rats were briefly nociceptive response [4]. Finally, we evaluate the efficacy anesthetized with halothane. Behavioral testing was perand potency of morphine in this model by comparing formed in two treatment groups: primed and unprimed. nociceptive thresholds in LPS-treated and saline-treated Primed rats received an i. Morphine antinociception was evaluated in both primed (2.0 mg followed by 20 mg) and unprimed (saline followed 2 .2. Behavioral testing by 20 mg) LPS-treated rats. The degree of anti-allodynic and thermal antinociception produced by morphine was 2 .2.1. Thermal hyperalgesia testing determined by comparing the response thresholds prior to Thermal nociceptive thresholds were determined using a (i.e. that produced by i.t. LPS (40-min onset of hyconstant temperature hot plate (48 8C). The response peralgesic effect)) and following morphine administration. latency to a hind paw lick or vigorous shaking was Behavioral response thresholds were assessed 30 min recorded (baseline 25-35 s; cutoff 50 s). Animals were following morphine administration (70 min after the initial habituated to the testing apparatus for at least 15 min on LPS injection), a time point that was shown to produce the day prior to testing. maximum antinociceptive efficacy in the hot plate test [5], and the percent maximum possible effect (% MPE) was followed by Wilcoxon signed rank test for post hoc calculated, as indicated below. An experiment was also comparisons. EC values produced by i.t. morphine were 50 designed to determine whether the anti-allodynic and obtained by regression analysis and interpolation calcuantinociceptive effects produced by morphine could be lated using Graphpad software (Prism 3.0). blocked by pretreatment with the non-selective opioid 21 receptor antagonist naloxone (1 mg kg , s.c.).

.4. Statistical analysis
3 .1. Thermal hyperalgesia and mechanical allodynia Data are presented as raw scores or converted to % MPE following intrathecal administration of LPS where % MPE5[(response value2baseline) /(cutoff2 baseline)]3100% for thermal thresholds. For mechanical The nociceptive effects produced by i.t. administration thresholds data are converted to % return to baseline of LPS were evaluated using both thermal (hot plate test) defined as [(baseline2threshold) / baseline]3100% because and mechanical (von Frey hair stimulation) behavioral the % MPE equation can not be used due to instances testing parameters (Fig. 1). Two groups were examined: when the baseline and the cutoff are the same value due to unprimed control rats received i.t. saline (Fig. 1A,B), and the nature of the stimulus (non-noxious). All data are The potential antinociceptive effects of morphine were detected between baseline responses when compared beexamined in LPS-treated (primed) compared to control tween each group (the mean and S.E.M. for the combined rats. Fig. 3 demonstrates the dose-response curves obgroups was 2962.7 s for thermal latencies to response and tained for morphine-induced antinociception on the hot 9.861.4 g for the mechanical response threshold). For plate test in both groups following various routes of LPS-induced decreases in mechanical thresholds, challenge morphine administration. Subcutaneous administration of doses of endotoxin ranging from 0.2 to 20 mg produced morphine produced similar antinociceptive dose-response dose-dependent decreases in mechanical response threscurves in both LPS-treated and saline groups, where ED holds compared to their baseline response. The Friedman's 50 21 non-parametric test demonstrated a statistical significance values for the control group were 1.91 (1.44-6.7) mg kg LPS-treated rats (Fig. 3C). Regression analysis confirmed that no significant difference in morphine-induced antinociception was evident between LPS-treated and control rats for either peripheral or i.c.v. administration of the opioid. However, the dose-response curve for the antinociceptive effects produced by i.t. administration of morphine were shifted significantly to the left in LPStreated rats compared to saline control rats (Fig. 3B). ED 50 values calculated for i.t. morphine-induced antinociception were 0.054 (0.032-0.071) mg in LPS versus 2.01 (1.83-2.79) mg in control rats, representing a 40-fold shift in potency.
Morphine administered by i.t. (Fig. 4A), s.c. (Fig. 4B) or i.c.v. (Fig. 4C) routes in saline-pretreated / LPS-challenged rats produced no significant changes in paw withdrawal thresholds in the von Frey test, when compared to baseline values obtained prior to LPS administration (Fig.  4, right column). In contrast, all three routes of morphine administration ( Fig. 4A-C) increased mechanical response thresholds in i.t. LPS-treated (pretreated and challenge injections consisted of LPS) rats compared to mechanical thresholds following LPS (`P,0.05). The effects of i.c.v. morphine were modest (Fig. 4C, left column), as the mechanical threshold following morphine administration remained significantly attenuated compared to baseline values (*P,0.05). The anti-allodynic effects produced by higher doses of i.c.v. morphine could not determined due to the appearance of motor side-effects.
The anti-hyperalgesic and anti-allodynic effects of i.t. morphine are also illustrated in Fig. 5. Baseline thermal (bottom panel) and mechanical thresholds (top panel) were recorded for all three groups prior to LPS injection (time: 0 min). Each group was assessed for hyperalgesia and allodynia after LPS treatment but prior to drug treatment (time: 40 min). There was no difference between baseline values for all groups, and LPS treatment (first arrow) produced an equivalent reduction in mechanical thresholds and hot plate latency for all groups. Mechanical (Fig. 5B) or thermal (Fig. 5E) thresholds were shown to be significantly elevated 30 min following i.t. morphine administration (3 mg) compared to response values following LPS administration (*P,0.05), an effect that was effectively antagonized by pretreatment with s.c. naloxone (1 mg 21 kg ; Fig. 5C,F).

. Discussion
In this study, we demonstrate the incidence of thermal hyperalgesia and tactile allodynia following spinal ad-  comparing EC s obtained using a thermal threshold test. rats at doses that had no effect on baseline mechanical 50 Finally, morphine administered either spinally or sys-thresholds (and little or no effect on thermal thresholds) in temically produced anti-allodynic effects in LPS-treated control animals. The central mechanisms underlying hyperalgesia elicited responses following a central inflammatory response, by a peripheral inflammatory injury have been extensively however, nitric oxide [30] and bradykinin [50] have been studied, confirming the relevance of various neurochemical implicated. In addition, our own studies have demonstrated systems and the necessity of neuronal plasticity in the that i.c.v. LPS-induced thermal hyperalgesia and mechanidevelopment of hyperalgesia and chronic pain [10,14,15]. cal allodynia were correlated with an increased activation It is now well established from various types of central of microglia [4]. nervous system injuries (traumas, infections, ischemia, The present study demonstrates enhanced antinocicepetc.), that the pro-inflammatory cytokines are rapidly tive potency and anti-allodynic effects of morphine in a produced in the brain [16]. This effect can also be model where the inflammatory response is initiated within produced by direct administration of endotoxin into the the central nervous system. It has been well established central nervous system. Thus, i.c.v. administration of that the antinociceptive potency of systemically adminisendotoxin has been shown to increase circulating levels of tered opioids [26,27,32] and spinal opioids [24,46] is IL-1 and IL-6 [19]. Moreover, there is intense monocyte enhanced in inflammatory states generated following a recruitment and microglial activation after LPS injection peripheral insult. Here we report for the first time the [20,31,40]. enhanced antinociceptive effectiveness of morphine in a In this study, we demonstrate that hyperalgesia and central model of inflammation. It is also interesting to note allodynia can also occur following spinal administration of that morphine, which normally has no effect on altering an endotoxin. Other studies have reported the incidence of thresholds to innocuous stimuli, produces an anti-allodynic hyperalgesia following central administration of endotox-effect following either peripheral or spinal administration. ins. For example, i.t. administration of LPS was shown to The effects of supraspinal administration of morphine induce thermal hyperalgesia [30] and increase the activity appeared to be only partially effective in producing antiof dorsal horn neurons [41], and i.c.v. administration of allodynic effects, however, this may be due to the con-LPS elicits both thermal hyperalgesia and a decrease in founding incidence of motor impairment associated with mechanical thresholds [4,50]. Little is known regarding the this route of administration. This is in agreement with sensory pathways within the central nervous system or the previous studies demonstrating that i.t., s.c. or local chemical mediators that are responsible for the nociceptive administration of morphine into the paw does not sig-nificantly affect baseline mechanical nociceptive thresholds enkephalins, in the spinal dorsal horn, that could then act [21,28], whereas others have demonstrated the morphine-synergistically with exogenously administered morphine to induced anti-allodynia in inflammation pain models. Thus, produce the enhanced effectiveness seen in this study. It local administration of mu opioid agonists in the rat paw has been well established that antinociceptive synergy inhibited prostaglandin E -induced hyperalgesia [29], and occurs following combined mu and delta opioid receptor 2 s.c. morphine elicited dose-dependent naloxone-reversible activation [45]. antinociception in the paw of monoarthritic rats [21].
The increase in opioid potency during inflammation may Furthermore, morphine produced mechanical anti-al-in part be attributed to an increase in the number of spinal lodynic effects in a model of neuropathic pain when opioid receptors. All three opioid receptors are differentialtolerance to morphine was blocked, or when spinal afferent ly regulated following peripheral inflammation [1,2,12]. drive was reduced with an N-methyl-D-aspartate receptor Importantly, in situ hybridization has shown that inflamantagonist [33,36,37]. mation increases the percentage of mu-receptor expressing It is intriguing that the thermal antinociceptive potency neurons in dorsal root ganglia [25]. of i.t. morphine, but not s.c. or i.c.v. morphine, is enhanced Recently, it has been demonstrated that peripheral in this model of LPS-induced inflammation. Consequently, inflammation [18], like i.t. administration of LPS [4], LPS-induced hyperalgesia and allodynia may be partially produces an activation of spinal microglia. Thus, periphermediated by changes at the level of the spinal cord. The al and central inflammation may produce a similar increase basis for this increase in morphine potency remains to be in cytokines and resultant changes in opioid peptide or determined, and will be investigated in future studies. One receptors, as discussed above, that may contribute to could speculate that the increase in the effectiveness of enhanced morphine efficacy. However, these findings are morphine might be due to changes in endogenous opioids.
not consistent with the observation that morphine analgesia Numerous reports have indicated the presence of opioid is attenuated in neuropathic rats, despite similar observapeptides in cells of the immune system [44] and opioid tions of increased microglial activation in the spinal cords peptides are released from immunocompetent cells infil-of neuropathic rats [11]. Recent finding in our lab and trating inflamed tissue within the periphery [47]. Pre-others suggest this discrepancy may be explained by proenkephalin mRNA in the spinal cord has been shown to differences in the influence of peripheral inflammation and increase following intraplantar formalin [38] and car-nerve injury on spinal levels of nerve growth factor rageenan injections [13]. Intraplantar formalin was also (NGF), and the ability of NGF to regulate anti-opioid reported to induce preprodynorphin mRNA in the lumbar peptide levels in the spinal cord dorsal horn. We hypothspinal cord ipsilateral to the side of injection [23]. In rat esize that despite similarities in microglial activation, models of chronic arthritis there is an increase synthesis of inflammation and nerve injury produce opposite effects on opioid peptides in the spinal cord [8]. In addition, pre-NGF levels in spinal cord, with a resultant differential proenkephalin mRNA was abundant in cells of hind paws effect on antiopioid peptides such as cholecystokinin inflamed with complete Freund's adjuvant, but absent from (CCK). Importantly chronic spinal infusion of NGF both non-inflamed tissue [39]. In this latter study, numerous lowers the heightened spinal CCK levels in neuropathic cells infiltrating the inflamed tissue stained positively for rats to normal levels [49], and restores opioid efficacy in b-endorphin and met-enkephalin. It was found that bneuropathic rats [6]. endorphin is present in T and B lymphocytes, monocytes In conclusion, intrathecal administration of bacterial and macrophages, and was concluded that opioid peptides endotoxin (LPS) appears to be an effective method to are synthesized and processed in various types of immune produce nociceptive behaviors associated with a central cells at the site of inflammation.
inflammatory response. We propose that spinal administra-It is now well established that the central nervous tion of LPS induces central inflammation, and produces a system is not an 'immune privileged' organ, but that model of persistent pain of inflammatory origin. By interactions exist between nervous and immune systems providing a protocol for a behavioral measurement that is [42]. Immune cells constantly survey the brain for foreign effective and reproducible, similar studies will likely aid material, and thus provide a potential source for opioid better understanding of central inflammatory pathological peptides. Several lines of evidence indicate that endogen-states. Consequently, this model may provide an underous opioid peptides mediate physiological consequences of standing of pain associated with inflammation within the LPS administration: (i) the levels of opioid peptides are central nervous system, a common process associated with higher in plasma of LPS-treated animals [7]; (ii) mononu-many neurodegenerative diseases, headaches and migraine. clear cells synthesize opioids upon exposure to LPS [3]; It has been well established that both centrally and and (iii) peripheral administration of LPS produces a peripherally administered opioids are more potent analsignificant antinociceptive response in phasic nociceptive gesics during states of peripheral inflammation. To our tests that was completely blocked by naltrexone [55].
knowledge, this is the first study to demonstrate the Taken together, i.t. LPS may increase the levels of enhanced effectiveness of morphine in a model of central immune-derived endogenous opioid peptides, such as inflammation.